It is unnecessary to perform axillary staging in patients with ductal carcinoma in situ (DCIS) of the breast because of the low incidence of axillary metastasis. However, diagnosis of DCIS by core needle biopsy showed a high rate of underestimation of invasive cancer. Thus, it is necessary to predict invasiveness in DCIS patients on core before surgery. We analyzed 340 patients with DCIS diagnosed by needle biopsy. The cases were divided into training and validation sets. Logistic regression was performed to predict the presence of invasive cancer in the final pathology, and a nomogram was constructed from the training set using the presence of palpability, the presence of ultrasonographic calcification and mass, the biopsy tools, and the presence of microinvasion. The model was subsequently applied to the validation set. The nomogram for the training set was both accurate and discriminating, with an area under the receiver operating characteristic curve (AUC) of 0.75. When applied to the validation group, the model accurately predicted the likelihood of invasive cancer (AUC: 0.71). Our nomogram will allow surgeons to easily and accurately estimate the likelihood of invasive cancer in patients with DCIS as diagnosed by preoperative needle biopsy.
All Science Journal Classification (ASJC) codes