From pixels to prediction: CT morphometrics and artificial intelligence in complex ventral hernia repair

Complex ventral hernia repair requires more than conventional assessment of defect width and location. Computed tomography (CT) provides quantitative information on hernia morphology, loss of domain, abdominal wall musculature, adipose distribution, and body composition that can improve preoperative risk stratification and surgical planning. Concurrently, artificial intelligence (AI) offers opportunities to automate image analysis, standardize morphometric assessment, and develop predictive models for individualized decision-making. This review critically evaluates the current evidence regarding CT-derived morphometrics and AI applications in the assessment and management of complex ventral and incisional hernias.

A structured literature review was conducted using PubMed/MEDLINE, Scopus, and Web of Science from database inception to the final search date before manuscript submission. Studies evaluating CT-derived morphometric parameters, loss-of-domain assessment, body-composition analysis, radiomics, machine learning, and AI applications in ventral and incisional hernia surgery were included. Eligible studies underwent methodological quality assessment using validated tools appropriate to study design, and findings were synthesized narratively because of substantial heterogeneity in imaging protocols, morphometric definitions, AI methodologies, and reported outcomes.

CT-derived morphometrics provide objective characterization of hernia anatomy beyond conventional defect dimensions. Volumetric assessment, standardized loss-of-domain measurements, abdominal wall muscle morphology, sarcopenia, myosteatosis, and adipose distribution have demonstrated associations with fascial closure, requirement for component separation, surgical-site occurrences, and postoperative outcomes. Emerging AI applications-including automated image segmentation, radiomics, and machine-learning prediction models-have shown promising performance for quantitative image analysis and individualized risk prediction. However, the current evidence remains limited by predominantly retrospective single-centre studies, heterogeneous methodologies, inconsistent imaging definitions, small datasets, and limited external validation, precluding widespread clinical implementation.

Quantitative CT morphometrics combined with AI has the potential to transform complex ventral hernia assessment from descriptive imaging to evidence-based, patient-specific surgical planning. Future progress will depend on standardized CT acquisition and reporting, consensus morphometric definitions, multicentre collaborative datasets, rigorous external validation of predictive models, and prospective evaluation of their clinical utility before routine implementation in abdominal wall reconstruction.