ASA Main Site  |  Past & Future Meetings
American Surgical Association

Back to 2022 Abstracts


Preoperative Computed Tomography Morphological Features Associated with Hernia Formation
John P. Fischer1, Phoebe McAuliffe1, Abhishek Desai1, Ankoor Talwar1, Robyn Broach1, Joseph Serletti1, Yubing Tong2, Jayaram Udupa2, Drew Torigian2
1Surgery, University of Pennsylvania, Philadelphia, Pennsylvania, United States, 2Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Objectives: Incisional hernia (IH) is a pervasive surgical complication impacting all surgical disciplines operating through the abdomen and affecting 1 in 8 patients undergoing surgery. Despite the significant costs and disability of IH, an incomplete understanding of surgical risk has impeded progress in the fight against hernia. We aim to identify key morphometric features associated with hernia formation using routine pre-operative computed tomography (CT) imaging.

Methods: A matched-pair, case-control study was performed. A cohort of patients (n=14,345) who underwent colorectal surgery between 2005-2017 was identified; outcome of IH repair, demographics, and clinical data were collected. All index patients underwent open, elective surgery. Patients who subsequently developed IH were paired and matched with those who did not. Optimal paired matching based on Euclidean distance was performed across age, gender, body mass index (BMI), seven diagnoses/procedures at index procedure and two prior to index. Pairs were exactly matched on race, gender and nine diagnoses/procedures, and were matched with calipers of 4 and 2 for age and BMI. Preoperative abdominopelvic CT scans were blinded and manually segmented to obtain anatomic measurements and define boundaries including skin and inner and outer borders of the body wall musculature. Automatic image morphologic feature extraction was performed to obtain subcutaneous fat, visceral fat, and skeletal muscle volumes. Body wall volume was defined as the volume of the region including body wall muscle, adipose tissue, and bone. Measures were analyzed with Pearsonís correlation and two-sided paired t-test.

Results: 100 matched patients were analyzed. Three predictive morphometric domains (figure 1) were associated with IH formation: 1) structural widening of the rectus complex, specifically greater linea alba width (p<0.01) and rectus muscle width (p<0.05); 2) larger intra-abdominal volume, specifically fat, represented by volume deep to the body wall musculature (7069.9cc vs 6299.2cc, p<0.05), height-normalized abdominopelvic volume (p<0.01), and total pelvic volume (p<0.01); and 3) abdominopelvic skeletal musculature atrophy, specifically a lower quantity of height-normalized abdominopelvic skeletal muscle tissue (0.74 vs 0.90, p<0.05) and a lower pelvic body wall volume (6465.7cc vs 6986.3cc p<0.01). Multivariate Pearson correlation demonstrated low correlation between morphometric domains supporting featural predictive independence.

Conclusions: We report previously-unidentified unique, independent, morphometric domains present on routine pre-operative CT imaging predictive of subsequent post-operative IH formation after abdominal surgery. This work highlights an innovative pathophysiologic mechanism for IH formation hallmarked by structural attenuation of the rectus complex coupled with generalized abdominopelvic muscle atrophy and increased three-dimensional intra-abdominal volume.


Back to 2022 Abstracts