The DIET study: A randomized controlled trial of a high fiber diet intervention (HFDI) in patients (pts) with melanoma receiving immune checkpoint blockade (ICB).

9558 Background: Multiple studies support that the gut microbiome plays a key role in response to ICB. Habitual high dietary fiber intake is associated with improved response to ICB and high fiber diet interventions have been shown to favorably modulate the microbiome in non-cancer populations. We therefore conducted a randomized trial of HFDI versus healthy control diet in melanoma pts receiving ICB. Methods: 45 melanoma pts starting ICB (21 adjuvant, 12 neoadjuvant, 12 advanced unresectable) were enrolled. Pts were randomized 2:1 to HFDI (30 to 50 g/day by ramp-up) or control (20 g) stratified by BMI and treatment (tx) intent. Inclusion criteria included BMI 18.5-40 kg/m2 and willing to exclusively eat the provided diet. Exclusion criteria include diabetes, major gastrointestinal disease/surgery, probiotic, antibiotic, or steroid use within 14 days, smoker or heavy drinker, and average dietary fiber intake > 20g/day. As a controlled feeding study, the weekly menu in both arms were isocaloric (matched to patient’s energy needs via 30% fat, 50% carbohydrates, 20% protein) and provided by MD Anderson Bionutrition Research Core for up to 11 weeks. Both diets followed cancer prevention recommendations with little to no processed meats or added sugars; no alcohol consumption; and dietary fiber intake scaled through whole, plant foods. Blood and stool samples were collected longitudinally. Compliance was defined by the proportion of calories consumed vs. provided. Results: 65 pts were consented. 20 pts (31%) screen failed largely due to current fiber intake >20 g/day. 45 pts were randomized, and 44 pts have completed the study to date. ICB tx were PD1 monotherapy (n=22), ipilimumab + nivolumab (n=16), and nivolumab + relatlimab (n=7). 53% of pts were female and average BMI was 29.7 kg/m 2 . Average baseline fiber intake was 15.5 g/day. Compliance with the diet while on study was 83% (95% CI 79%-87%); however, 15 pts (33%) withdrew consent early due to challenges with adherence to the study. The rate of diet-related AEs was 45%, all Grade I/II including abdominal pain (11%), anorexia (5%), bloating (6%), constipation (18%), diarrhea (23%), flatulence (18%), hyperglycemia (2%), nausea (2%) and weight loss (9%). The rate of Grade III/IV immune related adverse events was 25%. Conclusions: In this fully-controlled feeding study among melanoma pts undergoing ICB tx, both arms of the intervention were well-tolerated. In contrast to our prior studies in cancer survivor populations, we noted a significant discontinuation rate, perhaps due to the challenges of participating in a fully-controlled feeding study while undergoing active cancer therapy. Ongoing analyses are interrogating the effects of the dietary intervention on the structure and function of the gut microbiome, circulating and stool metabolites, and systemic and anti-tumor immunity. Clinical trial information: NCT04645680 .