2026 Proffered Presentations
S204: VOLUMETRIC ANALYSIS OF SKULL BASE CHORDOMAS TO GUIDE RESECTION STRATEGY AND PREDICT CLINICAL OUTCOMES
Sakibul Huq, MD; Zachary C Gersey, MD; Yassine A Idrissi, MD; Afrooz Zandifar, MD; Nasim Batavani, MD; Priyadarshini Mamindla, MSc; Mohammadreza Amjadzadeh, MD; Serafettin Zenkin, MD; Ayla Buyukkaya, MD; Ozlem T Kalayci, MD; Vishal Peddagangireddy, MSc; Ramazan Buyukkaya, MD; Walaa Hamza, MD; Sinef H Aksoy, MD; Hussein Abdallah, MD; Garret Choby, MD; Eric W Wang, MD; Carl H Snyderman, MD, MBA; Paul A Gardner, MD; Georgios A Zenonos, MD; University of Pittsburgh Medical Center
Introduction: Skull base chordomas (SBCs) are surgically challenging tumors with high rates of recurrence. While the goal of surgery is maximal safe resection, it is not known if there are clinically relevant extent of resection (EOR) thresholds that are associated with progression-free survival (PFS). We sought to address this gap using volumetric analysis of treatment-naïve SBCs at a high volume, experienced center.
Methods: All patients underwent magnetic resonance imaging (MRI) using standard protocols with a 1.5T or 3.0T scanner. An experienced team of radiologists performed image analysis and segmentation of pre- and postoperative T2-weighted MRIs using 3D slicer 4.10.2, a previously-validated open-source software package. Volumetric EOR was calculated for each patient. The optimal EOR cutpoint for predicting PFS after surgery was determined using maximally selected rank statistics with 10,000 bootstrap samples, and the median cutpoint from this analysis was applied for binary categorization. Kaplan-Meier survival curves and log-rank tests were used to compare PFS between groups defined by this binary EOR cutpoint. Multivariable Cox proportional hazards regression was then performed to assess the independent predictive value of EOR upon PFS. All analysis was conducted in R using the survminer, ggplot2, and dplyr packages.
Results: Our cohort included 104 unique patients undergoing first-time surgery at our institution with mean age of 44.2 years and male predominance (n=59). The mean and median EOR for the overall cohort were 95.9% and 100%, respectively. The optimal EOR cutpoint for predicting improved PFS was greater than 98.9% resection, separating 21 patients at or below this threshold from 83 patients above it. In this dataset, the cutpoint effectively distinguished patients who had gross total resection from those with any residual tumor volume. On Kaplan-Meier analysis, median PFS for patients above the EOR cutpoint was not reached, while median PFS for patients below the cutpoint was 30.2 months. On multivariable Cox proportional hazards regression, EOR above the cutpoint was a significant predictor of PFS after adjusting for age, sex, molecular group, and post-operative radiation.
Conclusions: Using volumetric analysis of pre- and postoperative MRIs and maximally selected rank statistics, we identified a clinically relevant EOR threshold of 98.9% which was associated with improved PFS in patients with SBC.
