Photo Credit: Md Saiful Islam Khan
A machine-learning model that combines multisequence magnetic resonance imaging radiomics with clinical parameters predicted lymph node metastasis (LNM) in patients with type I endometrial cancer (EC), researchers reported in the International Journal of Radiation Oncology, Biology, Physics. The model, trained on data from 118 patients and tested on 49 patients, demonstrated an area under the curve of 0.978 compared with 0.926 for a radiomics-only and 0.679 for a clinical parameter-only model. The clinical-radiomics model, the radiomics model, and the clinical model had a sensitivity of 1.000, 1.000, and 0.364; specificity of 0.974, 0.921, and 0.711; accuracy of 0.980, 0.939, and 0.633; positive predictive value of 0.917, 0.786, and 0.267; and a negative predictive value of 1.000, 1.000, and 0.794, respectively. Analysis showed the combined model predicted LNM and yielded benefits. The researchers wrote that their proposed nomogram could improve accuracy when predicting LNM and support lymphadenectomy decision-making.