A Novel Multimodal Combining Radiomics and Tumor-Stroma Ratio (TSR) Improves Diagnosis of Gastric Cancer Peritoneal Metastasis

A Novel Multimodal Combining Radiomics and Tumor-Stroma Ratio (TSR) Improves Diagnosis of Gastric Cancer Peritoneal Metastasis | 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 A Novel Multimodal Combining Radiomics and Tumor-Stroma Ratio (TSR) Improves Diagnosis of Gastric Cancer Peritoneal Metastasis Lin Zhong, Ting Lin, Dong Hou, Hongyun Huang, Shihai Zhou, Yu Lin, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7336599/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 Peritoneal metastasis (PM) is the most common form of metastasis in gastric cancer (GC), frequently leading to severe complications and a significantly poor prognosis. Prompt and early diagnosis of PM in GC is crucial. However, diagnostic laparoscopy and CT scans, while being the primary methods for identifying PM in GC, have notable limitations, such as being invasive and having low sensitivity. Therefore, developing a diagnostic model for PM in GC based on routine examination results holds substantial importance. In this retrospective study, we enrolled 813 patients from two medical centers and developed a robust diagnostic model by integrating various routine examination results, including CT scans, endoscopy, pathology, and hematological tests. To further explore the potential significance of various examination results, we conducted radiomic analysis of CT images, analyzed histopathological results via the Tumor-Stroma Ratio (TSR), and examined hematological results through parameters such as the Prognostic Nutritional Index (PNI), Neutrophil to Lymphocyte Ratio (NLR), and Albumin to Globulin Ratio (AGR). A novel diagnostic model, incorporating CA125, CA724, Borrmann classification, radiomics features, and the TSR, was successfully constructed. This model demonstrated strong performance in diagnosing synchronous PM (AUC = 0.874) and predicting metachronous (AUC = 0.784) PM in GC. To facilitate clinical application, a nomogram was constructed. Consequently, the study presents a novel and comprehensive diagnostic model for PM in GC patients, offering significant promise for clinical applicability based on routine examination results. Radiomics Tumor-Stroma Ratio (TSR) Diagnostic model Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Gastric cancer (GC) is one of the most common malignant tumors worldwide, posing a serious threat to public health. Although the incidence and mortality rates of GC have declined in recent years, its incidence still ranks 5th, and cancer-related mortality is 4th [ 1 ] . Due to the hidden symptoms of GC, the vast majority of patients are in the middle or late stages at the time of diagnosis [ 2 ] .The most common forms of metastasis for gastric cancer are peritoneal, liver, and lymph node metastases [ 3 ] , with peritoneal metastasis (PM) being one of the most common in advanced stages. PM can be categorized into synchronous and metachronous PM. Approximately 30% of patients with advanced GC develop synchronous PM, and this proportion increases to 45% in patients with poorly cohesive carcinomas (PCCs). Furthermore, 40–50% of GC patients still progress to PM even after undergoing standardized systemic treatment [ 4 , 5 ] . Patients with PM from GC face a markedly poor prognosis, characterized by a short survival duration, frequent and severe complications, and a lack of effective palliative therapeutic options [ 6 ] . It has been confirmed that a low Peritoneal Cancer Index (PCI), complete Cytoreductive Surgery (CRS), and comprehensive systemic therapy can effectively improve the prognosis of patients with PM of GC [ 4 , 7 – 9 ] . Neoadjuvant therapies, such as intraperitoneal chemotherapy and systemic therapy, can effectively downstage patients with PM of GC, leading to the disappearance of peritoneal metastatic lesions and negative free abdominal tumor cells, thereby increasing the median survival to 21.6–34.6 months [ 10 – 12 ] . Early diagnosis of PM in GC is crucial for improving patient prognosis. Currently, the diagnosis of PM in GC mainly relies on clinical signs, imaging examinations, and surgical exploration results [ 13 ] . In GC patients with PM, initial clinical manifestations frequently lack diagnostic specificity. The neoplastic process characteristically maintains clinical silence until attaining advanced disease status, at which juncture pathognomonic indicators such as refractory ascites, progressive abdominal pain, and measurable abdominal circumference expansion manifest as objectively verifiable clinical signs. [ 13 ] . CT, a routine imaging examination for GC, shows severely insufficient diagnostic sensitivity for peritoneal metastases, ranging only between 28% and 51% [ 14 , 15 ] . Positron emission tomography CT (PET-CT) is commonly used for diagnosing distant metastases of GC. However, Fluoro-2-deoxy-D-glucose PET-CT, a specific form of this imaging, can only detect 3% of occult PM in GC [ 16 ] . Despite the National Comprehensive Cancer Network (NCCN) guidelines assigning diagnostic laparoscopy a Category 2B recommendation (indicating a lower level of evidence), this procedure still requires general anesthesia, an intervention that imposes both physiological stress and psychological distress on patients [ 17 ] . The development of diagnostic models based on routine clinical parameters for PM detection in GC patients holds significant clinical value, particularly in resource-limited settings where such cost-effective diagnostic strategies are particularly advantageous. Radiomics is a novel diagnostic tool that extracts hundreds of quantitative features from medical images, combining key features into image-based biomarkers (known as radiomics signatures) to enhance diagnostic efficacy [ 18 ] . Tumor-Stroma Ratio (TSR), a quantitative histopathological biomarker defined as the proportional area of tumor cells relative to stromal components within the tumor microenvironment, is characterized by its technical simplicity, rapid assessment, cost-effectiveness, and high reproducibility. Emerging evidence highlights TSR as a robust prognostic indicator, demonstrating significant associations with tumor progression, metastatic dissemination, chemotherapy resistance, and diminished overall survival. Notably, our previous study has further established TSR as an independent diagnostic predictor for PM in GC [ 19 ] , underscoring its pivotal role in clinical decision-making for advanced malignancies. This study aims to develop an integrated multimodal diagnostic model that incorporates radiomics, TSR, and routine examination results to improve the accuracy and effectiveness of PM detection in GC. The multimodal model is expected to enhance the sensitivity and specificity of PM diagnosis, while also providing more comprehensive clinical information to support decision-making, ultimately improving treatment outcomes and survival for patients with PM in GC. Methods Patients The patients who were treated at Zhujiang Hospital, Southern Medical University, from December 2010 to March 2021, were enrolled in the study as the training cohort based on the following inclusion and exclusion criteria. Following these criteria, the patients who were treated at Zhongshan People's Hospital during the same period were enrolled as the external validation cohort. The patients who were identified as non-peritoneal metastasis (non-PM) from these two centers were combined into the metachronous PM cohort ( Fig. 1 ) . Inclusion criteria: All patients were pathologically diagnosed with GC; All patients underwent standard radical gastrectomy for gastric cancer (D2) and systemic therapy; All patients with PM were based on pathological diagnosis. Exclusion criteria: Patients with tumor cell residues at the margins of the resected tissue in the postoperative pathology report; Patients who received radiotherapy, chemotherapy, or immunotherapy before surgery; Patients with a history of other malignant tumors; Patients whose clinical data are seriously incomplete. Follow-up The metachronous PM cohort consists of 648 patients from the training and external validation cohorts, all diagnosed with non-PM. The follow-up method was conducted through telephone or outpatient visits. The deadline for the follow-up was March 2021. Ultimately, 62 patients were diagnosed with PM based on CT imaging or pathological confirmation. Hematological indicators All hematological indicators included in this study are commonly used in clinical practice. Additionally, this study incorporates hematological indicators that have been found to be associated with the prognosis of GC, intended for secondary analysis. These include white blood cell count (WBC, g/L), percentage of neutrophils (Neut%), percentage of lymphocytes (Lymph%), hemoglobin (Hb, g/L), platelet count (PLT, g/L), mean platelet volume (MPV, fl), platelet distribution width (PDW, %), neutrophil count (Neu, g/L), lymphocyte count (Lymph, g/L), albumin (Alb, g/L), globulin (Glo, g/L), alpha-fetoprotein (AFP, ng/ml), carcinoembryonic antigen (CEA, µg/L), CA125 (kU/L), CA153 (U/ml), CA199 (kU/L), and CA724 (U/mL). The neutrophil to lymphocyte ratio (NLR) is associated with overall survival (OS) of gastric cancer patients, tumor invasion depth, and peritoneal metastasis [ 20 ] . The prognostic nutritional index (PNI) was calculated using the formula: 1×serum albumin level (g/L) + 5×absolute lymphocyte count (10^9/L) [ 21 ] . Additionally, the albumin to globulin ratio (AGR, A/G) is also commonly considered[22]. Radiomics based on CT imaging The maximum diameter (Size, cm) of the tumor was measured on its largest cross-section. The T stage is determined based on the diagnoses from CT and endoscopic ultrasonography. To further enhance the diagnostic efficacy of CT for peritoneal metastasis of gastric cancer, we conducted radiomics analysis using preoperative CT imaging data. And incorporate radiomics features as independent parameters for further analysis. The detailed methods, as shown in the supplementary materials ( Fig. 2 a and Supplemental Fig. 1) . Gastroscopy indicators The gastroscopy examination indicators included in the study include the tumor occurrence site (Site: 0 = entire stomach; 1 = cardia, 2 = gastric body, 3 = antrum), Borrmann classification, history of Helicobacter pylori infection (Hp), and T stage (according to the diagnosis from endoscopic ultrasonography). Pathological indicators The pathological indicators included in the study comprise tumor differentiation (Diff: 0 = undifferentiated, 1 = poorly differentiated, 2 = moderately differentiated, 3 = well differentiated) and HER-2 expression status (HER-2, Negative or Positive). Our previous research also confirmed that TSR (Tumor-Stroma Ratio) is related to metastatic nodules in gastric cancer (GC) [ 23 ] .Other studies have verified that TSR is closely related to the prognosis of gastric cancer [ 24 , 25 ] . Further research has found that TSR is also an important predictive indicator for peritoneal metastasis of gastric cancer. Therefore, we have included this simple, rapid, low-cost, and reproducible diagnostic indicator in our analysis. The specific research methods are as follows: TSR analysis Hematoxylin and eosin (H&E) sections were collected from all patients for TSR analysis. Conventionally, TSR is assessed using postoperative pathological specimens to provide prognostic insights. To evaluate the feasibility of TSR assessment via endoscopic biopsy, we conducted a comparative analysis of paired endoscopic biopsy specimens and corresponding surgically resected tissues from 35 consecutive patients. Initially, a microscope with a 5× microscope was used to pinpoint the area of deepest tumor infiltration. Subsequently, we identified over three distinct observation zones, each exhibiting both tumor and stromal components, ensuring the presence of surrounding tumor cells. The TSR was then scored using a 10× magnification microscope by calculating the stroma's percentage within the observed field. Scores were assigned in increments of ten percent (e.g., 10%, 20%, 30%, etc.). Based on these TSR values, the samples were categorized into two groups: a Low TSR group (< 50%), indicating a stroma-poor tumor environment, and a High TSR group (≥ 50%), signifying a stroma-rich environment ( Fig. 3 ) . Statistics In this study, all statistical analyses were executed using R software, version 4.2.2. Comparative tables of baseline patient data for the training cohort, external validation cohort, and metachronous PM cohort were created using the 'tableone' package. The normality of continuous variables was tested with the Shapiro-Wilk test, and the Mann-Whitney U test was used for those not normally distributed. Chi-square and Fisher's exact tests were applied for comparing categorical and ordinal variables. Univariate and multivariate logistic regression analyses, crucial for assessing the influence of preoperative clinical data on the incidence of PM in GC, were conducted using the 'rms' package. A predictive nomogram was developed, and its efficacy was evaluated using ROC curves generated by the 'pROC', 'car', and 'rmda' packages to illustrate the area under the curve (AUC). Additionally, the 'rmda' package was employed for calibration curve construction, ensuring the prediction model's accuracy, while the 'car' package facilitated clinical decision curve analysis by quantifying the net benefit across various threshold probabilities, thereby assessing the nomogram's clinical utility. Statistical significance was determined at P < 0.05. Results Characteristics of the Patients. Based on the inclusion and exclusion criteria set forth, the study encompassed 813 participants: 432 patients from Zhujiang Hospital, Southern Medical University, and 381 from Zhongshan City People's Hospital. Patients from Zhujiang Hospital served as the training cohort, whereas those from Zhongshan City were designated as the external validation cohort. Since PM of GC is classified into synchronous and metachronous metastasis, we aimed to further validate the predictive efficacy of our model for the relapse of PM in GC. To achieve this, we combined patients from both centers who did not experience PM into metachronous PM cohort. The training cohort was classified into non-PM (321 patients) and PM (111 patients) groups. The gender distribution was 67% male and 33% female, whereas the age distribution showed that 64% were less than 65 years old and 36% were 65 or above. Significant differences were found between the non-PM and PM groups in BMI, WBC, Neut%, Lmph%, Hb, PLT, MPV, PDW, Neu, Lymph, NLR, Alb, PNI, A/G, AFP, CEA, CA125, CA199, CA724, Site, Size, T stage, Borrmann, Diff, Radiomics Features and TSR ( P < 0.05) ( Tab. 1 ). The validation cohort included 327 non-PM patients and 54 PM patients. The gender distribution was the same as the training cohort, and the age breakdown was 62% under 65 years and 38% 65 or older. There were statistical differences between the two groups in indicators such as Neut%, Lmph%, PLT, Lymph, NLR, A/G, CA125, CA724, Size, T stage, Radiomics Features, and TSR ( P < 0.05)( Tab. 1 ). The metachronous PM cohort included 648 patients. The gender distribution was the same as in the above cohorts, and the age breakdown was 63% <65 years (n=408) and 37% ≥65 years (n=240). Based on the follow-up results, patients in the metachronous PM cohort were also divided into Non-PM and PM groups. Statistical differences were observed in the levels of WBC, Neut%, Lmph%, Neu, Lymph, NLR, Alb, PNI, CA125, CA724, T stage, Borrmann type, Diff, Radiomics Features, and TSR ( P <0.05) ( Tab. 1 ). Comprehensive analysis indicates that the above factors may all be associated with the occurrence of PM in GC. Neut%, Lmph%, Lymph, NLR, CA125, CA724, T stage, Borrmann type, Radiomics Features, and TSR showed statistical differences across all three cohorts and may act as important predictors of PM in GC. Radiomics features and performance of the radiomics model. Through intraclass correlation coefficient (ICC) analysis (threshold >0.75), 626 radiomic features demonstrated measurement stability from the initial 1,316 candidates, indicating high inter-observer concordance in feature extraction. These validated features underwent LASSO regression for dimensionality reduction prior to principal component analysis (PCA), which ultimately derived 498 components accounting for 85% cumulative variance. Following rigorous feature selection via ANOVA (FDR-corrected p<0.01) from 498 candidate features, 26 optimal radiomic features were retained for model development. The Gaussian Process (GP) classifier demonstrated robust discriminative capacity with area under the curve (AUC) values of 0.821 (95% CI: 0.785-0.857) in the training cohort, 0.719 (95% CI:0.677-0.761) in the internal test cohort, and 0.701 (95% CI:0.658-0.744) in the external validation cohort. Fig.2b- Fig.2d illustrate the corresponding receiver operating characteristic (ROC) curves across all cohorts. TSR score All H&E-stained histological sections were initially evaluated in a blinded manner by two independent pathologists. In cases of discrepant interpretations, consensus was reached through joint microscopic review. Inter-observer agreement was quantified using Cohen’s kappa coefficient (κ = 0.87, 95% CI: 0.79–0.94), indicating excellent diagnostic concordance. Among 35 matched endoscopic–surgical specimen pairs, TSR assessment demonstrated substantial agreement using a 50% stromal cutoff (AC1 = 0.722; overall concordance rate: 85.7%), supporting the reliability of endoscopic specimens for TSR evaluation. In the training cohort, the non-PM group (n = 321) exhibited a higher proportion of TSR-low cases (64.8%, 208/321) compared to TSR-high cases (35.2%, 113/321). Conversely, the PM group (n = 111) showed a predominance of TSR-high cases (64.9%, 72/111) over TSR-low cases (35.1%, 39/111), with statistically significant stratification (χ² = 5.89, p = 0.02). In the external validation cohort, a similar distribution pattern was observed: TSR-low cases comprised 55.4% (181/327) of the non-PM group, while only 33.3% (18/54) of the PM group, maintaining statistical significance (χ² = 4.92, p = 0.03). In the metachronous PM cohort, pooled data from 648 patients revealed a significantly higher proportion of TSR-high cases in PM patients (62.9%, 39/62) compared to non-PM controls (40.4%, 237/586), demonstrating robust discriminative capacity (χ² = 10.24, p < 0.01). Risk factors associated with PM in GC. To identify risk factors associated with PM in GC, we conducted a univariate logistic regression analysis incorporating parameters that exhibited statistical differences between the non-PM group and the PM group. These parameters included BMI, WBC, Neut%, Lymph%, Hb, PLT, MPV, PDW, Neu, Lymph, NLR, Alb, PNI, A/G, AFP, CEA, CA125, CA199, CA724, Site, Size, T stage, Borrmann classification, Diff, Radiomics Features, and TSR. The analysis revealed significant statistical differences in parameters such as WBC, Neut%, Lymph%, PLT, Alb, PNI, CA125, CA724, tumor Size, T stage, Borrmann classification, Radiomics Features, and TSR ( P < 0.01) ( Tab. 2 ).Further, through multivariate regression analysis, we found that that CA125 (OR = 1.025, P < 0.001), CA724 (OR = 1.026, P = 0.002), Borrmann (OR = 10.191, P = 0.003), Radiomics Features (OR = 207.221, P < 0.001), and TSR (OR = 2.351, P = 0.003) have a significant correlation with the incidence of PM in GC ( Tab. 2 ). Therefore, we consider CA125, CA724, Borrmann, Radiomics Features, and TSR to be independent influencing factors for PM in GC. Diagnostic model construction based on the preoperative data. According to the results above, we constructed a risk score model for all patients, resulting in the following predictive equation: Risk Score = 0.0217* CA125 + 0.0286 * CA724 +2.0504 * Borrmann + 5.0457 * Radiomics Features + 0.8119 * (TSR≥50%) - 6.6743. Consequently, the probability of PM can be calculated by the formula: Probability of PM = 1 / (1 + e^ (- Risk Score)) To enhance the intuitive understanding of our model, we developed a Nomogram informed by pertinent findings. Illustrated in Fig 4a , the process involves aggregating the values assigned to each indicator, culminating in the Total Points. Correspondingly, the value located on the bottom horizontal axis of the Nomogram, aligned with these Total Points, indicates the risk probability of PM in GC. Model performance and validation The performance of this individualized nomogram model is evaluated by drawing the ROC curve and calculating the area under the curve. The area under the ROC curve for the training set is presented in Fig 4b , with an area of 0.874 (95% CI: 0.839-0.909). The AUC for the external validation cohort is 0.797 (95% CI: 0.733-0.862) ( Fig 4c ), and for the prediction cohort is 0.784 (95% CI: 0.724-0.845) ( Fig 4d ) . The calibration curve reveals good consistency between the nomogram and actual observations (Fig 4e and Supplemental Fig. 1). The maximum Youden index of 0.612 of the ROC curves of the nomogram was selected as the optimal cutoff value in the training cohort, and patients were divided into high-risk and low-risk groups. We found that the sensitivity, specificity, accuracy of the nomogram in the training cohort were 72.0 %, 89.2%, 80.6% respectively (Fig 4b ). In the validation cohort, the sensitivity was 79.8 %, specificity was 68.5%, and the accuracy was74.2% (Fig 4c) . In the metachronous PM cohort , the sensitivity was 66.0 %, the specificity was70.6%, and the accuracy was 68.3% (Fig 4d) . Clinical usefulness To evaluate the clinical usefulness of this model, a clinical decision curve was plotted in the training set cohort. This Decision curve included the model and independent factors significantly associated with PM in GC, and quantified their clinical net benefit. As shown in Fig 4f , the net benefit of this model is higher than that of all or each independent factor. Therefore, intervening with patients based on this model yields a good net benefit, and using this model for the diagnosis and treatment of PM in GC has great clinical benefit. Discussion In recent years, the incidence of GC has shown a downward trend, but at the same time, the proportion of patients with simultaneous PM is gradually increasing [ 26 ] . Despite continuous improvements in treatment, the prognosis for GC patients remains unsatisfactory, especially for those with PM, which is even more concerning [ 27 ] . Although there is some controversy over intraperitoneal chemotherapy in the treatment of PM in GC, it is a consensus that the completeness of CRS and a low PCI index are key factors in the prognosis of patients with PM [ 9 , 28 , 29 ] .The International Peritoneal Surface Oncology Group (PSOGI) recommends an initial pathological diagnosis through diagnostic laparoscopy, along with an assessment of the PCI. They advocate for conducting intraoperative HIPEC treatment following neoadjuvant chemotherapy. This is complemented by real-time combined CRS-HIPEC surgery, subsequent intraperitoneal chemotherapy, and sequential systemic chemotherapy [ 30 ] . For patients with low PCI undergoing neoadjuvant intraperitoneal systemic chemotherapy (NIPS), over 50% become cytologically negative, and 70% undergo complete CRS [ 19 ] . Therefore, timely and accurate diagnosis, along with treatment at experienced medical centers, significantly improves the prognosis of patients with PM in GC. According to NCCN guidelines, diagnostic laparoscopy combined with peritoneal lavage fluid examination is categorized as a 2B recommendation, acknowledging significant drawbacks of this invasive diagnostic method [ 17 ] . The ideal diagnostic approach should possess attributes such as high sensitivity, high specificity, accuracy, simplicity, non-invasiveness, and affordability. In this study, based on routine preoperative examinations for GC, deep insights from various examination results were explored. Using data from hematological, CT scans, gastroscopic, and pathological examinations of gastroscopic tissue biopsies, a model was constructed incorporating levels of CA125, CA724, Radiomics Features, Boman, and TSR. This model can effectively diagnose the occurrence of PM in GC (AUC = 0.874) and predict the metachronous PM in GC (AUC = 0.784). Moreover, the model is characterized by its simplicity, good repeatability, ease of implementation, and cost-effectiveness. Hematological examination is the most common clinical test. CA125 and CA724 are commonly used tumor markers for GC, but due to their low sensitivity and specificity, their diagnostic efficacy is limited [ 31 ] . The results of this study are consistent with many others, confirming that CA125 and CA724 are related to the occurrence of PM in GC [ 32 – 34 ] . NLR, PNI, and A/G have been proven to be new predictive factors for PM of GC. Hideaki Shimada et al. found that the NLR reflects the systemic inflammatory state and is an independent risk factor for poor prognosis in GC patients [ 20 ] . NLR > 3.5 is a risk factor for PM in GC [ 35 ] . In this study, significant differences in NLR were observed in all three cohorts ( P < 0.05) but regression analysis found it could not serve as an effective predictor for the occurrence of PM. The PNI is also a useful tool for assessing the prognosis of patients with metastatic GC [ 36 ] , and has been proven to predict PM of GC [ 37 ] . Univariate retrospective analysis also confirmed that PNI is an independent risk factor for PM of GC, but this predictive factor was not included in the multivariate retrospective analysis. The A/G mainly serves as a clinical marker for multiple myeloma or other immunoproliferative diseases, and studies have found that A/G is a new independent indicator for predicting early recurrence of curative gastric cancer [ 38 ] . In the training cohort, there is a significant difference in the presence of A/G between the PM group and the non-PM group ( P < 0.05); however, A/G is not an independent influencing factor in the logistic regression analysis (Univariate: P = 0.05; Multivariate: P = 0.507). We attribute the primary reason for this phenomenon to the insufficient predictive efficacy of these markers. CT is the most commonly used imaging examination for GC. In this study, it was also confirmed that CT-based radiomics has significant efficacy in diagnosing PM of GC. The sensitivity of CT scans is severely inadequate for diagnosing PM of GC, especially for nodules smaller than 5mm [ 39 ] . Radiomics is a new diagnostic tool that extracts hundreds of quantitative features from medical images and combines key features into image-based biomarkers (known as radiomic signatures) for cancer diagnosis [ 18 ] . Previous studies have constructed predictive models for PM of GC based on radiomics of the primary tumor [ 40 ] , omental features [ 41 ] , and a combination of primary tumors with adjacent peritoneum [ 42 ] . However, previous studies often selected a plane to outline the Region of Interest (ROI) on the 2D level for CT image feature extraction, leading to uncertainty in the final features included. Although efforts to increase the number of reading planes have been made, they still have limitations such as low repeatability and one-sidedness. In our study, we outlined the Volume of Interest (VOI) within the possible range of metastasis and extracted corresponding texture information from a 3D spatial perspective. This helps improve image characterization properties in consideration of factors like the heterogeneity of gastric cancer. The comprehensive predictive model constructed with radiomic features in this study has significantly enhanced the efficacy of predicting PM in GC compared to previous studies that solely relied on CT radiomics [ 43 ] . Additionally, all the analysis tools used in this study are open-source and accessible online, increasing the applicability of the model. Gastroscopy is the primary method for the screening and diagnosis of GC, allowing for direct observation of changes in the gastric mucosa and facilitating pathological sampling. Borrmann type 3 and 4 GC often exhibit strong invasiveness, making the Borrmann classification a significant risk factor for PM in GC. Our analysis results are consistent with previous studies, confirming the significance of Borrmann classification in predicting the risk of PM in GC [ 44 , 45 ] . Pathomics captures a vast amount of data from digital pathological images to generate quantitative features that describe various phenotypes of tissue samples. These data are analyzed to identify key indicators for diagnosis or prognosis prediction [ 46 ] . However, pathomics requires digitization of pathological images, necessitating specialized equipment and time-consuming scanning processes, thereby increasing time and economic costs, as well as specific analysis tools. In our study, we incorporated TSR on top of routine pathological indicators. TSR, as an easy, fast, cost-neutral, and easily repeatable pathological parameter, can intuitively describe the relationship between tumor cells and tumor-associated stroma. Our previous studies have demonstrated that stroma-rich metastatic nodules are independent determinants of poor prognosis in gastric cancer (GC) [ 23 ] . We further discovered that TSR is closely related to the occurrence of PM in GC. Although TSR is primarily assessed from postoperative pathological specimens, we found that TSR grading of tissues obtained via endoscopy is highly consistent with postoperative histopathology, as indicated by a Kappa value of 0.706, an AC1 of 0.722, and an agreement rate of 85.7%. Therefore, TSR can be entirely obtained from endoscopic biopsy tissue pathology as a prognostic indicator. Lauren diffuse type GC is also a high-risk factor for PM [ 47 ] , but in our study, effective analysis was hindered due to severe data missing for Lauren classification. However, Lauren diffuse type is characterized by poorly adhesive cells diffusely infiltrating gastric wall structures, with single cells encapsulated in fibroproliferative stroma, showing few or no gland formations, features that bear significant similarities to high TSR. As research progresses, there has been a paradigm shift toward molecular-level prediction of PM in GC. Lee IS, et al. have found that a transcriptomic profile of six genes (ZBTB1, CAVIN2, CHCHD, LTBP3, SLITPK6, STT3B) can effectively predict the occurrence of PM in GC (AUC = 0.72). When combined with Borrmann type and T stage in the model, the predictive power significantly increases (AUC = 0.84) [ 44 ] . Yanyan Chen, et al. identified through proteomics analysis that 10 proteins (DUOXA2, ITGA7, LIMS1, MSRB3, PLCB1, RAB6B, SEMA3C, SMTN, TADA1, TBC1D14) can effectively predict the occurrence of PM (ROC = 0.83) [ 48 ] . Other research teams have focused on the tumor microenvironment (TME), predicting gastric cancer peritoneal metastasis from aspects like metabolism [ 49 ] , immune microenvironment [ 50 ] , and tumor stroma characteristics [ 51 ] . However, the cost and complexity of gene expression analysis hinder its widespread clinical application, especially in developing countries. The integration of multi-omics, including pathomics, radiomics, transcriptomics, and proteomics, is beneficial for the diagnosis and determination of treatment strategies for diseases. Such a synergistic combination is likely to enhance the sensitivity, specificity, and accuracy of diagnosing PM in GC. We anticipate that this multimodal approach will improve diagnostic outcomes. However, it will undoubtedly increase the economic burden on patients and the workload of clinical physicians, which is a significant consideration, especially for developing countries. The model constructed in our study is not only effective in diagnosing synchronous PM but also in predicting metachronous PM, yielding significant clinical benefits. However, there are certain limitations to our research. First, our data were derived solely from two medical centers, which necessitates further validation with more external data, especially from Western countries. Second, as a retrospective study, there is an inherent risk of missing clinical feature data. Despite using Monte Carlo simulations to impute missing data, biases are unavoidable. Third, to ensure the adequate sample size of the data, considering the good consistency of TSR between endoscopic and postoperative tissue, we performed the evaluation of TSR in postoperative tissue when TSR failed to be evaluated in endoscopic tissue. Hence, comprehensive preoperative data are essential to validate this result. Fourth, the range of preoperative examination parameters covered in this study is limited. Therefore, it is necessary to enrich the examination parameters and conduct prospective, multicenter studies based on our research. Conclusion The model, integrating radiomics, TSR, and routine preoperative examination results, exhibits robust diagnostic performance for detecting PM in GC, thereby informing clinical treatment strategies. Declarations Conflict of interest The authors declare that they have no conflict of interest. Ethics Statement The study was approved by the Medical Ethics and Institutional Committee of Zhujiang Hospital of Southern Medical University and Zhongshan People's Hospital (Approval No. 2024-KY-417-01). All research procedures involving human participants conformed to the Declaration of Helsinki. Given the retrospective nature of the study and the use of anonymized clinical data without direct patient contact or intervention, the requirement for informed consent was waived by the Medical Ethics and Institutional Committee in accordance with relevant national regulations. Funding This research was supported by China Postdoctoral Science Foundation (2022M723656). Author Contribution Lin Zhong: Conceptualization; formal analysis; writing – original draft. Ting Lin: Data curation; investigation; methodology; writing – original draft. Dong Hou: Formal Analysis; methodology. Hongyun Huang: Data Curation. Shihai Zhou: Formal Analysis. Yu Lin: Investigation. Yue Yu: Resources. Liangquan Liu: data curation. Jing luo: writing – review and editing. Fanghai Han: Methodology; writing – review and editing. Lang Xie: Conceptualization; data curation; investigation; methodology; validation; writing – review and editing. Data Availability The datasets generated and/or analysed during the current study have been deposited in the Science Data Bank (SciDB) repository ( DOI: 10.57760/sciencedb.29034). References Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49. Layke JC, Lopez PP. Gastric cancer: diagnosis and treatment options. Am Fam Physician. 2004;69(5):1133–40. Zhang C, Yang J, Chen Y, Jiang F, Liao H, Liu X, Wang Y, Kong G, Zhang X, Li J, et al. miRNAs derived from plasma small extracellular vesicles predict organo-tropic metastasis of gastric cancer. Gastric Cancer. 2022;25(2):360–74. Bonnot PE, Lintis A, Mercier F, Benzerdjeb N, Passot G, Pocard M, Meunier B, Bereder JM, Abboud K, Marchal F, et al. Prognosis of poorly cohesive gastric cancer after complete cytoreductive surgery with or without hyperthermic intraperitoneal chemotherapy (CYTO-CHIP study). Br J Surg. 2021;108(10):1225–35. Gong X, Hou D, Zhou S, Tan J, Zhong G, Yang B, Xie L, Han F, Zhong L. FMO family may serve as novel marker and potential therapeutic target for the peritoneal metastasis in gastric cancer. Front Oncol. 2023;13:1144775. Wang R, Song S, Harada K, Ghazanfari Amlashi F, Badgwell B, Pizzi MP, Xu Y, Zhao W, Dong X, Jin J, et al. Multiplex profiling of peritoneal metastases from gastric adenocarcinoma identified novel targets and molecular subtypes that predict treatment response. Gut. 2020;69(1):18–31. Coccolini F, Catena F, Glehen O, Yonemura Y, Sugarbaker PH, Piso P, Montori G, Ansaloni L. Complete versus incomplete cytoreduction in peritoneal carcinosis from gastric cancer, with consideration to PCI cut-off. Systematic review and meta-analysis. Eur J Surg Oncol. 2015;41(7):911–9. Yonemura Y, Canbay E, Li Y, Coccolini F, Glehen O, Sugarbaker PH, Morris D, Moran B, Gonzaletz-Moreno S, Deraco M, et al. A comprehensive treatment for peritoneal metastases from gastric cancer with curative intent. Eur J Surg Oncol. 2016;42(8):1123–31. Rudloff U, Langan RC, Mullinax JE, Beane JD, Steinberg SM, Beresnev T, Webb CC, Walker M, Toomey MA, Schrump D, et al. Impact of maximal cytoreductive surgery plus regional heated intraperitoneal chemotherapy (HIPEC) on outcome of patients with peritoneal carcinomatosis of gastric origin: results of the GYMSSA trial. J Surg Oncol. 2014;110(3):275–84. Kitayama J, Ishigami H, Yamaguchi H, Yamashita H, Emoto S, Kaisaki S, Watanabe T. Salvage gastrectomy after intravenous and intraperitoneal paclitaxel (PTX) administration with oral S-1 for peritoneal dissemination of advanced gastric cancer with malignant ascites. Ann Surg Oncol. 2014;21(2):539–46. Chan DY, Syn NL, Yap R, Phua JN, Soh TI, Chee CE, Nga ME, Shabbir A, So JB, Yong WP. Conversion Surgery Post-Intraperitoneal Paclitaxel and Systemic Chemotherapy for Gastric Cancer Carcinomatosis Peritonei. Are We Ready? J Gastrointest Surg. 2017;21(3):425–33. Ishigami H, Yamaguchi H, Yamashita H, Asakage M, Kitayama J. Surgery after intraperitoneal and systemic chemotherapy for gastric cancer with peritoneal metastasis or positive peritoneal cytology findings. Gastric Cancer. 2017;20(Suppl 1):128–34. Gwee YX, Chia DKA, So J, Ceelen W, Yong WP, Tan P, Ong CJ, Sundar R. Integration of Genomic Biology Into Therapeutic Strategies of Gastric Cancer Peritoneal Metastasis. J Clin Oncol. 2022;40(24):2830. Yajima K, Kanda T, Ohashi M, Wakai T, Nakagawa S, Sasamoto R, Hatakeyama K. Clinical and diagnostic significance of preoperative computed tomography findings of ascites in patients with advanced gastric cancer. Am J Surg. 2006;192(2):185–90. Kim SJ, Kim HH, Kim YH, Hwang SH, Lee HS, Park DJ, Kim SY, Lee KH. Peritoneal metastasis: detection with 16- or 64-detector row CT in patients undergoing surgery for gastric cancer. Radiology. 2009;253(2):407–15. Gertsen EC, Brenkman HJF, van Hillegersberg R, van Sandick JW, van Berge Henegouwen MI, Gisbertz SS, Luyer MDP, Nieuwenhuijzen GAP, van Lanschot JJB, Lagarde SM, et al. 18F-Fludeoxyglucose-Positron Emission Tomography/Computed Tomography and Laparoscopy for Staging of Locally Advanced Gastric Cancer: A Multicenter Prospective Dutch Cohort Study (PLASTIC). JAMA Surg. 2021;156(12):e215340. Ajani JA, D'Amico TA, Bentrem DJ, Chao J, Cooke D, Corvera C, Das P, Enzinger PC, Enzler T, Fanta P, et al. Gastric Cancer, Version 2.2022, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2022;20(2):167–92. Limkin EJ, Sun R, Dercle L, Zacharaki EI, Robert C, Reuze S, Schernberg A, Paragios N, Deutsch E, Ferte C. Promises and challenges for the implementation of computational medical imaging (radiomics) in oncology. Ann Oncol. 2017;28(6):1191–206. Zhong L, Huang H, Hou D, Zhou S, Lin Y, Yu Y, Yu J, Han F, Xie L. Tumor-Stroma Ratio is a Critical Indicator of Peritoneal Metastasis in Gastric Cancer. Clin Exp Gastroenterol. 2025;18:11–24. Shimada H, Takiguchi N, Kainuma O, Soda H, Ikeda A, Cho A, Miyazaki A, Gunji H, Yamamoto H, Nagata M. High preoperative neutrophil-lymphocyte ratio predicts poor survival in patients with gastric cancer. Gastric Cancer. 2010;13(3):170–6. Pan Y, Ma Y, Dai G. The Prognostic Value of the Prognostic Nutritional Index in Patients with Advanced or Metastatic Gastric Cancer Treated with Immunotherapy. Nutrients 2023, 15(19). Wei C, Yu Z, Wang G, Zhou Y, Tian L. Low Pretreatment Albumin-to-Globulin Ratio Predicts Poor Prognosis in Gastric Cancer: Insight From a Meta-Analysis. Front Oncol. 2020;10:623046. Huang J, Yang B, Tan J, Zhou S, Chen Z, Zhong G, Gao H, Zhu J, Zeng J, Zhong L, et al. Gastric cancer nodal tumour-stroma ratios influence prognosis. Br J Surg. 2020;107(13):1713–8. Aurello P, Berardi G, Giulitti D, Palumbo A, Tierno SM, Nigri G, D'Angelo F, Pilozzi E, Ramacciato G. Tumor-Stroma Ratio is an independent predictor for overall survival and disease free survival in gastric cancer patients. Surgeon. 2017;15(6):329–35. Kim EY, Abdul-Ghafar J, Chong Y, Yim K. Calculated Tumor-Associated Neutrophils Are Associated with the Tumor-Stroma Ratio and Predict a Poor Prognosis in Advanced Gastric Cancer. Biomedicines 2022, 10(3). Koemans WJ, Lurvink RJ, Grootscholten C, Verhoeven RHA, de Hingh IH, van Sandick JW. Synchronous peritoneal metastases of gastric cancer origin: incidence, treatment and survival of a nationwide Dutch cohort. Gastric Cancer. 2021;24(4):800–9. Salati M, Orsi G, Smyth E, Aprile G, Beretta G, De Vita F, Di Bartolomeo M, Fanotto V, Lonardi S, Morano F, et al. Gastric cancer: Translating novels concepts into clinical practice. Cancer Treat Rev. 2019;79:101889. Bonnot PE, Piessen G, Kepenekian V, Decullier E, Pocard M, Meunier B, Bereder JM, Abboud K, Marchal F, Quenet F, et al. Cytoreductive Surgery With or Without Hyperthermic Intraperitoneal Chemotherapy for Gastric Cancer With Peritoneal Metastases (CYTO-CHIP study): A Propensity Score Analysis. J Clin Oncol. 2019;37(23):2028–40. Sato Y, Mizusawa J, Katayama H, Nakamura K, Fukagawa T, Katai H, Haruta S, Yamada M, Takagi M, Tamura S, et al. Diagnosis of invasion depth in resectable advanced gastric cancer for neoadjuvant chemotherapy: An exploratory analysis of Japan clinical oncology group study: JCOG1302A. Eur J Surg Oncol. 2020;46(6):1074–9. Yu HH, Yonemura Y, Ng HJ, Lee MC, Su BC, Hsieh MC. Benefit of Neoadjuvant Laparoscopic Hyperthermic Intraperitoneal Chemotherapy and Bidirectional Chemotherapy for Patients with Gastric Cancer with Peritoneal Carcinomatosis Considering Cytoreductive Surgery. Cancers (Basel) 2023, 15(13). Kim TH, Kim IH, Kang SJ, Choi M, Kim BH, Eom BW, Kim BJ, Min BH, Choi CI, Shin CM, et al. Korean Practice Guidelines for Gastric Cancer 2022: An Evidence-based, Multidisciplinary Approach. J Gastric Cancer. 2023;23(1):3–106. Emoto S, Ishigami H, Yamashita H, Yamaguchi H, Kaisaki S, Kitayama J. Clinical significance of CA125 and CA72-4 in gastric cancer with peritoneal dissemination. Gastric Cancer. 2012;15(2):154–61. Gao H, Ji K, Bao L, Chen H, Lin C, Feng M, Tao L, Wang M. Establishment and verification of prediction model of occult peritoneal metastasis in advanced gastric cancer. World J Surg Oncol. 2023;21(1):320. Yamamoto M, Yoshinaga K, Matsuyama A, Tsutsui S, Ishida T. CEA/CA72-4 levels in peritoneal lavage fluid are predictive factors in patients with gastric carcinoma. J Cancer Res Clin Oncol. 2014;140(4):607–12. Nakamura N, Kinami S, Fujii Y, Miura S, Fujita J, Kaida D, Tomita Y, Miyata T, Fujita H, Ueda N, et al. The neutrophil/lymphocyte ratio as a predictor of peritoneal metastasis during staging laparoscopy for advanced gastric cancer: a retrospective cohort analysis. World J Surg Oncol. 2019;17(1):108. Sachlova M, Majek O, Tucek S. Prognostic value of scores based on malnutrition or systemic inflammatory response in patients with metastatic or recurrent gastric cancer. Nutr Cancer. 2014;66(8):1362–70. Nie R, Yuan S, Chen S, Chen X, Chen Y, Zhu B, Qiu H, Zhou Z, Peng J, Chen Y. Prognostic nutritional index is an independent prognostic factor for gastric cancer patients with peritoneal dissemination. Chin J Cancer Res. 2016;28(6):570–8. Toiyama Y, Yasuda H, Ohi M, Yoshiyama S, Araki T, Tanaka K, Inoue Y, Mohri Y, Kusunoki M. Clinical impact of preoperative albumin to globulin ratio in gastric cancer patients with curative intent. Am J Surg. 2017;213(1):120–6. Giandola T, Maino C, Marrapodi G, Ratti M, Ragusi M, Bigiogera V, Talei Franzesi C, Corso R, Ippolito D. Imaging in Gastric Cancer: Current Practice and Future Perspectives. Diagnostics (Basel) 2023, 13(7). Liu S, Liu S, Ji C, Zheng H, Pan X, Zhang Y, Guan W, Chen L, Guan Y, Li W, et al. Application of CT texture analysis in predicting histopathological characteristics of gastric cancers. Eur Radiol. 2017;27(12):4951–9. Kim HY, Kim YH, Yun G, Chang W, Lee YJ, Kim B. Could texture features from preoperative CT image be used for predicting occult peritoneal carcinomatosis in patients with advanced gastric cancer? PLoS ONE. 2018;13(3):e0194755. Chen Y, Xi W, Yao W, Wang L, Xu Z, Wels M, Yuan F, Yan C, Zhang H. Dual-Energy Computed Tomography-Based Radiomics to Predict Peritoneal Metastasis in Gastric Cancer. Front Oncol. 2021;11:659981. Liu S, He J, Liu S, Ji C, Guan W, Chen L, Guan Y, Yang X, Zhou Z. Radiomics analysis using contrast-enhanced CT for preoperative prediction of occult peritoneal metastasis in advanced gastric cancer. Eur Radiol. 2020;30(1):239–46. Lee IS, Lee H, Hur H, Kanda M, Yook JH, Kim BS, Woo Y, Kodera Y, Kim K, Goel A. Transcriptomic Profiling Identifies a Risk Stratification Signature for Predicting Peritoneal Recurrence and Micrometastasis in Gastric Cancer. Clin Cancer Res. 2021;27(8):2292–300. Lee JH, Son SY, Lee CM, Ahn SH, Park DJ, Kim HH. Factors predicting peritoneal recurrence in advanced gastric cancer: implication for adjuvant intraperitoneal chemotherapy. Gastric Cancer. 2014;17(3):529–36. Chen D, Lai J, Cheng J, Fu M, Lin L, Chen F, Huang R, Chen J, Lu J, Chen Y, et al. Predicting peritoneal recurrence in gastric cancer with serosal invasion using a pathomics nomogram. iScience. 2023;26(3):106246. Lee JH, Chang KK, Yoon C, Tang LH, Strong VE, Yoon SS. Lauren Histologic Type Is the Most Important Factor Associated With Pattern of Recurrence Following Resection of Gastric Adenocarcinoma. Ann Surg. 2018;267(1):105–13. Chen Y, Cai G, Jiang J, He C, Chen Y, Ding Y, Lu J, Zhao W, Yang Y, Zhang Y, et al. Proteomic profiling of gastric cancer with peritoneal metastasis identifies a protein signature associated with immune microenvironment and patient outcome. Gastric Cancer. 2023;26(4):504–16. Kaji S, Irino T, Kusuhara M, Makuuchi R, Yamakawa Y, Tokunaga M, Tanizawa Y, Bando E, Kawamura T, Kami K, et al. Metabolomic profiling of gastric cancer tissues identified potential biomarkers for predicting peritoneal recurrence. Gastric Cancer. 2020;23(5):874–83. Zhang C, Li D, Yu R, Li C, Song Y, Chen X, Fan Y, Liu Y, Qu X. Immune Landscape of Gastric Carcinoma Tumor Microenvironment Identifies a Peritoneal Relapse Relevant Immune Signature. Front Immunol. 2021;12:651033. Chen D, Liu Z, Liu W, Fu M, Jiang W, Xu S, Wang G, Chen F, Lu J, Chen H, et al. Predicting postoperative peritoneal metastasis in gastric cancer with serosal invasion using a collagen nomogram. Nat Commun. 2021;12(1):179. Tables Table 1 and 2 are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Tab.1.docx Tab.2.docx supplementarymaterials.docx 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-7336599","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":508606584,"identity":"adca0b6b-3415-4f17-a62f-2939546f654c","order_by":0,"name":"Lin Zhong","email":"","orcid":"","institution":"Department of Breast Surgery, Sichuan Provincial People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Zhong","suffix":""},{"id":508606585,"identity":"f33ff4ff-e667-49a9-abdc-5e2794691598","order_by":1,"name":"Ting Lin","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ting","middleName":"","lastName":"Lin","suffix":""},{"id":508606587,"identity":"4ffa8116-3d6e-4aee-8f76-308ae8bc797a","order_by":2,"name":"Dong Hou","email":"","orcid":"","institution":"Sun Yat-sen Memorial Hospital, Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Dong","middleName":"","lastName":"Hou","suffix":""},{"id":508606592,"identity":"505b0203-8bac-40da-856c-226bd33aa2ec","order_by":3,"name":"Hongyun Huang","email":"","orcid":"","institution":"Chinese Academy of Medical Science and Peking Union Medical College","correspondingAuthor":false,"prefix":"","firstName":"Hongyun","middleName":"","lastName":"Huang","suffix":""},{"id":508606594,"identity":"6dea46e6-01ab-4b5c-a96c-69a9d427e97f","order_by":4,"name":"Shihai Zhou","email":"","orcid":"","institution":"Zhongshan City People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Shihai","middleName":"","lastName":"Zhou","suffix":""},{"id":508606595,"identity":"7111d954-1261-4c9c-819d-6ebaf769bf98","order_by":5,"name":"Yu Lin","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Lin","suffix":""},{"id":508606596,"identity":"760fc8ce-b89a-4541-80a3-3b3b04a5e24e","order_by":6,"name":"Yue Yu","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Yu","suffix":""},{"id":508606597,"identity":"f81eabf0-12ee-44c8-8f70-1cbd9b53d3fe","order_by":7,"name":"Liangquan Liu","email":"","orcid":"","institution":"Department of Breast Surgery, Sichuan Provincial People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Liangquan","middleName":"","lastName":"Liu","suffix":""},{"id":508606598,"identity":"0ad3ab6a-926b-4c55-8b96-1e43f4d272ae","order_by":8,"name":"Jing Luo","email":"","orcid":"","institution":"Department of Breast Surgery, Sichuan Provincial People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Luo","suffix":""},{"id":508606599,"identity":"d528aebe-4f06-457f-8399-6c6bfe72db77","order_by":9,"name":"Fanghai Han","email":"","orcid":"","institution":"The Affiliated Guangdong Second Provincial General Hospital of Jinan University","correspondingAuthor":false,"prefix":"","firstName":"Fanghai","middleName":"","lastName":"Han","suffix":""},{"id":508606600,"identity":"4f5cf86d-ff9d-4c03-8f3c-1c81fab24779","order_by":10,"name":"Lang Xie","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6UlEQVRIie3PMQrCMBSA4YRAXIJdCwpe4ZVAQRB7FaHgFMRVHBQcXHqASu/RuaWDS90d1Y5WqLgK2gR0TDsK5h8eKbyPpAiZTD9YV01ACG8DdWTdJkK/JMgRSmpC2xGZLRRBzaRzyK5sPvLI7pYWdzHuU0TOl6OOsNl0GMKU4GjmQxr79cMo50L7MOFCBRnFkXDtNCY1YbSnJVbpwgReDO9ySVYtiC34qYLExiGTJGtDSheH4AMOBIdDvGeUNPyLZQn+YM+x52xz57SIl57V2ZwLHVEXyemsP99Ev65WKjkHzYsmk8n0r70Bo8U/ZeAL7UwAAAAASUVORK5CYII=","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":true,"prefix":"","firstName":"Lang","middleName":"","lastName":"Xie","suffix":""}],"badges":[],"createdAt":"2025-08-10 03:38:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7336599/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7336599/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90541584,"identity":"420cebd0-e9d7-45d8-bf1d-3f4a0f70d684","added_by":"auto","created_at":"2025-09-03 23:57:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":9121814,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlow diagram depicting participant screening, eligibility assessment, and allocation. \u003c/strong\u003eThe numbers in parentheses represent the number of patients.\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/fb9a40da3981ac8aedbb7e32.png"},{"id":90541583,"identity":"7dd9803c-0bad-4a88-99bb-74b73eecccd6","added_by":"auto","created_at":"2025-09-03 23:57:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":4246434,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRadiomics-Based diagnostic model development and evaluation. \u003c/strong\u003ea. Radiomics workflow: VOI delineation, feature extraction, and model fitting; b. ROC curves of the radiomics model in the training, testing, and validation cohorts\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/4213f67031dc4c7bf5546532.png"},{"id":90541596,"identity":"30f0d57f-451d-42da-b047-783d7545cf0b","added_by":"auto","created_at":"2025-09-03 23:57:20","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":38884654,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRepresentative diagram of Tumor-Stroma Ratio (TSR) scoring. \u003c/strong\u003eIdentify the target area using a 5× magnification field of view, then conduct TSR assessment under a 10× magnification field of view. The selected area for scoring should guarantee a peripheral distribution of tumor cells. A threshold of 50% is established as the cutoff value to categorize the samples into a high TSR group (a) and a low TSR group (b).\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/2d35f468e80187fd4aff6439.png"},{"id":90541589,"identity":"739a30cf-ed2b-425c-982b-70f65815345e","added_by":"auto","created_at":"2025-09-03 23:57:19","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":4056589,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eModel performance and validation. \u003c/strong\u003ea. Competitive risk nomogram including CA125, CA724, Borrmann, TSR and Radiomics Feature; b. Diagnostic performance of the model in the training cohort; c. Diagnostic performance of the model in the external validation cohort; d. Diagnostic performance of the model in the metachronous PM cohort; e. Calibration curve of the model in the training cohort. The blue line represents the performance of the nomogram, while the red line corrects any bias present in the nomogram. The dotted line represents the reference line, which indicates the ideal nomogram position; f. Decision curve analysis showing the decision benefit of the model in practical clinical application. The x-axis represents the threshold probability, and the y-axis measures net benefit.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/eef63572a860f6e99b3828fb.png"},{"id":92071165,"identity":"52991306-e986-497b-b487-f2964a76f04f","added_by":"auto","created_at":"2025-09-24 09:47:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":54225211,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/1653f3cb-9640-4071-8e2d-fafa24f444e4.pdf"},{"id":90542579,"identity":"81454002-f38e-4614-b4d1-06203bf43c17","added_by":"auto","created_at":"2025-09-04 00:05:19","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":26811,"visible":true,"origin":"","legend":"","description":"","filename":"Tab.1.docx","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/5c206b59f8c42b4f3a4c1aad.docx"},{"id":90541579,"identity":"8a3b819c-83eb-43a5-835c-7df4afcfd0db","added_by":"auto","created_at":"2025-09-03 23:57:19","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":25167,"visible":true,"origin":"","legend":"","description":"","filename":"Tab.2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/0db71cfe415a3ad990b1e90a.docx"},{"id":90542580,"identity":"1040c983-0db8-4618-bc23-7258fb8943f3","added_by":"auto","created_at":"2025-09-04 00:05:19","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":287245,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarymaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-7336599/v1/e32eded5b18237c62c2e261b.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Novel Multimodal Combining Radiomics and Tumor-Stroma Ratio (TSR) Improves Diagnosis of Gastric Cancer Peritoneal Metastasis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGastric cancer (GC) is one of the most common malignant tumors worldwide, posing a serious threat to public health. Although the incidence and mortality rates of GC have declined in recent years, its incidence still ranks 5th, and cancer-related mortality is 4th \u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. Due to the hidden symptoms of GC, the vast majority of patients are in the middle or late stages at the time of diagnosis \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e.The most common forms of metastasis for gastric cancer are peritoneal, liver, and lymph node metastases \u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e, with peritoneal metastasis (PM) being one of the most common in advanced stages. PM can be categorized into synchronous and metachronous PM. Approximately 30% of patients with advanced GC develop synchronous PM, and this proportion increases to 45% in patients with poorly cohesive carcinomas (PCCs). Furthermore, 40\u0026ndash;50% of GC patients still progress to PM even after undergoing standardized systemic treatment \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. Patients with PM from GC face a markedly poor prognosis, characterized by a short survival duration, frequent and severe complications, and a lack of effective palliative therapeutic options\u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eIt has been confirmed that a low Peritoneal Cancer Index (PCI), complete Cytoreductive Surgery (CRS), and comprehensive systemic therapy can effectively improve the prognosis of patients with PM of GC \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. Neoadjuvant therapies, such as intraperitoneal chemotherapy and systemic therapy, can effectively downstage patients with PM of GC, leading to the disappearance of peritoneal metastatic lesions and negative free abdominal tumor cells, thereby increasing the median survival to 21.6\u0026ndash;34.6 months \u003csup\u003e[\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Early diagnosis of PM in GC is crucial for improving patient prognosis.\u003c/p\u003e\u003cp\u003eCurrently, the diagnosis of PM in GC mainly relies on clinical signs, imaging examinations, and surgical exploration results \u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. In GC patients with PM, initial clinical manifestations frequently lack diagnostic specificity. The neoplastic process characteristically maintains clinical silence until attaining advanced disease status, at which juncture pathognomonic indicators such as refractory ascites, progressive abdominal pain, and measurable abdominal circumference expansion manifest as objectively verifiable clinical signs. \u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. CT, a routine imaging examination for GC, shows severely insufficient diagnostic sensitivity for peritoneal metastases, ranging only between 28% and 51% \u003csup\u003e[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. Positron emission tomography CT (PET-CT) is commonly used for diagnosing distant metastases of GC. However, Fluoro-2-deoxy-D-glucose PET-CT, a specific form of this imaging, can only detect 3% of occult PM in GC \u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e. Despite the National Comprehensive Cancer Network (NCCN) guidelines assigning diagnostic laparoscopy a Category 2B recommendation (indicating a lower level of evidence), this procedure still requires general anesthesia, an intervention that imposes both physiological stress and psychological distress on patients \u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. The development of diagnostic models based on routine clinical parameters for PM detection in GC patients holds significant clinical value, particularly in resource-limited settings where such cost-effective diagnostic strategies are particularly advantageous.\u003c/p\u003e\u003cp\u003eRadiomics is a novel diagnostic tool that extracts hundreds of quantitative features from medical images, combining key features into image-based biomarkers (known as radiomics signatures) to enhance diagnostic efficacy \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. Tumor-Stroma Ratio (TSR), a quantitative histopathological biomarker defined as the proportional area of tumor cells relative to stromal components within the tumor microenvironment, is characterized by its technical simplicity, rapid assessment, cost-effectiveness, and high reproducibility. Emerging evidence highlights TSR as a robust prognostic indicator, demonstrating significant associations with tumor progression, metastatic dissemination, chemotherapy resistance, and diminished overall survival. Notably, our previous study has further established TSR as an independent diagnostic predictor for PM in GC\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e, underscoring its pivotal role in clinical decision-making for advanced malignancies.\u003c/p\u003e\u003cp\u003eThis study aims to develop an integrated multimodal diagnostic model that incorporates radiomics, TSR, and routine examination results to improve the accuracy and effectiveness of PM detection in GC. The multimodal model is expected to enhance the sensitivity and specificity of PM diagnosis, while also providing more comprehensive clinical information to support decision-making, ultimately improving treatment outcomes and survival for patients with PM in GC.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003ePatients\u003c/h2\u003e\u003cp\u003eThe patients who were treated at Zhujiang Hospital, Southern Medical University, from December 2010 to March 2021, were enrolled in the study as the training cohort based on the following inclusion and exclusion criteria. Following these criteria, the patients who were treated at Zhongshan People's Hospital during the same period were enrolled as the external validation cohort. The patients who were identified as non-peritoneal metastasis (non-PM) from these two centers were combined into the metachronous PM cohort \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eInclusion criteria:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAll patients were pathologically diagnosed with GC;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAll patients underwent standard radical gastrectomy for gastric cancer (D2) and systemic therapy;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAll patients with PM were based on pathological diagnosis.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003eExclusion criteria:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003ePatients with tumor cell residues at the margins of the resected tissue in the postoperative pathology report;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003ePatients who received radiotherapy, chemotherapy, or immunotherapy before surgery;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003ePatients with a history of other malignant tumors;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003ePatients whose clinical data are seriously incomplete.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eFollow-up\u003c/h3\u003e\n\u003cp\u003eThe metachronous PM cohort consists of 648 patients from the training and external validation cohorts, all diagnosed with non-PM. The follow-up method was conducted through telephone or outpatient visits. The deadline for the follow-up was March 2021. Ultimately, 62 patients were diagnosed with PM based on CT imaging or pathological confirmation.\u003c/p\u003e\n\u003ch3\u003eHematological indicators\u003c/h3\u003e\n\u003cp\u003eAll hematological indicators included in this study are commonly used in clinical practice. Additionally, this study incorporates hematological indicators that have been found to be associated with the prognosis of GC, intended for secondary analysis. These include white blood cell count (WBC, g/L), percentage of neutrophils (Neut%), percentage of lymphocytes (Lymph%), hemoglobin (Hb, g/L), platelet count (PLT, g/L), mean platelet volume (MPV, fl), platelet distribution width (PDW, %), neutrophil count (Neu, g/L), lymphocyte count (Lymph, g/L), albumin (Alb, g/L), globulin (Glo, g/L), alpha-fetoprotein (AFP, ng/ml), carcinoembryonic antigen (CEA, \u0026micro;g/L), CA125 (kU/L), CA153 (U/ml), CA199 (kU/L), and CA724 (U/mL). The neutrophil to lymphocyte ratio (NLR) is associated with overall survival (OS) of gastric cancer patients, tumor invasion depth, and peritoneal metastasis \u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e. The prognostic nutritional index (PNI) was calculated using the formula: 1\u0026times;serum albumin level (g/L)\u0026thinsp;+\u0026thinsp;5\u0026times;absolute lymphocyte count (10^9/L)\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. Additionally, the albumin to globulin ratio (AGR, A/G) is also commonly considered[22].\u003c/p\u003e\n\u003ch3\u003eRadiomics based on CT imaging\u003c/h3\u003e\n\u003cp\u003eThe maximum diameter (Size, cm) of the tumor was measured on its largest cross-section. The T stage is determined based on the diagnoses from CT and endoscopic ultrasonography. To further enhance the diagnostic efficacy of CT for peritoneal metastasis of gastric cancer, we conducted radiomics analysis using preoperative CT imaging data. And incorporate radiomics features as independent parameters for further analysis. The detailed methods, as shown in the supplementary materials \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea \u003cb\u003eand Supplemental Fig.\u0026nbsp;1)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eGastroscopy indicators\u003c/h3\u003e\n\u003cp\u003eThe gastroscopy examination indicators included in the study include the tumor occurrence site (Site: 0\u0026thinsp;=\u0026thinsp;entire stomach; 1\u0026thinsp;=\u0026thinsp;cardia, 2\u0026thinsp;=\u0026thinsp;gastric body, 3\u0026thinsp;=\u0026thinsp;antrum), Borrmann classification, history of Helicobacter pylori infection (Hp), and T stage (according to the diagnosis from endoscopic ultrasonography).\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003ePathological indicators\u003c/h2\u003e\u003cp\u003eThe pathological indicators included in the study comprise tumor differentiation (Diff: 0\u0026thinsp;=\u0026thinsp;undifferentiated, 1\u0026thinsp;=\u0026thinsp;poorly differentiated, 2\u0026thinsp;=\u0026thinsp;moderately differentiated, 3\u0026thinsp;=\u0026thinsp;well differentiated) and HER-2 expression status (HER-2, Negative or Positive). Our previous research also confirmed that TSR (Tumor-Stroma Ratio) is related to metastatic nodules in gastric cancer (GC) \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e.Other studies have verified that TSR is closely related to the prognosis of gastric cancer \u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. Further research has found that TSR is also an important predictive indicator for peritoneal metastasis of gastric cancer. Therefore, we have included this simple, rapid, low-cost, and reproducible diagnostic indicator in our analysis. The specific research methods are as follows:\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eTSR analysis\u003c/h3\u003e\n\u003cp\u003eHematoxylin and eosin (H\u0026amp;E) sections were collected from all patients for TSR analysis. Conventionally, TSR is assessed using postoperative pathological specimens to provide prognostic insights. To evaluate the feasibility of TSR assessment via endoscopic biopsy, we conducted a comparative analysis of paired endoscopic biopsy specimens and corresponding surgically resected tissues from 35 consecutive patients.\u003c/p\u003e\u003cp\u003eInitially, a microscope with a 5\u0026times; microscope was used to pinpoint the area of deepest tumor infiltration. Subsequently, we identified over three distinct observation zones, each exhibiting both tumor and stromal components, ensuring the presence of surrounding tumor cells. The TSR was then scored using a 10\u0026times; magnification microscope by calculating the stroma's percentage within the observed field. Scores were assigned in increments of ten percent (e.g., 10%, 20%, 30%, etc.). Based on these TSR values, the samples were categorized into two groups: a Low TSR group (\u0026lt;\u0026thinsp;50%), indicating a stroma-poor tumor environment, and a High TSR group (\u0026ge;\u0026thinsp;50%), signifying a stroma-rich environment \u003cb\u003e(\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eStatistics\u003c/h3\u003e\n\u003cp\u003eIn this study, all statistical analyses were executed using R software, version 4.2.2. Comparative tables of baseline patient data for the training cohort, external validation cohort, and metachronous PM cohort were created using the 'tableone' package. The normality of continuous variables was tested with the Shapiro-Wilk test, and the Mann-Whitney U test was used for those not normally distributed. Chi-square and Fisher's exact tests were applied for comparing categorical and ordinal variables. Univariate and multivariate logistic regression analyses, crucial for assessing the influence of preoperative clinical data on the incidence of PM in GC, were conducted using the 'rms' package. A predictive nomogram was developed, and its efficacy was evaluated using ROC curves generated by the 'pROC', 'car', and 'rmda' packages to illustrate the area under the curve (AUC). Additionally, the 'rmda' package was employed for calibration curve construction, ensuring the prediction model's accuracy, while the 'car' package facilitated clinical decision curve analysis by quantifying the net benefit across various threshold probabilities, thereby assessing the nomogram's clinical utility. Statistical significance was determined at \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eCharacteristics of the Patients.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBased on the inclusion and exclusion criteria set forth, the study encompassed 813 participants: 432 patients from Zhujiang Hospital, Southern Medical University, and 381 from Zhongshan City People's Hospital. Patients from Zhujiang Hospital served as the training cohort, whereas those from Zhongshan City were designated as the external validation cohort. Since PM of GC is classified into synchronous and metachronous metastasis, we aimed to further validate the predictive efficacy of our model for the relapse of PM in GC. To achieve this, we combined patients from both centers who did not experience PM into metachronous PM cohort.\u003c/p\u003e\n\u003cp\u003eThe training cohort was classified into non-PM (321 patients) and PM (111 patients) groups. The gender distribution was 67% male and 33% female, whereas the age distribution showed that 64% were less than 65 years old and 36% were 65 or above. Significant differences were found between the non-PM and PM groups in BMI, WBC, Neut%, Lmph%, Hb, PLT, MPV, PDW, Neu, Lymph, NLR, Alb, PNI, A/G, AFP, CEA, CA125, CA199, CA724, Site, Size, T stage, Borrmann, Diff, Radiomics Features and TSR (\u003cem\u003eP\u0026nbsp;\u003c/em\u003e\u0026lt; 0.05) (\u003cstrong\u003eTab. 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe validation cohort included 327 non-PM patients and 54 PM patients. The gender distribution was the same as the training cohort, and the age breakdown was 62% under 65 years and 38% 65 or older. There were statistical differences between the two groups in indicators such as Neut%, Lmph%, PLT, Lymph, NLR, A/G, CA125, CA724, Size, T stage, Radiomics Features, and TSR (\u003cem\u003eP\u0026nbsp;\u003c/em\u003e\u0026lt; 0.05)(\u003cstrong\u003eTab. 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe metachronous PM cohort included 648 patients. The gender distribution was the same as in the above cohorts, and the age breakdown was 63% \u0026lt;65 years (n=408) and 37% ≥65 years (n=240). Based on the follow-up results, patients in the metachronous PM cohort were also divided into Non-PM and PM groups. Statistical differences were observed in the levels of WBC, Neut%, Lmph%, Neu, Lymph, NLR, Alb, PNI, CA125, CA724, T stage, Borrmann type, Diff, Radiomics Features, and TSR (\u003cem\u003eP\u0026nbsp;\u003c/em\u003e\u0026lt;0.05) (\u003cstrong\u003eTab. 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eComprehensive analysis indicates that the above factors may all be associated with the occurrence of PM in GC. Neut%, Lmph%, Lymph, NLR, CA125, CA724, T stage, Borrmann type, Radiomics Features, and TSR showed statistical differences across all three cohorts and may act as important predictors of PM in GC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRadiomics features and performance of the radiomics model.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThrough intraclass correlation coefficient (ICC) analysis (threshold \u0026gt;0.75), 626 radiomic features demonstrated measurement stability from the initial 1,316 candidates, indicating high inter-observer concordance in feature extraction. These validated features underwent LASSO regression for dimensionality reduction prior to principal component analysis (PCA), which ultimately derived 498 components accounting for 85% cumulative variance.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFollowing rigorous feature selection via ANOVA (FDR-corrected p\u0026lt;0.01) from 498 candidate features, 26 optimal radiomic features were retained for model development. The Gaussian Process (GP) classifier demonstrated robust discriminative capacity with area under the curve (AUC) values of 0.821 (95% CI: 0.785-0.857) in the training cohort, 0.719 (95% CI:0.677-0.761) in the internal test cohort, and 0.701 (95% CI:0.658-0.744) in the external validation cohort. \u003cstrong\u003eFig.2b- Fig.2d\u0026nbsp;\u003c/strong\u003eillustrate the corresponding receiver operating characteristic (ROC) curves across all cohorts.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTSR\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003escore\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll H\u0026amp;E-stained histological sections were initially evaluated in a blinded manner by two independent pathologists. In cases of discrepant interpretations, consensus was reached through joint microscopic review. Inter-observer agreement was quantified using Cohen’s kappa coefficient (κ = 0.87, 95% CI: 0.79–0.94), indicating excellent diagnostic concordance.\u003c/p\u003e\n\u003cp\u003eAmong 35 matched endoscopic–surgical specimen pairs, TSR assessment demonstrated substantial agreement using a 50% stromal cutoff (AC1 = 0.722; overall concordance rate: 85.7%), supporting the reliability of endoscopic specimens for TSR evaluation.\u003c/p\u003e\n\u003cp\u003eIn the training cohort, the non-PM group (n = 321) exhibited a higher proportion of TSR-low cases (64.8%, 208/321) compared to TSR-high cases (35.2%, 113/321). Conversely, the PM group (n = 111) showed a predominance of TSR-high cases (64.9%, 72/111) over TSR-low cases (35.1%, 39/111), with statistically significant stratification (χ² = 5.89, \u003cem\u003ep\u0026nbsp;\u003c/em\u003e= 0.02).\u003c/p\u003e\n\u003cp\u003eIn the external validation cohort, a similar distribution pattern was observed: TSR-low cases comprised 55.4% (181/327) of the non-PM group, while only 33.3% (18/54) of the PM group, maintaining statistical significance (χ² = 4.92, \u003cem\u003ep\u003c/em\u003e = 0.03).\u003c/p\u003e\n\u003cp\u003eIn the metachronous PM cohort, pooled data from 648 patients revealed a significantly higher proportion of TSR-high cases in PM patients (62.9%, 39/62) compared to non-PM controls (40.4%, 237/586), demonstrating robust discriminative capacity (χ² = 10.24, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRisk factors associated with PM in GC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo identify risk factors associated with PM in GC, we conducted a univariate logistic regression analysis incorporating parameters that exhibited statistical differences between the non-PM group and the PM group. These parameters included BMI, WBC, Neut%, Lymph%, Hb, PLT, MPV, PDW, Neu, Lymph, NLR, Alb, PNI, A/G, AFP, CEA, CA125, CA199, CA724, Site, Size, T stage, Borrmann classification, Diff, Radiomics Features, and TSR. The analysis revealed significant statistical differences in parameters such as WBC, Neut%, Lymph%, PLT, Alb, PNI, CA125, CA724, tumor Size, T stage, Borrmann classification, Radiomics Features, and TSR (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01) (\u003cstrong\u003eTab. 2\u003c/strong\u003e).Further, through multivariate regression analysis, we found that that CA125 (OR = 1.025, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), CA724 (OR = 1.026, \u003cem\u003eP\u003c/em\u003e = 0.002), Borrmann (OR = 10.191, \u003cem\u003eP\u003c/em\u003e = 0.003), Radiomics Features (OR = 207.221, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), and TSR (OR = 2.351, \u003cem\u003eP\u003c/em\u003e = 0.003) have a significant correlation with the incidence of PM in GC (\u003cstrong\u003eTab. 2\u003c/strong\u003e). Therefore, we consider CA125, CA724, Borrmann, Radiomics Features, and TSR to be independent influencing factors for PM in GC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiagnostic model construction based on the preoperative data.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccording to the results above, we constructed a risk score model for all patients, resulting in the following predictive equation:\u003c/p\u003e\n\u003cp\u003eRisk Score = 0.0217* CA125 + 0.0286 * CA724 +2.0504 * Borrmann + 5.0457 * Radiomics Features + 0.8119 * (TSR≥50%) - 6.6743.\u003c/p\u003e\n\u003cp\u003eConsequently, the probability of PM can be calculated by the formula:\u003c/p\u003e\n\u003cp\u003eProbability of PM = 1 / (1 + e^ (- Risk Score))\u003c/p\u003e\n\u003cp\u003eTo enhance the intuitive understanding of our model, we developed a Nomogram informed by pertinent findings. Illustrated in\u003cstrong\u003e\u0026nbsp;Fig 4a\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e the process involves aggregating the values assigned to each indicator, culminating in the Total Points. Correspondingly, the value located on the bottom horizontal axis of the Nomogram, aligned with these Total Points, indicates the risk probability of PM in GC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel performance and validation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe performance of this individualized nomogram model is evaluated by drawing the ROC curve and calculating the area under the curve. The area under the ROC curve for the training set is presented in \u003cstrong\u003eFig 4b\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e with an area of 0.874 (95% CI: 0.839-0.909). The AUC for the external validation cohort is 0.797 (95% CI: 0.733-0.862) (\u003cstrong\u003eFig 4c\u003c/strong\u003e), and for the prediction cohort is 0.784 (95% CI: 0.724-0.845) (\u003cstrong\u003eFig 4d\u003c/strong\u003e\u003cstrong\u003e)\u003c/strong\u003e. The calibration curve reveals good consistency between the nomogram and actual observations \u003cstrong\u003e(Fig 4e and Supplemental Fig. 1).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe maximum Youden index of 0.612 of the ROC curves of the nomogram was selected as the optimal cutoff value in the training cohort, and patients were divided into high-risk and low-risk groups. We found that the sensitivity, specificity, accuracy of the nomogram in the training cohort were 72.0 %, 89.2%, 80.6% respectively \u003cstrong\u003e(Fig 4b\u003c/strong\u003e). In the validation cohort, the sensitivity was 79.8 %, specificity was 68.5%, and the accuracy was74.2%\u003cstrong\u003e\u0026nbsp;(Fig 4c)\u003c/strong\u003e. In the metachronous PM cohort , the sensitivity was 66.0 %, the specificity was70.6%, and the accuracy was 68.3% \u003cstrong\u003e(Fig 4d)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical usefulness\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the clinical usefulness of this model, a clinical decision curve was plotted in the training set cohort. This Decision curve included the model and independent factors significantly associated with PM in GC, and quantified their clinical net benefit. As shown in \u003cstrong\u003eFig 4f\u003c/strong\u003e, the net benefit of this model is higher than that of all or each independent factor. Therefore, intervening with patients based on this model yields a good net benefit, and using this model for the diagnosis and treatment of PM in GC has great clinical benefit.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn recent years, the incidence of GC has shown a downward trend, but at the same time, the proportion of patients with simultaneous PM is gradually increasing \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e. Despite continuous improvements in treatment, the prognosis for GC patients remains unsatisfactory, especially for those with PM, which is even more concerning \u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e. Although there is some controversy over intraperitoneal chemotherapy in the treatment of PM in GC, it is a consensus that the completeness of CRS and a low PCI index are key factors in the prognosis of patients with PM \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e.The International Peritoneal Surface Oncology Group (PSOGI) recommends an initial pathological diagnosis through diagnostic laparoscopy, along with an assessment of the PCI. They advocate for conducting intraoperative HIPEC treatment following neoadjuvant chemotherapy. This is complemented by real-time combined CRS-HIPEC surgery, subsequent intraperitoneal chemotherapy, and sequential systemic chemotherapy \u003csup\u003e[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/sup\u003e. For patients with low PCI undergoing neoadjuvant intraperitoneal systemic chemotherapy (NIPS), over 50% become cytologically negative, and 70% undergo complete CRS \u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. Therefore, timely and accurate diagnosis, along with treatment at experienced medical centers, significantly improves the prognosis of patients with PM in GC.\u003c/p\u003e\u003cp\u003eAccording to NCCN guidelines, diagnostic laparoscopy combined with peritoneal lavage fluid examination is categorized as a 2B recommendation, acknowledging significant drawbacks of this invasive diagnostic method \u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. The ideal diagnostic approach should possess attributes such as high sensitivity, high specificity, accuracy, simplicity, non-invasiveness, and affordability. In this study, based on routine preoperative examinations for GC, deep insights from various examination results were explored. Using data from hematological, CT scans, gastroscopic, and pathological examinations of gastroscopic tissue biopsies, a model was constructed incorporating levels of CA125, CA724, Radiomics Features, Boman, and TSR. This model can effectively diagnose the occurrence of PM in GC (AUC\u0026thinsp;=\u0026thinsp;0.874) and predict the metachronous PM in GC (AUC\u0026thinsp;=\u0026thinsp;0.784). Moreover, the model is characterized by its simplicity, good repeatability, ease of implementation, and cost-effectiveness.\u003c/p\u003e\u003cp\u003eHematological examination is the most common clinical test. CA125 and CA724 are commonly used tumor markers for GC, but due to their low sensitivity and specificity, their diagnostic efficacy is limited \u003csup\u003e[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e. The results of this study are consistent with many others, confirming that CA125 and CA724 are related to the occurrence of PM in GC \u003csup\u003e[\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e. NLR, PNI, and A/G have been proven to be new predictive factors for PM of GC. Hideaki Shimada et al. found that the NLR reflects the systemic inflammatory state and is an independent risk factor for poor prognosis in GC patients \u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e. NLR\u0026thinsp;\u0026gt;\u0026thinsp;3.5 is a risk factor for PM in GC \u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]\u003c/sup\u003e. In this study, significant differences in NLR were observed in all three cohorts (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) but regression analysis found it could not serve as an effective predictor for the occurrence of PM. The PNI is also a useful tool for assessing the prognosis of patients with metastatic GC \u003csup\u003e[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]\u003c/sup\u003e, and has been proven to predict PM of GC \u003csup\u003e[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e. Univariate retrospective analysis also confirmed that PNI is an independent risk factor for PM of GC, but this predictive factor was not included in the multivariate retrospective analysis. The A/G mainly serves as a clinical marker for multiple myeloma or other immunoproliferative diseases, and studies have found that A/G is a new independent indicator for predicting early recurrence of curative gastric cancer \u003csup\u003e[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. In the training cohort, there is a significant difference in the presence of A/G between the PM group and the non-PM group (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05); however, A/G is not an independent influencing factor in the logistic regression analysis (Univariate: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.05; Multivariate: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.507). We attribute the primary reason for this phenomenon to the insufficient predictive efficacy of these markers.\u003c/p\u003e\u003cp\u003eCT is the most commonly used imaging examination for GC. In this study, it was also confirmed that CT-based radiomics has significant efficacy in diagnosing PM of GC. The sensitivity of CT scans is severely inadequate for diagnosing PM of GC, especially for nodules smaller than 5mm \u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e. Radiomics is a new diagnostic tool that extracts hundreds of quantitative features from medical images and combines key features into image-based biomarkers (known as radiomic signatures) for cancer diagnosis \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. Previous studies have constructed predictive models for PM of GC based on radiomics of the primary tumor \u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e, omental features\u003csup\u003e[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]\u003c/sup\u003e, and a combination of primary tumors with adjacent peritoneum \u003csup\u003e[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]\u003c/sup\u003e. However, previous studies often selected a plane to outline the Region of Interest (ROI) on the 2D level for CT image feature extraction, leading to uncertainty in the final features included. Although efforts to increase the number of reading planes have been made, they still have limitations such as low repeatability and one-sidedness. In our study, we outlined the Volume of Interest (VOI) within the possible range of metastasis and extracted corresponding texture information from a 3D spatial perspective. This helps improve image characterization properties in consideration of factors like the heterogeneity of gastric cancer. The comprehensive predictive model constructed with radiomic features in this study has significantly enhanced the efficacy of predicting PM in GC compared to previous studies that solely relied on CT radiomics \u003csup\u003e[\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]\u003c/sup\u003e. Additionally, all the analysis tools used in this study are open-source and accessible online, increasing the applicability of the model.\u003c/p\u003e\u003cp\u003eGastroscopy is the primary method for the screening and diagnosis of GC, allowing for direct observation of changes in the gastric mucosa and facilitating pathological sampling. Borrmann type 3 and 4 GC often exhibit strong invasiveness, making the Borrmann classification a significant risk factor for PM in GC. Our analysis results are consistent with previous studies, confirming the significance of Borrmann classification in predicting the risk of PM in GC \u003csup\u003e[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003ePathomics captures a vast amount of data from digital pathological images to generate quantitative features that describe various phenotypes of tissue samples. These data are analyzed to identify key indicators for diagnosis or prognosis prediction \u003csup\u003e[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]\u003c/sup\u003e. However, pathomics requires digitization of pathological images, necessitating specialized equipment and time-consuming scanning processes, thereby increasing time and economic costs, as well as specific analysis tools. In our study, we incorporated TSR on top of routine pathological indicators. TSR, as an easy, fast, cost-neutral, and easily repeatable pathological parameter, can intuitively describe the relationship between tumor cells and tumor-associated stroma. Our previous studies have demonstrated that stroma-rich metastatic nodules are independent determinants of poor prognosis in gastric cancer (GC) \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e. We further discovered that TSR is closely related to the occurrence of PM in GC. Although TSR is primarily assessed from postoperative pathological specimens, we found that TSR grading of tissues obtained via endoscopy is highly consistent with postoperative histopathology, as indicated by a Kappa value of 0.706, an AC1 of 0.722, and an agreement rate of 85.7%. Therefore, TSR can be entirely obtained from endoscopic biopsy tissue pathology as a prognostic indicator. Lauren diffuse type GC is also a high-risk factor for PM \u003csup\u003e[\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]\u003c/sup\u003e, but in our study, effective analysis was hindered due to severe data missing for Lauren classification. However, Lauren diffuse type is characterized by poorly adhesive cells diffusely infiltrating gastric wall structures, with single cells encapsulated in fibroproliferative stroma, showing few or no gland formations, features that bear significant similarities to high TSR.\u003c/p\u003e\u003cp\u003eAs research progresses, there has been a paradigm shift toward molecular-level prediction of PM in GC. Lee IS, et al. have found that a transcriptomic profile of six genes (ZBTB1, CAVIN2, CHCHD, LTBP3, SLITPK6, STT3B) can effectively predict the occurrence of PM in GC (AUC\u0026thinsp;=\u0026thinsp;0.72). When combined with Borrmann type and T stage in the model, the predictive power significantly increases (AUC\u0026thinsp;=\u0026thinsp;0.84) \u003csup\u003e[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]\u003c/sup\u003e. Yanyan Chen, et al. identified through proteomics analysis that 10 proteins (DUOXA2, ITGA7, LIMS1, MSRB3, PLCB1, RAB6B, SEMA3C, SMTN, TADA1, TBC1D14) can effectively predict the occurrence of PM (ROC\u0026thinsp;=\u0026thinsp;0.83) \u003csup\u003e[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]\u003c/sup\u003e. Other research teams have focused on the tumor microenvironment (TME), predicting gastric cancer peritoneal metastasis from aspects like metabolism \u003csup\u003e[\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]\u003c/sup\u003e, immune microenvironment \u003csup\u003e[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]\u003c/sup\u003e, and tumor stroma characteristics \u003csup\u003e[\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]\u003c/sup\u003e. However, the cost and complexity of gene expression analysis hinder its widespread clinical application, especially in developing countries.\u003c/p\u003e\u003cp\u003eThe integration of multi-omics, including pathomics, radiomics, transcriptomics, and proteomics, is beneficial for the diagnosis and determination of treatment strategies for diseases. Such a synergistic combination is likely to enhance the sensitivity, specificity, and accuracy of diagnosing PM in GC. We anticipate that this multimodal approach will improve diagnostic outcomes. However, it will undoubtedly increase the economic burden on patients and the workload of clinical physicians, which is a significant consideration, especially for developing countries.\u003c/p\u003e\u003cp\u003eThe model constructed in our study is not only effective in diagnosing synchronous PM but also in predicting metachronous PM, yielding significant clinical benefits. However, there are certain limitations to our research. First, our data were derived solely from two medical centers, which necessitates further validation with more external data, especially from Western countries. Second, as a retrospective study, there is an inherent risk of missing clinical feature data. Despite using Monte Carlo simulations to impute missing data, biases are unavoidable. Third, to ensure the adequate sample size of the data, considering the good consistency of TSR between endoscopic and postoperative tissue, we performed the evaluation of TSR in postoperative tissue when TSR failed to be evaluated in endoscopic tissue. Hence, comprehensive preoperative data are essential to validate this result. Fourth, the range of preoperative examination parameters covered in this study is limited. Therefore, it is necessary to enrich the examination parameters and conduct prospective, multicenter studies based on our research.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe model, integrating radiomics, TSR, and routine preoperative examination results, exhibits robust diagnostic performance for detecting PM in GC, thereby informing clinical treatment strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eConflict of interest\u003c/h2\u003e\u003cp\u003eThe authors declare that they have no conflict of interest.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003ch2\u003eEthics Statement\u003c/h2\u003e\u003cp\u003e The study was approved by the Medical Ethics and Institutional Committee of Zhujiang Hospital of Southern Medical University and Zhongshan People's Hospital (Approval No. 2024-KY-417-01). All research procedures involving human participants conformed to the Declaration of Helsinki. Given the retrospective nature of the study and the use of anonymized clinical data without direct patient contact or intervention, the requirement for informed consent was waived by the Medical Ethics and Institutional Committee in accordance with relevant national regulations.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eThis research was supported by China Postdoctoral Science Foundation (2022M723656).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eLin Zhong: Conceptualization; formal analysis; writing \u0026ndash; original draft. Ting Lin: Data curation; investigation; methodology; writing \u0026ndash; original draft. Dong Hou: Formal Analysis; methodology. Hongyun Huang: Data Curation. Shihai Zhou: Formal Analysis. Yu Lin: Investigation. Yue Yu: Resources. Liangquan Liu: data curation. Jing luo: writing \u0026ndash; review and editing. Fanghai Han: Methodology; writing \u0026ndash; review and editing. Lang Xie: Conceptualization; data curation; investigation; methodology; validation; writing \u0026ndash; review and editing.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets generated and/or analysed during the current study have been deposited in the Science Data Bank (SciDB) repository ( DOI: 10.57760/sciencedb.29034).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209\u0026ndash;49.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLayke JC, Lopez PP. Gastric cancer: diagnosis and treatment options. Am Fam Physician. 2004;69(5):1133\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang C, Yang J, Chen Y, Jiang F, Liao H, Liu X, Wang Y, Kong G, Zhang X, Li J, et al. miRNAs derived from plasma small extracellular vesicles predict organo-tropic metastasis of gastric cancer. Gastric Cancer. 2022;25(2):360\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBonnot PE, Lintis A, Mercier F, Benzerdjeb N, Passot G, Pocard M, Meunier B, Bereder JM, Abboud K, Marchal F, et al. Prognosis of poorly cohesive gastric cancer after complete cytoreductive surgery with or without hyperthermic intraperitoneal chemotherapy (CYTO-CHIP study). Br J Surg. 2021;108(10):1225\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGong X, Hou D, Zhou S, Tan J, Zhong G, Yang B, Xie L, Han F, Zhong L. FMO family may serve as novel marker and potential therapeutic target for the peritoneal metastasis in gastric cancer. Front Oncol. 2023;13:1144775.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang R, Song S, Harada K, Ghazanfari Amlashi F, Badgwell B, Pizzi MP, Xu Y, Zhao W, Dong X, Jin J, et al. Multiplex profiling of peritoneal metastases from gastric adenocarcinoma identified novel targets and molecular subtypes that predict treatment response. Gut. 2020;69(1):18\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCoccolini F, Catena F, Glehen O, Yonemura Y, Sugarbaker PH, Piso P, Montori G, Ansaloni L. Complete versus incomplete cytoreduction in peritoneal carcinosis from gastric cancer, with consideration to PCI cut-off. Systematic review and meta-analysis. Eur J Surg Oncol. 2015;41(7):911\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYonemura Y, Canbay E, Li Y, Coccolini F, Glehen O, Sugarbaker PH, Morris D, Moran B, Gonzaletz-Moreno S, Deraco M, et al. A comprehensive treatment for peritoneal metastases from gastric cancer with curative intent. Eur J Surg Oncol. 2016;42(8):1123\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRudloff U, Langan RC, Mullinax JE, Beane JD, Steinberg SM, Beresnev T, Webb CC, Walker M, Toomey MA, Schrump D, et al. Impact of maximal cytoreductive surgery plus regional heated intraperitoneal chemotherapy (HIPEC) on outcome of patients with peritoneal carcinomatosis of gastric origin: results of the GYMSSA trial. J Surg Oncol. 2014;110(3):275\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKitayama J, Ishigami H, Yamaguchi H, Yamashita H, Emoto S, Kaisaki S, Watanabe T. Salvage gastrectomy after intravenous and intraperitoneal paclitaxel (PTX) administration with oral S-1 for peritoneal dissemination of advanced gastric cancer with malignant ascites. Ann Surg Oncol. 2014;21(2):539\u0026ndash;46.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChan DY, Syn NL, Yap R, Phua JN, Soh TI, Chee CE, Nga ME, Shabbir A, So JB, Yong WP. Conversion Surgery Post-Intraperitoneal Paclitaxel and Systemic Chemotherapy for Gastric Cancer Carcinomatosis Peritonei. Are We Ready? J Gastrointest Surg. 2017;21(3):425\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIshigami H, Yamaguchi H, Yamashita H, Asakage M, Kitayama J. Surgery after intraperitoneal and systemic chemotherapy for gastric cancer with peritoneal metastasis or positive peritoneal cytology findings. Gastric Cancer. 2017;20(Suppl 1):128\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGwee YX, Chia DKA, So J, Ceelen W, Yong WP, Tan P, Ong CJ, Sundar R. Integration of Genomic Biology Into Therapeutic Strategies of Gastric Cancer Peritoneal Metastasis. J Clin Oncol. 2022;40(24):2830.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYajima K, Kanda T, Ohashi M, Wakai T, Nakagawa S, Sasamoto R, Hatakeyama K. Clinical and diagnostic significance of preoperative computed tomography findings of ascites in patients with advanced gastric cancer. Am J Surg. 2006;192(2):185\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim SJ, Kim HH, Kim YH, Hwang SH, Lee HS, Park DJ, Kim SY, Lee KH. Peritoneal metastasis: detection with 16- or 64-detector row CT in patients undergoing surgery for gastric cancer. Radiology. 2009;253(2):407\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGertsen EC, Brenkman HJF, van Hillegersberg R, van Sandick JW, van Berge Henegouwen MI, Gisbertz SS, Luyer MDP, Nieuwenhuijzen GAP, van Lanschot JJB, Lagarde SM, et al. 18F-Fludeoxyglucose-Positron Emission Tomography/Computed Tomography and Laparoscopy for Staging of Locally Advanced Gastric Cancer: A Multicenter Prospective Dutch Cohort Study (PLASTIC). JAMA Surg. 2021;156(12):e215340.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAjani JA, D'Amico TA, Bentrem DJ, Chao J, Cooke D, Corvera C, Das P, Enzinger PC, Enzler T, Fanta P, et al. Gastric Cancer, Version 2.2022, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2022;20(2):167\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLimkin EJ, Sun R, Dercle L, Zacharaki EI, Robert C, Reuze S, Schernberg A, Paragios N, Deutsch E, Ferte C. Promises and challenges for the implementation of computational medical imaging (radiomics) in oncology. Ann Oncol. 2017;28(6):1191\u0026ndash;206.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhong L, Huang H, Hou D, Zhou S, Lin Y, Yu Y, Yu J, Han F, Xie L. Tumor-Stroma Ratio is a Critical Indicator of Peritoneal Metastasis in Gastric Cancer. Clin Exp Gastroenterol. 2025;18:11\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShimada H, Takiguchi N, Kainuma O, Soda H, Ikeda A, Cho A, Miyazaki A, Gunji H, Yamamoto H, Nagata M. High preoperative neutrophil-lymphocyte ratio predicts poor survival in patients with gastric cancer. Gastric Cancer. 2010;13(3):170\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePan Y, Ma Y, Dai G. The Prognostic Value of the Prognostic Nutritional Index in Patients with Advanced or Metastatic Gastric Cancer Treated with Immunotherapy. \u003cem\u003eNutrients\u003c/em\u003e 2023, 15(19).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWei C, Yu Z, Wang G, Zhou Y, Tian L. Low Pretreatment Albumin-to-Globulin Ratio Predicts Poor Prognosis in Gastric Cancer: Insight From a Meta-Analysis. Front Oncol. 2020;10:623046.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang J, Yang B, Tan J, Zhou S, Chen Z, Zhong G, Gao H, Zhu J, Zeng J, Zhong L, et al. Gastric cancer nodal tumour-stroma ratios influence prognosis. Br J Surg. 2020;107(13):1713\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAurello P, Berardi G, Giulitti D, Palumbo A, Tierno SM, Nigri G, D'Angelo F, Pilozzi E, Ramacciato G. Tumor-Stroma Ratio is an independent predictor for overall survival and disease free survival in gastric cancer patients. Surgeon. 2017;15(6):329\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim EY, Abdul-Ghafar J, Chong Y, Yim K. Calculated Tumor-Associated Neutrophils Are Associated with the Tumor-Stroma Ratio and Predict a Poor Prognosis in Advanced Gastric Cancer. Biomedicines 2022, 10(3).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKoemans WJ, Lurvink RJ, Grootscholten C, Verhoeven RHA, de Hingh IH, van Sandick JW. Synchronous peritoneal metastases of gastric cancer origin: incidence, treatment and survival of a nationwide Dutch cohort. Gastric Cancer. 2021;24(4):800\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSalati M, Orsi G, Smyth E, Aprile G, Beretta G, De Vita F, Di Bartolomeo M, Fanotto V, Lonardi S, Morano F, et al. Gastric cancer: Translating novels concepts into clinical practice. Cancer Treat Rev. 2019;79:101889.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBonnot PE, Piessen G, Kepenekian V, Decullier E, Pocard M, Meunier B, Bereder JM, Abboud K, Marchal F, Quenet F, et al. Cytoreductive Surgery With or Without Hyperthermic Intraperitoneal Chemotherapy for Gastric Cancer With Peritoneal Metastases (CYTO-CHIP study): A Propensity Score Analysis. J Clin Oncol. 2019;37(23):2028\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSato Y, Mizusawa J, Katayama H, Nakamura K, Fukagawa T, Katai H, Haruta S, Yamada M, Takagi M, Tamura S, et al. Diagnosis of invasion depth in resectable advanced gastric cancer for neoadjuvant chemotherapy: An exploratory analysis of Japan clinical oncology group study: JCOG1302A. Eur J Surg Oncol. 2020;46(6):1074\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYu HH, Yonemura Y, Ng HJ, Lee MC, Su BC, Hsieh MC. Benefit of Neoadjuvant Laparoscopic Hyperthermic Intraperitoneal Chemotherapy and Bidirectional Chemotherapy for Patients with Gastric Cancer with Peritoneal Carcinomatosis Considering Cytoreductive Surgery. \u003cem\u003eCancers (Basel)\u003c/em\u003e 2023, 15(13).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim TH, Kim IH, Kang SJ, Choi M, Kim BH, Eom BW, Kim BJ, Min BH, Choi CI, Shin CM, et al. Korean Practice Guidelines for Gastric Cancer 2022: An Evidence-based, Multidisciplinary Approach. J Gastric Cancer. 2023;23(1):3\u0026ndash;106.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEmoto S, Ishigami H, Yamashita H, Yamaguchi H, Kaisaki S, Kitayama J. Clinical significance of CA125 and CA72-4 in gastric cancer with peritoneal dissemination. Gastric Cancer. 2012;15(2):154\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGao H, Ji K, Bao L, Chen H, Lin C, Feng M, Tao L, Wang M. Establishment and verification of prediction model of occult peritoneal metastasis in advanced gastric cancer. World J Surg Oncol. 2023;21(1):320.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYamamoto M, Yoshinaga K, Matsuyama A, Tsutsui S, Ishida T. CEA/CA72-4 levels in peritoneal lavage fluid are predictive factors in patients with gastric carcinoma. J Cancer Res Clin Oncol. 2014;140(4):607\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNakamura N, Kinami S, Fujii Y, Miura S, Fujita J, Kaida D, Tomita Y, Miyata T, Fujita H, Ueda N, et al. The neutrophil/lymphocyte ratio as a predictor of peritoneal metastasis during staging laparoscopy for advanced gastric cancer: a retrospective cohort analysis. World J Surg Oncol. 2019;17(1):108.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSachlova M, Majek O, Tucek S. Prognostic value of scores based on malnutrition or systemic inflammatory response in patients with metastatic or recurrent gastric cancer. Nutr Cancer. 2014;66(8):1362\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNie R, Yuan S, Chen S, Chen X, Chen Y, Zhu B, Qiu H, Zhou Z, Peng J, Chen Y. Prognostic nutritional index is an independent prognostic factor for gastric cancer patients with peritoneal dissemination. Chin J Cancer Res. 2016;28(6):570\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eToiyama Y, Yasuda H, Ohi M, Yoshiyama S, Araki T, Tanaka K, Inoue Y, Mohri Y, Kusunoki M. Clinical impact of preoperative albumin to globulin ratio in gastric cancer patients with curative intent. Am J Surg. 2017;213(1):120\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGiandola T, Maino C, Marrapodi G, Ratti M, Ragusi M, Bigiogera V, Talei Franzesi C, Corso R, Ippolito D. Imaging in Gastric Cancer: Current Practice and Future Perspectives. \u003cem\u003eDiagnostics (Basel)\u003c/em\u003e 2023, 13(7).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu S, Liu S, Ji C, Zheng H, Pan X, Zhang Y, Guan W, Chen L, Guan Y, Li W, et al. Application of CT texture analysis in predicting histopathological characteristics of gastric cancers. Eur Radiol. 2017;27(12):4951\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim HY, Kim YH, Yun G, Chang W, Lee YJ, Kim B. Could texture features from preoperative CT image be used for predicting occult peritoneal carcinomatosis in patients with advanced gastric cancer? PLoS ONE. 2018;13(3):e0194755.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen Y, Xi W, Yao W, Wang L, Xu Z, Wels M, Yuan F, Yan C, Zhang H. Dual-Energy Computed Tomography-Based Radiomics to Predict Peritoneal Metastasis in Gastric Cancer. Front Oncol. 2021;11:659981.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu S, He J, Liu S, Ji C, Guan W, Chen L, Guan Y, Yang X, Zhou Z. Radiomics analysis using contrast-enhanced CT for preoperative prediction of occult peritoneal metastasis in advanced gastric cancer. Eur Radiol. 2020;30(1):239\u0026ndash;46.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee IS, Lee H, Hur H, Kanda M, Yook JH, Kim BS, Woo Y, Kodera Y, Kim K, Goel A. Transcriptomic Profiling Identifies a Risk Stratification Signature for Predicting Peritoneal Recurrence and Micrometastasis in Gastric Cancer. Clin Cancer Res. 2021;27(8):2292\u0026ndash;300.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee JH, Son SY, Lee CM, Ahn SH, Park DJ, Kim HH. Factors predicting peritoneal recurrence in advanced gastric cancer: implication for adjuvant intraperitoneal chemotherapy. Gastric Cancer. 2014;17(3):529\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen D, Lai J, Cheng J, Fu M, Lin L, Chen F, Huang R, Chen J, Lu J, Chen Y, et al. Predicting peritoneal recurrence in gastric cancer with serosal invasion using a pathomics nomogram. iScience. 2023;26(3):106246.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee JH, Chang KK, Yoon C, Tang LH, Strong VE, Yoon SS. Lauren Histologic Type Is the Most Important Factor Associated With Pattern of Recurrence Following Resection of Gastric Adenocarcinoma. Ann Surg. 2018;267(1):105\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen Y, Cai G, Jiang J, He C, Chen Y, Ding Y, Lu J, Zhao W, Yang Y, Zhang Y, et al. Proteomic profiling of gastric cancer with peritoneal metastasis identifies a protein signature associated with immune microenvironment and patient outcome. Gastric Cancer. 2023;26(4):504\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKaji S, Irino T, Kusuhara M, Makuuchi R, Yamakawa Y, Tokunaga M, Tanizawa Y, Bando E, Kawamura T, Kami K, et al. Metabolomic profiling of gastric cancer tissues identified potential biomarkers for predicting peritoneal recurrence. Gastric Cancer. 2020;23(5):874\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang C, Li D, Yu R, Li C, Song Y, Chen X, Fan Y, Liu Y, Qu X. Immune Landscape of Gastric Carcinoma Tumor Microenvironment Identifies a Peritoneal Relapse Relevant Immune Signature. Front Immunol. 2021;12:651033.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen D, Liu Z, Liu W, Fu M, Jiang W, Xu S, Wang G, Chen F, Lu J, Chen H, et al. Predicting postoperative peritoneal metastasis in gastric cancer with serosal invasion using a collagen nomogram. Nat Commun. 2021;12(1):179.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1 and 2 are available in the Supplementary Files section.\u003c/p\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":"Radiomics, Tumor-Stroma Ratio (TSR), Diagnostic model","lastPublishedDoi":"10.21203/rs.3.rs-7336599/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7336599/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePeritoneal metastasis (PM) is the most common form of metastasis in gastric cancer (GC), frequently leading to severe complications and a significantly poor prognosis. Prompt and early diagnosis of PM in GC is crucial. However, diagnostic laparoscopy and CT scans, while being the primary methods for identifying PM in GC, have notable limitations, such as being invasive and having low sensitivity. Therefore, developing a diagnostic model for PM in GC based on routine examination results holds substantial importance.\u003c/p\u003e\u003cp\u003eIn this retrospective study, we enrolled 813 patients from two medical centers and developed a robust diagnostic model by integrating various routine examination results, including CT scans, endoscopy, pathology, and hematological tests. To further explore the potential significance of various examination results, we conducted radiomic analysis of CT images, analyzed histopathological results via the Tumor-Stroma Ratio (TSR), and examined hematological results through parameters such as the Prognostic Nutritional Index (PNI), Neutrophil to Lymphocyte Ratio (NLR), and Albumin to Globulin Ratio (AGR). A novel diagnostic model, incorporating CA125, CA724, Borrmann classification, radiomics features, and the TSR, was successfully constructed.\u003c/p\u003e\u003cp\u003eThis model demonstrated strong performance in diagnosing synchronous PM (AUC\u0026thinsp;=\u0026thinsp;0.874) and predicting metachronous (AUC\u0026thinsp;=\u0026thinsp;0.784) PM in GC. To facilitate clinical application, a nomogram was constructed. Consequently, the study presents a novel and comprehensive diagnostic model for PM in GC patients, offering significant promise for clinical applicability based on routine examination results.\u003c/p\u003e","manuscriptTitle":"A Novel Multimodal Combining Radiomics and Tumor-Stroma Ratio (TSR) Improves Diagnosis of Gastric Cancer Peritoneal Metastasis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-03 23:57:14","doi":"10.21203/rs.3.rs-7336599/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":"c1abd0e1-22ba-4256-b860-cd3bd95dabe6","owner":[],"postedDate":"September 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-09-24T09:38:16+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-03 23:57:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7336599","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7336599","identity":"rs-7336599","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}