Objective To assess the efficacy and prognosis of gastrointestinal stromal tumors (GIST) after preoperative targeted therapy using the Choi criteria compared to pathological effects, and to observe the consistency between them.

Methods The clinicopathological data of 37 patients, who underwent preoperative treatment with targeted imatinib therapy for GIST, were retrospectively analyzed. Survival analysis of the Choi criteria and pathological effects was conducted using the Kaplan-Meier method and Log-rank test. The consistency between the Choi criteria and the pathological effects was assessed using Spearman's correlation and Kappa tests.

Results The median preoperative treatment duration for the 37 patients was 10 months (range, 2-36 months). According to the Choi criteria, there were no cases of complete response (CR), 26 cases of partial response (PR), five cases of stable disease (SD), and six progressive disease (PD) cases. The difference in overall survival (OS) between the effective group (CR+PR) and the ineffective group (SD+PD) was statistically significant (P<0.01). Pathological effects were evaluated as one complete effect, 11 high effects, 18 partial effects, and seven zero effect cases. The OS significantly differed between the effective (full effect+high effect+partial effect) and ineffective (zero effect) groups was statistically significant (P<0.01). The Choi showed moderate consistency with the pathological effects (r=0.592, P<0.01) with a (κappa=0.566).

Conclusions The Choi criteria were moderately correlated and consistent with the pathological effects. Both can be used to evaluate the efficacy and prognosis of preoperative targeted therapy for GIST. The combined use of these two criteria has better clinical application value than that of either alone.