Multimodal CT Radiomics and Deep Learning Signature Predicts Chondrosarcoma Survival

A multimodal signature combining CT radiomics, deep learning, and pathomics achieved a C-index of 0.821 for progression-free survival prediction in chondrosarcoma, outperforming individual signatures (validation cohort).

• Multicenter retrospective study: 166 patients (training n=117, validation n=49).
• The multimodal signature integrated preoperative CT radiomics and deep learning with postoperative pathology.
• High-risk vs low-risk patients by signature had significantly different progression-free survival (P<0.001).

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.