Active Surveillance for DCIS: The Importance of Selection Criteria and Monitoring
- 2.7k Downloads
More than 50,000 women in the United States this year alone and 1 of 33 women in the United States over their lifetime will be diagnosed with DCIS, almost all of these diagnoses made in completely asymptomatic individuals.1 Current guidelines dictate that all DCIS be treated upon diagnosis with some combination of surgery, radiation, and hormonal therapy—treatments very similar to those recommended to patients with invasive cancer. The goal of these treatment recommendations is to prevent breast cancer mortality. However, there remains considerable controversy over whether this represents overtreatment for those women whose DCIS may not progress to invasive cancer within the patient’s lifetime, or for whom an invasive cancer diagnosis may result in comparable treatment and outcome if it had been treated at the time of a DCIS diagnosis.
Active surveillance is a management approach that has been implemented for other cancers, most notably for early-stage invasive prostate cancer. For DCIS, this strategy could reduce overtreatment by avoiding surgical excision and possibly radiation in patients with low-risk disease, with treatment only if invasive progression is identified during surveillance. A meta-analysis of women who had DCIS that was not recognized and thus not excised found that only 22 % of DCIS progressed to invasive cancer without treatment at 1–31 years of follow-up, suggesting that many, if not most, women treated for DCIS may not benefit.2 To address these concerns, there are currently 3 randomized controlled trials for low-risk DCIS underway in Europe and the United States designed to test the safety, efficacy, and trade-offs of active surveillance compared with usual care: LORIS (multicenter UK study), LORD (EORTC study), and COMET (cooperative group U.S. study). All 3 trials seek to identify a subset of patients with DCIS with low risk of both occult invasive disease at initial presentation and subsequent progression to invasive disease. The endpoint of all studies is to evaluate invasive cancer progression and quality of life.
Critical to the successful implementation of any active surveillance program is the careful selection of eligible patients. In this journal, Pilewskie et al. report long-term outcomes in a population of women with low-risk DCIS, defined as those meeting published LORIS trial registration eligibility criteria.3 Among the 193 LORIS-eligible women who had lumpectomy alone, there were only 6 invasive cancer recurrences during follow-up, for an estimated 6 % 10-year rate of invasive cancer. Importantly, none of the patients in this cohort had occult invasive disease and all underwent surgery; thus this population represents a best-case estimate of invasive recurrence risk.
In the absence of surgery, the risk of invasive progression will undoubtedly be higher than in the series reported by Pilewskie et al. However, the degree to which this rate is increased and most importantly, whether this may impact the primary outcome of breast cancer specific survival remain unknown. This question is reasonable, as an EBTCG meta-analysis of 4 randomized controlled trials of lumpectomy for DCIS with or without radiation demonstrated that although radiation reduced recurrence following lumpectomy for DCIS by at least 50 %, breast cancer specific mortality was not impacted.4 In these studies of non-risk-stratified patients who likely represent a higher-risk population than LORIS-eligible patients, reduction of invasive recurrence did not result in either improved breast cancer specific survival or overall survival.
Currently, most physicians and patients equate DCIS with all of the implications of an invasive cancer diagnosis. However, an alternate strategy could be to consider DCIS as an opportunity for close surveillance and possible chemoprevention for both the ipsilateral and contralateral breast, much as how many patients with other high-risk histologies such as ADH or LCIS are commonly managed. Moreover, it is important to closely examine the trade-offs of any intervention for conditions that carry a low risk of morbidity and mortality. There are notable examples where data from prospective randomized trials have allowed the safe reduction of long-entrenched surgical patterns of treatment for breast cancer. ACOSOG Z011 provided evidence that sentinel node biopsy alone in node-positive patients undergoing lumpectomy and whole breast radiation resulted in excellent locoregional control, despite the likelihood that at least 20 % of patients who did not undergo axillary node dissection probably harbored residual axillary nodal disease.5 CALGB 9343 randomized women with early-stage invasive breast cancer aged 70 years and older to lumpectomy alone or lumpectomy with radiation and found that omission of RT did not lead to differences in breast cancer specific survival or overall survival.6 These studies provide examples of a rational approach to de-escalation of treatment for breast cancer. Moreover, such findings must be contextualized in light of continual improvements in outcomes for all patients with breast cancer. Moreover, more breast cancers than ever before are diagnosed in patients with major competing causes of mortality who may never derive any benefit from treatment of low-risk conditions.
Invariably, some patients undergoing active surveillance for DCIS will develop invasive cancer. Thus, all patients who elect to undergo active surveillance must be followed with close screening so that every patient has the best opportunity for invasion to be detected at an early stage, at which point the treatments and outcomes may not differ substantially from those they would have had at the time of DCIS diagnosis. An area of active investigation is to use detailed imaging features during surveillance. Currently, only basic imaging criteria are used for inclusion (e.g., calcifications) or exclusion (e.g., mass) criteria. Since pathology features only capture information from the small area of DCIS that was sampled during the initial biopsy they will always be limited. In contrast, imaging features apply to the entirety of the DCIS lesion and can be noninvasively followed. Preliminary work from several investigators have shown initial success using computer vision algorithms to extract features from DCIS microcalcifications to predict upstaging to invasive disease at surgical excision.7 Further work in this area is needed, but the use of more advanced imaging features, in conjunction with demographic, clinical, and pathology data, will be an important component of safe surveillance going forward.
De-escalation of therapy can be frightening for both patients and providers. Doing away with treatments is always more challenging than adding more therapies, even if these treatments offer little benefit and come at substantial cost and/or morbidity. However, the full promise of personalized care can only be realized if treatment is tailored to the magnitude of clinical significance. Further, the burden of overtreatment must be acknowledged as an important consideration. Among the LORIS-eligible patients in the Pilewskie study, 401 patients underwent breast conserving surgery, with 97 patients undergoing reexcision and 18 having 3 or more excisions; 207 women had radiation therapy, and 79 women had some component of endocrine therapy. In practice, a considerable proportion of LORIS-eligible women elect unilateral or bilateral mastectomy, further adding to this burden of treatment. One study has estimated that the annual national expenditure incurred by DCIS overtreatment exceeds $240 million.8 Arguably, each patient will have an individual tipping point between treatment and surveillance.
Pilewskie et al. conclude by emphasizing that, “additional research is paramount to identify DCIS subsets lacking capacity for disease progression.” Studies such as Pilewskie’s provide valuable insights into potential outcomes with active surveillance. However, it is only through successful accrual to prospective clinical trials that we will be able to reach definitive conclusions as well as a more nuanced understanding of the risks and benefits of usual care versus active surveillance. Outcomes collected from the LORIS, LORD, and COMET Trials aim to provide answers to important questions regarding the safety, efficacy, appropriate patient selection and quality of life considerations of an active surveillance approach, and each study protocol has been specifically designed to maximize patient safety. Ultimately, each patient with DCIS should be supported and empowered to make the best personal choice about their care, but in order to do so, it is imperative that in future, they be provided with accurate data about the critical trade-offs for each approach.
- 3.Pilewskie M, Olcese C, Patil S, Van Zee KJ. Women with low-risk DCIS eligible for the LORIS Trail after complete surgical excision: How low is their risk after standard therapy? Ann Surg Oncol. 2016. doi: 10.1245/S10434-016-5595-3.Google Scholar
- 5.Giuliano AE, Ballman K, McCall L, et al. Locoregional recurrence after sentinel lymph node dissection with or without axillary dissection in patients with sentinel lymph node metastases: long-term follow-up from the American College of Surgeons Oncology Group (Alliance) ACOSOG Z0011 randomized trial. Ann Surg. 2016;264:413–20.PubMedGoogle Scholar
- 7.Shi B, Grimm LJ, Mazurowski MA, et al. Prediction of invasion in ductal carcinoma in situ using computer-extracted mammographic features, SPIE Medical Imaging (Submitted).Google Scholar