Objective To explore the nomogram based on multi-parameter MRI and clinicopathological characteristics to predict Tumour-infiltrating lymphocytes(TILs)level in breast cancer.Methods A retrospective study of 222 patients with surgically and pathologically confirmed breast cancer,MRI features and clinicopathologic features of patients were eval-uated.TIL＜10％was considered to be low TIL levels and TIL≥ 10％was considered to be high TIL levels.The clinico-pathological prediction model,MRI prediction model and clinicopathic-MRI combined prediction model were constructed based on multivariate logistic regression analysis.The(AUC)area under receiver operating characteristic(ROC)curve was used to evaluate the prediction efficiency of the models,and the delong test was used to compare the models.A com-bined clinicopathological and MRI prediction model was selected to construct a visual nomogram,and the calibration curve was used to evaluate the nomogram.Results A total of 222 patients with breast cancer were enrolled,including 156 in the training group and 66 in the validation group.In the training group,85 cases(54.5％)in the low TIL levels group and 71 cases(45.5％)in the high TIL levels group.The internal enhancement features(P=0.004),multifocal/multicentric(P=0.018),ADC value of tumor(P=0.001),Ki-67 index(P=0.001),and molecular subtype(P=0.010)were associ-ated with TIL level.In the training set and validation set,the combined clinicopathological and MRI model had the best per-formance,with AUC values of 0.814 and 0.790,respectively.The combined prediction model was visualized using a nomo-gram.Conclusion The nomogram based on multi-parameter MRI combining with clinicopathological features could pre-dict TIL levels of breast cancer,which providing an accurate,comprehensive and efficient evaluation method for clinical practice.
Breast cancerMagnetic resonance imagingTumor-infiltrating lymphocytesNomogram
NETL
NSTL
万方数据
