Abstract Primary bone malignancies are a rare musculoskeletal tumor that carries a high mortality and significant patient burden. The most common primary bone malignancies are osteosarcoma, Ewing's sarcoma, and chondrosarcoma, each with its own unique radiological presentation. These tumors are difficult to diagnose due to their relative rarity and their ability to mimic other conditions, such as infection. As a result, diagnosis can be delayed, which worsens prognosis. Radiomics poses a solution to this problem, quantifying previously subjective changes and detecting small nuances on imaging. While small studies populate the literature field, there is no formal synthesis of evidence. This systematic review and meta-analysis aims to bridge this knowledge gap and evaluate the utility of radiomics in primary bone tumors. Parallel searches were conducted in PubMed, Embase, Scopus, and Web of Science for studies assessing the use of radiomics in the diagnosis of primary bone tumors. Papers were screened and extracted by two independent reviewers, with conflicts resolved through consensus. A bivariate random-effects meta-analysis was conducted to achieve pooled estimates of the diagnostic capabilities of radiomics. Searches retrieved 330 eligible papers, 13 of which were included in the meta-analysis. Pooled estimates show a sensitivity of 0.81 (95% confidence interval CI: 0.75–0.85), a specificity of 0.86 (95% CI: 0.79–0.90), and a diagnostic odds ratio of 30.63 (95% CI: 16.35–57.40). These results demonstrate the significant potential of radiomics in the diagnosis of primary bone malignancies. While the pooled estimates demonstrate the promise radiomics offers to aid radiological diagnosis of primary bone tumors, there are a few limitations to their integration in clinical workflow. Due to the rarity of primary bone malignancies, studies are often restricted to smaller populations, limiting the transferability and introducing publication bias in pooled analysis. To further strengthen the literature field and the clinical integration of radiomics in primary bone tumors, large multicenter prospective cohort studies should be conducted to best assess the role of radiomics in diagnostic workflows.
Gibson et al. (Tue,) studied this question.