Division Head of Vascular Surgery, Professor of Surgery University of Toronto Toronto, Ontario
Objectives: This study aims to describe the outcomes and evaluate the predictive utility of preoperative radiological characteristics on postoperative complications in patients undergoing carotid body tumor resection at a tertiary care center.
Methods: A retrospective analysis was conducted on 106 patients who underwent CBT resection between 2003 and 2023. Patient demographics, tumor characteristics, and operative details were collected. The primary outcomes were an estimated blood loss (EBL) >250 mL and cranial nerve (CN) injury.
Results: Most patients were female (72.6%), with a mean age of 46.2 ±15.8 years. Preoperative embolization and carotid balloon occlusion tests were performed in 7.5% and 13.2% of patients, respectively. The Shamblin classification Type III was encountered at 34% of patients, and the mean distance to the base of the skull (DTBOS) was 3.9 ±0.99 cm. Higher Shamblin and PUMCH grades were significantly associated with increased EBL and CN injury. Specifically for CN injury, the Shamblin grade alone had an R² of 0.16, which significantly improved to 0.27 with the addition of DTBOS and further to 0.29 with tumor volume. Additionally, for CN injury, the PUMCH grade alone had an R² of 0.14, improving to 0.21 with DTBOS and to 0.22 with tumor volume. Furthermore, a 1-cm decrease in DTBOS significantly increased the odds of requiring a blood transfusion (OR = 2.26, 95% CI: 1.28-4.01, p=0.0051) and the risk of CN injury (OR = 3.65, 95% CI: 1.98-6.73, p< 0.0001). The multivariate analysis revealed that younger age (p=0.022), left tumor side (p=0.033), preoperative carotid balloon occlusion test (p=0.009), shorter DTBOS (p=0.042), and presence of metastasis (p=0.042) were significant predictors of CN injuries. (Table 4. and 5.)
Conclusions: This study identified novel preoperative radiological predictors that enhance the predictive accuracy of standard classification systems. While the Shamblin and PUMCH classifications are useful tools on their own, our findings demonstrate that incorporating additional radiological features, such as DTBOS and tumor volume, can substantially increase their predictive utility. Surgeons are encouraged to incorporate multiple preoperative radiological variables alongside traditional classification systems to better assess the risk of postoperative complications. Further research with larger, multi-institutional cohorts are necessary to validate these findings and refine predictive models.
