Feasibility analysis of metabolic parameters based on baseline 18 F-FDG PET/CT to predict heterogeneity and recurrence of diffuse large B-cell lymphoma

Feasibility analysis of metabolic parameters based on baseline 18 F-FDG PET/CT to predict heterogeneity and recurrence of diffuse large B-cell lymphoma | 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 Feasibility analysis of metabolic parameters based on baseline 18 F-FDG PET/CT to predict heterogeneity and recurrence of diffuse large B-cell lymphoma Fan Ge, Tingting Wu, Xinyue Yang, Mengye Peng, Chen Yang, Kezheng Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6070367/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Jun, 2025 Read the published version in Annals of Hematology → Version 1 posted 6 You are reading this latest preprint version Abstract Objective This study aimed to evaluate the predictive value of intra-tumoral 18 F-FDG metabolic heterogeneity in patients with diffuse large B cell lymphoma (DLBCL) in terms of survival. Methods We retrospectively included 245 patients with DLBCL who underwent 18 F-FDG PET/CT prior to treatment and analyzed using total metabolic tumor volume (TMTV) and total lesion glycolysis (TLG) as metabolic volume parameters. The linear regression slopes of TMTV and TLG were calculated according to different percentages of SUV thresholds (i.e., 40%, 50%, 60%, 70%, and 80%), respectively, defined as Heterogeneity Factor-1 (HF1) and Heterogeneity Factor-2 (HF2). These indices of heterogeneity were used to predict progression-free survival (PFS). Based on the results of the Cox proportional hazards model, we constructed a multi-parameter prediction model and evaluated the model in the training and validation cohorts by calibration curve, consistency index (C-index) and decision curve analysis (DCA). Results Clinicopathological and PET/CT data from 245 patients were reviewed. 153 patients (62.4%) experienced relapse after treatment. Comparing relapsed and non-relapse patients, all 18 F-FDG PET/CT parameters and heterogeneity index showed significant differences. There were significant differences in survival risk stratification according to HF1 and HF2 cut-off classifications ( P <0.0001). In multivariate Cox regression analysis, SUVmax ( P <0.0001), TLG ( P <0.0001), HF1 ( P =0.004), and HF2 ( P <0.0001) showed significant results. Among the clinicopathological parameters, IPI ( P =0.027) and Size ( P <0.0001) were selected as important parameters. Conclusions HF1 and HF2 obtained by the linear regression slope of MTV and TLG may be a novel and useful prognostic marker in DLBCL, which can achieve survival-risk stratification of patients. In addition, multiparametric models have the potential to effectively predict the risk of recurrence in patients. [18F]FDG PET/CT Metabolic parameters Diffuse large B-cell lymphoma Progression-free survival Heterogeneity factor Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Diffuse large B-cell lymphoma (DLBCL) is a B-cell malignancy that is the predominant form of non-Hodgkin's lymphoma (NHL) in adults [ 1 ] . Currently, immunotherapies, including rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), remain the mainstay of treatment for patients with DLBCL [ 2 ] . Despite this, in fact, 30 to 40 percent of patients experience relapse or death due to drug resistance, for example, and therefore remain uncured [ 3 ] . Therefore, early identification of patients with high-risk DLBCL with refractory relapse is essential for the development of tailored and personalized treatment strategies. The International Prognostic Index (IPI) is widely used in clinical practice, however, the discriminatory power of IPI decreases with the use of rituximab. During the rituximab period, very high-risk categories could not be identified [ 4 ] . Although IPI modifications to DLBCL have shown higher predictive value in recent years, new prognostic biomarkers are needed to better identify patients who could benefit from a more aggressive approach to treatment in order to provide optimal treatment decisions for patients with DLBCL [ 5 ] . Malignant tumor cells are made up of heterogeneous components, including not only biological components, but also gene expression, metabolic, and behavioral characteristics. Heterogeneity varies within the same cancer type and has a broad spectrum even at the same stage, as characteristics such as growth rate, vascularization, and necrosis vary within the same tumor cell population [ 6 – 8 ] . Studies have shown that DLBCL is a genetically heterogeneous disease with a variety of low-frequency mutations, somatic copy number alterations, and structural variations. Currently, these tumors are thought to be produced by antigen-exposed B cells that cross germinal centers. Aspects of the germinal center environment, including high proliferation rates, physiological activation-induced cytidine deaminase-mediated immunoglobulin receptor editing, and aberrant somatic hypermutations favor malignant transformation. The heterogeneity of DLBCL is reflected in transcriptionally defined subtypes that provide insight into disease pathogenesis and therapeutic candidate targets [ 9 ] . Due to the heterogeneity of DLBCL in various biological behaviors such as faster proliferation and skipping metastasis, there is also extensive heterogeneity in its clinical manifestations, response to treatment, and survival outcomes. Intra-tumor heterogeneity of DLBCL has been shown to be an important prognostic factor leading to disease recurrence and mortality [ 10 ] . Therefore, the establishment of new prognostic imaging biomarkers that accurately reflect the intricate internal heterogeneity of DLBCL is essential for accurately predicting prognosis. Over the past decade, 18 F-fluorodeoxyglucose positron emission tomography/computed tomography ( 18 F-FDG PET/CT) has been widely used as a useful imaging tool as a standard noninvasive method for assessing DLBCL pretreatment staging due to its ability to correctly assess nodules and extra-nodal involvement [ 11 ] . However, the function of baseline PET/CT goes beyond anatomical mapping, as it is also used to quantify FDG uptake by calculating the Standardized uptake value (SUV). Many ongoing studies have used the semi-quantitative metabolic parameters of PET/CT to predict patient prognosis, and tumor metabolism is typically assessed by Maximum normalized uptake values (SUVmax), a semi-quantitative index that represents the fraction of tumor proliferation that is the fastest proliferating. Total metabolic tumor volume (TMTV) is a measure of the volume of metabolic activity of the tumor as assessed by whole-body PET/CT and is the sum of the local tumor regions that absorb FDG. TMTV provides an additional dimension, which is the volume of metabolically active tumors. TMTV has shown strong prognostic value before initiation of treatment for Hodgkin and various non-Hodgkin lymphomas [ 12 ] . Total lesion glycolysis (TLG), which is the sum of all individual lesion volumes multiplied by the mean lesion SUV), represents the total glycolytic burden of the lesion, tumor heterogeneity, texture characteristics, total tumor surface, and spatial dispersion [ 13 ] . For DLBCL, the current International Prognostic Index (IPI) includes only surrogate measures of actual tumor burden, such as Ann Arbor stage and lactate dehydrogenase levels [ 14 ] . However, in many studies, TMTV has been shown to be an independent prognostic factor compared to IPI or IPI-related factors, as well as volume measurement (maximum mass diameter). Recent studies have shown that the researchers have successfully constructed the International Metabolic Prognostic Index (IMPI) by systematically comparing the total metabolic tumor volume (TMTV) with the International Prognostic Index (IPI) classification and its individual components, and the research data show that IMPI is significantly better than traditional IPI in predicting the risk of recurrence and survival prognosis, and has demonstrated higher predictive power [ 15 ] . In addition, the metabolic tumor area (MTA) calculated based on the metabolically active area of the largest tumor lesion, as a new prognostic indicator, has shown important clinical value in diffuse large B-cell lymphoma (DLBCL), which can effectively predict the recurrence risk and survival outcome of patients [ 16 ] . Further studies have found that the maximum distance between the two lesions (i.e., Dmax) and total metabolic tumor volume (TMTV) can be used to characterize the degree of lesion dissemination and metabolic tumor burden, respectively, and risk stratification of patients with DLBCL [ 17 ] . These prognostic models based on PET-derived parameters have shown significant efficacy in predicting patient prognosis and survival outcomes. In recent years, studies have further combined tumor heterogeneity-related parameters to optimize prognostic models, such as calculating metabolic heterogeneity (MH) by the cumulative SUV volume histogram (AUC-CSH) method, and combining it with metabolic tumor volume (MTV) and international prognostic index (IPI) to construct a joint model, which can more accurately identify high-risk patients and provide an important basis for the formulation of individualized treatment strategies [ 45 ] . However, to date, studies evaluating tumor heterogeneity based on linear regression slope methods to predict the prognosis of diffuse large B-cell lymphoma (DLBCL) have not been reported, which provides a different direction for future research. The intratumoral distribution of FDG uptake reflects glucose metabolism in the tumor and microenvironment, as well as other processes including necrosis, apoptosis, proliferation, and angiogenesis [ 19 ] . The enhancement of glycolysis is a major metabolic alteration in tumor cells, and in the 1920s, Otto Warburg experimentally discovered the presence of excessive glucose uptake in tumor tissues [ 20 ] . Heterogeneity of glucose uptake in tumors can be visualized by 18 FDG-PET/CT. However, metabolic heterogeneity does not necessarily represent heterogeneity in the innate metabolic profile of cells, which may vary at the single-cell level and may depend on the genetic background of each tumor cell, while extrinsic factors on 18 FDG-PET/CT, such as altered oxygen and substrate delivery depending on the quantity and quality of tumor vasculature, as well as tumor cell adaptation mechanisms to the altered microenvironment, may result in high metabolic heterogeneity and affect their sensitivity to therapeutic agents, and may affect treatment outcomes [ 21 ] . For example, the hypoxic state and cell cycle of tumor cells may enhance glycolysis and glucose uptake, thereby inhibiting tumor-infiltrating T lymphocytes due to competition for glucose. In addition, the metabolic activity of stromal cells and immune cells in the tumor microenvironment may have an impact on this [ 22 ] . Therefore, the metabolic parameters of 18 F-FDG PET/CT can characterize the metabolic heterogeneity within the tumor to a certain extent. However, the metabolic and volume parameters of 18 F-FDG PET/CT do not accurately represent the spatial heterogeneity of tumor metabolism. Whereas, diffuse large B-cell lymphoma (DLBCL) is a highly heterogeneous tumor [ 8 , 23 ] , and these metabolic parameters cannot fully capture the nuances of its internal structural components. In fact, tumor heterogeneity is thought to be a key factor in drug resistance and tumor recurrence [ 9 ] . Therefore, there is an urgent need for a novel method that can reflect the metabolic heterogeneity within tumors. In recent years, a variety of new methods have been developed to characterize metabolic heterogeneity within tumors. For example, the coefficient of variation, which is the ratio of the variance of the SUV within the tumor to the mean, can effectively reflect the degree of dispersion of the metabolic distribution; In addition, the cumulative normalized uptake-volume histogram (AUC-CSH) provides a more comprehensive quantitative approach to metabolic heterogeneity by plotting the percentage curves of tumor volume at different SUV thresholds (thresholds ranging from 0–100% of SUVmax). These methods have significantly improved the accuracy and reliability of tumor metabolic heterogeneity studies. Accumulating evidence suggests that high metabolic heterogeneity as assessed by 18 F-FDG PET/CT is an important predictor of poor prognosis in a variety of malignancies, such as oropharyngeal squamous cell carcinoma, pancreatic adenocarcinoma, non-small cell lung cancer and uterine leiomyosarcoma [ 24 – 26 , 30 ] . A recent study further confirmed that high metabolic heterogeneity has significant prognostic value in patients with newly diagnosed primary mediastinal B-cell lymphoma [ 27 ] , which is the first pivotal study in which metabolic heterogeneity has been identified as a prognostic biomarker for malignant lymphoma. The difference between our study of DLBCL and previous studies is that the method of assessing metabolic heterogeneity is different, and we use a linear regression slope method to assess the metabolic heterogeneity of DLBCL. Because TMTV and TLG under different SUV thresholds can reflect the inherent heterogeneity of DLBCL to a certain extent. This study uses only the 40%-80% SUV threshold, excluding non-existent intra-tumoral regions or areas with very low FDG uptake due to tumor necrosis, etc., to obtain a more accurate estimate of the metabolic distribution. Several studies have shown that histopathological features of malignancies such as non-small cell lung cancer, oligodendroglioma, head and neck squamous cell carcinoma and oral cavity cancer are significantly associated with intra-tumoral heterogeneity of 18 F-FDG uptake on PET [ 24 , 27 , 28 , 29 ] . In addition, 18 F-FDG heterogeneity in malignant sarcoma correlates with patient prognosis [ 11 ] . Recently, a study on the heterogeneity of intra-tumoral 18 F-FDG uptake in nasopharyngeal carcinoma showed that it was significantly associated with tumor aggressiveness and was associated with various outcome measures [ 31 ] . Therefore, by utilizing different PET metabolic parameters to reflect the intricate internal heterogeneity of DLBCL, it is possible to facilitate the generation of innovative and prognostic imaging biomarkers. However, to the best of our knowledge, there are currently very limited studies on the application of heterogeneity factors in PET image analysis to identify high-risk DLBCL patients, and there are no studies using a linear regression slope approach to evaluate the prognostic value of intra-tumoral heterogeneity in PET patients with DLBCL. Therefore, the purpose of this study was to extract heterogeneous factors of metabolic parameters from PET images to evaluate the prognostic value of DLBCL patients, and to study the relationship between heterogeneous factors and prognostic parameters such as maximum standardized uptake value (SUVmax), total metabolic tumor volume (TMTV), and total lesion glycolysis (TLG). Materials and methods Patients From January 2013 to December 2020, we retrospectively enrolled 245 histologically confirmed patients with DLBCL who had undergone PET/CT imaging scans at the Affiliated Cancer Hospital of Harbin Medical University prior to treatment. As this study is retrospective, it is not necessary to submit evidence of informed consent, and the hospital's institutional review board approved the study. The following criteria were required for inclusion: (1) histopathologically confirmed DLBCL; (2) no previous history of cancer; (3) 18 F-FDG PET/CT performed less than two weeks prior to the first treatment; (4) No anti-tumor therapy before the scan; and (5) availability of clinical and follow-up data. To determine their disease status, the case received anthracycline-based chemotherapy followed by PET/CT scans and scans using 18 F-FDG. Following the first 2 years of treatment, followed by follow-up assessments every 3 months and then every 6 months, the primary endpoint of the study is the patient's PFS, which can be defined as the period between diagnosis and the date of first recurrence, progression, or death due to any cause. At the last known follow-up, cases without any events were removed. PET/CT images Prior to the PET/CT examination, patients were required to fast for 6–8 hours and maintain blood glucose levels below 11.1 mmol/L to ensure optimal 18 F-FDG uptake. The imaging procedure was performed using a Discovery 690 Elite PET/CT scanner (GE Healthcare) with 18 F-FDG of radiochemical purity > 98%. Following intravenous administration of 18 F-FDG (3.7 MBq/kg), patients rested for approximately 60 minutes to allow adequate radiotracer distribution before imaging. The PET/CT acquisition protocol encompassed anatomical coverage from the skull base to the mid-thigh. Initial CT scanning was performed using the following parameters: 120 kV, 140 mA, 1.25 mm slice spacing, and 3.75 mm slice thickness, with a scan duration of 20–30 seconds. Subsequently, PET imaging was acquired in 3D mode across 6–7 bed positions, with 2.5 minutes per bed position while maintaining consistent patient positioning. Image reconstruction was performed using the ordered subset expectation maximization (OSEM) algorithm with CT-based attenuation correction. The reconstructed PET and CT datasets were transferred to a Xeleris™ workstation (GE Healthcare) for co-registration and fusion processing, enabling comprehensive anatomical and metabolic evaluation. Measurement of PET metabolic parameters PET/CT images were transferred to Advantage Workstation 4.5 (GE Healthcare) and reviewed by two experienced radiologists. The region of interest (ROI) is then extracted at a threshold of 41% SUVmax and manually adjusted if there is a high background or high uptake area (e.g., bladder, heart) in the vicinity of the lesion, or if the uptake rate in the lesion is low [ 18 ] . Volumetric metabolic parameters based on systemic multi-focus were achieved using Life-x software ( http://www.lifexsoft.org/.version 7.6), which automatically represented the volume of interest (VOI) using an SUV-based contour threshold method. Various thresholds are used to determine VOI boundaries; The relative thresholds are 40–80% of SUVmax, respectively. In this study, TMTV was automatically calculated by taking 40–80% of SUVmax in VOI, expressed as TMTV (40%), TMTV (50%), TMTV (60%), TMTV (70%), and TMTV (80%), respectively. TLG is calculated using 40–80% of the SUVmax threshold; The results were specified as TLG (40%), TLG (50%), TLG (60%), TLG (70%) and TLG (80%), respectively. In addition, we assessed the internal heterogeneity of 18 F-FDG uptake in DLBCL, represented by the heterogeneity factor (HF). HF is a heterogeneity index calculated based on the slope of PET/CT metabolic parameters (TMTV/TLG) as a function of SUV threshold, which is used to assess the non-uniformity of metabolic distribution within DLBCL tumors and may provide additional information for prognostic prediction. Figure 1 a shows a representative image of the linear regression slope using various SUV thresholds. The TMTV at each SUV threshold was recorded, and then HF1 was calculated as a negative slope for linear regression on the threshold-volume curve (Fig. 1 a). Figure 1 b records the TLG at each SUV threshold and then calculates the negative slope of HF2 as the linear regression on the threshold-glycolysis curve (Fig. 1 b). Thus, higher HF values indicate a faster decline in TMTV or TLG with an increase in SUV threshold, i.e., a more uneven distribution of metabolic activity within the tumor (higher heterogeneity), and lower HF values indicate a more homogeneous metabolic distribution (lower heterogeneity). In addition, other metabolic parameters within the ROI were calculated using Life-x software, including the maximum lesion diameter (Size), the maximum distance between lesions (Dmax), and the maximum normalized uptake value (SUVmax). Statistical analysis Statistical analysis was performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R statistical software (version 4.4.1). The statistical significance is set to a P -value of less than 0.05. Progression-free survival (PFS) was used as the endpoint to evaluate the prognosis of patients with DLBCL. The optimal thresholds for SUVmax, TMTV, TLG, HF1, and HF2 were determined using the Receiver operating characteristic (ROC) curve in the training cohort. Cox regression analysis was used to explore the prognostic value of potential independent factors, and the model was constructed. Survival was assessed by the Kaplan-Meier method and compared by the log-rank test. Results Patients characteristics The basic statistics of the cases are shown in Table 1 . A total of 245 patients with DLBCL were enrolled, including 114 males and 131 females. The mean age of participants was 57 years (range 13 to 93 years). Cases were divided into training cohorts and validation cohorts in a 7:3 ratio. According to the data analysis in Table 1 , there was no statistical difference between the training cohort and the validation cohort ( P > 0.05). The median follow-up time was 37.98 months and 32.31 months, respectively. During the follow-up period, the ratio of recurrence to non-recurrence cases was 3:2, and the information for patients with recurrence and non-recurrence was shown in Table 2 , and the analysis showed that there were statistically significant differences between the recurrence and non-recurrence groups in all variables except sex, age, and platelets ( P < 0.05). Table 1 The baseline characteristics of patients with DLBCL in the training and validation cohorts Characteristic Validation(n = 74) Training(n = 171) Overall (n = 245) P value* ECOG 0–1 56(22.90%) 146(59.60%) 202(82.40%) 0.067 ≥ 2 18(7.30%) 25(10.20%) 43(17.60%) B 0 58(23.70%) 153(62.40%) 211(86.10%) 0.121 1 16(6.50%) 18(7.30%) 34(13.90%) Gender male 27(11.00%) 87(35.50%) 114(46.50%) 0.138 female 47(19.20%) 84(34.30%) 131(53.50%) Ann-Arbor I-II 24(9.80%) 85(34.70%) 109(44.50%) 0.112 III-IV 50(20.40%) 86(35.10%) 136(55.50%) IPI 0–2 38(15.50%) 116(47.30%) 154(62.90%) 0.214 ≥ 3 36(14.70%) 55(22.40%) 91(37.10%) LDH ≤ 271 39(15.90%) 120(49.00%) 159(64.90%) 0.109 >271 35(14.30%) 51(20.80%) 86(35.10%) Platelet ≤ 300 50(20.40%) 130(53.10%) 180(73.50%) 0.169 >300 24(9.80%) 41(16.70%) 65(26.50%) Recurrence no 22(9.00%) 70(28.60%) 92(37.60%) 0.196 yes 52(21.20%) 101(41.20%) 153(62.40%) Age <60y 40(16.30%) 94(38.40%) 134(54.70%) 0.895 ≥ 60y 34(13.90%) 77(31.40%) 111(45.30%) Size 17.10(6.28, 29.75) 25.60(11.10, 33.5) 22.90(8.80, 32.10) 0.136 SUVmax 5.45(2.85, 9.78) 4.10(2.60, 6.70) 4.40(2.70, 8.3) 0.134 Dmax40 29.57(13.56, 52.55) 26.79(7.45, 57.59) 27.49(9.66, 56.15) 0.536 TMTV40 158.34(158.34, 307.64) 100.49(34.73, 201.85) 103.75(38.99, 241.26) 0.121 TLG40 2219.30(465.39, 5055.24) 1143.93(344.92, 2901.05) 1513.33(359.60, 3234.78) 0.141 HF1 3.77(0.99, 6.93) 2.15(0.80, 4.80) 2.30(0.87, 5.60) 0.118 HF2 47.78(10.00, 114.34) 24.21(6.78, 61.86) 28.51(7.51, 77.81) 0.131 * P > 0.05, Chi-square test. Table 2 The baseline characteristics of patients with DLBCL in the recurrence and no-recurrence cohorts Characteristic No recurrence(n = 92) Recurrence(n = 153) 0verall(n = 245) P value* ECOG 0.013 0–1 83(33.90%) 119(48.60%) 202(82.40%) ≥ 2 9(3.70%) 34(13.90%) 43(17.60%) B 0.01 0 86(35.10%) 125(51.00%) 211(86.10%) 1 6(2.40%) 28(11.40%) 34(13.90%) Gender 0.831 male 42(17.10%) 72(29.40%) 114(46.50%) female 50(20.40%) 81(33.10%) 131(53.50%) AnnArbor < 0.0001 I-II 57(23.30%) 52(21.20%) 109(44.50%) III-IV 35(14.30%) 101(41.20%) 136(55.50%) IPI < 0.0001 0–2 71(29.00%) 83(33.90%) 154(62.90%) ≥ 3 21(8.60%) 70(28.60%) 91(37.10%) LDH 271 16(6.50%) 70(28.60%) 86(35.10%) Platelet 0.056 ≤ 300 74(30.20%) 106(43.30%) 180(73.50%) >300 18(7.30%) 47(19.20%) 65(26.50%) Age 0.132 <60y 56(22.90%) 78(31.80%) 134(54.70%) ≥ 60y 36(14.70%) 75(30.60%) 111(45.30%) Size 19.90(9.30, 28.15) 23.75(8.30, 35.25) 22.90(8.80, 32.10) 0.036 SUVmax 4.72(1.925, 4.50) 11.01(3.20, 13.25) 4.40(2.70, 8.3) < 0.0001 Dmax40 23.01(5.52, 37.76) 38.17(14.08, 61.17) 27.49(9.66, 56.15) < 0.0001 TMTV40 96.09(22.45, 106.46) 257.08(62.63, 312.45) 103.75(38.99, 241.26) < 0.0001 TLG40 1309.56(167.23, 1192.02) 3646.45(889.70, 5108.14) 1513.33(359.60, 3234.78) < 0.0001 HF1 2.27(0.51, 2.42) 5.91(1.40, 7.20) 2.30(0.87, 5.60) < 0.0001 HF2 30.96(3.58, 25.41) 81.68(18.47, 115.50) 28.51(7.51, 77.81) < 0.0001 * P <0.05, Chi-square test. HF predicts the performance of recurrence and survival risk stratification In the training and validation cohorts, HF1 and HF2 showed high accuracy in predicting disease recurrence. The optimal cut-off values for HF1 and HF2 were determined by analysis, which were 0.390 and 0.232, respectively. Based on these cut-off values, HF1 and HF2 were divided into low and high groups, and the results showed that there was a statistically significant difference in progression-free survival (PFS) between the low and high HF1 and HF2 groups in the training and validation cohorts (all P < 0.0001) (Fig. 2 ). In addition, the optimal cut-off values for SUVmax, Size, Dmax, TMTV and TLG were 27.1, 4.1 cm, 46.6 cm, 133.1 cm³ and 827.6, respectively. Correlation of HF with other variables The heterogeneity factor HF1 was significantly associated with TMTV (r = 0.9976; P < 0.0001), TLG (r = 0.9217; P < 0.0001), and SUVmax (r = 0.5425; P < 0.0001), but not with size (Fig. 3 ). The heterogeneity factor HF2 was significantly correlated with TLG (r = 0.9965; P < 0.0001), TMTV (r = 0.9288; P < 0.0001), SUVmax (r = 0.5874; P < 0.0001), and Size (r = 0.2531; P < 0.0001) (Fig. 3 ). Heterogeneity factor HF1 was also significantly correlated with heterogeneity factor HF2 (r = 0.9296; P < 0.0001) (Fig. 4 ). HF1 and HF2 were strongly correlated with TMTV and TLG, respectively, and HF1 and TLG and HF2 were also strongly correlated with TMTV (Fig. 3 ). Single/multivariate analysis results Univariate Cox regression analysis was performed based on the clinical variables, metabolic parameters and heterogeneity factor HF of the training cohort samples, and the results are detailed in Table 3 . On this basis, multivariate Cox regression analysis was further carried out, including forward, backward and stepwise regression methods. Statistically significant variables in univariate analyses were included in multivariate analyses, and the results are summarized in Table 3 . Multivariate analysis showed that IPI, Size, SUVmax, TLG, HF1 and HF2 were all independent prognostic factors for predicting PFS and were statistically significant. Table 3 Univariate and multivariate analyses results on variables of the training cohort samples Variable HR HR(95%CI) P value* HR HR(95%CI) P value* Age 0.37 0.20–1.45 0.063 N/A N/A N/A AnnArbor 0.79 0.21–2.20 < 0.0001 N/A N/A N/A B 0.37 0.29–1.44 0.203 N/A N/A N/A Dmax40 0.15 0.00-1.01 0 N/A N/A N/A ECOG 0.08 0.27–1.08 0.76 N/A N/A N/A Gender 0.07 0.20–1.07 0.72 N/A N/A N/A IPI 0.77 0.20–2.15 < 0.0001 1.63 1.06–2.52 0.027 LDH 0.69 0.21–1.99 0.001 N/A N/A N/A Platelet 0.39 0.22–1.47 0.08 N/A N/A N/A Size 0.02 0.01–1.02 0.007 1.06 1.04–1.08 < 0.0001 SUVmax 0.03 0.01–1.03 0.002 1.07 1.04–1.1 < 0.0001 TLG40 0 0.00–1.00 < 0.0001 1 1 < 0.0001 TMTV40 0 0.00–1.00 < 0.0001 N/A N/A N/A HF1 0.05 0.01–1.05 0.0006 1.2 1.06–1.36 0.004 HF2 0 0.00–1.00 0.001 0.97 0.95–0.98 < 0.0001 * P <0.05, Chi-square test. N/A, not assessed Construction and evaluation of multi-parameter model nomograms The multiparametric model was constructed by integrating heterogeneity indices (HF1, HF2), metabolic indexes (Size, SUVmax, and TLG), and clinical variables (IPI). To evaluate the gain-effect of heterogeneity indices in prognostic prediction, we also constructed a traditional model that included only metabolic indicators and clinical variables. In addition, a model that included only IPI was established for comparison with the International Prognostic Index (IPI), which is routinely used in clinical practice. To facilitate routine use by clinicians, the nomogram demonstrates the ability of a multiparametric model to predict the risk of disease recurrence at 2, 3, and 5 years (Fig. 5). To demonstrate the clinical application of the nomogram, we show maximum intensity projection images of 18 F-FDG PET scans for two typical DLBCL cases (Fig. 6 ). As successfully predicted by the nomogram, the prognosis of the first case (Figs. 6 a, b, and c) showed no recurrence after standard treatment 5.2 years after diagnosis. Similar to the nomogram prediction for the second case with a higher risk of recurrence (Fig. 6 d, e, and f), disease progression was observed at 2 months after standard treatment. By plotting a calibration curve, calculate the degree of fit between the actual case results and the nomogram predictions. The results show calibration curves for 2-, 3-, and 5-year progression-free survival (Fig. 7 a). The calibration curve of the model with higher accuracy is closer to the diagonal dashed line, indicating that the predicted values are in good agreement with the clinical observations. In addition, we configured a time-dependent AUC curve for the prediction model (Fig. 7 b). The C-index of the multiparameter model was 0.729 (95% CI: 0.680–0.778), which was higher than that of the traditional model and the IPI model (Table 4 ). The analysis showed that both the multi-parameter model and the traditional model reached the highest C-index at 2 years. In order to prevent the model from overfitting, the model results were corrected several times by using the Bootstrap method, and the C index was stable at about 0.7. In addition, decision curve analysis further demonstrates that the multiparameter model results in a greater overall net benefit than other competing models across most risk thresholds (Fig. 7 c). The prediction performance of the model in the validation cohort is still better than that of the traditional model and the IPI model (Supplementary Fig. 1). These results indicate that the multi-parameter model significantly improves the accuracy and clinical practicability of recurrence risk prediction in patients with DLBCL after integrating heterogeneity indicators and provides strong support for the implementation of precision medicine. Table 4 The Harrell’s C-index results in the training and validation cohorts Progression-free survival Training cohort validation cohort C-index 95%CI C-index 95%CI Multiparametric model 0.729 0.680–0.778 0.685 0.609–0.761 Conventional model 0.694 0.643–0.745 0.636 0.560–0.712 IPI 0.598 0.549–0.647 0.539 0.464–0.613 At the same time, in order to more intuitively compare the prediction results of 2-year and 3-year progression-free survival (PFS), Fig. 8 shows the prediction performance of the HF model for 2-year and 3-year PFS. The results showed that the HF1 model was better than the 3-year PFS (AUC = 0.712) in predicting 2-year PFS (AUC = 0.729), while the HF2 model was better than the 2-year PFS (AUC = 0.768) in predicting 3-year PFS (AUC = 0.774). It is worth noting that the prediction performance of the HF2 model is better than that of the HF1 model, whether it is predicting 2-year or 3-year PFS. In order to further evaluate the predictive power of the HF model combined with the International Prognostic Index (IPI) for 2-year and 3-year PFS, we constructed a combined HF1-IPI and HF2-IPI models, respectively. Figure 8 shows the prediction results of the two joint models: the predicted AUC values of the HF1-IPI model for 2-year and 3-year PFS were 0.717 and 0.698, respectively, and the predicted AUC values of the HF2-IPI model for 2-year and 3-year PFS were 0.739 and 0.736, respectively. The results show that compared with the HF model alone, the combination of IPI indicators does not significantly improve the prediction performance of the model. In addition, similar results were obtained in the validation cohort (Supplementary Fig. 2). Discussion In summary, in this retrospective study of 245 patients, we developed and validated imaging biomarkers based on heterogeneous metabolic parameters that predict survival outcomes after treatment based on PET images. In this study, tumor metabolic parameters and clinical indexes were confirmed, including SUVmax, TLG, Size and IPI (HR = 1.07, P < 0.001; HR = 1.00, P < 0.001; HR = 1.06, P < 0.001; HR = 1.63, P < 0.001), which was an independent predictor of PFS in patients with DLBCL. To improve prediction performance, we integrated heterogeneous factors (i.e., HF1, HF2), metabolic parameters (i.e., Size, SUVmax, and TLG), and clinical indicators (i.e., IPI), and designed a multiparameter model containing multidimensional prognostic information. Calibration curves showed a strong correlation between predicted and actual outcomes of PFS at 2, 3, and 5 years for multiparametric models, and DLBCL results were traditionally assessed only on the basis of PFS and/or overall survival (OS). Previous studies have used two-year PFS as the endpoint for disease-associated DLBCL immunochemotherapy outcomes [ 32 ] . However, based on recent studies suggesting that three-year event-free survival (EFS) is a better indicator of long-term prognosis than two-year EFS in patients with newly diagnosed DLBCL [ 33 ] , this study focused on evaluating the predictive power of two- and three-year progression-free survival (PFS) rather than overall survival (OS). The results showed that the HF1 model performed better than the 3-year PFS in predicting 2-year PFS, while the HF2 model showed better predictive power in predicting 3-year PFS. It is worth noting that the overall prediction performance of the HF2 model is significantly better than that of the HF1 model in both 2-year and 3-year PFS predictions. Therefore, the HF2 model is considered to be a superior prediction tool compared with the HF1 model. However, unfortunately, the prediction performance of the model was not significantly improved after combining the HF model with the IPI index. Furthermore, according to the results of the correlation between the variables in this study, the heterogeneity factor HF1 was significantly correlated with TMTV, TLG and SUVmax, but not with tumor size (Table 2 and Fig. 2 ). Notably, HF1 correlates more strongly with TMTV and TLG than with other variables. The heterogeneity factor HF2 was significantly correlated with TLG, TMTV, SUVmax and Size (Table 2 and Fig. 2 ). Since HF2 exhibited a strong correlation with all variables included in the model, this may explain the results that HF2 is better predictive than HF1. In addition, due to the calculation method of heterogeneity factors, the correlation between HF1 and HF2 and TMTV and TLG, respectively, was significantly higher than that of other variables. At the same time, significant correlations between HF1 and TLG and between HF2 and TMTV were also observed. This may indicate that there is some similarity in the trends of TMTV and TLG with SUV thresholds, leading to a strong correlation between HF1 and HF2, which is consistent with previous studies [ 26 , 29 ] . In addition, in the multivariate prediction model, the 2-year PFS performed satisfactorily in both the validation and training sets (training cohort: 2-year PFS AUC = 0.729, validation cohort: 2-year PFS AUC = 0.685). At the same time, compared with the traditional model and IPI, the C-index results of the multi-parameter model showed better prognostic performance, and the C-index values of the training cohort and validation cohort predicted PFS were 0.729 and 0.685, respectively. By introducing the variables of the heterogeneity factor HF to explain the impact of false-positive and false-negative patients on prognostic judgment, we observed that the multiparametric model has superior clinical utility over the traditional model and IPI, so as to achieve personalized prognostic assessment and customized tumor treatment. Our results highlight the potential of PET-based heterogeneous metabolic parameter analysis to advance the prognostic assessment of patients with DLBCL and ultimately improve clinical outcomes. The microenvironment of human tumors is inherently heterogeneous. This heterogeneity increases the complexity of our understanding of tumor biological behavior and increases the challenges of treatment planning [ 34 ] . Since FDG PET depicts voxel-based glucose metabolism in tumors, it can be used to study this aspect of tumor heterogeneity. Studies have been conducted to investigate the relationship between heterogeneity in FDG uptake and histopathological heterogeneity in vivo and ex vivo tumors. Henriksson et al. found that intra-tumoral heterogeneity in FDG uptake in nude mice with head and neck squamous cell carcinoma xenografts depended on the distribution of different tissue components [ 35 ] . Specifically, regions that predominantly contain tumor cells (and therefore fewer stromal cells) and regions with less necrosis have significantly higher FDG uptake. Pugachev et al. reported that higher FDG uptake indicated the presence of hypoxic regions in nude mouse prostate tumors [ 19 ] . This heterogeneity in FDG uptake suggests that there are cell populations with different metabolic rates within the tumor, which may be due to differences in characteristics such as growth rate, vascularization, necrosis, etc. Among the 18 F-FDG PET/CT parameters, SUVmax is the most commonly used quantitative analysis parameter, and SUVmax has shown significant prognostic value in DLBCL. In addition, TMTV and TLG are of great significance in predicting the prognosis of DLBCL [ 42 ] . However, there is some controversy regarding the use of metabolic parameters such as SUVmax for assessment. Heterogeneity of tumors, partial volume effects, timing of SUV assessment, and body size may disrupt the assessment of metabolic parameters to reflect accurate tumor characteristics [ 43 ] . Volume parameters (TMTV and TLG) are expected to be more efficient than metabolic parameters (SUVmax) because tumor burden is considered using volume parameters. Previous studies using direct comparisons of metabolic and volumetric parameters have maintained that volumetric parameters are superior to metabolic parameters in predicting DLBCL [ 44 ] . A number of studies have shown that metabolic heterogeneity indexes based on volume parameters can predict tumor heterogeneity and disease treatment effect to a certain extent. As an example, heterogeneity in FDG uptake has been reported in the literature to predict treatment outcomes for certain cancer types, including cervical cancer and sarcoma [ 36 , 37 ] . Tisier et al. found that tumor metabolic heterogeneity with pre-treatment baseline FDG PET stratified treatment response prediction better than SUVmax and SUVmean in patients with esophageal cancer [ 38 ] . In other studies, pretreatment metabolic heterogeneity has been found to be significantly associated with tumor volume, SUVmax, and T stage of nasopharyngeal carcinoma, and can be used to predict response to treatment and patient survival (overall and disease-free survival) [ 39 ] . DLBCL is a clinically and biologically diverse disease entity, and patient response to treatment exhibits heterogeneity. Given the 30–40% recurrence rate and mortality rate, it is critical to identify patients with DLBCL at high risk of recurrence who could benefit from intensive systemic chemotherapy. The unique role of heterogeneity factors in the assessment of DLBCL response may be due to the correlation between heterogeneity in tumor metabolism and spatial heterogeneity in tumor response to chemotherapy [ 36 , 37 ] . The resulting heterogeneous response may have a further impact on treatment outcomes and survival. If we are able to predict a poor prognosis by preoperative 18 F-FDG PET/CT, prolonging the duration of first-line therapy or changing the agent is another treatment option, and close follow-up can be performed to detect recurrence early. There are several possible methods to assess heterogeneity using 18 F-FDG PET/CT. One of the most widely used methods is texture analysis. Accumulating evidence suggests that tumor heterogeneity as measured by 18 F-FDG PET/CT texture analysis is associated with treatment response and prognosis for esophageal, lung, head and neck, breast, and cervical cancers [ 40 ] . In addition, entropy is considered the best predictive parameter for pancreatic ductal adenocarcinoma among several structural analysis parameters [ 41 ] . However, there are currently no PET structural analysis parameters that are widely accepted for measuring tumor heterogeneity. In addition, texture analysis is difficult to evaluate in clinical practice due to the difficulty of obtaining measurement results. Compared with texture analysis, our method of using linear regression slope is more convenient and suitable for application in clinical practice. In addition, methods such as coefficient of variance, coefficient of variation, and AUC-CSH have been used to assess metabolic heterogeneity within tumors [ 26 , 27 , 31 ] . However, in lymphoma, there have been no studies using a linear regression slope to characterize disease heterogeneity. The heterogeneity index (HF) derived from the linear regression slope has a unique added value in the assessment of tumor heterogeneity. Compared with simple prognostic parameters such as Dmax (i.e., the maximum tumor diameter), HF can not only quantify metabolic heterogeneity within tumors, but also reflect the spatial heterogeneity distribution characteristics of tumors, thereby providing more comprehensive heterogeneity information, whereas Dmax can only reflect the unidimensional characteristics of tumors and cannot fully assess tumor heterogeneity [ 17 ] . In addition, compared to complex analytical methods employing artificial intelligence (AI) or radiomics, HF is computationally simple and easy to perform clinically, while still providing predictive performance comparable to complex methods. Therefore, HF has significant advantages in balancing computational complexity with clinical practicability, especially in clinical scenarios where rapid assessment of tumor heterogeneity and prognosis is required. However, AI and radiomics approaches are able to integrate more dimensions of data and may have more potential when dealing with highly complex patterns of heterogeneity. Consistent with previous studies [ 26 , 29 , 31 ] , the heterogeneous factors HF1 (HR = 1.2, P < 0.001) and HF2 (HR = 0.97, P < 0.001) derived from PET images in this study were reliable prognostic biomarkers for PFS. Heterogeneous metabolic parameters extracted from medical images may be able to reflect multi-level disease heterogeneity and thus may have higher power in diagnosis and outcome prediction compared with a single metabolic profile [ 45 ] . In conclusion, our study supports the use of heterogeneous metabolic parameters from PET/CT images as a valuable tool to identify high-risk patients with DLBCL who require intensive treatment or close monitoring. Overall, HF has important application value in clinical practice as a heterogeneity assessment tool of medium complexity. However, there are limitations and shortcomings in this study. First, due to the nature of retrospective studies, inherent selection limitations are unavoidable. Secondly, with regard to the calculation methods of TMTV, different calculation methods of TMTV may have an impact on the results of heterogeneity analysis, especially in the low SUV threshold region, TMTV is very sensitive to threshold changes [ 27 ] . Small differences may be magnified, leading to "exponential" variations in heterogeneous results. Although the selection of a threshold range of 40 to 80 percent for SUVs (which may not include a region of low SUV thresholds), the use of relative thresholds for SUVs instead of absolute thresholds (which may be closer to the actual distribution of metabolic activity) [ 18 ] , and the use of a uniform TMTV calculation method may reduce the impact of differences in TMTV calculation methods on the results, this approach will inevitably deviate from the true value to some extent. Thirdly, with regard to the calculation method of heterogeneity factors, a potential drawback of heterogeneity factor calculation is the association between intra-tumoral heterogeneity and tumor volume and total lesion glycolysis, which may affect the independence of parametric heterogeneity factors and the robustness of the method. Fourth, although histological studies have shown a good spatial correlation between FDG uptake regions and histological outcomes, there is no reference standard for the overall measurement of heterogeneity. We were unable to verify the accuracy of our results as reflecting the true extent of tumor heterogeneity. Future work may focus on improving heterogeneity factor calculations and assessing their impact on treatment planning by addressing these shortcomings. Finally, our data is limited to cases in one medical center and is smaller. Therefore, clinical support for this predictive model is limited. Future studies will require larger-scale external validation from multiple medical centers. Conclusion Heterogeneity factors (HFs) can reflect the metabolic heterogeneity of FDG to a certain extent. Heterogeneity factors (HFs) can be used as a stable imaging biomarker for prognosis prediction of DLBCL patients and can achieve survival risk stratification of patients. The prediction results of the multi-parameter model are better than those of the traditional model and the IPI model and may accurately predict the recurrence risk of DLBCL patients. Declarations Funding This paper is supported by the Provincial Key Research and Development Program of Heilongjiang Province (GA21C001 KW), Distinguished Young Scientist Funding of Harbin Medical University Affiliated Tumor Hospital (JCQN2019-02), Key Project of Harbin Medical University Cancer Hospital Climbing Funding (PDYS2024-03), Key Innovation Technology Project Harbin Medical University Cancer Hospital Innovation Technology Funding (CXJSZD-2023-04),Heilongjiang Province's "unveiling and leading" technology research and development project (2022ZXJ03C01).The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data sharing statement The datasets generated during and analyzed during the current study are not publicly available due to patient privacy concerns but are available from the corresponding author on reasonable request. Conflict of Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper Informed Consent The need of written informed consent was waived in this study due to the retrospective nature. Guarantor The scientific guarantor of this publication is Kezheng Wang. Statistics and Biometry No complex statistical methods were necessary for this paper Author contributions Fan Ge and Tingting Wu contributed equally to this work. Tingting Wu contributed to the study conception and design. Material preparation and data collection were performed by Tingting Wu, Xinyue Yang and Fan Ge. The first draft of the manuscript and statistics analysis were performed by Fan Ge. Data curation was performed by Mengye Peng, Chen Yang and Fan Ge. Reviewing and editing were performed by Kezheng Wang. And all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Statistics and biometry One of the authors (Fan Ge) has significant statistical expertise. Ethical approval and consent to participate The retrospective study was reviewed and approved by Harbin Medical University Cancer Hospital Medical Ethics Committee. All methods were in accordance with the ethical standards as laid down in the Declaration of Helsinki and its later amendments or comparable ethical standards. As this was a retrospective study, formal consent was not obtained. Acknowledgements We would like to thank GE Healthcare Company for providing technical support for our manuscript. References Li S, Young KH, Medeiros LJ (2018) Diffuse Large B-Cell Lymphoma Pathol 50(1):74–87 Vodicka P, Klener P, Trneny M (2022) Diffuse Large B-Cell Lymphoma (DLBCL): Early Patient Management and Emerging Treatment Options. OTT Volume 15, 1481–1501 Poletto S, Novo M, Paruzzo L et al (2022) Treatment Strategies for Patients with Diffuse Large B-Cell Lymphoma. Cancer Treat Rev 110:102443 Sehn LH, Berry B, Chhanabhai M et al (2007) The Revised International Prognostic Index (R-IPI) Is a Better Predictor of Outcome than the Standard IPI for Patients with Diffuse Large B-Cell Lymphoma Treated with R-CHOP. Blood 109(5):1857–1861 Ceriani L, Gritti G, Cascione L et al (2020) SAKK38/07 Study: Integration of Baseline Metabolic Heterogeneity and Metabolic Tumor Volume in DLBCL Prognostic Model. Blood Adv 4(6):1082–1092 Nicolson GL (1984) Generation of Phenotypic Diversity and Progression in Metastatic Tumor Cells. Cancer Metast Rev 3(1):25–42 Tumor Heterogeneity Morphological, Molecular and Clinical Implications. Histology and Histopathology 2000, No. 33, 881–898 Gillies RJ, Schomack PA, Secomb TW et al (1999) Causes and Effects of Heterogeneous Perfusion in Tumors. Neoplasia 1(3):197–207 Chapuy B, Stewart C, Dunford AJ et al (2018) Molecular Subtypes of Diffuse Large B Cell Lymphoma Are Associated with Distinct Pathogenic Mechanisms and Outcomes. Nat Med 24(5):679–690 McGranahan N, Swanton C (2017) Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell 168(4):613–628 Kostakoglu L, Chauvie S (2019) PET-Derived Quantitative Metrics for Response and Prognosis in Lymphoma. PET Clin 14(3):317–329 Chen S, He K, Feng F et al (2019) Metabolic Tumor Burden on Baseline 18F-FDG PET/CT Improves Risk Stratification in Pediatric Patients with Mature B-Cell Lymphoma. Eur J Nucl Med Mol Imaging 46(9):1830–1839 Vergote VKJ, Verhoef G, Janssens A et al (2024) International Metabolic Prognostic Index Is Superior to Other Metabolic Tumor Volume-Based Prognostication Methods in a Real-Life Cohort of Diffuse Large B-Cell Lymphoma. Journal of nuclear medicine: official publication. Society of Nuclear Medicine Shagera QA, Cheon GJ, Koh Y et al (2019) Prognostic Value of Metabolic Tumour Volume on Baseline 18F-FDG PET/CT in Addition to NCCN-IPI in Patients with Diffuse Large B-Cell Lymphoma: Further Stratification of the Group with a High-Risk NCCN-IPI. Eur J Nucl Med Mol Imaging 46(7):1417–1427 Mikhaeel NG, Heymans MW, Eertink JJ et al (2022) Proposed New Dynamic Prognostic Index for Diffuse Large B-Cell Lymphoma: International Metabolic Prognostic Index. J Clin Oncol 40(21):2352–2360 Cui S, Xin W, Wang F et al (2024) Metabolic tumour area: a novel prognostic indicator based on 18F-FDG PET/CT in patients with diffuse large B-cell lymphoma in the R-CHOP era. BMC Cancer 24(1):895 Cottereau AS, Meignan M, Nioche C et al (2021) Risk stratification in diffuse large B-cell lymphoma using lesion dissemination and metabolic tumor burden calculated from baseline PET/CT†. Ann Oncol 32(3):404–411 Boellaard R, O'Doherty MJ, Weber WA et al (2010) FDG PET and PET/CT: EANM procedure guidelines for tumour PET imaging: version 1.0. Eur J Nucl Med Mol Imaging 37(1):181–200 Pugachev A, Ruan S, Carlin S et al (2005) Dependence of FDG Uptake on Tumor Microenvironment. Int J Radiation Oncology*Biology*Physics 62(2):545–553 Warburg O, Wind F, Negelein E, THE METABOLISM OF TUMORS IN THE BODY (1927) J Gen Physiol 8(6):519–530 Vander Heiden MG (2011) Targeting Cancer Metabolism: A Therapeutic Window Opens. Nat Rev Drug Discov 10(9):671–684 Al Tameemi W, Dale TP, Al-Jumaily RMK et al (2019) Hypoxia-Modified Cancer Cell Metabolism Front Cell Dev Biol 7:4 Morin RD, Arthur SE, Hodson DJ (2022) Molecular Profiling in Diffuse Large B-cell Lymphoma: Why so Many Types of Subtypes? Br J Haematol 196(4):814–829 Van Velden FHP, Cheebsumon P, Yaqub M et al (2011) Evaluation of a Cumulative SUV-Volume Histogram Method for Parameterizing Heterogeneous Intratumoural FDG Uptake in Non-Small Cell Lung Cancer PET Studies. Eur J Nucl Med Mol Imaging 38(9):1636–1647 Mena E, Taghipour M, Sheikhbahaei S et al (2017) Value of Intratumoral Metabolic Heterogeneity and Quantitative 18F-FDG PET/CT Parameters to Predict Prognosis in Patients With HPV-Positive Primary Oropharyngeal Squamous Cell Carcinoma. Clin Nucl Med 42(5):e227–e234 Kim Y, Kim YJ, Paeng JC et al (2017) Heterogeneity Index Evaluated by Slope of Linear Regression on 18F-FDG PET/CT as a Prognostic Marker for Predicting Tumor Recurrence in Pancreatic Ductal Adenocarcinoma. Eur J Nucl Med Mol Imaging 44(12):1995–2003. https://doi.org/10.1007/s00259-017-3755-8 Ceriani L, Milan L, Martelli M et al (2018) Metabolic Heterogeneity on Baseline 18FDG-PET/CT Scan Is a Predictor of Outcome in Primary Mediastinal B-Cell Lymphoma. Blood 132(2):179–186 Cheng N-M, Dean Fang Y-H, Tung-Chieh Chang J et al (2013) Textural Features of Pretreatment18 F-FDG PET/CT Images: Prognostic Significance in Patients with Advanced T-Stage Oropharyngeal Squamous Cell Carcinoma. J Nucl Med 54(10):1703–1709 Kwon SH, Yoon J, An Y, Shin YS et al (2014) Prognostic Significance of the Intratumoral Heterogeneity of 18 F-FDG Uptake in Oral Cavity Cancer. J Surg Oncol 110(6):702–706 Lee J-W, Park J-Y, Lee HJ et al (2018) Preoperative [18F]FDG PET/CT Tumour Heterogeneity Index in Patients with Uterine Leiomyosarcoma: A Multicentre Retrospective Study. Eur J Nucl Med Mol Imaging 45(8):1309–1316 Huang B, Chan T, Kwong DL-W et al (2012) Nasopharyngeal Carcinoma: Investigation of Intratumoral Heterogeneity With FDG PET/CT. Am J Roentgenol 199(1):169–174 Aide N, Fruchart C, Nganoa C et al (2020) Baseline 18F-FDG PET Radiomic Features as Predictors of 2-Year Event-Free Survival in Diffuse Large B Cell Lymphomas Treated with Immunochemotherapy. Eur Radiol 30(8):4623–4632 Izumiyama K, Inao T, Goto H et al (2024) Event-Free Survival at 36 Months Is a Suitable Endpoint for Diffuse Large B-Cell Lymphoma Patients Treated with Immunochemotherapy: Real-World Evidence from the North Japan Hematology Study Group. Haematologica Heppner GH, Miller BE (1983) Tumor Heterogeneity: Biological Implications and Therapeutic Consequences. Cancer Metast Rev 2(1):5–23 Henriksson E, Kjellen E, Wahlberg P et al (2007) 2-Deoxy-2-[18F] Fluoro-D-Glucose Uptake and Correlation to Intratumoral Heterogeneity. Anticancer Res 27(4B):2155–2159 Kidd EA, Grigsby PW (2008) Intratumoral Metabolic Heterogeneity of Cervical Cancer. Clin Cancer Res 14(16):5236–5241 Eary JF, O’Sullivan F, O’Sullivan J et al (2008) Spatial Heterogeneity in Sarcoma18 F-FDG Uptake as a Predictor of Patient Outcome. J Nucl Med 49(12):1973–1979 Tixier F, Le Rest CC, Hatt M et al (2011) Intratumor Heterogeneity Characterized by Textural Features on Baseline18 F-FDG PET Images Predicts Response to Concomitant Radiochemotherapy in Esophageal Cancer. J Nucl Med 52(3):369–378 Chan WKS, Mak HKF, Huang B et al (2010) Nasopharyngeal Carcinoma: Relationship between 18F-FDG PET-CT Maximum Standardized Uptake Value, Metabolic Tumour Volume and Total Lesion Glycolysis and TNM Classification. Nucl Med Commun 31(3):206–210 Hatt M, Majdoub M, Vallières M et al (2015) 18F-FDG PET Uptake Characterization Through Texture Analysis: Investigating the Complementary Nature of Heterogeneity and Functional Tumor Volume in a Multi–Cancer Site Patient Cohort. J Nucl Med 56(1):38–44 Hyun SH, Kim HS, Choi SH et al (2016) Intratumoral Heterogeneity of 18F-FDG Uptake Predicts Survival in Patients with Pancreatic Ductal Adenocarcinoma. Eur J Nucl Med Mol Imaging 43(8):1461–1468 Kanoun S, Rossi C, Berriolo-Riedinger A et al (2014) Baseline Metabolic Tumour Volume Is an Independent Prognostic Factor in Hodgkin Lymphoma. Eur J Nucl Med Mol Imaging 41(9):1735–1743 Thie JA (2004) Understanding the Standardized Uptake Value, Its Methods, and Implications for Usage. J Nucl Med 45(9):1431–1434 Sasanelli M, Meignan M, Haioun C et al (2014) Pretherapy Metabolic Tumour Volume Is an Independent Predictor of Outcome in Patients with Diffuse Large B-Cell Lymphoma. Eur J Nucl Med Mol Imaging 41(11):2017–2022 Senjo H, Hirata K, Izumiyama K, North Japan Hematology Study Group et al (2020) High Metabolic Heterogeneity on Baseline 18FDG-PET/CT Scan as a Poor Prognostic Factor for Newly Diagnosed Diffuse Large B-Cell Lymphoma. Blood Adv 4(10):2286–2296 Additional Declarations No competing interests reported. Supplementary Files Supplementary.docx Cite Share Download PDF Status: Published Journal Publication published 11 Jun, 2025 Read the published version in Annals of Hematology → Version 1 posted Editorial decision: Accepted 12 May, 2025 Reviews received at journal 21 Apr, 2025 Reviewers agreed at journal 21 Apr, 2025 Reviewers invited by journal 17 Apr, 2025 Submission checks completed at journal 17 Apr, 2025 First submitted to journal 11 Apr, 2025 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-6070367","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":444291258,"identity":"032acd1a-78b0-4843-b4b5-7291914f3d0c","order_by":0,"name":"Fan Ge","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Fan","middleName":"","lastName":"Ge","suffix":""},{"id":444291259,"identity":"2c041d02-ddd1-4873-9252-bf7ef7f9a77f","order_by":1,"name":"Tingting Wu","email":"","orcid":"","institution":"The First People's Hospital of Shuangliu District, Chengdu, West China Airport Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Tingting","middleName":"","lastName":"Wu","suffix":""},{"id":444291260,"identity":"172edf26-034a-4574-b491-70849dd74e83","order_by":2,"name":"Xinyue Yang","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xinyue","middleName":"","lastName":"Yang","suffix":""},{"id":444291261,"identity":"581d016e-d8ab-464f-ae0e-92400b2e1dec","order_by":3,"name":"Mengye Peng","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mengye","middleName":"","lastName":"Peng","suffix":""},{"id":444291262,"identity":"edfea2f3-7713-411d-8ee8-04703f042f0d","order_by":4,"name":"Chen Yang","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chen","middleName":"","lastName":"Yang","suffix":""},{"id":444291263,"identity":"da3ae500-6172-4989-aaff-4955c6e8fd57","order_by":5,"name":"Kezheng Wang","email":"data:image/png;base64,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","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":true,"prefix":"","firstName":"Kezheng","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-02-20 09:08:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6070367/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6070367/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00277-025-06409-8","type":"published","date":"2025-06-11T15:57:49+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":81009770,"identity":"155127f0-6391-4df7-a09d-957f50c89906","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":78524,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCurves by plotting thresholds to metabolic factors automatically calculated with variable thresholds of 40, 50, 60, 70, and 80% of SUVmax. \u003c/strong\u003e(a.) Heterogeneity factor-1(2.59) is calculated as the negative slop from TMTV’s linear regression. (b) Heterogeneity factor-2(60.6) is calculated as the negative slop from TLG’s linear regression.\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/4595251b1124d18dd0b2f811.jpeg"},{"id":81009771,"identity":"fb844f6c-3ff3-4fc2-aebe-bd55a4bc8dad","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":222797,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKaplan–Meier survival analyses of PFS according to the HF1(a.) and HF2 (b.) in the training cohort. Kaplan–Meier survival analyses of PFS according to the HF1(c.) and HF2 (d.) in the validation cohort.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/89d179eb8fb06b2ce7880381.jpeg"},{"id":81009772,"identity":"58c38da7-d59c-4ca6-b92d-892dcee33529","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":82610,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelations between heterogeneity factor-1 and (a) metabolic tumor volume, (b) total lesion glycolysis, (c) SUVmax, (d) Size. Correlations between heterogeneity factor-2 and(d) metabolic tumor volume, (e) total lesion glycolysis, (f) SUVmax, (h) Size.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/8808464716906085d79a3474.jpeg"},{"id":81010460,"identity":"5744a2a7-ca44-406e-8b1d-4c8c44b8c4d9","added_by":"auto","created_at":"2025-04-21 08:17:52","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":36450,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelations between heterogeneity factor-1 and heterogeneity factor-2.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/8b7d1bde9bdc8b7297790c30.png"},{"id":81009784,"identity":"5f6a96d4-a7e7-4456-b055-98c8fbb5b207","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":10673,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe nomogram combined with the heterogeneity factors, the metabolic risk factors and the independent clinical risk factors, to predict the risk of PFS at 2, 3, and 5 years.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/466b6fbfab1fd65d8cd4bf52.png"},{"id":81695854,"identity":"4359ee44-2a01-487d-be73-df10e6c016d3","added_by":"auto","created_at":"2025-04-30 12:06:10","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":432446,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTwo typical cases with DLBCL to show the clinical application of the nomogram.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ea–c：Female, 64 years old, who underwent 5 cycles of R-CHOP regimen chemotherapy after firstly diagnosed of DLBCL (a) and was confirmed as completely response (b). SIZE:2, SUVmax:7, TLG:212, HF1:0, HF2:0, IPI:1. The red dots of each variable were drawn (c) and total points (0 + 1.2 + 2.1 + 2.8 + 11.4+ 18= 35.5).\u003c/p\u003e\n\u003cp\u003ed–f：Female, 46 years old, who underwent 4 cycles of R-CHOP regimen chemotherapy after firstly diagnosed of DLBCL(d) and was confirmed as progression disease(e).SIZE:42 SUVmax:8, TLG:10913, HF1:1, HF2:1, IPI:0.The red dotsof each variable were drawn (f) and total points(0 + 1.5 + 7.8+ 17.2 + 35 + 59.5 = 121).\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/1b7ef434859fb4211611791c.png"},{"id":81009778,"identity":"3e14fe19-9cdf-4138-be3f-3ff5207e56fc","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":129349,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a.) Calibration curves of the model for predicting PFS in the training cohort. (b.) The time-dependent area under the ROC curve of the models for predicting PFS in the training cohort. (c.) Decision curve analysis of the models for predicting PFS in the training.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/cca7e65a6d011336f7e985bb.jpeg"},{"id":81010467,"identity":"30e528a4-bfa4-40f5-8fb2-76d313b1c123","added_by":"auto","created_at":"2025-04-21 08:17:52","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":142212,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a.) The ROC curve of HF1 for predicting 2-PFS and 3-PFS in the training cohort. (b.) The ROC curve of HF1 and IPI for predicting 2-PFS and 3-PFS in the training cohort. (c.) The ROC curve of HF2 for predicting 2-PFS and 3-PFS in the training cohort. (d.) The ROC curve of HF2 and IPI for predicting 2-PFS and 3-PFS in the training cohort.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/14d3531af8bb6915e3541c92.jpeg"},{"id":84726557,"identity":"168e66ea-2c2d-4d89-87ec-49716b6e9e37","added_by":"auto","created_at":"2025-06-16 16:07:00","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2985435,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/c14e4148-2eba-41aa-8d98-43d61f18e8a5.pdf"},{"id":81009776,"identity":"617aa580-9d36-4fbd-a2c8-9362a2faddda","added_by":"auto","created_at":"2025-04-21 08:09:52","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":272179,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementary.docx","url":"https://assets-eu.researchsquare.com/files/rs-6070367/v1/8ae2d32b20ceb6f14debdbf2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Feasibility analysis of metabolic parameters based on baseline 18 F-FDG PET/CT to predict heterogeneity and recurrence of diffuse large B-cell lymphoma","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDiffuse large B-cell lymphoma (DLBCL) is a B-cell malignancy that is the predominant form of non-Hodgkin's lymphoma (NHL) in adults \u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. Currently, immunotherapies, including rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), remain the mainstay of treatment for patients with DLBCL \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Despite this, in fact, 30 to 40 percent of patients experience relapse or death due to drug resistance, for example, and therefore remain uncured \u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e. Therefore, early identification of patients with high-risk DLBCL with refractory relapse is essential for the development of tailored and personalized treatment strategies. The International Prognostic Index (IPI) is widely used in clinical practice, however, the discriminatory power of IPI decreases with the use of rituximab. During the rituximab period, very high-risk categories could not be identified \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. Although IPI modifications to DLBCL have shown higher predictive value in recent years, new prognostic biomarkers are needed to better identify patients who could benefit from a more aggressive approach to treatment in order to provide optimal treatment decisions for patients with DLBCL \u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMalignant tumor cells are made up of heterogeneous components, including not only biological components, but also gene expression, metabolic, and behavioral characteristics. Heterogeneity varies within the same cancer type and has a broad spectrum even at the same stage, as characteristics such as growth rate, vascularization, and necrosis vary within the same tumor cell population \u003csup\u003e[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. Studies have shown that DLBCL is a genetically heterogeneous disease with a variety of low-frequency mutations, somatic copy number alterations, and structural variations. Currently, these tumors are thought to be produced by antigen-exposed B cells that cross germinal centers. Aspects of the germinal center environment, including high proliferation rates, physiological activation-induced cytidine deaminase-mediated immunoglobulin receptor editing, and aberrant somatic hypermutations favor malignant transformation. The heterogeneity of DLBCL is reflected in transcriptionally defined subtypes that provide insight into disease pathogenesis and therapeutic candidate targets \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. Due to the heterogeneity of DLBCL in various biological behaviors such as faster proliferation and skipping metastasis, there is also extensive heterogeneity in its clinical manifestations, response to treatment, and survival outcomes. Intra-tumor heterogeneity of DLBCL has been shown to be an important prognostic factor leading to disease recurrence and mortality \u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. Therefore, the establishment of new prognostic imaging biomarkers that accurately reflect the intricate internal heterogeneity of DLBCL is essential for accurately predicting prognosis.\u003c/p\u003e \u003cp\u003eOver the past decade, \u003csup\u003e18\u003c/sup\u003eF-fluorodeoxyglucose positron emission tomography/computed tomography (\u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT) has been widely used as a useful imaging tool as a standard noninvasive method for assessing DLBCL pretreatment staging due to its ability to correctly assess nodules and extra-nodal involvement \u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. However, the function of baseline PET/CT goes beyond anatomical mapping, as it is also used to quantify FDG uptake by calculating the Standardized uptake value (SUV). Many ongoing studies have used the semi-quantitative metabolic parameters of PET/CT to predict patient prognosis, and tumor metabolism is typically assessed by Maximum normalized uptake values (SUVmax), a semi-quantitative index that represents the fraction of tumor proliferation that is the fastest proliferating. Total metabolic tumor volume (TMTV) is a measure of the volume of metabolic activity of the tumor as assessed by whole-body PET/CT and is the sum of the local tumor regions that absorb FDG. TMTV provides an additional dimension, which is the volume of metabolically active tumors. TMTV has shown strong prognostic value before initiation of treatment for Hodgkin and various non-Hodgkin lymphomas \u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Total lesion glycolysis (TLG), which is the sum of all individual lesion volumes multiplied by the mean lesion SUV), represents the total glycolytic burden of the lesion, tumor heterogeneity, texture characteristics, total tumor surface, and spatial dispersion \u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. For DLBCL, the current International Prognostic Index (IPI) includes only surrogate measures of actual tumor burden, such as Ann Arbor stage and lactate dehydrogenase levels \u003csup\u003e[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e. However, in many studies, TMTV has been shown to be an independent prognostic factor compared to IPI or IPI-related factors, as well as volume measurement (maximum mass diameter). Recent studies have shown that the researchers have successfully constructed the International Metabolic Prognostic Index (IMPI) by systematically comparing the total metabolic tumor volume (TMTV) with the International Prognostic Index (IPI) classification and its individual components, and the research data show that IMPI is significantly better than traditional IPI in predicting the risk of recurrence and survival prognosis, and has demonstrated higher predictive power \u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. In addition, the metabolic tumor area (MTA) calculated based on the metabolically active area of the largest tumor lesion, as a new prognostic indicator, has shown important clinical value in diffuse large B-cell lymphoma (DLBCL), which can effectively predict the recurrence risk and survival outcome of patients \u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e. Further studies have found that the maximum distance between the two lesions (i.e., Dmax) and total metabolic tumor volume (TMTV) can be used to characterize the degree of lesion dissemination and metabolic tumor burden, respectively, and risk stratification of patients with DLBCL \u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. These prognostic models based on PET-derived parameters have shown significant efficacy in predicting patient prognosis and survival outcomes. In recent years, studies have further combined tumor heterogeneity-related parameters to optimize prognostic models, such as calculating metabolic heterogeneity (MH) by the cumulative SUV volume histogram (AUC-CSH) method, and combining it with metabolic tumor volume (MTV) and international prognostic index (IPI) to construct a joint model, which can more accurately identify high-risk patients and provide an important basis for the formulation of individualized treatment strategies \u003csup\u003e[\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/sup\u003e. However, to date, studies evaluating tumor heterogeneity based on linear regression slope methods to predict the prognosis of diffuse large B-cell lymphoma (DLBCL) have not been reported, which provides a different direction for future research.\u003c/p\u003e \u003cp\u003eThe intratumoral distribution of FDG uptake reflects glucose metabolism in the tumor and microenvironment, as well as other processes including necrosis, apoptosis, proliferation, and angiogenesis \u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. The enhancement of glycolysis is a major metabolic alteration in tumor cells, and in the 1920s, Otto Warburg experimentally discovered the presence of excessive glucose uptake in tumor tissues \u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e. Heterogeneity of glucose uptake in tumors can be visualized by \u003csup\u003e18\u003c/sup\u003eFDG-PET/CT. However, metabolic heterogeneity does not necessarily represent heterogeneity in the innate metabolic profile of cells, which may vary at the single-cell level and may depend on the genetic background of each tumor cell, while extrinsic factors on \u003csup\u003e18\u003c/sup\u003eFDG-PET/CT, such as altered oxygen and substrate delivery depending on the quantity and quality of tumor vasculature, as well as tumor cell adaptation mechanisms to the altered microenvironment, may result in high metabolic heterogeneity and affect their sensitivity to therapeutic agents, and may affect treatment outcomes \u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. For example, the hypoxic state and cell cycle of tumor cells may enhance glycolysis and glucose uptake, thereby inhibiting tumor-infiltrating T lymphocytes due to competition for glucose. In addition, the metabolic activity of stromal cells and immune cells in the tumor microenvironment may have an impact on this \u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e. Therefore, the metabolic parameters of \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT can characterize the metabolic heterogeneity within the tumor to a certain extent.\u003c/p\u003e \u003cp\u003eHowever, the metabolic and volume parameters of \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT do not accurately represent the spatial heterogeneity of tumor metabolism. Whereas, diffuse large B-cell lymphoma (DLBCL) is a highly heterogeneous tumor \u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e, and these metabolic parameters cannot fully capture the nuances of its internal structural components. In fact, tumor heterogeneity is thought to be a key factor in drug resistance and tumor recurrence \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. Therefore, there is an urgent need for a novel method that can reflect the metabolic heterogeneity within tumors.\u003c/p\u003e \u003cp\u003eIn recent years, a variety of new methods have been developed to characterize metabolic heterogeneity within tumors. For example, the coefficient of variation, which is the ratio of the variance of the SUV within the tumor to the mean, can effectively reflect the degree of dispersion of the metabolic distribution; In addition, the cumulative normalized uptake-volume histogram (AUC-CSH) provides a more comprehensive quantitative approach to metabolic heterogeneity by plotting the percentage curves of tumor volume at different SUV thresholds (thresholds ranging from 0\u0026ndash;100% of SUVmax). These methods have significantly improved the accuracy and reliability of tumor metabolic heterogeneity studies. Accumulating evidence suggests that high metabolic heterogeneity as assessed by \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT is an important predictor of poor prognosis in a variety of malignancies, such as oropharyngeal squamous cell carcinoma, pancreatic adenocarcinoma, non-small cell lung cancer and uterine leiomyosarcoma \u003csup\u003e[\u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/sup\u003e. A recent study further confirmed that high metabolic heterogeneity has significant prognostic value in patients with newly diagnosed primary mediastinal B-cell lymphoma \u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e, which is the first pivotal study in which metabolic heterogeneity has been identified as a prognostic biomarker for malignant lymphoma.\u003c/p\u003e \u003cp\u003eThe difference between our study of DLBCL and previous studies is that the method of assessing metabolic heterogeneity is different, and we use a linear regression slope method to assess the metabolic heterogeneity of DLBCL. Because TMTV and TLG under different SUV thresholds can reflect the inherent heterogeneity of DLBCL to a certain extent. This study uses only the 40%-80% SUV threshold, excluding non-existent intra-tumoral regions or areas with very low FDG uptake due to tumor necrosis, etc., to obtain a more accurate estimate of the metabolic distribution. Several studies have shown that histopathological features of malignancies such as non-small cell lung cancer, oligodendroglioma, head and neck squamous cell carcinoma and oral cavity cancer are significantly associated with intra-tumoral heterogeneity of \u003csup\u003e18\u003c/sup\u003eF-FDG uptake on PET \u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. In addition, \u003csup\u003e18\u003c/sup\u003eF-FDG heterogeneity in malignant sarcoma correlates with patient prognosis \u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. Recently, a study on the heterogeneity of intra-tumoral \u003csup\u003e18\u003c/sup\u003eF-FDG uptake in nasopharyngeal carcinoma showed that it was significantly associated with tumor aggressiveness and was associated with various outcome measures \u003csup\u003e[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTherefore, by utilizing different PET metabolic parameters to reflect the intricate internal heterogeneity of DLBCL, it is possible to facilitate the generation of innovative and prognostic imaging biomarkers. However, to the best of our knowledge, there are currently very limited studies on the application of heterogeneity factors in PET image analysis to identify high-risk DLBCL patients, and there are no studies using a linear regression slope approach to evaluate the prognostic value of intra-tumoral heterogeneity in PET patients with DLBCL. Therefore, the purpose of this study was to extract heterogeneous factors of metabolic parameters from PET images to evaluate the prognostic value of DLBCL patients, and to study the relationship between heterogeneous factors and prognostic parameters such as maximum standardized uptake value (SUVmax), total metabolic tumor volume (TMTV), and total lesion glycolysis (TLG).\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eFrom January 2013 to December 2020, we retrospectively enrolled 245 histologically confirmed patients with DLBCL who had undergone PET/CT imaging scans at the Affiliated Cancer Hospital of Harbin Medical University prior to treatment. As this study is retrospective, it is not necessary to submit evidence of informed consent, and the hospital's institutional review board approved the study. The following criteria were required for inclusion: (1) histopathologically confirmed DLBCL; (2) no previous history of cancer; (3) \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT performed less than two weeks prior to the first treatment; (4) No anti-tumor therapy before the scan; and (5) availability of clinical and follow-up data. To determine their disease status, the case received anthracycline-based chemotherapy followed by PET/CT scans and scans using \u003csup\u003e18\u003c/sup\u003eF-FDG. Following the first 2 years of treatment, followed by follow-up assessments every 3 months and then every 6 months, the primary endpoint of the study is the patient's PFS, which can be defined as the period between diagnosis and the date of first recurrence, progression, or death due to any cause. At the last known follow-up, cases without any events were removed.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePET/CT images\u003c/h3\u003e\n\u003cp\u003ePrior to the PET/CT examination, patients were required to fast for 6\u0026ndash;8 hours and maintain blood glucose levels below 11.1 mmol/L to ensure optimal \u003csup\u003e18\u003c/sup\u003eF-FDG uptake. The imaging procedure was performed using a Discovery 690 Elite PET/CT scanner (GE Healthcare) with \u003csup\u003e18\u003c/sup\u003eF-FDG of radiochemical purity\u0026thinsp;\u0026gt;\u0026thinsp;98%. Following intravenous administration of \u003csup\u003e18\u003c/sup\u003eF-FDG (3.7 MBq/kg), patients rested for approximately 60 minutes to allow adequate radiotracer distribution before imaging. The PET/CT acquisition protocol encompassed anatomical coverage from the skull base to the mid-thigh. Initial CT scanning was performed using the following parameters: 120 kV, 140 mA, 1.25 mm slice spacing, and 3.75 mm slice thickness, with a scan duration of 20\u0026ndash;30 seconds. Subsequently, PET imaging was acquired in 3D mode across 6\u0026ndash;7 bed positions, with 2.5 minutes per bed position while maintaining consistent patient positioning. Image reconstruction was performed using the ordered subset expectation maximization (OSEM) algorithm with CT-based attenuation correction. The reconstructed PET and CT datasets were transferred to a Xeleris\u0026trade; workstation (GE Healthcare) for co-registration and fusion processing, enabling comprehensive anatomical and metabolic evaluation.\u003c/p\u003e\n\u003ch3\u003eMeasurement of PET metabolic parameters\u003c/h3\u003e\n\u003cp\u003ePET/CT images were transferred to Advantage Workstation 4.5 (GE Healthcare) and reviewed by two experienced radiologists. The region of interest (ROI) is then extracted at a threshold of 41% SUVmax and manually adjusted if there is a high background or high uptake area (e.g., bladder, heart) in the vicinity of the lesion, or if the uptake rate in the lesion is low \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. Volumetric metabolic parameters based on systemic multi-focus were achieved using Life-x software (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.lifexsoft.org/.version\u003c/span\u003e\u003cspan address=\"http://www.lifexsoft.org/.version\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e 7.6), which automatically represented the volume of interest (VOI) using an SUV-based contour threshold method. Various thresholds are used to determine VOI boundaries; The relative thresholds are 40\u0026ndash;80% of SUVmax, respectively. In this study, TMTV was automatically calculated by taking 40\u0026ndash;80% of SUVmax in VOI, expressed as TMTV (40%), TMTV (50%), TMTV (60%), TMTV (70%), and TMTV (80%), respectively. TLG is calculated using 40\u0026ndash;80% of the SUVmax threshold; The results were specified as TLG (40%), TLG (50%), TLG (60%), TLG (70%) and TLG (80%), respectively.\u003c/p\u003e \u003cp\u003eIn addition, we assessed the internal heterogeneity of \u003csup\u003e18\u003c/sup\u003eF-FDG uptake in DLBCL, represented by the heterogeneity factor (HF). HF is a heterogeneity index calculated based on the slope of PET/CT metabolic parameters (TMTV/TLG) as a function of SUV threshold, which is used to assess the non-uniformity of metabolic distribution within DLBCL tumors and may provide additional information for prognostic prediction. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea shows a representative image of the linear regression slope using various SUV thresholds. The TMTV at each SUV threshold was recorded, and then HF1 was calculated as a negative slope for linear regression on the threshold-volume curve (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb records the TLG at each SUV threshold and then calculates the negative slope of HF2 as the linear regression on the threshold-glycolysis curve (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). Thus, higher HF values indicate a faster decline in TMTV or TLG with an increase in SUV threshold, i.e., a more uneven distribution of metabolic activity within the tumor (higher heterogeneity), and lower HF values indicate a more homogeneous metabolic distribution (lower heterogeneity). In addition, other metabolic parameters within the ROI were calculated using Life-x software, including the maximum lesion diameter (Size), the maximum distance between lesions (Dmax), and the maximum normalized uptake value (SUVmax).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eStatistical analysis was performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R statistical software (version 4.4.1). The statistical significance is set to a \u003cem\u003eP\u003c/em\u003e-value of less than 0.05. Progression-free survival (PFS) was used as the endpoint to evaluate the prognosis of patients with DLBCL. The optimal thresholds for SUVmax, TMTV, TLG, HF1, and HF2 were determined using the Receiver operating characteristic (ROC) curve in the training cohort. Cox regression analysis was used to explore the prognostic value of potential independent factors, and the model was constructed. Survival was assessed by the Kaplan-Meier method and compared by the log-rank test.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePatients characteristics\u003c/h2\u003e \u003cp\u003eThe basic statistics of the cases are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. A total of 245 patients with DLBCL were enrolled, including 114 males and 131 females. The mean age of participants was 57 years (range 13 to 93 years). Cases were divided into training cohorts and validation cohorts in a 7:3 ratio. According to the data analysis in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, there was no statistical difference between the training cohort and the validation cohort (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The median follow-up time was 37.98 months and 32.31 months, respectively. During the follow-up period, the ratio of recurrence to non-recurrence cases was 3:2, and the information for patients with recurrence and non-recurrence was shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, and the analysis showed that there were statistically significant differences between the recurrence and non-recurrence groups in all variables except sex, age, and platelets (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe baseline characteristics of patients with DLBCL in the training and validation cohorts\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eValidation(n\u0026thinsp;=\u0026thinsp;74)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTraining(n\u0026thinsp;=\u0026thinsp;171)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOverall (n\u0026thinsp;=\u0026thinsp;245)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value*\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eECOG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56(22.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e146(59.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e202(82.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.067\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e18(7.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e25(10.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e43(17.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e58(23.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e153(62.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e211(86.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.121\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16(6.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e18(7.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e34(13.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003emale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e27(11.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e87(35.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e114(46.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.138\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003efemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e47(19.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e84(34.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e131(53.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnn-Arbor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eI-II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e24(9.80%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e85(34.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e109(44.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.112\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIII-IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50(20.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e86(35.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e136(55.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIPI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e38(15.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e116(47.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e154(62.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.214\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e36(14.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e55(22.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e91(37.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e39(15.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e120(49.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e159(64.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.109\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e35(14.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e51(20.80%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e86(35.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50(20.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e130(53.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e180(73.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.169\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e24(9.80%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e41(16.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e65(26.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRecurrence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e22(9.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e70(28.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92(37.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.196\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e52(21.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e101(41.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e153(62.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;60y\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40(16.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e94(38.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e134(54.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.895\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;60y\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e34(13.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e77(31.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e111(45.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e17.10(6.28, 29.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e25.60(11.10, 33.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22.90(8.80, 32.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.136\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSUVmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.45(2.85, 9.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.10(2.60, 6.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.40(2.70, 8.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.134\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDmax40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e29.57(13.56, 52.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e26.79(7.45, 57.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e27.49(9.66, 56.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.536\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTMTV40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e158.34(158.34, 307.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e100.49(34.73, 201.85)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e103.75(38.99, 241.26)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.121\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTLG40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2219.30(465.39, 5055.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1143.93(344.92, 2901.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1513.33(359.60, 3234.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.141\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3.77(0.99, 6.93)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.15(0.80, 4.80)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.30(0.87, 5.60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.118\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e47.78(10.00, 114.34)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e24.21(6.78, 61.86)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e28.51(7.51, 77.81)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.131\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e*\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05, Chi-square test.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe baseline characteristics of patients with DLBCL in the recurrence and no-recurrence cohorts\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo recurrence(n\u0026thinsp;=\u0026thinsp;92)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRecurrence(n\u0026thinsp;=\u0026thinsp;153)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0verall(n\u0026thinsp;=\u0026thinsp;245)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value*\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eECOG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e83(33.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e119(48.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e202(82.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9(3.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34(13.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e43(17.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e86(35.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e125(51.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e211(86.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6(2.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28(11.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e34(13.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.831\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003emale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e42(17.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e72(29.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e114(46.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003efemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50(20.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e81(33.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e131(53.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnArbor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eI-II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e57(23.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e52(21.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e109(44.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIII-IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e35(14.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e101(41.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e136(55.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIPI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e71(29.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e83(33.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e154(62.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e21(8.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e70(28.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e91(37.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e76(31.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e83(33.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e159(64.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16(6.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e70(28.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e86(35.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.056\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e74(30.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e106(43.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e180(73.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e18(7.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e47(19.20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e65(26.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.132\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;60y\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56(22.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e78(31.80%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e134(54.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;60y\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e36(14.70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e75(30.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e111(45.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e19.90(9.30, 28.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e23.75(8.30, 35.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e22.90(8.80, 32.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.036\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSUVmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.72(1.925, 4.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.01(3.20, 13.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.40(2.70, 8.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDmax40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e23.01(5.52, 37.76)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e38.17(14.08, 61.17)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e27.49(9.66, 56.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTMTV40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e96.09(22.45, 106.46)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e257.08(62.63, 312.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e103.75(38.99, 241.26)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTLG40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1309.56(167.23, 1192.02)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3646.45(889.70, 5108.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1513.33(359.60, 3234.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.27(0.51, 2.42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.91(1.40, 7.20)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.30(0.87, 5.60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e30.96(3.58, 25.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e81.68(18.47, 115.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e28.51(7.51, 77.81)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e*\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05, Chi-square test.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eHF predicts the performance of recurrence and survival risk stratification\u003c/h3\u003e\n\u003cp\u003eIn the training and validation cohorts, HF1 and HF2 showed high accuracy in predicting disease recurrence. The optimal cut-off values for HF1 and HF2 were determined by analysis, which were 0.390 and 0.232, respectively. Based on these cut-off values, HF1 and HF2 were divided into low and high groups, and the results showed that there was a statistically significant difference in progression-free survival (PFS) between the low and high HF1 and HF2 groups in the training and validation cohorts (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In addition, the optimal cut-off values for SUVmax, Size, Dmax, TMTV and TLG were 27.1, 4.1 cm, 46.6 cm, 133.1 cm\u0026sup3; and 827.6, respectively.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eCorrelation of HF with other variables\u003c/h3\u003e\n\u003cp\u003eThe heterogeneity factor HF1 was significantly associated with TMTV (r\u0026thinsp;=\u0026thinsp;0.9976; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), TLG (r\u0026thinsp;=\u0026thinsp;0.9217; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), and SUVmax (r\u0026thinsp;=\u0026thinsp;0.5425; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), but not with size (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). The heterogeneity factor HF2 was significantly correlated with TLG (r\u0026thinsp;=\u0026thinsp;0.9965; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), TMTV (r\u0026thinsp;=\u0026thinsp;0.9288; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), SUVmax (r\u0026thinsp;=\u0026thinsp;0.5874; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), and Size (r\u0026thinsp;=\u0026thinsp;0.2531; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Heterogeneity factor HF1 was also significantly correlated with heterogeneity factor HF2 (r\u0026thinsp;=\u0026thinsp;0.9296; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). HF1 and HF2 were strongly correlated with TMTV and TLG, respectively, and HF1 and TLG and HF2 were also strongly correlated with TMTV (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSingle/multivariate analysis results\u003c/h2\u003e \u003cp\u003eUnivariate Cox regression analysis was performed based on the clinical variables, metabolic parameters and heterogeneity factor HF of the training cohort samples, and the results are detailed in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. On this basis, multivariate Cox regression analysis was further carried out, including forward, backward and stepwise regression methods. Statistically significant variables in univariate analyses were included in multivariate analyses, and the results are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Multivariate analysis showed that IPI, Size, SUVmax, TLG, HF1 and HF2 were all independent prognostic factors for predicting PFS and were statistically significant.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariate and multivariate analyses results on variables of the training cohort samples\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHR(95%CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP value*\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHR(95%CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value*\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20\u0026ndash;1.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.063\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnArbor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.21\u0026ndash;2.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.29\u0026ndash;1.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.203\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDmax40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.00-1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eECOG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.27\u0026ndash;1.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20\u0026ndash;1.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIPI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20\u0026ndash;2.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.06\u0026ndash;2.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.21\u0026ndash;1.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.22\u0026ndash;1.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.01\u0026ndash;1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.04\u0026ndash;1.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSUVmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.01\u0026ndash;1.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.04\u0026ndash;1.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTLG40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.00\u0026ndash;1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTMTV40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.00\u0026ndash;1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.01\u0026ndash;1.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.0006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.06\u0026ndash;1.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHF2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.00\u0026ndash;1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.95\u0026ndash;0.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003e*\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05, Chi-square test. N/A, not assessed\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eConstruction and evaluation of multi-parameter model nomograms\u003c/h2\u003e \u003cp\u003eThe multiparametric model was constructed by integrating heterogeneity indices (HF1, HF2), metabolic indexes (Size, SUVmax, and TLG), and clinical variables (IPI). To evaluate the gain-effect of heterogeneity indices in prognostic prediction, we also constructed a traditional model that included only metabolic indicators and clinical variables. In addition, a model that included only IPI was established for comparison with the International Prognostic Index (IPI), which is routinely used in clinical practice.\u003c/p\u003e \u003cp\u003eTo facilitate routine use by clinicians, the nomogram demonstrates the ability of a multiparametric model to predict the risk of disease recurrence at 2, 3, and 5 years (Fig.\u0026nbsp;5). To demonstrate the clinical application of the nomogram, we show maximum intensity projection images of \u003csup\u003e18\u003c/sup\u003eF-FDG PET scans for two typical DLBCL cases (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003e). As successfully predicted by the nomogram, the prognosis of the first case (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003ea, b, and c) showed no recurrence after standard treatment 5.2 years after diagnosis. Similar to the nomogram prediction for the second case with a higher risk of recurrence (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003ed, e, and f), disease progression was observed at 2 months after standard treatment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBy plotting a calibration curve, calculate the degree of fit between the actual case results and the nomogram predictions. The results show calibration curves for 2-, 3-, and 5-year progression-free survival (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e7\u003c/span\u003ea). The calibration curve of the model with higher accuracy is closer to the diagonal dashed line, indicating that the predicted values are in good agreement with the clinical observations. In addition, we configured a time-dependent AUC curve for the prediction model (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e7\u003c/span\u003eb). The C-index of the multiparameter model was 0.729 (95% CI: 0.680\u0026ndash;0.778), which was higher than that of the traditional model and the IPI model (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The analysis showed that both the multi-parameter model and the traditional model reached the highest C-index at 2 years. In order to prevent the model from overfitting, the model results were corrected several times by using the Bootstrap method, and the C index was stable at about 0.7. In addition, decision curve analysis further demonstrates that the multiparameter model results in a greater overall net benefit than other competing models across most risk thresholds (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e7\u003c/span\u003ec). The prediction performance of the model in the validation cohort is still better than that of the traditional model and the IPI model (Supplementary Fig.\u0026nbsp;1). These results indicate that the multi-parameter model significantly improves the accuracy and clinical practicability of recurrence risk prediction in patients with DLBCL after integrating heterogeneity indicators and provides strong support for the implementation of precision medicine.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe Harrell\u0026rsquo;s C-index results in the training and validation cohorts\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProgression-free survival\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTraining cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003evalidation cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eC-index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95%CI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eC-index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e95%CI\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMultiparametric model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.729\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.680\u0026ndash;0.778\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.685\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.609\u0026ndash;0.761\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConventional model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.643\u0026ndash;0.745\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.636\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.560\u0026ndash;0.712\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIPI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.598\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.549\u0026ndash;0.647\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.539\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.464\u0026ndash;0.613\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAt the same time, in order to more intuitively compare the prediction results of 2-year and 3-year progression-free survival (PFS), Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e8\u003c/span\u003e shows the prediction performance of the HF model for 2-year and 3-year PFS. The results showed that the HF1 model was better than the 3-year PFS (AUC\u0026thinsp;=\u0026thinsp;0.712) in predicting 2-year PFS (AUC\u0026thinsp;=\u0026thinsp;0.729), while the HF2 model was better than the 2-year PFS (AUC\u0026thinsp;=\u0026thinsp;0.768) in predicting 3-year PFS (AUC\u0026thinsp;=\u0026thinsp;0.774). It is worth noting that the prediction performance of the HF2 model is better than that of the HF1 model, whether it is predicting 2-year or 3-year PFS. In order to further evaluate the predictive power of the HF model combined with the International Prognostic Index (IPI) for 2-year and 3-year PFS, we constructed a combined HF1-IPI and HF2-IPI models, respectively. Figure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e8\u003c/span\u003e shows the prediction results of the two joint models: the predicted AUC values of the HF1-IPI model for 2-year and 3-year PFS were 0.717 and 0.698, respectively, and the predicted AUC values of the HF2-IPI model for 2-year and 3-year PFS were 0.739 and 0.736, respectively. The results show that compared with the HF model alone, the combination of IPI indicators does not significantly improve the prediction performance of the model. In addition, similar results were obtained in the validation cohort (Supplementary Fig.\u0026nbsp;2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn summary, in this retrospective study of 245 patients, we developed and validated imaging biomarkers based on heterogeneous metabolic parameters that predict survival outcomes after treatment based on PET images. In this study, tumor metabolic parameters and clinical indexes were confirmed, including SUVmax, TLG, Size and IPI (HR\u0026thinsp;=\u0026thinsp;1.07, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; HR\u0026thinsp;=\u0026thinsp;1.00, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; HR\u0026thinsp;=\u0026thinsp;1.06, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; HR\u0026thinsp;=\u0026thinsp;1.63, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which was an independent predictor of PFS in patients with DLBCL. To improve prediction performance, we integrated heterogeneous factors (i.e., HF1, HF2), metabolic parameters (i.e., Size, SUVmax, and TLG), and clinical indicators (i.e., IPI), and designed a multiparameter model containing multidimensional prognostic information. Calibration curves showed a strong correlation between predicted and actual outcomes of PFS at 2, 3, and 5 years for multiparametric models, and DLBCL results were traditionally assessed only on the basis of PFS and/or overall survival (OS). Previous studies have used two-year PFS as the endpoint for disease-associated DLBCL immunochemotherapy outcomes \u003csup\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e. However, based on recent studies suggesting that three-year event-free survival (EFS) is a better indicator of long-term prognosis than two-year EFS in patients with newly diagnosed DLBCL \u003csup\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e, this study focused on evaluating the predictive power of two- and three-year progression-free survival (PFS) rather than overall survival (OS). The results showed that the HF1 model performed better than the 3-year PFS in predicting 2-year PFS, while the HF2 model showed better predictive power in predicting 3-year PFS. It is worth noting that the overall prediction performance of the HF2 model is significantly better than that of the HF1 model in both 2-year and 3-year PFS predictions. Therefore, the HF2 model is considered to be a superior prediction tool compared with the HF1 model. However, unfortunately, the prediction performance of the model was not significantly improved after combining the HF model with the IPI index. Furthermore, according to the results of the correlation between the variables in this study, the heterogeneity factor HF1 was significantly correlated with TMTV, TLG and SUVmax, but not with tumor size (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Notably, HF1 correlates more strongly with TMTV and TLG than with other variables. The heterogeneity factor HF2 was significantly correlated with TLG, TMTV, SUVmax and Size (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Since HF2 exhibited a strong correlation with all variables included in the model, this may explain the results that HF2 is better predictive than HF1. In addition, due to the calculation method of heterogeneity factors, the correlation between HF1 and HF2 and TMTV and TLG, respectively, was significantly higher than that of other variables. At the same time, significant correlations between HF1 and TLG and between HF2 and TMTV were also observed. This may indicate that there is some similarity in the trends of TMTV and TLG with SUV thresholds, leading to a strong correlation between HF1 and HF2, which is consistent with previous studies \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. In addition, in the multivariate prediction model, the 2-year PFS performed satisfactorily in both the validation and training sets (training cohort: 2-year PFS AUC\u0026thinsp;=\u0026thinsp;0.729, validation cohort: 2-year PFS AUC\u0026thinsp;=\u0026thinsp;0.685). At the same time, compared with the traditional model and IPI, the C-index results of the multi-parameter model showed better prognostic performance, and the C-index values of the training cohort and validation cohort predicted PFS were 0.729 and 0.685, respectively. By introducing the variables of the heterogeneity factor HF to explain the impact of false-positive and false-negative patients on prognostic judgment, we observed that the multiparametric model has superior clinical utility over the traditional model and IPI, so as to achieve personalized prognostic assessment and customized tumor treatment. Our results highlight the potential of PET-based heterogeneous metabolic parameter analysis to advance the prognostic assessment of patients with DLBCL and ultimately improve clinical outcomes.\u003c/p\u003e \u003cp\u003eThe microenvironment of human tumors is inherently heterogeneous. This heterogeneity increases the complexity of our understanding of tumor biological behavior and increases the challenges of treatment planning \u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e. Since FDG PET depicts voxel-based glucose metabolism in tumors, it can be used to study this aspect of tumor heterogeneity. Studies have been conducted to investigate the relationship between heterogeneity in FDG uptake and histopathological heterogeneity in vivo and ex vivo tumors. Henriksson et al. found that intra-tumoral heterogeneity in FDG uptake in nude mice with head and neck squamous cell carcinoma xenografts depended on the distribution of different tissue components \u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]\u003c/sup\u003e. Specifically, regions that predominantly contain tumor cells (and therefore fewer stromal cells) and regions with less necrosis have significantly higher FDG uptake. Pugachev et al. reported that higher FDG uptake indicated the presence of hypoxic regions in nude mouse prostate tumors \u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. This heterogeneity in FDG uptake suggests that there are cell populations with different metabolic rates within the tumor, which may be due to differences in characteristics such as growth rate, vascularization, necrosis, etc. Among the \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT parameters, SUVmax is the most commonly used quantitative analysis parameter, and SUVmax has shown significant prognostic value in DLBCL. In addition, TMTV and TLG are of great significance in predicting the prognosis of DLBCL \u003csup\u003e[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]\u003c/sup\u003e. However, there is some controversy regarding the use of metabolic parameters such as SUVmax for assessment. Heterogeneity of tumors, partial volume effects, timing of SUV assessment, and body size may disrupt the assessment of metabolic parameters to reflect accurate tumor characteristics \u003csup\u003e[\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]\u003c/sup\u003e. Volume parameters (TMTV and TLG) are expected to be more efficient than metabolic parameters (SUVmax) because tumor burden is considered using volume parameters. Previous studies using direct comparisons of metabolic and volumetric parameters have maintained that volumetric parameters are superior to metabolic parameters in predicting DLBCL \u003csup\u003e[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]\u003c/sup\u003e. A number of studies have shown that metabolic heterogeneity indexes based on volume parameters can predict tumor heterogeneity and disease treatment effect to a certain extent. As an example, heterogeneity in FDG uptake has been reported in the literature to predict treatment outcomes for certain cancer types, including cervical cancer and sarcoma \u003csup\u003e[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e. Tisier et al. found that tumor metabolic heterogeneity with pre-treatment baseline FDG PET stratified treatment response prediction better than SUVmax and SUVmean in patients with esophageal cancer \u003csup\u003e[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. In other studies, pretreatment metabolic heterogeneity has been found to be significantly associated with tumor volume, SUVmax, and T stage of nasopharyngeal carcinoma, and can be used to predict response to treatment and patient survival (overall and disease-free survival) \u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e. DLBCL is a clinically and biologically diverse disease entity, and patient response to treatment exhibits heterogeneity. Given the 30\u0026ndash;40% recurrence rate and mortality rate, it is critical to identify patients with DLBCL at high risk of recurrence who could benefit from intensive systemic chemotherapy. The unique role of heterogeneity factors in the assessment of DLBCL response may be due to the correlation between heterogeneity in tumor metabolism and spatial heterogeneity in tumor response to chemotherapy \u003csup\u003e[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e. The resulting heterogeneous response may have a further impact on treatment outcomes and survival. If we are able to predict a poor prognosis by preoperative \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT, prolonging the duration of first-line therapy or changing the agent is another treatment option, and close follow-up can be performed to detect recurrence early.\u003c/p\u003e \u003cp\u003eThere are several possible methods to assess heterogeneity using \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT. One of the most widely used methods is texture analysis. Accumulating evidence suggests that tumor heterogeneity as measured by \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT texture analysis is associated with treatment response and prognosis for esophageal, lung, head and neck, breast, and cervical cancers \u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e. In addition, entropy is considered the best predictive parameter for pancreatic ductal adenocarcinoma among several structural analysis parameters \u003csup\u003e[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]\u003c/sup\u003e. However, there are currently no PET structural analysis parameters that are widely accepted for measuring tumor heterogeneity. In addition, texture analysis is difficult to evaluate in clinical practice due to the difficulty of obtaining measurement results. Compared with texture analysis, our method of using linear regression slope is more convenient and suitable for application in clinical practice. In addition, methods such as coefficient of variance, coefficient of variation, and AUC-CSH have been used to assess metabolic heterogeneity within tumors \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e. However, in lymphoma, there have been no studies using a linear regression slope to characterize disease heterogeneity.\u003c/p\u003e \u003cp\u003eThe heterogeneity index (HF) derived from the linear regression slope has a unique added value in the assessment of tumor heterogeneity. Compared with simple prognostic parameters such as Dmax (i.e., the maximum tumor diameter), HF can not only quantify metabolic heterogeneity within tumors, but also reflect the spatial heterogeneity distribution characteristics of tumors, thereby providing more comprehensive heterogeneity information, whereas Dmax can only reflect the unidimensional characteristics of tumors and cannot fully assess tumor heterogeneity \u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. In addition, compared to complex analytical methods employing artificial intelligence (AI) or radiomics, HF is computationally simple and easy to perform clinically, while still providing predictive performance comparable to complex methods. Therefore, HF has significant advantages in balancing computational complexity with clinical practicability, especially in clinical scenarios where rapid assessment of tumor heterogeneity and prognosis is required. However, AI and radiomics approaches are able to integrate more dimensions of data and may have more potential when dealing with highly complex patterns of heterogeneity. Consistent with previous studies \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e, the heterogeneous factors HF1 (HR\u0026thinsp;=\u0026thinsp;1.2, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and HF2 (HR\u0026thinsp;=\u0026thinsp;0.97, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) derived from PET images in this study were reliable prognostic biomarkers for PFS. Heterogeneous metabolic parameters extracted from medical images may be able to reflect multi-level disease heterogeneity and thus may have higher power in diagnosis and outcome prediction compared with a single metabolic profile \u003csup\u003e[\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/sup\u003e. In conclusion, our study supports the use of heterogeneous metabolic parameters from PET/CT images as a valuable tool to identify high-risk patients with DLBCL who require intensive treatment or close monitoring. Overall, HF has important application value in clinical practice as a heterogeneity assessment tool of medium complexity.\u003c/p\u003e \u003cp\u003eHowever, there are limitations and shortcomings in this study. First, due to the nature of retrospective studies, inherent selection limitations are unavoidable. Secondly, with regard to the calculation methods of TMTV, different calculation methods of TMTV may have an impact on the results of heterogeneity analysis, especially in the low SUV threshold region, TMTV is very sensitive to threshold changes \u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e. Small differences may be magnified, leading to \"exponential\" variations in heterogeneous results. Although the selection of a threshold range of 40 to 80 percent for SUVs (which may not include a region of low SUV thresholds), the use of relative thresholds for SUVs instead of absolute thresholds (which may be closer to the actual distribution of metabolic activity) \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e, and the use of a uniform TMTV calculation method may reduce the impact of differences in TMTV calculation methods on the results, this approach will inevitably deviate from the true value to some extent. Thirdly, with regard to the calculation method of heterogeneity factors, a potential drawback of heterogeneity factor calculation is the association between intra-tumoral heterogeneity and tumor volume and total lesion glycolysis, which may affect the independence of parametric heterogeneity factors and the robustness of the method. Fourth, although histological studies have shown a good spatial correlation between FDG uptake regions and histological outcomes, there is no reference standard for the overall measurement of heterogeneity. We were unable to verify the accuracy of our results as reflecting the true extent of tumor heterogeneity. Future work may focus on improving heterogeneity factor calculations and assessing their impact on treatment planning by addressing these shortcomings. Finally, our data is limited to cases in one medical center and is smaller. Therefore, clinical support for this predictive model is limited. Future studies will require larger-scale external validation from multiple medical centers.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003col\u003e\n\u003cli\u003e\n\u003cp\u003eHeterogeneity factors (HFs) can reflect the metabolic heterogeneity of FDG to a certain extent.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eHeterogeneity factors (HFs) can be used as a stable imaging biomarker for prognosis prediction of DLBCL patients and can achieve survival risk stratification of patients.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThe prediction results of the multi-parameter model are better than those of the traditional model and the IPI model and may accurately predict the recurrence risk of DLBCL patients.\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis paper is supported by the Provincial Key Research and Development Program of Heilongjiang Province (GA21C001 KW), Distinguished Young Scientist Funding of Harbin Medical University Affiliated Tumor Hospital (JCQN2019-02), Key Project of Harbin Medical University Cancer Hospital Climbing Funding (PDYS2024-03), Key Innovation Technology Project Harbin Medical University Cancer Hospital Innovation Technology Funding (CXJSZD-2023-04),Heilongjiang Province's \"unveiling and leading\" technology research and development project (2022ZXJ03C01).The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData sharing statement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated during and analyzed during the current study are not publicly available due to patient privacy concerns but are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed Consent\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe need of written informed consent was waived in this study due to the retrospective nature.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGuarantor\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe scientific guarantor of this publication is Kezheng Wang.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistics and Biometry\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;No complex statistical methods were necessary for this paper\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFan Ge and Tingting Wu contributed equally to this work. Tingting Wu contributed to the study conception and design. Material preparation and data collection were performed by Tingting Wu, Xinyue Yang and Fan Ge. The first draft of the manuscript and statistics analysis were performed by Fan Ge. Data curation was performed by Mengye Peng, Chen Yang and Fan Ge. Reviewing and editing were performed by Kezheng Wang. And all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistics\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and biometry\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOne of the authors (Fan Ge) has significant statistical expertise.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe retrospective study was reviewed and approved by Harbin Medical University Cancer Hospital Medical Ethics Committee. All methods were in accordance with the ethical standards as laid down in the Declaration of Helsinki and its later amendments or comparable ethical standards. As this was a retrospective study, formal consent was not obtained.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank GE Healthcare Company for providing technical support for our manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLi S, Young KH, Medeiros LJ (2018) Diffuse Large B-Cell Lymphoma Pathol 50(1):74\u0026ndash;87\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVodicka P, Klener P, Trneny M (2022) Diffuse Large B-Cell Lymphoma (DLBCL): Early Patient Management and Emerging Treatment Options. OTT Volume 15, 1481\u0026ndash;1501\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoletto S, Novo M, Paruzzo L et al (2022) Treatment Strategies for Patients with Diffuse Large B-Cell Lymphoma. Cancer Treat Rev 110:102443\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSehn LH, Berry B, Chhanabhai M et al (2007) The Revised International Prognostic Index (R-IPI) Is a Better Predictor of Outcome than the Standard IPI for Patients with Diffuse Large B-Cell Lymphoma Treated with R-CHOP. Blood 109(5):1857\u0026ndash;1861\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCeriani L, Gritti G, Cascione L et al (2020) SAKK38/07 Study: Integration of Baseline Metabolic Heterogeneity and Metabolic Tumor Volume in DLBCL Prognostic Model. Blood Adv 4(6):1082\u0026ndash;1092\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNicolson GL (1984) Generation of Phenotypic Diversity and Progression in Metastatic Tumor Cells. Cancer Metast Rev 3(1):25\u0026ndash;42\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTumor Heterogeneity Morphological, Molecular and Clinical Implications. Histology and Histopathology 2000, No. 33, 881\u0026ndash;898\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGillies RJ, Schomack PA, Secomb TW et al (1999) Causes and Effects of Heterogeneous Perfusion in Tumors. Neoplasia 1(3):197\u0026ndash;207\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChapuy B, Stewart C, Dunford AJ et al (2018) Molecular Subtypes of Diffuse Large B Cell Lymphoma Are Associated with Distinct Pathogenic Mechanisms and Outcomes. Nat Med 24(5):679\u0026ndash;690\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcGranahan N, Swanton C (2017) Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell 168(4):613\u0026ndash;628\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKostakoglu L, Chauvie S (2019) PET-Derived Quantitative Metrics for Response and Prognosis in Lymphoma. PET Clin 14(3):317\u0026ndash;329\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen S, He K, Feng F et al (2019) Metabolic Tumor Burden on Baseline 18F-FDG PET/CT Improves Risk Stratification in Pediatric Patients with Mature B-Cell Lymphoma. Eur J Nucl Med Mol Imaging 46(9):1830\u0026ndash;1839\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVergote VKJ, Verhoef G, Janssens A et al (2024) International Metabolic Prognostic Index Is Superior to Other Metabolic Tumor Volume-Based Prognostication Methods in a Real-Life Cohort of Diffuse Large B-Cell Lymphoma. Journal of nuclear medicine: official publication. Society of Nuclear Medicine\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShagera QA, Cheon GJ, Koh Y et al (2019) Prognostic Value of Metabolic Tumour Volume on Baseline 18F-FDG PET/CT in Addition to NCCN-IPI in Patients with Diffuse Large B-Cell Lymphoma: Further Stratification of the Group with a High-Risk NCCN-IPI. Eur J Nucl Med Mol Imaging 46(7):1417\u0026ndash;1427\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMikhaeel NG, Heymans MW, Eertink JJ et al (2022) Proposed New Dynamic Prognostic Index for Diffuse Large B-Cell Lymphoma: International Metabolic Prognostic Index. J Clin Oncol 40(21):2352\u0026ndash;2360\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui S, Xin W, Wang F et al (2024) Metabolic tumour area: a novel prognostic indicator based on 18F-FDG PET/CT in patients with diffuse large B-cell lymphoma in the R-CHOP era. BMC Cancer 24(1):895\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCottereau AS, Meignan M, Nioche C et al (2021) Risk stratification in diffuse large B-cell lymphoma using lesion dissemination and metabolic tumor burden calculated from baseline PET/CT\u0026dagger;. Ann Oncol 32(3):404\u0026ndash;411\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoellaard R, O'Doherty MJ, Weber WA et al (2010) FDG PET and PET/CT: EANM procedure guidelines for tumour PET imaging: version 1.0. Eur J Nucl Med Mol Imaging 37(1):181\u0026ndash;200\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePugachev A, Ruan S, Carlin S et al (2005) Dependence of FDG Uptake on Tumor Microenvironment. Int J Radiation Oncology*Biology*Physics 62(2):545\u0026ndash;553\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWarburg O, Wind F, Negelein E, THE METABOLISM OF TUMORS IN THE BODY (1927) J Gen Physiol 8(6):519\u0026ndash;530\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVander Heiden MG (2011) Targeting Cancer Metabolism: A Therapeutic Window Opens. Nat Rev Drug Discov 10(9):671\u0026ndash;684\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl Tameemi W, Dale TP, Al-Jumaily RMK et al (2019) Hypoxia-Modified Cancer Cell Metabolism Front Cell Dev Biol 7:4\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorin RD, Arthur SE, Hodson DJ (2022) Molecular Profiling in Diffuse Large B-cell Lymphoma: Why so Many Types of Subtypes? Br J Haematol 196(4):814\u0026ndash;829\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Velden FHP, Cheebsumon P, Yaqub M et al (2011) Evaluation of a Cumulative SUV-Volume Histogram Method for Parameterizing Heterogeneous Intratumoural FDG Uptake in Non-Small Cell Lung Cancer PET Studies. Eur J Nucl Med Mol Imaging 38(9):1636\u0026ndash;1647\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMena E, Taghipour M, Sheikhbahaei S et al (2017) Value of Intratumoral Metabolic Heterogeneity and Quantitative 18F-FDG PET/CT Parameters to Predict Prognosis in Patients With HPV-Positive Primary Oropharyngeal Squamous Cell Carcinoma. Clin Nucl Med 42(5):e227\u0026ndash;e234\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim Y, Kim YJ, Paeng JC et al (2017) Heterogeneity Index Evaluated by Slope of Linear Regression on 18F-FDG PET/CT as a Prognostic Marker for Predicting Tumor Recurrence in Pancreatic Ductal Adenocarcinoma. Eur J Nucl Med Mol Imaging 44(12):1995\u0026ndash;2003. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00259-017-3755-8\u003c/span\u003e\u003cspan address=\"10.1007/s00259-017-3755-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCeriani L, Milan L, Martelli M et al (2018) Metabolic Heterogeneity on Baseline 18FDG-PET/CT Scan Is a Predictor of Outcome in Primary Mediastinal B-Cell Lymphoma. Blood 132(2):179\u0026ndash;186\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng N-M, Dean Fang Y-H, Tung-Chieh Chang J et al (2013) Textural Features of Pretreatment18 F-FDG PET/CT Images: Prognostic Significance in Patients with Advanced T-Stage Oropharyngeal Squamous Cell Carcinoma. J Nucl Med 54(10):1703\u0026ndash;1709\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwon SH, Yoon J, An Y, Shin YS et al (2014) Prognostic Significance of the Intratumoral Heterogeneity of 18 F-FDG Uptake in Oral Cavity Cancer. J Surg Oncol 110(6):702\u0026ndash;706\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee J-W, Park J-Y, Lee HJ et al (2018) Preoperative [18F]FDG PET/CT Tumour Heterogeneity Index in Patients with Uterine Leiomyosarcoma: A Multicentre Retrospective Study. Eur J Nucl Med Mol Imaging 45(8):1309\u0026ndash;1316\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang B, Chan T, Kwong DL-W et al (2012) Nasopharyngeal Carcinoma: Investigation of Intratumoral Heterogeneity With FDG PET/CT. Am J Roentgenol 199(1):169\u0026ndash;174\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAide N, Fruchart C, Nganoa C et al (2020) Baseline 18F-FDG PET Radiomic Features as Predictors of 2-Year Event-Free Survival in Diffuse Large B Cell Lymphomas Treated with Immunochemotherapy. Eur Radiol 30(8):4623\u0026ndash;4632\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIzumiyama K, Inao T, Goto H et al (2024) Event-Free Survival at 36 Months Is a Suitable Endpoint for Diffuse Large B-Cell Lymphoma Patients Treated with Immunochemotherapy: Real-World Evidence from the North Japan Hematology Study Group. Haematologica\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeppner GH, Miller BE (1983) Tumor Heterogeneity: Biological Implications and Therapeutic Consequences. Cancer Metast Rev 2(1):5\u0026ndash;23\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenriksson E, Kjellen E, Wahlberg P et al (2007) 2-Deoxy-2-[18F] Fluoro-D-Glucose Uptake and Correlation to Intratumoral Heterogeneity. Anticancer Res 27(4B):2155\u0026ndash;2159\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKidd EA, Grigsby PW (2008) Intratumoral Metabolic Heterogeneity of Cervical Cancer. Clin Cancer Res 14(16):5236\u0026ndash;5241\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEary JF, O\u0026rsquo;Sullivan F, O\u0026rsquo;Sullivan J et al (2008) Spatial Heterogeneity in Sarcoma18 F-FDG Uptake as a Predictor of Patient Outcome. J Nucl Med 49(12):1973\u0026ndash;1979\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTixier F, Le Rest CC, Hatt M et al (2011) Intratumor Heterogeneity Characterized by Textural Features on Baseline18 F-FDG PET Images Predicts Response to Concomitant Radiochemotherapy in Esophageal Cancer. J Nucl Med 52(3):369\u0026ndash;378\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan WKS, Mak HKF, Huang B et al (2010) Nasopharyngeal Carcinoma: Relationship between 18F-FDG PET-CT Maximum Standardized Uptake Value, Metabolic Tumour Volume and Total Lesion Glycolysis and TNM Classification. Nucl Med Commun 31(3):206\u0026ndash;210\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHatt M, Majdoub M, Valli\u0026egrave;res M et al (2015) 18F-FDG PET Uptake Characterization Through Texture Analysis: Investigating the Complementary Nature of Heterogeneity and Functional Tumor Volume in a Multi\u0026ndash;Cancer Site Patient Cohort. J Nucl Med 56(1):38\u0026ndash;44\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHyun SH, Kim HS, Choi SH et al (2016) Intratumoral Heterogeneity of 18F-FDG Uptake Predicts Survival in Patients with Pancreatic Ductal Adenocarcinoma. Eur J Nucl Med Mol Imaging 43(8):1461\u0026ndash;1468\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanoun S, Rossi C, Berriolo-Riedinger A et al (2014) Baseline Metabolic Tumour Volume Is an Independent Prognostic Factor in Hodgkin Lymphoma. Eur J Nucl Med Mol Imaging 41(9):1735\u0026ndash;1743\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThie JA (2004) Understanding the Standardized Uptake Value, Its Methods, and Implications for Usage. J Nucl Med 45(9):1431\u0026ndash;1434\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSasanelli M, Meignan M, Haioun C et al (2014) Pretherapy Metabolic Tumour Volume Is an Independent Predictor of Outcome in Patients with Diffuse Large B-Cell Lymphoma. Eur J Nucl Med Mol Imaging 41(11):2017\u0026ndash;2022\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSenjo H, Hirata K, Izumiyama K, North Japan Hematology Study Group et al (2020) High Metabolic Heterogeneity on Baseline 18FDG-PET/CT Scan as a Poor Prognostic Factor for Newly Diagnosed Diffuse Large B-Cell Lymphoma. Blood Adv 4(10):2286\u0026ndash;2296\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"annals-of-hematology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aohe","sideBox":"Learn more about [Annals of Hematology](http://link.springer.com/journal/277)","snPcode":"277","submissionUrl":"https://submission.nature.com/new-submission/277/3","title":"Annals of Hematology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"[18F]FDG PET/CT, Metabolic parameters, Diffuse large B-cell lymphoma, Progression-free survival, Heterogeneity factor","lastPublishedDoi":"10.21203/rs.3.rs-6070367/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6070367/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eObjective\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study aimed to evaluate the predictive value of intra-tumoral \u003csup\u003e18\u003c/sup\u003eF-FDG metabolic heterogeneity in patients with diffuse large B cell lymphoma (DLBCL) in terms of survival.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe retrospectively included 245 patients with DLBCL who underwent \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT prior to treatment and analyzed using total metabolic tumor volume (TMTV) and total lesion glycolysis (TLG) as metabolic volume parameters. The linear regression slopes of TMTV and TLG were calculated according to different percentages of SUV thresholds (i.e., 40%, 50%, 60%, 70%, and 80%), respectively, defined as Heterogeneity Factor-1 (HF1) and Heterogeneity Factor-2 (HF2). These indices of heterogeneity were used to predict progression-free survival (PFS). Based on the results of the Cox proportional hazards model, we constructed a multi-parameter prediction model and evaluated the model in the training and validation cohorts by calibration curve, consistency index (C-index) and decision curve analysis (DCA).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eClinicopathological and PET/CT data from 245 patients were reviewed. 153 patients (62.4%) experienced relapse after treatment. Comparing relapsed and non-relapse patients, all \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT parameters and heterogeneity index showed significant differences. There were significant differences in survival risk stratification according to HF1 and HF2 cut-off classifications (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001). In multivariate Cox regression analysis, SUVmax (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001), TLG (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001), HF1 (\u003cem\u003eP\u003c/em\u003e=0.004), and HF2 (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001) showed significant results. Among the clinicopathological parameters, IPI (\u003cem\u003eP\u003c/em\u003e=0.027) and Size (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001) were selected as important parameters.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHF1 and HF2 obtained by the linear regression slope of MTV and TLG may be a novel and useful prognostic marker in DLBCL, which can achieve survival-risk stratification of patients. In addition, multiparametric models have the potential to effectively predict the risk of recurrence in patients.\u003c/p\u003e","manuscriptTitle":"Feasibility analysis of metabolic parameters based on baseline 18 F-FDG PET/CT to predict heterogeneity and recurrence of diffuse large B-cell lymphoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-21 08:09:47","doi":"10.21203/rs.3.rs-6070367/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accepted","date":"2025-05-12T14:15:45+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-21T09:09:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"132884968232080406082698145421514486288","date":"2025-04-21T08:19:22+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-17T09:11:54+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-04-17T07:03:12+00:00","index":"","fulltext":""},{"type":"submitted","content":"Annals of Hematology","date":"2025-04-11T09:00:15+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"annals-of-hematology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aohe","sideBox":"Learn more about [Annals of Hematology](http://link.springer.com/journal/277)","snPcode":"277","submissionUrl":"https://submission.nature.com/new-submission/277/3","title":"Annals of Hematology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"55cfa659-c611-46a6-96d9-90769f3c3a46","owner":[],"postedDate":"April 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-06-16T16:02:18+00:00","versionOfRecord":{"articleIdentity":"rs-6070367","link":"https://doi.org/10.1007/s00277-025-06409-8","journal":{"identity":"annals-of-hematology","isVorOnly":false,"title":"Annals of Hematology"},"publishedOn":"2025-06-11 15:57:49","publishedOnDateReadable":"June 11th, 2025"},"versionCreatedAt":"2025-04-21 08:09:47","video":"","vorDoi":"10.1007/s00277-025-06409-8","vorDoiUrl":"https://doi.org/10.1007/s00277-025-06409-8","workflowStages":[]},"version":"v1","identity":"rs-6070367","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6070367","identity":"rs-6070367","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}