Machine Learning-Based Prediction of Consistency and Histological Characteristics in Renal Cell Carcinoma Venous Tumor Thrombus Through Volumetric Radiomics

Machine Learning-Based Prediction of Consistency and Histological Characteristics in Renal Cell Carcinoma Venous Tumor Thrombus Through Volumetric Radiomics | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine Learning-Based Prediction of Consistency and Histological Characteristics in Renal Cell Carcinoma Venous Tumor Thrombus Through Volumetric Radiomics Paweł Kowal, Krzysztof Ratajczyk, Paulina Miernikiewicz, Wiktor Bursiewicz, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6254932/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background. Renal cell carcinoma (RCC) possesses a distinctive inclination to infiltrate the inferior vena cava, resulting in the formation of a venous tumour thrombus (VTT). Accurately assessing the consistency of the VTT prior to surgery is essential for effective treatment strategizing and favourable results. The study aimed to investigate the performance of volumetric radiomic MRI analysis in prediction of consistency and histomorphological vascular patterns of RCC venous tumour thrombus (VTT) with the assistance of machine learning. Methods. Twenty-four RCC patients with VTT underwent nephrectomy and thrombectomy in this study. Preoperatively abdominal DW-MRI was conducted, followed by the creation of 3D model of the thrombus. First-order radiomic features were computed from the complete thrombus volume utilizing ADC maps. The immunohistochemical staining of VTT was performed using CD34, SMA and VEGFR. The machine learning analysis was employed to develop predictive models for VTT histologic features. Patients were grouped based on the thrombus consistency into either solid or friable categories. Results. The solid and friable thrombi were detected in 13 (54.2%) and 11 (45.8%) cases, respectively. Large vessels were predominantly observed in solid VTTs (73.3%; p=0.015). Rich vascularization was a main pattern in solid VTT at 51.5%, contrasting with the friable at 9.1% (p=0.008). There was a strong association between thrombus vessel size and following radiomic features: entropy (r=0.722), skewness (r=0.635), and ADC mean (r=0.610). ADC entropy outperformed in distinguishing between VTT with large and small vessels, achieving the highest performance (AUC 0.930; p<0.001). In distinguishing between VTT with rich and poor vascularization, ADC median showed the best performance (AUC = 0.881; p < 0.001). Using machine learning analysis, we've developed two models predicting crucial histologic traits of VTT with prognostic accuracies of 89% for consistency and 75% for vessel size. Conclusions. Leveraging volumetric radiomic data from MR-DWI, along with machine learning models, we identified unique vascular patterns in VTTs among RCC patients. These models were developed to predict VTT consistency and vessel size using volumetric ADC data from DWI. renal cell carcinoma MRI thrombus consistency diffusion-weighted imaging volumetric immunohistochemistry machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 BACKGROUND Renal cell carcinoma (RCC) accounts for about 3% of all cancers, with 99,200 newly diagnosed cases and 39,100 kidney cancer-related deaths in the European Union [ 1 ]. Over the last few decades, there has been a 2% annual increase in the incidence of this disease globally [ 2 ]. RCC is notably prone to extending into the inferior vena cava (IVC), observed in 4–10% of all RCC cases [ 3 ]. Involvement of the IVC by the tumour is linked to advanced disease stages, posing a risk for recurrence and a poor prognosis [ 4 ]. Currently, aggressive surgical intervention is the gold standard for managing most patients with IVC thrombus, providing the only chance for long-term survival [ 5 , 6 ]. RCC extending into the IVC can complicate surgery significantly, so the operative approach to the tumour thrombus profoundly influences treatment outcomes. Thrombus volume is a crucial factor, among others, affecting the choice of surgical approach, complexity of the surgery, and postoperative complications [ 7 ]. Moreover, thrombi described as solid with regular surfaces are easier to manage than irregular, friable surfaces that are more likely to break into smaller pieces, leading to potentially fatal thromboembolic complications [ 8 ]. Studies have shown that the presence of friable thrombus in the IVC complicates surgery and is associated with worse survival outcomes in RCC patients. At the time, the existence of a friable thrombus markedly reduces median cancer-specific survival in comparison to a solid thrombus [ 9 – 12 ]. Accurately predicting thrombus consistency through presurgical imaging studies could help design surgical procedures more effectively and prevent pulmonary embolism during nephrectomy and thrombectomy [ 13 ]. However, distinguishing between tumour thrombus consistencies on preoperative imaging remains challenging, and friable thrombus can be misdiagnosed as bland [ 14 , 15 ]. Scientific data on this topic is scarce. Various imaging modalities, such as preoperative contrast-enhanced CT and intraoperative contrast-enhanced ultrasound, have been explored to differentiate between solid and friable IVC thrombus. However, none have been widely accepted [ 16 – 18 ]. Earlier studies demonstrated that qualitative analysis of diffusion-weighted imaging (DWI) and the apparent diffusion coefficient (ADC) of MRI accurately differentiated solid from bland thrombus in the portal vein of patients with hepatocellular carcinoma [ 19 , 20 ]. However, the diagnostic performance of MR-DWI in distinguishing bland from friable thrombus in the IVC of RCC patients has not been comprehensively evaluated. Research studies have indicated that elevated tumour expression of immunohistochemical markers like CD44 and HNF1B, along with specific histologic characteristics of the tumour, appear to be linked with the formation of tumour thrombus in the renal vein or vena cava [ 21 , 22 ]. Individuals diagnosed with RCC that extends into the inferior vena cava and exhibits a papillary subtype experience significantly shorter survival when compared to those with a clear cell subtype [ 23 ]. Furthermore, there was a significantly elevated expression of several immunohistochemical markers in the metastatic lesions in contrast to the primary and renal vein tumour thrombi. In locally advanced renal cell carcinoma, the relevance of VEGFR1 and VEGFD in univariate analysis underscores the significance of the hypoxia pathway in RCC pathogenesis. The variations in Ki67, p53, VEGFR1, SLUG, and SNAIL expressions between the primary tumour and metastases emphasize the roles of proliferation, angiogenesis, and epithelial-mesenchymal transition (EMT) in RCC pathogenesis as well [ 24 ]. Such data indicate the potentially important role of the immuno-histologic markers in different variants of tumour thrombi depending on their consistency in the natural history and prognosis of the disease. However, the utilization of immunohistochemical and radiomic markers for distinguishing between solid and fragile tumour thrombi in renal cell carcinoma and assessing different variants of their histological architecture, potentially significantly influencing treatment decisions and prognosis, remains largely unexplored. METHODS Aim. The study aimed to investigate the performance of volumetric radiomic MRI analysis in prediction of consistency and histomorphologicalvascularpatterns of RCC venous tumour thrombus (VTT) with the assistance of machine learning. Compliance with Ethical Standards This retrospective study was approved by the Local Bioethical Committee in the Research and Development Center, Regional Specialist Hospital in Wroclaw (no. KB/12/2021) and was conducted during 2022-2023 (based on clinical and radiologic data from 2011-2024). All procedures followed the ethical guidelines set by the institutional and national research committee, adhering to the principles outlined in the 1964 Helsinki Declaration and its subsequent revisions or equivalent ethical standards. All patients signed the written informed consent for enrolment in the study. All authors declare that they have no conflicts of interest. General Data This study showcases how incorporating volumetric radiomic analysis and machine learning techniques could improve the expertise of healthcare practitioners, such as urologists, radiologists, and oncologists, in evaluating the consistency and vascular pattern of VTT. This improvement is achieved through the utilization of algorithms developed in this research, as illustrated in Figure 1. The inclusion criteria were: patients with pathologically confirmed RCC with the spread of the tumour thrombus into the renal vein or IVC, surgically treated with nephrectomy and thrombectomy and in whom preoperatively abdominal MRI including the DWI sequence as an integral part was conducted. Based on the histological report, all patients were stratified into one of the two groups depending on the consistency of the VTT – solid or friable. Furthermore, the surgeon assessed the correlation between the surgical and pathological thrombus consistency during the nephrectomy procedure. Histological examination For histologic examination, the leading edge of the tumour thrombus was chosen. The tissue was fixed in 4% formaldehyde and embedded in paraffin using routine procedures, 4 μm thin sections were cut and stained with hematoxylin and eosin. Immunohistochemical (IHC) staining was performed on a VENTANA BenchMark XT automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA). IHC stains CD 34 and smooth muscle actin (SMA) were employed to visualize the vascular network better and determine the thickness of vascular walls. Vascular endothelial growth factor receptor (VEGFR) immunohistochemical stain was used to assess the degree of angiogenesis in VTT. Primary antibodies CD34 (Monoclonal Mouse Anti-Human, Class II, Clone QBEnd 10), SMA (Monoclonal Mouse Anti-Human, Clone 1A4) VEGFR (A-3):sc-6251 were performed. Positive and negative controls were also performed for each antibody. Slides were assessed using an Olympus BX 43 microscope with a camera Olympus SC 50. The main focus of the microscopic analysis was to evaluate the percentage of viable tumour cells present in each thrombus. Anatomic properties of the thrombus were assessed under low magnification (4X objective), focusing on the distribution of vessels, whereas the size and thickness of the vessel were additionally assessed under high magnification (40X objective). Quantification of predominant vessel type was assessed by conventional eyeballing by a single experienced pathologist. MR Imaging Technique MRI scans were conducted using a 1.5 T body scanner (Signa HDxt, General Electric, USA) equipped with an 8-channel phased-array body coil. The MRI protocol comprised several sequences with specific parameters: 1. Coronal T2-weighted single-shot fast spin-echo (SSFSE): TR=2625 ms, TE=90 ms, flip angle (FA)=90°, FOV=40×40 cm, matrix=200×192, breath-hold; 2. Axial 2D fast imaging employing steady-state acquisition with fat saturation (FIESTA FAT SAT): TR=4.1 ms, TE=1.8 ms, FA=90°, FOV=40×40 cm, matrix=224×320; 3. Sagittal T2-weighted SSFSE: TR=1760 ms, TE=87.4 ms, FA=90°, FOV=37×37 cm, matrix=384×256; 4. Axial T1-weighted fast spoiled gradient-recalled echo dual-echo (FSPGR-DE): TR=130 ms, TE=2.1 ms and 4.3 ms, FA=70°, FOV=43×43 cm, matrix=320×192, breath-hold; 5. Axial diffusion-weighted imaging (DWI): single-shot echo-planar, with parallel imaging and fat saturation during one breath-hold, prior to contrast media administration, TR=12000 ms, TE=90 ms, FOV=40×40 cm, matrix=200×192, NEX=3, bandwidth=250 kHz, diffusion direction=slice, slice thickness=6.0 mm, interscan gap=1.0 mm, b-values=50, 200, 800 s/mm²; 6. Axial 3D fat-saturated T1-weighted spoiled gradient echo liver acquisition with volume acquisition (LAVA): TR=4.5 ms, TE=2.2 ms, FA=15°, FOV=38×38 cm, matrix=320×192, administration of gadopentetate dimeglumine, in a dose of 0.1 mmol/kg of body weight as a bolus injection with 20 s between each breath-hold acquisition. MRI volumetric analysis The MRI data interpretation involved a qualitative analysis through visual assessment of T1-weighted (T1-WI), T2-weighted (T2-WI), and DWI, along with the corresponding apparent diffusion coefficient (ADC) map. A colour ADC map was generated using Functool 4.5 software on the Advantage Windows workstation from GE Healthcare. An experienced radiologist with 10 years of expertise in urogenital imaging evaluated the MR images. The 3D Slicer v.5.0.2 software was utilized to extract volumetric data and texture analysis. A region of interest (ROI) was meticulously placed over the thrombus area, encompassing the renal vein. This ROI was carefully traced on each slice of the ADC maps to ensure accuracy. This segmentation technique was employed to create a detailed 3D model of the thrombus (Figures 2 and 3). Within the domain of 3D texture analysis, the ADC map served as the foundation, enabling the calculation of radiomic first-order features throughout the entire volume of the thrombus. These features included ADC mean, median, range, 10th percentile, 90th percentile, interquartile range, entropy, kurtosis, skewness, uniformity, and variance. Special attention was given to excluding IVC wall from the ROI to maintain the precision of the analysis. Statistical Analysis Data processing was conducted using SPSS 22.0 software. Radiomic features were presented as mean ± standard deviation (SD). Categorical variables were compared using the chi-square test. The normality of the data was evaluated through the Kolmogorov-Smirnov and Shapiro-Wilk tests. Due to the non-normal distribution of the data, radiomic features in solid and friable thrombi cohorts were compared using the Mann-Whitney test. Pearson and Spearman methods were employed for correlation analysis. The diagnostic performance of radiomic features was assessed using receiver operating characteristics (ROC) analysis. Statistical significance was defined as a p-value < 0.05. Hierarchical clustering was employed to discern and illustrate the natural groupings of histologic features of VTTs within the dataset based on microscopic and IHC data using the hclust function in R (R Core Team, 2021). A cluster hierarchy based on data similarity was visualized with a dendrogram. Pairwise distances were computed using the Gower distance metric, and an agglomerative approach was employed, with each data point starting as a single cluster. Ward's method quantified cluster similarity. Cluster validity indices such as the silhouette coefficient were utilized to evaluate the robustness and relevance of the clusters. Machine Learning Analysis To evaluate the potential for predicting the IHC vascular pattern of VTT using radiomic features extracted from MRI data, we employed a machine learning approach comprising the following steps: Random Forest Analysis used to construct predictive models for determining the histologic characteristics of VTTs. This was achieved by leveraging volumetric radiomic MRI data and utilizing the randomForest package within the R programming environment (R Core Team, 2021). This method improves prediction accuracy and controls over-fitting by combining multiple decision trees. The steps involved were: tree generation, feature selection, node splitting, aggregation and model evaluation. A Confusion Matrix was generated, facilitating the determination of sensitivity, specificity, and the F1-score for the prediction model. Furthermore, a Variable Importance plot was produced, aiding in the identification of the most influential predictors for the VTT histologic features. Classification and Regression Tree (CART) analysis was utilized as a non-parametric decision tree learning technique to model and predict categorical outcomes from our data and to confirm the results of Random forest method. The visualization of trees was conducted using the scikit-learn library in Python, along with the dtreeviz library. This method simplifies the modelling of complex interactions and nonlinear relationships between variables. The decision tree's graphical representation also provided an intuitive visualization of the decision-making process. RESULTS Twenty-four cases were selected for analysis: 12 males and 12 females. The mean age of patients was 62.08±6.61 years (range, 47-74 years). In 14 (58,3%) and 10 (41.7%) cases, tumours involved the right and the left kidney, respectively. All patients were distributed following the 8 th edition of AJCC Cancer Staging Manual/TNM classification: 9 (37.5%) patients with T3a stage, 9 (37.5%) patients with T3b stage and 6 (25.0%) patients with T3c stage. The metastatic lymphatic nodules involvement and distant metastasis were observed in 3 (12.5%) and 4 (16.7%), respectively. All tumours were classified as clear cell renal cell carcinoma (ccRCC). The grade of the RCCs according to the WHO/International Society of Urological Pathology (ISUP) grading system was as follows: grade 2 – 4 (16.7%), grade 3 – 16 (66.7%) and grade 4 – 4 (16.7%) of cases. Histomorphological and immunohistochemical features of the VTTs Upon microscopic analysis, the proportion of viable tumour cells within each thrombus ranged from 0% (suggesting the absence of viable cells) to as high as 90% of the thrombus surface. Microscopically, the distribution of vessels was variable. Large vessels were either evenly distributed or more pronounced in the central part of the thrombus. Small vessels were relatively evenly distributed. Both vessel types exhibited an open lumen that was easily discernible or a collapsed state, visualized only through endothelial CD 34 immunostaining. The large vessel was defined as one overpassing that size (Figure 4A, 4B). A small vessel was defined as one fitting one high power field (1HPF) with a diameter of ≤0.055 mm (Figure 4C, 4D). The thick-walled vessels were defined as vessels showing multiple layers of SMA-positive cells, occupying more than ½ circumference of the vessel (Figure 4A, 4C). The examples of thin-walled vessels showing up to two linear SMA positive layers are presented in Figure 4B and 4D. The case of IVC thrombus with evenly distributed thick-walled small vessels with additional surrounding SMA-positive meshwork is presented in Figure 4E, 4F. The subsequent phase of investigation involved selecting the primary histologic characteristics of the thrombi for hierarchical cluster analysis. These included the percentage of viable tumour cells, vessel size, vessel wall thickness, presence of nested and trabecular patterns as per CD34 and SMA staining individually, the presence of a vascular-rich pattern on CD34 staining, and VEGF expression. The hierarchical cluster analysis revealed the identification of two primary clusters, delineating distinct histologic profiles among patients with VTT (Figure 5). Cluster 1 (С 1 ) exhibited a lower proportion of viable RCC cells within the thrombus, with a mean of 36.67% (1st quartile: 25%, 3rd quartile: 50%). Additionally, this cluster was characterized by a such as small vessel size, thin vascular wall, and a nested vascular pattern as indicated by SMA staining. On the other hand, Cluster 2 (С 2 ) showed a higher proportion of viable RCC cells within the thrombus, with a mean of 65.33% (1st quartile: 50%, 3rd quartile: 80%). This cluster was marked by characteristics including large vessel size, vascular richness and trabecular vascular patterns observed through CD34 staining. Furthermore, a trabecular vascular pattern noted in SMA staining and VEGF expression exceeding 75% were predominant features of this cluster. Based on the findings from hierarchical clustering analysis, VTT histologic consistency was classified as "solid" if it consisted of more than 50% viable tumour cells. According to mentioned above histologic criteria, the solid thrombus was detected in 13 (54.2%) cases, while the friable variant was evidenced in 11 (45.8%) patients. The mean percentage of tumour cells within VTT in groups with solid and friable variants was 70.0±12.91% and 36.36±16.29%, respectively (p<0.001). In all cases, the histologic subtype of RCC was clear-cell. There was no difference in age, sex, tumour side, T-stage and grade in groups of patients with solid and friable thrombus, Table 1. During the intraoperative palpation, a complete match was observed between the surgical and pathological consistency in every case (Cohen's kappa coefficient=1). Table 1. Patient characteristics and descriptive statistics Variable All patients, n=24 Solid VTT, n=13 Friable VTT, n=11 p -value Age, yr Mean (median) Range 62.08 (6.61) 47-74 63.08 (6.96) 47-74 60.91 (6.30) 49-70 0.713* Sex, no. (%) Male Female 12 (50) 12 (50) 6 (46.2) 7 (53.8) 6 (54.5) 5 (45.5) 0.682 § Tumour side, no. (%) Right Left 14 (58,3) 10 (41.7) 7 (53.8) 6 (46.2) 7 (63.6) 4 (36.4) 0.628 § T-stage, no. (%) T3a T3b T3c 9 (37.5) 9 (37.5) 6 (25.0) 5 (38.5) 5 (38.5) 3 (23.1) 4 (36.4) 4 (36.4) 3 (27.3) 0.972 § N-stage, no. (%) N0 N1 21 (87.5) 3 (12.5) 11 (84.6) 2 (15.4) 10 (90.9) 1 (9.1) 0.642 § M-stage, no. (%) M0 M1 20 (83.3) 4 (16.7) 11 (84.6) 2 (15.4) 9 (81.8) 2 (18.2) 0.855 § Tumour grade Grade 2 Grade 3 Grade 4 4 (16.7) 16 (66.7) 4 (16.7) 2 (15.4) 9 (69.2) 2 (15.4) 2 (18.2) 7 (63.6) 2 (18.2) 0.959 § VTT venous tumour thrombus, SD standard deviation, * - Student t test, § - x 2 test. In groups with different VTT consistency, there was a dissimilarity in predominant vessel size histomorphological pattern: the large vessels were more often observed in its solid variant (73.3%) than in friable (26.7%), p=0.015. However, there was no significant difference in the vascular wall thickness in groups with solid and friable thrombi (p=0.098), Table 2. Table 2. Characteristics of the venous tumour thrombus vessel size and wall thickness Variable Solid VTT, n=13 Friable VTT, n=11 p -value Vessell size, n (%) Small Large 2 (15.4) 11 (84.6) 7 (63.6) 4 (36.4) 0.015 Vascular wall thickness, n (%) Thin Thick 8 (61.5) 5 (38.5) 10 (90.9) 1 (9.1) 0.098 VTT venous tumour thrombus As a result of CD 34 immunostaining, there was no difference in the presence of the nested vascular histomorphological pattern in groups of patients with solid and friable thrombi (p=0.973). Likewise, there was no distinction in the occurrence of the trabecular vascular histomorphological VTT pattern among the mentioned patient groups (p=0.239). Rich vascularization emerged as the predominant pattern in solid VTT at 51.5%, contrasting with the friable consistency at 9.1% (p=0.008) and was mainly observed in conjunction with a nested vascular architecture. With SMA immunostaining, there was no discernible difference in the presence of nested (p=0.219) and trabecular (p=0.239) vascular architecture among patient groups with solid and friable thrombi. However, the unique attributes of the solid VTTs appeared to be microvascular hyperplasia (1 case; 7.69%), the meshwork of the vessels (3 cases; 23.08%) and abortive vessels (1 case; 7.69%). Additionally, one solid and one friable VTT case showed a peculiar pattern of dense SMA cell networks bridging small and large vessels. Immunostaining with VEGFR allowed the detection of a difference in the percentage of the VEGFR-positive cells in solid and friable thrombi: in most cases, solid VTTs demonstrated more pronounced positivity to VEGFR in comparison to friable, p=0.044, Table 3. Table 3. Characteristics of CD34, SMA and VEGFR immunostaining and vascular patterns of the venous tumour thrombi Variable Solid VTT, n=13 Friable VTT, n=11 p -value CD34 immunostaining vascular pattern/architecture Nested Trabecular Vascular rich 7 (53.8) 9 (69.2) 8 (61,5) 6 (54.5) 5 (45.5) 1 (9.1) 0.973 0.239 0.008 SMA immunostaining vascular pattern Nested Trabecular 8 (61.5) 9 (69.2) 4 (36.4) 5 (45.5) 0.219 0.239 VEGFR immunostaining >50-75% positive cells >75% positive cells 0 (0) 13 (100) 3 (27,3) 8 (72.7) 0.044 VTT venous tumour thrombus In the correlation analysis, there was a strong direct association between the percentage of the RCC cells within the VTT and the vessel size (Spearman r=0.613; p=0.001) and a moderate direct association between the percentage of the RCC cells in the VTT and the vascular wall thickness (Spearman r=0.427; p=0.038). Moreover, there was a moderate direct association between the vessel size and their wall thickness (Spearman r=0.447; p=0.028). There was a moderate association between the percentage of the RCC cells within the VTT and the vascular-rich CD 34 immunostaining pattern (Spearman r=0.475; p=0.019). Also, there were associations between the nested CD34 and SMA immunostaining vascular patterns in tumour thrombus (Spearman r=0.585; p=0.003). Moreover, there was a direct association between the percentage of VEGFR-positive cells and the VTT vessel size (Spearman r=0.488; p=0.016). Volumetric VTT analysis using MRI and histologic data The mean volume of the thrombi did not differ substantially in both groups and was 54.47±46,84 cm 3 (range, 12.7-157.03 cm 3 ) in solid vs 62.22±30.59 cm 3 (range, 23.56-112.33 cm 3 ) in friable (p=0.643). As a result of volumetric analysis, no significant difference in mean values of such radiomic features as range, 10th percentile, 90th percentile, interquartile range, kurtosis, uniformity and variance was found between groups (p>0.05). The mean ADC value of the whole volume of the thrombus was significantly higher in the group of patients with solid thrombi compared to friable. It amounted to 1457.16±253.48 vs 1199.81±99.23 mm 2 /sec, accordingly (p=0.004). The mean median value in solid compared to friable thrombi demonstrated a similar tendency and was 1459.73±264.77 vs 1221.82±123.76 mm 2 /sec, respectively (p=0.012). Likewise, the mean entropy value was significantly higher in the group with solid thrombi compared to friable: 5.71±0.45 vs 5.32±0.32 mm 2 /sec, accordingly (p=0.022). Also, there was a difference in mean skewness values between solid and friable thrombus groups: -0.26±0.48 vs 0.43±0.38 mm 2 /sec, respectively (p<0.001), Figure 6. When comparing MRI radiomic features in subgroups with large and small VTT vessels' histopathological patterns, there was a significant difference in ADC mean, mean entropy, and mean skewness. The ADC mean was higher in VTTs with large vessels compared to small: 1423.59±250.93 vs 1199.59±113.15 mm 2 /sec, respectively (p=0.003). Likewise, mean entropy was higher in VTTs with large vessels than in small vascular patterns: 5.74±0.42 vs 5.19±0.16 mm 2 /sec, respectively (p=0.001). The mean skewness value demonstrated an opposite tendency and was -0.20±0.48 in large vessels vs 0.48±0.40 mm 2 /sec in small vessels (p=0.002), Figure 7. The detailed statistical characteristics of the radiomic features in both subgroups of patients are presented in Table 4. Table 4. ADC-map radiomic features in RCC patients with VTT: large vs. small vessel groups' statistical characteristics Radiomic feature Large vessels (n=15) Small vessels (n=9) p -value Mean SD 95% CI Mean SD 95% CI Mean, mm 2 /sec 1423.59 250.93 1284.63-1562.55 1199.59 113.15 1112.61-1286.57 0.003 Median, mm 2 /sec 1421.83 263.16 1276.10-1567.57 1232.11 141.26 1123.53-1340.70 0.05 Range, mm 2 /sec 2559.73 335.07 2374.18-2745.29 2586.0 233.51 2406.51-2765.49 0.765 10th percentile, mm 2 /sec 1106.85 308.06 936.25-1277.45 850.89 189.95 704.88-996.9 0.05 90th percentile, mm 2 /sec 1665.01 412.19 1436.74-1893.27 1454.11 88.53 1386.06-1522.16 0.310 Interquartile range, mm 2 /sec 413.88 82.80 368.03-459.74 444.89 69.87 391.18-498.60 0.244 Entropy 5.74 0.42 5.51-5.97 5.19 0.16 5.06-5.31 0.001 Kurtosis 4.24 1.17 3.59-4.88 3.46 0.77 2.87-4.06 0.092 Skewness -0.20 0.48 -0.47-0.06 0.48 0.40 0.17-0. .78 0.002 Uniformity 0.02 0.01 0.02-0.03 0.02 0.01 0.02-0.03 0.918 Variance 110022.79 69089.28 71762.42-148283.16 87895.27 29453.69 65255.16-110535.38 0.454 SD standard deviation, CI confidence interval In an analysis of MRI radiomic features in subgroups with histomorphological patterns with thick and thin-walled vessels of VTT, we found a significant difference in ADC mean, mean median, mean 90th percentile, and mean skewness values. The ADC mean was higher in VTTs with thick-walled vessels compared to thin and amounted to 1592.50±319.60 vs 1255.29±119.76 mm 2 /sec, respectively (p=0.005). Furthermore, the mean median was also higher in VTTs with thick-walled vessels than in thin-walled vascular pattern: 1590.58±346.07 vs 1270.72±126.60 mm 2 /sec, respectively (p=0.02). The mean 90th percentile value demonstrated a similar propensity and was 2051.63±397.13 in thick-walled vessels vs 1430.68±99.36 mm 2 /sec in thin-walled vessels (p<0.001). Concurrently, mean skewness values demonstrated an inverted tendency, Figure 8. Table 5 displays the comprehensive statistical analysis of radiomic feature characteristics within both patient subgroups. Table 5. ADC-map radiomic features in RCC patients with VTT: thick vs. thin-walled vessel groups' detailed statistical characteristics Radiomic feature Thin-walled vessels (n=15) Thick-walled vessels (n=9) p -value Mean SD 95% CI Mean SD 95% CI Mean, mm 2 /sec 1255.29 119.76 1195.73-1314.85 1592.50 319.60 1257.10-1927.89 0.005 Median, mm 2 /sec 1270.72 126.60 1207.76-1333.68 1590.58 346.07 1227.41-1953.76 0.020 Range, mm 2 /sec 2567.33 265.31 2435.40-2699.27 2576.33 403.72 2152.66-3000.01 0.947 10th percentile, mm 2 /sec 956.59 284.90 814.92-1098.27 1173.67 280.68 879.11-1468.22 0.062 90th percentile, mm 2 /sec 1430.68 99.36 1381.27-1480.09 2051.63 397.13 1634.88-2468.39 <0.001 Interquartile range, mm 2 /sec 420.72 67.19 387.31-454.13 439.88 111.40 322.96-556.79 0.789 Entropy 5.45 0.45 5.23-5.68 5.77 0.29 5.46-6.08 0.05 Kurtosis 3.89 0.95 3.42-4.36 4.11 1.53 2.50-5.72 0.682 Skewness 0.19 0.44 -0.03-0.41 -0.36 0.70 -1.09-0.38 0.033 Uniformity 0.03 0.01 0.02-0.03 0.01 0.02-0.03 0.02-0.03 0.487 Variance 86984.52 24978.02 74563.25-99405.80 145946.30 100270.02 40719.37-251173.23 0.180 SD standard deviation, CI confidence interval There was no difference in mean radiomic values of all investigated first-order features in nested vs non-nested and trabecular vs non-trabecular CD34 immunostaining vascular patterns/architecture (p>0.05). Despite this, we found a difference in mean values of ADC, median and skewness in vascular-rich CD34 immunostaining vascular pattern compared to vascular poor, Figure 9. Although there was no dissimilarity in all studied radiomic features in trabecular or non-trabecular SMA immunostaining vascular patterns, there was a variation in mean uniformity values in subgroups with nested and non-nested variants (p=0.042), Table 6. Moreover, there was no difference in radiomic features in subgroups with different percentages of VEGFR-positive cells (p>0.05). Table 6. ADC-map radiomic features in patients with varying CD34 and SMA vascular patterns: detailed statistical characteristics. Radiomic feature Immunostaining vascular pattern p -value Mean SD Mean SD CD34 immunostaining Vascular rich Vascular poor ADC mean, mm 2 /sec 1496.38 296.47 1245.52 122.21 0.008 ADC median, mm 2 /sec 1529.94 286.42 1243.13 123.25 0.002 Skewness -0.24 0.61 0.23 0.45 0.044 SMA immunostaining Non-nested Nested Uniformity 0.027 0.004 0.023 0.004 0.042 SD standard deviation, CI confidence interval There was a strong association between thrombus vessel size and such radiomic features as entropy (Spearman r=0.722; p<0.001), skewness (Spearman r=0.635; p<0.001), and ADC mean (Spearman r=0.610; p=0.002). In addition, there was an association between thrombus vascular wall thickness and 90th percentile (Spearman r=0.751; p<0.001), ADC mean (Spearman r=0.584; p=0.003) and median (Spearman r=0.487; p=0.016). Moreover, there were associations detected between the presence of nested VTT CD34 immunostaining vascular pattern and ADC median (Spearman r=0.825; p<0.001) and entropy (Spearman r=0.705; p<0.001); between presence highly vascularized VTT CD34 immunostaining vascular pattern and ADC median (Spearman r=0.640; p<0.001); between nested SMA VTT immunostaining vascular pattern and uniformity (Spearman r=-0.418; p=0.042). Also, we observed a moderate association between the percentage of VEGFR-positive cells and such radiomic features of the thrombi as ADC mean (Spearman r=0.446; p=0.026). In the ROC analysis, the most effective performance in distinguishing between VTT with large and small vessels was showcased by ADC entropy, achieving an area under the curve (AUC) of 0.930; p<0.001. Simultaneously, the differentiation between venous thrombi with rich vascularization and those with poor vascularization exhibited optimal performance with the radiomic feature of ADC median (AUC = 0.881; p < 0.001), Table 7, Figure 10. Table 7. The diagnostic performance of mean radiomic features in differentiation between different VTT vascular patterns Radiomic feature The threshold value Sensitivity Specificity AUC P value Venous thrombus with large vs small vessels ADC mean , mm 2 /sec 1191.46 93.3 66.7 0.863 <0.001 Entropy 5.36 86.7 88.9 0.930 <0.001 Skewness 0.19 77.8 80.0 0.878 <0.001 Vascular rich vs vascular poor venous thrombus ADC mean, mm 2 /sec 1316.50 88.9 73.3 0.800 0.003 ADC median, mm 2 /sec 1348.50 88.9 80.0 0.881 <0.001 Skewness 0.09 60.0 66.7 0.700 0.072 AUC area under the curve Machine learning analysis of radiomic data Application of the Random Forest method facilitated the development of two robust predictive models capable of forecasting key immunohistochemical features of VTTs, specifically consistency (Model 1) and vessel size (Model 2). The Model 1 achieved a recall of 75% and precision of 100%, resulting in an F1-score of 86% for predicting solid VTT consistency. For friable consistency, it demonstrated a recall of 100%, precision of 80%, and an F1-score of 89%. Similarly, Model 2 exhibited a recall of 60% and precision of 100%, yielding an F1-score of 75% for predicting large VTT vascular pattern. Conversely, for small vascular pattern prediction, it attained a recall of 100%, precision of 60%, and an F1-score of 75%. The depiction of the classification Model 1 and Model 2's performance, illustrating their accuracy in classifying instances into their respective categories, is represented through Confusion Matrices in Figure 11. According to the variable importance plot, which was constructed to determine the predictors with the most influence on the target variable, it is evident that for Model 1, the most important variables are skewness and entropy. Meanwhile, for Model 2, the key predictors include entropy, mean, and skewness (Figure 12). Through validation utilizing the CART approach, the Random Forest method's results were confirmed, leading to the identification of the most critical radiomic predictors and the establishment of their respective cut-off values. Analysis of these trees revealed that skewness, interquartile range, mean, and the 10th percentile emerged as the primary predictors derived from the radiomic MRI data based on ADC image sequences for predicting VTT consistency. Additionally, for predicting VTT vessel size, entropy, variance, median, and skewness were identified as the most influential radiomic predictors. Figures 13 and 14 illustrate the visual representation of Regression Trees for Models 1 and 2, along with the corresponding cut-off values of radiomic predictors. DISCUSSION Locally advanced RCC with venous tumour thrombus presents a severe condition with a grim prognosis. Even if initial staging shows no signs of metastasis, approximately 29% of cases lead to cancer-specific death. Surgical resection remains the sole curative option for non-metastatic disease and is a component of multimodal treatment in metastatic cases, as targeted therapy alone is deemed less effective [25]. However, for most patients with RCC involving the IVC, survival prospects are limited. Numerous studies have highlighted the technical challenges in handling tumour thrombus due to its fragile nature. Thrombi are characterized as solid with smooth surfaces that are easier to handle than irregular, fragile surfaces prone to breaking into smaller pieces [8,13]. This fragility increases the risk of potentially fatal thromboembolic complications. Considering the high risk of intraoperative complications, tumour recurrence or progression post-surgery in RCC patients with IVC involvement, identifying prognostic markers to pinpoint individuals requiring specific surgery planning, further treatment, or closer monitoring is crucial, especially in light of recently published data. Recently, there has been an increasing emphasis on the prognostic importance of tumour thrombus consistency. In a retrospective study involving 174 patients, friable thrombus consistency emerged as an independent predictor of survival, correlating with significantly worse Cancer-Specific Survival (CSS) and overall survival (OS) outcomes. The thrombus was pathologically categorized as solid when a minimum of 90% of the samples exhibited solid characteristics, including compact and cohesive tumour growth, a rounded linear profile, and, occasionally, a partial endothelial lining resembling a pseudo capsule [26]. Another retrospective review of 200 patients validated a significantly shorter OS in individuals with friable thrombus consistency. Notably, thrombus consistency demonstrated predictive value in the subgroup of non-metastasized patients. However, the analysis did not establish the independent predictive significance of thrombus consistency in overall survival [27]. The morphological appearance of the VTT was categorized as either friable or solid, following the criteria by Bertini et al. [26]. In a cohort of 413 patients, thrombus consistency also does not appear to be independently linked to survival in patients with renal cell carcinoma. Nevertheless, this study did not assess the thrombus using pathological examination. Instead, its consistency was categorized simply as friable or solid, determined by the surgeon's intraoperative observation of pliable and slithery versus hard and barely compressible thrombotic tissue. This method makes it challenging, if only partially feasible, to compare this data with studies that employ objective morphological criteria of thrombus consistency [28]. Hence, the absence of a unified system for assessing thrombus consistency contributes to the conflicting results observed when analysing the survival outcomes in this patient category. In our study, the histologic consistency of the thrombus was morphologically classified as solid when it consisted of >50% of viable tumour cells, which differs from the criteria previously proposed by Bertini. The criteria proposed by our team for thrombus solidity are derived from the outcomes of hierarchical cluster analysis, which identified two distinct clusters with varying histologic patterns. Additionally, during an intraoperative palpation test, complete agreement was noted between the surgical and pathological consistency in all cases, thereby validating the proposed criteria. Previously, a robust correlation was observed between the formation of new blood and lymphatic vessels in RCC tissues and critical prognostic parameters, including clinical stage, pathological grade, and lymph node involvement [29]. To date, no studies profoundly examine vascular growth architecture besides the morphological assessment of thrombus consistency. Our research observed variations in the predominant vessel size histomorphological pattern among groups with different VTT consistency. Large vessels were more frequently observed in the solid variant than the friable one (p=0.015), while there was no significant difference in the vascular wall thickness between those groups. Additionally, a robust positive correlation was identified between the percentage of RCC cells within the VTT and vessel size. Although the expression of specific tissue immunohistochemical markers has been linked to thrombus formation in the inferior vena cava in RCC and prognosis [21–24], previous studies have not explored their role in distinguishing histological vascular patterns within solid and friable thrombi. The immunohistochemical biomarker CD34 is commonly utilized to identify vascular endothelial cells and hematopoietic stem and progenitor cells. It is broadly present in these cell types. Although initially identified as a marker for kidney glomerular epithelial cells and crucial for kidney development, abnormal CD34 expression has been linked to various malignancies [30]. For example, Kuroda and colleagues noted that myofibroblasts constituted the primary stromal components in invasive renal pelvic and ureteral cancers, with CD34-positive stromal cells consistently absent or lost in the stroma [31]. As per Yilmazer's findings, tumour vascular density, assessed through CD34 staining, was elevated in conventional RCC (P < 0.05). Additionally, this density showed a significant correlation with both the distribution and intensity of VEGF expression and the tumour stage (P < 0.05) [32]. In the recent study conducted by Kang et al., the endothelial cells of RCC were labelled with the CD34 for immunohistochemical analysis, assessment of microvessel density (MVD), and examination of the expression of carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) and endoglin (CD105). The findings revealed a significant association between CD105 expression levels and the clinical stages of RCC. Additionally, average MVD was linked to advanced histologic grades (P < 0.05) and clinical stages (P < 0.01), indicating that the down-regulation of CEACAM1 promotes angiogenesis in RCC, while the up-regulation of CD105 contributes to the progression of RCC [33]. In our study, immunostaining with CD34 revealed that, in contrast to the nested or trabecular vascular histomorphological pattern, a prominent feature of solid VTT was its rich vascularization, which was significantly more prevalent than in cases with friable consistency (p=0.008). Smooth muscle actin is a substance found in myofibroblasts, and it serves as a marker for epithelial to mesenchymal transition, which is recognized as a crucial process during which normal cells transform into cancerous, enabling them to metastasize [34]. The stromal cell populations in the tumour microenvironment play a crucial role in the progression and dissemination of RCC. Research has recognized the existence of SMA-positive cells (SMA+) in the stromal region of ccRCC tissues, and their abundance is strongly associated with unfavourable survival outcomes. Interleukin 6 emerges as a key driver, promoting the presence of SMA+ cells in ccRCC tissues by inducing EMT [35]. In the study conducted by Yu and colleagues, the expression of SMA was closely linked to adverse histologic features in RCC, indicating a poor prognosis, P = 0.005 [36]. Following SMA immunostaining, our study found no disparity in the presence of nested and trabecular vascular architecture within VTT in patients with solid and friable thrombi. Nevertheless, the solid VTTs exhibited distinctive characteristics, including microvascular hyperplasia, a meshwork of vessels, and abortive vessels. Angiogenesis plays a crucial role in the vascularization, growth, and metastasis of tumour tissue. In RCC, various growth factors influence different stages of angiogenesis. Among these factors, VEGF has received extensive attention due to its powerful angiogenic properties during both embryological and adult vasculogenesis and angiogenesis stages: a spectrum of studies has identified elevated levels of VEGF expression in both the tumour tissue and the blood and urine of patients with RCC. Moreover, recent research has revealed connections between VEGF expression in RCC and factors such as microvascular density, tumour size, nuclear grade, stage, and prognosis [37–39]. While tumour vascularization is not the exclusive factor affecting its size, lesions with dense vascularization tend to exhibit increased growth. In a study by Yilmazer et al., tumours with elevated and widespread VEGF expression were associated with larger sizes [32]. Patients with stage III and IV tumours exhibited more intense VEGF staining (P < 0.05). Moreover, the distribution of VEGF was elevated in advanced-stage tumours (P < 0.05). These results align with the discoveries made by Lee and Chang [40,41]. However, according to Raica et al., the immunohistochemical expression of VEGF does not align with MVD determined using slides stained for CD31 and endoglin. Most blood vessels within the tumour region exhibited a mature phenotype, characterized by perivascular cells testing positive for SMA [42]. We established a direct correlation between the percentage of VEGFR-positive cells and the size of VTT vessels in RCC patients (p=0.016). Diffusion-weighted imaging utilizing MRI remains the sole method for measuring water diffusion in vivo and holds the potential to offer non-invasive insights into the tumour microenvironment. DWI signal captures the diffusion of water molecules, with signal attenuation corrected by the b value, indicating the degree of diffusion weighting. This b value facilitates the calculation of the ADC value, enabling DWI to offer qualitative and quantitative data. Earlier research has demonstrated that the ADC can differentiate viable tumour areas from necrotic regions [43]. Apart from distinguishing viable tumour regions from necrotic areas, DWI has been employed to quantify tumour-associated neovascularization (angiogenesis), a crucial metric for assessing histologic structure and prognosis [44]. Hence, a notable focus has been on non-invasive and repeatable imaging techniques for evaluating malignant tumour angiogenesis. Hu and co-authors found a correlation of quantitative parameters MR perfusion-weighted imaging with VEGF, MVD and hypoxia-inducible factor-1α in nasopharyngeal carcinoma [45]. A recent discovery highlights a positive correlation between tumour vascularity and cellularity in ccRCC. Additionally, a confirmed negative correlation exists between tumour diffusion and cellularity [46]. Such results were supported by Lee et al. [47]. In a recent study, researchers effectively employed volumetric MRI histogram analysis to distinguish between RCC and oncocytoma [48]. Previously, it was shown that preoperative magnetic resonance venography improved the accuracy of distinguishing solid from bland RCC venous involvement by using FLASH-enhanced MR images (with a sensitivity of 89% and specificity of 96%). This outperformed signal intensity and precontrast FLASH images (with a sensitivity of 79% and specificity of 94%) according to McNemar's test (p < 0.05) [49]. In their study, Catalano et al. demonstrated the potential of ADC values in diffusion-weighted MR Imaging for distinguishing venous malignant thrombus from friable thrombus in patients with HCC. The mean ADC for solid and friable venous thrombus was 0.88 × 10−3 mm2/sec and 2.89 × 10−3 mm2/sec, respectively (P = 0.0003) [19]. No research has investigated the relationship between DWI and thrombus consistency and the histological vascular patterns of VTT in patients with RCC. In our study, we employed qualitative volumetric analysis of MR-DWI data, relying on ADC maps, to explore the histomorphological vascular patterns of solid and friable VTT in RCC patients. We conducted correlation analyses with the outcomes of immunohistochemical staining of the thrombus involving CD34, SMA, and VEGFR and radiomic first-order features. In groups with solid and friable VTT, we found a significant difference in mean values of radiomic features calculated from the whole volume of the thrombus as ADC value, median, entropy and skewness. When examining radiomic features from MRI scans within subgroups categorized by the size of VTT vessels and their histomorphological patterns, our analysis revealed that the ADC mean (p=0.003) and entropy (p=0.001) values were notably elevated in VTTs with larger vessels as compared to those with smaller vessels. The mean skewness value demonstrated an opposite tendency and was lower in thrombi with large vessels vs small (p=0.002). Furthermore, when scrutinizing MRI radiomic features within subgroups distinguished by the histomorphological patterns of VTT, explicitly focusing on thick and thin-walled vessels, we observed a noteworthy difference in ADC mean, mean median, mean 90th percentile, and mean skewness values. The correlation analysis demonstrated a robust association between thrombus vessel size and entropy, skewness, and ADC mean. Additionally, an association was found between thrombus vascular wall thickness and the 90th percentile, ADC mean, and median values. Furthermore, we identified correlations between the presence of a nested VTT CD34 immunostaining vascular pattern and ADC median and entropy. Similarly, associations were noted between a highly vascularized VTT CD34 immunostaining vascular pattern and ADC median and between the nested SMA VTT immunostaining vascular pattern and uniformity. Also, we noticed a moderate correlation between the percentage of VEGFR-positive cells and ADC mean. We believe that the differentiation of the histomorphological vascular patterns mentioned above, through volumetric analysis of ADC-maps, was achievable because of variations in the ratio of viable tumour cells with a constricted extracellular space and the presence of cell membranes resulting from aggressive and disorganized growth, VTT necrotic regions, and the presence of fibrin. The data we acquired, through a combination of immunohistochemistry and volumetric analysis of diffusion-weighted images, may reveal substantial distinctions in the pathogenesis and potentially the growth trajectory of tumour thrombi in renal cell cancer. Recent studies have demonstrated advancements in RCC diagnostics and prognostication through machine learning analysis utilizing radiomic and genomic data. These approaches have enabled the prediction of tumour grade, histologic subtypes, and prognosis [50–54]. However, there is a lack of research employing machine learning methods to predict VTT consistency in RCC patients. Through the application of machine learning analysis, we have successfully developed two models capable of predicting essential histologic characteristics of VTT. These models leverage volumetric MRI analysis data, specifically ADC derived from DWI, to predict VTT consistency and vessel size. The prognostic accuracy of these models stands at 89% for consistency and 75% for vessel size, respectively. Furthermore, this approach facilitated the identification of key predictors for VTT consistency, encompassing radiomic features such as skewness, interquartile range, mean, and the 10th percentile. Similarly, predictors for VTT vessel size, including entropy, variance, median, and skewness, were discerned. Additionally, we have established cut-off values for each significant predictor, enhancing the practical utility of the models in clinical settings. Our study has certain limitations , primarily arising from the absence of histologic subtypes of RCC beyond the conventional ones. Further comprehensive research is necessary to assess the survival rates among patients exhibiting these diverse vascular patterns. CONCLUSIONS In summary, our study identified unique vascular patterns in solid and fragile VTTs among RCC patients, utilizing volumetric radiomic data from MR-DWI. These distinctions may offer new insights into diagnostics, disease progression, and the development of patient-tailored treatment strategies for individuals with varying venous thrombus consistency. Additionally, two machine learning models were developed using volumetric ADC data from DWI to predict both VTT consistency and vessel size. Abbreviations ADC – apparent diffusion coefficient AUC – area under the curve CCRCC – clear cell renal cell carcinoma CI – confidence interval CSS – cancer-specific survival CT – computed tomography DWI – diffusion-weighted imaging IVC – inferior vena cava MRI – magnetic resonance imaging OS – overall survival RCC – renal cell carcinoma ROC – receiver operating characteristic ROI – region of interest SD – standard deviation SMA – smooth muscle actin VEGFR – vascular endothelial growth factor receptor VTT – venous tumour thrombus Declarations Ethics approval This study was approved by the Local Bioethical Committee in the Research and Development Center, Regional Specialist Hospital in Wroclaw (no. KB/12/2021) and was conducted during 2022-2024. All procedures followed the ethical guidelines set by the institutional and national research committee, adhering to the principles outlined in the 1964 Helsinki Declaration and its subsequent revisions or equivalent ethical standards. All patients signed the written informed consent for enrolment in the study. Consent to Participate declaration: not applicable Consent for publication: not applicable. Availability of data and materials The datasets used and analysed during the current study available from the corresponding author on reasonable request. Competing interests The authors have no relevant financial or nonfinancial interests to disclose. Funding This research was funded in whole by the National Science Centre, Poland, Grant number MINIATURA DEC-2022/06/X/NZ5/00677. For the purpose of Open Access, the author has applied a CC-BY public copyright license to any Author Accepted Manuscript (AAM) version arising from this submission. Authors' contributions PK conceived the conception and design of the study, performed acquisition of data, data analysis, final approval. KR performed acquisition of data. PM conceived the conception and design of the study. WB performed literature research, drafting the article. MT performed acquisition of data. KMB performed statistic data analysis. JR performed pathologic specimen analysis. YK performed acquisition of data, drafting the article. ML performed literature research. DS performed statistic data analysis. YM performed analysis and interpretation of data, paper preparation. All authors edited and approved the final version of the manuscript. Acknowledgements Not applicable. Footnotes Not applicable. References Capitanio U, Bensalah K, Bex A, Boorjian SA, Bray F, Coleman J, et al. Epidemiology of Renal Cell Carcinoma. Eur Urol. 2019;75:74–84. Ferlay J, Colombet M, Soerjomataram I, Dyba T, Randi G, Bettio M, et al. Cancer incidence and mortality patterns in Europe: Estimates for 40 countries and 25 major cancers in 2018. Eur J Cancer. 2018;103:356–87. Hevia V, Ciancio G, Gómez V, Álvarez S, Díez-Nicolás V, Burgos FJ. Surgical technique for the treatment of renal cell carcinoma with inferior vena cava tumor thrombus: tips, tricks and oncological results. SpringerPlus. 2016;5:132. Moinzadeh A, Libertino JA. Prognostic significance of tumor thrombus level in patients with renal cell carcinoma and venous tumor thrombus extension. Is all T3b the same? J Urol. 2004;171:598–601. Haferkamp A, Bastian PJ, Jakobi H, Pritsch M, Pfitzenmaier J, Albers P, et al. Renal cell carcinoma with tumor thrombus extension into the vena cava: prospective long-term followup. J Urol. 2007;177:1703–8. Ljungberg B, Albiges L, Abu-Ghanem Y, Bedke J, Capitanio U, Dabestani S, et al. European Association of Urology Guidelines on Renal Cell Carcinoma: The 2022 Update. Eur Urol. 2022;82:399–410. Baheen Q, Liu Z, Hao Y, Sawh RRR, Li Y, Zhao X, et al. The Significant Role of Tumor Volume on the Surgical Approach Choice, Surgical Complexity, and Postoperative Complications in Renal Cell Carcinoma With Venous Tumor Thrombus From a Large Chinese Center Experience. Front Oncol. 2022;12:869891. Ayyathurai R, Garcia-Roig M, Gorin MA, González J, Manoharan M, Kava BR, et al. Bland thrombus association with tumour thrombus in renal cell carcinoma: analysis of surgical significance and role of inferior vena caval interruption. BJU Int. 2012;110:E449-455. Hutchinson R, Rew C, Chen G, Woldu S, Krabbe L-M, Meissner M, et al. The Adverse Survival Implications of Bland Thrombus in Renal Cell Carcinoma With Venous Tumor Thrombus. Urology. 2018;115:119–24. Liu Z, Zhang L, Hong P, Li L, Tang S, Zhao X, et al. The influence of venous tumor thrombus combined with bland thrombus on the surgical treatment and prognosis of renal cell carcinoma patients. Cancer Med. 2020;9:5860–8. Liu * Zhuo, Zhang L, Li L, Zhang H, Tang S, Zhao X, et al. Mp57-06 bland thrombus as an independent risk factor for survival implications in renal cell carcinoma with venous tumor thrombus. Journal of Urology. 2020;203:e859–61. Blute ML, Leibovich BC, Lohse CM, Cheville JC, Zincke H. The Mayo Clinic experience with surgical management, complications and outcome for patients with renal cell carcinoma and venous tumour thrombus. BJU Int. 2004;94:33–41. Chen K, Liu Z, Li Y, Zhao X, Zhang S, Liu C, et al. Diagnosis and treatment strategies for intraoperative pulmonary embolism caused by renal tumor thrombus shedding. J Card Surg. 2022;37:3973–83. Rohatgi S, Howard SA, Tirumani SH, Ramaiya NH, Krajewski KM. Multimodality Imaging of Tumour Thrombus. Can Assoc Radiol J. 2015;66:121–9. Tabbara MM, Farag A, Ciancio G. Renal cell carcinoma with inferior vena cava tumor thrombus initially misdiagnosed as bland thrombus due to hypercoagulable state. SAGE Open Med Case Rep. 2022;10:2050313X221102019. Li Q-Y, Li N, Huang Q-B, Luo Y-K, Wang B-J, Guo A-T, et al. Contrast-enhanced ultrasound in detecting wall invasion and differentiating bland from tumor thrombus during robot-assisted inferior vena cava thrombectomy for renal cell carcinoma. Cancer Imaging. 2019;19:79. Zhang L, Yang G, Shen W, Qi J. Spectrum of the inferior vena cava: MDCT findings. Abdom Imaging. 2007;32:495–503. Wang K, Liu Z, Hong P, Qin Y-C, Zhao X, Zhang H-X, et al. T.H.R.O.B.V.S. Score - A Comprehensive Model to Predict the Surgical Complexity of Renal Cell Carcinoma With Tumor Thrombus. Front Oncol. 2022;12:900550. Catalano OA, Choy G, Zhu A, Hahn PF, Sahani DV. Differentiation of malignant thrombus from bland thrombus of the portal vein in patients with hepatocellular carcinoma: application of diffusion-weighted MR imaging. Radiology. 2010;254:154–62. Ahn J-H, Yu J-S, Cho E-S, Chung J-J, Kim JH, Kim KW. Diffusion-Weighted MRI of Malignant versus Benign Portal Vein Thrombosis. Korean J Radiol. 2016;17:533–40. Altinel M, Arpali E, Turhan N, Gonenc F, Yazicioglu A. Increased expression of CD44s in conventional renal cell carcinomas with renal vein or vena cava thrombosis. a retrospective evaluation of the expression of CD44s by immunohistochemical analysis in conventional RCC patients and in conventional RCC patients with renal vein or vena cava thrombosis. Urol Int. 2008;81:452–5. Cuff J, Salari K, Clarke N, Esheba GE, Forster AD, Huang S, et al. Integrative bioinformatics links HNF1B with clear cell carcinoma and tumor-associated thrombosis. PLoS One. 2013;8:e74562. Kondo T, Ikezawa E, Takagi T, Kobayashi H, Hashimoto Y, Iizuka J, et al. Negative impact of papillary histological subtype in patients with renal cell carcinoma extending into the inferior vena cava: single-center experience. Int J Urol. 2013;20:1072–7. Laird A, O’Mahony FC, Nanda J, Riddick ACP, O’Donnell M, Harrison DJ, et al. Differential expression of prognostic proteomic markers in primary tumour, venous tumour thrombus and metastatic renal cell cancer tissue and correlation with patient outcome. PLoS One. 2013;8:e60483. Heng DYC, Wells JC, Rini BI, Beuselinck B, Lee J-L, Knox JJ, et al. Cytoreductive nephrectomy in patients with synchronous metastases from renal cell carcinoma: results from the International Metastatic Renal Cell Carcinoma Database Consortium. Eur Urol. 2014;66:704–10. Bertini R, Roscigno M, Freschi M, Strada E, Angiolilli D, Petralia G, et al. Impact of venous tumour thrombus consistency (solid vs friable) on cancer-specific survival in patients with renal cell carcinoma. Eur Urol. 2011;60:358–65. Weiss VL, Braun M, Perner S, Harz A, Vorreuther R, Kristiansen G, et al. Prognostic significance of venous tumour thrombus consistency in patients with renal cell carcinoma (RCC). BJU Int. 2014;113:209–17. Mager R, Daneshmand S, Evans CP, Palou J, Martínez-Salamanca JI, Master VA, et al. Renal cell carcinoma with inferior vena cava involvement: Prognostic effect of tumor thrombus consistency on cancer specific survival. J Surg Oncol. 2016;114:764–8. Ohe C, Yoshida T, Amin MB, Atsumi N, Ikeda J, Saiga K, et al. Development and validation of a vascularity-based architectural classification for clear cell renal cell carcinoma: correlation with conventional pathological prognostic factors, gene expression patterns, and clinical outcomes. Modern Pathology. 2022;35:816–24. Nielsen JS, McNagny KM. Novel functions of the CD34 family. J Cell Sci. 2008;121:3683–92. Kuroda N, Shimasaki N, Miyazaki E, Hamauzu T, Toi M, Hiroi M, et al. The distribution of myofibroblasts and CD34-positive stromal cells in normal renal pelvis and ureter and their cancers. Histol Histopathol. 2006;21:1303–7. Yilmazer D, Han U, Onal B. A comparison of the vascular density of VEGF expression with microvascular density determined with CD34 and CD31 staining and conventional prognostic markers in renal cell carcinoma. Int Urol Nephrol. 2007;39:691–8. Kang Z-P, Wang L-X, Liu J, Ouyang H, Xie S, Tan Y. Expression of CEACAM1 and CD105 in Renal Cell Carcinoma and Its Correlation with Microvessel Density. Crit Rev Eukaryot Gene Expr. 2021;31:1–9. Anggorowati N, Ratna Kurniasari C, Damayanti K, Cahyanti T, Widodo I, Ghozali A, et al. Histochemical and Immunohistochemical Study of α-SMA, Collagen, and PCNA in Epithelial Ovarian Neoplasm. Asian Pac J Cancer Prev. 2017;18:667–71. Chen Q, Yang D, Zong H, Zhu L, Wang L, Wang X, et al. Growth-induced stress enhances epithelial-mesenchymal transition induced by IL-6 in clear cell renal cell carcinoma via the Akt/GSK-3β/β-catenin signaling pathway. Oncogenesis. 2017;6:e375. Yu W, Wang Y, Jiang Y, Zhang W, Li Y. Distinct immunophenotypes and prognostic factors in renal cell carcinoma with sarcomatoid differentiation: a systematic study of 19 immunohistochemical markers in 42 cases. BMC Cancer. 2017;17:293. Paradis V, Lagha NB, Zeimoura L, Blanchet P, Eschwege P, Ba N, et al. Expression of vascular endothelial growth factor in renal cell carcinomas. Virchows Arch. 2000;436:351–6. Choueiri TK, Kaelin WG. Targeting the HIF2-VEGF axis in renal cell carcinoma. Nat Med. 2020;26:1519–30. Situ Y, Xu Q, Deng L, Zhu Y, Gao R, Lei L, et al. System analysis of VEGFA in renal cell carcinoma: The expression, prognosis, gene regulation network and regulation targets. Int J Biol Markers. 2022;37:90–101. Lee JS, Kim HS, Jung JJ, Park CS, Lee MC. Expression of vascular endothelial growth factor in renal cell carcinoma and the relation to angiogenesis and p53 protein expression. J Surg Oncol. 2001;77:55–60. Chang SG, Jeon SH, Lee SJ, Choi JM, Kim YW. Clinical significance of urinary vascular endothelial growth factor and microvessel density in patients with renal cell carcinoma. Urology. 2001;58:904–8. Raica M, Cimpean AM, Anghel A. Immunohistochemical expression of vascular endothelial growth factor (VEGF) does not correlate with microvessel density in renal cell carcinoma. Neoplasma. 2007;54:278–84. Herneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol. 2003;45:208–13. Qian T, Chen M, Gao F, Meng F, Gao X, Yin H. Diffusion-weighted magnetic resonance imaging to evaluate microvascular density after transarterial embolization ablation in a rabbit VX2 liver tumor model. Magn Reson Imaging. 2014;32:1052–7. Hu Y, E H, Yu X, Li F, Zeng L, Lu Q, et al. Correlation of quantitative parameters of magnetic resonance perfusion-weighted imaging with vascular endothelial growth factor, microvessel density and hypoxia-inducible factor-1α in nasopharyngeal carcinoma: Evaluation on radiosensitivity study. Clin Otolaryngol. 2018;43:425–33. Yuan Q, Kapur P, Zhang Y, Xi Y, Carvo I, Signoretti S, et al. Intratumor Heterogeneity of Perfusion and Diffusion in Clear-Cell Renal Cell Carcinoma: Correlation With Tumor Cellularity. Clin Genitourin Cancer. 2016;14:e585–94. Lee H-J, Rha SY, Chung YE, Shim HS, Kim YJ, Hur J, et al. Tumor perfusion-related parameter of diffusion-weighted magnetic resonance imaging: correlation with histological microvessel density. Magn Reson Med. 2014;71:1554–8. Akinci O, Turkoglu F, Nalbant MO, Inci E. Differentiating renal cell carcinoma and oncocytoma with volumetric MRI histogram analysis. North Clin Istanb. 2023;10:636–41. Laissy JP, Menegazzo D, Debray MP, Toublanc M, Ravery V, Dumont E, et al. Renal carcinoma: diagnosis of venous invasion with Gd-enhanced MR venography. Eur Radiol. 2000;10:1138–43. Uhlig A, Uhlig J, Leha A, Biggemann L, Bachanek S, Stöckle M, et al. Radiomics and machine learning for renal tumor subtype assessment using multiphase computed tomography in a multicenter setting. Eur Radiol. 2024; Ji J, Liu Y, Bao Y, Men Y, Hui Z. Network analysis of histopathological image features and genomics data improving prognosis performance in clear cell renal cell carcinoma. Urol Oncol. 2024;S1078-1439(24)00400-9. Farias E, Terrematte P, Stransky B. Machine Learning Gene Signature to Metastatic ccRCC Based on ceRNA Network. Int J Mol Sci. 2024;25:4214. Buhas BA, Toma V, Beauval J-B, Andras I, Couți R, Muntean LA-M, et al. Label-Free SERS of Urine Components: A Powerful Tool for Discriminating Renal Cell Carcinoma through Multivariate Analysis and Machine Learning Techniques. Int J Mol Sci. 2024;25:3891. Alhussaini AJ, Steele JD, Jawli A, Nabi G. Radiomics Machine Learning Analysis of Clear Cell Renal Cell Carcinoma for Tumour Grade Prediction Based on Intra-Tumoural Sub-Region Heterogeneity. Cancers (Basel). 2024;16:1454. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6254932","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":452191849,"identity":"3a51b4af-64b6-47f9-9b07-cd5be3314705","order_by":0,"name":"Paweł Kowal","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Paweł","middleName":"","lastName":"Kowal","suffix":""},{"id":452191850,"identity":"b2ab8b3e-352b-49db-bd58-bccb04db5e44","order_by":1,"name":"Krzysztof Ratajczyk","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Krzysztof","middleName":"","lastName":"Ratajczyk","suffix":""},{"id":452191852,"identity":"7e26dd6a-57a4-4efd-8b58-b3e75a41139a","order_by":2,"name":"Paulina Miernikiewicz","email":"","orcid":"","institution":"Hirszfeld Institute of Immunology and Experimental Therapy Polish Academy of Sciences, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Paulina","middleName":"","lastName":"Miernikiewicz","suffix":""},{"id":452191854,"identity":"9cd20cb5-6943-4832-ad26-20ae5fdf3224","order_by":3,"name":"Wiktor Bursiewicz","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Wiktor","middleName":"","lastName":"Bursiewicz","suffix":""},{"id":452191856,"identity":"b262acdd-79d5-41aa-aff0-efe6c9782741","order_by":4,"name":"Maciej Trzciniecki","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Maciej","middleName":"","lastName":"Trzciniecki","suffix":""},{"id":452191857,"identity":"8d4457f1-88f9-4e5c-8026-335af2d90023","order_by":5,"name":"Karolina Marek-Bukowiec","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Karolina","middleName":"","lastName":"Marek-Bukowiec","suffix":""},{"id":452191858,"identity":"79b2b125-9f78-4b2a-aa0b-bece7943aa03","order_by":6,"name":"Joanna Rogala","email":"","orcid":"","institution":"Regional Specialist Hospital, Wroclaw","correspondingAuthor":false,"prefix":"","firstName":"Joanna","middleName":"","lastName":"Rogala","suffix":""},{"id":452191859,"identity":"f69c0b89-ac59-4e04-9739-efdf073f8c9f","order_by":7,"name":"Yuriy Kobilnyk","email":"","orcid":"","institution":"St. Padre Pio Regional Hospital in Przemysl","correspondingAuthor":false,"prefix":"","firstName":"Yuriy","middleName":"","lastName":"Kobilnyk","suffix":""},{"id":452191860,"identity":"71b01b98-5c2d-429a-8cc4-8e3e6f0bd648","order_by":8,"name":"Mateusz Lesny","email":"","orcid":"","institution":"St. Padre Pio Regional Hospital in Przemysl","correspondingAuthor":false,"prefix":"","firstName":"Mateusz","middleName":"","lastName":"Lesny","suffix":""},{"id":452191861,"identity":"6eb7ab61-a904-4702-95a7-553ea2f386f8","order_by":9,"name":"Dmytro Stroy","email":"","orcid":"","institution":"Bogomoletz Institute of Physiology of National Academy of Sciences of Ukraine","correspondingAuthor":false,"prefix":"","firstName":"Dmytro","middleName":"","lastName":"Stroy","suffix":""},{"id":452191862,"identity":"81e1660d-b7d5-49ff-88b0-128fd2f9b967","order_by":10,"name":"Yulian Mytsyk","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9UlEQVRIiWNgGAWjYNCCA0DMztjAwFABZDAzNxBQzgzVwgzScgbGIE4LEDO2gUQIaNGdkX/4M88ZhsQNh5kbPxfOq43mbwdq+VGxDacWsxvJbNI8N0BaGJulZ247njvjMGMDY8+Z23i1MPN8AGtpkObddiy3AchgZmzDq4X5M1RL82/eOcdy5xOhhQHmsDZp3oaa3A0EtZx5bCY554yE8UygFmueYwdyNwK1HMTrl+OJjz+8OWYj23e8/fFtnpq63HnnDx988KMCtxYGgQQGJh4GCccGCPcwmDyAWz0Q8B9gYPzBwGAP5dbhVTwKRsEoGAUjEwAAKB1gngu20LkAAAAASUVORK5CYII=","orcid":"","institution":"Voxel Medical Diagnostic Centers, Katowice","correspondingAuthor":true,"prefix":"","firstName":"Yulian","middleName":"","lastName":"Mytsyk","suffix":""}],"badges":[],"createdAt":"2025-03-18 16:08:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6254932/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6254932/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82176297,"identity":"7fb124d4-9a54-4b72-b807-b71763bc0bce","added_by":"auto","created_at":"2025-05-07 11:12:45","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":209490,"visible":true,"origin":"","legend":"\u003cp\u003eThe Figure illustrates the study's workflow\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/247a4970886c3c738146a435.jpg"},{"id":82179452,"identity":"8622562b-69c1-405a-ada1-7bb35ec594ff","added_by":"auto","created_at":"2025-05-07 11:36:45","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1138567,"visible":true,"origin":"","legend":"\u003cp\u003eA 3D model of the tumour thrombus reconstructed from the ADC-map segmentation\u003c/p\u003e\n\u003cp\u003eLegend to Figure 2: Abdominal MRI images of a 67-year-old patient diagnosed with conventional RCC in the right kidney and a tumour thrombus in the inferior vena cava (IVC), T3cN0M0, G3. (A) The ADC-map with a region of interest (ROI) contouring the tumour thrombus in the IVC; (B) a 3D model of the tumour thrombus reconstructed from the ADC-map segmentation, thrombus volume – 121 cm³; (C) and (D) T2-weighted images in coronal and axial views, respectively, along with the corresponding ROIs\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/903416a1f8c35fb72fb0a3f8.jpg"},{"id":82176322,"identity":"c5d251a1-c04a-426c-9b84-682e699800b6","added_by":"auto","created_at":"2025-05-07 11:12:45","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3514888,"visible":true,"origin":"","legend":"\u003cp\u003e3D model of the tumour thrombus generated from the ADC-map segmentation\u003c/p\u003e\n\u003cp\u003eLegend to Figure 3: Abdominal MRI images of a 68-year-old patient diagnosed with conventional RCC in the right kidney and a tumour thrombus in the inferior vena cava (IVC), T3bN0M0, G2. Thrombus demonstrated solid consistency, according to CD 34 immunostaining - rich vascularization, large vessels, according to SMA immunostaining - meshwork of the vessels. (A) Shows the ADC-map with a region of interest (ROI) delineating the tumour thrombus in the IVC; (B) a 3D model of the tumour thrombus generated from the ADC-map segmentation, with a thrombus volume of 113 cm³; (C) and (D) T2-weighted images in coronal and axial views, respectively, along with their corresponding ROIs\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/8b3459385d1bbab7946a7094.jpg"},{"id":82177426,"identity":"6b881e49-2fcb-4164-9530-b5028c9a35f2","added_by":"auto","created_at":"2025-05-07 11:20:45","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":15880165,"visible":true,"origin":"","legend":"\u003cp\u003eThe image illustrates the immunohistochemistry of the venous thrombi\u003c/p\u003e\n\u003cp\u003eLegend to Figure 4: (A) Objective 4x SMA immunohistochemistry highlighting thick-walled large vessels, evenly distributed; (B) Objective 4x SMA immunohistochemistry, large thin-walled vessel, centrally located; (C) Objective 4x, SMA immunohistochemistry, thick-walled small vessels, evenly distributed, vascular-rich; (D) Objective 4x, SMA immunohistochemistry, thin-walled small vessels, evenly distributed, vascular-rich; (E) the case of IVC thrombus with thick-walled small vessels and surrounding SMA-positive meshwork, objective 4x, SMA immunohistochemistry, thick-walled, small vessels, evenly distributed and additional surrounding SMA (+) meshwork; (F) the same patient, macrophotography of kidney and tumour thrombus specimen after nephrectomy and thrombectomy (venous thrombus is marked with the arrowhead)\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/70156a07d8dd19eb7cabfcd3.jpg"},{"id":82177417,"identity":"115bfd2c-dc41-47ad-9e9b-eda98f1464c1","added_by":"auto","created_at":"2025-05-07 11:20:45","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":280658,"visible":true,"origin":"","legend":"\u003cp\u003eThe heatmap generated through clustering analysis illustrates the variation in histologic profiles among patients with VTT, it delineates two distinct clusters, labelled as C\u003csub\u003e1\u003c/sub\u003e and C\u003csub\u003e2\u003c/sub\u003e, indicating clear groupings within the dataset\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/a2e9bcbab52baa3f71e50054.jpg"},{"id":82177412,"identity":"04c19124-b60d-4131-9b09-989e9338faa3","added_by":"auto","created_at":"2025-05-07 11:20:45","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":275115,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of radiomic features of the whole-volume solid and friable thrombus\u003c/p\u003e\n\u003cp\u003eLegend to Figure 6: (A): ADC mean and median of the whole-volume solid and friable thrombus; (B) entropy and skewness of the whole-volume solid and friable thrombus\u003c/p\u003e","description":"","filename":"6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/abef008a8eaf5985c74e069f.jpg"},{"id":82176308,"identity":"c93a0a2c-ae5f-487a-95ee-886c56a1b278","added_by":"auto","created_at":"2025-05-07 11:12:45","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":287159,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of radiomic features of the whole-volume thrombus with small and large vessels\u003c/p\u003e\n\u003cp\u003eLegend to Figure 7: (A): ADC mean of the whole-volume thrombus with small and large vessels; (B) entropy and skewness of the thrombus with small and large vessels\u003c/p\u003e","description":"","filename":"7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/4d0f4b4914b0aa53b4e8e67a.jpg"},{"id":82178829,"identity":"84c51d07-35c9-4c91-baf8-39b2de6a107b","added_by":"auto","created_at":"2025-05-07 11:28:45","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":295749,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of radiomic features of the whole-volume thrombus with thick and thin vessels\u003c/p\u003e\n\u003cp\u003eLegend to Figure 8: (A): ADC mean, median, and 90th percentile in thrombus with thick and thin vessels; (B) skewness of the whole-volume thrombus with thick and thin vessels\u003c/p\u003e","description":"","filename":"8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/6112e82e575375e06de4b529.jpg"},{"id":82176310,"identity":"4bf20d6c-d034-48ec-ad7f-8c65061f1519","added_by":"auto","created_at":"2025-05-07 11:12:45","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":291131,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of radiomic features in vascular reach and poor thrombus vascular patterns\u003c/p\u003e\n\u003cp\u003eLegend to Figure 9: (A): ADC mean and median in vascular reach and poor thrombus vascular patterns; (B) skewness in vascular reach and poor thrombus vascular patterns\u003c/p\u003e","description":"","filename":"9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/76564b2a409138490a9c297a.jpg"},{"id":82177422,"identity":"392d431e-46ac-46b0-9d8b-10485e31d9e9","added_by":"auto","created_at":"2025-05-07 11:20:45","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":206293,"visible":true,"origin":"","legend":"\u003cp\u003eROC-curves of radiomic features in differentiation between thrombus with different vascular patterns\u003c/p\u003e\n\u003cp\u003eLegend to Figure 10: (A): ADC mean and entropy in differentiation between thrombus with large vs small vessels; (B) ADC mean and median in differentiation between vascular rich vs vascular poor thrombus\u003c/p\u003e","description":"","filename":"10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/6c3b4bc37691151f072a5997.jpg"},{"id":82176318,"identity":"89008a13-06b5-4eb4-b628-94eb34bf8a46","added_by":"auto","created_at":"2025-05-07 11:12:45","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":171191,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion Matrix analysis for predicting consistency (A) and vessel size (B) of venous vascular thrombus\u003c/p\u003e","description":"","filename":"11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/11441a924d6014de8010409a.jpg"},{"id":82178832,"identity":"1c645aa4-da83-4215-9464-57b3f706b9af","added_by":"auto","created_at":"2025-05-07 11:28:45","extension":"jpg","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":378007,"visible":true,"origin":"","legend":"\u003cp\u003eThe variable importance plot highlights the most influential radiomic predictors for the consistency (A) and vessel size (B) of venous thrombus\u003c/p\u003e","description":"","filename":"12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/a9270b200f250d65d7a68c21.jpg"},{"id":82177420,"identity":"ed4404c3-cfd1-4fdd-92c7-e73ea20449e5","added_by":"auto","created_at":"2025-05-07 11:20:45","extension":"jpg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":197386,"visible":true,"origin":"","legend":"\u003cp\u003eClassification and Regression Tree for predicting VTT consistency using radiomic data\u003c/p\u003e","description":"","filename":"13.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/64139a89eb2c407082031711.jpg"},{"id":82178833,"identity":"9cd7aaff-bbce-41fb-ae2c-6b90d13efa69","added_by":"auto","created_at":"2025-05-07 11:28:46","extension":"jpg","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":197197,"visible":true,"origin":"","legend":"\u003cp\u003eClassification and Regression Tree for predicting VTT vessel size using radiomic data\u003c/p\u003e","description":"","filename":"14.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/278f195b5f1dd648848e0105.jpg"},{"id":101847298,"identity":"cbc9856e-7576-48e0-a23c-3994ead5a4ed","added_by":"auto","created_at":"2026-02-04 09:28:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":24557058,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6254932/v1/964c6b77-31d4-4d5a-a684-d2b4cdd63348.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning-Based Prediction of Consistency and Histological Characteristics in Renal Cell Carcinoma Venous Tumor Thrombus Through Volumetric Radiomics","fulltext":[{"header":"BACKGROUND","content":"\u003cp\u003eRenal cell carcinoma (RCC) accounts for about 3% of all cancers, with 99,200 newly diagnosed cases and 39,100 kidney cancer-related deaths in the European Union [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Over the last few decades, there has been a 2% annual increase in the incidence of this disease globally [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. RCC is notably prone to extending into the inferior vena cava (IVC), observed in 4\u0026ndash;10% of all RCC cases [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Involvement of the IVC by the tumour is linked to advanced disease stages, posing a risk for recurrence and a poor prognosis [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Currently, aggressive surgical intervention is the gold standard for managing most patients with IVC thrombus, providing the only chance for long-term survival [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. RCC extending into the IVC can complicate surgery significantly, so the operative approach to the tumour thrombus profoundly influences treatment outcomes. Thrombus volume is a crucial factor, among others, affecting the choice of surgical approach, complexity of the surgery, and postoperative complications [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Moreover, thrombi described as solid with regular surfaces are easier to manage than irregular, friable surfaces that are more likely to break into smaller pieces, leading to potentially fatal thromboembolic complications [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Studies have shown that the presence of friable thrombus in the IVC complicates surgery and is associated with worse survival outcomes in RCC patients. At the time, the existence of a friable thrombus markedly reduces median cancer-specific survival in comparison to a solid thrombus [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Accurately predicting thrombus consistency through presurgical imaging studies could help design surgical procedures more effectively and prevent pulmonary embolism during nephrectomy and thrombectomy [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, distinguishing between tumour thrombus consistencies on preoperative imaging remains challenging, and friable thrombus can be misdiagnosed as bland [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Scientific data on this topic is scarce. Various imaging modalities, such as preoperative contrast-enhanced CT and intraoperative contrast-enhanced ultrasound, have been explored to differentiate between solid and friable IVC thrombus. However, none have been widely accepted [\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Earlier studies demonstrated that qualitative analysis of diffusion-weighted imaging (DWI) and the apparent diffusion coefficient (ADC) of MRI accurately differentiated solid from bland thrombus in the portal vein of patients with hepatocellular carcinoma [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. However, the diagnostic performance of MR-DWI in distinguishing bland from friable thrombus in the IVC of RCC patients has not been comprehensively evaluated.\u003c/p\u003e \u003cp\u003eResearch studies have indicated that elevated tumour expression of immunohistochemical markers like CD44 and HNF1B, along with specific histologic characteristics of the tumour, appear to be linked with the formation of tumour thrombus in the renal vein or vena cava [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Individuals diagnosed with RCC that extends into the inferior vena cava and exhibits a papillary subtype experience significantly shorter survival when compared to those with a clear cell subtype [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Furthermore, there was a significantly elevated expression of several immunohistochemical markers in the metastatic lesions in contrast to the primary and renal vein tumour thrombi. In locally advanced renal cell carcinoma, the relevance of VEGFR1 and VEGFD in univariate analysis underscores the significance of the hypoxia pathway in RCC pathogenesis. The variations in Ki67, p53, VEGFR1, SLUG, and SNAIL expressions between the primary tumour and metastases emphasize the roles of proliferation, angiogenesis, and epithelial-mesenchymal transition (EMT) in RCC pathogenesis as well [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Such data indicate the potentially important role of the immuno-histologic markers in different variants of tumour thrombi depending on their consistency in the natural history and prognosis of the disease. However, the utilization of immunohistochemical and radiomic markers for distinguishing between solid and fragile tumour thrombi in renal cell carcinoma and assessing different variants of their histological architecture, potentially significantly influencing treatment decisions and prognosis, remains largely unexplored.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cp\u003e\u003cstrong\u003eAim.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study aimed to investigate the performance of volumetric radiomic MRI analysis in prediction of consistency and histomorphologicalvascularpatterns of RCC venous tumour thrombus (VTT) with the assistance of machine learning.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompliance with Ethical Standards\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study was approved by the Local Bioethical Committee in the Research and Development Center, Regional Specialist Hospital in Wroclaw (no. KB/12/2021) and was conducted during 2022-2023 (based on clinical and radiologic data from 2011-2024). All procedures followed the ethical guidelines set by the institutional and national research committee, adhering to the principles outlined in the 1964 Helsinki Declaration and its subsequent revisions or equivalent ethical standards. All patients signed the written informed consent for enrolment in the study. All authors declare that they have no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGeneral Data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study showcases how incorporating volumetric radiomic analysis and machine learning techniques could improve the expertise of healthcare practitioners, such as urologists, radiologists, and oncologists, in evaluating the consistency and vascular pattern of VTT. This improvement is achieved through the utilization of algorithms developed in this research, as illustrated in Figure 1. The inclusion criteria were: patients with pathologically confirmed RCC with the spread of the tumour thrombus into the renal vein or IVC, surgically treated with nephrectomy and thrombectomy and in whom preoperatively abdominal MRI including the DWI sequence as an integral part was conducted. Based on the histological report, all patients were stratified into one of the two groups depending on the consistency of the VTT \u0026ndash; solid or friable. Furthermore, the surgeon assessed the correlation between the surgical and pathological thrombus consistency during the nephrectomy procedure.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHistological examination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor histologic examination, the leading edge of the tumour thrombus was chosen. The tissue was fixed in 4% formaldehyde and embedded in paraffin using routine procedures, 4 \u0026mu;m thin sections were cut and stained with hematoxylin and eosin. Immunohistochemical (IHC) staining was performed on a VENTANA BenchMark XT automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA). IHC stains CD 34 and smooth muscle actin (SMA) were employed to visualize the vascular network better and determine the thickness of vascular walls. Vascular endothelial growth factor receptor (VEGFR) immunohistochemical stain was used to assess the degree of angiogenesis in VTT. Primary antibodies CD34 (Monoclonal Mouse Anti-Human, Class II, Clone QBEnd 10), SMA (Monoclonal Mouse Anti-Human, Clone 1A4) VEGFR (A-3):sc-6251 were performed. Positive and negative controls were also performed for each antibody. Slides were assessed using an Olympus BX 43 microscope with a camera Olympus SC 50. The main focus of the microscopic analysis was to evaluate the percentage of viable tumour cells present in each thrombus. Anatomic properties of the thrombus were assessed under low magnification (4X objective), focusing on the distribution of vessels, whereas the size and thickness of the vessel were additionally assessed under high magnification (40X objective). Quantification of predominant vessel type was assessed by conventional eyeballing by a single experienced pathologist.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMR Imaging Technique\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMRI scans were conducted using a 1.5 T body scanner (Signa HDxt, General Electric, USA) equipped with an 8-channel phased-array body coil. The MRI protocol comprised several sequences with specific parameters:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e1. Coronal T2-weighted single-shot fast spin-echo (SSFSE): TR=2625 ms, TE=90 ms, flip angle (FA)=90\u0026deg;, FOV=40\u0026times;40 cm, matrix=200\u0026times;192, breath-hold;\u003c/p\u003e\n\u003cp\u003e2. Axial 2D fast imaging employing steady-state acquisition with fat saturation (FIESTA FAT SAT): TR=4.1 ms, TE=1.8 ms, FA=90\u0026deg;, FOV=40\u0026times;40 cm, matrix=224\u0026times;320;\u003c/p\u003e\n\u003cp\u003e3. Sagittal T2-weighted SSFSE: TR=1760 ms, TE=87.4 ms, FA=90\u0026deg;, FOV=37\u0026times;37 cm, matrix=384\u0026times;256;\u003c/p\u003e\n\u003cp\u003e4. Axial T1-weighted fast spoiled gradient-recalled echo dual-echo (FSPGR-DE): TR=130 ms, TE=2.1 ms and 4.3 ms, FA=70\u0026deg;, FOV=43\u0026times;43 cm, matrix=320\u0026times;192, breath-hold;\u003c/p\u003e\n\u003cp\u003e5. Axial diffusion-weighted imaging (DWI): single-shot echo-planar, with parallel imaging and fat saturation during one breath-hold, prior to contrast media administration, TR=12000 ms, TE=90 ms, FOV=40\u0026times;40 cm, matrix=200\u0026times;192, NEX=3, bandwidth=250 kHz, diffusion direction=slice, slice thickness=6.0 mm, interscan gap=1.0 mm, b-values=50, 200, 800 s/mm\u0026sup2;;\u003c/p\u003e\n\u003cp\u003e6. Axial 3D fat-saturated T1-weighted spoiled gradient echo liver acquisition with volume acquisition (LAVA): TR=4.5 ms, TE=2.2 ms, FA=15\u0026deg;, FOV=38\u0026times;38 cm, matrix=320\u0026times;192, administration of gadopentetate dimeglumine, in a dose of 0.1 mmol/kg of body weight as a bolus injection with 20 s between each breath-hold acquisition.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMRI volumetric analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe MRI data interpretation involved a qualitative analysis through visual assessment of T1-weighted (T1-WI), T2-weighted (T2-WI), and DWI, along with the corresponding apparent diffusion coefficient (ADC) map. A colour ADC map was generated using Functool 4.5 software on the Advantage Windows workstation from GE Healthcare. An experienced radiologist with 10 years of expertise in urogenital imaging evaluated the MR images. The 3D Slicer v.5.0.2 software was utilized to extract volumetric data and texture analysis. A region of interest (ROI) was meticulously placed over the thrombus area, encompassing the renal vein. This ROI was carefully traced on each slice of the ADC maps to ensure accuracy. This segmentation technique was employed to create a detailed 3D model of the thrombus (Figures 2 and 3). Within the domain of 3D texture analysis, the ADC map served as the foundation, enabling the calculation of radiomic first-order features throughout the entire volume of the thrombus. These features included ADC mean, median, range, 10th percentile, 90th percentile, interquartile range, entropy, kurtosis, skewness, uniformity, and variance. Special attention was given to excluding IVC wall from the ROI to maintain the precision of the analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData processing was conducted using SPSS 22.0 software. Radiomic features were presented as mean \u0026plusmn; standard deviation (SD). Categorical variables were compared using the chi-square test. The normality of the data was evaluated through the Kolmogorov-Smirnov and Shapiro-Wilk tests. Due to the non-normal distribution of the data, radiomic features in solid and friable thrombi cohorts were compared using the Mann-Whitney test. Pearson and Spearman methods were employed for correlation analysis. The diagnostic performance of radiomic features was assessed using receiver operating characteristics (ROC) analysis. Statistical significance was defined as a p-value \u0026lt; 0.05. Hierarchical clustering was employed to discern and illustrate the natural groupings of histologic features of VTTs within the dataset based on microscopic and IHC data using the hclust function in R (R Core Team, 2021). A cluster hierarchy based on data similarity was visualized with a dendrogram. Pairwise distances were computed using the Gower distance metric, and an agglomerative approach was employed, with each data point starting as a single cluster. Ward\u0026apos;s method quantified cluster similarity. Cluster validity indices such as the silhouette coefficient were utilized to evaluate the robustness and relevance of the clusters.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMachine Learning Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the potential for predicting the IHC vascular pattern of VTT using radiomic features extracted from MRI data, we employed a machine learning approach comprising the following steps:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eRandom Forest Analysis used to construct predictive models for determining the histologic characteristics of VTTs. This was achieved by leveraging volumetric radiomic MRI data and utilizing the randomForest package within the R programming environment (R Core Team, 2021). This method improves prediction accuracy and controls over-fitting by combining multiple decision trees. The steps involved were: tree generation, feature selection, node splitting, aggregation and model evaluation.\u003c/li\u003e\n \u003cli\u003eA Confusion Matrix was generated, facilitating the determination of sensitivity, specificity, and the F1-score for the prediction model. Furthermore, a Variable Importance plot was produced, aiding in the identification of the most influential predictors for the VTT histologic features.\u003c/li\u003e\n \u003cli\u003eClassification and Regression Tree (CART) analysis was utilized as a non-parametric decision tree learning technique to model and predict categorical outcomes from our data and to confirm the results of Random forest method. The visualization of trees was conducted using the scikit-learn library in Python, along with the dtreeviz library. This method simplifies the modelling of complex interactions and nonlinear relationships between variables. The decision tree\u0026apos;s graphical representation also provided an intuitive visualization of the decision-making process.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"RESULTS ","content":"\u003cp\u003eTwenty-four cases were selected for analysis: 12 males and 12 females. The mean age of patients was 62.08\u0026plusmn;6.61 years (range, 47-74 years). In 14 (58,3%) and 10 (41.7%) cases, tumours involved the right and the left kidney, respectively. All patients were distributed following the 8\u003csup\u003eth\u003c/sup\u003e edition of AJCC Cancer Staging Manual/TNM classification: 9 (37.5%) patients with T3a stage, 9 (37.5%) patients with T3b stage and 6 (25.0%) patients with T3c stage. The metastatic lymphatic nodules involvement and distant metastasis were observed in 3 (12.5%) and 4 (16.7%), respectively. All tumours were classified as clear cell renal cell carcinoma (ccRCC). The grade of the RCCs according to the WHO/International Society of Urological Pathology (ISUP) grading system was as follows: grade 2 \u0026ndash; 4 (16.7%), grade 3 \u0026ndash; 16 (66.7%) and grade 4 \u0026ndash; 4 (16.7%) of cases.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHistomorphological and immunohistochemical features of the VTTs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUpon microscopic analysis, the proportion of viable tumour cells within each thrombus ranged from 0% (suggesting the absence of viable cells) to as high as 90% of the thrombus surface. Microscopically, the distribution of vessels was variable. Large vessels were either evenly distributed or more pronounced in the central part of the thrombus. Small vessels were relatively evenly distributed. Both vessel types exhibited an open lumen that was easily discernible or a collapsed state, visualized only through endothelial CD 34 immunostaining. The large vessel was defined as one overpassing that size (Figure 4A, 4B). A small vessel was defined as one fitting one high power field (1HPF) with a diameter of \u0026le;0.055 mm (Figure 4C, 4D). The thick-walled vessels were defined as vessels showing multiple layers of SMA-positive cells, occupying more than \u0026frac12; circumference of the vessel (Figure 4A, 4C). The examples of thin-walled vessels showing up to two linear SMA positive layers are presented in Figure 4B and 4D. The case of IVC thrombus with evenly distributed thick-walled small vessels with additional surrounding SMA-positive meshwork is presented in Figure 4E, 4F. The subsequent phase of investigation involved selecting the primary histologic characteristics of the thrombi for hierarchical cluster analysis. These included the percentage of viable tumour cells, vessel size, vessel wall thickness, presence of nested and trabecular patterns as per CD34 and SMA staining individually, the presence of a vascular-rich pattern on CD34 staining, and VEGF expression.\u003c/p\u003e\n\u003cp\u003eThe hierarchical cluster analysis revealed the identification of two primary clusters, delineating distinct histologic profiles among patients with VTT (Figure 5). Cluster 1 (С\u003csub\u003e1\u003c/sub\u003e) exhibited a lower proportion of viable RCC cells within the thrombus, with a mean of 36.67% (1st quartile: 25%, 3rd quartile: 50%). Additionally, this cluster was characterized by a such as small vessel size, thin vascular wall, and a nested vascular pattern as indicated by SMA staining. On the other hand, Cluster 2 (С\u003csub\u003e2\u003c/sub\u003e) showed a higher proportion of viable RCC cells within the thrombus, with a mean of 65.33% (1st quartile: 50%, 3rd quartile: 80%). This cluster was marked by characteristics including large vessel size, vascular richness and trabecular vascular patterns observed through CD34 staining. Furthermore, a trabecular vascular pattern noted in SMA staining and VEGF expression exceeding 75% were predominant features of this cluster. \u0026nbsp; Based on the findings from hierarchical clustering analysis, VTT histologic consistency was classified as \u0026quot;solid\u0026quot; if it consisted of more than 50% viable tumour cells.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAccording to mentioned above histologic criteria, the solid thrombus was detected in 13 (54.2%) cases, while the friable variant was evidenced in 11 (45.8%) patients. The mean percentage of tumour cells within VTT in groups with solid and friable variants was 70.0\u0026plusmn;12.91% and 36.36\u0026plusmn;16.29%, respectively (p\u0026lt;0.001). In all cases, the histologic subtype of RCC was clear-cell. There was no difference in age, sex, tumour side, T-stage and grade in groups of patients with solid and friable thrombus, Table 1. During the intraoperative palpation, a complete match was observed between the surgical and pathological consistency in every case (Cohen\u0026apos;s kappa coefficient=1).\u003c/p\u003e\n\u003cp\u003eTable 1. Patient characteristics and descriptive statistics\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll patients, n=24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSolid VTT, n=13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFriable VTT, n=11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eAge, yr\u003c/p\u003e\n \u003cp\u003eMean (median)\u003c/p\u003e\n \u003cp\u003eRange\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e62.08 (6.61)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e47-74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e63.08 (6.96)\u003c/p\u003e\n \u003cp\u003e47-74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e60.91 (6.30)\u003c/p\u003e\n \u003cp\u003e49-70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.713*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eSex, no. (%)\u003c/p\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e12 (50)\u003c/p\u003e\n \u003cp\u003e12 (50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e6 (46.2)\u003c/p\u003e\n \u003cp\u003e7 (53.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e6 (54.5)\u003c/p\u003e\n \u003cp\u003e5 (45.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.682\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eTumour side, no. (%)\u003c/p\u003e\n \u003cp\u003eRight\u003c/p\u003e\n \u003cp\u003eLeft\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e14 (58,3)\u003c/p\u003e\n \u003cp\u003e10 (41.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e7 (53.8)\u003c/p\u003e\n \u003cp\u003e6 (46.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e7 (63.6)\u003c/p\u003e\n \u003cp\u003e4 (36.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.628\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eT-stage, no. (%)\u003c/p\u003e\n \u003cp\u003eT3a\u003c/p\u003e\n \u003cp\u003eT3b\u003c/p\u003e\n \u003cp\u003eT3c\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e9 (37.5)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e9 (37.5)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e6 (25.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e5 (38.5)\u003c/p\u003e\n \u003cp\u003e5 (38.5)\u003c/p\u003e\n \u003cp\u003e3 (23.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e4 (36.4)\u003c/p\u003e\n \u003cp\u003e4 (36.4)\u003c/p\u003e\n \u003cp\u003e3 (27.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.972\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eN-stage, no. (%)\u003c/p\u003e\n \u003cp\u003eN0\u003c/p\u003e\n \u003cp\u003eN1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e21 (87.5)\u003c/p\u003e\n \u003cp\u003e3 (12.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e11 (84.6)\u003c/p\u003e\n \u003cp\u003e2 (15.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e10 (90.9)\u003c/p\u003e\n \u003cp\u003e1 (9.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.642\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eM-stage, no. (%)\u003c/p\u003e\n \u003cp\u003eM0\u003c/p\u003e\n \u003cp\u003eM1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e20 (83.3)\u003c/p\u003e\n \u003cp\u003e4 (16.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e11 (84.6)\u003c/p\u003e\n \u003cp\u003e2 (15.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e9 (81.8)\u003c/p\u003e\n \u003cp\u003e2 (18.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.855\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003e\n \u003cp\u003eTumour grade\u003c/p\u003e\n \u003cp\u003eGrade 2\u003c/p\u003e\n \u003cp\u003eGrade 3\u003c/p\u003e\n \u003cp\u003eGrade 4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e4 (16.7)\u003c/p\u003e\n \u003cp\u003e16 (66.7)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e4 (16.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e2 (15.4)\u003c/p\u003e\n \u003cp\u003e9 (69.2)\u003c/p\u003e\n \u003cp\u003e2 (15.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e2 (18.2)\u003c/p\u003e\n \u003cp\u003e7 (63.6)\u003c/p\u003e\n \u003cp\u003e2 (18.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.959\u003csup\u003e\u0026sect;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eVTT\u003c/em\u003e venous tumour thrombus, \u003cem\u003eSD\u003c/em\u003e standard deviation, * - Student t test, \u0026sect; - x\u003csup\u003e2\u003c/sup\u003e test.\u003c/p\u003e\n\u003cp\u003eIn groups with different VTT consistency, there was a dissimilarity in predominant vessel size histomorphological\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003epattern: the large vessels were more often observed in its solid variant (73.3%) than in friable (26.7%), p=0.015. However, there was no significant difference in the vascular wall thickness in groups with solid and friable thrombi (p=0.098), Table 2. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 2. Characteristics of the venous tumour thrombus vessel size and wall thickness\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 208px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 180px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSolid VTT, n=13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFriable VTT, n=11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 88px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 208px;\"\u003e\n \u003cp\u003eVessell size, n (%)\u003c/p\u003e\n \u003cp\u003eSmall\u003c/p\u003e\n \u003cp\u003eLarge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 180px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e2 (15.4)\u003c/p\u003e\n \u003cp\u003e11 (84.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e7 (63.6)\u003c/p\u003e\n \u003cp\u003e4 (36.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 88px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 208px;\"\u003e\n \u003cp\u003eVascular wall thickness, n (%)\u003c/p\u003e\n \u003cp\u003eThin\u003c/p\u003e\n \u003cp\u003eThick\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 180px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e8 (61.5)\u003c/p\u003e\n \u003cp\u003e5 (38.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e10 (90.9)\u003c/p\u003e\n \u003cp\u003e1 (9.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 88px;\"\u003e\n \u003cp\u003e0.098\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eVTT\u003c/em\u003e venous tumour thrombus\u003c/p\u003e\n\u003cp\u003eAs a result of CD 34 immunostaining, there was no difference in the presence of the nested vascular histomorphological\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003epattern in groups of patients with solid and friable thrombi (p=0.973). Likewise, there was no distinction in the occurrence of the trabecular vascular histomorphological\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eVTT\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003epattern among the mentioned patient groups (p=0.239). Rich vascularization emerged as the predominant pattern in solid VTT at 51.5%, contrasting with the friable consistency at 9.1% (p=0.008) and was mainly observed in conjunction with a nested vascular architecture. With SMA immunostaining, there was no discernible difference in the presence of nested (p=0.219) and trabecular (p=0.239) vascular architecture among patient groups with solid and friable thrombi. However, the unique attributes of the solid VTTs appeared to be microvascular hyperplasia (1 case; 7.69%), the meshwork of the vessels (3 cases; 23.08%) and abortive vessels (1 case; 7.69%). Additionally, one solid and one friable VTT case showed a peculiar pattern of dense SMA cell networks bridging small and large vessels. Immunostaining with VEGFR allowed the detection of a difference in the percentage of the VEGFR-positive cells in solid and friable thrombi: in most cases, solid VTTs demonstrated more pronounced positivity to VEGFR in comparison to friable, p=0.044, Table 3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 3. Characteristics of CD34, SMA and VEGFR immunostaining and vascular patterns of the venous tumour thrombi\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 312px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSolid\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eVTT, n=13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFriable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eVTT, n=11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 312px;\"\u003e\n \u003cp\u003eCD34 immunostaining vascular pattern/architecture\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eNested\u003c/p\u003e\n \u003cp\u003eTrabecular\u003c/p\u003e\n \u003cp\u003eVascular rich\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e7 (53.8)\u003c/p\u003e\n \u003cp\u003e9 (69.2)\u003c/p\u003e\n \u003cp\u003e8 (61,5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e6 (54.5)\u003c/p\u003e\n \u003cp\u003e5 (45.5)\u003c/p\u003e\n \u003cp\u003e1 (9.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.973\u003c/p\u003e\n \u003cp\u003e0.239\u003c/p\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 312px;\"\u003e\n \u003cp\u003eSMA immunostaining vascular pattern\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eNested\u003c/p\u003e\n \u003cp\u003eTrabecular\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e8 (61.5)\u003c/p\u003e\n \u003cp\u003e9 (69.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e4 (36.4)\u003c/p\u003e\n \u003cp\u003e5 (45.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.219\u003c/p\u003e\n \u003cp\u003e0.239\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 312px;\"\u003e\n \u003cp\u003eVEGFR immunostaining\u003c/p\u003e\n \u003cp\u003e\u0026gt;50-75% positive cells\u003c/p\u003e\n \u003cp\u003e\u0026gt;75% positive cells\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0 (0)\u003c/p\u003e\n \u003cp\u003e13 (100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e3 (27,3)\u003c/p\u003e\n \u003cp\u003e8 (72.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eVTT\u003c/em\u003e venous tumour thrombus\u003c/p\u003e\n\u003cp\u003eIn the correlation analysis, there was a strong direct association between the percentage of the RCC cells within the VTT and the vessel size (Spearman r=0.613; p=0.001) and a moderate direct association between the percentage of the RCC cells in the VTT and the vascular wall thickness (Spearman r=0.427; p=0.038). Moreover, there was a moderate direct association between the vessel size and their wall thickness (Spearman r=0.447; p=0.028). There was a moderate association between the percentage of the RCC cells within the VTT and the vascular-rich CD 34 immunostaining pattern (Spearman r=0.475; p=0.019). Also, there were associations between the nested CD34 and SMA immunostaining vascular patterns in tumour thrombus (Spearman r=0.585; p=0.003). Moreover, there was a direct association between the percentage of VEGFR-positive cells and the VTT vessel size (Spearman r=0.488; p=0.016).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVolumetric VTT analysis using MRI and histologic data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe mean volume of the thrombi did not differ substantially in both groups and was 54.47\u0026plusmn;46,84 cm\u003csup\u003e3\u003c/sup\u003e (range, 12.7-157.03 cm\u003csup\u003e3\u003c/sup\u003e) in solid vs 62.22\u0026plusmn;30.59 cm\u003csup\u003e3\u003c/sup\u003e (range, 23.56-112.33 cm\u003csup\u003e3\u003c/sup\u003e) in friable (p=0.643). As a result of volumetric analysis, no significant difference in mean values of such radiomic features as\u0026nbsp;range, 10th percentile, 90th percentile, interquartile range, kurtosis, uniformity and variance was found between groups (p\u0026gt;0.05).\u0026nbsp;The mean ADC value of the\u0026nbsp;whole volume of the thrombus was significantly higher\u0026nbsp;in the group of patients with\u0026nbsp;solid thrombi compared to\u0026nbsp;friable. It amounted to 1457.16\u0026plusmn;253.48 vs 1199.81\u0026plusmn;99.23 mm\u003csup\u003e2\u003c/sup\u003e/sec, accordingly (p=0.004). The mean\u0026nbsp;median\u0026nbsp;value in solid compared to friable\u0026nbsp;thrombi demonstrated a similar tendency and was 1459.73\u0026plusmn;264.77 vs 1221.82\u0026plusmn;123.76 mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p=0.012). Likewise, the mean entropy value was significantly higher in the group with solid thrombi compared to friable: 5.71\u0026plusmn;0.45 vs 5.32\u0026plusmn;0.32 mm\u003csup\u003e2\u003c/sup\u003e/sec, accordingly (p=0.022). Also, there was a difference in mean\u0026nbsp;skewness values between solid and\u0026nbsp;friable thrombus groups: -0.26\u0026plusmn;0.48 vs 0.43\u0026plusmn;0.38\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p\u0026lt;0.001), Figure 6.\u003c/p\u003e\n\u003cp\u003eWhen comparing MRI radiomic features in subgroups with large and small VTT vessels\u0026apos; histopathological\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003epatterns, there was a significant difference in ADC mean, mean entropy, and mean skewness. The ADC mean was higher in VTTs with large vessels compared to small: 1423.59\u0026plusmn;250.93 vs 1199.59\u0026plusmn;113.15 mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p=0.003). Likewise, mean entropy was higher in VTTs with large vessels than in small vascular patterns: 5.74\u0026plusmn;0.42 vs 5.19\u0026plusmn;0.16\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p=0.001). The mean skewness value demonstrated an opposite tendency and was -0.20\u0026plusmn;0.48 in large vessels vs 0.48\u0026plusmn;0.40 mm\u003csup\u003e2\u003c/sup\u003e/sec in small vessels (p=0.002), Figure 7. The detailed statistical characteristics of the radiomic features in both subgroups of patients are presented in Table 4.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 4. ADC-map radiomic features in RCC patients with VTT: large vs. small vessel groups\u0026apos; statistical characteristics\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"670\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRadiomic feature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 244px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLarge vessels (n=15)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 244px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSmall vessels (n=9)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eMean,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1423.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e250.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1284.63-1562.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1199.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e113.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1112.61-1286.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eMedian,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1421.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e263.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1276.10-1567.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1232.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e141.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1123.53-1340.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eRange,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2559.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e335.07\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2374.18-2745.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2586.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e233.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2406.51-2765.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.765\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e10th percentile,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1106.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e308.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e936.25-1277.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e850.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e189.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e704.88-996.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e90th percentile,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1665.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e412.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1436.74-1893.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1454.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e88.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1386.06-1522.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eInterquartile range,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e413.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e82.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e368.03-459.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e444.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e69.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e391.18-498.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.244\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eEntropy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.51-5.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.06-5.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eKurtosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e4.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e3.59-4.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e3.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2.87-4.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.092\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-0.47-0.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.17-0. .78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eUniformity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.918\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eVariance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e110022.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e69089.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e71762.42-148283.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e87895.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e29453.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e65255.16-110535.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.454\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eSD\u003c/em\u003e standard deviation, \u003cem\u003eCI\u003c/em\u003e confidence interval\u003c/p\u003e\n\u003cp\u003eIn an analysis of MRI radiomic features in subgroups with histomorphological patterns with thick and thin-walled vessels of VTT, we found a significant difference in ADC mean, mean median, mean 90th percentile, and\u0026nbsp;mean skewness values. The ADC mean was higher in VTTs with thick-walled vessels compared to thin and amounted to 1592.50\u0026plusmn;319.60 vs 1255.29\u0026plusmn;119.76 mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p=0.005). Furthermore, the mean median was also higher in VTTs with thick-walled vessels than in thin-walled vascular pattern: 1590.58\u0026plusmn;346.07 vs 1270.72\u0026plusmn;126.60\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec, respectively (p=0.02). The mean 90th percentile value demonstrated a similar propensity and was 2051.63\u0026plusmn;397.13 in thick-walled vessels vs 1430.68\u0026plusmn;99.36 mm\u003csup\u003e2\u003c/sup\u003e/sec in thin-walled vessels (p\u0026lt;0.001). Concurrently, mean skewness values demonstrated an inverted tendency, Figure 8. Table 5 displays the comprehensive statistical analysis of radiomic feature characteristics within both patient subgroups.\u003c/p\u003e\n\u003cp\u003eTable 5. ADC-map radiomic features in RCC patients with VTT: thick vs. thin-walled vessel groups\u0026apos; detailed statistical characteristics\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"670\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRadiomic feature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 244px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eThin-walled vessels (n=15)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 244px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eThick-walled vessels (n=9)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eMean,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1255.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e119.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1195.73-1314.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1592.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e319.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1257.10-1927.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eMedian,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1270.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e126.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1207.76-1333.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1590.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e346.07\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1227.41-1953.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.020\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eRange,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2567.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e265.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2435.40-2699.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2576.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e403.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2152.66-3000.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.947\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e10th percentile,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e956.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e284.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e814.92-1098.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1173.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e280.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e879.11-1468.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.062\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e90th percentile,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1430.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e99.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1381.27-1480.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2051.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e397.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1634.88-2468.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eInterquartile range,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e420.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e67.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e387.31-454.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e439.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e111.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e322.96-556.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.789\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eEntropy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.23-5.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e5.46-6.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eKurtosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e3.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e3.42-4.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e4.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e2.50-5.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.682\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-0.03-0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-1.09-0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eUniformity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.02-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.487\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eVariance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e86984.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e24978.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e74563.25-99405.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e145946.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e100270.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e40719.37-251173.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.180\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eSD\u003c/em\u003e standard deviation, \u003cem\u003eCI\u003c/em\u003e confidence interval\u003c/p\u003e\n\u003cp\u003eThere was no difference in mean radiomic values of all investigated first-order features in nested vs non-nested and trabecular vs non-trabecular CD34 immunostaining vascular patterns/architecture (p\u0026gt;0.05). Despite this, we found a difference in mean values of ADC, median and skewness in vascular-rich CD34 immunostaining vascular pattern compared to vascular poor, Figure 9. Although there was no dissimilarity in all studied radiomic features in trabecular or non-trabecular SMA immunostaining vascular patterns, there was a variation in mean uniformity values in subgroups with nested and non-nested variants (p=0.042), Table 6. Moreover, there was no difference in radiomic features in subgroups with different percentages of VEGFR-positive cells (p\u0026gt;0.05).\u003c/p\u003e\n\u003cp\u003eTable 6. ADC-map radiomic features in patients with varying CD34 and SMA vascular patterns: detailed statistical characteristics.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRadiomic feature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 397px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eImmunostaining vascular pattern\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 664px;\"\u003e\n \u003cp\u003eCD34 immunostaining\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eVascular\u0026nbsp;rich\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eVascular\u0026nbsp;poor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eADC mean,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e1496.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e296.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e1245.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e122.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eADC median,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e1529.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e286.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e1243.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e123.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e-0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 664px;\"\u003e\n \u003cp\u003eSMA immunostaining\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eNon-nested\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eNested\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\n \u003cp\u003eUniformity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e0.042\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eSD\u003c/em\u003e standard deviation, \u003cem\u003eCI\u003c/em\u003e confidence interval\u003c/p\u003e\n\u003cp\u003eThere was a strong association between thrombus vessel size and such radiomic features as entropy (Spearman r=0.722; p\u0026lt;0.001), skewness (Spearman r=0.635; p\u0026lt;0.001), and ADC mean (Spearman r=0.610; p=0.002). In addition, there was an association between thrombus vascular wall thickness and 90th percentile (Spearman r=0.751; p\u0026lt;0.001), ADC mean (Spearman r=0.584; p=0.003) and median (Spearman r=0.487; p=0.016). Moreover, there were associations detected between the presence of nested VTT CD34 immunostaining vascular pattern and ADC median (Spearman r=0.825; p\u0026lt;0.001) and entropy (Spearman r=0.705; p\u0026lt;0.001); between presence highly vascularized VTT CD34 immunostaining vascular pattern and ADC median (Spearman r=0.640; p\u0026lt;0.001); between nested SMA VTT immunostaining vascular pattern and uniformity (Spearman r=-0.418; p=0.042). Also, we observed a moderate association between the percentage of VEGFR-positive cells and such radiomic features of the thrombi as ADC mean (Spearman r=0.446; p=0.026).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the ROC analysis, the most effective performance in distinguishing between VTT with large and small vessels was showcased by ADC entropy, achieving an area under the curve (AUC) of 0.930; p\u0026lt;0.001. Simultaneously, the differentiation between venous thrombi with rich vascularization and those with poor vascularization exhibited optimal performance with the radiomic feature of ADC median (AUC = 0.881; p \u0026lt; 0.001), Table 7, Figure 10.\u003c/p\u003e\n\u003cp\u003eTable 7. The diagnostic performance of mean radiomic features\u0026nbsp;in differentiation between\u0026nbsp;different VTT vascular patterns\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"633\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRadiomic feature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eThe threshold value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 633px;\"\u003e\n \u003cp\u003eVenous thrombus with\u0026nbsp;large vs small vessels\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eADC mean\u003cstrong\u003e,\u0026nbsp;\u003c/strong\u003emm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e1191.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e93.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e66.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eEntropy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e5.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e86.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e88.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.930\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e77.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e80.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.878\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 633px;\"\u003e\n \u003cp\u003eVascular\u0026nbsp;rich vs vascular poor venous thrombus\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eADC mean,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e1316.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e88.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e73.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.800\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eADC median,\u0026nbsp;mm\u003csup\u003e2\u003c/sup\u003e/sec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e1348.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e88.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e80.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.881\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e60.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e66.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e0.700\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 75px;\"\u003e\n \u003cp\u003e0.072\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eAUC\u003c/em\u003e area under the curve\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMachine learning analysis of radiomic data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eApplication of the Random Forest method facilitated the development of two robust predictive models capable of forecasting key immunohistochemical features of VTTs, specifically consistency (Model 1) and vessel size (Model 2). The Model 1 achieved a recall of 75% and precision of 100%, resulting in an F1-score of 86% for predicting solid VTT consistency. For friable consistency, it demonstrated a recall of 100%, precision of 80%, and an F1-score of 89%. Similarly, Model 2 exhibited a recall of 60% and precision of 100%, yielding an F1-score of 75% for predicting large VTT vascular pattern. Conversely, for small vascular pattern prediction, it attained a recall of 100%, precision of 60%, and an F1-score of 75%. The depiction of the classification Model 1 and Model 2\u0026apos;s performance, illustrating their accuracy in classifying instances into their respective categories, is represented through Confusion Matrices in Figure 11.\u003c/p\u003e\n\u003cp\u003eAccording to the variable importance plot, which was constructed to determine the predictors with the most influence on the target variable, it is evident that for Model 1, the most important variables are skewness and entropy. Meanwhile, for Model 2, the key predictors include entropy, mean, and skewness (Figure 12).\u003c/p\u003e\n\u003cp\u003eThrough validation utilizing the CART approach, the Random Forest method\u0026apos;s results were confirmed, leading to the identification of the most critical radiomic predictors and the establishment of their respective cut-off values. Analysis of these trees revealed that skewness, interquartile range, mean, and the 10th percentile emerged as the primary predictors derived from the radiomic MRI data based on ADC image sequences for predicting VTT consistency. Additionally, for predicting VTT vessel size, entropy, variance, median, and skewness were identified as the most influential radiomic predictors. Figures 13 and 14 illustrate the visual representation of Regression Trees for Models 1 and 2, along with the corresponding cut-off values of radiomic predictors.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eLocally advanced RCC with venous tumour thrombus presents a severe condition with a grim prognosis. Even if initial staging shows no signs of metastasis, approximately 29% of cases lead to cancer-specific death. Surgical resection remains the sole curative option for non-metastatic disease and is a component of multimodal treatment in metastatic cases, as targeted therapy alone is deemed less effective [25]. However, for most patients with RCC involving the IVC, survival prospects are limited. Numerous studies have highlighted the technical challenges in handling tumour thrombus due to its fragile nature. Thrombi are characterized as solid with smooth surfaces that are easier to handle than irregular, fragile surfaces prone to breaking into smaller pieces [8,13]. This fragility increases the risk of potentially fatal thromboembolic complications. Considering the high risk of intraoperative complications, tumour recurrence or progression post-surgery in RCC patients with IVC involvement, identifying prognostic markers to pinpoint individuals requiring specific surgery planning, further treatment, or closer monitoring is crucial, especially in light of recently published data. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRecently, there has been an increasing emphasis on the prognostic importance of tumour thrombus consistency. In a retrospective study involving 174 patients, friable thrombus consistency emerged as an independent predictor of survival, correlating with significantly worse Cancer-Specific Survival (CSS) and overall survival (OS) outcomes. The thrombus was pathologically categorized as solid when a minimum of 90% of the samples exhibited solid characteristics, including compact and cohesive tumour growth, a rounded linear profile, and, occasionally, a partial endothelial lining resembling a pseudo capsule [26]. Another retrospective review of 200 patients validated a significantly shorter OS in individuals with friable thrombus consistency. Notably, thrombus consistency demonstrated predictive value in the subgroup of non-metastasized patients. However, the analysis did not establish the independent predictive significance of thrombus consistency in overall survival [27]. The morphological appearance of the VTT was categorized as either friable or solid, following the criteria by Bertini et al. [26]. In a cohort of 413 patients, thrombus consistency also does not appear to be independently linked to survival in patients with renal cell carcinoma. Nevertheless, this study did not assess the thrombus using pathological examination. Instead, its consistency was categorized simply as friable or solid, determined by the surgeon\u0026apos;s intraoperative observation of pliable and slithery versus hard and barely compressible thrombotic tissue. This method makes it challenging, if only partially feasible, to compare this data with studies that employ objective morphological criteria of thrombus consistency [28]. Hence, the absence of a unified system for assessing thrombus consistency contributes to the conflicting results observed when analysing the survival outcomes in this patient category. In our study, the histologic consistency of the thrombus was morphologically classified as solid when it consisted of \u0026gt;50% of viable tumour cells, which differs from the criteria previously proposed by Bertini. The criteria proposed by our team for thrombus solidity are derived from the outcomes of hierarchical cluster analysis, which identified two distinct clusters with varying histologic patterns. Additionally, during an intraoperative palpation test, complete agreement was noted between the surgical and pathological consistency in all cases, thereby validating the proposed criteria.\u003c/p\u003e\n\u003cp\u003ePreviously, a robust correlation was observed between the formation of new blood and lymphatic vessels in RCC tissues and critical prognostic parameters, including clinical stage, pathological grade, and lymph node involvement [29]. To date, no studies profoundly examine vascular growth architecture besides the morphological assessment of thrombus consistency. Our research observed variations in the predominant vessel size histomorphological pattern among groups with different VTT consistency. Large vessels were more frequently observed in the solid variant than the friable one (p=0.015), while there was no significant difference in the vascular wall thickness between those groups. Additionally, a robust positive correlation was identified between the percentage of RCC cells within the VTT and vessel size.\u003c/p\u003e\n\u003cp\u003eAlthough the expression of specific tissue immunohistochemical markers has been linked to thrombus formation in the inferior vena cava in RCC and prognosis [21\u0026ndash;24], previous studies have not explored their role in distinguishing histological vascular patterns within solid and friable thrombi. The immunohistochemical biomarker CD34 is commonly utilized to identify vascular endothelial cells and hematopoietic stem and progenitor cells. It is broadly present in these cell types. Although initially identified as a marker for kidney glomerular epithelial cells and crucial for kidney development, abnormal CD34 expression has been linked to various malignancies [30]. For example, Kuroda and colleagues noted that myofibroblasts constituted the primary stromal components in invasive renal pelvic and ureteral cancers, with CD34-positive stromal cells consistently absent or lost in the stroma [31]. As per Yilmazer\u0026apos;s findings, tumour vascular density, assessed through CD34 staining, was elevated in conventional RCC (P \u0026lt; 0.05). Additionally, this density showed a significant correlation with both the distribution and intensity of VEGF expression and the tumour stage (P \u0026lt; 0.05) [32]. In the recent study conducted by Kang et al., the endothelial cells of RCC were labelled with the CD34 for immunohistochemical analysis, assessment of microvessel density (MVD), and examination of the expression of carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) and endoglin (CD105). The findings revealed a significant association between CD105 expression levels and the clinical stages of RCC. Additionally, average MVD was linked to advanced histologic grades (P \u0026lt; 0.05) and clinical stages (P \u0026lt; 0.01), indicating that the down-regulation of CEACAM1 promotes angiogenesis in RCC, while the up-regulation of CD105 contributes to the progression of RCC [33]. In our study, immunostaining with CD34 revealed that, in contrast to the nested or trabecular vascular histomorphological pattern, a prominent feature of solid VTT was its rich vascularization, which was significantly more prevalent than in cases with friable consistency (p=0.008).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSmooth muscle actin is a substance found in myofibroblasts, and it serves as a marker for epithelial to mesenchymal transition, which is recognized as a crucial process during which normal cells transform into cancerous, enabling them to metastasize [34]. The stromal cell populations in the tumour microenvironment play a crucial role in the progression and dissemination of RCC. Research has recognized the existence of SMA-positive cells (SMA+) in the stromal region of ccRCC tissues, and their abundance is strongly associated with unfavourable survival outcomes. Interleukin 6 emerges as a key driver, promoting the presence of SMA+ cells in ccRCC tissues by inducing EMT [35]. In the study conducted by Yu and colleagues, the expression of SMA was closely linked to adverse histologic features in RCC, indicating a poor prognosis, P = 0.005 [36]. Following SMA immunostaining, our study found no disparity in the presence of nested and trabecular vascular architecture within VTT in patients with solid and friable thrombi. Nevertheless, the solid VTTs exhibited distinctive characteristics, including microvascular hyperplasia, a meshwork of vessels, and abortive vessels.\u003c/p\u003e\n\u003cp\u003eAngiogenesis plays a crucial role in the vascularization, growth, and metastasis of tumour tissue. In RCC, various growth factors influence different stages of angiogenesis. Among these factors, VEGF has received extensive attention due to its powerful angiogenic properties during both embryological and adult vasculogenesis and angiogenesis stages: a spectrum of studies has identified elevated levels of VEGF expression in both the tumour tissue and the blood and urine of patients with RCC. Moreover, recent research has revealed connections between VEGF expression in RCC and factors such as microvascular density, tumour size, nuclear grade, stage, and prognosis [37\u0026ndash;39]. While tumour vascularization is not the exclusive factor affecting its size, lesions with dense vascularization tend to exhibit increased growth. In a study by Yilmazer et al., tumours with elevated and widespread VEGF expression were associated with larger sizes [32]. Patients with stage III and IV tumours exhibited more intense VEGF staining (P \u0026lt; 0.05). Moreover, the distribution of VEGF was elevated in advanced-stage tumours (P \u0026lt; 0.05). These results align with the discoveries made by Lee and Chang [40,41]. However, according to Raica et al., the immunohistochemical expression of VEGF does not align with MVD determined using slides stained for CD31 and endoglin. Most blood vessels within the tumour region exhibited a mature phenotype, characterized by perivascular cells testing positive for SMA [42]. We established a direct correlation between the percentage of VEGFR-positive cells and the size of VTT vessels in RCC patients (p=0.016).\u003c/p\u003e\n\u003cp\u003eDiffusion-weighted imaging utilizing MRI remains the sole method for measuring water diffusion in vivo and holds the potential to offer non-invasive insights into the tumour microenvironment. DWI signal captures the diffusion of water molecules, with signal attenuation corrected by the b value, indicating the degree of diffusion weighting. This b value facilitates the calculation of the ADC value, enabling DWI to offer qualitative and quantitative data. Earlier research has demonstrated that the ADC can differentiate viable tumour areas from necrotic regions [43]. Apart from distinguishing viable tumour regions from necrotic areas, DWI has been employed to quantify tumour-associated neovascularization (angiogenesis), a crucial metric for assessing histologic structure and prognosis [44]. Hence, a notable focus has been on non-invasive and repeatable imaging techniques for evaluating malignant tumour angiogenesis. Hu and co-authors found a correlation of quantitative parameters MR perfusion-weighted imaging with VEGF, MVD and hypoxia-inducible factor-1\u0026alpha; in nasopharyngeal carcinoma [45]. A recent discovery highlights a positive correlation between tumour vascularity and cellularity in ccRCC. Additionally, a confirmed negative correlation exists between tumour diffusion and cellularity [46]. Such results were supported by Lee et al. [47]. In a recent study, researchers effectively employed volumetric MRI histogram analysis to distinguish between RCC and oncocytoma [48]. Previously, it was shown that preoperative magnetic resonance venography improved the accuracy of distinguishing solid from bland RCC venous involvement by using FLASH-enhanced MR images (with a sensitivity of 89% and specificity of 96%). This outperformed signal intensity and precontrast FLASH images (with a sensitivity of 79% and specificity of 94%) according to McNemar\u0026apos;s test (p \u0026lt; 0.05) [49]. In their study, Catalano et al. demonstrated the potential of ADC values in diffusion-weighted MR Imaging for distinguishing venous malignant thrombus from friable thrombus in patients with HCC. The mean ADC for solid and friable venous thrombus was 0.88 \u0026times; 10\u0026minus;3 mm2/sec and 2.89 \u0026times; 10\u0026minus;3 mm2/sec, respectively (P = 0.0003) [19]. No research has investigated the relationship between DWI and thrombus consistency and the histological vascular patterns of VTT in patients with RCC. In our study, we employed qualitative volumetric analysis of MR-DWI data, relying on ADC maps, to explore the histomorphological vascular patterns of solid and friable VTT in RCC patients. We conducted correlation analyses with the outcomes of immunohistochemical staining of the thrombus involving CD34, SMA, and VEGFR and\u0026nbsp;radiomic first-order features. In groups with\u0026nbsp;solid and\u0026nbsp;friable\u0026nbsp;VTT,\u0026nbsp;we found a significant difference in mean values of radiomic features calculated from the whole volume of the thrombus as\u0026nbsp;ADC value,\u0026nbsp;median,\u0026nbsp;entropy and\u0026nbsp;skewness.\u0026nbsp;When examining radiomic features from MRI scans within subgroups categorized by the size of VTT vessels and their histomorphological patterns, our analysis revealed that the ADC mean (p=0.003) and entropy (p=0.001) values were notably elevated in VTTs with larger vessels as compared to those with smaller vessels. The mean skewness value demonstrated an opposite tendency and was lower in thrombi with large vessels vs small (p=0.002).\u0026nbsp;Furthermore, when scrutinizing MRI radiomic features within subgroups distinguished by the histomorphological patterns of VTT, explicitly focusing on thick and thin-walled vessels, we observed a noteworthy difference in ADC mean, mean median, mean 90th percentile, and mean skewness values.\u0026nbsp;The correlation analysis demonstrated a robust association between thrombus vessel size and entropy, skewness, and ADC mean. Additionally, an association was found between thrombus vascular wall thickness and the 90th percentile, ADC mean, and median values. Furthermore, we identified correlations between the presence of a nested VTT CD34 immunostaining vascular pattern and ADC median and entropy. Similarly, associations were noted between a highly vascularized VTT CD34 immunostaining vascular pattern and ADC median and between the nested SMA VTT immunostaining vascular pattern and uniformity. Also, we noticed a moderate correlation between the percentage of VEGFR-positive cells and ADC mean. We believe that the differentiation of the histomorphological vascular patterns mentioned above, through volumetric analysis of ADC-maps, was achievable because of variations in the ratio of viable tumour cells with a constricted extracellular space and the presence of cell membranes resulting from aggressive and disorganized growth, VTT necrotic regions, and the presence of fibrin. The data we acquired, through a combination of immunohistochemistry and volumetric analysis of diffusion-weighted images, may reveal substantial distinctions in the pathogenesis and potentially the growth trajectory of tumour thrombi in renal cell cancer.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRecent studies have demonstrated advancements in RCC diagnostics and prognostication through machine learning analysis utilizing radiomic and genomic data. These approaches have enabled the prediction of tumour grade, histologic subtypes, and prognosis [50\u0026ndash;54]. However, there is a lack of research employing machine learning methods to predict VTT consistency in RCC patients. Through the application of machine learning analysis, we have successfully developed two models capable of predicting essential histologic characteristics of VTT. These models leverage volumetric MRI analysis data, specifically ADC derived from DWI, to predict VTT consistency and vessel size. The prognostic accuracy of these models stands at 89% for consistency and 75% for vessel size, respectively. \u0026nbsp;Furthermore, this approach facilitated the identification of key predictors for VTT consistency, encompassing radiomic features such as skewness, interquartile range, mean, and the 10th percentile. Similarly, predictors for VTT vessel size, including entropy, variance, median, and skewness, were discerned. Additionally, we have established cut-off values for each significant predictor, enhancing the practical utility of the models in clinical settings.\u003c/p\u003e\n\u003cp\u003eOur study has certain \u003cstrong\u003elimitations\u003c/strong\u003e, primarily arising from the absence of histologic subtypes of RCC beyond the conventional ones. Further comprehensive research is necessary to assess the survival rates among patients exhibiting these diverse vascular patterns.\u003c/p\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eIn summary, our study identified unique vascular patterns in solid and fragile VTTs among RCC patients, utilizing volumetric radiomic data from MR-DWI. These distinctions may offer new insights into diagnostics, disease progression, and the development of patient-tailored treatment strategies for individuals with varying venous thrombus consistency. Additionally, two machine learning models were developed using volumetric ADC data from DWI to predict both VTT consistency and vessel size.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eADC \u0026ndash; apparent diffusion coefficient\u003c/p\u003e\n\u003cp\u003eAUC \u0026ndash; area under the curve\u003c/p\u003e\n\u003cp\u003eCCRCC \u0026ndash; clear cell renal cell carcinoma\u003c/p\u003e\n\u003cp\u003eCI \u0026ndash; confidence interval\u003c/p\u003e\n\u003cp\u003eCSS \u0026ndash; cancer-specific survival\u003c/p\u003e\n\u003cp\u003eCT \u0026ndash; computed tomography\u003c/p\u003e\n\u003cp\u003eDWI \u0026ndash; diffusion-weighted imaging\u003c/p\u003e\n\u003cp\u003eIVC \u0026ndash; inferior vena cava\u003c/p\u003e\n\u003cp\u003eMRI \u0026ndash; magnetic resonance imaging\u003c/p\u003e\n\u003cp\u003eOS \u0026ndash; overall survival\u003c/p\u003e\n\u003cp\u003eRCC \u0026ndash; renal cell carcinoma\u003c/p\u003e\n\u003cp\u003eROC \u0026ndash; receiver operating characteristic\u003c/p\u003e\n\u003cp\u003eROI \u0026ndash; region of interest\u003c/p\u003e\n\u003cp\u003eSD \u0026ndash; standard deviation\u003c/p\u003e\n\u003cp\u003eSMA \u0026ndash; smooth muscle actin\u003c/p\u003e\n\u003cp\u003eVEGFR \u0026ndash; vascular endothelial growth factor receptor\u003c/p\u003e\n\u003cp\u003eVTT \u0026ndash; venous tumour thrombus\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Local Bioethical Committee in the Research and Development Center, Regional Specialist Hospital in Wroclaw (no. KB/12/2021) and was conducted during 2022-2024. All procedures followed the ethical guidelines set by the institutional and national research committee, adhering to the principles outlined in the 1964 Helsinki Declaration and its subsequent revisions or equivalent ethical standards. All patients signed the written informed consent for enrolment in the study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate declaration:\u003c/strong\u003e not applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u0026nbsp;\u003c/strong\u003enot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and analysed during the current study available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or nonfinancial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded in whole by the National Science Centre, Poland, Grant number MINIATURA DEC-2022/06/X/NZ5/00677. For the purpose of Open Access, the author has applied a CC-BY public copyright license to any Author Accepted Manuscript (AAM) version arising from this submission.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePK conceived the conception and design of the study, performed acquisition of data, data analysis, final approval. KR performed acquisition of data. PM conceived the conception and design of the study. WB performed literature research, drafting the article. MT performed acquisition of data. KMB performed statistic data analysis. JR performed pathologic specimen analysis. YK performed acquisition of data, drafting the article. ML performed literature research. \u0026nbsp;DS performed statistic data analysis. YM performed analysis and interpretation of data, paper preparation. All authors edited and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFootnotes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCapitanio U, Bensalah K, Bex A, Boorjian SA, Bray F, Coleman J, et al. Epidemiology of Renal Cell Carcinoma. Eur Urol. 2019;75:74\u0026ndash;84. \u003c/li\u003e\n\u003cli\u003eFerlay J, Colombet M, Soerjomataram I, Dyba T, Randi G, Bettio M, et al. Cancer incidence and mortality patterns in Europe: Estimates for 40 countries and 25 major cancers in 2018. Eur J Cancer. 2018;103:356\u0026ndash;87. \u003c/li\u003e\n\u003cli\u003eHevia V, Ciancio G, G\u0026oacute;mez V, \u0026Aacute;lvarez S, D\u0026iacute;ez-Nicol\u0026aacute;s V, Burgos FJ. Surgical technique for the treatment of renal cell carcinoma with inferior vena cava tumor thrombus: tips, tricks and oncological results. SpringerPlus. 2016;5:132. \u003c/li\u003e\n\u003cli\u003eMoinzadeh A, Libertino JA. Prognostic significance of tumor thrombus level in patients with renal cell carcinoma and venous tumor thrombus extension. Is all T3b the same? J Urol. 2004;171:598\u0026ndash;601. \u003c/li\u003e\n\u003cli\u003eHaferkamp A, Bastian PJ, Jakobi H, Pritsch M, Pfitzenmaier J, Albers P, et al. Renal cell carcinoma with tumor thrombus extension into the vena cava: prospective long-term followup. J Urol. 2007;177:1703\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eLjungberg B, Albiges L, Abu-Ghanem Y, Bedke J, Capitanio U, Dabestani S, et al. European Association of Urology Guidelines on Renal Cell Carcinoma: The 2022 Update. Eur Urol. 2022;82:399\u0026ndash;410. \u003c/li\u003e\n\u003cli\u003eBaheen Q, Liu Z, Hao Y, Sawh RRR, Li Y, Zhao X, et al. The Significant Role of Tumor Volume on the Surgical Approach Choice, Surgical Complexity, and Postoperative Complications in Renal Cell Carcinoma With Venous Tumor Thrombus From a Large Chinese Center Experience. Front Oncol. 2022;12:869891. \u003c/li\u003e\n\u003cli\u003eAyyathurai R, Garcia-Roig M, Gorin MA, Gonz\u0026aacute;lez J, Manoharan M, Kava BR, et al. Bland thrombus association with tumour thrombus in renal cell carcinoma: analysis of surgical significance and role of inferior vena caval interruption. BJU Int. 2012;110:E449-455. \u003c/li\u003e\n\u003cli\u003eHutchinson R, Rew C, Chen G, Woldu S, Krabbe L-M, Meissner M, et al. The Adverse Survival Implications of Bland Thrombus in Renal Cell Carcinoma With Venous Tumor Thrombus. Urology. 2018;115:119\u0026ndash;24. \u003c/li\u003e\n\u003cli\u003eLiu Z, Zhang L, Hong P, Li L, Tang S, Zhao X, et al. The influence of venous tumor thrombus combined with bland thrombus on the surgical treatment and prognosis of renal cell carcinoma patients. Cancer Med. 2020;9:5860\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eLiu * Zhuo, Zhang L, Li L, Zhang H, Tang S, Zhao X, et al. Mp57-06\u0026emsp;bland thrombus as an independent risk factor for survival implications in renal cell carcinoma with venous tumor thrombus. Journal of Urology. 2020;203:e859\u0026ndash;61. \u003c/li\u003e\n\u003cli\u003eBlute ML, Leibovich BC, Lohse CM, Cheville JC, Zincke H. The Mayo Clinic experience with surgical management, complications and outcome for patients with renal cell carcinoma and venous tumour thrombus. BJU Int. 2004;94:33\u0026ndash;41. \u003c/li\u003e\n\u003cli\u003eChen K, Liu Z, Li Y, Zhao X, Zhang S, Liu C, et al. Diagnosis and treatment strategies for intraoperative pulmonary embolism caused by renal tumor thrombus shedding. J Card Surg. 2022;37:3973\u0026ndash;83. \u003c/li\u003e\n\u003cli\u003eRohatgi S, Howard SA, Tirumani SH, Ramaiya NH, Krajewski KM. Multimodality Imaging of Tumour Thrombus. Can Assoc Radiol J. 2015;66:121\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eTabbara MM, Farag A, Ciancio G. Renal cell carcinoma with inferior vena cava tumor thrombus initially misdiagnosed as bland thrombus due to hypercoagulable state. SAGE Open Med Case Rep. 2022;10:2050313X221102019. \u003c/li\u003e\n\u003cli\u003eLi Q-Y, Li N, Huang Q-B, Luo Y-K, Wang B-J, Guo A-T, et al. Contrast-enhanced ultrasound in detecting wall invasion and differentiating bland from tumor thrombus during robot-assisted inferior vena cava thrombectomy for renal cell carcinoma. Cancer Imaging. 2019;19:79. \u003c/li\u003e\n\u003cli\u003eZhang L, Yang G, Shen W, Qi J. Spectrum of the inferior vena cava: MDCT findings. Abdom Imaging. 2007;32:495\u0026ndash;503. \u003c/li\u003e\n\u003cli\u003eWang K, Liu Z, Hong P, Qin Y-C, Zhao X, Zhang H-X, et al. T.H.R.O.B.V.S. Score - A Comprehensive Model to Predict the Surgical Complexity of Renal Cell Carcinoma With Tumor Thrombus. Front Oncol. 2022;12:900550. \u003c/li\u003e\n\u003cli\u003eCatalano OA, Choy G, Zhu A, Hahn PF, Sahani DV. Differentiation of malignant thrombus from bland thrombus of the portal vein in patients with hepatocellular carcinoma: application of diffusion-weighted MR imaging. Radiology. 2010;254:154\u0026ndash;62. \u003c/li\u003e\n\u003cli\u003eAhn J-H, Yu J-S, Cho E-S, Chung J-J, Kim JH, Kim KW. Diffusion-Weighted MRI of Malignant versus Benign Portal Vein Thrombosis. Korean J Radiol. 2016;17:533\u0026ndash;40. \u003c/li\u003e\n\u003cli\u003eAltinel M, Arpali E, Turhan N, Gonenc F, Yazicioglu A. Increased expression of CD44s in conventional renal cell carcinomas with renal vein or vena cava thrombosis. a retrospective evaluation of the expression of CD44s by immunohistochemical analysis in conventional RCC patients and in conventional RCC patients with renal vein or vena cava thrombosis. Urol Int. 2008;81:452\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eCuff J, Salari K, Clarke N, Esheba GE, Forster AD, Huang S, et al. Integrative bioinformatics links HNF1B with clear cell carcinoma and tumor-associated thrombosis. PLoS One. 2013;8:e74562. \u003c/li\u003e\n\u003cli\u003eKondo T, Ikezawa E, Takagi T, Kobayashi H, Hashimoto Y, Iizuka J, et al. Negative impact of papillary histological subtype in patients with renal cell carcinoma extending into the inferior vena cava: single-center experience. Int J Urol. 2013;20:1072\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eLaird A, O\u0026rsquo;Mahony FC, Nanda J, Riddick ACP, O\u0026rsquo;Donnell M, Harrison DJ, et al. Differential expression of prognostic proteomic markers in primary tumour, venous tumour thrombus and metastatic renal cell cancer tissue and correlation with patient outcome. PLoS One. 2013;8:e60483. \u003c/li\u003e\n\u003cli\u003eHeng DYC, Wells JC, Rini BI, Beuselinck B, Lee J-L, Knox JJ, et al. Cytoreductive nephrectomy in patients with synchronous metastases from renal cell carcinoma: results from the International Metastatic Renal Cell Carcinoma Database Consortium. Eur Urol. 2014;66:704\u0026ndash;10. \u003c/li\u003e\n\u003cli\u003eBertini R, Roscigno M, Freschi M, Strada E, Angiolilli D, Petralia G, et al. Impact of venous tumour thrombus consistency (solid vs friable) on cancer-specific survival in patients with renal cell carcinoma. Eur Urol. 2011;60:358\u0026ndash;65. \u003c/li\u003e\n\u003cli\u003eWeiss VL, Braun M, Perner S, Harz A, Vorreuther R, Kristiansen G, et al. Prognostic significance of venous tumour thrombus consistency in patients with renal cell carcinoma (RCC). BJU Int. 2014;113:209\u0026ndash;17. \u003c/li\u003e\n\u003cli\u003eMager R, Daneshmand S, Evans CP, Palou J, Mart\u0026iacute;nez-Salamanca JI, Master VA, et al. Renal cell carcinoma with inferior vena cava involvement: Prognostic effect of tumor thrombus consistency on cancer specific survival. J Surg Oncol. 2016;114:764\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eOhe C, Yoshida T, Amin MB, Atsumi N, Ikeda J, Saiga K, et al. Development and validation of a vascularity-based architectural classification for clear cell renal cell carcinoma: correlation with conventional pathological prognostic factors, gene expression patterns, and clinical outcomes. Modern Pathology. 2022;35:816\u0026ndash;24. \u003c/li\u003e\n\u003cli\u003eNielsen JS, McNagny KM. Novel functions of the CD34 family. J Cell Sci. 2008;121:3683\u0026ndash;92. \u003c/li\u003e\n\u003cli\u003eKuroda N, Shimasaki N, Miyazaki E, Hamauzu T, Toi M, Hiroi M, et al. The distribution of myofibroblasts and CD34-positive stromal cells in normal renal pelvis and ureter and their cancers. Histol Histopathol. 2006;21:1303\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eYilmazer D, Han U, Onal B. A comparison of the vascular density of VEGF expression with microvascular density determined with CD34 and CD31 staining and conventional prognostic markers in renal cell carcinoma. Int Urol Nephrol. 2007;39:691\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eKang Z-P, Wang L-X, Liu J, Ouyang H, Xie S, Tan Y. Expression of CEACAM1 and CD105 in Renal Cell Carcinoma and Its Correlation with Microvessel Density. Crit Rev Eukaryot Gene Expr. 2021;31:1\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eAnggorowati N, Ratna Kurniasari C, Damayanti K, Cahyanti T, Widodo I, Ghozali A, et al. Histochemical and Immunohistochemical Study of \u0026alpha;-SMA, Collagen, and PCNA in Epithelial Ovarian Neoplasm. Asian Pac J Cancer Prev. 2017;18:667\u0026ndash;71. \u003c/li\u003e\n\u003cli\u003eChen Q, Yang D, Zong H, Zhu L, Wang L, Wang X, et al. Growth-induced stress enhances epithelial-mesenchymal transition induced by IL-6 in clear cell renal cell carcinoma via the Akt/GSK-3\u0026beta;/\u0026beta;-catenin signaling pathway. Oncogenesis. 2017;6:e375. \u003c/li\u003e\n\u003cli\u003eYu W, Wang Y, Jiang Y, Zhang W, Li Y. Distinct immunophenotypes and prognostic factors in renal cell carcinoma with sarcomatoid differentiation: a systematic study of 19 immunohistochemical markers in 42 cases. BMC Cancer. 2017;17:293. \u003c/li\u003e\n\u003cli\u003eParadis V, Lagha NB, Zeimoura L, Blanchet P, Eschwege P, Ba N, et al. Expression of vascular endothelial growth factor in renal cell carcinomas. Virchows Arch. 2000;436:351\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eChoueiri TK, Kaelin WG. Targeting the HIF2-VEGF axis in renal cell carcinoma. Nat Med. 2020;26:1519\u0026ndash;30. \u003c/li\u003e\n\u003cli\u003eSitu Y, Xu Q, Deng L, Zhu Y, Gao R, Lei L, et al. System analysis of VEGFA in renal cell carcinoma: The expression, prognosis, gene regulation network and regulation targets. Int J Biol Markers. 2022;37:90\u0026ndash;101. \u003c/li\u003e\n\u003cli\u003eLee JS, Kim HS, Jung JJ, Park CS, Lee MC. Expression of vascular endothelial growth factor in renal cell carcinoma and the relation to angiogenesis and p53 protein expression. J Surg Oncol. 2001;77:55\u0026ndash;60. \u003c/li\u003e\n\u003cli\u003eChang SG, Jeon SH, Lee SJ, Choi JM, Kim YW. Clinical significance of urinary vascular endothelial growth factor and microvessel density in patients with renal cell carcinoma. Urology. 2001;58:904\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eRaica M, Cimpean AM, Anghel A. Immunohistochemical expression of vascular endothelial growth factor (VEGF) does not correlate with microvessel density in renal cell carcinoma. Neoplasma. 2007;54:278\u0026ndash;84. \u003c/li\u003e\n\u003cli\u003eHerneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol. 2003;45:208\u0026ndash;13. \u003c/li\u003e\n\u003cli\u003eQian T, Chen M, Gao F, Meng F, Gao X, Yin H. Diffusion-weighted magnetic resonance imaging to evaluate microvascular density after transarterial embolization ablation in a rabbit VX2 liver tumor model. Magn Reson Imaging. 2014;32:1052\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eHu Y, E H, Yu X, Li F, Zeng L, Lu Q, et al. Correlation of quantitative parameters of magnetic resonance perfusion-weighted imaging with vascular endothelial growth factor, microvessel density and hypoxia-inducible factor-1\u0026alpha; in nasopharyngeal carcinoma: Evaluation on radiosensitivity study. Clin Otolaryngol. 2018;43:425\u0026ndash;33. \u003c/li\u003e\n\u003cli\u003eYuan Q, Kapur P, Zhang Y, Xi Y, Carvo I, Signoretti S, et al. Intratumor Heterogeneity of Perfusion and Diffusion in Clear-Cell Renal Cell Carcinoma: Correlation With Tumor Cellularity. Clin Genitourin Cancer. 2016;14:e585\u0026ndash;94. \u003c/li\u003e\n\u003cli\u003eLee H-J, Rha SY, Chung YE, Shim HS, Kim YJ, Hur J, et al. Tumor perfusion-related parameter of diffusion-weighted magnetic resonance imaging: correlation with histological microvessel density. Magn Reson Med. 2014;71:1554\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eAkinci O, Turkoglu F, Nalbant MO, Inci E. Differentiating renal cell carcinoma and oncocytoma with volumetric MRI histogram analysis. North Clin Istanb. 2023;10:636\u0026ndash;41. \u003c/li\u003e\n\u003cli\u003eLaissy JP, Menegazzo D, Debray MP, Toublanc M, Ravery V, Dumont E, et al. Renal carcinoma: diagnosis of venous invasion with Gd-enhanced MR venography. Eur Radiol. 2000;10:1138\u0026ndash;43. \u003c/li\u003e\n\u003cli\u003eUhlig A, Uhlig J, Leha A, Biggemann L, Bachanek S, St\u0026ouml;ckle M, et al. Radiomics and machine learning for renal tumor subtype assessment using multiphase computed tomography in a multicenter setting. Eur Radiol. 2024; \u003c/li\u003e\n\u003cli\u003eJi J, Liu Y, Bao Y, Men Y, Hui Z. Network analysis of histopathological image features and genomics data improving prognosis performance in clear cell renal cell carcinoma. Urol Oncol. 2024;S1078-1439(24)00400-9. \u003c/li\u003e\n\u003cli\u003eFarias E, Terrematte P, Stransky B. Machine Learning Gene Signature to Metastatic ccRCC Based on ceRNA Network. Int J Mol Sci. 2024;25:4214. \u003c/li\u003e\n\u003cli\u003eBuhas BA, Toma V, Beauval J-B, Andras I, Couți R, Muntean LA-M, et al. Label-Free SERS of Urine Components: A Powerful Tool for Discriminating Renal Cell Carcinoma through Multivariate Analysis and Machine Learning Techniques. Int J Mol Sci. 2024;25:3891. \u003c/li\u003e\n\u003cli\u003eAlhussaini AJ, Steele JD, Jawli A, Nabi G. Radiomics Machine Learning Analysis of Clear Cell Renal Cell Carcinoma for Tumour Grade Prediction Based on Intra-Tumoural Sub-Region Heterogeneity. Cancers (Basel). 2024;16:1454. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"renal cell carcinoma, MRI, thrombus, consistency, diffusion-weighted imaging, volumetric, immunohistochemistry, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-6254932/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6254932/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground. \u003c/strong\u003eRenal cell carcinoma (RCC) possesses a distinctive inclination to infiltrate the inferior vena cava, resulting in the formation of a venous tumour thrombus (VTT). Accurately assessing the consistency of the VTT prior to surgery is essential for effective treatment strategizing and favourable results. The study aimed to investigate the performance of volumetric radiomic MRI analysis in prediction of consistency and histomorphological\u003cstrong\u003e \u003c/strong\u003evascular\u003cstrong\u003e \u003c/strong\u003epatterns of RCC venous tumour thrombus (VTT) with the assistance of machine learning.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods. \u003c/strong\u003eTwenty-four RCC patients with VTT underwent nephrectomy and thrombectomy in this study. Preoperatively abdominal DW-MRI was conducted, followed by the creation of 3D model of the thrombus. First-order radiomic features were computed from the complete thrombus volume utilizing ADC maps.\u003cstrong\u003e \u003c/strong\u003eThe immunohistochemical staining of VTT was performed using CD34, SMA and VEGFR. The machine learning analysis was employed to develop predictive models for VTT histologic features. Patients were grouped based on the thrombus consistency into either solid or friable categories.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults. \u003c/strong\u003eThe solid and friable thrombi were detected in 13 (54.2%) and 11 (45.8%) cases, respectively. \u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eLarge vessels were predominantly observed in solid VTTs (73.3%; p=0.015). Rich vascularization was a main pattern in solid VTT at 51.5%, contrasting with the friable at 9.1% (p=0.008). There was a strong association between thrombus vessel size and following radiomic features: entropy (r=0.722), skewness (r=0.635), and ADC mean (r=0.610). ADC entropy outperformed in distinguishing between VTT with large and small vessels, achieving the highest performance (AUC 0.930; p\u0026lt;0.001). In distinguishing between VTT with rich and poor vascularization, ADC median showed the best performance (AUC = 0.881; p \u0026lt; 0.001). Using machine learning analysis, we've developed two models predicting crucial histologic traits of VTT with prognostic accuracies of 89% for consistency and 75% for vessel size.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions. \u0026nbsp;\u003c/strong\u003eLeveraging volumetric radiomic data from MR-DWI, along with machine learning models, we identified unique vascular patterns in VTTs among RCC patients. These models were developed to predict VTT consistency and vessel size using volumetric ADC data from DWI.\u003c/p\u003e","manuscriptTitle":"Machine Learning-Based Prediction of Consistency and Histological Characteristics in Renal Cell Carcinoma Venous Tumor Thrombus Through Volumetric Radiomics","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-07 11:12:40","doi":"10.21203/rs.3.rs-6254932/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"29b524bc-3b53-41ff-af22-3cad170467e9","owner":[],"postedDate":"May 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-04T09:27:29+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-07 11:12:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6254932","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6254932","identity":"rs-6254932","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}