Diagnostic performance and reliability of ACR Bone-RADS versus REST for reporting bone tumors on conventional radiographs
Abstract Background and purpose Conventional radiographs are the primary imaging modality for evaluation of bone tumors. This study aimed to compare the reliability and validity of the radiological evaluation score for bone tumors (REST) and the Bone reporting and data system (Bone-RADS) by American...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2025-07-01
|
| Series: | The Egyptian Journal of Radiology and Nuclear Medicine |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s43055-025-01530-0 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background and purpose Conventional radiographs are the primary imaging modality for evaluation of bone tumors. This study aimed to compare the reliability and validity of the radiological evaluation score for bone tumors (REST) and the Bone reporting and data system (Bone-RADS) by American college of radiology (ACR) in assessment of bony tumors on conventional radiographs. Methods A retrospective study involved 177 primary bony lesions. Three radiologists independently evaluated radiographs to assess features of bony lesions and set REST and Bone-RADS categories. One rater reviewed the radiographs twice. Inter- and intra-rater agreements and the validity of REST and ACR Bone-RADS were calculated. Results One seventy-seven (138, 78% benign, 10, 5.6% intermediate, 29, 16.4% malignant) bone lesions were involved. There were almost perfect inter- and intra-rater agreements for most of the features as well as benign/malignant classification according to REST and Bone-RADS, agreements for periosteal reaction, extra-osseous soft tissue, scalloping, and categories of Bone-RADS were substantial (k: 0.79, 0.75, 0.64, 0.78), respectively, with higher rates for intra-rater one. Both systems exhibited high sensitivity, negative predictive value (NPV) and accuracy with relatively low positive predictive value (PPV) [87.2–100%, 95.7–100%, 82–85.8%, 50.7–63%], respectively, in predicting malignant potential with REST showing relatively higher accuracy (82.5–85.8%). Conclusions REST and ACR Bone-RADS demonstrated perfect inter- and intra-rater agreements in assessing the malignancy risk of bone tumors, with REST showing higher readings. Additionally, both systems exhibited high sensitivity, NPV and accuracy, but relatively low PPV in predicting malignant potential with REST showing relatively higher accuracy. Thus, ACR Bone-RADS may benefit from further studies to inform potential updates. |
|---|---|
| ISSN: | 2090-4762 |