Archive-Recorded Volume Change in Matched Renal Target Lesions: A Secondary Analysis of Public Pediatric Oncology Annotations

Background: Public tumor annotations permit secondary measurement studies, but longitudinal analyses require consistent lesion identity, examination selection, and data provenance.Objective: This study aimed to describe changes in the archive-recorded volumes of matched renal target lesions between two timepoints in three Children's Oncology Group trial archives and to document the matching and verification steps that such secondary analyses require.Materials and methods: We analyzed the corrected version 2 annotation releases for AREN0532, AREN0533, and AREN0534.The primary analysis required computed tomography (CT) at baseline and the first numbered follow-up, one examination per visit, identical renal tracking identifiers and compatible labels, and verification of each intended region against its source annotation object.The endpoint was percentage change in summed archive-reported target volume.Collection-specific medians and nonparametric 95% confidence intervals (CIs) were calculated.Magnetic resonance imaging (MRI) pairs and CT pairs separated by no more than 90 days were evaluated separately.Results: Of 1,047 subjects with renal segmentations, 227 (21.7%) met all the primary criteria; 720 lacked a CT renal segmentation at the first numbered follow-up, including 479 with no post-baseline annotation record.Median percentage changes were AREN0532: -85.2% (95% CI -92.8% to -63.8%; n = 18); AREN0533: -79.6% (95% CI -82.5% to -74.7%; n = 81); AREN0534: -71.8% (95% CI -78.4% to -65.4%; n = 128).Median examination intervals were 47 days, 45 days, and 43 days, respectively.Recorded volume decreased in 212 of 227 patients.In 57 patients with complete contour recomputation, archived and recomputed median changes were similar (-78.0%and -77.4%), which supports computational consistency rather than segmentation accuracy.The separate MRI analysis included 39 patients.Clinical treatment, surgical timing, and outcomes were not linked.Conclusions: In this secondary analysis of public trial-archive annotations, the recorded renal target volume decreased substantially among targets that remained annotated and evaluable at both timepoints.Because this cohort is a selected minority and source images, treatment, pathology, and outcomes were not reviewed or linked, these measurements describe archived annotations and do not establish clinical response, treatment efficacy, or prognostic utility.