Tumour Imaging: The Fundamentals (2025)-W8-CMK03-2025

Video Transcription

Video Summary

The transcript covers several musculoskeletal radiology initiatives aimed at standardizing reporting, improving management recommendations, and reducing radiology–pathology mismatch.<br /><br />First, speakers discuss multiple “RADS” systems for bone lesions, focusing on the ACR Bone-RADS designed for radiographs to guide especially non–MSK radiologists. They note practical limitations: unclear definitions (e.g., what counts as “history of primary cancer”), difficulty mirroring BI-RADS management pathways, uncertain surveillance intervals, and confusion where features like pathologic fracture alter management more than malignancy risk. Validation studies show generally good diagnostic accuracy (AUC often >0.8), weaker agreement for features like endosteal scalloping, and persistent challenges with cartilage tumors—prompting suggestions for follow-up imaging. CT “applicability” work suggests consistency but not accuracy. Machine learning and large language models can help apply these systems using structured inputs.<br /><br />Next, Soft Tissue RADS (ST-RADS) is presented as COVID-era work to standardize risk stratification for soft-tissue masses. An ACR-branded update adds clearer terminology (tumor vs tumor-like lesions), evidence-based tables, refined categories, algorithms for generalists vs MSK experts, structured templates, and validation showing improved reliability and reader confidence, especially among juniors. A GPT-based calculator using text features performed similarly to expert readers.<br /><br />Finally, a talk on imaging–pathology discordance emphasizes causes (missing clinical info, suboptimal protocols, sampling error, poor communication) and recommends optimized imaging, structured reports, multidisciplinary discussion, and RADPATH addenda labeling cases as concordant or discordant.<br /><br />A brief pediatric section highlights osteosarcoma staging: full-bone MRI (T1 for intramedullary extent, skip metastases), chemical shift for red marrow vs tumor, limited utility of ADC thresholds in children, potential roles for DCE-MRI and PET, and emerging radiomics; limb-sparing surgery is now common.

Keywords

ACR Bone-RADS

bone lesion reporting standardization

ST-RADS soft tissue mass risk stratification

structured radiology reporting templates

radiology–pathology discordance (RADPATH)

cartilage tumor surveillance and follow-up imaging

machine learning and large language models in radiology

pediatric osteosarcoma staging MRI