Author Affiliations: Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York (Dr Morrow); and Departments of Medicine and Health Management and Policy, University of Michigan, Ann Arbor (Dr Katz).
Although the treatment of cancer is increasingly multimodal, virtually all patients with solid tumors undergo surgery. In breast cancer, survival after mastectomy, or breast-conserving surgery (BCS) and radiotherapy, are equivalent,1 and both surgical approaches have a negligible risk of major adverse events. Surgical quality performance measures for cancer have generally focused on 30-day morbidity and mortality rates for higher-risk procedures such as pancreatectomy and esophagectomy.2 Thus, breast surgery has largely been excluded from the cancer surgical procedures for which quality measures have been developed. Indeed, the focus of quality improvement in breast cancer has primarily been on process measures tied to nonsurgical treatments, including systemic therapies and radiation.3
The framework used to assess potential quality indicators has been well described and should be incorporated into all studies that investigate new candidate measures. Key concepts include high reliability and validity of the indicator, low cost of collecting the necessary information, clear potential for action based on a direct link to patient health, and a well-defined objective (eg, quality improvement vs selective referral to physicians or hospitals with better performance).4,5
In this issue of JAMA, McCahill et al6 explore the possibility of using reexcision after BCS as a quality indicator for breast cancer surgery. This metric is fairly common, relatively easy to measure, and may be associated with adverse health outcomes and excess costs. The investigators performed an observational study to examine patterns and correlates of reexcision among 2206 patients with invasive breast cancer treated in a variety of practice settings. The overall reexcision rate was 22.9% and almost half of the reexcisions (n = 242, 47.5%) were performed in patients with negative margins, defined as tumor not touching ink (ie, tumor cells not touching the inked surface of the specimen). Rates of reexcision varied significantly by surgeon (ranging between 0% and 70%) and by institution (range, 1.7%-20.9%). The variation was not explained by the basic clinical and demographic factors available in the data set. The authors concluded that the wide range of unexplained variation indicated a potential barrier to high-quality, cost-effective care.
The article by McCahill et al6 underscores the challenge in developing surgical quality indicators for patients with cancer, especially for procedures with very low risk of major complications. While there is strong evidence that positive margins (tumor touching ink) are associated with an increased rate of local recurrence,7,8 a substantial number of reexcisions are performed among patients with negative margins to obtain a more widely clear margin. There is no consensus among surgeons and radiation oncologists as to what constitutes an optimal negative margin width9,10 because the question has not been addressed in prospective randomized trials. The observational design used in the McCahill et al6 study is valuable for illuminating the nature of potential quality gaps but cannot be used to inform the validity of candidate quality measures.
Despite the limitations in study design, the results of the study suggest that reexcision after BCS fails the litmus test as a quality measure for a number of reasons. Validity is low because many factors drive the decision to perform margin reexcision, including clinical variables such as the method of pathological margin processing, quality of preoperative imaging, histological tumor type, and patient age.11,12 Surgeon attitudes toward contraindications to initial BCS and cosmetic outcomes influence reexcision rates. For instance, some surgeons may offer BCS only to patients with tumors of 1 cm or less in size, despite definitive evidence that BCS is equally effective for tumors up to 4 cm in size.1,7 These surgeons' reexcision rates will be low because small tumors are easy to remove, but the low reexcision rate is the result of many unnecessary initial mastectomies. Similarly, surgeons who remove very large amounts of normal breast tissue during lumpectomy will have low reexcision rates, but poor cosmetic outcomes.8,13 Collecting all the information to adjust for these clinical, surgeon, and institutional factors would be both difficult and expensive.
Another problem with breast reexcision is that this measure cannot be strongly linked to health outcomes. A meta-analysis was unable to identify a significant relationship between margin widths and local recurrence after adjusting for other treatment factors such as a boost dose of radiotherapy and the use of endocrine therapy.14 In the absence of a decrease in local recurrence rates, reexcision cannot be expected to be associated with a survival benefit. Although we are unaware of studies focusing specifically on the relationship between reexcision and quality of life, the long-term effect has to be small. A number of studies have shown generally high quality of life and decision satisfaction after breast surgery with no association with receipt of different locoregional treatment options, despite the use of reexcision with BCS but not with primary mastectomy.15,16
Perhaps the most important limitation of using reexcision as a quality measure is the potential consequences of its adoption. The focus in the study by McCahill et al6 on variations in reexcision rates across individual surgeons raises the possibility that this information might be used to direct patients to different surgeons based on having the “right rate.” But there are flaws to the logic of using reexcision for this purpose. Breast surgery is widely disseminated in the community, and there are many lower-volume hospitals and surgeons. Patient volume may be too low to validly discriminate between hospitals after appropriate risk adjustment, let alone individual surgeons. Uncertainty about the link between process and patient health outcomes underscores the challenge of determining “which rate is right.” In addition, behavioral responses to profiling the rates of reexcision of individual surgeons may result in greater use of mastectomy as surgeons restrict BCS to patients for whom negative margins are easily achieved.
The focus of McCahill et al6 on potential surgical overtreatment in breast cancer is important. Advances in the understanding of tumor biology and the availability of increasingly effective targeted therapies make this an ideal time to reconsider approaches to locoregional management. In particular, the substantial increase in the diagnosis of breast cancer with relatively favorable prognosis has motivated a number of initiatives to reduce the burden and morbidity of locoregional treatment in selected subsets of patients. These initiatives are (1) omitting axillary dissection in women with metastases in 1 or 2 sentinel nodes who are undergoing treatment with whole-breast irradiation and systemic therapy,17 (2) the study of partial breast irradiation in subsets of patients known to be at low risk for local recurrence,18 and (3) research that addresses the causes and consequences of the increasing use of contralateral prophylactic mastectomy in the modern era.19 In this context, quality measures include not only the application of appropriate treatment, but also the avoidance of unnecessary or unproven evaluative tests and treatments. Although it may be premature to determine quality indicators for the strategies targeting reduction of therapy for favorable patient groups, any proposed quality measures must be evidence based and relevant to the current understanding of breast cancer biology, and the subset of patients for whom the measure applies must be rigorously defined for those measures to be useful to clinicians and beneficial to patients.
Corresponding Author: Monica Morrow, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 300 E 66th St, New York, NY 10065 (email@example.com).
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Editorials represent the opinions of the authors and JAMA and not those of the American Medical Association.
Morrow M, Katz SJ. The Challenge of Developing Quality Measures for Breast Cancer Surgery. JAMA. 2012;307(5):509-510. doi:10.1001/jama.2012.74