Distinguishing very high-risk patients among high-risk gastrointestinal stromal tumor cases: development and validation of a nomogram based on a multicenter population-based retrospective cohort study

Abstract

Background: High-risk gastrointestinal stromal tumors (GISTs) are highly heterogeneous. This study aimed to developa nomogram for predicting recurrence in patients with high-risk GISTand to provide guidance for adjuvant therapy.

Methods: Data were retrospectively collected from 971 patients with high-risk GIST who underwent genetic testingat 13 centers in China. A nomogram was constructed in the training cohort and validated in the validation cohort.

Rseults: The training and validation cohorts included 696 and 275 patients, respectively. The nomogram incorporated blood plateletlevels, total protein, tumor location, tumor size, mitotic count, tumor rupture, Ki-67 index, and gene mutations. The C-index and AUC were 0.758 and 0.781 in the training cohort and 0.841 and 0.755, respectively, in the validation cohort. Calibration and DCA curves confirmed favorable discrimination, calibration accuracy, and clinical benefits. Using the nomogram, patients were categorized into a general high-risk groupand a very high-risk group. Among the general high-risk group, no significant difference in RFS was observed between patients who received adjuvant imatinib for 2.5 years or longer. Conversely, in the very high-risk group, patients receiving adjuvant imatinib for five or more years had markedly improved RFS.

Conclusions: Based on the largest nomogram-related multicenter study in high-risk GIST, the nomogram accurately predicted RFS in patients with high-risk GIST and provided guidance on adjuvant therapy for the first time. For general high-risk patients, 3 years of adjuvant imatinib is adequate, whereas very high-risk patients benefit significantly from more than 5 years of adjuvant imatinib.

Keywords: Adjuvant therapy; gastrointestinal stromal tumors; gene mutations; nomogram; very high-risk.