Development and Validation of a Cancer-Specific Early Death Prediction Model for Patients with Gastric Cancer with Liver Metastasis: Based on Machine Learning

Development and Validation of a Cancer-Specific Early Death Prediction Model for Patients with Gastric Cancer with Liver Metastasis: Based on Machine Learning | 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 Development and Validation of a Cancer-Specific Early Death Prediction Model for Patients with Gastric Cancer with Liver Metastasis: Based on Machine Learning Yulan Zhu, Xiaolong Chen, Peiling Ye, Ka Li, Min LIAO, Yu LUO, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4485633/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Gastric cancer with liver metastasis (GCLM) patients typically have a grim prognosis and are at high risk of early mortality. This study aimed to predict cancer-specific early mortality and risk factors for GCLM patients through machine learning (ML) methods. Methods The data of patients with GCLM were obtained from the SEER database. LASSO regression, univariate and multivariate logistic regression analyses were employed to identify significant independent risk factors for cancer-specific early death (CSED). Models such as logistic regression (LR), decision tree (DT), K-nearest neighbors (KNN), light gradient boosting machine (LightGBM), random forest (RF), support vector machine (SVM), and extreme gradient boosting (XGBoost) were used to predict the CSED and extract important features. Tenfold cross-validation, receiver operating characteristic (ROC) curve analysis, accuracy, balance accuracy, precision, sensitivity, specificity, F1-score, precision‒recall (PR) curve analysis, calibration curve analysis and decision curve analysis (DCA) were utilized to assess the performance of the models. The DALEX package was used to compute feature importance. Results The study recruited a total of 3661 patients. A total of 1648 (45%) patients experienced CSED. Among the 7 ML models, the XGBoost model achieved the best performance. The top 6 most influential factors were chemotherapy, months from diagnosis to therapy, age, grade, N stage, and surgery in the XGBoost model, with chemotherapy being the most significant. Conclusion The XGBoost model might be applied to predict the CSED of GCLM patients, and chemotherapy was the most important feature in the XGBoost model. These results could offer crucial reference data to assist clinicians in making informed decisions beforehand. Health sciences/Oncology Health sciences/Oncology/Surgical oncology Gastric cancer with liver metastasis (GCLM) SEER cancer-specific early death (CSED) machine learning (ML) Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Gastric cancer (GC) is a major health concern worldwide. In 2020, global statistics revealed 1,089,103 new cases of GC and 768,793 deaths, making it the fifth most common cancer in terms of incidence and the fourth leading cause of cancer death( 1 ). Distant metastases account for the majority of GC-related deaths, with a 5-year survival rate of less than 5% for metastatic GC patients( 2 , 3 ). The liver is the primary site for distant metastasis in patients with gastric cancer (GC)( 4 , 5 ). Upon diagnosis, 35% of patients exhibit distant metastases, with liver metastases occurring in 4–14% of patients with GC( 6 ). Patients with GCLM have a grim prognosis, typically surviving only approximately 4 months on average( 7 ). Despite advancements in medical care that have improved the prognosis for metastatic GC patients in recent decades, many individuals with GCLM still die early. Therefore, accurately identifying the risk factors linked to CSED in GCLM patients is essential. Although there have been studies reporting on early death in metastatic GC patients and stage IV GC patients and the prognosis of GCLM( 5 , 8 , 9 ), there is limited literature on the risk factors for CSED in patients with GCLM. The Surveillance, Epidemiology, and End Results (SEER) database is an authoritative source of cancer statistics in the United States. In this study, we utilized data from the SEER registry for patients diagnosed with GCLM between 2010 and 2015, developing predictive models for cancer-specific early death (CSED) using machine learning (ML) techniques, with the aim of aiding clinicians in making treatment decisions. 2 Methods 2.1 Database and patient selection This study was a retrospective analysis utilizing the SEER database, covering more than 28% of the U.S. population. SEERStat, provided by SEER, was utilized to select eligible patients, and this information is available on the official website ( https://seer.cancer.gov/ ). The study included patients diagnosed with GCLM between 2010 and 2015, identified according to the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) pathology types (8010–8231, 8255–8576), and tumor sites (C16.0 ~ C16.6, C16.8 ~ C16.9). The inclusion criteria were as follows: ( 1 ) age ≥ 18 years at diagnosis; ( 2 ) gastric cancer as the sole primary malignant tumor; ( 3 ) confirmed liver metastasis; ( 4 ) clear survival time available; and ( 5 ) diagnosis based on pathological specimens rather than death certificates or autopsies. The exclusion criteria were ( 1 ) the presence of cancers other than gastric cancer; ( 2 ) unknown distant metastasis information; ( 3 ) incomplete follow-up; ( 4 ) absence of primary cancer; ( 5 ) unclear survival time; and ( 6 ) a diagnosis according to an autopsy or death certificate. Ultimately, the study included 3661 patients with GCLM who were randomly assigned to two cohorts (7:3): a training cohort (2561 patients) and a validation cohort (1100 patients). The process of patient selection is illustrated in Fig. 1. The SEER database is an open-access resource containing deidentified patient data. This study did not require approval from the Institutional Review Board of West China Hospital, Sichuan University. 2.2 Variables and definition of early death Early death was characterized as death occurring within three months of the initial diagnosis, as reported in previous studies ( 5 ). Each patient was monitored for a minimum duration of three months, or the date of death was recorded. Descriptive statistics were employed to outline the baseline characteristics of the patients, categorized by age (< 65 and ≥ 65), sex (male and female), marital status (married, unmarried, unknown), year of diagnosis (2010–2012 and 2013–2015), race (white, black, other), primary site (C16.0Cardia, C16.1Fundus of stomach, C16.2Body of stomach, C16.3Gastric antrum, C16.4Pylorus, C16.5Lesser curvature of stomach, C16.6Greater curvature of stomach, C16.8Overlapping lesion of stomach, C16.9Stomach), Grade classification (I-II, III-IV, Unknown), histological type (adenocarcinoma, Signet ring cell, Other), T stage (T1,T2, T3,T4, TX), N stage (N0, N1,N2,N3, NX), surgery (yes or no), nonprimary surgery (yes or no), radiotherapy (yes or no), chemotherapy (yes or no), bone metastasis (yes or no), brain metastasis (yes or no), lung metastasis (yes or no), and median household income( $ 74,999) and months from diagnosis to therapy(0,≥1,unknown). 2.3 Statistical analysis All categorical data are presented as frequencies (percentages). To mitigate overfitting in the multifactorial model, we performed least absolute shrinkage and selection operator (LASSO) regression analysis on the training dataset to select variables. Subsequently, we conducted univariate analysis on the variables with nonzero coefficients identified from the LASSO regression to further exclude irrelevant variables. Variables that showed statistical significance in the univariate analysis were subsequently included in the multivariate analysis to identify the independent variables for constructing the ML model. The final candidates for the ML model were determined based on variables with p < 0.05. All the statistical analyses were performed using R software (version 4.3.0). 2.4 Model Construction and Evaluation Seven early death prediction models were developed using LR, DT, KNN, LightGBM, RF, SVM and XGBoost. We employed tenfold cross-validation to mitigate the randomness in model training and prevent overfitting. Subsequently, we assessed and compared the performance of each model using the validation dataset. In our scenario, we utilized metrics such as the area under the curve (AUC), accuracy, balance accuracy, precision, sensitivity, specificity and F1 score to assess model performance. Calibration curves were generated to assess the consistency of the models. Moreover, DCA was employed to assess the net clinical benefit. DCA represents an innovative approach for evaluating the predictive performance of various models. In essence, it utilizes analysis metrics termed net benefits, calculated using true positives and false positives of certain models, to place the advantages and disadvantages of models on the same scale. To further elucidate the performance of various models, we utilized the DALEX package to enhance the comprehensibility of different algorithms and contrast the forecasting capability of various models. We also calculated the significance of variables in the best-performing model using the DALEX package. 3 Results 3.1 Demographic and Clinical Characteristics Among the 3661 GCML patients, 1648 (45%) experienced CSED. Table 1 shows the baseline patient characteristics. With respect to demographic features, patients prone to CSED were aged ≥ 65 years (61.8%), male (69.6%), white (70.8%), brain metastasis (97.6%), and had a median income between $ 60,000 and $ 74,999 (42.2%). In addition, patients with the following tumor characteristics were at greater risk of CSED: cardia (35.6%), Grade III-IV (54.9%), adenocarcinoma (77.5%), and TX stage (49.6%). Regarding treatment, patients with CSED died from non-surgery (94.5%), non-primary site surgery (97.3%), non-radiation therapy (88%, including patients with unknown radiation therapy), months from diagnosis to therapy ≥ 1 (23.4%, except for the unknown months) and non-chemotherapy (71.1%, including patients with unknown chemotherapy). Table 1 Demographic and clinical characteristics in the SEER database Characteristics Total(N = 3661) No early death(N = 2013) Cancer-specific early death (N = 1648) Age =65 1957 (53.5%) 939 (46.6%) 1018 (61.8%) Sex Male 2615 (71.4%) 1468 (72.9%) 1147 (69.6%) Female 1046 (28.6%) 545 (27.1%) 501 (30.4%) Marital.status Married 2119 (57.9%) 1258 (62.5%) 861 (52.2%) Unmarried 1393 (38.0%) 674 (33.5%) 719 (43.6%) Unknown 149 (4.1%) 81 (4%) 68 (4.1%) Year.of.diagnosis 2010–2012 1734 (47.4%) 930 (46.2%) 804 (48.8%) 2013–2015 1927 (52.6%) 1083 (53.8%) 844 (51.2%) Race White 2626 (71.7%) 1460 (72.5%) 1166 (70.8%) Black 599 (16.4%) 316 (15.7%) 283 (17.2%) Other 436 (11.9%) 237 (11.8%) 199 (12.1%) Primary.Site Cardia, NOS 1513 (41.3%) 926 (46%) 587 (35.6%) Fundus of stomach 200 (5.5%) 111 (5.5%) 89 (5.4%) Body of stomach 310 (8.5%) 164 (8.1%) 146 (8.9%) Gastric antrum 457 (12.5%) 231 (11.5%) 226 (13.7%) Pylorus 57 (1.6%) 30 (1.5%) 27 (1.6%) Lesser curvature of stomach NOS 207 (5.7%) 123 (6.1%) 84 (5.1%) Greater curvature of stomach NOS 114 (3.1%) 43 (2.1%) 71 (4.3%) Overlapping lesion of stomach 272 (7.4%) 126 (6.3%) 146 (8.9%) Stomach, NOS 531 (14.5%) 259 (12.9%) 272 (16.5%) Grade I-II 1009 (27.6%) 624 (31%) 385 (23.4%) III-IV 1883 (51.4%) 979 (48.6%) 904 (54.9%) Unknown 769 (21.0%) 410 (20.4%) 359 (21.8%) Histologic.type Adenocarcinoma 2869 (78.4%) 1592 (79.1%) 1277 (77.5%) Signet ring cell 255 (7.0%) 120 (6%) 135 (8.2%) Other 537 (14.7%) 301 (15%) 236 (14.3%) T T1 726 (19.8%) 387 (19.2%) 339 (20.6%) T2 127 (3.5%) 85 (4.2%) 42 (2.5%) T3 438 (12.0%) 307 (15.3%) 131 (7.9%) T4 674 (18.4%) 356 (17.7%) 318 (19.3%) TX 1696 (46.3%) 878 (43.6%) 818 (49.6%) N N0 1296 (35.4%) 702 (34.9%) 594 (36%) N1 1431 (39.1%) 832 (41.3%) 599 (36.3%) N2 167 (4.6%) 110 (5.5%) 57 (3.5%) N3 161 (4.4%) 98 (4.9%) 63 (3.8%) NX 606 (16.6%) 271 (13.5%) 335 (20.3%) Surgery No 3378 (92.3%) 1820 (90.4%) 1558 (94.5%) Yes 283 (7.7%) 193 (9.6%) 90 (5.5%) Non.primary.Surgery No 3529 (96.4%) 1926 (95.7%) 1603 (97.3%) Yes 132 (3.6%) 87 (4.3%) 45 (2.7%) Radiation No/Unknown 3093 (84.5%) 1643 (81.6%) 1450 (88%) Yes 568 (15.5%) 370 (18.4%) 198 (12%) Chemotherapy No/Unknown 1562 (42.7%) 391 (19.4%) 1171 (71.1%) Yes 2099 (57.3%) 1622 (80.6%) 477 (28.9%) Bone.metastasis No 3290 (89.9%) 1828 (90.8%) 1462 (88.7%) Yes 371 (10.1%) 185 (9.2%) 186 (11.3%) Brain.metastasis No 3595 (98.2%) 1986 (98.7%) 1609 (97.6%) Yes 66 (1.8%) 27 (1.3%) 39 (2.4%) Lung.metastasis No 2997 (81.9%) 1707 (84.8%) 1290 (78.3%) Yes 664 (18.1%) 306 (15.2%) 358 (21.7%) Median.household.income $ 74,999 1205 (32.9%) 712 (35.4%) 493 (29.9%) Months.from.diagnosis.to.therapy 0 808 (22.1%) 516 (25.6%) 292 (17.7%) >=1 1606 (43.9%) 1220 (60.6%) 386 (23.4%) Unknown 1247 (34.1%) 277 (13.8%) 970 (58.9%) No significant differences were detected between the training and validation cohorts in terms of age, sex, marital status, year of diagnosis, race, primary site, grade, histological type, TN stage, surgery, non-primary site surgery, radiation therapy, chemotherapy, bone metastasis, brain metastasis, lung metastasis, median household income, or months from diagnosis to therapy ( P >0.05) (Table 2 ). Thus, both the training and validation cohorts were suitable for subsequent analysis. Table 2 Demographic information of patients with GCML in training and validation cohorts. Characteristics Total(N = 3661) Train(N = 2561) Test(N = 1100) P Age =65 1957 (53.5%) 1384 (54%) 573 (52.1%) Sex Male 2615 (71.4%) 1812 (70.8%) 803 (73%) 0.18 Female 1046 (28.6%) 749 (29.2%) 297 (27%) Marital.status Married 2119 (57.9%) 1490 (58.2%) 629 (57.2%) 0.85 Unmarried 1393 (38.0%) 967 (37.8%) 426 (38.7%) Unknown 149 (4.1%) 104 (4.1%) 45 (4.1%) Year.of.diagnosis 2010–2012 1734 (47.4%) 1212 (47.3%) 522 (47.5%) 0.972 2013–2015 1927 (52.6%) 1349 (52.7%) 578 (52.5%) Race White 2626 (71.7%) 1839 (71.8%) 787 (71.5%) 0.471 Black 599 (16.4%) 409 (16%) 190 (17.3%) Other 436 (11.9%) 313 (12.2%) 123 (11.2%) Primary.Site Cardia, NOS 1513 (41.3%) 1069 (41.7%) 444 (40.4%) 0.68 Fundus of stomach 200 (5.5%) 143 (5.6%) 57 (5.2%) Body of stomach 310 (8.5%) 219 (8.6%) 91 (8.3%) Gastric antrum 457 (12.5%) 306 (11.9%) 151 (13.7%) Pylorus 57 (1.6%) 42 (1.6%) 15 (1.4%) Lesser curvature of stomach NOS 207 (5.7%) 149 (5.8%) 58 (5.3%) Greater curvature of stomach NOS 114 (3.1%) 82 (3.2%) 32 (2.9%) Overlapping lesion of stomach 272 (7.4%) 194 (7.6%) 78 (7.1%) Stomach, NOS 531 (14.5%) 357 (13.9%) 174 (15.8%) Grade I-II 1009 (27.6%) 694 (27.1%) 315 (28.6%) 0.467 III-IV 1883 (51.4%) 1334 (52.1%) 549 (49.9%) Unknown 769 (21.0%) 533 (20.8%) 236 (21.5%) Histologic.type Adenocarcinoma 2869 (78.4%) 2003 (78.2%) 866 (78.7%) 0.511 Signet ring cell 255 (7.0%) 173 (6.8%) 82 (7.5%) Other 537 (14.7%) 385 (15%) 152 (13.8%) T T1 726 (19.8%) 499 (19.5%) 227 (20.6%) 0.737 T2 127 (3.5%) 95 (3.7%) 32 (2.9%) T3 438 (12.0%) 307 (12%) 131 (11.9%) T4 674 (18.4%) 470 (18.4%) 204 (18.5%) TX 1696 (46.3%) 1190 (46.5%) 506 (46%) N N0 1296 (35.4%) 895 (34.9%) 401 (36.5%) 0.49 N1 1431 (39.1%) 1000 (39%) 431 (39.2%) N2 167 (4.6%) 117 (4.6%) 50 (4.5%) N3 161 (4.4%) 108 (4.2%) 53 (4.8%) NX 606 (16.6%) 441 (17.2%) 165 (15%) Surgery No 3378 (92.3%) 2359 (92.1%) 1019 (92.6%) 0.634 Yes 283 (7.7%) 202 (7.9%) 81 (7.4%) Non.primary.Surgery No 3529 (96.4%) 2469 (96.4%) 1060 (96.4%) 1 Yes 132 (3.6%) 92 (3.6%) 40 (3.6%) Radiation No/Unknown 3093 (84.5%) 2166 (84.6%) 927 (84.3%) 0.855 Yes 568 (15.5%) 395 (15.4%) 173 (15.7%) Chemotherapy No/Unknown 1562 (42.7%) 1088 (42.5%) 474 (43.1%) 0.761 Yes 2099 (57.3%) 1473 (57.5%) 626 (56.9%) Bone.metastasis No 3290 (89.9%) 2310 (90.2%) 980 (89.1%) 0.338 Yes 371 (10.1%) 251 (9.8%) 120 (10.9%) Brain.metastasis No 3595 (98.2%) 2519 (98.4%) 1076 (97.8%) 0.32 Yes 66 (1.8%) 42 (1.6%) 24 (2.2%) Lung.metastasis No 2997 (81.9%) 2084 (81.4%) 913 (83%) 0.261 Yes 664 (18.1%) 477 (18.6%) 187 (17%) Median.household.income $ 74,999 1205 (32.9%) 843 (32.9%) 362 (32.9%) Months.from.diagnosis.to.therapy 0 808 (22.1%) 558 (21.8%) 250 (22.7%) 0.742 >=1 1606 (43.9%) 1133 (44.2%) 473 (43%) Unknown 1247 (34.1%) 870 (34%) 377 (34.3%) 3.2 Correlation of variables with clinical outcomes Pearson correlation tests were conducted among all variables, and the correlation heatmap indicated no significant relationships between variables (Fig. 2A). The variance inflation factor (VIF) of all variables was < 10, indicating the absence of multicollinearity between variables (Fig. 2B). When CSED was the outcome, LASSO regression was used to select 7 of the 19 variables with non-zero coefficients (Fig. 3A, Fig. 3B). Univariate analysis demonstrated that age, grade, N stage, surgery, chemotherapy, and months from diagnosis to therapy were associated with CSED ( P < 0.05). According to multivariate logistic regression analysis, age, grade, N stage, surgery, chemotherapy, and months from diagnosis to therapy were determined to be independent risk factors for CSED in patients with GCLM. Specifically, age ≥ 65 years (1.28, 95% CI: 1.06–1.55, P = 0.010), Grade III-IV (1.94, 95% CI: 1.54–2.43, P < 0.001), and unknown stage (1.59, 95% CI: 1.21–2.10, P = 0.001) and NX (1.33, 95% CI: 1.01–1.76, P = 0.040) were found to be independent risk factors for CSED in GCLM patients. Surgery (0.35, 95% CI: 0.22–0.54, P < 0.001), chemotherapy (0.13, 95% CI: 0.09–0.18, P < 0.001) and months from diagnosis to therapy ≥ 1 (0.51, 95% CI: 0.40–0.65, P < 0.001) could reduce the risk of CSED in GCLM patients (Table 3 ). Table 3 The univariable and multivariable logistic regression analysis of cancer-specific early death in GCML patients. Characteristics Univariable OR (95%, P ) Multivariable OR (95%, P ) Age =65 1.75(1.50–2.05, P < 0.001) 1.28 (1.06–1.55, P = 0.010) Grade I-II Reference Reference III-IV 1.58 (1.31–1.91, P < 0.001) 1.94 (1.54–2.43, P < 0.001) Unknown 1.45 (1.16–1.83, P = 0.001) 1.59 (1.21–2.10, P = 0.001) N NO Reference Reference N1 0.75 (0.62–0.90, P = 0.002) 0.92 (0.73–1.14, P = 0.430) N2 0.61 (0.41–0.91, P = 0.015) 0.85 (0.53–1.36, P = 0.499) N3 0.66 (0.44-1.00, P = 0.051) 1.10 (0.67–1.81, P = 0.705) NX 1.31 (1.04–1.65, P = 0.021) 1.33 (1.01–1.76, P = 0.040) Surgery No Reference Reference Yes 0.51 (0.37–0.69, P < 0.001) 0.35 (0.22–0.54, P < 0.001) Chemotherapy No/Unknown Reference Reference Yes 0.10 (0.08–0.12, P < 0.001) 0.13 (0.09–0.18, P =1 0.51 (0.41–0.64, P < 0.001) 0.51 (0.40–0.65, P < 0.001) Unknown 5.89 (4.65–7.44, P < 0.001) 0.94 (0.63–1.41, P = 0.764) OR, odds ratio; CI, confidence interval. 3.3 Development and validation of predictive models Seven ML models—LR, DT, KNN, RF, SVM, LightGBM and XGBoost—were constructed to predict the 6 independent risk factors via multivariate analysis. The predictive performance in cross-validation is shown in Fig. 4. After 10-fold cross-validation, the XGBoost model exhibited the best performance (AUC = 0.799, SD = 0.008), followed by LR (AUC = 0.792, SD = 0.008), LightGBM (AUC = 0.792, SD = 0.013), SVM (AUC = 0.780, SD = 0.013), RF (AUC = 0.776, SD = 0.008) and DT (AUC = 0.771, SD = 0.014), with KNN showing the poorest performance (AUC = 0.589, SD = 0.011). Furthermore, model performance was comprehensively evaluated using accuracy, balanced accuracy, precision, sensitivity, specificity, and F1-scores. The XGBoost model achieved the highest scores, with an accuracy of 0.781, balance accuracy of 0.774, precision of 0.786, sensitivity of 0.705, specificity of 0.843 and F1-score of 0.743 (Table 4). Additionally, Fig. 5A displays the AUC values of the ROC curves for each model, with XGBoost (AUC = 0.814) performing the best, followed by LR (AUC = 0.813), LightGBM (AUC = 0.807), DT (AUC = 0.792), RF (AUC = 0.796), SVM (AUC = 0.779), and KNN (AUC = 0.592). Furthermore, we further evaluated the ability of the 7 models to identify positives at different thresholds through PR curves (Fig. 5B). The RF model exhibited the best performance, with an AUC value of 0.794, particularly for balancing precision and recall, followed by LightGBM (AUC = 0.788), LR (AUC = 0.756), XGBoost (AUC = 0.752), SVM (AUC = 0.732), DT (AUC = 0.712) and KNN (AUC = 0.596). Table 4 Comparison of CSED prediction performance of different models in GCLM patients Models F1 score Accuracy Bal_accuracy Precision Sensitivity Specificity Xgboost 0.743 0.781 0.774 0.786 0.705 0.843 Lightgbm 0.740 0.781 0.773 0.794 0.693 0.853 Logistic 0.738 0.767 0.764 0.747 0.729 0.798 DT 0.736 0.771 0.765 0.764 0.711 0.820 KNN 0.420 0.602 0.577 0.612 0.319 0.834 RF 0.741 0.775 0.770 0.770 0.715 0.825 SVM 0.742 0.776 0.770 0.772 0.713 0.828 The calibration plots were constructed for the 7 models, and the Brier scores (BS) were compared, where a lower BS indicated greater predictive accuracy. The results showed that the XGBoost model had the strongest consistency with the observed outcomes in predicting CSED (BS = 0.168), followed by LR (BS = 0.17), DT (BS = 0.171), SVM (BS = 0.174), LightGBM (BS = 0.196) and RF (BS = 0.206), with the KNN model exhibiting the lowest predictive accuracy (BS = 0.362) (Fig. 5C). Decision-based clinical analysis (DCA) curves of the validation data (Fig. 5D) showed that the XGBoost model had superior performance across a range of probability thresholds from 12–81%. The average net benefit of the XGBoost model exceeds that of the other models. We utilized the DALEX package to compute feature importance and displayed the rankings of the top 6 clinical variables with the greatest importance according to the XGBoost model (Fig. 6). Additionally, we generated partial dependence plots of XGBoost, which offer a graphical depiction of the marginal impact of features on the ML model's prediction outcomes. In these plots, the x-axis indicates the respective variables, while the y-axis denotes the occurrence of CSED. This provides a method for quantifying the relationship between features and risk. A key benefit of partial dependence plots is their capacity to illustrate the relationship between features and outcomes. Factors such as age ≥ 65 years, Grade III-IV and unknown, NX stage, not undergoing surgery, on-chemotherapy or unknown, and unknown months from diagnosis to therapy contribute to predicting the risk of CSED in GCLM patients (Supplementary Fig. 1). To further explore the contribution of these clinical features to individual patient predictions, we randomly selected a patient from the validation cohort for demonstration purposes. Through interpretable algorithms, we visualized which specific features increased the prediction of CSED for this patient and which variables decreased the prediction. This patient was aged ≥ 65 years, had Grade III-IV disease, had N stage N1 disease, did not undergo surgery, had an unknown or not undergone chemotherapy status, and had an unknown duration from diagnosis to therapy. The XGBoost model predicted a CSED risk of 81.4% for this GCLM patient based on clinical features. Among these features, age ≥ 65 years, Grade III-IV stage, no surgery, and chemotherapy status (No/Unknown) were factors contributing to an increased risk of CSED. Conversely, N stage N1 decreased the model's ability to predict CSED (Supplementary Fig. 2). Combining these results, our study suggested that the XGBoost prediction model, constructed using the aforementioned factors, demonstrates greater accuracy and clinical applicability for predicting CSED in GCLM patients than the other six models. 4 Discussion Despite a decrease in the incidence of GC in recent years, it continues to be one of the most prevalent cancer types globally and a leading cause of cancer-related mortality worldwide.( 1 , 10 ) According to the SEER database, despite advancements in treatment, the overall early mortality among GC patients remains high at 32.57%, with a cancer-specific mortality of 30.73%( 9 ). Distant organ metastasis significantly impacts the prognosis of gastric cancer patients, with the liver being one of the most frequent sites of such metastasis. Patients with GCLM generally experience unfavorable outcomes, often encounter early death, and encounter significant controversy regarding comprehensive treatment approaches.( 11 , 12 , 13 , 14 ) In recent years, the early mortality of GCLM patients has attracted widespread attention among medical researchers. However, reports on cancer-specific early mortality and associated factors in GCLM patients are scarce. In our study, we observed that 45% of GCLM patients experienced CSED. A prior study utilizing the SEER database reported an early mortality rate of 49.6% among GCLM patients( 15 ), albeit reporting overall early mortality and not specifically cancer-specific early mortality, which is slightly higher than the results of our study. Furthermore, we found that age ≥ 65 years, Grade III-IV and unknown, NX stage, non-surgery, on-chemotherapy or unknown, and unknown months from diagnosis to therapy increased the risk of CSED in GCLM patients. Feng et al( 15 ) reported that age, primary site, histologic grade, tumor size, surgery, and distant metastasis were independent predictors of early death in patients with stage IV GC. Similarly, Zhu et al( 5 ) reported that race, grade, surgery, chemotherapy, and bone, brain, liver, or lung metastasis were found to be independent risk factors for early death in metastatic GC patients. Both studies utilized traditional statistical methods to establish nomograms. In contrast to traditional models, ML algorithms offer advantages such as greater adaptability, the ability to handle larger data volumes, higher dimensionality, and more frequent information exchange. These benefits have made ML algorithms a popular tool among medical researchers for detecting and predicting cancer.( 16 , 17 ) In our study, for the first time, we used 7 ML models trained with different feature sets to establish a predictive model for CSED in GCLM patients based on population data and evaluated the models' performance through 10-fold cross-validation, AUC, accuracy, balanced accuracy, precision, sensitivity, specificity, F1-score, calibration curves, and DCA. Additionally, we employed the DALEX package to compute feature importance and demonstrated the contribution of these features to the outcomes, ensuring the reliability of the results. According to the feature ranking plots of the XGBoost model, chemotherapy was the most significant influencing factor. Despite previous research indicating higher toxicity and risk of complications associated with chemotherapy, we discovered that GCLM patients who underwent chemotherapy exhibited a markedly lower early mortality rate than those who did not ( P < 0.001). Chemotherapy is a vital component of the treatment strategy for GCLM. Chemotherapy can kill or inhibit the progression of cancer cells, reduce tumor size, and even stabilize the disease, ultimately positively enhancing the overall survival of GC patients ( 18 ). In particular, combination chemotherapy offers significant advantages in treatment outcomes( 19 , 20 , 21 ). Currently, chemotherapy is currently the primary treatment for advanced GC globally( 20 , 22 ). With the aging of the global population, the incidence of GC among elderly patients is also increasing( 23 ). Statistics indicate that more than 60% of GC patients are aged ≥ 65 years( 24 ). A study on elderly patients with locally advanced GC indicated that age is a standalone risk factor for cancer-specific survival( 25 ). This may be attributed to the higher incidence of severe complications among elderly patients, particularly the postoperative pulmonary complications( 26 ), as also demonstrated in studies by Takeuchi et al( 27 ) and Yang et al( 28 ). Additionally, research suggests that younger patients are more willing to receive alternative treatments and exhibit greater tolerance to the side effects and adverse reactions of surgery and adjuvant therapy( 29 ). It has also been reported that elderly GC patients experience reduced physiological reserves and weaker tumor immune responses, ultimately leading to immune escape and tumor metastasis( 30 ). Tumor differentiation is an important prognostic indicator that reflects the biological behavior of the tumor itself. In our study, patients with poorly differentiated tumors exhibited significantly higher early mortality ( P < 0.001). This factor was also incorporated into a nomogram for predicting the prognosis of metastatic GC patients who underwent palliative gastrectomy.( 31 ) In addition, tumor differentiation was factored into a nomogram developed by Gao et al.( 32 ) that predicts the prognosis of patients with stage III/IV GC. Poor tumor grading reflects high invasiveness and poor treatment response, which severely impact patient prognosis( 33 ). Furthermore, our study revealed that patients who postponed therapy for more than one month had a notably increased occurrence of early mortality compared to those who did not delay therapy ( P < 0.001). Previous research has shown that treatment delay is associated with increased overall mortality( 34 ). A systematic review and meta-analysis also demonstrated that a four-week delay in cancer therapy increased the risk of death from various cancer types ( 35 ). However, our study has several limitations. First, this study is retrospective, potentially introducing selection bias. In the future, prospective clinical data are needed to establish more reliable evidence regarding the clinical applicability of our work. Second, research has demonstrated that the degrees of liver metastasis and peritoneal metastasis are separate criteria that might predict the prognosis of GCLM patients( 36 ). As the SEER database does not include these data in its reports, it may affect the predictive ability of the model. Third, the SEER database contains only 30% of the entire US population, potentially limiting the extensiveness of the study sample. 5 Conclusion In this study, 45% of GCLM patients experienced CSED. Furthermore, age, grade, N stage, surgery, chemotherapy and months from diagnosis to therapy were identified as independent risk factors for CSED in GCLM patients. This study provides the first analysis of CSED prediction in a population of GCLM patients, which will offer doctors a reliable point of reference for effective screening for CSED in GCLM patients, thereby developing better treatment strategies and providing assistance for the clinical treatment of GCLM to improve patient prognosis. Declarations A uthor contributions All the authors contributed to the study conception and design. The material preparation, data collection and analysis were performed by Y.Z., X.C., M.L., Y.L., Z.L.and P.Y.. The first draft of the manuscript was written by Y.Z., and all the authors commented on previous versions of the manuscript. All the authors have read and approved the final manuscript. Competing Interests The authors have no relevant financial or non-financial interests to disclose. Funding This study was supported by the Sichuan Science and Technology Program, Sichuan, Chengdu (No. 2023YFS0066). Acknowledgments The authors thank the SEER database for the availability of the data. Data availability statement The datasets generated and/or analyzed during the current study are available in the SEER database (https://seer.cancer.gov/). Ethics approval We received permission to access the research data in the SEER program from the National Cancer Institute, US. Approval was waived by the local ethics committee, as SEER data are publicly available and deidentified. References Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-249. Rawla P, Barsouk A. Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol. 2019;14(1):26-38. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin. 2019;69(1):7-34. Yoo CH, Noh SH, Shin DW, Choi SH, Min JS. Recurrence following curative resection for gastric carcinoma. Br J Surg. 2000;87(2):236-242. Zhu Y, Fang X, Wang L, Zhang T, Yu D. A Predictive Nomogram for Early Death of Metastatic Gastric Cancer: A Retrospective Study in the SEER Database and China. J Cancer. 2020;11(18):5527-5535. Shin A, Kim J, Park S. Gastric Cancer Epidemiology in Korea. J Gastric Cancer. 2011;11(3):135-140. Qiu MZ, Shi SM, Chen ZH, Yu HE, Sheng H, Jin Y, et al. Frequency and clinicopathological features of metastasis to liver, lung, bone, and brain from gastric cancer: A SEER-based study. Cancer Med. 2018;7(8):3662-3672. An W, Bao L, Wang C, Zheng M, Zhao Y. Analysis of Related Risk Factors and Prognostic Factors of Gastric Cancer with Liver Metastasis: A SEER and External Validation Based Study. Int J Gen Med. 2023;16:5969-5978. Yang Y, Chen ZJ, Yan S. The incidence, risk factors and predictive nomograms for early death among patients with stage IV gastric cancer: a population-based study. J Gastrointest Oncol. 2020;11(5):964-982. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA: a cancer journal for clinicians. 2022;72(1). Thrift AP, El-Serag HB. Burden of Gastric Cancer. Clin Gastroenterol Hepatol. 2020;18(3):534-542. Minciuna CE, Tudor S, Micu A, Diaconescu A, Alexandrescu ST, Vasilescu C. Safety and Efficacy of Simultaneous Resection of Gastric Carcinoma and Synchronous Liver Metastasis-A Western Center Experience. Medicina (Kaunas). 2022;58(12). Kawahara K, Makino H, Kametaka H, Hoshino I, Fukada T, Seike K, et al. Outcomes of surgical resection for gastric cancer liver metastases: a retrospective analysis. World J Surg Oncol. 2020;18(1):41. Cheng J, Cai M, Shuai X, Gao J, Wang G, Tao K. Multimodal treatments for resectable gastric cancer: A systematic review and network meta-analysis. Eur J Surg Oncol. 2019;45(10):1796-1805. Feng Y, Guo K, Jin H, Xiang Y, Zhang Y, Ruan S. A Predictive Nomogram for Early Mortality in Stage IV Gastric Cancer. Med Sci Monit. 2020;26:e923931. Ngiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery. Lancet Oncol. 2019;20(5):e262-e273. Li S, Yi H, Leng Q, Wu Y, Mao Y. New perspectives on cancer clinical research in the era of big data and machine learning. Surg Oncol. 2024;52:102009. Hironaka S, Sugimoto N, Yamaguchi K, Moriwaki T, Komatsu Y, Nishina T, et al. S-1 plus leucovorin versus S-1 plus leucovorin and oxaliplatin versus S-1 plus cisplatin in patients with advanced gastric cancer: a randomised, multicentre, open-label, phase 2 trial. Lancet Oncol. 2016;17(1):99-108. Pernot S, Mitry E, Samalin E, Dahan L, Dalban C, Ychou M, et al. Biweekly docetaxel, fluorouracil, leucovorin, oxaliplatin (TEF) as first-line treatment for advanced gastric cancer and adenocarcinoma of the gastroesophageal junction: safety and efficacy in a multicenter cohort. Gastric Cancer. 2014;17(2):341-347. Nishikawa K, Kawakami H, Shimokawa T, Fujitani K, Tamura S, Endo S, et al. Meta-analysis of three randomized trials of capecitabine plus cisplatin (XP) versus S-1 plus cisplatin (SP) as first-line treatment for advanced gastric cancer. Int J Clin Oncol. 2023;28(11):1501-1510. Ma X, Zhang Y, Wang C, Yu J. Efficacy and safety of combination chemotherapy regimens containing taxanes for first-line treatment in advanced gastric cancer. Clin Exp Med. 2023;23(2):381-396. Hsieh M-c, Wang S-H, Rau K-M. Real world analysis of adjuvant chemotherapy for advanced gastric cancer after D2 radical surgery. Annals of Oncology. 2017;28:iii36-iii37. Fujiwara Y, Fukuda S, Tsujie M, Ishikawa H, Kitani K, Inoue K, et al. Effects of age on survival and morbidity in gastric cancer patients undergoing gastrectomy. World J Gastrointest Oncol. 2017;9(6):257-262. Zheng Y, Wu C. [Prevalence and trend of gastrointestinal malignant tumors in the elderly over 75 years old in China]. Zhonghua Wei Chang Wai Ke Za Zhi. 2016;19(5):481-485. Sun Y, Li Z, Tian Y, Gao C, Liang B, Cao S, et al. Development and validation of nomograms for predicting overall survival and cancer-specific survival in elderly patients with locally advanced gastric cancer: a population-based study. BMC Gastroenterol. 2023;23(1):117. Wong JU, Tai FC, Huang CC. An examination of surgical and survival outcomes in the elderly (65-79 years of age) and the very elderly (≥80 years of age) who received surgery for gastric cancer. Curr Med Res Opin. 2020;36(2):229-233. Takeuchi D, Koide N, Suzuki A, Ishizone S, Shimizu F, Tsuchiya T, et al. Postoperative complications in elderly patients with gastric cancer. J Surg Res. 2015;198(2):317-326. Yang JY, Lee HJ, Kim TH, Huh YJ, Son YG, Park JH, et al. Short- and Long-Term Outcomes After Gastrectomy in Elderly Gastric Cancer Patients. Ann Surg Oncol. 2017;24(2):469-477. O'Connell JB, Maggard MA, Livingston EH, Yo CK. Colorectal cancer in the young. Am J Surg. 2004;187(3):343-348. Schlesinger-Raab A, Mihaljevic AL, Egert S, Emeny R, Jauch KW, Kleeff J, et al. Outcome of gastric cancer in the elderly: a population-based evaluation of the Munich Cancer Registry. Gastric Cancer. 2016;19(3):713-722. Ma T, Wu ZJ, Xu H, Wu CH, Xu J, Peng WR, et al. Nomograms for predicting survival in patients with metastatic gastric adenocarcinoma who undergo palliative gastrectomy. BMC Cancer. 2019;19(1):852. Gao Z, Ni J, Ding H, Yan C, Ren C, Li G, et al. A nomogram for prediction of stage III/IV gastric cancer outcome after surgery: A multicenter population-based study. Cancer Med. 2020;9(15):5490-5499. Johncilla M, Chen Z, Sweeney J, Yantiss RK. Tumor Grade Is Prognostically Relevant Among Mismatch Repair Deficient Colorectal Carcinomas. Am J Surg Pathol. 2018;42(12):1686-1692. Cone EB, Marchese M, Paciotti M, Nguyen D-D, Nabi J, Cole AP, et al. Assessment of Time-to-Treatment Initiation and Survival in a Cohort of Patients With Common Cancers. JAMA Network Open. 2020;3(12):e2030072-e2030072. Hanna TP, King WD, Thibodeau S, Jalink M, Paulin GA, Harvey-Jones E, et al. Mortality due to cancer treatment delay: systematic review and meta-analysis. BMJ. 2020;371:m4087. Ueda K, Iwahashi M, Nakamori M, Nakamura M, Naka T, Ishida K, et al. Analysis of the prognostic factors and evaluation of surgical treatment for synchronous liver metastases from gastric cancer. Langenbecks Arch Surg. 2009;394(4):647-653. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigure1.jpg SupplementaryFigure2.jpg 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-4485633","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":312679521,"identity":"e74baad1-0b04-4330-b667-2b83bc884b14","order_by":0,"name":"Yulan Zhu","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yulan","middleName":"","lastName":"Zhu","suffix":""},{"id":312679522,"identity":"724eeba4-d857-4e53-838a-83e1a15c6f27","order_by":1,"name":"Xiaolong Chen","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Xiaolong","middleName":"","lastName":"Chen","suffix":""},{"id":312679526,"identity":"d27ef03b-73fd-4322-8b26-54699df8d55e","order_by":2,"name":"Peiling Ye","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Peiling","middleName":"","lastName":"Ye","suffix":""},{"id":312679528,"identity":"230b2a69-c969-49ea-b436-761951db8db3","order_by":3,"name":"Ka Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAtElEQVRIie3PIQ7CUAyA4b68pDMvm32KcYSRGY7TGWY2P0my5CnCEeAW6LeQTI0DEMwIF0AOAwyJIOkcol9t/6QFEOIPhRoBqHrNMKg9L8FP0nc6DU1LzGQcdXU629lizkwCzO6EmDsoAIbqwDrsaMmY0sHJq0134STB2pK1pVNb0srxknqgJMlRm4SbYGtphDghWS3J+4UzSA3rlyhq0/Pj6eN4f2v6oWIkX/zEfSGEEL+8AUegM79e1ahTAAAAAElFTkSuQmCC","orcid":"","institution":"Sichuan University","correspondingAuthor":true,"prefix":"","firstName":"Ka","middleName":"","lastName":"Li","suffix":""},{"id":312679530,"identity":"8988fd5f-5fa5-463f-9036-2f06e83c79c3","order_by":4,"name":"Min LIAO","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"LIAO","suffix":""},{"id":312679533,"identity":"0efe0c19-960b-4d4d-8072-14794c3e1829","order_by":5,"name":"Yu LUO","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"LUO","suffix":""},{"id":312679534,"identity":"5b754605-0b36-40a0-afb1-e4c19bd9048c","order_by":6,"name":"ZhiYu LI","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"ZhiYu","middleName":"","lastName":"LI","suffix":""},{"id":312679535,"identity":"7eb766c9-f2db-4beb-ba15-29aed39b66fe","order_by":7,"name":"Yuwei LIU","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yuwei","middleName":"","lastName":"LIU","suffix":""}],"badges":[],"createdAt":"2024-05-27 14:14:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4485633/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4485633/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":58248468,"identity":"209bad8b-51de-46c4-a45b-d1cf130c9484","added_by":"auto","created_at":"2024-06-13 02:47:26","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1120610,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/ca6087ade9bbf05e93ef667f.jpg"},{"id":58247636,"identity":"da03d395-7e8a-463b-8ad8-5cf88fd47ecd","added_by":"auto","created_at":"2024-06-13 02:39:26","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":188923,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/f0ee3f7d21602abac74095de.jpg"},{"id":58247633,"identity":"709b620e-ff8d-4ca1-a58b-7e25197850cb","added_by":"auto","created_at":"2024-06-13 02:39:26","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":116492,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/e41c35ba781d99f7983de2b0.jpg"},{"id":58247637,"identity":"2c92d180-c052-46f7-86fe-72ed3f05e32a","added_by":"auto","created_at":"2024-06-13 02:39:26","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":122553,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/7066ed07a11845a22f55f5aa.jpg"},{"id":58247639,"identity":"52bbbcd2-d3fe-4d12-af09-e03a8335d1f0","added_by":"auto","created_at":"2024-06-13 02:39:27","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2356059,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/45f796f4f1a7fd2452e5fa8c.jpg"},{"id":58247638,"identity":"85cbaff1-936b-447e-aa78-a1a44fa9234c","added_by":"auto","created_at":"2024-06-13 02:39:27","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":474354,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/1ee9146a85dba37c4b2b0be1.jpg"},{"id":61158919,"identity":"8d600520-d87a-483c-98be-b91b1400a19f","added_by":"auto","created_at":"2024-07-26 11:39:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5454674,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/6f88986b-8759-44f5-8f67-6aba0b1a55e0.pdf"},{"id":58247631,"identity":"80bd4545-de92-469b-ba4f-b94067eeb703","added_by":"auto","created_at":"2024-06-13 02:39:26","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1149995,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/e6bb06dd1bc3c2f721274d5d.jpg"},{"id":58247635,"identity":"2b5a762a-d636-4711-9005-72d0c760e31f","added_by":"auto","created_at":"2024-06-13 02:39:26","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":964122,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4485633/v1/c82042cb12e028a0a16b2e62.jpg"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Cancer-Specific Early Death Prediction Model for Patients with Gastric Cancer with Liver Metastasis: Based on Machine Learning","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eGastric cancer (GC) is a major health concern worldwide. In 2020, global statistics revealed 1,089,103 new cases of GC and 768,793 deaths, making it the fifth most common cancer in terms of incidence and the fourth leading cause of cancer death(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Distant metastases account for the majority of GC-related deaths, with a 5-year survival rate of less than 5% for metastatic GC patients(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). The liver is the primary site for distant metastasis in patients with gastric cancer (GC)(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Upon diagnosis, 35% of patients exhibit distant metastases, with liver metastases occurring in 4\u0026ndash;14% of patients with GC(\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). Patients with GCLM have a grim prognosis, typically surviving only approximately 4 months on average(\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Despite advancements in medical care that have improved the prognosis for metastatic GC patients in recent decades, many individuals with GCLM still die early. Therefore, accurately identifying the risk factors linked to CSED in GCLM patients is essential. Although there have been studies reporting on early death in metastatic GC patients and stage IV GC patients and the prognosis of GCLM(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e), there is limited literature on the risk factors for CSED in patients with GCLM.\u003c/p\u003e \u003cp\u003eThe Surveillance, Epidemiology, and End Results (SEER) database is an authoritative source of cancer statistics in the United States. In this study, we utilized data from the SEER registry for patients diagnosed with GCLM between 2010 and 2015, developing predictive models for cancer-specific early death (CSED) using machine learning (ML) techniques, with the aim of aiding clinicians in making treatment decisions.\u003c/p\u003e"},{"header":"2 Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Database and patient selection\u003c/h2\u003e \u003cp\u003eThis study was a retrospective analysis utilizing the SEER database, covering more than 28% of the U.S. population. SEERStat, provided by SEER, was utilized to select eligible patients, and this information is available on the official website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://seer.cancer.gov/\u003c/span\u003e\u003cspan address=\"https://seer.cancer.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The study included patients diagnosed with GCLM between 2010 and 2015, identified according to the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) pathology types (8010\u0026ndash;8231, 8255\u0026ndash;8576), and tumor sites (C16.0\u0026thinsp;~\u0026thinsp;C16.6, C16.8\u0026thinsp;~\u0026thinsp;C16.9). The inclusion criteria were as follows: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) age\u0026thinsp;\u0026ge;\u0026thinsp;18 years at diagnosis; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) gastric cancer as the sole primary malignant tumor; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) confirmed liver metastasis; (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) clear survival time available; and (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) diagnosis based on pathological specimens rather than death certificates or autopsies. The exclusion criteria were (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) the presence of cancers other than gastric cancer; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) unknown distant metastasis information; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) incomplete follow-up; (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) absence of primary cancer; (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) unclear survival time; and (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e) a diagnosis according to an autopsy or death certificate. Ultimately, the study included 3661 patients with GCLM who were randomly assigned to two cohorts (7:3): a training cohort (2561 patients) and a validation cohort (1100 patients). The process of patient selection is illustrated in Fig.\u0026nbsp;1. The SEER database is an open-access resource containing deidentified patient data. This study did not require approval from the Institutional Review Board of West China Hospital, Sichuan University.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Variables and definition of early death\u003c/h2\u003e \u003cp\u003eEarly death was characterized as death occurring within three months of the initial diagnosis, as reported in previous studies (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Each patient was monitored for a minimum duration of three months, or the date of death was recorded. Descriptive statistics were employed to outline the baseline characteristics of the patients, categorized by age (\u0026lt;\u0026thinsp;65 and \u0026ge;\u0026thinsp;65), sex (male and female), marital status (married, unmarried, unknown), year of diagnosis (2010\u0026ndash;2012 and 2013\u0026ndash;2015), race (white, black, other), primary site (C16.0Cardia, C16.1Fundus of stomach, C16.2Body of stomach, C16.3Gastric antrum, C16.4Pylorus, C16.5Lesser curvature of stomach, C16.6Greater curvature of stomach, C16.8Overlapping lesion of stomach, C16.9Stomach), Grade classification (I-II, III-IV, Unknown), histological type (adenocarcinoma, Signet ring cell, Other), T stage (T1,T2, T3,T4, TX), N stage (N0, N1,N2,N3, NX), surgery (yes or no), nonprimary surgery (yes or no), radiotherapy (yes or no), chemotherapy (yes or no), bone metastasis (yes or no), brain metastasis (yes or no), lung metastasis (yes or no), and median household income(\u0026lt;\u003cspan\u003e$\u003c/span\u003e45,000/\u003cspan\u003e$\u003c/span\u003e45,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e59,999/\u003cspan\u003e$\u003c/span\u003e60,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e74,999/\u0026gt;\u003cspan\u003e$\u003c/span\u003e74,999) and months from diagnosis to therapy(0,\u0026ge;1,unknown).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Statistical analysis\u003c/h2\u003e \u003cp\u003eAll categorical data are presented as frequencies (percentages). To mitigate overfitting in the multifactorial model, we performed least absolute shrinkage and selection operator (LASSO) regression analysis on the training dataset to select variables. Subsequently, we conducted univariate analysis on the variables with nonzero coefficients identified from the LASSO regression to further exclude irrelevant variables. Variables that showed statistical significance in the univariate analysis were subsequently included in the multivariate analysis to identify the independent variables for constructing the ML model. The final candidates for the ML model were determined based on variables with \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05. All the statistical analyses were performed using R software (version 4.3.0).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Model Construction and Evaluation\u003c/h2\u003e \u003cp\u003eSeven early death prediction models were developed using LR, DT, KNN, LightGBM, RF, SVM and XGBoost. We employed tenfold cross-validation to mitigate the randomness in model training and prevent overfitting. Subsequently, we assessed and compared the performance of each model using the validation dataset. In our scenario, we utilized metrics such as the area under the curve (AUC), accuracy, balance accuracy, precision, sensitivity, specificity and F1 score to assess model performance. Calibration curves were generated to assess the consistency of the models. Moreover, DCA was employed to assess the net clinical benefit. DCA represents an innovative approach for evaluating the predictive performance of various models. In essence, it utilizes analysis metrics termed net benefits, calculated using true positives and false positives of certain models, to place the advantages and disadvantages of models on the same scale. To further elucidate the performance of various models, we utilized the DALEX package to enhance the comprehensibility of different algorithms and contrast the forecasting capability of various models. We also calculated the significance of variables in the best-performing model using the DALEX package.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Demographic and Clinical Characteristics\u003c/h2\u003e \u003cp\u003eAmong the 3661 GCML patients, 1648 (45%) experienced CSED. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the baseline patient characteristics. With respect to demographic features, patients prone to CSED were aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years (61.8%), male (69.6%), white (70.8%), brain metastasis (97.6%), and had a median income between \u003cspan\u003e$\u003c/span\u003e60,000 and \u003cspan\u003e$\u003c/span\u003e74,999 (42.2%). In addition, patients with the following tumor characteristics were at greater risk of CSED: cardia (35.6%), Grade III-IV (54.9%), adenocarcinoma (77.5%), and TX stage (49.6%). Regarding treatment, patients with CSED died from non-surgery (94.5%), non-primary site surgery (97.3%), non-radiation therapy (88%, including patients with unknown radiation therapy), months from diagnosis to therapy\u0026thinsp;\u0026ge;\u0026thinsp;1 (23.4%, except for the unknown months) and non-chemotherapy (71.1%, including patients with unknown chemotherapy).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic and clinical characteristics in the SEER database\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal(N\u0026thinsp;=\u0026thinsp;3661)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo early death(N\u0026thinsp;=\u0026thinsp;2013)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCancer-specific early death (N\u0026thinsp;=\u0026thinsp;1648)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1704 (46.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1074 (53.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e630 (38.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1957 (53.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e939 (46.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1018 (61.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2615 (71.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1468 (72.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1147 (69.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1046 (28.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e545 (27.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e501 (30.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eMarital.status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2119 (57.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1258 (62.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e861 (52.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnmarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1393 (38.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e674 (33.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e719 (43.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e149 (4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e81 (4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68 (4.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eYear.of.diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2010\u0026ndash;2012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1734 (47.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e930 (46.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e804 (48.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2013\u0026ndash;2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1927 (52.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1083 (53.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e844 (51.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRace\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2626 (71.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1460 (72.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1166 (70.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBlack\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e599 (16.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e316 (15.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e283 (17.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e436 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e237 (11.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e199 (12.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePrimary.Site\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCardia, NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1513 (41.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e926 (46%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e587 (35.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFundus of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e200 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e111 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e89 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e310 (8.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e164 (8.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e146 (8.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGastric antrum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e457 (12.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e231 (11.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e226 (13.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePylorus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e57 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 (1.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesser curvature of stomach NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e207 (5.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e123 (6.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e84 (5.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreater curvature of stomach NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e114 (3.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43 (2.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e71 (4.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverlapping lesion of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e272 (7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e126 (6.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e146 (8.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStomach, NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e531 (14.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e259 (12.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e272 (16.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI-II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1009 (27.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e624 (31%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e385 (23.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIII-IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1883 (51.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e979 (48.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e904 (54.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e769 (21.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e410 (20.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e359 (21.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHistologic.type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdenocarcinoma\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2869 (78.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1592 (79.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1277 (77.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSignet ring cell\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e255 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e120 (6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e135 (8.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e537 (14.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e301 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e236 (14.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e726 (19.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e387 (19.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e339 (20.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e85 (4.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e42 (2.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e438 (12.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e307 (15.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e131 (7.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e674 (18.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e356 (17.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e318 (19.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1696 (46.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e878 (43.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e818 (49.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1296 (35.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e702 (34.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e594 (36%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1431 (39.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e832 (41.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e599 (36.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e167 (4.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e110 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e57 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e161 (4.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e98 (4.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e63 (3.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e606 (16.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e271 (13.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e335 (20.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3378 (92.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1820 (90.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1558 (94.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e283 (7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e193 (9.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eNon.primary.Surgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3529 (96.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1926 (95.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1603 (97.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e132 (3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e87 (4.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45 (2.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo/Unknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3093 (84.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1643 (81.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1450 (88%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e568 (15.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e370 (18.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e198 (12%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eChemotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo/Unknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1562 (42.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e391 (19.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1171 (71.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2099 (57.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1622 (80.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e477 (28.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBone.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3290 (89.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1828 (90.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1462 (88.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e371 (10.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e185 (9.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e186 (11.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBrain.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3595 (98.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1986 (98.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1609 (97.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e66 (1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (1.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e39 (2.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eLung.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2997 (81.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1707 (84.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1290 (78.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e664 (18.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e306 (15.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e358 (21.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003eMedian.household.income\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u003cspan\u003e$\u003c/span\u003e45,000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e258 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e123 (6.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e135 (8.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cspan\u003e$\u003c/span\u003e45,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e59,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e677 (18.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e353 (17.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e324 (19.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cspan\u003e$\u003c/span\u003e60,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e74,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1521 (41.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e825 (41%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e696 (42.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u003cspan\u003e$\u003c/span\u003e74,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1205 (32.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e712 (35.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e493 (29.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003eMonths.from.diagnosis.to.therapy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e808 (22.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e516 (25.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e292 (17.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1606 (43.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1220 (60.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e386 (23.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1247 (34.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e277 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e970 (58.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNo significant differences were detected between the training and validation cohorts in terms of age, sex, marital status, year of diagnosis, race, primary site, grade, histological type, TN stage, surgery, non-primary site surgery, radiation therapy, chemotherapy, bone metastasis, brain metastasis, lung metastasis, median household income, or months from diagnosis to therapy (\u003cem\u003eP\u003c/em\u003e\u0026gt;0.05) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Thus, both the training and validation cohorts were suitable for subsequent analysis.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic information of patients with GCML in training and validation cohorts.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal(N\u0026thinsp;=\u0026thinsp;3661)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTrain(N\u0026thinsp;=\u0026thinsp;2561)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTest(N\u0026thinsp;=\u0026thinsp;1100)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1704 (46.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1177 (46%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e527 (47.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.294\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1957 (53.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1384 (54%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e573 (52.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2615 (71.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1812 (70.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e803 (73%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1046 (28.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e749 (29.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e297 (27%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eMarital.status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2119 (57.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1490 (58.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e629 (57.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnmarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1393 (38.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e967 (37.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e426 (38.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e149 (4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e104 (4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45 (4.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eYear.of.diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2010\u0026ndash;2012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1734 (47.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1212 (47.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e522 (47.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.972\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2013\u0026ndash;2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1927 (52.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1349 (52.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e578 (52.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRace\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2626 (71.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1839 (71.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e787 (71.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.471\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBlack\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e599 (16.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e409 (16%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e190 (17.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e436 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e313 (12.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e123 (11.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePrimary.Site\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCardia, NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1513 (41.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1069 (41.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e444 (40.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFundus of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e200 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e143 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e57 (5.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e310 (8.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e219 (8.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e91 (8.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGastric antrum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e457 (12.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e306 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e151 (13.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePylorus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e57 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15 (1.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesser curvature of stomach NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e207 (5.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e149 (5.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e58 (5.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreater curvature of stomach NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e114 (3.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e82 (3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32 (2.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverlapping lesion of stomach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e272 (7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e194 (7.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e78 (7.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStomach, NOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e531 (14.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e357 (13.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e174 (15.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI-II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1009 (27.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e694 (27.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e315 (28.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.467\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIII-IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1883 (51.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1334 (52.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e549 (49.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e769 (21.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e533 (20.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e236 (21.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eHistologic.type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdenocarcinoma\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2869 (78.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2003 (78.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e866 (78.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.511\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSignet ring cell\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e255 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e173 (6.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e82 (7.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e537 (14.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e385 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e152 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e726 (19.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e499 (19.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e227 (20.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.737\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95 (3.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32 (2.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e438 (12.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e307 (12%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e131 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e674 (18.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e470 (18.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e204 (18.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1696 (46.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1190 (46.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e506 (46%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1296 (35.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e895 (34.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e401 (36.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1431 (39.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1000 (39%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e431 (39.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e167 (4.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e117 (4.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50 (4.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e161 (4.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e108 (4.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e53 (4.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e606 (16.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e441 (17.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e165 (15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3378 (92.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2359 (92.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1019 (92.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e283 (7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e202 (7.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81 (7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eNon.primary.Surgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3529 (96.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2469 (96.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1060 (96.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e132 (3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e92 (3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e40 (3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo/Unknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3093 (84.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2166 (84.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e927 (84.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.855\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e568 (15.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e395 (15.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e173 (15.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eChemotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo/Unknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1562 (42.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1088 (42.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e474 (43.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.761\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2099 (57.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1473 (57.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e626 (56.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBone.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3290 (89.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2310 (90.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e980 (89.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.338\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e371 (10.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e251 (9.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e120 (10.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBrain.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3595 (98.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2519 (98.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1076 (97.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.32\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e66 (1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24 (2.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eLung.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2997 (81.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2084 (81.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e913 (83%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.261\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e664 (18.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e477 (18.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e187 (17%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003eMedian.household.income\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u003cspan\u003e$\u003c/span\u003e45,000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e258 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e172 (6.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e86 (7.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cspan\u003e$\u003c/span\u003e45,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e59,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e677 (18.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e481 (18.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e196 (17.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cspan\u003e$\u003c/span\u003e60,000\u0026ndash;\u003cspan\u003e$\u003c/span\u003e74,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1521 (41.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1065 (41.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e456 (41.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u003cspan\u003e$\u003c/span\u003e74,999\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1205 (32.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e843 (32.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e362 (32.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMonths.from.diagnosis.to.therapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e808 (22.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e558 (21.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e250 (22.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.742\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1606 (43.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1133 (44.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e473 (43%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1247 (34.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e870 (34%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e377 (34.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Correlation of variables with clinical outcomes\u003c/h2\u003e \u003cp\u003ePearson correlation tests were conducted among all variables, and the correlation heatmap indicated no significant relationships between variables (Fig.\u0026nbsp;2A). The variance inflation factor (VIF) of all variables was \u0026lt;\u0026thinsp;10, indicating the absence of multicollinearity between variables (Fig.\u0026nbsp;2B). When CSED was the outcome, LASSO regression was used to select 7 of the 19 variables with non-zero coefficients (Fig.\u0026nbsp;3A, Fig.\u0026nbsp;3B). Univariate analysis demonstrated that age, grade, N stage, surgery, chemotherapy, and months from diagnosis to therapy were associated with CSED (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). According to multivariate logistic regression analysis, age, grade, N stage, surgery, chemotherapy, and months from diagnosis to therapy were determined to be independent risk factors for CSED in patients with GCLM. Specifically, age\u0026thinsp;\u0026ge;\u0026thinsp;65 years (1.28, 95% CI: 1.06\u0026ndash;1.55, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.010), Grade III-IV (1.94, 95% CI: 1.54\u0026ndash;2.43, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and unknown stage (1.59, 95% CI: 1.21\u0026ndash;2.10, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001) and NX (1.33, 95% CI: 1.01\u0026ndash;1.76, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.040) were found to be independent risk factors for CSED in GCLM patients. Surgery (0.35, 95% CI: 0.22\u0026ndash;0.54, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), chemotherapy (0.13, 95% CI: 0.09\u0026ndash;0.18, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and months from diagnosis to therapy\u0026thinsp;\u0026ge;\u0026thinsp;1 (0.51, 95% CI: 0.40\u0026ndash;0.65, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) could reduce the risk of CSED in GCLM patients (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe univariable and multivariable logistic regression analysis of cancer-specific early death in GCML patients.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnivariable\u003c/p\u003e \u003cp\u003eOR (95%, \u003cem\u003eP\u003c/em\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMultivariable\u003c/p\u003e \u003cp\u003eOR (95%, \u003cem\u003eP\u003c/em\u003e)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.75(1.50\u0026ndash;2.05, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.28 (1.06\u0026ndash;1.55, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.010)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI-II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIII-IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.58 (1.31\u0026ndash;1.91, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.94 (1.54\u0026ndash;2.43, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.45 (1.16\u0026ndash;1.83, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.59 (1.21\u0026ndash;2.10, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.75 (0.62\u0026ndash;0.90, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.92 (0.73\u0026ndash;1.14, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.430)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.61 (0.41\u0026ndash;0.91, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.015)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.85 (0.53\u0026ndash;1.36, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.499)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.66 (0.44-1.00, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.051)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.10 (0.67\u0026ndash;1.81, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.705)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.31 (1.04\u0026ndash;1.65, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.021)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.33 (1.01\u0026ndash;1.76, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.040)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.51 (0.37\u0026ndash;0.69, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.35 (0.22\u0026ndash;0.54, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eChemotherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo/Unknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.10 (0.08\u0026ndash;0.12, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.13 (0.09\u0026ndash;0.18, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBone.metastasis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.08 (0.83\u0026ndash;1.40, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.574)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eMonths.from.diagnosis.to.therapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;=1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.51 (0.41\u0026ndash;0.64, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.51 (0.40\u0026ndash;0.65, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnknown\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.89 (4.65\u0026ndash;7.44, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.94 (0.63\u0026ndash;1.41, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.764)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eOR, odds ratio; CI, confidence interval.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Development and validation of predictive models\u003c/h2\u003e \u003cp\u003eSeven ML models\u0026mdash;LR, DT, KNN, RF, SVM, LightGBM and XGBoost\u0026mdash;were constructed to predict the 6 independent risk factors via multivariate analysis.\u003c/p\u003e \u003cp\u003eThe predictive performance in cross-validation is shown in Fig.\u0026nbsp;4. After 10-fold cross-validation, the XGBoost model exhibited the best performance (AUC\u0026thinsp;=\u0026thinsp;0.799, SD\u0026thinsp;=\u0026thinsp;0.008), followed by LR (AUC\u0026thinsp;=\u0026thinsp;0.792, SD\u0026thinsp;=\u0026thinsp;0.008), LightGBM (AUC\u0026thinsp;=\u0026thinsp;0.792, SD\u0026thinsp;=\u0026thinsp;0.013), SVM (AUC\u0026thinsp;=\u0026thinsp;0.780, SD\u0026thinsp;=\u0026thinsp;0.013), RF (AUC\u0026thinsp;=\u0026thinsp;0.776, SD\u0026thinsp;=\u0026thinsp;0.008) and DT (AUC\u0026thinsp;=\u0026thinsp;0.771, SD\u0026thinsp;=\u0026thinsp;0.014), with KNN showing the poorest performance (AUC\u0026thinsp;=\u0026thinsp;0.589, SD\u0026thinsp;=\u0026thinsp;0.011). Furthermore, model performance was comprehensively evaluated using accuracy, balanced accuracy, precision, sensitivity, specificity, and F1-scores. The XGBoost model achieved the highest scores, with an accuracy of 0.781, balance accuracy of 0.774, precision of 0.786, sensitivity of 0.705, specificity of 0.843 and F1-score of 0.743 (Table\u0026nbsp;4). Additionally, Fig.\u0026nbsp;5A displays the AUC values of the ROC curves for each model, with XGBoost (AUC\u0026thinsp;=\u0026thinsp;0.814) performing the best, followed by LR (AUC\u0026thinsp;=\u0026thinsp;0.813), LightGBM (AUC\u0026thinsp;=\u0026thinsp;0.807), DT (AUC\u0026thinsp;=\u0026thinsp;0.792), RF (AUC\u0026thinsp;=\u0026thinsp;0.796), SVM (AUC\u0026thinsp;=\u0026thinsp;0.779), and KNN (AUC\u0026thinsp;=\u0026thinsp;0.592). Furthermore, we further evaluated the ability of the 7 models to identify positives at different thresholds through PR curves (Fig.\u0026nbsp;5B). The RF model exhibited the best performance, with an AUC value of 0.794, particularly for balancing precision and recall, followed by LightGBM (AUC\u0026thinsp;=\u0026thinsp;0.788), LR (AUC\u0026thinsp;=\u0026thinsp;0.756), XGBoost (AUC\u0026thinsp;=\u0026thinsp;0.752), SVM (AUC\u0026thinsp;=\u0026thinsp;0.732), DT (AUC\u0026thinsp;=\u0026thinsp;0.712) and KNN (AUC\u0026thinsp;=\u0026thinsp;0.596).\u003c/p\u003e \u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e\u0026nbsp; Comparison of CSED prediction performance of different models in GCLM patients\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eModels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003eF1 score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003eBal_accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003ePrecision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003eSensitivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003eSpecificity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eXgboost\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.743\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.781\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.774\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.786\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.705\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.843\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eLightgbm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.740\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.781\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.773\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.794\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.693\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.853\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eLogistic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.738\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e0.767\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.764\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.747\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.729\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.798\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eDT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.736\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e0.771\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.765\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.764\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.711\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.820\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eKNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.420\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e0.602\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.577\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.612\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.319\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.834\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.741\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e0.775\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.770\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.770\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.715\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.825\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eSVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.706349206349206%\" valign=\"top\"\u003e\n \u003cp\u003e0.742\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.492063492063492%\" valign=\"top\"\u003e\n \u003cp\u003e0.776\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.857142857142858%\" valign=\"top\"\u003e\n \u003cp\u003e0.770\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.095238095238095%\" valign=\"top\"\u003e\n \u003cp\u003e0.772\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.682539682539682%\" valign=\"top\"\u003e\n \u003cp\u003e0.713\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.880952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e0.828\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\u003cp\u003eThe calibration plots were constructed for the 7 models, and the Brier scores (BS) were compared, where a lower BS indicated greater predictive accuracy. The results showed that the XGBoost model had the strongest consistency with the observed outcomes in predicting CSED (BS\u0026thinsp;=\u0026thinsp;0.168), followed by LR (BS\u0026thinsp;=\u0026thinsp;0.17), DT (BS\u0026thinsp;=\u0026thinsp;0.171), SVM (BS\u0026thinsp;=\u0026thinsp;0.174), LightGBM (BS\u0026thinsp;=\u0026thinsp;0.196) and RF (BS\u0026thinsp;=\u0026thinsp;0.206), with the KNN model exhibiting the lowest predictive accuracy (BS\u0026thinsp;=\u0026thinsp;0.362) (Fig.\u0026nbsp;5C).\u003c/p\u003e \u003cp\u003eDecision-based clinical analysis (DCA) curves of the validation data (Fig.\u0026nbsp;5D) showed that the XGBoost model had superior performance across a range of probability thresholds from 12\u0026ndash;81%. The average net benefit of the XGBoost model exceeds that of the other models.\u003c/p\u003e \u003cp\u003e We utilized the DALEX package to compute feature importance and displayed the rankings of the top 6 clinical variables with the greatest importance according to the XGBoost model (Fig.\u0026nbsp;6). Additionally, we generated partial dependence plots of XGBoost, which offer a graphical depiction of the marginal impact of features on the ML model's prediction outcomes. In these plots, the x-axis indicates the respective variables, while the y-axis denotes the occurrence of CSED. This provides a method for quantifying the relationship between features and risk. A key benefit of partial dependence plots is their capacity to illustrate the relationship between features and outcomes. Factors such as age\u0026thinsp;\u0026ge;\u0026thinsp;65 years, Grade III-IV and unknown, NX stage, not undergoing surgery, on-chemotherapy or unknown, and unknown months from diagnosis to therapy contribute to predicting the risk of CSED in GCLM patients (Supplementary Fig.\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eTo further explore the contribution of these clinical features to individual patient predictions, we randomly selected a patient from the validation cohort for demonstration purposes. Through interpretable algorithms, we visualized which specific features increased the prediction of CSED for this patient and which variables decreased the prediction. This patient was aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years, had Grade III-IV disease, had N stage N1 disease, did not undergo surgery, had an unknown or not undergone chemotherapy status, and had an unknown duration from diagnosis to therapy. The XGBoost model predicted a CSED risk of 81.4% for this GCLM patient based on clinical features. Among these features, age\u0026thinsp;\u0026ge;\u0026thinsp;65 years, Grade III-IV stage, no surgery, and chemotherapy status (No/Unknown) were factors contributing to an increased risk of CSED. Conversely, N stage N1 decreased the model's ability to predict CSED (Supplementary Fig.\u0026nbsp;2).\u003c/p\u003e \u003cp\u003eCombining these results, our study suggested that the XGBoost prediction model, constructed using the aforementioned factors, demonstrates greater accuracy and clinical applicability for predicting CSED in GCLM patients than the other six models.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eDespite a decrease in the incidence of GC in recent years, it continues to be one of the most prevalent cancer types globally and a leading cause of cancer-related mortality worldwide.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e) According to the SEER database, despite advancements in treatment, the overall early mortality among GC patients remains high at 32.57%, with a cancer-specific mortality of 30.73%(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Distant organ metastasis significantly impacts the prognosis of gastric cancer patients, with the liver being one of the most frequent sites of such metastasis. Patients with GCLM generally experience unfavorable outcomes, often encounter early death, and encounter significant controversy regarding comprehensive treatment approaches.(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) In recent years, the early mortality of GCLM patients has attracted widespread attention among medical researchers. However, reports on cancer-specific early mortality and associated factors in GCLM patients are scarce.\u003c/p\u003e \u003cp\u003eIn our study, we observed that 45% of GCLM patients experienced CSED. A prior study utilizing the SEER database reported an early mortality rate of 49.6% among GCLM patients(\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e), albeit reporting overall early mortality and not specifically cancer-specific early mortality, which is slightly higher than the results of our study.\u003c/p\u003e \u003cp\u003eFurthermore, we found that age\u0026thinsp;\u0026ge;\u0026thinsp;65 years, Grade III-IV and unknown, NX stage, non-surgery, on-chemotherapy or unknown, and unknown months from diagnosis to therapy increased the risk of CSED in GCLM patients. Feng et al(\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e) reported that age, primary site, histologic grade, tumor size, surgery, and distant metastasis were independent predictors of early death in patients with stage IV GC. Similarly, Zhu et al(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) reported that race, grade, surgery, chemotherapy, and bone, brain, liver, or lung metastasis were found to be independent risk factors for early death in metastatic GC patients. Both studies utilized traditional statistical methods to establish nomograms. In contrast to traditional models, ML algorithms offer advantages such as greater adaptability, the ability to handle larger data volumes, higher dimensionality, and more frequent information exchange. These benefits have made ML algorithms a popular tool among medical researchers for detecting and predicting cancer.(\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e) In our study, for the first time, we used 7 ML models trained with different feature sets to establish a predictive model for CSED in GCLM patients based on population data and evaluated the models' performance through 10-fold cross-validation, AUC, accuracy, balanced accuracy, precision, sensitivity, specificity, F1-score, calibration curves, and DCA. Additionally, we employed the DALEX package to compute feature importance and demonstrated the contribution of these features to the outcomes, ensuring the reliability of the results.\u003c/p\u003e \u003cp\u003eAccording to the feature ranking plots of the XGBoost model, chemotherapy was the most significant influencing factor. Despite previous research indicating higher toxicity and risk of complications associated with chemotherapy, we discovered that GCLM patients who underwent chemotherapy exhibited a markedly lower early mortality rate than those who did not (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Chemotherapy is a vital component of the treatment strategy for GCLM. Chemotherapy can kill or inhibit the progression of cancer cells, reduce tumor size, and even stabilize the disease, ultimately positively enhancing the overall survival of GC patients (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). In particular, combination chemotherapy offers significant advantages in treatment outcomes(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). Currently, chemotherapy is currently the primary treatment for advanced GC globally(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWith the aging of the global population, the incidence of GC among elderly patients is also increasing(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Statistics indicate that more than 60% of GC patients are aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years(\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). A study on elderly patients with locally advanced GC indicated that age is a standalone risk factor for cancer-specific survival(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). This may be attributed to the higher incidence of severe complications among elderly patients, particularly the postoperative pulmonary complications(\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e), as also demonstrated in studies by Takeuchi et al(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e) and Yang et al(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e). Additionally, research suggests that younger patients are more willing to receive alternative treatments and exhibit greater tolerance to the side effects and adverse reactions of surgery and adjuvant therapy(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). It has also been reported that elderly GC patients experience reduced physiological reserves and weaker tumor immune responses, ultimately leading to immune escape and tumor metastasis(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTumor differentiation is an important prognostic indicator that reflects the biological behavior of the tumor itself. In our study, patients with poorly differentiated tumors exhibited significantly higher early mortality (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). This factor was also incorporated into a nomogram for predicting the prognosis of metastatic GC patients who underwent palliative gastrectomy.(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e) In addition, tumor differentiation was factored into a nomogram developed by Gao et al.(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e) that predicts the prognosis of patients with stage III/IV GC. Poor tumor grading reflects high invasiveness and poor treatment response, which severely impact patient prognosis(\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). Furthermore, our study revealed that patients who postponed therapy for more than one month had a notably increased occurrence of early mortality compared to those who did not delay therapy (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Previous research has shown that treatment delay is associated with increased overall mortality(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). A systematic review and meta-analysis also demonstrated that a four-week delay in cancer therapy increased the risk of death from various cancer types (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHowever, our study has several limitations. First, this study is retrospective, potentially introducing selection bias. In the future, prospective clinical data are needed to establish more reliable evidence regarding the clinical applicability of our work. Second, research has demonstrated that the degrees of liver metastasis and peritoneal metastasis are separate criteria that might predict the prognosis of GCLM patients(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). As the SEER database does not include these data in its reports, it may affect the predictive ability of the model. Third, the SEER database contains only 30% of the entire US population, potentially limiting the extensiveness of the study sample.\u003c/p\u003e"},{"header":"5 Conclusion","content":"\u003cp\u003eIn this study, 45% of GCLM patients experienced CSED. Furthermore, age, grade, N stage, surgery, chemotherapy and months from diagnosis to therapy were identified as independent risk factors for CSED in GCLM patients. This study provides the first analysis of CSED prediction in a population of GCLM patients, which will offer doctors a reliable point of reference for effective screening for CSED in GCLM patients, thereby developing better treatment strategies and providing assistance for the clinical treatment of GCLM to improve patient prognosis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003cstrong\u003euthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the authors contributed to the study conception and design. The material preparation, data collection and analysis were performed by Y.Z., X.C., M.L., Y.L., Z.L.and P.Y.. The first draft of the manuscript was written by Y.Z., and all the authors commented on previous versions of the manuscript. All the authors have read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Sichuan Science and Technology Program, Sichuan, Chengdu (No. 2023YFS0066).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors thank the SEER database for the availability of the data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analyzed during the current study are available in the SEER database (https://seer.cancer.gov/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe received permission to access the research data in the SEER program from the National Cancer Institute, US. Approval was waived by the local ethics committee, as SEER data are publicly available and deidentified.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-249.\u003c/li\u003e\n\u003cli\u003eRawla P, Barsouk A. Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol. 2019;14(1):26-38.\u003c/li\u003e\n\u003cli\u003eSiegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin. 2019;69(1):7-34.\u003c/li\u003e\n\u003cli\u003eYoo CH, Noh SH, Shin DW, Choi SH, Min JS. Recurrence following curative resection for gastric carcinoma. Br J Surg. 2000;87(2):236-242.\u003c/li\u003e\n\u003cli\u003eZhu Y, Fang X, Wang L, Zhang T, Yu D. A Predictive Nomogram for Early Death of Metastatic Gastric Cancer: A Retrospective Study in the SEER Database and China. J Cancer. 2020;11(18):5527-5535.\u003c/li\u003e\n\u003cli\u003eShin A, Kim J, Park S. Gastric Cancer Epidemiology in Korea. J Gastric Cancer. 2011;11(3):135-140.\u003c/li\u003e\n\u003cli\u003eQiu MZ, Shi SM, Chen ZH, Yu HE, Sheng H, Jin Y, et al. Frequency and clinicopathological features of metastasis to liver, lung, bone, and brain from gastric cancer: A SEER-based study. Cancer Med. 2018;7(8):3662-3672.\u003c/li\u003e\n\u003cli\u003eAn W, Bao L, Wang C, Zheng M, Zhao Y. Analysis of Related Risk Factors and Prognostic Factors of Gastric Cancer with Liver Metastasis: A SEER and External Validation Based Study. Int J Gen Med. 2023;16:5969-5978.\u003c/li\u003e\n\u003cli\u003eYang Y, Chen ZJ, Yan S. The incidence, risk factors and predictive nomograms for early death among patients with stage IV gastric cancer: a population-based study. J Gastrointest Oncol. 2020;11(5):964-982.\u003c/li\u003e\n\u003cli\u003eSiegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA: a cancer journal for clinicians. 2022;72(1).\u003c/li\u003e\n\u003cli\u003eThrift AP, El-Serag HB. Burden of Gastric Cancer. Clin Gastroenterol Hepatol. 2020;18(3):534-542.\u003c/li\u003e\n\u003cli\u003eMinciuna CE, Tudor S, Micu A, Diaconescu A, Alexandrescu ST, Vasilescu C. Safety and Efficacy of Simultaneous Resection of Gastric Carcinoma and Synchronous Liver Metastasis-A Western Center Experience. Medicina (Kaunas). 2022;58(12).\u003c/li\u003e\n\u003cli\u003eKawahara K, Makino H, Kametaka H, Hoshino I, Fukada T, Seike K, et al. Outcomes of surgical resection for gastric cancer liver metastases: a retrospective analysis. World J Surg Oncol. 2020;18(1):41.\u003c/li\u003e\n\u003cli\u003eCheng J, Cai M, Shuai X, Gao J, Wang G, Tao K. Multimodal treatments for resectable gastric cancer: A systematic review and network meta-analysis. Eur J Surg Oncol. 2019;45(10):1796-1805.\u003c/li\u003e\n\u003cli\u003eFeng Y, Guo K, Jin H, Xiang Y, Zhang Y, Ruan S. A Predictive Nomogram for Early Mortality in Stage IV Gastric Cancer. Med Sci Monit. 2020;26:e923931.\u003c/li\u003e\n\u003cli\u003eNgiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery. Lancet Oncol. 2019;20(5):e262-e273.\u003c/li\u003e\n\u003cli\u003eLi S, Yi H, Leng Q, Wu Y, Mao Y. New perspectives on cancer clinical research in the era of big data and machine learning. Surg Oncol. 2024;52:102009.\u003c/li\u003e\n\u003cli\u003eHironaka S, Sugimoto N, Yamaguchi K, Moriwaki T, Komatsu Y, Nishina T, et al. S-1 plus leucovorin versus S-1 plus leucovorin and oxaliplatin versus S-1 plus cisplatin in patients with advanced gastric cancer: a randomised, multicentre, open-label, phase 2 trial. Lancet Oncol. 2016;17(1):99-108.\u003c/li\u003e\n\u003cli\u003ePernot S, Mitry E, Samalin E, Dahan L, Dalban C, Ychou M, et al. Biweekly docetaxel, fluorouracil, leucovorin, oxaliplatin (TEF) as first-line treatment for advanced gastric cancer and adenocarcinoma of the gastroesophageal junction: safety and efficacy in a multicenter cohort. Gastric Cancer. 2014;17(2):341-347.\u003c/li\u003e\n\u003cli\u003eNishikawa K, Kawakami H, Shimokawa T, Fujitani K, Tamura S, Endo S, et al. Meta-analysis of three randomized trials of capecitabine plus cisplatin (XP) versus S-1 plus cisplatin (SP) as first-line treatment for advanced gastric cancer. Int J Clin Oncol. 2023;28(11):1501-1510.\u003c/li\u003e\n\u003cli\u003eMa X, Zhang Y, Wang C, Yu J. Efficacy and safety of combination chemotherapy regimens containing taxanes for first-line treatment in advanced gastric cancer. Clin Exp Med. 2023;23(2):381-396.\u003c/li\u003e\n\u003cli\u003eHsieh M-c, Wang S-H, Rau K-M. Real world analysis of adjuvant chemotherapy for advanced gastric cancer after D2 radical surgery. Annals of Oncology. 2017;28:iii36-iii37.\u003c/li\u003e\n\u003cli\u003eFujiwara Y, Fukuda S, Tsujie M, Ishikawa H, Kitani K, Inoue K, et al. Effects of age on survival and morbidity in gastric cancer patients undergoing gastrectomy. World J Gastrointest Oncol. 2017;9(6):257-262.\u003c/li\u003e\n\u003cli\u003eZheng Y, Wu C. [Prevalence and trend of gastrointestinal malignant tumors in the elderly over 75 years old in China]. Zhonghua Wei Chang Wai Ke Za Zhi. 2016;19(5):481-485.\u003c/li\u003e\n\u003cli\u003eSun Y, Li Z, Tian Y, Gao C, Liang B, Cao S, et al. Development and validation of nomograms for predicting overall survival and cancer-specific survival in elderly patients with locally advanced gastric cancer: a population-based study. BMC Gastroenterol. 2023;23(1):117.\u003c/li\u003e\n\u003cli\u003eWong JU, Tai FC, Huang CC. An examination of surgical and survival outcomes in the elderly (65-79 years of age) and the very elderly (\u0026ge;80 years of age) who received surgery for gastric cancer. Curr Med Res Opin. 2020;36(2):229-233.\u003c/li\u003e\n\u003cli\u003eTakeuchi D, Koide N, Suzuki A, Ishizone S, Shimizu F, Tsuchiya T, et al. Postoperative complications in elderly patients with gastric cancer. J Surg Res. 2015;198(2):317-326.\u003c/li\u003e\n\u003cli\u003eYang JY, Lee HJ, Kim TH, Huh YJ, Son YG, Park JH, et al. Short- and Long-Term Outcomes After Gastrectomy in Elderly Gastric Cancer Patients. Ann Surg Oncol. 2017;24(2):469-477.\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Connell JB, Maggard MA, Livingston EH, Yo CK. Colorectal cancer in the young. Am J Surg. 2004;187(3):343-348.\u003c/li\u003e\n\u003cli\u003eSchlesinger-Raab A, Mihaljevic AL, Egert S, Emeny R, Jauch KW, Kleeff J, et al. Outcome of gastric cancer in the elderly: a population-based evaluation of the Munich Cancer Registry. Gastric Cancer. 2016;19(3):713-722.\u003c/li\u003e\n\u003cli\u003eMa T, Wu ZJ, Xu H, Wu CH, Xu J, Peng WR, et al. Nomograms for predicting survival in patients with metastatic gastric adenocarcinoma who undergo palliative gastrectomy. BMC Cancer. 2019;19(1):852.\u003c/li\u003e\n\u003cli\u003eGao Z, Ni J, Ding H, Yan C, Ren C, Li G, et al. A nomogram for prediction of stage III/IV gastric cancer outcome after surgery: A multicenter population-based study. Cancer Med. 2020;9(15):5490-5499.\u003c/li\u003e\n\u003cli\u003eJohncilla M, Chen Z, Sweeney J, Yantiss RK. Tumor Grade Is Prognostically Relevant Among Mismatch Repair Deficient Colorectal Carcinomas. Am J Surg Pathol. 2018;42(12):1686-1692.\u003c/li\u003e\n\u003cli\u003eCone EB, Marchese M, Paciotti M, Nguyen D-D, Nabi J, Cole AP, et al. Assessment of Time-to-Treatment Initiation and Survival in a Cohort of Patients With Common Cancers. JAMA Network Open. 2020;3(12):e2030072-e2030072.\u003c/li\u003e\n\u003cli\u003eHanna TP, King WD, Thibodeau S, Jalink M, Paulin GA, Harvey-Jones E, et al. Mortality due to cancer treatment delay: systematic review and meta-analysis. BMJ. 2020;371:m4087.\u003c/li\u003e\n\u003cli\u003eUeda K, Iwahashi M, Nakamori M, Nakamura M, Naka T, Ishida K, et al. Analysis of the prognostic factors and evaluation of surgical treatment for synchronous liver metastases from gastric cancer. Langenbecks Arch Surg. 2009;394(4):647-653.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Gastric cancer with liver metastasis (GCLM), SEER, cancer-specific early death (CSED), machine learning (ML)","lastPublishedDoi":"10.21203/rs.3.rs-4485633/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4485633/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eGastric cancer with liver metastasis (GCLM) patients typically have a grim prognosis and are at high risk of early mortality. This study aimed to predict cancer-specific early mortality and risk factors for GCLM patients through machine learning (ML) methods.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThe data of patients with GCLM were obtained from the SEER database. LASSO regression, univariate and multivariate logistic regression analyses were employed to identify significant independent risk factors for cancer-specific early death (CSED). Models such as logistic regression (LR), decision tree (DT), K-nearest neighbors (KNN), light gradient boosting machine (LightGBM), random forest (RF), support vector machine (SVM), and extreme gradient boosting (XGBoost) were used to predict the CSED and extract important features. Tenfold cross-validation, receiver operating characteristic (ROC) curve analysis, accuracy, balance accuracy, precision, sensitivity, specificity, F1-score, precision‒recall (PR) curve analysis, calibration curve analysis and decision curve analysis (DCA) were utilized to assess the performance of the models. The DALEX package was used to compute feature importance.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe study recruited a total of 3661 patients. A total of 1648 (45%) patients experienced CSED. Among the 7 ML models, the XGBoost model achieved the best performance. The top 6 most influential factors were chemotherapy, months from diagnosis to therapy, age, grade, N stage, and surgery in the XGBoost model, with chemotherapy being the most significant.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe XGBoost model might be applied to predict the CSED of GCLM patients, and chemotherapy was the most important feature in the XGBoost model. These results could offer crucial reference data to assist clinicians in making informed decisions beforehand.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Cancer-Specific Early Death Prediction Model for Patients with Gastric Cancer with Liver Metastasis: Based on Machine Learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-13 02:39:21","doi":"10.21203/rs.3.rs-4485633/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":"5172dceb-66ff-45bd-a1e0-25519b1f8581","owner":[],"postedDate":"June 13th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":33052147,"name":"Health sciences/Oncology"},{"id":33052148,"name":"Health sciences/Oncology/Surgical oncology"}],"tags":[],"updatedAt":"2024-07-26T11:31:14+00:00","versionOfRecord":[],"versionCreatedAt":"2024-06-13 02:39:21","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4485633","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4485633","identity":"rs-4485633","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}