Predicting long-term outcome in anorexia nervosa: a machine learning analysis of brain structure at different stages of weight recovery

Abstract

Background Anorexia nervosa (AN) is characterized by sizable, widespread gray matter (GM) reductions in the acutely underweight state. However, evidence for persistent alterations after weight-restoration has been surprisingly scarce despite high relapse rates, frequent transitions to other psychiatric disorders, and generally unfavorable outcome. While most studies investigated brain regions separately (univariate analysis), psychiatric disorders can be conceptualized as brain network disorders characterized by multivariate alterations with only subtle local effects. We tested for persistent multivariate structural brain alterations in weight-restored individuals with a history of AN, investigated their putative biological substrate and relation with 1-year treatment outcome. Methods We trained machine learning models on regional GM measures to classify healthy controls (HC) (N = 289) from individuals at three stages of AN: underweight patients starting intensive treatment (N = 165, used as baseline), patients after partial weight-restoration (N = 115), and former patients after stable and full weight-restoration (N = 89). Alterations after weight-restoration were related to treatment outcome and characterized both anatomically and functionally. Results Patients could be classified from HC when underweight (ROC-AUC = 0.90) but also after partial weight-restoration (ROC-AUC = 0.64). Alterations after partial weight-restoration were more pronounced in patients with worse outcome and were not detected in long-term weight-recovered individuals, i.e. those with favorable outcome. These alterations were more pronounced in regions with greater functional connectivity, not merely explained by body mass index, and even increases in cortical thickness were observed (insula, lateral orbitofrontal, temporal pole). Conclusions Analyzing persistent multivariate brain structural alterations after weight-restoration might help to develop personalized interventions after discharge from inpatient treatment.

Lower region‐specific gray matter volume in females with atypical anorexia nervosa and anorexia nervosa

“…This may also explain our divergent findings for the insula as well. Several previous studies observed significant alterations in gray matter in the insula in individuals with AN, which we were likely not powered to observe (Arold et al, 2023;Frank et al, 2013;Lavagnino et al, 2018).…”

“…While the etiology of atypical AN and AN is still unclear, studies utilizing neuroimaging approaches have shown that the severe caloric restriction and concomitant reduction in body weight have a negative impact (i.e., reduction) on cortical gray matter (Arold et al, 2023;Bahnsen et al, 2022;King et al, 2018). The recent meta-analysis conducted by Walton and colleagues from the Enhancing Neuro Imaging Genetics through Meta Analysis (ENIGMA) Eating Disorders Working Group (2023) provides the most conclusive evidence of the negative impact of the illness on cortical gray matter thus far (Walton et al, 2022).…”

Examining the role of artificial intelligence to advance knowledge and address barriers to research in eating disorders

“…In another study published in 2024, researchers were able to successfully distinguish between AN and BN utilizing ML models based on diffusion tensor images (DTI), with the left middle temporal gyrus (MTG_L) and the left superior temporal gyrus (STG_L) suggested as potentially useful neuroimaging biomarkers (Zheng et al, 2024). In a further ML study that sought to predict longerterm outcomes in patients with AN using analysis of regional grey matter (GM) structure at different stages of weight recovery, scientists showed that underweight and partially weight restored (WR) patients could be differentiated from healthy controls (Arold et al, 2023). Of interest, the ML-based risk score was a predictor of future outcomes, and partially WR patients who had poor outcomes showed alterations in regions of the brain with greater functional connectivity, which could not be fully explained by body mass index.…”