Reliability study of radiology interpretation in osteosarcoma

Introduction: Radiography interpretation of osteosarcoma is almost similar with the benign and other malignant bone tumors. The radiologists’ ability to accurately diagnose osteosarcoma will determine the therapy and diagnostic consistency. Currently, there is lack of evidence on radiography interpretation reliability of osteosarcoma performed by radiologists. This study aimed to evaluate the interobserver and intraobserver radiologists’ compatibility to confirm the diagnosis of osteosarcoma and explore the interpretation variability of radiological description. Methods: This was an interobserver and intraobserver reliability study for osteosarcoma radiograph interpretation. The osteosarcoma radiographs, which have been histopathologically diagnosed, were evaluated retrospectively by two observers and twice by one observer in a different timeline. The components of osteosarcoma identified through the radiographs are bone destruction and its pattern, zone of transition, type of periosteal reaction, type of matrix calcification, skip metastasis, and soft tissue mass. The data analyzed with reliability testing through agreement analysis and Kappa value. Results: After evaluation of 30 radiographs of osteosarcoma, substantial reliability score for periosteal reaction (Kappa 0.79) to almost perfect for bone destruction, types of bone destruction, transitional zone, types of periosteal reaction, calcification matrix, and presence of soft tissue mass (Kappa 0.84-1) were found. There is a strong level of agreement (96.7%) found for skip metastasis. Intraobserver reliability also showed substantial value for the presence of soft tissue mass (Kappa 0.78) and almost perfect for other characteristics of osteosarcoma (0.84-0.93). Conclusion: Intraobserver and interobserver reliability of osteosarcoma radiograph interpretation was substantial and almost perfect.