Solitary bone tumor imaging reporting and data system (BTI-RADS). Initial assessment of a systematic imaging evaluation and comprehensive reporting method

Identify the most pertinent imaging features for solitary bone tumor characterization using a multimodality approach and propose a systematic evaluation system. Data from a prospective trial, including 230 participants with histologically confirmed bone tumors, typical ''do not touch'' lesions, and stable chondral lesions, were retrospectively evaluated. Clinical data, CT, and MR imaging features were analyzed by a musculoskeletal radiologist blinded to the diagnosis using a structured report. The benign-malignant distribution of lesions bearing each image feature evaluated was compared to the benign-malignant distribution in the study sample. Benign and malignant indicators were identified. Two additional readers with different expertise levels independently evaluated the study sample. The sample included 140 men and 90 women (mean age 40.7 $\pm$ 18.3 years). The global benign-malignant distribution was 67-33%. Seven imaging features reached the criteria for benign indicators with a mean frequency of benignancy of 94%. Six minor malignant indicators were identified with a mean frequency of malignancy of 60.5%. Finally, three major malignant indicators were identified (Lodwick-Madewell grade III, aggressive periosteal reaction, and suspected metastatic disease) with a mean frequency of malignancy of 82.4%. A bone tumor imaging reporting and data system (BTI-RADS) was proposed. The reproducibility of the BTI-RADS was considered fair (kappa = 0.67) with a mean frequency of malignancy in classes I, II, III, and IV of 0%, 2.2%, 20.1%, and 71%, respectively. BTI-RADS is an evidence-based systematic approach to solitary bone tumor characterization with a fair reproducibility, allowing lesion stratification in classes of increasing malignancy frequency.