Abstract
Introduction:Conventional classification systems for giant cell tumors (GCTs) lack robust correlation with management and clinical outcomes. We propose a new radiologic classification system based on surgically relevant features to address this shortcoming.Methods:This IRB-approved single-institution retrospective study involved 35 extremity GCTs from 2013 to 2023 with preoperative radiographs and cross-sectional imaging (MRI and/or CT). An experienced musculoskeletal (MSK) radiologist and orthopaedic oncologist independently assessed tumors according to the Campanacci or new grading system, defined on 1 to 3 scale: (1) intraosseous contained tumor, (2) intraosseous noncontained tumor with extraosseous implant accessible through single incision, and (3) intraosseous noncontained tumor with an extraosseous soft tissue implant nonaccessible from single incision alone. Interrater agreement was determined through the intraclass correlation coefficient. The two-way Friedman test with rater and grading system as factors was used to compare system grading similarity.Results:Thirty patients underwent curettage, five underwent resection; 10 experienced local recurrence. Intraclass correlation coefficients between raters for the Campanacci and novel grading systems were 0.83 and 0.79, respectively. However, compared with the novel system, Campanacci grades were significantly higher by an average of 0.34 ± 0.68 and 0.46 ± 0.70 for the first and second raters, respectively (P = 0.003). None of the patients who underwent resection experienced local recurrence, but in patients who underwent curettage, recurrence rates were higher in Campanacci versus novel grade 1 tumors (29% vs. 17%).Discussion:The novel GCT grading system demonstrates excellent interrater agreement, and classified more nonrecurrent curetted tumors as low grade, suggesting improved predictive performance compared with the Campanacci classification.