Clinicopathological and Prognostic Analysis of Primary Gastrointestinal Stromal Tumor Presenting with Gastrointestinal Bleeding: a 10-Year Retrospective Study

The objectives of this paper were to investigate the clinicopathological characteristics and prognostic factors of GI-bleeding GIST patients and explore whether GI bleeding is a risk factor for GIST relapse. Primary GIST patients with initial symptoms of GI bleeding or no GI bleeding were retrospectively studied. Up to 178 GI-bleeding GIST patients including 108 (60.7%) males and 70 (39.3%) females were evaluated for the clinicopathological characteristics. The stomach, small bowel, and colorectum were the tumor sites in 82 (46.1%), 85 (47.8%), and 11 (6.2%) patients. Of the 178 patients, 163 GI-bleeding patients had follow-up while another 363 patients from the total population presented without GI bleeding were followed up. Up to 526 patients who received postoperative follow-up were included in the survival analysis. Compared with the 363 non-GI-bleeding patients, GI-bleeding patients developed smaller tumors (P = 0.015) and had a longer relapse-free survival (RFS; P = 0.014). For the 163 GI-bleeding patients, a Cox regression analysis showed that the mitotic count and the platelet-lymphocyte ratio before surgery were independent prognostic predictors for poor outcome regarding RFS. For all 526 patients, a Cox regression analysis indicated that tumor location, mitotic index, platelet-lymphocyte ratio, and GI bleeding were independent prognosis predictors. Compared to non-GI-bleeding GIST patients, patients with GI bleeding were more likely to be male and to have more small intestine GISTs, smaller tumors, and a longer RFS. For GI-bleeding patients, mitotic count and platelet-lymphocyte ratio were independent prognostic indicators. GI bleeding served as a surrogate for smaller GIST and was a protective factor for GIST recurrence.