Clinicopathological predictors of postoperative upstaging to invasive ductal carcinoma (IDC) in patients preoperatively diagnosed with ductal carcinoma in situ (DCIS): a multi-institutional retrospective cohort study

Background We conducted a prospective study with the intention to omit surgery for patients with ductal carcinoma in situ (DCIS) of the breast. We aimed to identify clinicopathological predictors of postoperative upstaging to invasive ductal carcinoma (IDC) in patients preoperatively diagnosed with DCIS. Patients and methods We retrospectively analyzed patients with DCIS diagnosed through biopsy between April 1, 2010 and December 31, 2014, from 16 institutions. Clinical, radiological, and histological variables were collected from medical records. Results We identified 2,293 patients diagnosed with DCIS through biopsy, including 1,663 DCIS (72.5%) cases and 630 IDC (27.5%) cases. In multivariate analysis, the presence of a palpable mass (odds ratio [OR] 1.8; 95% confidence interval [CI] 1.2–2.6), mammography findings (≥ category 4; OR 1.8; 95% CI 1.2–2.6), mass formations on ultrasonography (OR 1.8; 95% CI 1.2–2.5), and tumor size on MRI (> 20 mm; OR 1.7; 95% CI 1.2–2.4) were independent predictors of IDC. Among patients with a tumor size on MRI of ≤ 20 mm, the possibility of postoperative upstaging to IDC was 22.1%. Among the 258 patients with non-palpable mass, nuclear grade 1/2, and positive for estrogen receptor, the possibility was 18.1%, even if the upper limit of the tumor size on MRI was raised to ≤ 40 mm. Conclusion We identified four independent predictive factors of upstaging to IDC after surgery among patients with DCIS diagnosed by biopsy. The combined use of various predictors of IDC reduces the possibility of postoperative upstaging to IDC, even if the tumor size on MRI is larger than 20 mm.


Introduction
The increase in breast cancer screening programs has contributed to a dramatic increase in the incidence of ductal carcinoma in situ (DCIS), and more than 20% of breast cancers diagnosed by screening mammography (MMG) are DCIS according to a recent study [1]. It has also been reported that approximately 80% of breast cancers diagnosed by calcifications on screening MMG are DCIS [1].
Surgical management is the current standard approach for DCIS. For breast lesions, breast-conserving surgery followed by radiotherapy or total mastectomy with or without reconstruction is performed. For sentinel lymph nodes, the Japan Breast Cancers Guideline recommends that sentinel lymph node biopsy (SLNB) can be omitted in DCIS patients treated with breast-conserving surgery and predicted to have no invasion [2]; in daily practice, SLNB is sometimes omitted. DCIS has a very good prognosis, and especially for patients with low-risk DCIS, the current standard surgery does not contribute to the improvement of life prognosis [3]. Several randomized controlled trials, such as the COMET [4,5], LORD [6], and LORIS [7] trials, are currently investigating the feasibility and non-inferiority of active surveillance with or without endocrine therapy for managing low-risk DCIS. In Japan, the single-arm JCOG1505 (LORETTA trial, UMIN 000028298) [8] has begun to confirm non-inferiority of endocrine therapy alone compared to surgery for estrogen receptor-positive, low-risk DCIS.
A problem in omitting surgery is that among patients with preoperatively diagnosed DCIS, 8.3-43.6% presents upstaging to invasive carcinoma as determined by examination of postoperative specimens [9][10][11][12][13][14]. Furthermore, the frequency of axillary node-positive among patients preoperatively diagnosed with DCIS is 2.5-6.8% [15][16][17]. Thus, better preoperative information is important to predict DCIS in the final pathological diagnosis so as not to administer overly intensive treatment to patients.
The current study aimed to understand the diagnostic accuracy and treatments of DCIS in institutions with the intention to research individualized DCIS treatment for the future. In addition, we sought to identify clinicopathological predictors of postoperative upstaging to IDC in patients preoperatively diagnosed with DCIS to assist with the provision of adequate surgical procedures.

Patients
We retrospectively reviewed patients diagnosed with DCIS through core needle or vacuum-assisted biopsy between April 1, 2010 and December 31, 2014, from 16 institutions of the Breast Cancer Study Group in Japan Clinical Oncology Group (JCOG). This study was approved by the Institutional Review Boards of each institution. The need for written informed consent was waived due to the retrospective nature of the study, and the patients were provided with a means to opt out.

Preoperative radiological assessment
All patients routinely underwent clinical examination, MMG, ultrasonography (US), and dynamic magnetic resonance imaging (MRI). The collection items were as follows: presence or absence of a palpable mass as a clinical examination finding, category classifications [18] and the presence or absence of calcification as MMG findings, presence or absence of mass formation, low echoic area and mammary duct ectasia as US findings, tumor size including non-mass enhancement and presence or absence of mass formation as MRI findings. All data were collected from medical records or clinical database by breast oncologists in each institution.

Pathological assessment
The pre-and postoperative pathological findings, including estrogen receptor (ER), progesterone receptor (PgR), human epithelial growth factor receptor 2 (HER2), DCIS grade (low, intermediate, or high), nuclear grade, and the presence or absence of comedo necrosis, were collected from the pathological reports in each institution. The tumors on both pre-and postoperative specimens were histologically classified using the World Health Organization criteria [19]. ER and PgR were considered positive if reported as a total Allred score of 3-8 or a positive cell occupancy of 1% or more on immunohistochemical analysis. HER2 positivity was defined as a receptor overexpression score of 3 + on immunohistochemical analysis [20]. The Van Nuys classification system was used for DCIS grade, and final postoperative pathological results were classified using the TNM classification.

Surgical procedure
The breast (partial or total mastectomy) and axillary lymph node (none, SLNB, or axillary lymph node dissection) surgical procedures were collected.

Adjuvant treatments and follow-up
Adjuvant treatment, including endocrine therapy, radiotherapy, chemotherapy, and additional surgery, were collected from medical records. The recurrence status was also assessed.

Statistical analysis
Preoperative clinicopathological findings were extracted to determine their association with a postoperative diagnosis upstaging from DCIS to IDC, and logistic regression analysis was used to assess the factors. Variables with a p value < 0.0001 in the univariate analysis were included in the multivariate analysis, and a p value ≤ 0.05 was considered statistically significant. Statistical analyses were performed using JMP ® 12.1 (SAS Institute Inc., Cary, NC, USA).
We considered the relationship between tumor size, including non-mass enhancement on MRI, plus preoperative clinicopathological factors and the possibility of postoperative upstaging to IDC on postoperative specimens. The possibility of postoperative upstaging to IDC was calculated by the number of patients who were preoperatively diagnosed with DCIS as the denominator, and the number 1 3 of patients postoperatively diagnosed with IDC cancer as the numerator. First, a graph was created by setting the upper limit of the tumor diameter at 5-mm intervals and calculating the ratio of postoperative upstaging to IDC using the dynamic MRI tumor diameter data. Next, in patients with data on dynamic MRI tumor diameter and preoperative clinicopathological factors, the possibility of postoperative upstaging to IDC was calculated in the same manner.

Results
We identified 2,317 patients diagnosed with DCIS through preoperative biopsies. Among the patients, postoperative diagnosis was special type (mucinous carcinoma) in 2 patients, lobular carcinoma in situ in 5 patients, invasive lobular carcinoma in 2 patients, benign tumor in 8 patients, and no postoperative report in 7 patients. Therefore, excluding these 24 patients, further analysis was performed in a total of 2,293 patients whose final pathological results were DCIS and IDC.
The axillary operation methods were SLNB in 1,807 patients (78.8%), axillary lymph node dissection in 105 (4.6%), omission in 258 (11.3%), and no data in 123 (5.4%). Ninety-seven patients (4.2%) had lymph node metastasis, including 16 (Fig. 1). When a threshold of tumor size, including non-mass enhancement, on MRI was ≥ 50 mm, the possibility of postoperative upstaging to IDC was almost the same. Among patients with a tumor size on MRI of ≤ 20 mm, the possibility of postoperative upstaging to IDC was 22.1%. Among the 258 patients with non-palpable mass, NG1/2, ER-positive DCIS, and detailed tumor diameter data on dynamic MRI, the possibility was 10.8% when the tumor size, including non-mass enhancement, on MRI was ≤ 20 mm. In addition, the possibility was 18.1%, even when the upper limit of the tumor size on MRI was raised to ≤ 40 mm.
In previous meta-analyses [13], the tumor size was usually measured by MMG and MRI, or US only if impossible to  [25] reported that approximately 40% of DCIS were MRI-only detected lesions, while some later studies [26,27] have shown the superiority of MRI over MMG for the detection of DCIS (sensitivity of 92% versus 56%, respectively), as well as for the determination of the spread of the disease. On the other hand, one of the weak points of dynamic MRI is overdiagnosis due to background parenchymal enhancement. In daily practice, a dynamic MRI for preoperative assessment in patients with diagnosed DCIS is not routinely performed. Indeed, Roozendaal et al. [28] showed that 409 of 910 patients with preoperatively diagnosed DCIS in four institutions in the Netherlands underwent MRI (average 44.9%; range 5.7-68.2%), while dynamic MRI is more frequently performed in Japan. In the current study, 1,538 of 2,317 patients (66.1%) with preoperatively diagnosed DCIS in 16 institutions underwent MRI (range 26.7-100.0%). We demonstrated a larger threshold of tumor size, including non-mass enhancement on MRI, is associated with an increased possibility of postoperative upstaging to IDC (Fig. 1). The combined use of various IDC predictors to select patients reduces the possibility of postoperative upstaging to IDC, even if the tumor size on MRI is more than 20 mm. Thus, we concluded that dynamic MRI and clinicopathological factors could assist not only with the identification of the extent of resection but also in predicting the possibility of IDC for patients with preoperatively diagnosed DCIS through biopsy to determine the appropriate surgical procedure. This study has several limitations. First, this is a retrospective study, and there were some missing data of clinical, radiological, and histological variables. However, this study remains one of the largest studies of retrospectively collected data across 16 institutions. Second, the follow-up period was short.
Four active surveillance clinical trials for low-risk DCIS have commenced in the United Kingdom (LORIS), Europe (LORD), United States (COMET), and Japan (JCOG1505, LORETTA trial) [4][5][6][7][8]. These studies are non-inferiority prospective trials to examine the effectiveness and safety of active surveillance compared to surgical based treatment approaches for low-risk DCIS patients, and each of these studies specifies low-risk DCIS with multiple factors. These studies will be important in prospective validation of prognostic factors.

Conclusion
In conclusion, we identified the following four independent clinicopathological predictive factors of postoperative upstaging to IDC among patients with DCIS diagnosed by biopsy in this retrospective study: presence of a palpable mass, MMG findings (≥ category 4), mass formations on US, and tumor size on MRI (> 20 mm). The combined use of various predictors of IDC reduces the possibility of postoperative upstaging to IDC, even if the tumor size on MRI is larger than20 mm. Thus, we consider that the eligibility criteria of prospective study (JCOG1505) are appropriate. In addition, we could also consider the omission of SLNB among patients with low risk of postoperative upstaging to IDC using the four predictive factors.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required.
Informed consent Informed consent was obtained from all individual participants included in the study by opt-out.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.