Effect of tumor location on the efficacy of chemotherapy for gastrointestinal stromal tumors.

Abstract

790 Background: Gastrointestinal stromal tumors (GISTs) arise from multiple sites within the gastrointestinal tract. Since the introduction of imatinib in 2002, patients have seen improved overall survival. However, a recent National Cancer Database (NCDB) study suggested no survival benefit to adjuvant imatinib for colorectal GIST patients. Another recent study showed that small bowel GISTs have larger median tumor size and higher incidence of tumors with at least 5 mitoses compared to gastric GISTs. These findings suggest a differential tumor biology for GISTs arising from different primary tumor sites, potentially conferring differential response to imatinib. In this NCDB study, we compare overall survival outcomes with and without chemotherapy, stratifying by primary tumor site. Methods: We queried the NCDB from 2004-2018 for patients with GIST histology. We excluded patients with age <40, age>80, high comorbidity score, and patients who received additional therapy other than surgery or chemotherapy. Of the 22,846 patients meeting criteria for our study, 27 patients had tumors originating from the anus, 454 from the colon, 173 from the esophagus, 15,296 from the stomach, 556 from the rectum, 32 from the rectosigmoid junction, and 6,308 from the small intestine. 8,983 (39%) received chemotherapy. Treatment with chemotherapy was used to extrapolate response to imatinib, the mainstay of treatment. Results: Cox proportional hazards analysis demonstrated increased hazard of death for chemotherapy, increasing age, and comorbidity score of 1. Positive predictive factors for survival included female sex and later year of diagnosis. For anal, esophageal, gastric, and rectosigmoid GIST patients, there was no statistically significant overall survival benefit for any treatment modality. For colonic GISTs, neoadjuvant chemotherapy combined with surgery improved overall survival. For rectal GISTs, adjuvant chemotherapy with surgery also provided survival benefit. For small intestinal GISTs, surgery alone or surgery with neoadjuvant, adjuvant, or both neoadjuvant and adjuvant chemotherapy improved overall survival. Conclusions: While our study is retrospective and limited as a NCDB study, our Kaplan-Meier survival analysis demonstrated survival benefit for chemotherapy only in patients with colon, rectal, or small intestinal GIST tumors. Our Cox proportional analysis demonstrated an overall negative survival impact for chemotherapy. These findings lend further support to the idea that differential tumor biology at each primary site may lead to differential response to imatinib. Overall, our study provides rationale for further investigating the heterogeneity of GIST tumor microenvironments, as these differences could contribute to the differential treatment outcomes.