1 INTRODUCTION

Breast cancer has replaced lung cancer as the most newly diagnosed cancer worldwide, and it is the second leading cause of cancer-related death in women.¹ With the advancement of diagnosis and treatment technology, the survival rate of early breast cancer has climbed steadily in recent years. Unfortunately, some early breast cancers recur and metastasize in a short time, which complicates the clinical management of the disease.

Among these early breast cancers, ductal carcinoma in situ with microinvasion (DCIS-MI) is defined as ductal carcinoma in situ (DCIS) with a microscopic invasive focus ≤1 mm in the longest diameter.² DCIS-MI accounts for approximately 1% of all breast cancer cases and 5%–10% of DCIS cases.³ With respect to epidemiology, the median age of women with DCIS and DCIS-MI is 59, which is significantly different from T1a invasive ductal carcinoma (T1a-IDC) (62 years).⁴ White is still the highest risk race (67.9%), black, Asian or Pacific Islander, Hispanic, and American Indian/Alaska Native account for 11.1%, 10.7%, 9.6%, and 0.4%, respectively.⁴

According to the American Joint Committee on Cancer (AJCC) staging system, DCIS-MI belongs to the earliest T staging and should be classified as the earliest invasive breast cancer. Due to the low incidence of DCIS-MI and the lack of prospective clinical studies, there is currently no guideline or consensus on its diagnosis, treatment, and prognosis. Some scholars believe that DCIS-MI is a transitional stage in the development of DCIS to IDC^{5, 6} and that recognition of the changes between DCIS and DCIS-MI will help clinicians to understand the progression of breast cancer and provide individualized treatment for patients with DCIS-MI.

Fortunately, a review⁷ in recent years described relevant literature regarding the characteristics, diagnosis, treatment, and prognosis of DCIS-MI, particularly with regard to surgical options (local treatment and management of axillary lymph nodes).

Inspired by it, the author updated some representative and newly published research to strive for accuracy and timeliness. Furthermore, we tried to explore the relationship between adverse prognostic factors and treatment of DCIS-MI, and combined it with National Comprehensive Cancer Network (NCCN) guidelines to provide a reference for clinicians when making treatment decisions.

2 LITERATURE SEARCH

We searched PubMed for relevant articles covering the period of 1982 to 2021 using the following terms by MeSH and free-word: breast cancer, microinvasion, DCIS, DCIS-MI, and IDC. Valid publications included full-text studies in English involving women diagnosed with primary DCIS, DCIS-MI, and IDC. It is worth noting that we select relevant literature on microinvasive breast carcinoma (MIBC) to illustrate the relationship between MIBC and DCIS-MI. Animal studies, conference abstracts, case reports/case series, commentaries, and letters to the editor were excluded.

3 EVOLUTION OF THE DEFINITION OF MIBC AND DCIS-MI

Early in 1982, Lagios introduced the term “microinvasion” in breast pathology as synonymous with invasion less than 1 mm,⁸ but this term has never been applied in a consistent, standardized manner. The fifth edition of the AJCC staging system published in 1997⁹ was the first staging system that recognized a specific T stage for microinvasive carcinoma (MC), defined as “the extension of cancer cells beyond the basement membrane into the adjacent tissues with no focus more than 0.1 cm in greatest dimension” and formally reported it as pT1mic. The AJCC staging system further pointed out that “when there are multiple foci of microinvasion, the size of only the largest focus is used to classify the microinvasion, and the size of the individual foci should not be added together.” The seventh edition of the AJCC staging system published in 2010 rounded the tumor size in mm but did not clearly define the maximum diameter of invasive cancer. They instead gave a range of 1.1 mm to 1.4 mm, which caused confusion in clinical staging and affected the evaluation of MC. The eighth edition of the AJCC staging system published in 2016 was improved to clarify that the maximum diameter of MC was ≤1.0 mm and that the maximum diameters of the foci >1.0 mm and <2.0 mm were characterized as 2.0 mm and classified as pT1a to avoid underestimation of tumor stage.

The fifth edition of the Word Health of Organization (WHO) Classification of Tumors published in 2019 pointed out that microinvasive carcinoma is the earliest morphologically recognized form of invasive carcinoma with a dominant carcinoma in situ, generally DCIS of high-grade, but sometimes also lower-grade DCIS, lobular carcinoma in situ (LCIS) or Paget disease of the nipple. In 2011, Ross et al.¹⁰ studied 75,250 patients with breast cancer and found that only 16 cases (0.02%) were confirmed to be lobular carcinoma in situ with microinvasion (LCIS-MI). All of these cases had no recurrence or metastasis, and all survived at follow-up (median of 24 months, range of 1–72 months).

In summary, the diagnostic criteria for MIBC have been standardized, but those few clinical studies that do exist on DCIS-MI did not clearly distinguish between DCIS-MI, LCIS-MI, and Paget disease with microinvasion and collectively termed all of them together as “MIBC.” This review focuses on DCIS-MI since it is the most common type above them and has the worst prognosis.

4 CLINICOPATHOLOGICAL CHARACTERISTICS OF DCIS-MI

Most of the current research^{4, 5, 11-19} show that DCIS-MI tends to express more aggressive pathological features, such as necrosis, human epidermal growth factor receptor 2 (HER2) overexpression, and estrogen receptor (ER) or progesterone receptor (PR) negative, compared to DCIS and even IDC in some indicators. Costarelli et al.¹⁹ examined 17,431 cases of ductal carcinoma treated between 2011 and 2016 by 10 Italian breast units and classified 12.1% (2,107) as DCIS, 1.3% (233) as DCIS-MI, and 86.6% (15,091) as IDC. The mean age at diagnosis did not differ between DCIS (57.8 ± 12.2 years) and DCIS-MI (56.4 ± 13.0 years) but was higher in IDC patients (60.6 ± 13.7 years) (p < 0.001). Additionally, the mean tumor diameter was significantly higher for DCIS-MI (28.2 ± 17.0 mm) compared to DCIS (19.4 ± 12.0 mm) and IDC (22.0 ± 16.0 mm) (p < 0.05). In the aspect of lymph node status, 12.2% (25) of DCIS-MI cases had macrometastasis compared to 36.6% (5132) of IDC cases (p < 0.001). The presence of necrosis was more frequent with DCIS-MI (51.3%; 61/119) than with DCIS (32.9%; 301/914) or IDC (21.3%; 802/3758) (p < 0.01), and having a high tumor grade of the infiltrating component was also more frequent with DCIS-MI (49%) than with IC (28.4%) (p < 0.001). Finally, for ER, PR, HER2, and Ki67, the results for the DCIS-MI group had statistical differences compared to DCIS and IDC (Table 1).

While a few studies have suggested different conclusions from the above.^{20, 21} A study from The Fourth Hospital of Hebei Medical University²⁰ analyzed 1,886 patients with ductal carcinoma and found that the expression of ER and PR showed no statistical differences between groups of DCIS, DCIS-MI, and T1-IDC patients (p = 0.079, 0.150), and the expression of Ki67 (>20%) showed no statistical differences between groups of DCIS, DCIS-MI, and T1a-IDC, but was significantly lower in these groups than in T1b or T1c patients (p < 0.05).

Regarding axillary lymph node status, a meta-analysis²² of sentinel lymph node (SLN) biopsies in MIBC involving 968 patients with MIBC from 24 studies, showed that positive SLN cases accounted for 10.1% of the total. This conclusion is similar to a study from Champion et al.,⁴ whose positive lymph node cases accounted for 9.68% of DCIS-MI patients and 6.48% of T1a-IDC cases. However, in the SEER Database, from which the study of Champion et al.⁴ selected patients, any DCIS patient with positive lymph nodes is recorded as an “invasive” tumor, and it is uncertain whether these patients were included in the DCIS-MI or T1a-IDC cohorts.

5 ADVERSE BIOMARKERS AND TREATMENT OF DCIS-MI

6 PROGNOSIS OF DCIS-MI

Overall, patients with DCIS-MI have a good prognosis. A retrospective analysis involving 1299 patients with postoperative DCIS-MI from 2008 to 2019 was conducted by Fudan University.³² The proportions of patients who received radiotherapy, chemotherapy, and targeted therapy were 12.4%, 43.9%, and 17.5%, respectively, and 92% of HR+ patients received endocrine therapy. After a median follow-up time of 54.57 months, the 5-year local recurrence-free survival (LRFS), distant metastasis-free survival (DMFS), and OS were 98.6%, 97.1%, and 99.4%, respectively.

However, there are still conflicting results about the prognosis of DCIS-MI compared to DCIS and T1a-IDC. Of note, a substantial proportion of studies consider DCIS-MI to have a worse prognosis than DCIS, or one similar to T1a-IDC, including several large retrospective analyses. For example, a study from Champion et al.⁴ screened 134,569 postoperative non-metastatic breast cancer cases with stage ≤ T1a from the SEER database between 2004 and 2015, aged 18–90 years, where 3.2% (4361) were classified as DCIS-MI, 70.9% (95,393) as DCIS, and 25.9% (34,815) as T1a-IDC, and where 10.9% of DCIS-MI and 11.8% of T1a-IDC received chemotherapy (no information provided on endocrine therapy and targeted therapy). After a median follow-up time of 66 months, the results showed that after adjustment for age, race, and other cancer histories, the CSS of DCIS-MI differed significantly from the other two cohorts: worse than DCIS (p < 0.001) but better than T1a-IDC (p = 0.03). Adjusted OS was not significantly different between DCIS-MI and T1a-IDC (p = 0.66) and was better for DCIS than for DCIS-MI (p < 0.001). The study also compared the DCIS-MI and T1a-IDC cohorts using the latest prognostic staging in the 8th edition of the AJCC and found that the 5-year survival of patients with stage IA in both groups was similar (the CSS was 99% for both groups and the OS was 94% for both). Although the CSS in stage IB patients was slightly different from IA patients (p = 0.002), there was no significant difference in OS (p = 0.12). Besides, there was also no significant difference in CSS or OS in stage II/III between the two groups (p > 0.05).

These findings are consistent with two previous large studies that evaluated patients with DCIS-MI in the SEER database.^{33, 34} In this study that featured a large cohort of 87,695 DCIS and 8863 DCIS-MI patients, DCIS-MI had worse CSS (HR = 2.475, p < 0.001) and OS (HR = 1.263, p < 0.001)³³ than DCIS. In the other large cohort study of 525,395 patients with node-negative invasive breast cancer with DCIS or small tumors (≤2 cm), the actuarial rates of breast cancer mortality at 20 years were 3.8% for DCIS, 6.9% for DCIS-MI and 6.8% for DCIS-IDC.³⁴

In contrast, several other studies have suggested that the prognosis of DCIS-MI is not significantly different from that of DCIS.^{14, 18, 20, 24, 32, 35-40} A summary of some studies with large sample sizes is shown in Table 3.

7 DISCUSSION

MIBC is usually seen in DCIS of high grade but sometimes also in lower-grade DCIS, LCIS, or Paget disease of the nipple. Except for DCIS-MI, other types of MIBC are rare and have not been found to have poor prognoses. DCIS-MI, which has a low incidence (approximately 1% of all breast cancer cases and 5%–10% of DCIS cases), is generally considered to have a good prognosis. However, some patients are prone to relapse and metastasis, which brings confusion to clinical management. It is necessary to further summarize and study, providing a reference for clinicians to choose treatments and follow-up after surgery.

According to our summary, DCIS-MI tends to express more aggressive pathological features than DCIS and IDC, such as necrosis, HER2 overexpression, ER or PR negativity and high nuclear grade. Ki67 overexpression in DCIS-MI is similar to that of IDC but significantly higher than DCIS and the rate of lymph node metastasis of DCIS-MI is similar to that of T1a-IDC. These features may suggest that the biological behavior of DCIS-MI is worse than that of DCIS and similar to early IDC.

The overall prognosis of DCIS-MI is typically good, but there are also conflicting results on the differences between the prognoses of DCIS-MI and DCIS and T1a-IDC. Interestingly, the study from Champion et al.⁴ compared DCIS-MI and T1a-IDC cohorts using the latest prognostic staging in the 8th edition of the AJCC and found similar survival outcomes for both. Therefore, exploring the prognosis and adverse prognostic factors of DCIS-MI may be very useful for guiding treatment.

Our summary also found that some clinicopathological features and biological indicators, such as young age, HR-, HER2+, and multimicroinvasive lesions, were associated with worse prognoses of DCIS-MI. Other risk factors, such as larger tumor diameter, metastatic lymph nodes, and high Ki67 expression, which have been found to be associated with worse breast cancer prognoses in previous studies, have not been shown to have similar effects on DCIS-MI by any existing current retrospective studies, although in some cases the results of such studies have been inconsistent. The possible reasons for this difference are as follows. First, the pathological characteristics of DCIS-MI are different from DCIS and IDC in many aspects, which may make its biological behavior and prognosis have independent characteristics. Second, DCIS-MI is a cohort that lies outside the inclusion criteria of most clinical studies, and breast cancer-related prognostic factors may not be fully applicable to this population. Third, the natural history and the exploration of prognostic factors may be influenced by referring to existing results that are then applied to DCIS-MI patients. For example, DCIS-MI patients who are lymph node-positive are more likely to receive intensive adjuvant therapy. Therefore, prospective studies are needed in the future.

The prognostic factor most worthy of attention is the relationship between lymph node metastasis and the number of microinvasion foci. The study from Ruijin Hospital¹⁷ found significant differences in lymph node involvement rates in DCIS-MI with different numbers of microinvasion lesions (2.1% for one focus and 15.6% for multiple foci, p = 0.037), and the univariate analysis of another study²⁵ also suggested that patients with multiple invasive foci had a higher rate of lymphatic invasion (p = 0.034). If it is hypothesized that the presence of multifocality of DCIS-MI may represent the driving force of DCIS to penetrate the basement membrane and that metastasis to axillary lymph nodes may represent the ability of cancer to invade, but the inconsistent prognostic between multifocality and lymph node metastasis does not support this hypothesis. Moreover, compared to the lymph-node-negative patients with DCIS-MI, there are no relevant research results for reference on whether lymph-node-positive patients have the clinicopathological characteristics of poor prognosis. However, even if lymph node positivity is associated with other poor prognostic factors, it has not been shown to have an independent prognostic value. Therefore, further analysis of prognostic correlations related to axillary lymph node status in DCIS-MI patients is warranted, especially in patients with multiple microinvasions.

Regarding the treatment of DCIS-MI, adjuvant endocrine therapy for HR+ patients has been shown to provide survival benefits. But various clinical studies do not agree on the chemotherapy or targeted therapy of patients with DCIS-MI. Because it is classified into the earliest T stage and belongs to the group of invasive cancers, the NCCN guidelines for treatment recommendations for early stage invasive breast cancer can be used as a reference.

To this end, patients with N+ are recommended to receive adjuvant chemotherapy combined with anti-HER2 therapy (if HER2+). Among them, HR+ and HER2- patients can be assessed for recurrence risk based on relevant genetic testing and given individual therapy.

Additionally, for patients with N0, treatment needs to be decided according to different subtypes. In our summary, we found that chemotherapy or targeted therapy was more inclined to be used in HR- and/or HER2+ patients because of their worse outcomes. In the exploration of adverse prognostic factors, the mean survival for DFS of HER2+ DCIS-MI patients without chemotherapy and targeted therapy was 42.6 months, whereas HER2- patients were 70 months (p = 0.019)¹⁷; and ER-/PR- patients had cumulative survival of 79.3% compared with 97.5% for ER+ or PR+ patients.²⁹

In NCCN guidelines, for HR- patients, if they are HER2- and the tumor size ≤ 0.5 cm, adjuvant chemotherapy is not required, but this treatment is considered appropriate for high-risk groups (such as young patients with high-grade cancer). The remaining HR- patients are recommended to receive adjuvant chemotherapy combined with anti-HER2 therapy (if HER2+). Similarly, in baseline characteristics of the nomogram constructed by Zheng et al.³⁰ N0 accounted for 92.6%, which provides a theoretical basis for the universality of the existence of chemotherapy benefits in DCIS-MI patients with HR- and N0. Furthermore, a large meta-analysis⁴¹ which evaluated the efficacy of trastuzumab in patients with HER2+, ≤2 cm tumors found that anti-HER2 therapy had a significant benefit in the HR- population, whose 8-year cumulative recurrence rate decreased from 33.4% to 24% (p < 0.001) and 8-year cumulative mortality rate decreased from 21.2% to 12.4% (p < 0.001).

For HR + patients, conversely, chemotherapy and targeted therapy will be put behind under certain conditions due to the clinical benefits of endocrine therapy. For example, the NCCN guidelines point out that if the tumor size borders on T1mic, the estimated recurrence risk is less than 5% and the absolute benefit of HER2-based systemic therapy may be negligible. The guidelines further suggest that the prognosis of patients with pT1a and pT1b tumors that are pN0 is uncertain even when HER2 is amplified or overexpressed. The decision to use trastuzumab in this patient cohort must balance the toxicity and the uncertain, absolute benefits. Similarly, the above meta-analysis⁴¹ also found the efficacy of trastuzumab in patients with HR+ and ≤1 positive lymph node, however, because their untreated outcomes, whose 8-year cumulative recurrence rate was 19.4% and the 8-year cumulative mortality rate was 7.4%, was better than treated outcomes of patients with HR- and ≤1 positive lymph node, less aggressive treatment approaches ought to be assessed. Treatments with less toxicity, such as weekly paclitaxel and trastuzumab therapy, can be considered.

In conclusion, for the adjuvant treatment of DCIS-MI patients, clinicians need to give priority to the hormone receptor status, HER2 expression, and axillary lymph node status of patients, because these may affect the prognosis and treatment response.

At present, our understanding of DCIS-MI is still limited, and basic research is still needed to reveal its pathogenesis, evolution process, and tumor microenvironment, among other characteristics, so as to broaden research horizons and provide ideas for clinical diagnosis and treatment. Clinical prospective studies are also necessary to explore its prognosis and treatment options to help guide clinicians in making individualized treatment decisions.

AUTHOR CONTRIBUTIONS

Ge Song collected the literature and wrote this review, Yongqiang Zhang guided the revision of the structure and content of this review.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

ETHICS STATEMENT

None.