Three-Dimensional (3D) Periarterial Costoclavicular Distance on Computed Tomography Angiography in Thoracic Outlet Syndrome: Deep Learning-Based 3D Mapping Approach

Background: Objective measurement of costoclavicular distance (CCD) in thoracic outlet syndrome (TOS) is not standardized.We aimed to define a 3-dimensional periarterial CCD on computed tomography angiography (CTA), determine demographic factors associated with CCD, and evaluate its association with photoplethysmography (PPG)-defined arterial compromise. Methods:This cross-sectional study included patients with TOS who underwent preoperative upper extremity CTA and subsequent surgery from January 2021 to December 2023.For each patient, the symptomatic arm, or the arm with greater neuropathic symptoms in bilateral cases, was analyzed.A deep-learning algorithm segmented the clavicle, first rib, and subclavian artery.Straight-line clavicle-first-rib surface distances were calculated, and only lines passing within 1 mm of the segmented subclavian artery were retained.The minimum retained distance was defined as the periarterial CCD.Positive digital PPG was defined as a 90% reduction in arterial pulse amplitude during full shoulder abduction and external rotation.Associations of CCD with age, sex, and body mass index (BMI) were assessed using univariable and multivariable linear regression.Sex-stratified receiver operating characteristic (ROC) analyses evaluated CCD for predicting positive digital PPG.Results: Fifty-one extremities were analyzed; 33 (65%) had positive PPG.Mean CCD was 10.1 3.1 mm overall and was higher in males than females (11.3 3.4 vs 8.7 3.0 mm; p = 0.007).In multivariable analysis, higher BMI ( = 0.40, p < 0.001) and male sex ( = 1.85, p = 0.026) were associated with larger CCD, whereas age was not ( = 0.03, p = 0.239).The area under the ROC curve was 0.72 (95% CI, 0.48-0.91) in males and 0.75 (95% CI, 0.52-0.92) in females.Sex-specific cutoffs for positive digital PPG were 12.5 mm in males (sensitivity, 77%; specificity, 64%) and 7.9 mm in females (sensitivity, 63%; specificity, 100%). Conclusions:We developed an automated CTA-based method to quantify 3D periarterial CCD and assessed its performance for identifying PPG-defined arterial compromise during provocative positioning.Sex-specific CCD cutoffs suggest that a universal threshold may be inappropriate.This metric may support objective anatomical assessment and surgical planning for first-rib resection.