Tao Chen1, Lili Xu2, Liangying Ye2, Haibo Qiu3, Yanfeng Hu2, Hao Liu2, Zhiwei Zhou3, Guoxin Li2, Jiang Yu4. 1. Department of General Surgery, Nanfang Hospital, Southern Medical University, Guangdong Provincial Engineering Technology Research Center of Minimally Invasive Surgery, Guangzhou, 510515, Guangdong Province, China. Electronic address: drchentao@163.com. 2. Department of General Surgery, Nanfang Hospital, Southern Medical University, Guangdong Provincial Engineering Technology Research Center of Minimally Invasive Surgery, Guangzhou, 510515, Guangdong Province, China. 3. Department of Gastric Surgery, Sun Yat-sen University Cancer Center, Sun Yat-sen University, Guangzhou, 510060, Guangdong Province, China. 4. Department of General Surgery, Nanfang Hospital, Southern Medical University, Guangdong Provincial Engineering Technology Research Center of Minimally Invasive Surgery, Guangzhou, 510515, Guangdong Province, China. Electronic address: balbc@163.com.
Abstract
BACKGROUND: This study aimed to build a new risk stratification nomogram for gastrointestinal stromal tumors (GISTs) focused on a popular factor Ki-67 to enable individualized and precise predictions of the most suitable candidates for imatinib therapy. METHODS: We retrospectively collected clinicopathologic data of the patients diagnosed with GISTs from January 1998 to December 2015 at Southern Medical University Nanfang Hospital as the experiment group. And patients with GISTs at the Sun Yat-sen University Cancer Center from January 2007 to December 2012 were included as the validation group. The nomogram was built using Kaplan-Meier method and the Cox proportional hazards regression model. The receiver operating characteristic (ROC) curves were established to compare the discriminative ability of the new nomogram with other risk stratification systems, including the modified National Institute of Health (modified NIH) criteria, Armed Forces Institute of Pathology (AFIP) criteria, Memorial Sloan Kettering Cancer Center (MSKCC) prognostic nomogram, and contour maps. RESULTS: In univariate analysis, the tumor size, site, mitotic count, tumor rupture and Ki-67 labeling index were significant factors (all P < 0.05) and included in the Cox model to build our nomogram. According to the ROC curve, our new nomogram showed the largest AUC value (0.778) compared with that of the other classification methods (contour maps, AUC = 0.743; AFIP, AUC = 0.719; MSKCC, AUC = 0.712; and modified NIH, AUC = 0.719). CONCLUSION: Our new nomogram exhibits an excellent performance and might become a potential risk stratification to support therapeutic decision-making for GISTs.
BACKGROUND: This study aimed to build a new risk stratification nomogram for gastrointestinal stromal tumors (GISTs) focused on a popular factor Ki-67 to enable individualized and precise predictions of the most suitable candidates for imatinib therapy. METHODS: We retrospectively collected clinicopathologic data of the patients diagnosed with GISTs from January 1998 to December 2015 at Southern Medical University Nanfang Hospital as the experiment group. And patients with GISTs at the Sun Yat-sen University Cancer Center from January 2007 to December 2012 were included as the validation group. The nomogram was built using Kaplan-Meier method and the Cox proportional hazards regression model. The receiver operating characteristic (ROC) curves were established to compare the discriminative ability of the new nomogram with other risk stratification systems, including the modified National Institute of Health (modified NIH) criteria, Armed Forces Institute of Pathology (AFIP) criteria, Memorial Sloan Kettering Cancer Center (MSKCC) prognostic nomogram, and contour maps. RESULTS: In univariate analysis, the tumor size, site, mitotic count, tumor rupture and Ki-67 labeling index were significant factors (all P < 0.05) and included in the Cox model to build our nomogram. According to the ROC curve, our new nomogram showed the largest AUC value (0.778) compared with that of the other classification methods (contour maps, AUC = 0.743; AFIP, AUC = 0.719; MSKCC, AUC = 0.712; and modified NIH, AUC = 0.719). CONCLUSION: Our new nomogram exhibits an excellent performance and might become a potential risk stratification to support therapeutic decision-making for GISTs.