Risk factors associated with de novo metastatic colorectal cancer in early onset colorectal cancer.

Abstract

3529 Background: Early onset colorectal cancer (EO CRC) is rising worldwide. Epidemiologic studies have shown that diets high in sugar and/or processed foods and limited exercise are associated with EO CRC when compared to healthy controls. However, risk factors associated with de novo metastatic disease are less well established. Identifying such risk factors may provide insight into modifiable lifestyle variables. Methods: All eligible EO CRC patients were enrolled in MSK’s Center for Young Onset Colorectal and Gastrointestinal Cancer and completed a risk factor questionnaire (MSK-approved IRB #20-315). We analyzed questionnaire responses and compared risk factors between patients with de novo metastatic disease vs localized disease using Wilcoxon rank sum test or Fisher’s exact test. Clinical outcomes, including overall survival (OS) from diagnosis and progression free survival (PFS) from 1 st -line chemotherapy, were estimated using Kaplan-Meier methods. The Cox regression model was used to assess association between risk factors and survival outcomes. Tumors from a subset of patients (n = 206) were sequenced using MSK-IMPACT (MSK-approved IRB #12-245) and underwent genomic analyses. Results: 303 patients completed the questionnaire (median age at diagnosis 42; 51% Female; 88% left-sided tumors). 112 had de novo stage IV disease and 191 had stage I-III disease. Patients with de novo metastatic disease were younger, 40.9 [95%CI: 36.8 - 44.8] vs. 43.0 [95%CI: 38.4 - 46.4] ( P -value 0.037) . Analysis of dietary factors showed no association with fruit, vegetable, fish, poultry, red meat, processed meat, or dairy intake. However, high sugar diets were significantly associated with de novo metastatic disease, with 30 (45%) vs. 37 (29%) ( P -value, 0.004) patients reporting daily consumption of high sugar foods. Daily consumption of high calorie foods was also frequently reported in patients with metastatic disease ( P- value, 0.057). 3-year OS in the metastatic population was 72% [95%CI: 62% - 84%] vs. 99% [95%CI: 98% - 100%]. Within the metastatic group, no association was observed between daily high sugar consumption and non-daily consumption in terms of PFS or OS. 3-year OS in the daily high sugar group was 79 % [95%CI: 62%-100%] vs 74% [95%CI: 57%-95%]. For early stage patients who later progressed (n = 27), median time to progression was not significantly shorter among patients who reported daily high sugar consumption (18 vs. 19 months, P -value 0.5). Genomic analyses revealed no significant differences in tumor mutational burden, fraction genome altered, frequency of oncogenic or signaling pathway alterations in de novo metastatic vs. non-metastatic patients. Conclusions: In a single center study, in EO CRC patients, high sugar diets may be associated with de novo metastatic disease. There were no significant genomic differences detected in patients with de novo metastatic vs. early stage disease.