Multicentric Bone Tumor Imaging Report and Data System

Bone Neoplasm

Bone Neoplasms, Multidetector Computed Tomography, Radiographs, Magnetic Resonance Imaging, Evidence-Based Medicine

In this study a previously described structured multimodality image report system for the characterization of focal bone lesions is evaluated in a larger patient population. The objective of this study is to evaluate the performance of this tool stratifying the malignancy risk of bone tumors.

The characterization of focal bone lesions by imaging can be difficult. Primary bone sarcomas are rare, accounting for 0.2% of all malignant tumors occurring at an estimated rate of one tenth that of soft tissue sarcomas. Focal bone lesions have a wide differential diagnosis, including benign and malignant neoplasms, metabolic disorders, degenerative changes, and tumor-like conditions. The precise differentiation between benign and malignant bone tumors is essential for optimal patient management, with a considerable impact on prognosis and survival rates. The relapse-free survival of patients with sarcoma is significantly better when treatment is guided by a multidisciplinary oncology committee. In addition, surgical treatment in referral centers reduces the risk of recurrence and death. Due to the rarity of primary malignant bone neoplasms and the varied imaging presentation of focal bone lesions, radiologists outside of cancer centers tend to have little experience with this type of abnormality. Thus, imaging reports can be unclear and misleading, increasing the risk of misdiagnosis and suboptimal patient management. Previous studies have largely addressed the specific imaging features of bone tumors, and a systematic approach to the assessment of bone tumors has been recommended. However, there is little information on how to combine these imaging results and which are most relevant for characterization of lesions. Various studies have demonstrated the value of structured analysis of medical information in clinical decision making and such an approach is currently used for the assessment of malignancy in various organs and systems. For bone tumors, this approach requires the combined analysis of several demographic, clinical and imaging characteristics. Of the large number of features described, 16 (seven benign and nine malignant) were considered relevant for the differentiation between benign and malignant tumors. Three signs (Lodwick-Madewell grade III, aggressive periosteal reaction and suspected or confirmed metastatic disease) had a mean frequency of associated malignancy greater than 80%. Thus, lesions showing any of these signs should be considered malignant until proven otherwise. It should be noted that these 3 signs are CT or standard radiography criteria, confirming the essential role of X-ray-based methods for the characterization of bone tumors. On the contrary, certain signs classically suggestive of malignancy, such as a history of pain, pathological fracture and endosteal scalloping, have been identified as non-determining indicators for the characterization of these lesions. These data could contribute to a more precise assessment of the aggressive potential of a bone tumor. Finally, on the basis of these results, an evidence-based classification of solitary bone lesions (BTI-RADS) was proposed, allowing the stratification of bone tumors into four classes with an average frequency of malignancy of 0%; 2.2% (1.1 - 3.1%), 20.1% (17 - 24.4%) and 71% (65.6 - 75%) for each class. This system was applicable for readers with different levels of expertise, including a general radiologist, with acceptable interobserver reproducibility (Kappa = 0.67). BTI-RADS could be particularly beneficial outside of specialized oncology centers. However, the BTI-RADS was established based on a single-center analysis of a relatively small patient population. The application of this system in a larger population in multicentric study is necessary to validate this tool and potentially refine it through the identification of additional pertinent criteria for lesion characterization.

France