Radiographic evaluation of the psoas and iliopsoas muscle as predictors for spinal cord ischemia after fenestrated and branched endovascular aortic repair.

Abstract

This study aimed to investigate the association between sarcopenia and spinal cord ischemia (SCI) after fenestrated and branched endovascular aortic repair (F/B-EVAR) using two- and three-dimensional measurements of the psoas and iliopsoas muscles on preoperative computed tomography angiography (CTA). A retrospective, two-center study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Data were collected from patients with Crawford type I-IV thoracoabdominal aortic aneurysms (TAAA) and pararenal abdominal aortic aneurysms treated with F/B-EVAR between December 2010 and January 2024. One center included all consecutive patients, whereas from the other center, patients with SCI were actively selected together with non-SCI patients in a 1:1 fashion based on patient and procedural characteristics. Preoperative CTAs were analyzed for surrogate markers of sarcopenia, including the psoas muscle area (cm²), lean psoas muscle area (cm²∗HU [Hounsfield units]), iliopsoas muscle volume (cm³), and lean iliopsoas muscle volume (cm³∗HU). Area measurements were performed manually, whereas volume measurements were performed using an artificial intelligence-based segmentation tool. The primary outcome was to evaluate the predictive value of the measured sarcopenia surrogate markers for SCI occurrence. A total of 138 patients (35.5% female; median age 72 years, interquartile range [IQR]: 68-75 years), with 16 Crawford type I (11.6%), 45 type II (32.6%), 30 type III (21.7%), and 47 type IV/pararenal (34.1%) aneurysms, were included. Fifty-one patients had postoperative SCI (all severities), and 87 had no SCI symptoms. Compared with non-SCI patients, patients with SCI had higher American Society of Anesthesiologists classification (P = .005), more commonly type II TAAA (P < .001), and symptomatic presentation (P = .016). Other patient characteristics were similar between the groups. Psoas muscle area (6.97 cm² [IQR: 5.22-8.73 cm²] vs 8.47 cm² [IQR: 6.39-10.03 cm²], P = .003), lean psoas muscle area (253.3 cm²∗HU [IQR: 204.9-333.8 cm²∗HU] vs 335.6 cm²∗HU [IQR: 256.3-409.7 cm²∗HU], P = .002), iliopsoas muscle volume (247.6 cm³ [IQR: 184.0-303.8 cm³] vs 277.7 cm³ [IQR: 234.1-331.5 cm³], P = .018), and lean iliopsoas muscle volume (10,879 cm³∗HU [IQR: 8589-14,497 cm³∗HU] vs 13,445 cm³∗HU [IQR: 10,777-16,396 cm³∗HU], P = .004) were lower in patients with SCI in the unadjusted analyses. On multivariable analysis, only psoas muscle area was independently associated with SCI (odds ratio: 0.815; 95% confidence interval: 0.680-0.977, P = .027). Psoas muscle area was independently associated with SCI after F/B-EVAR, indicating that patients with signs of sarcopenia on preoperative CTA may be at a higher risk for SCI. Volumetric iliopsoas muscle measurements were not better predictors of SCI than two-dimensional measurements of the psoas muscle area.