Functional MRI Patterns Predict 1-Year Bariatric Surgery Outcomes: A Pilot Study

• Published in: Obesity (Silver Spring, Md.) Md.), 2021-12, Vol.29, p.106-106
• Main Authors: Koenis, Marinka, Patel, Krishna, Papasavas, Pavlos, Stevens, Michael, Tishler, Darren, Stone, Andrea, Kober, Hedy, Pearlson, Godfrey
• Format: Article
• Language: English
• Subjects: Gastrointestinal surgery; Metabolism; Surgical outcomes
• ISSN: 1930-7381; 1930-739X

Background: Post-bariatric surgery, 15-30% of the patients fail to lose a substantial portion of their weight, or regain weight. Currently, research has failed to identify metabolic, social, hormonal, psychological, or eating-related predictors of bariatric surgical success. Neuroimaging, however, may provide predictive information. Methods: Participants with BMI > 35 were recruited from the Metabolic and Bariatric surgery center at Hartford Hospital. They were scanned with five food- and reward-related fMRI tasks after an >8-hour fast. Weight, metabolic, psychological, and eating pattern measures were acquired pre-surgery and 1 year post-surgery. N = 55 patients with follow-up weight-loss measurements were included in the analysis (mean (sd) BMI = 45.9 (8.2) kg/m2; age = 41.1 (11.1) years; 78% female). A regularized regression method (elastic net) and repeated cross validation were used to select variables that were associated with surgery outcome as measured by percentage total weight loss (%TWL) and a composite score derived from (mental) health-related indices. Results: Pre-surgical BMI was significantly reduced at 1-year postsurgery to 33.4 (7.0) kg/m2. Nine variables explained 36% of the variance in %TWL at 1 year post-surgery: baseline BMI, race, anterior cingulate (ACC), orbitofrontal cortex, and hippocampal activation during different tasks. Similarly, eight variables explained 37% of the variance in the health composite score: race, surgery type, hippocampal, insular, ACC, and inferior temporal gyrus activation during different tasks. Psychological and metabolic variables were not predictive of surgery outcome. Conclusions: Pre-surgical brain responses during food- and rewardrelated fMRI tasks were able to predict 1-year post-surgery weight loss in participants with morbid obesity. This has important implications for patients and can improve targeted patient outcome. Further investigation in a new sample is needed to replicate these findings and assess their translational significance.