DCIS is characterized by malignant epithelial cells within the breast ducts without invasion or breach of the myoepithelial layer or basement membrane [10]. Histopathologic features of DCIS include both architectural subtype and nuclear grade and vary widely, giving way to numerous classification systems, none of which is uniformly accepted [10, 11]. The architectural subtypes include comedo, cribriform, micropapillary, solid, and mixed. Comedo lesions, for example, represent plugged ducts with atypical cells and central necrosis. These subtypes may impart information on the nature of DCIS. In a study of 130 confirmed DCIS cases, Bellamy et al. [12] found that comedo DCIS had a higher rate of local recurrence than did the other DCIS types. Papillary DCIS, on the other hand, was more often multicentric, involving multiple quadrants of the breast. Ajisaka et al. [13] reviewed 24 cases of comedo DCIS and concluded that lesions containing comedo or cribriform elements were more likely than other types of DCIS to have microscopic spread. Challenges exist because there is much architectural heterogeneity among the types of DCIS, which inhibits reproducibility [14, 15]. Silverstein [16] argues that noncomedo lesions may express biologic markers, such as high proliferation rate and ErbB-2 (also known as HER2/neu) gene amplification, similar to comedo lesions, and may behave like high-grade comedo lesions.

There is a growing accord favoring nuclear grading over morphologic analysis [10, 11]. Nuclear grading involves primarily distinguishing nuclear atypia and mitotic activity, and the grades are categorized as low, intermediate, and high [17]. In fact, the College of American Pathologists has recognized that nuclear grade, the presence of necrosis, and the distance from margins are the most important determinants in assessing the likelihood of recurrence [18]. However, challenges again exist because of both unclear diagnostic criteria and intralesional heterogeneity, thus resulting in moderate-to-high interob-server variability [19, 20]. In 2015, Elmore et al. [21] conducted an investigation showing a low level of concordance between a patholo-gist's diagnosis and consensus-derived reference diagnosis for DCIS and atypia.

As already mentioned, the natural history of DCIS remains poorly understood. Many studies have supported the idea that DCIS is a precursor lesion with progression to invasive breast cancer in up to 53% of lesions untreated. Rosen et al. [5] performed a retrospective study of approximately 8000 biopsies with 30 cases of DCIS initially misdiagnosed as benign. Ultimately, at follow-up, 53% of those cases were found to have subsequent invasive carcinoma [5]. Likewise, Collins et al. [6] examined a total of 1877 cases, including 13 cases of DCIS initially misdiag-nosed as benign, with 46% resulting in subsequent invasive cancer. Another argument for the theory of DCIS as a precursor lesion stems from studies analyzing the risk of recurrence of DCIS as an invasive cancer after excision, with 50% of recurrences diagnosed as invasive cancers [22]. Boyages et al. [23] performed a meta-analysis examining treatment factors as predictors for local recurrence. The authors found a recurrence rate of 22.5% (95% CI, 16.9–28.2%) for treatment with breast-conversing surgery alone, 8.9% (95% CI, 6.8–11.0%) for combination breast-conserving surgery and radiation therapy, and 1.4% (95% CI, 0.7–2.1%) for mastectomy alone.

We know, however, that many cases of DCIS do not progress to invasive carcinoma in one's lifetime. Multiple autopsy studies exist showing that up to 14.7% of women have undetected DCIS [24]. Interestingly, Ozanne et al. [25] conducted an analysis of frequency trends, challenging the long-held theory. The authors noted that a decrease in frequency of invasive breast cancer should be expected if DCIS is a precursor of invasive disease because the frequency of diagnosis and treatment of DCIS have increased. However, the frequency of invasive breast cancer is increasing [25, 26]. Therefore, DCIS is most currently viewed as a nonobligate precursor to invasive disease [4].

What then causes progression to invasive cancer? There are molecular and histochemical similarities between DCIS and invasive breast cancer, specifically within the same histopathologic grade. This lends to the theory that the progression of DCIS to invasive cancer is the result of an accumulation of genetic abnormalities within two parallel pathways [27–29]. Multiple studies have found shared changes between low-grade DCIS and low-grade invasive carcinomas, often exhibiting estrogen receptor positivity. High-grade DCIS has been linked to high-grade invasive cancers, often exhibiting ErbB-2 positivity [22]. In addition, researchers have found microenvironmental factors to play a role in the progression of DCIS to invasive breast cancer, implicating alterations of myoepithelial and stromal cells [30–32].

No factor or combination of factors, however, has been shown to accurately predict the progression of DCIS. Erbas et al. [22] discussed limitations in the current data and concluded that exploration of prognostic determinants and models to better predict or estimate progression of DCIS are needed.

The risk factors for DCIS are similar to those of invasive breast cancer, with age playing a strong role. There is an increase in incidence of DCIS with older age, with a significant increase among women older than 50 years and a peak incidence of 96.7 cases per 100,000 women 65–69 years old [33–36]. DCIS was found to be uncommon in women younger than 30 years. A meta-analysis performed by Kerlikowske et al. [37] involving a cross-sectional study of 39,542 women with abnormal mammograms and subsequent follow-up showed increased risk of DCIS in those with a family history of breast cancer and, specifically, nulliparity among women 50 years or older. Claus et al. [38] conducted a large population-based case control study of all female patients diagnosed with DCIS in Connecticut between 1994 and 1998 and found a strong association between a history of breast biopsy and DCIS diagnosis. The authors concluded that this may be because of both increased surveillance and benign breast diseases that progressed to DCIS. There is a lack of evidence at this time to suggest associations between oral contraceptives and smoking with the risk of DCIS [35, 38–41]. There are mixed data on the association between hormone replacement therapy and DCIS [42].

According to the surveillance research from the American Cancer Society [43], the incidence rates of DCIS in the United States between 2007 and 2011 for non-Hispanic white and non-Hispanic black women were similar (26.6 and 25.6 cases per 100,000 women, respectively). Rates were lower among Hispanic women (17.9 cases per 100,000 women) and lowest for American Indian and Alaska Native women (14.4 cases per 100,000 women).

Risk factors for recurrence of DCIS have also been evaluated. An analysis of the European Organization for Research and Treatment of Cancer Trial 10853 performed by Bijker et al. [44] found multiple factors associated with an increased risk of local recurrence. These included young age (hazard ratio = 2.14; p = 0.2), involved margins (hazard ratio = 2.07; p = 0.0008), and treatment with local excision alone (hazard ratio = 1.74; p = 0.009). In 2016, Cronin et al. [45] conducted a meta-analysis of 2996 cases of DCIS and found that the risk of recurrence decreases with increasing age. The 10-year invasive recurrence rate for women younger than 40 years was 16%, versus 6.5% for women older than 40 years. In fact, 21% of patients younger than 40 years were found to have distant metastasis. Other risk factors involve histopathologic parameters, as discussed previously.

DCIS is primarily diagnosed at imaging because it is most often clinically occult. Most DCIS lesions found at mammography present as microcalcifications, with approximately 75% of lesions presenting only as calcifications [46] (Fig. 1). Up to 23% of DCIS may present as a mass or asymmetry, and roughly 12% are associated with a palpable abnormality [46, 47] (Fig. 2). Hence, the increased frequency of DCIS diagnosis is largely due to screening. Given that DCIS may never reach clinical significance over a woman's lifetime, critics argue that screening mammography may induce more harm than good, citing overdiagnosis, overtreatment, and increased radiation. On the other hand, because there is a lack of understanding of DCIS and its natural progression, proponents argue that limiting screening because of overdiagnosis of DCIS will cause undue harm. Numerous randomized controlled trials have found reduced mortality with screening, citing decreases in mortality in up to 32% of patients [48, 49]. The Gothenburg Breast Screening Trial, for example, found a significant 21% reduction in breast cancer–related mortality among 51,611 women [49]. There is, however, much debate over whether mammography is associated with significant decrease in mortality in patients 40–49 years old [50–53].

With regard to DCIS, numerous studies have found support for the use of screening mammography because of increased detection of high-grade DCIS lesions, detection of associated invasive cancer, or the possible subsequent prevention of invasive breast cancer. Weigel et al. [54] found that higher DCIS detection rates were strongly associated with intermediate- and high-grade DCIS (i.e., higher risk and rate of invasive progression) when compared with low-grade DCIS. Of the total 1074 screening-detected DCIS cases, the authors found that 432 (40.2%) cases were high grade. The authors concluded that the detection of low-grade DCIS, which are presumed to contribute to overdiagnosis, has a lesser effect on rates of invasive progression. In 2016, Duffy et al. [55] conducted a retrospective population-based study that found a statistically significant negative association between screening detections of DCIS and subsequent invasive cancer frequency. Evans et al. [56] found an increased rate of small grade 3 invasive tumors when DCIS was detected at screening mammography as linear or branching microcalcifications. The authors concluded that detection of calcification suggestive of DCIS should remain a part of mammographic screening.

Many studies have attempted to associate morphologic features and distribution of calcifications with pathologic findings of DCIS. Evans et al. [57] found an association between calcification (p < 0.001), calcification in a ductal distribution (p < 0.005), and rodlike calcification (p < 0.001) with necrotic DCIS. The authors conclude that radiologic characteristics may help identify necrotic DCIS, which is known to be more biologically aggressive than other types of DCIS. Other studies have shown fine linear or linear branching calcifications to be associated with high-grade DCIS [58]. In addition to clinical and molecular markers, perhaps a movement to identify additional radiographic findings to help better predict the progression of DCIS to invasive breast cancer is warranted.

At this time, mammography screening is well accepted and recommended in most Western countries, with attendance rates as high as 70% [59, 60] (Table 1). The American College of Radiology and Society of Breast Imaging recommend annual screening mammography at age 40 years for individuals at average risk [61]. The U.S. Preventive Services Task Force recently published their breast cancer screening guidelines in 2016, recommending biennial screening mammography for women aged 50–74 years [62]. For women 40–49 years old, the U.S. Preventive Services Task Force recommends a more individualized approach, urging individuals to weigh benefits against harm.

Ultrasound is an important modality that has many benefits in the detection and workup of DCIS. Given that DCIS has a variable appearance on imaging, ultrasound has been found to increase the specificity of mammography [63]. Ultrasound can be particularly useful in evaluation of an associated mass when calcifications are seen at mammography (Fig. 3). An associated mass may help indicate invasion [64]. Furthermore, visualization of calcifications at ultrasound can help direct biopsy. In a study of 111 lesions characterized as suspicious clusters of microcalcifications at mammography, Soo et al. [65] identified 26 lesions (23%) sonographically that were successfully biopsied. These lesions were more frequently found to be malignant than those not seen on ultrasound. Noncalcified DCIS may be detected sonographically in patients presenting with nipple discharge or in palpable lesions that are not seen on a mammogram [63].

MRI may help with treatment decision making; however, its role has not been clearly delineated [4]. DCIS most commonly presents as a ductal or linear clumped nonmasslike enhancement. MRI has been found to be useful in conjunction with other modalities for pre-operative evaluation, specifically in detection of residual or multicentric disease [66]. Although MRI may help detect mammographically occult DCIS lesions, this modality has been found to overestimate the extent of DCIS [67, 68]. The Comparative Effectiveness of MRI in Breast Cancer trial by Turnbull et al. [69] found no reduction in reexcision rates after lumpectomy and preoperative MRI. There is mixed evidence at this time that enhancement, morphologic characteristics, and kinetics may be useful predictors of differing grades of DCIS, given that the kinetic characteristics appear to be heterogeneous [70–72]. Nadrljanski et al. [73] found that, although nonmass lesions with focal or segmental distribution and plateau enhancement curves were frequently found with DCIS, there was no statistically significant difference in the frequency of enhancement or kinetic patterns of DCIS. Kim et al. [74] found no statistical difference between kinetic characteristics and histopathologic factors of DCIS.

The National Comprehensive Cancer Network has published guidelines pertaining to DCIS treatment. Currently, treatment of DCIS largely involves breast-conserving surgery, excising all disease with negative margins [75]. Mastectomy is reserved for cases with multifocal or extensive disease or patient decision. There has been no difference in overall survival between mastectomy and breast-conserving surgery for the treatment of DCIS [76]. Sentinel lymph node biopsy is not typically performed unless the patient is at high risk, if mastectomy is chosen, or a mass is identified at imaging [28, 77, 78]. Approximately two-thirds of patients who receive a diagnosis of DCIS will pursue breast-conserving surgery [4].

Adjuvant therapy involving radiation therapy is controversial [79]. Although much of the standard of care currently involves radiation therapy after breast-conserving surgery, several studies have questioned the need for radiation therapy in low-risk groups. Multiple randomized controlled trials have supported the benefit of radiation therapy after breast-conserving surgery. The National Surgical Adjuvant Breast and Bowel Project (NSABP) evaluated ipsilateral breast cancer recurrence in patients with DCIS after lumpectomy alone and lumpectomy with radiation therapy. The authors found that radiation therapy reduced ipsilateral breast tumor recurrence by 52% in the lumpectomy and radiation therapy group compared with lumpectomy only (hazard ratio = 0.48; p < 0.001) [80]. A meta-analysis performed by Goodwin et al. [81] of four randomized controlled trials (European Organization for Research and Treatment of Cancer, NSABP, Swedish DCIS Trial, and UK Co-ordinating Committee on Cancer Research) ultimately concluded that there is a statistically significant benefit from the addition of radiotherapy after breast-conserving surgery on all ipsilateral breast events (hazard ratio = 0.49; 95% CI, 0.41–0.58; p < 0.00001).

The Radiation Therapy Oncology group 9804 trial was a prospective randomized clinical trial of 1790 patients with favorable DCIS (mammographically detected low- or intermediate-grade DCIS measuring < 2.5 cm) that closed early because of limited accrual. However, they found that, in approximately 636 patients with low-risk DCIS, the local failure rate was low with observation and significantly decreased with radiation therapy [82]. The Eastern Cooperative Oncology Group and North Central Cancer Treatment Group in 2009 selected low- to intermediate-grade DCIS cases and found a low rate of ipsilateral breast events (6.1%; 95% CI, 4.1–8.2%), suggesting that excision alone may be sufficient [83]. The 12-year results from the Eastern Cooperative Oncology Group–American College of Radiology Imaging Network study published in 2015 enrolled two study cohorts, one with lowto intermediate-grade DCIS and the other with high-grade DCIS, treated with surgical excision without radiotherapy [84]. The 12-year rates of developing an invasive recurrent were 7.5% and 13.4%, respectively (p = 0.08), without a plateau. These studies depict the ongoing controversy about adjuvant radiation therapy, thereby suggesting at this time that the decision to pursue radiation therapy remains an individualized one.

Likewise, adjuvant endocrine therapy has been investigated and is currently recommended on an individualized basis, particularly in patients with estrogen receptor–positive DCIS. Given the varied evidence for survival benefit, the National Comprehensive Cancer Network recommends consideration of tamoxifen for 5 years after breast-conserving therapy and radiation therapy. The NSABP B-24 trial evaluated the effect of adjuvant tamoxifen in addition to the treatment of DCIS with excision and radiation therapy [85]. The authors found a 37% reduction in relative risk of subsequent ipsi-lateral and contralateral cancer with adjuvant tamoxifen. The NSABP-35 trial evaluated anastrazole versus tamoxifen in postmenopausal women and concluded that anastrozole significantly improved the breast cancer–free interval in postmenopausal women younger than 60 years [86].

Given the biologic heterogeneity of DCIS and questions about overtreatment, there is much discussion about using active surveillance over surgical treatment of low- and intermediate-risk DCIS subsets. Multiple trials are currently ongoing, such as the United States–based Comparison of Operative vs Medical Endocrine Therapy for Low Risk DCIS (COMET) trial, the phase 3 Low Risk DCIS Trial (LORIS), and the Low Risk DCIS (LORD) trial performed by the European Organization for Research and Treatment Center [4] (Table 2). Numerous studies have been performed in the last year evaluating the eligibility criteria for the LORIS trial, which was initiated in 2014, calling for additional risk stratification. Pilewskie et al. [87] found that 20% of LORIS-eligible women had their diagnoses upgraded to invasive carcinoma at surgical excision.

In the past decade, several studies and clinical trials have shed light on prognostic factors and factor combinations that can help us delineate the risk of DCIS progression and ultimately dictate appropriate treatment (Table 3). The University of Southern California Van Nuys Prognostic Index was the first prognostic tool created. The Van Nuys Prognostic Index incorporates lesion size, margin width, pathologic classification, and patient age to stratify patients into three groups [88]. Each factor is assigned a score from 1 to 3, with 3 reflecting the worst prognosis. The Van Nuys Prognostic Index recommends excision alone for patients with total scores of 4–6, excision and radiation therapy for scores of 7–9, and mastectomy for scores of 10–12. However, there are limitations to this prognostic index. The Van Nuys Prognostic Index is based on a retrospective single-institution study that lacks independent validation. Furthermore, the index takes into account only the clinical and histopathologic characteristics [79]. Data to support the Nuys Prognostic Index are mixed, with some studies supporting it as a prognostic tool and others finding minimal clinical utility [89–91].

Nomograms have also been constructed as a tool to help predict the risk of recurrence after breast-conserving surgery. In particular, the Memorial Sloan Kettering Cancer Center nomogram takes into account approximately 10 factors, including age, family history, radiation therapy, and surgical excision to help predict the 5- and 10-year recurrence after surgical excision [92, 93]. However, this tool also does not take into consideration potential molecular markers.

The Oncotype DX DCIS score was created in 2013 to help predict 10-year risks of local recurrence after breast-conserving surgery alone. Solin et al. [94] used breast assay findings of 12 genes in the Eastern Cooperative Oncology Group E5194 trial and determined the subsequent development of invasive breast cancer. A score of 0–100 was assigned, and patients were stratified into three risk groups. The authors found that the DCIS score was statistically significantly associated with the risk of developing DCIS recurrence or invasive cancer (hazard ratio = 2.31; p = 0.02). The 10-year risks of developing a local recurrence in low-, moderate-, and high-risk groups were 10.6%, 26.7%, and 25.9%, respectively, and those for an invasive recurrence were 3.7%, 12.3%, and 19.2%, respectively. Numerous other articles are finding validation of this score [95]. However, Martínez-Pérez et al. [79] point out some limitations, including cost, logistical constrictions, and limited applicability to intermediate- and high-risk patients with DCIS.

Numerous studies have been or are currently under way to evaluate other markers, such as molecular phenotypes (e.g., luminal A, luminal B, ErbB-2 type, and triple negative) and biomarkers (e.g., Ki-67 and COX-2) [96–98]. Further studies must be done for validation.

Although DCIS has an excellent prognosis overall, there is much controversy regarding management of the lesion given its biologic heterogeneity. Our goal was to review the current research in the natural history of DCIS, the benefit of mammography screening, current treatment options, prognostic indexes, and new research in biomarkers and molecular phenotypes to predict progression of DCIS to invasive breast cancer. There is a general consensus that DCIS is a spectrum of disease and that many variables play a role in this progression. Currently, studies and trials are being undertaken to evaluate these factors, with huge potential implications for management. Perhaps it would be of merit to further evaluate radiographic characteristics to help identify prognostic data on DCIS at the time of detection.

In addition, numerous active clinical trials are on the horizon to evaluate appropriate treatment of DCIS. Does active surveillance have a role given that up to a half of diagnosed DCIS cases remain indolent over a patient's lifetime? Because there are limited data thus far to accurately detect progressive disease, early detection with screening mammography is key. Larger multiinstitutional randomized studies and trials need to be pursued to better delineate the prognostic factors in DCIS lesions.