Tumor thickness and histological features as predictors of invasive foci within preoperatively diagnosed ductal carcinoma in situ

Abstract

Small invasion into ductal carcinoma in situ (DCIS) can easily be overlooked in resected breast specimens. To disclose useful markers predictive of invasive foci within preoperatively diagnosed DCIS lesions, a retrospective histopathological comparison was made between postoperatively diagnosed invasive ductal carcinoma with a predominant intraductal component (IDCPIC) (n=43) and pure DCIS (n=82). Through a multivariate logistic regression analysis model, 5 variables (DCIS grade, "tumor thickness," extent of retraction cleft, presence of lymph node metastasis, and HER2 score) were found to be significantly associated with the presence of invasive foci within DCIS; with a cutoff point of 0.315, sensitivity, specificity, positive predictive value, and negative predictive value were 0.93, 0.77, 0.68, and 0.95, respectively. No statistically significant difference was observed in recurrence-free survival between IDCPIC and pure DCIS, whereas the IDCPIC curve showed a slightly earlier decline than the DCIS one. In general, preoperative detection of lymph node metastasis in DCIS patients is elusive because of the extremely tiny metastatic size in most cases; thus, a 4-variable model, without lymph node metastasis, would be the actual working model. Furthermore, tumor "thickness" was found to be the most significant parameter predictive of invasive foci within DCIS. Although IDCPIC and pure DCIS showed similar recurrence-free survival curves, prediction of invasive foci within DCIS necessitates postoperative pathological analysis of surgically resected lesions.

Keywords: DCIS; Invasive ductal carcinoma with a predominant intraductal component (DICPIC); Logistic regression analysis; Predicting invasion within DCIS; Tumor thickness.