Quantification of vessel separation using the carotid-jugular angle to predict the nerve origin of neck peripheral nerve sheath tumors: A pooled analysis of cases from the literature and a single-center cohort.

Abstract

Postoperative nerve palsy is a major complication following resection of neck peripheral nerve sheath tumors (PNSTs). Accurate preoperative identification of the nerve origin (NO) can improve surgical outcomes and patient counseling. This study was a retrospective cohort and quantitative analysis of the literature. We introduced a parameter, the carotid-jugular angle (CJA), to differentiate the NO. A literature review of neck PNST cases from 2010 to 2022 was conducted. The CJA was measured from eligible imaging data, and quantitative analysis was performed to evaluate the ability of the CJA to predict the NO. External validation was performed using a single-center cohort from 2008 to 2021. In total, 17 patients from our single-center cohort and 88 patients from the literature were analyzed. Among them, 53, 45, and 7 patients had sympathetic, vagus, and cervical nerve PNSTs, respectively. Vagus nerve tumors had the largest CJA, followed by sympathetic tumors, whereas cervical nerve tumors had the smallest CJA (P<0.001). Multivariate logistic regression identified a larger CJA as a predictor of vagus NO (P<0.001), and receiver operating characteristic (ROC) analysis showed an area under the curve (AUC) of 0.907 (0.831-0.951) for the CJA to predict vagus NO (P<0.001). External validation showed an AUC of 0.928 (0.727-0.988) (P<0.001). Compared with the AUC of the previously proposed qualitative method (AUC=0.764, 0.673-0.839), that of the CJA was greater (P=0.011). The cutoff value identified to predict vagus NO was ≥100°. ROC analysis showed an AUC of 0.909 (0.837-0.956) for the CJA to predict cervical NO (P<0.001), with a cutoff value <38.5°. A CJA ≥ 100° predicted a vagus NO and a CJA < 100° predicted a non-vagus NO. Moreover, a CJA < 38.5 was associated with an increased likelihood of cervical NO.