Response Evaluation Criteria in Bone Metastases: Performance and Association of Response Classifications with...

Background Current bone metastasis evaluation relies on fragmented anatomic or metabolic criteria, often yielding discordant classifications. Purpose To develop Response Evaluation Criteria in Bone Metastases (termed RECIBM) and evaluate its performance against the University of Texas MD Anderson Cancer Center, PET Response Criteria in Solid Tumors (PERCIST), and European Organisation for Research and Treatment of Cancer (EORTC) criteria. Materials and Methods This retrospective study was conducted between October 2017 and September 2023. RECIBM was established based on multidisciplinary consensus and assessed on real-world patient data, with treatment responses evaluated using RECIBM, MD Anderson, PERCIST, and EORTC criteria. The paired subgroup comprised patients evaluable by all four criteria; the remaining patients underwent technetium 99m methylene diphosphonate bone scintigraphy only and were evaluable using the MD Anderson criteria and RECIBM. Agreement was quantified using the Cohen κ value, and prognostic performance for overall survival and bone progression-free survival was assessed using the Harrell C-index and multivariable Cox regression. Results Eighty-four patients (median age, 59.0 years [IQR, 52.0-68.3 years]; 59 male) were included. RECIBM-defined response classifications demonstrated high agreement with those of existing criteria (κ = 0.80-0.98; all P < .001) and reclassified 19% of patients (16 of 84). In the paired subgroup, RECIBM showed the highest C-index for overall survival (0.807 vs 0.771-0.782; P = .10-.14) and bone progression-free survival (0.849 vs 0.799-0.833; raw P = .03-.26; adjusted P = .088-.256), with no evidence of differences after adjustment. Subgroup analyses showed consistent performance across different imaging modalities and tumor types (P for interaction = .39 and .051, respectively). In multivariable analysis, RECIBM-defined response was associated with overall survival (hazard ratio, 0.07 [95% CI: 0.03, 0.19]; P < .001) and bone progression-free survival (hazard ratio, 0.03 [95% CI: 0.01, 0.08]; P < .001). Conclusion RECIBM-defined response classifications showed high agreement with established criteria, reclassified 19% of patients, and were associated with overall survival and bone progression-free survival. © RSNA, 2026 Supplemental material is available for this article. See also the editorial by Mohammadi and Guermazi in this issue.