Quantitative Impact of T1 Subtraction Maps on Enhancing Component Delineation and Measured Volumes in Minimally Enhancing Pediatric Brain Tumors

BACKGROUND AND PURPOSE: The delineation of contrast enhancement in pediatric brain tumors is crucial for effective surgical and treatment planning, as well as for monitoring treatment response per Response Assessment in Pediatric Neuro-Oncology (RAPNO) guidelines. Accurate delineation of enhancement is also important in ground truth generation for training automated deep learning models. However, many pediatric brain tumors frequently demonstrate borderline minimal to mild enhancement, which can be difficult to detect and lead to suboptimal reader agreement. This study aimed to quantitatively determine whether the use of T1-weighted subtraction maps during manual segmentation alters the measured volume of the enhancing component (EC) in mildly enhancing pediatric brain tumors. MATERIALS AND METHODS: Twenty treatment-naïve pediatric brain tumor cases with mild enhancement were selected from the Children's Brain Tumor Network (CBTN) data set. Pre-contrast T1-weighted images were co-registered to their corresponding post-contrast scans, and subtraction maps were generated by subtracting pre-from post-contrast images. Tumors were then segmented twice: first using pre- and post-contrast T1-weighted images alone and then with the addition of T1 subtraction maps. In both cases, the EC was delineated based on the consensus of two board-certified pediatric neuroradiologists. EC volumes (mm³) were compared using the two-tailed Wilcoxon signed-rank test. Percent change in volume was calculated and categorized as minimal (<25%), moderate (25-100%), or substantial (>100%). RESULTS: Segmentation incorporating T1 subtraction maps yielded larger EC volumes in all tumors. Median volume increased from 3,917 mm³ (IQR = 13,411) to 9,222 mm³ (IQR = 18,474), p<0.001. The mean ± SD percent increase was 181 ± 317% (median = 41.9%; range = 0.8-1,143%). Volume changes were minimal in 8 cases (40%), moderate in 5 (25%), and substantial in 7 (35%). CONCLUSIONS: The use of T1 subtraction maps resulted in greater measured EC volumes, with substantial increases in one-third of cases. These findings, based on expert consensus, indicate improved sensitivity for detecting mild enhancement and support further evaluation of subtraction imaging as an adjunctive tool in pediatric neuro-oncology practice, to improve consistency and accuracy in treatment planning and response assessment.