We assessed 613 cases of DCIS with or without microinvasion that were consecutively resected from 2003 to December 2014 at Seoul National University Bundang Hospital. At our institution, sampling for histologic examinations includes all sections of any grossly apparent lesions and margins. However, for large lesions in mastectomy specimens, sampling is performed from the whole section of the largest tumor slice and from representative sections with 0.5-cm intervals. Clinicopathological data were obtained by reviewing medical records and hematoxylin and eosin-stained sections. The following clinicopathological variables were determined: extent of DCIS, type of surgery, safety margin, axillary staging method, lymph nodal status, features of DCIS (nuclear grade, necrosis, and architectural pattern), microinvasion, and recurrence. For cases with microinvasive carcinoma, the number of microinvasive foci was also recorded. All cases were reviewed by two pathologists (M.K. and S.Y.P.). This study was approved by the Institutional Review Board, and informed consent was waived off (IRB number: B-1701/377-304).

Immunohistochemistry (IHC) for basic biomarkers, including the estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), p53, and Ki-67, was performed at the time of diagnosis. For cases with missing data, immunohistochemical staining was carried out using representative tissue sections and the BenchMark XT autostainer (Ventana Medical Systems, Tucson, USA) with the UltraView detection kit (Ventana Medical Systems). The following antibodies were used: ER (1:100, clone SP1; Labvision, Fremont, USA), PR (1:70, PgR 636; Dako, Carpinteria, USA), HER2 (ready to use, 4B5; Ventana Medical Systems), p53 (1:600, D07; Dako), and Ki-67 (1:250, MIB-1; Dako).

After reviewing the pathological reports and immunohistochemically stained slides for the basic biomarkers, ER and PR were regarded as positive if the tumor showed at least 10% positive nuclear staining. HER2 expression was scored according to the 2013 American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) guidelines. p53 was considered positive if the tumor showed 10% or more positive staining, and a Ki-67 proliferation index was considered to be high if 20% or more of the tumor cells showed positive staining.

Of the 136 cases of microinvasive carcinoma, slides of 86 cases were available for review of basic biomarker analyses. Among the 86 cases, 67 (77.9%) had microinvasive foci in the immunohistochemically stained slides, and the biomarker status was analyzed in the microinvasive foci. In these cases, the biomarker status was same in DCIS and microinvasive foci. When microinvasive foci were not present in the immunohistochemically stained slides, DCIS results were recorded. When slides were not available, biomarker statuses were recorded according to pathological reports.

We used tissue microarray (TMA) to determine the HER2 status of HER2 IHC 2+ cases. To exclude sampling errors from TMA platform evaluations, the section that was the most representative of the tumor was selected in each case. One tissue column with a diameter of 4.0 mm was taken from the representative tumor area, which included microinvasive foci for microinvasive carcinoma, and was arranged in TMA blocks using a trephine apparatus (Superbiochips Laboratories, Seoul, Korea).

HER2 fluorescence in situ hybridization was performed using TMAs to determine the presence/absence of HER2 amplification. Briefly, 4-µm-thick deparaffinized TMA sections were incubated in a pretreatment solution (Abbott Molecular, Downers Grove, USA) at 80℃ for 30 minutes, then in a protease solution (Abbott Molecular) for 20 minutes at 37℃. Probes were diluted in tDen-Hyb-2 hybridization buffer (InSitus Biotechnologies, Albuquerque, USA). Co-denaturation of the probes and DNA from the tissue sections was achieved by incubating them for 5 minutes in 73℃ using HYBrite™ (Abbott Molecular), followed by a 16-hour hybridization at 37℃. Posthybridization washes were performed according to the manufacturer protocols. Slides were mounted using 4′,6-diamidino-2-phenylindole, and at least 20 tumor cell nuclei were counted in each case under a fluorescence microscope.

According to the 2013 ASCO/CAP guidelines, a HER2 copy number of 6.0 or higher per cell or a HER2:CEP17 ratio of 2 or higher was considered to represent HER2 amplification. HER2/CEP17 ratios of <2 and HER2 copy numbers of 4–6 signals per cell were classified as equivocal, and HER2 copy numbers of <4 signals per cell and HER2/CEP17 ratios of <2 were considered nonamplified. In this study, HER2-equivocal cases were regarded as HER2-nonamplified for statistical analyses.

Breast cancer subtypes were categorized according to the 2011 St. Gallen Expert Consensus, in a similar manner to a previous study [17]: luminal A (ER+ and/or PR+, HER2–, Ki-67 <14%), luminal B (ER+ and/or PR+, HER2–, Ki-67 ≥14%; ER+ and/or PR+, HER2+), HER2+ (ER–, PR–, HER2+), and triple-negative (ER–, PR–, HER2–).

Statistical analysis was performed using SPSS version 15.0 (SPSS Inc., Chicago, USA). Categorical variables were analyzed by Pearson chi-square or Fisher exact tests. Continuous data were analyzed by independent sample t-test or the Mann-Whitney U-test. Univariate and multivariate binary logistic regression analyses were performed to identify independent predictive factors of microinvasion. Backward stepwise model selection was used to construct the multivariate logistic regression model, and odds ratios and their 95% confidence intervals (CIs) were calculated for each factor. Recurrence-free survival was defined as the time from the date of surgery to the date of recurrence. Survival curves were drawn using the Kaplan-Meier method and were compared using the log-rank test. The Cox proportional hazards regression model was also used for univariate survival analysis, and hazard ratios (HRs) and its 95% CIs were calculated for the variables. p-values less than 0.05 were considered statistically significant, and all reported p-values were two sided.

This study included 613 cases of DCIS with or without microinvasion, of which 136 (22.2%) were microinvasive carcinoma. The mean extent of DCIS was 3.2±2.3 cm. Axillary staging was performed for 300 cases (48.9%), of which 12 underwent axillary node dissections and 288 underwent sentinel node biopsies. The other baseline characteristics are summarized in Table 1.

We compared clinicopathological characteristics between microinvasive carcinoma and pure DCIS (Table 2). Microinvasive carcinoma was significantly associated with DCIS with a large extent (≥3.2 cm, dichotomized by the mean size); and thus, microinvasive carcinoma was more frequently treated by mastectomy (p<0.001). Microinvasive carcinoma was also associated with DCIS with high nuclear grade, necrosis, and comedo-type architectural pattern (all p<0.001) (Figure 1). In terms of biomarker expression, hormone receptor expression was high in pure DCIS. Positive HER2 status, high Ki-67 proliferation index, and p53 overexpression were more frequently observed in microinvasive carcinoma than in pure DCIS (all p<0.001) (Figure 1). As for breast cancer subtype, the luminal A subtype was more frequent in pure DCIS, whereas the HER2+ and triple-negative subtypes were more frequent in microinvasive carcinoma. Of the 300 patients who underwent axillary node staging operations, nodal metastases were found in only four patients with microinvasive carcinoma (p=0.017).

Among the 136 cases of microinvasive carcinoma, 41 had multifocal microinvasion. Clinicopathological characteristics, including age, extent of DCIS, nuclear grade, necrosis, and architectural pattern, were compared between cases with single invasion and those with multifocal invasion. We observed no significant differences in clinicopathological characteristics between microinvasive carcinoma with single and multifocal invasion, except for axillary node metastasis. Of the 110 patients who underwent axillary staging surgery, 72 patients with single microinvasions showed no lymph node metastases. Only four of the 38 patients with multifocal microinvasion showed lymph node metastases, of whom two patients were N1mi and two were N1a (p=0.013) (Figure 2). Of the four patients with axillary node metastases, three had the luminal A subtype, and the remaining one had luminal B subtype. All of the four patients had a tumor extent greater than 5.0 cm, and thus, received mastectomies. Patients with multifocal microinvasion tended to exhibit a higher Ki-67 proliferation index than those with a single microinvasion (43.9% vs. 28.4%, p=0.078).

In order to identify factors that independently predict microinvasion, univariate and multivariate binary logistic regression analyses were performed (Table 3). In the univariate analysis, large tumor extent (greater than 3.2 cm), high nuclear grade, necrosis, comedo type, negative ER, negative PR, positive HER2 status, p53 overexpression, and high Ki-67 proliferation index were found to be predictive factors of microinvasion (all p<0.001). Among these variables, large tumor extent, comedo-type architecture, and ER negativity were found to be independent predictors of microinvasion (all p<0.001).

We evaluated the clinical outcomes of patients with microinvasive carcinoma and pure DCIS. Most patients were treated according to standard guidelines and were followed up regularly after surgery. The median follow-up period was 4.0 years (range, 0.1–12.4 years), during which 12 patients had ipsilateral breast cancer recurrence. The clinicopathological characteristics of these patients are presented in Table 4. Of the 12 patients, nine were originally diagnosed with pure DCIS and three with microinvasive carcinoma. The histologic type of recurrent tumors varied from pure DCIS and microinvasive carcinomas to invasive ductal carcinomas. Interestingly, three patients with triple-negative DCIS or microinvasive carcinoma had invasive carcinoma recurrences. In survival analyses, only the triple-negative subtype was found to be a prognostic factor associated with all types of ipsilateral breast recurrences (HR, 4.136, 95% CI, 1.120–15.282, p=0.033, Cox proportional hazards regression analysis; p=0.021, log-rank test) and invasive recurrences (HR, 37.188, 95% CI, 3.867–357.599, p=0.002, Cox proportional hazards regression analysis; p<0.001, log-rank test) (Table 5, Figure 3). After ipsilateral breast recurrence, one (case 83) patient subsequently developed local recurrence in the chest wall, and another patient (case 588) developed distant metastasis.