An integrated prognosis prediction model based on real-word clinical characteristics for immunotherapy in advanced esophageal squamous cell carcinoma

An integrated prognosis prediction model based on real-word clinical characteristics for immunotherapy in advanced esophageal squamous cell carcinoma | 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 Article An integrated prognosis prediction model based on real-word clinical characteristics for immunotherapy in advanced esophageal squamous cell carcinoma Zhihao Lu, Liyuan Dong, Yue Ma, Guang Cao, Dongze Chen, Fengxiao Dong, and 15 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5322833/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction The efficacy of immune checkpoint inhibitors (ICIs) for advanced esophageal squamous cell carcinoma (ESCC) remains suboptimal. This study aims to construct and validate a clinically accessible model to better identify populations that may potentially benefit from ICIs. Methods This study enrolled advanced ESCC patients treated with ICIs at Peking University Cancer Hospital from January 14, 2016, to January 26, 2024, forming the training cohort. Combined positive score (CPS) was recorded to evaluate the predictive value of programmed cell death ligand-1 (PD-L1). Baseline clinical characteristics and laboratory test results were identified as predictors through a 2-phase selection based on Cox proportional hazard regression and minimization of Akaike information criterion (AIC). The prediction model was internally validated using bootstrapping and externally validated in patients from Harbin Medical University Cancer Hospital between January 10, 2019, and July 6, 2022. Results A total of 430 patients from Peking University Cancer Hospital and 184 patients from Harbin Medical University Cancer Hospital were ultimately enrolled. PD-L1 expression failed to discriminate survival outcomes (HR=0.94, 95% CI: 0.74-1.19, P = .6 ). The final model incorporates 10 variables: stage, bone metastasis, line of therapy, treatment, lactate dehydrogenase, carcinoembryonic antigen, carbohydrate antigen 199, lymphocyte count, prognostic nutritional index, and systemic immune-inflammation index. The C-index was 0.725 (95%CI: 0.694-0.756) in the training cohort, 0.722 (95%CI: 0.688-0.751) after bootstrapping, and 0.691 (95%CI: 0.650-0.733) in the external validation cohort, outperforming PD-L1 in prognostic prediction and risk stratification. An interactive online prediction tool (https://escc-survival.shinyapps.io/shiny_app/) was subsequently developed. Conclusions This is the first model for individualized survival prediction in advanced ESCC patients treated with ICIs based on large-scale, high-quality real-world data, potentially guiding clinical decision-making and optimize treatment strategies. Biological sciences/Cancer/Cancer therapy/Cancer immunotherapy Health sciences/Medical research/Biomarkers/Prognostic markers Health sciences/Oncology/Cancer/Gastrointestinal cancer/Oesophageal cancer Advanced esophageal squamous cell carcinoma immune checkpoint inhibitors prediction model real-world data Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Esophageal cancer ranks as the seventh most common and sixth leading cause of cancer death worldwide 1 . China accounts for over half of the global incidence and mortality rates of esophageal cancer, with 90% of cases being esophageal squamous cell carcinoma (ESCC) 2 , and the five-year survival rate remaining below 5% 3 . Although the advent of immune checkpoint inhibitors (ICIs), such as programmed cell death protein-1(PD-1)/programmed cell death ligand-1 (PD-L1) inhibitors and cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) inhibitors has revolutionized the treatment landscape 4 – 6 and emerged as the standard option for advanced ESCC 7 , only a subset of patients benefits from ICIs. Even when combined with chemotherapy, the 5-year survival rate remains at a mere 10–15% 8 , emphasizing the urgent need for effective predictors to identify populations that could potentially benefit from ICIs. Traditional biomarkers including tumor mutational load (TMB) 9 and PD-L1 expression 10 have been investigated. However, their predictive value in ESCC remains unsatisfactory. Particularly, the role of PD-L1 expression in the efficacy of ICIs is fraught with uncertainty and contradictions in different clinical trials 11 , 12 . Recently, several multi-gene signatures, such as fibroblast-associated signature 13 , immunogenic cell death (ICD)-associated gene panel 14 and angiogenesis-associated risk score 15 have been developed as predictive models and have demonstrated some prognostic utility. Nevertheless, these predictors either rely on tumor tissue and complex molecular analyses, or have not been trained and validated in large-scale and real-world datasets, leading to considerable limitations in effectiveness, clinical accessibility and reliability. Therefore, there is an urgent need to construct an integrated model that incorporates the most direct, accessible and non-invasive clinical biomarkers, enhancing efficiency and cost-effectiveness while enabling rapid and precise prognostic predictions for patients treated with ICIs. In this study, we aim to construct and validate a multivariable model for individualized prediction of prognosis and survival in ESCC patients treated with ICIs, based on large-scale and real-world clinical data with long-term follow-up. Furthermore, we expect to provide insights that will enable clinicians and patients to decide upon the optimal ICIs treatment for advanced ESCC. Materials and Methods Study Population and Follow-Up Data for the training cohort were obtained from patients with advanced ESCC treated at the Department of Gastrointestinal Oncology, Peking University Cancer Hospital, between January 14, 2016, and January 26, 2024. The external validation cohort comprised patients from Harbin Medical University Cancer Hospital, between January 10, 2019, and July 6, 2022. Inclusion criteria were: (1) pathologically confirmed ESCC; (2) unresectable advanced, recurrent (including postoperative recurrence), or metastatic ESCC without prior systemic therapy; (3) treated with mono-ICIs, ICIs combined with chemotherapy or targeted therapy; (4) at least one measurable lesion as assessed by response evaluation criteria in solid tumours1.1 (RECIST 1.1) 16 ; and (5) availability of complete clinical and follow-up data. Exclusion criteria included: (1) patients lost to follow-up after only one cycle of ICIs; (2) patients who did not undergo regular imaging evaluations (at least once every two cycles); and (3) patients with severe organ dysfunction or significant abnormalities in blood counts, liver or kidney function, or autoimmune diseases. Efficacy Assessment Patients were evaluated for efficacy by imaging scans after every two cycles of treatment. Clinical responses to treatment were determined according to RECIST 1.1, which includes complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Follow-up was conducted to ascertain patients' survival status (alive or deceased). The primary endpoint of this study was overall survival (OS), defined as the time from the first dose of ICIs to death from any cause. Variable Selection and Model Construction Potential variables were categorized into 4 groups: patient characteristics (age, sex, smoking, etc.), tumor characteristics (staging, metastasis of lymph node, liver, lung and bone), treatment-related data (line of therapy, treatment options, etc.) and baseline laboratory test results obtained within 7 days prior to ICIs [e.g., lactate dehydrogenase (LDH), hemoglobin (Hb), lymphocytes count (ALC)]. A total of 26 variables were initially identified as potential predictors for subsequent analysis. Variables with more than 10% of values missing were not included for analysis. The missing values for carcinoembryonic antigen (CEA) and carbohydrate antigen 199 (CA199) in the validation cohort were 4.3% and 5.4%, respectively, with missing data imputed using multiple interpolation methods. All other variables had complete data with no missing values. Optimal cutoff values were determined based on clinical reference ranges or were calculated using the “surv_cutpoint” algorithm from the “survminer” R package (Supplementary Table 1). Univariate Cox regression analyses assessed the independent prognostic value of these 26 variables. Variables finally included in multivariable Cox regression model were selected based on clinical significance and minimization of Akaike information criterion (AIC), which considers both the statistical goodness of fit and the number of parameters, imposing a penalty for excessive parameters 17 . A nomogram was subsequently developed to visualize individual OS probabilities at 1-, 2-, and 3-year. Assessment and Validation of the Model The discrimination of the prognostic model was assessed by calculating the Harrell’s concordance index (C-index) 18 . Calibration curves were plotted to compare model-predicted survival estimates with actual Kaplan-Meier survival probabilities 19 . Time-dependent ROC curves were plotted using the “timeROC” R package to evaluate predictive accuracy at specific time points (1-, 2-, and 3-year) 20 . The area under the curve (AUC) and its 95% confidence intervals were computed. Bootstrapping with 1000 resample was employed for internal validation 21 , and the accuracy and generalizability of the model were examined in the validation cohort. Risk stratification For each patient, the risk score was calculated based on the nomogram using the “predict” algorithm. Tertiles of risk score served as cutoff values for risk stratification, classifying patients into low-, moderate-, and high-risk groups. Kaplan-Meier survival curves for these three risk groups were plotted and compared using the log-rank test 22 . Data processing and statistical analysis All data processing, statistical analysis and figure plotting were conducted with SPSS 26.0 and R 4.3.0. Two-sided P-values < .05 were considered statistically significant. Results Patient’s characteristics We screened a total of 513 patients with advanced ESCC who received ICIs from Peking University Cancer Hospital between January 14, 2016, and January 26, 2024, along with 213 patients from Harbin Cancer Hospital between January 10, 2019, and July 6, 2022. Following the established inclusion and exclusion criteria, 430 patients were ultimately enrolled in the training cohort, while 184 patients constituted the validation cohort (Fig. 1 ). The median follow-up time was 29.7 months (95% CI: 28.1–31.3 months) in the training cohort, during which 282 (65.5%) deaths occurred. In the validation cohort, the median follow-up was 43.2 months (95% CI: 39.9–46.6 months), with 152 (82.6%) deaths recorded The background characteristics are detailed in Table 1 . The median age of patients in both cohorts was 61.0 years, with a higher proportion of females in the training cohort. No significant differences were observed in disease stage or metastasis (to the liver, lung, and bone) between the two cohorts. In the training cohort, 140 (32.6%) patients received second-line or later treatment, significantly more than 34 (18.5%) patients in the validation cohort. Additionally, in the training cohort, 95 (22.1%) received mono-ICIs, while 273 (63.5%) patients received ICIs combined with chemotherapy and 62 (14.4%) patients received ICIs combined with targeted therapy. In contrast, most patients (91.8%) in the validation cohort were treated with ICIs combined with chemotherapy (Table 1 ). Table 1 Selected Patient Characteristics in the Training and Validation Cohorts Characteristics Patients, No. (%) P-value Training Cohort (n = 430) Validation Cohort (n = 200) Age Median (IQR) 61 (56.0–66.0) 61 (56.0-66.8) 0.812 Sex Male 383 (89.1) 180 (97.8) Female 47 (10.9) 4 (2.2) < 0.001 Smoke No 106 (24.7) 66 (35.9) Yes 324 (75.3) 118 (64.1) 0.005 Alcohol No 148 (34.4) 61 (33.2) Yes 282 (65.6) 123 (66.8) 0.762 Stage III 24 (5.6) 10 (5.4) IV 406 (94.4) 174 (94.6) 0.942 Liver metastasis No 344 (80.0) 156 (84.8) Yes 86 (20.0) 28 (15.2) 0.163 Lung metastasis No 328 (76.3) 147 (79.9) Yes 102 (23.7) 37 (20.1) 0.327 Bone metastasis No 382 (88.8) 171 (92.9) Yes 48 (11.2) 13 (7.1) 0.120 Line of therapy First-line 290 (67.4) 150 (81.5) ≥Second-line 140 (32.6) 34 (18.5) < 0.001 Treatment Mono-ICIs/ ICIs + targeted therapy 157 (36.5) 15 (8.2) ICIs + chemotherapy 273 (63.5) 169 (91.8) 5 100 (23.3) 38 (20.7) 0.479 CA199 ≤ 37 389 (90.5) 167 (90.8) > 37 41 (9.5) 17 (9.2) 0.909 LDH ≤ 240 341 (79.3) 163 (88.6) > 240 89 (20.7) 21 (11.4) 0.006 ALC ≤ 1.7 309 (71.9) 110 (59.8) > 1.7 121 (28.1) 74 (40.2) 0.003 PNI ≤ 49.50 196 (45.6) 115 (62.5) > 49.50 234 (54.4) 69 (37.5) 589.41 281 (65.3) 111 (60.3) 0.235 The percentages of patients reflect the proportion of patients within each subgroup of specified characteristics. Categorical variables were analyzed using Chi-square test, while continuous variables were assessed using t-test. Survival outcomes. In the training cohort, the median OS for first-line treatment was 21.3 months (95% CI: 17.0-27.1 months), with 2-, 3-, and 5-year OS rates of 47.3%, 29.8%, and 20.2%, respectively. In contrast, the median OS for second-line or later treatment was 10.2 months (95% CI: 8.6–13.4 months), with corresponding OS rates of 16.6%, 8.2%, and 5.4%. (Fig. 2 A). In the validation cohort, the median OS for first-line treatment was 15.6 months (95% CI: 13.1–18.8 months), with 2-, 3-, and 5-year OS rates of 36.7%, 25.1%, and 12%. the median OS for second-line or later treatment was 8.5 months (95% CI: 6.3–12.5 months), with OS rates of 17.6%, 11.0%, and 0.0%. (Supplementary Fig. 2A). Among different treatment regions in the training cohort, the median OS was significantly higher in the ICIs combined with chemotherapy group (24.6 months, 95% CI: 17.6–28.5) compared to the ICIs combined with targeted therapy (10.3 months, 95% CI 8.4–13.2, P < .001 ) and mono-ICIs groups (10.2 months, 95% CI: 7.9–14.0, P < .001 ). However, no noticeable difference in OS was observed between the ICIs combined with targeted therapy group and the mono-ICIs group (HR = 1.07, 95% CI: 0.76–1.50, P = 0.701 , Fig. 2 B). A similar trend was noted in the validation cohort, although the treatment regimen was overwhelmingly ICIs combined with chemotherapy (Supplementary Fig. 2B). Additionally, we stratified patients into high PD-L1 expression [combined positive score (CPS) ≥ 10] and low PD-L1 expression (CPS < 10) groups based on a CPS cutoff value of 10 to evaluate the potential of PD-L1 as a prognostic biomarker. Surprisingly, PD-L1 failed to differentiate OS between the two groups (HR = 0.94, 95% CI: 0.74–1.19, P = .6 , Fig. 2 C), indicating that PD-L1 expression was not a reliable prognostic predictor in the training cohort. Construction and Validation of the Prognosis Prediction Model Previous studies have demonstrated that factors influencing survival and prognosis in ESCC extend beyond tumor characteristics (stage, metastasis, etc.) and treatment regimen. Tumor markers such as CEA and CA199 23 , along with laboratory markers such as inflammatory markers [LDH 24 , neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) 24 , systemic immune-inflammation index (SII) 25 , etc.], and nutritional markers [albumin, prognostic nutritional index (PNI) 26 , etc.], also play significant roles. However, these markers have not been widely discussed or utilized in predictive models for clinical immunotherapy. In our study, covering the above indicators, a total of 26 candidate predictive variables were entered into a multivariable Cox regression model (Supplementary Table 1). After stepwise selection based on the AIC, stage, bone metastasis, line of therapy, treatment, LDH, CEA, CA199, ALC, PNI, and SII were retained in the final prediction model. The model confirmed that stage III, first-line ICIs combined with chemotherapy and elevated levels of PNI (> 49.50) (HR = 0.73, 95%CI: 0.56–0.95, P = 0.019 ) were associated with superior survival outcomes. Furthermore, elevated levels of CEA (> 5) (HR = 1.35, 95%CI: 1.03–1.79, P = 0.030 ), CA199 (> 37) (HR = 1.75, 95%CI: 1.17–2.62, P = 0.006 ), LDH (> 240) (HR 1.45, 95%CI: 1.09–1.91, P = 0.010 ) and SII (> 589.41) (HR = 1.63, 95%CI: 1.23–2.16, P < 0.001 ) were associated with a higher risk of death and poorer prognosis. Bone metastasis (HR = 1.34, 95%CI 0.94–1.91, P = 0.108 ) and ALC (≤ 1.7) (HR = 0.73, 95%CI: 0.51–1.04, P = 0.078 ) was also linked to inferior survival, although not reaching statistical significance (Supplementary Table 2). A risk prediction nomogram was developed based on this model (Fig. 3 ). According to the points values to each variable, stage had the most significant impact on prognosis, despite only 24 (5.5%) patients being classified as stage III. The calibration plot for both the training and validation cohorts demonstrated optimal concordance between nomogram predictions and actual observation for 2-, and 3-year OS, remaining within an acceptable range for survival probability prediction. However, the 1-year OS prediction exhibited moderate discrimination, potentially due to the inherent characteristics of the training cohort (Supplementary Fig. 3). Further exploration with larger cohorts is necessary to enhance precision. The model in the training cohort achieved a C-index of 0.725 (95%CI 0.694–0.756) and the bootstrap-corrected C-index was 0.722 (95% CI 0.688–0.751), indicating a robust level of discriminative ability for survival prediction. The time-dependent ROC curve analysis further underscored the model's strong predictive performance for 1-, 2-, and 3-year OS, with AUC values reflecting excellent discrimination, thereby confirming the reliability of the model in prognostic predictions across these time points (Fig. 4 A). To test the generalizability of the prediction model, we evaluated its performance in the external validation cohort, which yielded a C-index of 0.691 (95%CI: 0.650–0.733). The AUC values were consistent with those observed in the training cohort, further validating the predictive value of the model in real-world settings (Fig. 4 B). These internal and external validation results highlight the robustness and reliability of this model, demonstrating consistent performance across various follow-up time intervals. Risk Stratification of Mortality for ESCC Patients To facilitate the practical application of this model in clinical settings, we calculated the risk score for each patient in both the training and validation cohorts based on the total points from the nomogram. Patients were subsequently stratified into low-, moderate-, and high-risk groups according to the tertiles of risk scores from the training cohort. In the training cohort, the high-risk group exhibited the lowest 1-, 2- and 3-year survival probabilities at 37.5%, 9.8% and 1.9%, compared to 69.2%, 38.1% and 21.4% for the moderate-risk group and 83.6%, 65.0% and 45.8% for the low-risk group. In the validation cohort, the high-risk group showed 29.2%, 8.3% and 0% survival probabilities, while the moderate-risk group had probabilities of 55.7%, 22.8% and 13.9% and the low-risk group for 77.8%, 50.6% and 36.7% (Supplementary Table 3). Kaplan-Meier curves for the 3 risk groups in both the training and validation cohorts demonstrated significant differentiation (both log-rank P < .001 ). (Fig. 5 ). Discussion In this study, we reviewed a large cohort of 430 patients with advanced ESCC treated with ICIs and described their survival outcomes. Subsequently, we constructed and validated an integrated model incorporating 10 variables to predict OS based on comprehensive clinical characteristics. This model demonstrated strong discrimination ability (C-index = 0.725), providing real-world evidence that patients treated with ICIs in combination with chemotherapy, along with high PNI (> 49.5), low LDH (≤ 240), CEA (≤ 5), CA199 (≤ 37) and SII levels (≤ 589.41), were more likely to benefit from the treatment. Finally, we integrated this model into an easy-to-use online prediction tool (Supplementary Fig. 5, https://escc-survival.shinyapps.io/shiny_app/ ) featuring an interactive interface for parameter entry and graphical output. PD-L1 expression has been investigated as a promising predictive marker for response and efficacy of ICIs in advanced ESCC 27 . However, current evidence regarding its predictive value is fraught with conflicts and ambiguities. For instance, CheckMate-648 28 indicated that ICIs extend OS only in patients with high PD-L1 expression [tumor cell proportion score (TPS) > 1%], while RATIONALE-302 29 , ORIENT-15 7 and two meta-analyses 11 , 12 suggested that those with low PD-L1 expression (TPS < 1% or CPS < 10) also experienced OS benefits. In this study, we evaluated various ICIs including sintilimab, pembrolizumab, nivolumab, and camrelizumab. Generally, PD-L1 failed to demonstrate any predictive value (Fig. 2 C, Supplementary Fig. 4), indicating that PD-L1 may not accurately and reliably predict prognosis for ICIs treatment, thus necessitating further investigation into its clinical utility as a biomarker. Conversely, our prediction model exhibited superior survival and prognostic capabilities for both mono-ICIs and in combination with chemotherapy, significantly outperforming PD-L1. This model integrates comprehensive array of factors across four categories, encompassing fundamental tumor characteristics, treatment strategies, and biological predictors such as LDH, CEA, CA199, ALC, PNI and SII. To our knowledge, this is the first model designed for individualized survival prediction in advanced ESCC patients treated with ICIs, based on large-scale, high-quality real-world data. These clinical predictors stemmed from essential tumor evaluations and laboratory tests, ensuring convenience, cost-effectiveness, and clinical accessibility. This study guarantees the representativeness of the patient population and facilitates a thorough exploration of the interactions between clinical characteristics and survival outcomes for immunotherapy, ultimately informing personalized therapeutic decision-making. Tumor characteristics, such as stage and metastasis, are critical in predicting prognosis for ESCC. In our training cohort, patients with stage IV and concomitant bone metastases exhibited a poorer prognosis. Additionally, line of therapy also greatly influenced patient outcomes. The median OS for first-line treatment in the training cohort reached 21 months, with 2-, 3-, and 5-year survival probabilities of 47.3%, 29.8%, and 20.2%, surpassing the average level. In contrast, the validation cohort exhibited slightly lower median OS, reflecting regional differences in clinical practice, therapeutic concepts and treatment experience between the two study centers. Regarding treatment regimens, ICIs combined with chemotherapy 7 , 8 , 12 , 30 , 31 or targeted inhibitors such as epidermal growth factor receptor (EGFR) inhibitors and tyrosine kinase inhibitors (TKIs) 32 , 33 , are widely recognized options. However, few studies have compared these therapies simultaneously. Our findings demonstrated that survival and prognosis were best in patients treated with ICIs combined with chemotherapy, while no considerable difference was observed between ICIs combined with targeted therapy and mono-ICIs, underscoring the need for further clinical trials to validate these targets. The role of biological predictors, including serum tumor markers, biochemical markers, inflammatory markers and nutritional markers, cannot be underestimated in assessing prognosis and survival, especially during ICI therapy. Limited research has focused on the association between tumor markers and therapeutic effects in ESCC. A small sample study suggested that CEA and CA199 might predict the efficacy of postoperative chemotherapy 23 . Our study further indicated that CEA and CA199 can serve as prognostic markers for ICIs therapy in advanced ESCC. Elevated circulating levels of LDH, a key enzyme in anaerobic glycolysis, correlate with adverse prognosis 24 , 34 , which was consistent with our findings. Additionally, circulating neutrophils, lymphocytes, and platelets can reflect the body’s immune inflammatory state 35 , linking to tumorigenesis, progression, and metastasis 36 . Considering discussions regarding inflammatory biomarkers such as PLR, NLR and SII in ESCC remain inconsistent due to sample size constraints 37 , 38 , our analysis enrolled ALC, PLR, NLR (dNLR), lymphocyte-to-white blood cell ratio (LWR) and SII in multivariable Cox regression. Ultimately, only ALC and SII were retained in the final model, with high SII (> 589.61) identified as a poor prognostic indicator for OS. PNI, which combines serum albumin levels and peripheral blood lymphocyte counts, reflects both immune and nutritional status 39 . Consistent with prior study 40 , we found that a high PNI (> 49.5) was associated with improved survival outcomes, highlighting the importance of timely nutritional and immunological interventions. The prediction model integrates above variables demonstrates strong performance. There was no deterioration in calibration during internal validation (C-index = 0.722), indicating that the model was likely not overfitted. More importantly, the model also performed well (C-index = 0.691) in the external validation cohort from a geographically distant area in China, despite some expected loss of discriminative performance. Using this prediction model, patients could be stratified into low-, moderate-, and high-risk groups based on their risk scores. OS demonstrated clear differentiation among the three groups in both the training and validation cohorts (Fig. 5 and Supplementary Table 3), highlighting the model's clinical significance, especially in the absence of standardized criteria on patient selection for ICIs. To facilitate the implementation of this model in clinical practice, we integrated our model into an interactive online risk calculator (Supplementary Fig. 5, https://escc-survival.shinyapps.io/shiny_app/ ), which provides automatic prediction of patient outcomes. This tool could serve as a valuable adjunct for treatment decision-making, particularly in institutions where ICIs are introduced as a novel therapeutic option and during discussions with patients. It can also aid in patients’ selection for clinical trials based on predicted outcomes, allowing for stratified randomization into low-, moderate-, and high-risk groups. Enriching trial participation with patients most likely to respond can lead to smaller sample sizes, thereby reducing costs and time, while minimizing the risk of exposing patients to ineffective treatments. Several limitations of the present study warrant acknowledgment. Firstly, the single center design, small sample size, and retrospective nature of the analysis introduce potential selection bias. Therefore, a larger, multi-center cohort is necessary to estimate and validate this prediction model with adequate statistical power. Additionally, while our findings align with previous studies, the lack of standardized cutoff values complicates the generalization of our conclusions to other studies. Consequently, establishing a unified testing method to confirm appropriate cutoff values is essential. In summary, based on real-world clinical data, we developed the first integrated prognosis prediction model for immunotherapy and integrated it into an interactive online tool. This model may provide individualized survival predictions for advanced ESCC patients treated with ICIs, enabling clinicians and patients to estimate the survival outcomes and make decisions more accurately and efficiently in the future. Declarations Acknowledgments This study was supported by Beijing Natural Science Foundation (Z220022) and Beijing Hospitals Authority Clinical Medicine Development of special funding support (ZLRK202326). The authors thank all research team members and participants of each cohort study for their contribution to this research. Ethics Statement This retrospective study was approved by the Medical Ethics Committee of Peking University Cancer Hospital and Harbin Medical University Cancer Hospital, and informed consent was obtained from all patients. Participants gave informed consent to participate in the study before taking part. Funding This study was supported by Beijing Natural Science Foundation (Z220022) and Beijing Hospitals Authority Clinical Medicine Development of special funding support (ZLRK202326). CRediT Authorship Contribution Statement Zhihao Lu: Corresponding author, Conceptualization, Funding acquisition, Writing-review and editing, Final approval of the manuscript. Zhonghu He and Yanqiao Zhang: Corresponding author, Conceptualization, Data curation, Writing-review and editing, Final approval of the manuscript. Liyuan Dong: Investigation, Conceptualization, Data curation, Methodology, Software, Formal analysis, Validation, Visualization, Writing-original draft, Writing-review and editing, Final approval of the manuscript. Yue Ma: Resources, Data curation, Methodology, Validation, Writing-review and editing, Final approval of the manuscript. Guang Cao: Data curation, Writing-review and editing, Final approval of the manuscript. Dongze Chen: Methodology, Writing-review and editing, Final approval of the manuscript. Xi Jiao, Fengxiao Dong, Yanshuo Cao, Chang Liu, Yanni Wang, Na Zhuo, Fengyuan Wang, Yixuan Guo, Tingting Dai, Shuwei Zhang, Hao Jiao, Xingyue Zou, Jian Li and Lin Shen:Data curation, Writing-review and editing, Final approval of the manuscript. No potential conflict of interest was reported by the authors References Sung H, Ferlay J, Siegel RL et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin May 71(3):209–249. 10.3322/caac.21660 Liang H, Fan JH, Qiao YL (2017) Epidemiology, etiology, and prevention of esophageal squamous cell carcinoma in China. Cancer Biol Med Feb 14(1):33–41. 10.20892/j.issn.2095-3941.2016.0093 Zeng H, Chen W, Zheng R et al (2018) Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health May 6(5):e555–e567. 10.1016/S2214-109X(18)30127-X Kojima T, Shah MA, Muro K et al (2020) Randomized Phase III KEYNOTE-181 Study of Pembrolizumab Versus Chemotherapy in Advanced Esophageal Cancer. J Clin Oncol Dec 10(35):4138–4148. 10.1200/JCO.20.01888 Doki Y, Ajani JA, Kato K et al (2022) Nivolumab Combination Therapy in Advanced Esophageal Squamous-Cell Carcinoma. N Engl J Med Feb 3(5):449–462. 10.1056/NEJMoa2111380 Luo H, Lu J, Bai Y et al (2021) Effect of Camrelizumab vs Placebo Added to Chemotherapy on Survival and Progression-Free Survival in Patients With Advanced or Metastatic Esophageal Squamous Cell Carcinoma: The ESCORT-1st Randomized Clinical Trial. JAMA Sep 14(10):916–925. 10.1001/jama.2021.12836 Lu Z, Wang J, Shu Y et al (2022) Sintilimab versus placebo in combination with chemotherapy as first line treatment for locally advanced or metastatic oesophageal squamous cell carcinoma (ORIENT-15): multicentre, randomised, double blind, phase 3 trial. Bmj Apr 19 377:e068714. 10.1136/bmj-2021-068714 Shah MA, Sun J-M, Shen L et al (2024) First-line pembrolizumab (pembro) plus chemotherapy (chemo) for advanced esophageal cancer: 5-year outcomes from the phase 3 KEYNOTE-590 study. American Society of Clinical Oncology Lu Z, Chen H, Li S et al (2020) Tumor copy-number alterations predict response to immune-checkpoint-blockade in gastrointestinal cancer. J Immunother Cancer Aug 8(2). 10.1136/jitc-2019-000374 Kojima T, Muro K, Francois E et al (2019) Pembrolizumab versus chemotherapy as second-line therapy for advanced esophageal cancer: Phase III KEYNOTE-181 study. American Society of Clinical Oncology Yap DWT, Leone AG, Wong NZH et al (2023) Effectiveness of Immune Checkpoint Inhibitors in Patients With Advanced Esophageal Squamous Cell Carcinoma: A Meta-analysis Including Low PD-L1 Subgroups. JAMA Oncol Feb 1(2):215–224. 10.1001/jamaoncol.2022.5816 Wu HX, Pan YQ, He Y et al (2023) Clinical Benefit of First-Line Programmed Death-1 Antibody Plus Chemotherapy in Low Programmed Cell Death Ligand 1-Expressing Esophageal Squamous Cell Carcinoma: A Post Hoc Analysis of JUPITER-06 and Meta-Analysis. J Clin Oncol Mar 20(9):1735–1746. 10.1200/JCO.22.01490 Ren Q, Zhang P, Zhang X et al (2023) A fibroblast-associated signature predicts prognosis and immunotherapy in esophageal squamous cell cancer. Front Immunol 14:1199040. 10.3389/fimmu.2023.1199040 Zhang H, Luo YB, Wu W et al (2021) The molecular feature of macrophages in tumor immune microenvironment of glioma patients. Comput Struct Biotechnol J 19:4603–4618. 10.1016/j.csbj.2021.08.019 Gao M, Wu X, Jiao X et al (2024) Prognostic and predictive value of angiogenesis-associated serum proteins for immunotherapy in esophageal cancer. J Immunother Cancer Feb 1(2). 10.1136/jitc-2022-006616 Schwartz LH, Litiere S, de Vries E et al (2016) RECIST 1.1-Update and clarification: From the RECIST committee. Eur J Cancer Jul 62:132–137. 10.1016/j.ejca.2016.03.081 Liu Q, Charleston MA, Richards SA, Holland BR (2023) Performance of Akaike Information Criterion and Bayesian Information Criterion in Selecting Partition Models and Mixture Models. Syst Biol May 19(1):92–105. 10.1093/sysbio/syac081 Harrell FE Jr., Lee KL, Mark DB (1996) Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med Feb 28(4):361–387. 10.1002/(SICI)1097-0258(19960229)15:43.0.CO;2-4 Steyerberg EW, Vickers AJ, Cook NR et al (2010) Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiol Jan 21(1):128–138. 10.1097/EDE.0b013e3181c30fb2 Blanche P, Dartigues JF, Jacqmin-Gadda H (2013) Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med Dec 30(30):5381–5397. 10.1002/sim.5958 Wang Y, Li J, Xia Y et al (2013) Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol Mar 20(9):1188–1195. 10.1200/JCO.2012.41.5984 Shakur AH, Huang S, Qian X, Chang X (2021) SURVFIT: Doubly sparse rule learning for survival data. J Biomed Inf May 117:103691. 10.1016/j.jbi.2021.103691 Yang Y, Huang X, Zhou L et al (2019) Clinical use of tumor biomarkers in prediction for prognosis and chemotherapeutic effect in esophageal squamous cell carcinoma. BMC Cancer May 31(1):526. 10.1186/s12885-019-5755-5 Shang H, Chen Y, Wang Q, Yang Y, Zhang J (2024) A Correlation Evaluation Between the Peripheral Blood Index and the Prognosis of Advanced Esophageal Squamous Cell Carcinoma Patients Treated with Camrelizumab. J Inflamm Res 17:2009–2021. 10.2147/jir.S450669 Hamai Y, Emi M, Ibuki Y et al (2023) Ability of Blood Cell Parameters to Predict Clinical Outcomes of Nivolumab Monotherapy in Advanced Esophageal Squamous Cell Carcinoma. Onco Targets Ther 16:263–273. 10.2147/OTT.S404926 Liu J, Hu G, Zhai C et al (2023) Predictive value of nutritional indicators with regard to the survival outcomes in patients with metastatic esophageal squamous cell carcinoma treated with camrelizumab. Oncol Lett May 25(5):198. 10.3892/ol.2023.13784 Doroshow DB, Bhalla S, Beasley MB et al (2021) PD-L1 as a biomarker of response to immune-checkpoint inhibitors. 18(6):345–362 Kato K, Doki Y, Ogata T et al (2023) First-line nivolumab plus ipilimumab or chemotherapy versus chemotherapy alone in advanced esophageal squamous cell carcinoma: a Japanese subgroup analysis of open-label, phase 3 trial (CheckMate 648/ONO-4538-50). Esophagus Apr 20(2):291–301. 10.1007/s10388-022-00970-1 Shen L, Kato K, Kim SB et al (2022) Tislelizumab Versus Chemotherapy as Second-Line Treatment for Advanced or Metastatic Esophageal Squamous Cell Carcinoma (RATIONALE-302): A Randomized Phase III Study. J Clin Oncol Sep 10(26):3065–3076. 10.1200/JCO.21.01926 Kato K, Doki Y, Chau I et al (May 2024) Nivolumab plus chemotherapy or ipilimumab versus chemotherapy in patients with advanced esophageal squamous cell carcinoma (CheckMate 648): 29-month follow-up from a randomized, open-label, phase III trial. Cancer Med 13(9):e7235. 10.1002/cam4.7235 He M, Wang Z, Lu J et al (2024) Final analysis of camrelizumab plus chemotherapy for untreated advanced or metastatic esophageal squamous cell carcinoma: The ESCORT-1st trial. Med Sep 13(9):1137–1149e3. 10.1016/j.medj.2024.05.008 Wang Y, Liu C, Chen H et al (2024) Clinical efficacy and identification of factors confer resistance to afatinib (tyrosine kinase inhibitor) in EGFR-overexpressing esophageal squamous cell carcinoma. Signal Transduct Target Ther Jun 28(1):153. 10.1038/s41392-024-01875-4 Zhang X, Zeng L, Li Y et al (2021) Anlotinib combined with PD-1 blockade for the treatment of lung cancer: a real-world retrospective study in China. Cancer Immunol Immunother Sep 70(9):2517–2528. 10.1007/s00262-021-02869-9 Liu C, Han J, Han D, Huang W, Li B (2023) A new risk score model based on lactate dehydrogenase for predicting prognosis in esophageal squamous cell carcinoma treated with chemoradiotherapy. J Thorac disease Apr 28(4):2116–2128. 10.21037/jtd-23-388 Greten FR, Grivennikov SI (2019) Inflammation and Cancer: Triggers, Mechanisms, and Consequences. Immun Jul 16(1):27–41. 10.1016/j.immuni.2019.06.025 Hanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell Mar 4(5):646–674. 10.1016/j.cell.2011.02.013 Liu J, Gao D, Li J, Hu G, Liu J, Liu D (2022) The Predictive Value of Systemic Inflammatory Factors in Advanced, Metastatic Esophageal Squamous Cell Carcinoma Patients Treated with Camrelizumab. Onco Targets Ther 15:1161–1170. 10.2147/OTT.S382967 Wu X, Han R, Zhong Y, Weng N, Zhang A (2021) Post treatment NLR is a predictor of response to immune checkpoint inhibitor therapy in patients with esophageal squamous cell carcinoma. Cancer cell Int Jul 7(1):356. 10.1186/s12935-021-02072-x Park SH, Lee S, Song JH et al (2020) Prognostic significance of body mass index and prognostic nutritional index in stage II/III gastric cancer. Eur J Surg Oncol Apr 46(4 Pt A):620–625. 10.1016/j.ejso.2019.10.024 Zhang L, Ma W, Qiu Z et al (2023) Prognostic nutritional index as a prognostic biomarker for gastrointestinal cancer patients treated with immune checkpoint inhibitors. Front Immunol 14:1219929. 10.3389/fimmu.2023.1219929 Additional Declarations There is NO Competing Interest. Supplementary Files Supplementarymaterias.pdf Thetrainingcohort.xlsx Dataset 1 Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5322833","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":371731014,"identity":"e18563fe-c644-4350-9fe6-cf2b0e007c14","order_by":0,"name":"Zhihao Lu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIiWNgGAWjYDACCcY2IGnBwMDDwPgYSICAATFaJEBamI2J1MLABtPCJg0Vw69FfnZz24MfFRKJ23nOmFUXyGxLbGBv3ibBUHMHpxbGOQfbDXvOSCTu7O0xuz2D53ZiA8+xMgmGY89wamGWSGyT4G2TSNxwnsfsNg9Ii0SOmQRjw2GcWtiAWiT/QrUUg7XIv8GvhQeoRRpsy9keM2aILTz4tUiAtMickTDe2XOsWBroF+M2nrRii4RjuLXIz0h/JvmmwkZ2O0/yxs+FPbdl+9kPb7zxoQa3FjgAxwVjDwM4mhgSCGuARd8PYpSOglEwCkbBSAMAoI5QOzJUBAUAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-5588-5657","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":true,"prefix":"","firstName":"Zhihao","middleName":"","lastName":"Lu","suffix":""},{"id":371731015,"identity":"64e8d55f-c37e-408d-8164-9a104656b639","order_by":1,"name":"Liyuan Dong","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Liyuan","middleName":"","lastName":"Dong","suffix":""},{"id":371731016,"identity":"fcd03f3d-f6f4-4a47-be12-d40d18d2bc3a","order_by":2,"name":"Yue Ma","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Ma","suffix":""},{"id":371731017,"identity":"ca379fef-5169-45bb-9e10-a76701e4a0c6","order_by":3,"name":"Guang Cao","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Guang","middleName":"","lastName":"Cao","suffix":""},{"id":371731018,"identity":"2e77291d-16c4-4e04-b3ee-a90e318e0f2c","order_by":4,"name":"Dongze Chen","email":"","orcid":"https://orcid.org/0000-0002-0112-5652","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Dongze","middleName":"","lastName":"Chen","suffix":""},{"id":371731019,"identity":"9f2028ca-887c-410e-b03b-6b0a30fcccae","order_by":5,"name":"Fengxiao Dong","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Fengxiao","middleName":"","lastName":"Dong","suffix":""},{"id":371731020,"identity":"30d50840-7a47-4dcf-b197-a7a1bef7fe6e","order_by":6,"name":"Xi Jiao","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute, Department of Gastrointestinal Oncology","correspondingAuthor":false,"prefix":"","firstName":"Xi","middleName":"","lastName":"Jiao","suffix":""},{"id":371731021,"identity":"6b3cf53f-e614-46b5-91cd-0160a6bb854f","order_by":7,"name":"Yanshuo Cao","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Yanshuo","middleName":"","lastName":"Cao","suffix":""},{"id":371731022,"identity":"4bedcbe4-17af-4346-8e45-1a8dd1031537","order_by":8,"name":"Chang Liu","email":"","orcid":"https://orcid.org/0000-0001-9388-4180","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Chang","middleName":"","lastName":"Liu","suffix":""},{"id":371731023,"identity":"9b79a28d-7385-4854-a484-135a8238b6d0","order_by":9,"name":"Yanni Wang","email":"","orcid":"","institution":"Peking University Cancer Hospital and Institute","correspondingAuthor":false,"prefix":"","firstName":"Yanni","middleName":"","lastName":"Wang","suffix":""},{"id":371731024,"identity":"48fa44a2-cbe4-4499-8415-28ac68f99ead","order_by":10,"name":"Na Zhuo","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Na","middleName":"","lastName":"Zhuo","suffix":""},{"id":371731025,"identity":"b5a76245-5080-4b52-aa39-4674ac349ab3","order_by":11,"name":"Fengyuan Wang","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Fengyuan","middleName":"","lastName":"Wang","suffix":""},{"id":371731026,"identity":"fecf9987-5ad6-4757-89ea-3cb41b7d5bb5","order_by":12,"name":"Yixuan Guo","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Yixuan","middleName":"","lastName":"Guo","suffix":""},{"id":371731027,"identity":"cd63f231-7163-4346-bd1b-6dee7a51bacf","order_by":13,"name":"Tingting Dai","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Tingting","middleName":"","lastName":"Dai","suffix":""},{"id":371731028,"identity":"f4e104d2-1f6f-4683-930f-9bf24c82a115","order_by":14,"name":"Shuwei Zhang","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Shuwei","middleName":"","lastName":"Zhang","suffix":""},{"id":371731029,"identity":"d372733b-649f-42b0-a2d6-0d558d2c9fcb","order_by":15,"name":"Hao Jiao","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Hao","middleName":"","lastName":"Jiao","suffix":""},{"id":371731030,"identity":"58db30e9-2c71-49b9-a1b1-0cdaf58c7680","order_by":16,"name":"Xingyue Zou","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Xingyue","middleName":"","lastName":"Zou","suffix":""},{"id":371731031,"identity":"a251542a-8485-47e4-be26-3a5b14276164","order_by":17,"name":"Jian Li","email":"","orcid":"https://orcid.org/0000-0002-9333-3255","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Li","suffix":""},{"id":371731032,"identity":"a0c88773-04e3-4ec0-ba40-582c2649535e","order_by":18,"name":"Lin Shen","email":"","orcid":"https://orcid.org/0000-0003-1134-2922","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Shen","suffix":""},{"id":371731033,"identity":"de22d9a0-299e-4e7d-9f3c-eea6de578103","order_by":19,"name":"Zhonghu He","email":"","orcid":"https://orcid.org/0000-0002-4444-7950","institution":"Beijing Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhonghu","middleName":"","lastName":"He","suffix":""},{"id":371731034,"identity":"3c5b7d6a-a065-40b8-86dd-47e8cca3e025","order_by":20,"name":"Yanqiao Zhang","email":"","orcid":"","institution":"Harbin Medical University Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yanqiao","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2024-10-24 05:30:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5322833/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5322833/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":68353281,"identity":"ed91e5be-e077-4dd6-85a9-aefd1232267e","added_by":"auto","created_at":"2024-11-06 11:03:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":40542,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePatient Enrollment Flowchart\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe enrollment process of patients in the training and validation cohorts set adhered strictly to the inclusion criteria and the exclusion criteria.\u003csup\u003e ‘*’\u003c/sup\u003e: ICIs used in the training and validation cohorts, include camrelizumab, sintilimab, pembrolizumab, toripalimab, atezolizumab and tislelizumab. \u003csup\u003e‘#’\u003c/sup\u003e:\u003csup\u003e \u003c/sup\u003eA portion of the patients came from clinical trials, include NCT03189719, NCT03748134, NCT03430843.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/7ff249c24e481f9fb2c4903b.png"},{"id":68353275,"identity":"e1d5bcf9-2023-4543-a92b-c66492c810bd","added_by":"auto","created_at":"2024-11-06 11:03:02","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":98023,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eClinical survival outcomes in training cohort\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOS in patient subgroups with (A) first line versus second-line or later therapy, (B) mono-ICIs versus ICIs + chemotherapy versus ICIs + targeted therapy, and (C) high PD-L1 expression (CPS ≥ 10) versus low PD-L1 expression (CPS \u0026lt; 10).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/f8992c50b6f0b95e829b5e82.png"},{"id":68353273,"identity":"a53f2eca-802a-4963-8d5d-d453eb7cd4b7","added_by":"auto","created_at":"2024-11-06 11:03:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":75060,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePostoperative prognostic nomogram for patients with advanced ESCC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo use this nomogram, first draw a vertical line from each factor to the points scale to determine the value for that factor. Then sum all the values and drop a vertical line from the total points to the 1-, 2-, and 3-year probability lines to obtain estimated OS.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/5844113b81338aaf986a791f.png"},{"id":68353274,"identity":"3be39492-b7ac-4d49-9735-91cdf0352c53","added_by":"auto","created_at":"2024-11-06 11:03:02","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":96514,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTime-dependent ROC curves for 1-, 2-, and 3-year OS predictions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTime-dependent ROC analysis illustrates the model's robust performance in stratifying patient outcomes at 1-, 2-, 3-year for (A) the training cohort and (B) the validation cohort.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/bf33f37f5423e57861c35ebe.png"},{"id":68353278,"identity":"2611604e-13c4-4c39-ad24-714b63943685","added_by":"auto","created_at":"2024-11-06 11:03:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":94094,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSurvival curves of three risk groups in the training and validation cohorts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe patients in (A) the training cohort and (B) the validation cohort were categorized into low-, moderate- and high-risk groups. The cutoff values were determined based on the tertiles of the risk scores calculated from the nomogram.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/41717b3e7c36559f7aa38acc.png"},{"id":68355307,"identity":"64078798-c618-4507-8281-69d1ca80a12b","added_by":"auto","created_at":"2024-11-06 11:19:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1151686,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/693843ad-9519-463d-8d86-5189ad9c89aa.pdf"},{"id":68353279,"identity":"29abbf9d-a8f8-4505-bce4-43c8870f7708","added_by":"auto","created_at":"2024-11-06 11:03:04","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":594663,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterias.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/43c684d32efa4ded6baadb52.pdf"},{"id":68353277,"identity":"bf0caca8-977e-4dbe-a642-0df8e062db24","added_by":"auto","created_at":"2024-11-06 11:03:03","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":180175,"visible":true,"origin":"","legend":"\u003cp\u003eDataset 1\u003c/p\u003e","description":"","filename":"Thetrainingcohort.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5322833/v1/74522530d5bd315d6960f322.xlsx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"An integrated prognosis prediction model based on real-word clinical characteristics for immunotherapy in advanced esophageal squamous cell carcinoma","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEsophageal cancer ranks as the seventh most common and sixth leading cause of cancer death worldwide\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. China accounts for over half of the global incidence and mortality rates of esophageal cancer, with 90% of cases being esophageal squamous cell carcinoma (ESCC)\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, and the five-year survival rate remaining below 5%\u003csup\u003e3\u003c/sup\u003e. Although the advent of immune checkpoint inhibitors (ICIs), such as programmed cell death protein-1(PD-1)/programmed cell death ligand-1 (PD-L1) inhibitors and cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) inhibitors has revolutionized the treatment landscape\u003csup\u003e\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003eand emerged as the standard option for advanced ESCC\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, only a subset of patients benefits from ICIs. Even when combined with chemotherapy, the 5-year survival rate remains at a mere 10\u0026ndash;15%\u003csup\u003e8\u003c/sup\u003e, emphasizing the urgent need for effective predictors to identify populations that could potentially benefit from ICIs.\u003c/p\u003e \u003cp\u003eTraditional biomarkers including tumor mutational load (TMB)\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e and PD-L1 expression\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e have been investigated. However, their predictive value in ESCC remains unsatisfactory. Particularly, the role of PD-L1 expression in the efficacy of ICIs is fraught with uncertainty and contradictions in different clinical trials\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Recently, several multi-gene signatures, such as fibroblast-associated signature\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e, immunogenic cell death (ICD)-associated gene panel\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e and angiogenesis-associated risk score\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e have been developed as predictive models and have demonstrated some prognostic utility. Nevertheless, these predictors either rely on tumor tissue and complex molecular analyses, or have not been trained and validated in large-scale and real-world datasets, leading to considerable limitations in effectiveness, clinical accessibility and reliability. Therefore, there is an urgent need to construct an integrated model that incorporates the most direct, accessible and non-invasive clinical biomarkers, enhancing efficiency and cost-effectiveness while enabling rapid and precise prognostic predictions for patients treated with ICIs.\u003c/p\u003e \u003cp\u003eIn this study, we aim to construct and validate a multivariable model for individualized prediction of prognosis and survival in ESCC patients treated with ICIs, based on large-scale and real-world clinical data with long-term follow-up. Furthermore, we expect to provide insights that will enable clinicians and patients to decide upon the optimal ICIs treatment for advanced ESCC.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Population and Follow-Up\u003c/h2\u003e \u003cp\u003eData for the training cohort were obtained from patients with advanced ESCC treated at the Department of Gastrointestinal Oncology, Peking University Cancer Hospital, between January 14, 2016, and January 26, 2024. The external validation cohort comprised patients from Harbin Medical University Cancer Hospital, between January 10, 2019, and July 6, 2022. Inclusion criteria were: (1) pathologically confirmed ESCC; (2) unresectable advanced, recurrent (including postoperative recurrence), or metastatic ESCC without prior systemic therapy; (3) treated with mono-ICIs, ICIs combined with chemotherapy or targeted therapy; (4) at least one measurable lesion as assessed by response evaluation criteria in solid tumours1.1 (RECIST 1.1)\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e; and (5) availability of complete clinical and follow-up data. Exclusion criteria included: (1) patients lost to follow-up after only one cycle of ICIs; (2) patients who did not undergo regular imaging evaluations (at least once every two cycles); and (3) patients with severe organ dysfunction or significant abnormalities in blood counts, liver or kidney function, or autoimmune diseases.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEfficacy Assessment\u003c/h3\u003e\n\u003cp\u003ePatients were evaluated for efficacy by imaging scans after every two cycles of treatment. Clinical responses to treatment were determined according to RECIST 1.1, which includes complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Follow-up was conducted to ascertain patients' survival status (alive or deceased). The primary endpoint of this study was overall survival (OS), defined as the time from the first dose of ICIs to death from any cause.\u003c/p\u003e\n\u003ch3\u003eVariable Selection and Model Construction\u003c/h3\u003e\n\u003cp\u003ePotential variables were categorized into 4 groups: patient characteristics (age, sex, smoking, etc.), tumor characteristics (staging, metastasis of lymph node, liver, lung and bone), treatment-related data (line of therapy, treatment options, etc.) and baseline laboratory test results obtained within 7 days prior to ICIs [e.g., lactate dehydrogenase (LDH), hemoglobin (Hb), lymphocytes count (ALC)]. A total of 26 variables were initially identified as potential predictors for subsequent analysis. Variables with more than 10% of values missing were not included for analysis. The missing values for carcinoembryonic antigen (CEA) and carbohydrate antigen 199 (CA199) in the validation cohort were 4.3% and 5.4%, respectively, with missing data imputed using multiple interpolation methods. All other variables had complete data with no missing values. Optimal cutoff values were determined based on clinical reference ranges or were calculated using the \u0026ldquo;surv_cutpoint\u0026rdquo; algorithm from the \u0026ldquo;survminer\u0026rdquo; R package (Supplementary Table\u0026nbsp;1). Univariate Cox regression analyses assessed the independent prognostic value of these 26 variables. Variables finally included in multivariable Cox regression model were selected based on clinical significance and minimization of Akaike information criterion (AIC), which considers both the statistical goodness of fit and the number of parameters, imposing a penalty for excessive parameters\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. A nomogram was subsequently developed to visualize individual OS probabilities at 1-, 2-, and 3-year.\u003c/p\u003e\n\u003ch3\u003eAssessment and Validation of the Model\u003c/h3\u003e\n\u003cp\u003eThe discrimination of the prognostic model was assessed by calculating the Harrell\u0026rsquo;s concordance index (C-index)\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Calibration curves were plotted to compare model-predicted survival estimates with actual Kaplan-Meier survival probabilities\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Time-dependent ROC curves were plotted using the \u0026ldquo;timeROC\u0026rdquo; R package to evaluate predictive accuracy at specific time points (1-, 2-, and 3-year)\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. The area under the curve (AUC) and its 95% confidence intervals were computed.\u003c/p\u003e \u003cp\u003eBootstrapping with 1000 resample was employed for internal validation \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, and the accuracy and generalizability of the model were examined in the validation cohort.\u003c/p\u003e\n\u003ch3\u003eRisk stratification\u003c/h3\u003e\n\u003cp\u003eFor each patient, the risk score was calculated based on the nomogram using the \u0026ldquo;predict\u0026rdquo; algorithm. Tertiles of risk score served as cutoff values for risk stratification, classifying patients into low-, moderate-, and high-risk groups. Kaplan-Meier survival curves for these three risk groups were plotted and compared using the log-rank test\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData processing and statistical analysis\u003c/h2\u003e \u003cp\u003eAll data processing, statistical analysis and figure plotting were conducted with SPSS 26.0 and R 4.3.0. Two-sided P-values\u0026thinsp;\u0026lt;\u0026thinsp;.05 were considered statistically significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\n \u003ch2\u003ePatient\u0026rsquo;s characteristics\u003c/h2\u003e\n \u003cp\u003eWe screened a total of 513 patients with advanced ESCC who received ICIs from Peking University Cancer Hospital between January 14, 2016, and January 26, 2024, along with 213 patients from Harbin Cancer Hospital between January 10, 2019, and July 6, 2022. Following the established inclusion and exclusion criteria, 430 patients were ultimately enrolled in the training cohort, while 184 patients constituted the validation cohort (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). The median follow-up time was 29.7 months (95% CI: 28.1\u0026ndash;31.3 months) in the training cohort, during which 282 (65.5%) deaths occurred. In the validation cohort, the median follow-up was 43.2 months (95% CI: 39.9\u0026ndash;46.6 months), with 152 (82.6%) deaths recorded\u003c/p\u003e\n \u003cp\u003eThe background characteristics are detailed in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. The median age of patients in both cohorts was 61.0 years, with a higher proportion of females in the training cohort. No significant differences were observed in disease stage or metastasis (to the liver, lung, and bone) between the two cohorts. In the training cohort, 140 (32.6%) patients received second-line or later treatment, significantly more than 34 (18.5%) patients in the validation cohort. Additionally, in the training cohort, 95 (22.1%) received mono-ICIs, while 273 (63.5%) patients received ICIs combined with chemotherapy and 62 (14.4%) patients received ICIs combined with targeted therapy. In contrast, most patients (91.8%) in the validation cohort were treated with ICIs combined with chemotherapy (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSelected Patient Characteristics in the Training and Validation Cohorts\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eCharacteristics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003ePatients, No. (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eTraining Cohort (n\u0026thinsp;=\u0026thinsp;430)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eValidation Cohort (n\u0026thinsp;=\u0026thinsp;200)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eMedian (IQR)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e61 (56.0\u0026ndash;66.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e61 (56.0-66.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.812\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eMale\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e383 (89.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e180 (97.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eFemale\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e47 (10.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (2.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSmoke\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e106 (24.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e66 (35.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e324 (75.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e118 (64.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAlcohol\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e148 (34.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e61 (33.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e282 (65.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e123 (66.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.762\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eStage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eIII\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e24 (5.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 (5.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eIV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e406 (94.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e174 (94.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.942\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eLiver metastasis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e344 (80.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e156 (84.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e86 (20.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e28 (15.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.163\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eLung metastasis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e328 (76.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e147 (79.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e102 (23.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e37 (20.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.327\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBone metastasis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e382 (88.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e171 (92.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e48 (11.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (7.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.120\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eLine of therapy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eFirst-line\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e290 (67.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e150 (81.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ge;Second-line\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e140 (32.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e34 (18.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eTreatment\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eMono-ICIs/\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eICIs\u0026thinsp;+\u0026thinsp;targeted therapy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e157 (36.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (8.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eICIs\u0026thinsp;+\u0026thinsp;chemotherapy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e273 (63.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e169 (91.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCEA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e330 (76.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e146 (79.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e100 (23.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e38 (20.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.479\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCA199\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e389 (90.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e167 (90.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e41 (9.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e17 (9.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.909\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eLDH\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;240\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e341 (79.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e163 (88.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;240\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e89 (20.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21 (11.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eALC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;1.7\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e309 (71.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e110 (59.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;1.7\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e121 (28.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e74 (40.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ePNI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;49.50\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e196 (45.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e115 (62.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;49.50\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e234 (54.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e69 (37.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSII\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026le;\u0026thinsp;589.41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e149 (34.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e73 (39.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026gt;\u0026thinsp;589.41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e281 (65.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e111 (60.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.235\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eThe percentages of patients reflect the proportion of patients within each subgroup of specified characteristics. Categorical variables were analyzed using Chi-square test, while continuous variables were assessed using t-test.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eSurvival outcomes.\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eIn the training cohort, the median OS for first-line treatment was 21.3 months (95% CI: 17.0-27.1 months), with 2-, 3-, and 5-year OS rates of 47.3%, 29.8%, and 20.2%, respectively. In contrast, the median OS for second-line or later treatment was 10.2 months (95% CI: 8.6\u0026ndash;13.4 months), with corresponding OS rates of 16.6%, 8.2%, and 5.4%. (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eA). In the validation cohort, the median OS for first-line treatment was 15.6 months (95% CI: 13.1\u0026ndash;18.8 months), with 2-, 3-, and 5-year OS rates of 36.7%, 25.1%, and 12%. the median OS for second-line or later treatment was 8.5 months (95% CI: 6.3\u0026ndash;12.5 months), with OS rates of 17.6%, 11.0%, and 0.0%. (Supplementary Fig.\u0026nbsp;2A).\u003c/p\u003e\n \u003cp\u003eAmong different treatment regions in the training cohort, the median OS was significantly higher in the ICIs combined with chemotherapy group (24.6 months, 95% CI: 17.6\u0026ndash;28.5) compared to the ICIs combined with targeted therapy (10.3 months, 95% CI 8.4\u0026ndash;13.2, \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/em\u003e) and mono-ICIs groups (10.2 months, 95% CI: 7.9\u0026ndash;14.0, \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/em\u003e). However, no noticeable difference in OS was observed between the ICIs combined with targeted therapy group and the mono-ICIs group (HR\u0026thinsp;=\u0026thinsp;1.07, 95% CI: 0.76\u0026ndash;1.50, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.701\u003c/em\u003e, Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eB). A similar trend was noted in the validation cohort, although the treatment regimen was overwhelmingly ICIs combined with chemotherapy (Supplementary Fig.\u0026nbsp;2B).\u003c/p\u003e\n \u003cp\u003eAdditionally, we stratified patients into high PD-L1 expression [combined positive score (CPS)\u0026thinsp;\u0026ge;\u0026thinsp;10] and low PD-L1 expression (CPS\u0026thinsp;\u0026lt;\u0026thinsp;10) groups based on a CPS cutoff value of 10 to evaluate the potential of PD-L1 as a prognostic biomarker. Surprisingly, PD-L1 failed to differentiate OS between the two groups (HR\u0026thinsp;=\u0026thinsp;0.94, 95% CI: 0.74\u0026ndash;1.19, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;.6\u003c/em\u003e, Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eC), indicating that PD-L1 expression was not a reliable prognostic predictor in the training cohort.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eConstruction and Validation of the Prognosis Prediction Model\u003c/h2\u003e\n \u003cp\u003ePrevious studies have demonstrated that factors influencing survival and prognosis in ESCC extend beyond tumor characteristics (stage, metastasis, etc.) and treatment regimen. Tumor markers such as CEA and CA199\u003csup\u003e23\u003c/sup\u003e, along with laboratory markers such as inflammatory markers [LDH\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, systemic immune-inflammation index (SII)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, etc.], and nutritional markers [albumin, prognostic nutritional index (PNI)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, etc.], also play significant roles. However, these markers have not been widely discussed or utilized in predictive models for clinical immunotherapy.\u003c/p\u003e\n \u003cp\u003eIn our study, covering the above indicators, a total of 26 candidate predictive variables were entered into a multivariable Cox regression model (Supplementary Table 1). After stepwise selection based on the AIC, stage, bone metastasis, line of therapy, treatment, LDH, CEA, CA199, ALC, PNI, and SII were retained in the final prediction model. The model confirmed that stage III, first-line ICIs combined with chemotherapy and elevated levels of PNI (\u0026gt;\u0026thinsp;49.50) (HR\u0026thinsp;=\u0026thinsp;0.73, 95%CI: 0.56\u0026ndash;0.95, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.019\u003c/em\u003e) were associated with superior survival outcomes. Furthermore, elevated levels of CEA (\u0026gt;\u0026thinsp;5) (HR\u0026thinsp;=\u0026thinsp;1.35, 95%CI: 1.03\u0026ndash;1.79, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.030\u003c/em\u003e), CA199 (\u0026gt;\u0026thinsp;37) (HR\u0026thinsp;=\u0026thinsp;1.75, 95%CI: 1.17\u0026ndash;2.62, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.006\u003c/em\u003e), LDH (\u0026gt;\u0026thinsp;240) (HR 1.45, 95%CI: 1.09\u0026ndash;1.91, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.010\u003c/em\u003e) and SII (\u0026gt;\u0026thinsp;589.41) (HR\u0026thinsp;=\u0026thinsp;1.63, 95%CI: 1.23\u0026ndash;2.16, \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/em\u003e) were associated with a higher risk of death and poorer prognosis. Bone metastasis (HR\u0026thinsp;=\u0026thinsp;1.34, 95%CI 0.94\u0026ndash;1.91, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.108\u003c/em\u003e) and ALC (\u0026le;\u0026thinsp;1.7) (HR\u0026thinsp;=\u0026thinsp;0.73, 95%CI: 0.51\u0026ndash;1.04, \u003cem\u003eP\u0026thinsp;=\u0026thinsp;0.078\u003c/em\u003e) was also linked to inferior survival, although not reaching statistical significance (Supplementary Table 2). A risk prediction nomogram was developed based on this model (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). According to the points values to each variable, stage had the most significant impact on prognosis, despite only 24 (5.5%) patients being classified as stage III.\u003c/p\u003e\n \u003cp\u003eThe calibration plot for both the training and validation cohorts demonstrated optimal concordance between nomogram predictions and actual observation for 2-, and 3-year OS, remaining within an acceptable range for survival probability prediction. However, the 1-year OS prediction exhibited moderate discrimination, potentially due to the inherent characteristics of the training cohort (Supplementary Fig. 3). Further exploration with larger cohorts is necessary to enhance precision. The model in the training cohort achieved a C-index of 0.725 (95%CI 0.694\u0026ndash;0.756) and the bootstrap-corrected C-index was 0.722 (95% CI 0.688\u0026ndash;0.751), indicating a robust level of discriminative ability for survival prediction. The time-dependent ROC curve analysis further underscored the model\u0026apos;s strong predictive performance for 1-, 2-, and 3-year OS, with AUC values reflecting excellent discrimination, thereby confirming the reliability of the model in prognostic predictions across these time points (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eA).\u003c/p\u003e\n \u003cp\u003eTo test the generalizability of the prediction model, we evaluated its performance in the external validation cohort, which yielded a C-index of 0.691 (95%CI: 0.650\u0026ndash;0.733). The AUC values were consistent with those observed in the training cohort, further validating the predictive value of the model in real-world settings (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eB). These internal and external validation results highlight the robustness and reliability of this model, demonstrating consistent performance across various follow-up time intervals.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003eRisk Stratification of Mortality for ESCC Patients\u003c/h2\u003e\n \u003cp\u003eTo facilitate the practical application of this model in clinical settings, we calculated the risk score for each patient in both the training and validation cohorts based on the total points from the nomogram. Patients were subsequently stratified into low-, moderate-, and high-risk groups according to the tertiles of risk scores from the training cohort. In the training cohort, the high-risk group exhibited the lowest 1-, 2- and 3-year survival probabilities at 37.5%, 9.8% and 1.9%, compared to 69.2%, 38.1% and 21.4% for the moderate-risk group and 83.6%, 65.0% and 45.8% for the low-risk group. In the validation cohort, the high-risk group showed 29.2%, 8.3% and 0% survival probabilities, while the moderate-risk group had probabilities of 55.7%, 22.8% and 13.9% and the low-risk group for 77.8%, 50.6% and 36.7% (Supplementary Table 3). Kaplan-Meier curves for the 3 risk groups in both the training and validation cohorts demonstrated significant differentiation (both log-rank \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/em\u003e). (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we reviewed a large cohort of 430 patients with advanced ESCC treated with ICIs and described their survival outcomes. Subsequently, we constructed and validated an integrated model incorporating 10 variables to predict OS based on comprehensive clinical characteristics. This model demonstrated strong discrimination ability (C-index\u0026thinsp;=\u0026thinsp;0.725), providing real-world evidence that patients treated with ICIs in combination with chemotherapy, along with high PNI (\u0026gt;\u0026thinsp;49.5), low LDH (\u0026le;\u0026thinsp;240), CEA (\u0026le;\u0026thinsp;5), CA199 (\u0026le;\u0026thinsp;37) and SII levels (\u0026le;\u0026thinsp;589.41), were more likely to benefit from the treatment. Finally, we integrated this model into an easy-to-use online prediction tool (Supplementary Fig.\u0026nbsp;5, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://escc-survival.shinyapps.io/shiny_app/\u003c/span\u003e\u003cspan address=\"https://escc-survival.shinyapps.io/shiny_app/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) featuring an interactive interface for parameter entry and graphical output.\u003c/p\u003e \u003cp\u003ePD-L1 expression has been investigated as a promising predictive marker for response and efficacy of ICIs in advanced ESCC\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. However, current evidence regarding its predictive value is fraught with conflicts and ambiguities. For instance, CheckMate-648\u003csup\u003e28\u003c/sup\u003e indicated that ICIs extend OS only in patients with high PD-L1 expression [tumor cell proportion score (TPS)\u0026thinsp;\u0026gt;\u0026thinsp;1%], while RATIONALE-302\u003csup\u003e29\u003c/sup\u003e, ORIENT-15\u003csup\u003e7\u003c/sup\u003e and two meta-analyses\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e suggested that those with low PD-L1 expression (TPS\u0026thinsp;\u0026lt;\u0026thinsp;1% or CPS\u0026thinsp;\u0026lt;\u0026thinsp;10) also experienced OS benefits. In this study, we evaluated various ICIs including sintilimab, pembrolizumab, nivolumab, and camrelizumab. Generally, PD-L1 failed to demonstrate any predictive value (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC, Supplementary Fig.\u0026nbsp;4), indicating that PD-L1 may not accurately and reliably predict prognosis for ICIs treatment, thus necessitating further investigation into its clinical utility as a biomarker.\u003c/p\u003e \u003cp\u003eConversely, our prediction model exhibited superior survival and prognostic capabilities for both mono-ICIs and in combination with chemotherapy, significantly outperforming PD-L1. This model integrates comprehensive array of factors across four categories, encompassing fundamental tumor characteristics, treatment strategies, and biological predictors such as LDH, CEA, CA199, ALC, PNI and SII. To our knowledge, this is the first model designed for individualized survival prediction in advanced ESCC patients treated with ICIs, based on large-scale, high-quality real-world data. These clinical predictors stemmed from essential tumor evaluations and laboratory tests, ensuring convenience, cost-effectiveness, and clinical accessibility. This study guarantees the representativeness of the patient population and facilitates a thorough exploration of the interactions between clinical characteristics and survival outcomes for immunotherapy, ultimately informing personalized therapeutic decision-making.\u003c/p\u003e \u003cp\u003eTumor characteristics, such as stage and metastasis, are critical in predicting prognosis for ESCC. In our training cohort, patients with stage IV and concomitant bone metastases exhibited a poorer prognosis. Additionally, line of therapy also greatly influenced patient outcomes. The median OS for first-line treatment in the training cohort reached 21 months, with 2-, 3-, and 5-year survival probabilities of 47.3%, 29.8%, and 20.2%, surpassing the average level. In contrast, the validation cohort exhibited slightly lower median OS, reflecting regional differences in clinical practice, therapeutic concepts and treatment experience between the two study centers. Regarding treatment regimens, ICIs combined with chemotherapy\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e or targeted inhibitors such as epidermal growth factor receptor (EGFR) inhibitors and tyrosine kinase inhibitors (TKIs)\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e, are widely recognized options. However, few studies have compared these therapies simultaneously. Our findings demonstrated that survival and prognosis were best in patients treated with ICIs combined with chemotherapy, while no considerable difference was observed between ICIs combined with targeted therapy and mono-ICIs, underscoring the need for further clinical trials to validate these targets.\u003c/p\u003e \u003cp\u003eThe role of biological predictors, including serum tumor markers, biochemical markers, inflammatory markers and nutritional markers, cannot be underestimated in assessing prognosis and survival, especially during ICI therapy. Limited research has focused on the association between tumor markers and therapeutic effects in ESCC. A small sample study suggested that CEA and CA199 might predict the efficacy of postoperative chemotherapy\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Our study further indicated that CEA and CA199 can serve as prognostic markers for ICIs therapy in advanced ESCC. Elevated circulating levels of LDH, a key enzyme in anaerobic glycolysis, correlate with adverse prognosis\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e, which was consistent with our findings. Additionally, circulating neutrophils, lymphocytes, and platelets can reflect the body\u0026rsquo;s immune inflammatory state\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e, linking to tumorigenesis, progression, and metastasis\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Considering discussions regarding inflammatory biomarkers such as PLR, NLR and SII in ESCC remain inconsistent due to sample size constraints\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e, our analysis enrolled ALC, PLR, NLR (dNLR), lymphocyte-to-white blood cell ratio (LWR) and SII in multivariable Cox regression. Ultimately, only ALC and SII were retained in the final model, with high SII (\u0026gt;\u0026thinsp;589.61) identified as a poor prognostic indicator for OS. PNI, which combines serum albumin levels and peripheral blood lymphocyte counts, reflects both immune and nutritional status\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Consistent with prior study\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, we found that a high PNI (\u0026gt;\u0026thinsp;49.5) was associated with improved survival outcomes, highlighting the importance of timely nutritional and immunological interventions.\u003c/p\u003e \u003cp\u003eThe prediction model integrates above variables demonstrates strong performance. There was no deterioration in calibration during internal validation (C-index\u0026thinsp;=\u0026thinsp;0.722), indicating that the model was likely not overfitted. More importantly, the model also performed well (C-index\u0026thinsp;=\u0026thinsp;0.691) in the external validation cohort from a geographically distant area in China, despite some expected loss of discriminative performance. Using this prediction model, patients could be stratified into low-, moderate-, and high-risk groups based on their risk scores. OS demonstrated clear differentiation among the three groups in both the training and validation cohorts (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and Supplementary Table\u0026nbsp;3), highlighting the model's clinical significance, especially in the absence of standardized criteria on patient selection for ICIs. To facilitate the implementation of this model in clinical practice, we integrated our model into an interactive online risk calculator (Supplementary Fig.\u0026nbsp;5, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://escc-survival.shinyapps.io/shiny_app/\u003c/span\u003e\u003cspan address=\"https://escc-survival.shinyapps.io/shiny_app/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which provides automatic prediction of patient outcomes. This tool could serve as a valuable adjunct for treatment decision-making, particularly in institutions where ICIs are introduced as a novel therapeutic option and during discussions with patients. It can also aid in patients\u0026rsquo; selection for clinical trials based on predicted outcomes, allowing for stratified randomization into low-, moderate-, and high-risk groups. Enriching trial participation with patients most likely to respond can lead to smaller sample sizes, thereby reducing costs and time, while minimizing the risk of exposing patients to ineffective treatments.\u003c/p\u003e \u003cp\u003eSeveral limitations of the present study warrant acknowledgment. Firstly, the single center design, small sample size, and retrospective nature of the analysis introduce potential selection bias. Therefore, a larger, multi-center cohort is necessary to estimate and validate this prediction model with adequate statistical power. Additionally, while our findings align with previous studies, the lack of standardized cutoff values complicates the generalization of our conclusions to other studies. Consequently, establishing a unified testing method to confirm appropriate cutoff values is essential.\u003c/p\u003e \u003cp\u003eIn summary, based on real-world clinical data, we developed the first integrated prognosis prediction model for immunotherapy and integrated it into an interactive online tool. This model may provide individualized survival predictions for advanced ESCC patients treated with ICIs, enabling clinicians and patients to estimate the survival outcomes and make decisions more accurately and efficiently in the future.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by Beijing Natural Science Foundation (Z220022) and Beijing Hospitals Authority Clinical Medicine Development of special funding support (ZLRK202326). The authors thank all research team members and participants of each cohort study for their contribution to this research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study was approved by the Medical Ethics Committee of Peking University Cancer Hospital and Harbin Medical University Cancer Hospital, and informed consent was obtained from all patients. Participants gave informed consent to participate in the study before taking part.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by Beijing Natural Science Foundation (Z220022) and Beijing Hospitals Authority Clinical Medicine Development of special funding support (ZLRK202326).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCRediT Authorship Contribution Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZhihao Lu: Corresponding author, Conceptualization, Funding acquisition, Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eZhonghu He and Yanqiao Zhang: Corresponding author, Conceptualization, Data curation, Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eLiyuan Dong: Investigation, Conceptualization, Data curation, Methodology, Software, Formal analysis, Validation, Visualization, Writing-original draft, Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eYue Ma: Resources, Data curation, Methodology, Validation, Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eGuang Cao: Data curation, Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eDongze Chen: Methodology,\u0026nbsp;Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eXi Jiao, Fengxiao Dong, Yanshuo Cao, Chang Liu, Yanni Wang, Na Zhuo, Fengyuan Wang, Yixuan Guo, Tingting Dai, Shuwei Zhang, Hao Jiao, Xingyue Zou, Jian Li and Lin Shen:Data curation,\u0026nbsp;Writing-review and editing, Final approval of the manuscript.\u003c/p\u003e\n\u003cp\u003eNo potential conflict of interest was reported by the authors\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin May 71(3):209\u0026ndash;249. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3322/caac.21660\u003c/span\u003e\u003cspan address=\"10.3322/caac.21660\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiang H, Fan JH, Qiao YL (2017) Epidemiology, etiology, and prevention of esophageal squamous cell carcinoma in China. Cancer Biol Med Feb 14(1):33\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.20892/j.issn.2095-3941.2016.0093\u003c/span\u003e\u003cspan address=\"10.20892/j.issn.2095-3941.2016.0093\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng H, Chen W, Zheng R et al (2018) Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health May 6(5):e555\u0026ndash;e567. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S2214-109X(18)30127-X\u003c/span\u003e\u003cspan address=\"10.1016/S2214-109X(18)30127-X\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKojima T, Shah MA, Muro K et al (2020) Randomized Phase III KEYNOTE-181 Study of Pembrolizumab Versus Chemotherapy in Advanced Esophageal Cancer. J Clin Oncol Dec 10(35):4138\u0026ndash;4148. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1200/JCO.20.01888\u003c/span\u003e\u003cspan address=\"10.1200/JCO.20.01888\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoki Y, Ajani JA, Kato K et al (2022) Nivolumab Combination Therapy in Advanced Esophageal Squamous-Cell Carcinoma. N Engl J Med Feb 3(5):449\u0026ndash;462. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa2111380\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa2111380\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo H, Lu J, Bai Y et al (2021) Effect of Camrelizumab vs Placebo Added to Chemotherapy on Survival and Progression-Free Survival in Patients With Advanced or Metastatic Esophageal Squamous Cell Carcinoma: The ESCORT-1st Randomized Clinical Trial. JAMA Sep 14(10):916\u0026ndash;925. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.2021.12836\u003c/span\u003e\u003cspan address=\"10.1001/jama.2021.12836\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu Z, Wang J, Shu Y et al (2022) Sintilimab versus placebo in combination with chemotherapy as first line treatment for locally advanced or metastatic oesophageal squamous cell carcinoma (ORIENT-15): multicentre, randomised, double blind, phase 3 trial. Bmj Apr 19 377:e068714. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmj-2021-068714\u003c/span\u003e\u003cspan address=\"10.1136/bmj-2021-068714\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShah MA, Sun J-M, Shen L et al (2024) First-line pembrolizumab (pembro) plus chemotherapy (chemo) for advanced esophageal cancer: 5-year outcomes from the phase 3 KEYNOTE-590 study. American Society of Clinical Oncology\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu Z, Chen H, Li S et al (2020) Tumor copy-number alterations predict response to immune-checkpoint-blockade in gastrointestinal cancer. J Immunother Cancer Aug 8(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jitc-2019-000374\u003c/span\u003e\u003cspan address=\"10.1136/jitc-2019-000374\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKojima T, Muro K, Francois E et al (2019) Pembrolizumab versus chemotherapy as second-line therapy for advanced esophageal cancer: Phase III KEYNOTE-181 study. American Society of Clinical Oncology\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYap DWT, Leone AG, Wong NZH et al (2023) Effectiveness of Immune Checkpoint Inhibitors in Patients With Advanced Esophageal Squamous Cell Carcinoma: A Meta-analysis Including Low PD-L1 Subgroups. JAMA Oncol Feb 1(2):215\u0026ndash;224. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamaoncol.2022.5816\u003c/span\u003e\u003cspan address=\"10.1001/jamaoncol.2022.5816\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu HX, Pan YQ, He Y et al (2023) Clinical Benefit of First-Line Programmed Death-1 Antibody Plus Chemotherapy in Low Programmed Cell Death Ligand 1-Expressing Esophageal Squamous Cell Carcinoma: A Post Hoc Analysis of JUPITER-06 and Meta-Analysis. J Clin Oncol Mar 20(9):1735\u0026ndash;1746. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1200/JCO.22.01490\u003c/span\u003e\u003cspan address=\"10.1200/JCO.22.01490\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRen Q, Zhang P, Zhang X et al (2023) A fibroblast-associated signature predicts prognosis and immunotherapy in esophageal squamous cell cancer. Front Immunol 14:1199040. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fimmu.2023.1199040\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2023.1199040\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang H, Luo YB, Wu W et al (2021) The molecular feature of macrophages in tumor immune microenvironment of glioma patients. Comput Struct Biotechnol J 19:4603\u0026ndash;4618. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.csbj.2021.08.019\u003c/span\u003e\u003cspan address=\"10.1016/j.csbj.2021.08.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao M, Wu X, Jiao X et al (2024) Prognostic and predictive value of angiogenesis-associated serum proteins for immunotherapy in esophageal cancer. J Immunother Cancer Feb 1(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jitc-2022-006616\u003c/span\u003e\u003cspan address=\"10.1136/jitc-2022-006616\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwartz LH, Litiere S, de Vries E et al (2016) RECIST 1.1-Update and clarification: From the RECIST committee. Eur J Cancer Jul 62:132\u0026ndash;137. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ejca.2016.03.081\u003c/span\u003e\u003cspan address=\"10.1016/j.ejca.2016.03.081\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Q, Charleston MA, Richards SA, Holland BR (2023) Performance of Akaike Information Criterion and Bayesian Information Criterion in Selecting Partition Models and Mixture Models. Syst Biol May 19(1):92\u0026ndash;105. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/sysbio/syac081\u003c/span\u003e\u003cspan address=\"10.1093/sysbio/syac081\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarrell FE Jr., Lee KL, Mark DB (1996) Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med Feb 28(4):361\u0026ndash;387. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/(SICI)1097-0258(19960229)15:4\u0026lt;361::AID-SIM168\u0026gt;3.0.CO;2-4\u003c/span\u003e\u003cspan address=\"10.1002/(SICI)1097-0258(19960229)15:4%3C361::AID-SIM168%3E3.0.CO;2-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteyerberg EW, Vickers AJ, Cook NR et al (2010) Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiol Jan 21(1):128\u0026ndash;138. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/EDE.0b013e3181c30fb2\u003c/span\u003e\u003cspan address=\"10.1097/EDE.0b013e3181c30fb2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlanche P, Dartigues JF, Jacqmin-Gadda H (2013) Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med Dec 30(30):5381\u0026ndash;5397. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/sim.5958\u003c/span\u003e\u003cspan address=\"10.1002/sim.5958\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Y, Li J, Xia Y et al (2013) Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol Mar 20(9):1188\u0026ndash;1195. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1200/JCO.2012.41.5984\u003c/span\u003e\u003cspan address=\"10.1200/JCO.2012.41.5984\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShakur AH, Huang S, Qian X, Chang X (2021) SURVFIT: Doubly sparse rule learning for survival data. J Biomed Inf May 117:103691. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jbi.2021.103691\u003c/span\u003e\u003cspan address=\"10.1016/j.jbi.2021.103691\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Y, Huang X, Zhou L et al (2019) Clinical use of tumor biomarkers in prediction for prognosis and chemotherapeutic effect in esophageal squamous cell carcinoma. BMC Cancer May 31(1):526. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12885-019-5755-5\u003c/span\u003e\u003cspan address=\"10.1186/s12885-019-5755-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShang H, Chen Y, Wang Q, Yang Y, Zhang J (2024) A Correlation Evaluation Between the Peripheral Blood Index and the Prognosis of Advanced Esophageal Squamous Cell Carcinoma Patients Treated with Camrelizumab. J Inflamm Res 17:2009\u0026ndash;2021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/jir.S450669\u003c/span\u003e\u003cspan address=\"10.2147/jir.S450669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHamai Y, Emi M, Ibuki Y et al (2023) Ability of Blood Cell Parameters to Predict Clinical Outcomes of Nivolumab Monotherapy in Advanced Esophageal Squamous Cell Carcinoma. Onco Targets Ther 16:263\u0026ndash;273. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/OTT.S404926\u003c/span\u003e\u003cspan address=\"10.2147/OTT.S404926\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu J, Hu G, Zhai C et al (2023) Predictive value of nutritional indicators with regard to the survival outcomes in patients with metastatic esophageal squamous cell carcinoma treated with camrelizumab. Oncol Lett May 25(5):198. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3892/ol.2023.13784\u003c/span\u003e\u003cspan address=\"10.3892/ol.2023.13784\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoroshow DB, Bhalla S, Beasley MB et al (2021) PD-L1 as a biomarker of response to immune-checkpoint inhibitors. 18(6):345\u0026ndash;362\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKato K, Doki Y, Ogata T et al (2023) First-line nivolumab plus ipilimumab or chemotherapy versus chemotherapy alone in advanced esophageal squamous cell carcinoma: a Japanese subgroup analysis of open-label, phase 3 trial (CheckMate 648/ONO-4538-50). Esophagus Apr 20(2):291\u0026ndash;301. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10388-022-00970-1\u003c/span\u003e\u003cspan address=\"10.1007/s10388-022-00970-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShen L, Kato K, Kim SB et al (2022) Tislelizumab Versus Chemotherapy as Second-Line Treatment for Advanced or Metastatic Esophageal Squamous Cell Carcinoma (RATIONALE-302): A Randomized Phase III Study. J Clin Oncol Sep 10(26):3065\u0026ndash;3076. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1200/JCO.21.01926\u003c/span\u003e\u003cspan address=\"10.1200/JCO.21.01926\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKato K, Doki Y, Chau I et al (May 2024) Nivolumab plus chemotherapy or ipilimumab versus chemotherapy in patients with advanced esophageal squamous cell carcinoma (CheckMate 648): 29-month follow-up from a randomized, open-label, phase III trial. Cancer Med 13(9):e7235. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/cam4.7235\u003c/span\u003e\u003cspan address=\"10.1002/cam4.7235\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe M, Wang Z, Lu J et al (2024) Final analysis of camrelizumab plus chemotherapy for untreated advanced or metastatic esophageal squamous cell carcinoma: The ESCORT-1st trial. Med Sep 13(9):1137\u0026ndash;1149e3. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.medj.2024.05.008\u003c/span\u003e\u003cspan address=\"10.1016/j.medj.2024.05.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Y, Liu C, Chen H et al (2024) Clinical efficacy and identification of factors confer resistance to afatinib (tyrosine kinase inhibitor) in EGFR-overexpressing esophageal squamous cell carcinoma. Signal Transduct Target Ther Jun 28(1):153. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41392-024-01875-4\u003c/span\u003e\u003cspan address=\"10.1038/s41392-024-01875-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang X, Zeng L, Li Y et al (2021) Anlotinib combined with PD-1 blockade for the treatment of lung cancer: a real-world retrospective study in China. Cancer Immunol Immunother Sep 70(9):2517\u0026ndash;2528. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00262-021-02869-9\u003c/span\u003e\u003cspan address=\"10.1007/s00262-021-02869-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu C, Han J, Han D, Huang W, Li B (2023) A new risk score model based on lactate dehydrogenase for predicting prognosis in esophageal squamous cell carcinoma treated with chemoradiotherapy. J Thorac disease Apr 28(4):2116\u0026ndash;2128. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.21037/jtd-23-388\u003c/span\u003e\u003cspan address=\"10.21037/jtd-23-388\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreten FR, Grivennikov SI (2019) Inflammation and Cancer: Triggers, Mechanisms, and Consequences. Immun Jul 16(1):27\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.immuni.2019.06.025\u003c/span\u003e\u003cspan address=\"10.1016/j.immuni.2019.06.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell Mar 4(5):646\u0026ndash;674. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2011.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2011.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu J, Gao D, Li J, Hu G, Liu J, Liu D (2022) The Predictive Value of Systemic Inflammatory Factors in Advanced, Metastatic Esophageal Squamous Cell Carcinoma Patients Treated with Camrelizumab. Onco Targets Ther 15:1161\u0026ndash;1170. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/OTT.S382967\u003c/span\u003e\u003cspan address=\"10.2147/OTT.S382967\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu X, Han R, Zhong Y, Weng N, Zhang A (2021) Post treatment NLR is a predictor of response to immune checkpoint inhibitor therapy in patients with esophageal squamous cell carcinoma. Cancer cell Int Jul 7(1):356. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12935-021-02072-x\u003c/span\u003e\u003cspan address=\"10.1186/s12935-021-02072-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark SH, Lee S, Song JH et al (2020) Prognostic significance of body mass index and prognostic nutritional index in stage II/III gastric cancer. Eur J Surg Oncol Apr 46(4 Pt A):620\u0026ndash;625. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ejso.2019.10.024\u003c/span\u003e\u003cspan address=\"10.1016/j.ejso.2019.10.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang L, Ma W, Qiu Z et al (2023) Prognostic nutritional index as a prognostic biomarker for gastrointestinal cancer patients treated with immune checkpoint inhibitors. Front Immunol 14:1219929. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fimmu.2023.1219929\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2023.1219929\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Advanced esophageal squamous cell carcinoma, immune checkpoint inhibitors, prediction model, real-world data ","lastPublishedDoi":"10.21203/rs.3.rs-5322833/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5322833/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe efficacy of immune checkpoint inhibitors (ICIs) for advanced esophageal squamous cell carcinoma (ESCC) remains suboptimal. This study aims to construct and validate a clinically accessible model to better identify populations that may potentially benefit from ICIs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study enrolled advanced ESCC patients treated with ICIs at Peking University Cancer Hospital from January 14, 2016, to January 26, 2024, forming the training cohort. Combined positive score (CPS) was recorded to evaluate the predictive value of programmed cell death ligand-1 (PD-L1). Baseline clinical characteristics and laboratory test results were identified as predictors through a 2-phase selection based on Cox proportional hazard regression and minimization of Akaike information criterion (AIC). The prediction model was internally validated using bootstrapping and externally validated in patients from Harbin Medical University Cancer Hospital between January 10, 2019, and July 6, 2022.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 430 patients from Peking University Cancer Hospital and 184 patients from Harbin Medical University Cancer Hospital were ultimately enrolled. PD-L1 expression failed to discriminate survival outcomes (HR=0.94, 95% CI: 0.74-1.19, \u003cem\u003eP = .6\u003c/em\u003e). The final model incorporates 10 variables: stage, bone metastasis, line of therapy, treatment, lactate dehydrogenase, carcinoembryonic antigen, carbohydrate antigen 199, lymphocyte count, prognostic nutritional index, and systemic immune-inflammation index. The C-index was 0.725 (95%CI: 0.694-0.756) in the training cohort, 0.722 (95%CI: 0.688-0.751) after bootstrapping, and 0.691 (95%CI: 0.650-0.733) in the external validation cohort, outperforming PD-L1 in prognostic prediction and risk stratification. An interactive online prediction tool (https://escc-survival.shinyapps.io/shiny_app/) was subsequently developed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis is the first model for individualized survival prediction in advanced ESCC patients treated with ICIs based on large-scale, high-quality real-world data, potentially guiding clinical decision-making and optimize treatment strategies.\u003c/p\u003e","manuscriptTitle":"An integrated prognosis prediction model based on real-word clinical characteristics for immunotherapy in advanced esophageal squamous cell carcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-06 11:02:55","doi":"10.21203/rs.3.rs-5322833/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f4666e3d-35df-4df0-932d-381a53003e4b","owner":[],"postedDate":"November 6th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":39560022,"name":"Biological sciences/Cancer/Cancer therapy/Cancer immunotherapy"},{"id":39560023,"name":"Health sciences/Medical research/Biomarkers/Prognostic markers"},{"id":39560024,"name":"Health sciences/Oncology/Cancer/Gastrointestinal cancer/Oesophageal cancer"}],"tags":[],"updatedAt":"2024-11-06T11:02:58+00:00","versionOfRecord":[],"versionCreatedAt":"2024-11-06 11:02:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5322833","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5322833","identity":"rs-5322833","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}