Identifying patients with ductal carcinoma in situ who are more likely to develop invasive breast cancer remains a challenge, despite decades of research and the development of stratification methods to predict progression and recurrence.
Norman Wolmark, MD
Identifying patients with ductal carcinoma in situ (DCIS) who are more likely to develop invasive breast cancer remains a challenge, despite decades of research and the development of stratification methods to predict progression and recurrence, experts say. Two major areas of study are the appropriate use of active surveillance and the use of biomarkers observed in the DCIS microenvironment to determine risk.
Most patients with DCIS are treated with surgery, often combined with radiation and hormone therapy. This results in some patients being treated for low-grade precancerous lesions that probably would not develop into invasive cancers.1 At the same time, longterm follow-up of women with minimally treated DCIS showed that invasive cancers and distant metastases are significant concerns, and management of the condition remains controversial.2
Norman Wolmark, MD, chairman of the National Surgical Adjuvant Breast and Bowel Project (NSABP), laments that DCIS risk is often categorized somewhat arbitrarily. Favorable-risk patients are variously defined by characteristics such as a favorable pattern of microcalcifications, low nuclear grade, estrogen receptor (ER) positivity, and mature age, he said. High-risk patients are said to be ER negative, HER2 positive, of high nuclear grade, and/or younger and to require aggressive treatment.
Wolmark, who discussed DCIS during a presentation at the 36th Annual Miami Breast Conference® (MBCC), Friday, said in his conference abstract that the debate over the treatment of DCIS has been “driven largely by social media, the nonmedical press, and big data mining.” “A conspicuous absence from this perturbing dilemma is data from well-conducted, randomized, prospective clinical trials, “ he said.
Some clarity on grading may eventually be provided by 3 noninferiority trials that are comparing standard therapy with watchful waiting, also known as active surveillance, Wolmark said. Each aims to ultimately include about 900 patients. The oldest is the phase III randomized LORIS trial launched in 2014 by Cancer Research UK.3 Women aged 46 years and older who recently received a diagnosis of DCIS not classified as high grade are assigned to surgery or active monitoring with annual mammograms. Recruitment is scheduled to end in March 2020.
The COMET trial, funded by the Patient- Centered Outcomes Research Institute, opened its first study site in June 2017.4,5 It is enrolling women aged 40 and older with low-risk DCIS and randomizing them to surgery with or without radiation and endocrine therapy or active surveillance with an option for endocrine therapy. The primary outcome measure is new diagnoses of ipsilateral invasive breast cancer after 2 years, with completion planned for 2023. The similar LORD trial, sponsored by the European Organisation for Research and Treatment of Cancer, sets its endpoint further out, to 10 years.6 Accrual began in 2017, and the trial’s estimated completion date is December 2029.
Wolmark, who helped develop the Oncotype DX and Oncotype DX DCIS risk of recurrence assays, said he would have liked to see these data-collection efforts integrate genetic testing, as well. “The major weakness of these 3 trials is that risk is defined based on subjective morphologic interpretations of histologic characteristics and patient age,” he said in his abstract. “The lessons learned from applying genomic algorithms to define risk in invasive breast cancer have not been applied to DCIS. The failure to utilize the DCIS recurrence score [RS] genomic algorithm to define risk is an opportunity lost.”
He also pointed to NSABP B-43, a study comparing breast-conserving surgery, radiation therapy, and 2 infusions of trastuzumab (Herceptin) with breast-conserving surgery and radiation only for HER2-positive DCIS.7 The study, which has recruited more than 2000 patients, will also provide an estimate of risk for HER2-positive disease, he said.The utility of the Oncotype DX DCIS test to predict the risk of recurrence in women diagnosed with DCIS and treated only with lumpectomy has been repeatedly confirmed.8 Yet Stuart J. Schnitt, MD, chief of Breast Oncologic Pathology at Dana-Farber/ Brigham and Women’s Cancer Center in Boston, Massachusetts, said the test has had much less uptake than the Oncotype DX assay for invasive breast cancer. The DCIS test would be more useful if it also looked at biomarkers in the tissues surrounding the cancer, he said.
“Just like invasive breast cancers, DCIS doesn’t exist in isolation. It exists in an entire microenvironment that includes the myoepithelial cells that surround the ducts, blood vessels, immune cells, fibroblasts, and collagen, and there are a lot of data that suggest that alterations in these components may actually have a bearing on the risk of DCIS progressing to invasive breast cancer,” Schnitt said in an interview in advance of his MBCC presentation on the topic Friday.
“Tumor cells themselves can acquire attributes that kind of make them invasive and make them break out of the ducts. The other possibility is that there are factors in the microenvironment that lead to alterations that permit the DCIS cells to sort of invade or, once they invade, make it easier for them to invade the breast or spread to distant sites,” he said.
Schnitt focused on 3 types of tissue in the microenvironment: myoepithelial cells that surround the ducts with DCIS, stromal components that include fibroblasts and blood vessels, and tumor-infiltrating lymphocytes (TILs). “We’ve found that large numbers of tumor-infiltrating lymphocytes tend to be associated with what we view as more aggressive types of DCIS, like those that are HER2 positive and those that are triple-negative,” he said. “Large numbers of PD-L1—positive tumor-infiltrating lymphocytes are also associated with DCIS that seems to have a more aggressive phenotype.”
One recent analysis concluded that TIL density level can distinguish 2 biologically different DCIS subgroups, 1 of which (DCIS with ≥30% TILs) is biologically very similar to microinvasive carcinoma.9 TIL-rich pure DCIS lesions are also characterized by lower ratios of T-cell to B-cell TILs. Investigators said the potential for development of immunotherapy-based prevention of DCIS progression is worth further investigation. Another study found that dense TILs were an independent predictor of shorter recurrence- free interval (P = .002) in patients treated with breast conservation.10 Schnitt said that myoepithelial cells that are associated with DCIS show abnormalities, including upregulation of mRNA and proteins associated with tumor progression and downregulation of molecules associated with tumor suppression.11 “So, myoepithelial cells in at least some DCIS cases seem to be sick and may provide a permissive environment that allows the cells to invade,” he said.
Tumor-associated fibroblasts in the stroma have a promoting effect on tumor growth and invasion, he noted. Experimental data show that combining these fibroblasts with DCIS cells in a mouse model provides an environment that permits the DCIS to become invasive.12
Investigators are even looking at alterations in the collagen around the DCIS using a variety of methods, including an imaging technique called second harmonic generation microscopy, Schnitt said. For example, collagen fibers perpendicular to the duct perimeter were found to be more frequent in DCIS lesions with features typical of poor prognosis, although collagen alignment and expression of the protein syndecan-1 in the stroma did not predict recurrence.13
A number of research groups are evaluating the DCIS microenvironment using multiplexed immunofluorescence methods to simultaneously identify different molecules on paraffinembedded specimens, Schnitt said.
His team uses cyclic immunofluorescence, which allows the study of up to 60 markers on a single slide using repetitive cycles of immunofluorescence. A colorful image of a DCIS sample stained from multiple immune cell marks shows how the methods can distinguish elements of the microenvironment, quantify the number of cells of each type, and elucidate their relationship to the DCIS, he said.
Much more research is needed to improve the understanding and treatment of DCIS, Schnitt said, and he wants to emphasize the importance of investigating a variety of cell types and markers. “My main point is really to be more provocative than definitive. It’s to suggest that maybe 1 of the reasons we haven’t made more progress in understanding why DCIS progresses to invasive breast cancer, in some cases, is that we focus largely on the DCIS cells themselves, and, in fact, focusing on the factors in the microenvironment may be just as important if not more important,” he said.