Development and internal validation of a risk scoring system for gastrointestinal events requiring surgery in gastrointestinal lymphoma patients.

ABSTRACT

BACKGROUND AND AIM: The predictors of severe gastrointestinal (GI) events in GI lymphoma patients are unclear. We aimed to develop a risk scoring system for GI events requiring surgery. METHODS: In this retrospective study of 192 patients with GI lymphoma, the state of lymphoma, macroscopic findings, examination results, and International Prognostic Index were assessed. We developed a risk score for GI events that required surgery and assessed its accuracy by calculating the area under the receiver operating characteristic curve (AUC). Internal validation was performed using bootstrap resampling. RESULTS: Severe GI events occurred in 21 (11%) patients. We developed a 4-point scoring system (the FLASH score) comprising the following three independent predictors (weighted by regression coefficients) (i) focal appearance and large size (≥ 40 mm), 1 point; (ii) aggressive lymphoma of the small bowel, 2 points; and (iii) high (18)F-fluorodeoxyglucose positron emission tomography uptake, 1 point. The score predicted severe GI events with an AUC value of 0.91 (internal validation; AUC, 0.86). Risk was classified into three categories the GI event rate was 0% in the low-risk group (0 points), 9% in the intermediate-risk group (1-2 points), and 61% in the high-risk group (3-4 points) (AUC, 0.89). CONCLUSIONS: We developed and internally validated a risk scoring system (the FLASH score) that included macroscopic findings to predict severe GI events in GI lymphoma patients. Patients with high scores are candidates for elective surgery to prevent GI events.