A single-center feasibility study showed that a machine learning model developed for upper limb soft-tissue sarcomas achieved C-indices of 0.79, 0.67, and 0.75 for predicting local recurrence, distant metastasis, and overall survival, respectively, in 188 patients. Its overall performance was comparable to that of nomograms such as PERSARC and Sarculator. The researchers concluded that model calibration still requires improvement and that external validation is essential before clinical use.
Source:AJMC · Read original
Leave a Reply