Development of an MRI-based surgical difficulty scoring system for anterior and deep lobe parotid tumors

Anterior and deep lobe parotid tumors present substantial technical challenges because of their close relationship to the facial nerve and deep parotid anatomy. We aimed to develop and internally validate a practical MRI-based scoring system to preoperatively predict surgical difficulty in these tumors.

We retrospectively reviewed patients who underwent parotidectomy for anterior or deep lobe tumors between March 2020 and December 2024. A five-domain MRI-based difficulty score (range, 3-15) was constructed based on tumor location, size, relationship to the facial nerve, fat-plane integrity, and multifocality. Surgical difficulty was defined as prolonged operative time exceeding the 75th percentile. Predictive validity was assessed using operative time, estimated blood loss, and postoperative facial nerve outcomes. Interobserver reliability was evaluated using intraclass correlation coefficients and Bland-Altman analysis. Internal validation was performed in a standardized single-surgeon cohort.

A total of 149 patients were included. Higher MRI-based difficulty scores were significantly associated with longer operative time and increased estimated blood loss (both p < 0.001). Postoperative facial nerve weakness occurred more frequently with increasing difficulty tiers (p = 0.01). These findings remained consistent in the standardized validation cohort (n = 36). Interobserver reliability was excellent (ICC = 0.87; 95% CI, 0.82-0.91), with minimal mean bias and narrow limits of agreement.

This MRI-based difficulty score reliably predicts surgical complexity in anterior and deep lobe parotid tumors and demonstrates excellent interobserver reproducibility. It may serve as a useful preoperative tool for operative planning and risk stratification. External validation is warranted.