Introduction

The efficacy of neoadjuvant chemotherapy (NAC) in achieving successful breast conservation for women initially considered ineligible for breast conservation therapy (BCT) was demonstrated more than 15 years ago [1]. Current practice patterns of NAC use, however, are not well-described, particularly as the surgical treatment of invasive breast cancer has changed in the past 20 years. Partial mastectomy (PM) has become more common as the initial surgical procedure for management of the primary breast tumor [2,3]. This shift is largely due to the results of several randomized controlled trials and meta-analyses that established BCT to be as effective as total mastectomy (TM) with regard to overall patient survival [4,5] and the 1990 National Institutes of Health Consensus Conference [6] recommendation of BCT as an alternative to TM. More recently, TM rates appear to be increasing secondary to a variety of proposed factors, although this remains controversial [3,7-9]. Clinical conditions or patient-related factors associated with greater utilization of initial mastectomy include larger tumor size, multicentric breast cancer, family history of breast cancer, race, younger age, pre-operative breast magnetic resonance imaging (MRI) utilization, lower socioeconomic status, greater living distance from a radiation facility, patient preference, and provider preference [10-14].

Commonly accepted indications for NAC include the treatment of patients presenting with inflammatory breast cancer or T4 lesions (direct invasion of tumor into chest wall or skin) as well as treatment of women with initially operable tumors (T1c-3, N0-2, M0) who desire BCT but would be considered ineligible based on tumor size or tumor to breast size ratio [15,16]. A significant amount of clinical research has demonstrated the efficacy and safety of NAC in allowing patients initially considered ineligible for BCT to successfully undergo BCT without compromising patient overall survival. The National Surgical Adjuvant Breast and Bowel Project-18 (NSABP B-18) trial found NAC to be equally efficacious to post-operative chemotherapy in overall survival rates, while demonstrating a moderate increase in the percentage of women who were candidates for BCT, as did NSABP B-27 [1,17]. Similarly, investigators at the Milan Cancer Institute reported even greater success with NAC use, allowing for BCT in >50% of patients with large tumors (>5.0 cm) who were initially considered ineligible for BCT [18]. The results of these and other trials have led to endorsement of NAC by the National Comprehensive Cancer Network (NCCN) and an international panel of experts [19,20].

Under-utilization of effective treatments and other forms of unexplained clinical variation are recognized barriers to efficient and optimal health care. Studies have demonstrated wide variation in both regional and provider-level practice patterns that can result in decreased quality of care due to underuse, overuse, or misuse of health care services [21,22]. Comparison of NAC utilization between different health care organizations outside of clinical trials has not been well reported; therefore, with our study, we sought to examine the current pattern of NAC utilization among four geographically diverse healthcare institutions and identify factors associated with NAC use, particularly in the scenario of failed breast conservation.

Methods

Results

A total of 4527 patient records were reviewed in the BRCASO database and consisted of 4630 diagnostic events (including additional tumors in the contralateral breast). Of these patients, 2871 women with 2907 diagnostic events met study entry criteria and were included in statistical analyses. Subjects were excluded for undergoing initial surgery without confirmed pre-operative malignancy (n=393), pre-operative diagnosis of DCIS (n=936), pre-operative malignancy type unknown (n=16), known Stage IV breast cancer (n=6), confirmed breast cancer in an axillary node with no known primary (n=5), inflammatory breast cancer (n=75), pre-operatively known multifocal breast cancer (n=96), previous ipsilateral breast cancer (n=85), previous history of radiation therapy (n=98), and patients receiving endocrine-only neoadjuvant therapy (n=13). The remaining patients did not have identifiable clinical factors limiting eligibility for BCT. Of these patients, 111/2907 (3.8%) were treated with NAC.

Patient demographics according to utilization of NAC are shown in Table 1. Younger women (< age 35) were more likely to receive NAC (P<0.0001). Additional pre-operative variables associated with higher utilization rates of NAC included known positive nodal status (P<0.0001), increasing tumor size (P<0.0001), negative estrogen receptor (ER) status (P<0.0001), and negative progesterone receptor (PR) status (P<0.0001). There was no statistically significant difference in NAC use between invasive ductal carcinoma and invasive lobular carcinoma (P=0.1041). Utilization of NAC varied significantly across the four treatment facilities (P<0.0001).

A multivariable analysis of the patient, institution, and pre-operative tumor characteristics related to NAC use in the univariate analysis described above was performed (Table 2). Pre-operative positive nodal status, treating institution, pre-operative tumor size, and age were significantly associated with increased NAC utilization after controlling for other variables in the model (Table 2). Of note, ER and PR status were not independently significant.

The frequencies of PM and TM as the initial or final surgical procedures were examined (Figure 1). An estimated clinical tumor size of 30–40 mm was noted to have an approximately 1:1 ratio for frequency of initial PM to initial TM. BCT was used more often for smaller tumors, while initial TM and final TM were more common for larger tumors. When these data were examined according to individual institution, the trend for increased initial TM (exceeding 50% in tumors >40 mm) was seen at each institution and was statistically significant (P<0.0001) (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Frequency of Surgery Modalities according to Tumor Size

As tumor size increases, the frequency of breast conservation therapy (BCT) decreases and the frequency of mastectomy as initial or final surgery increases. At a tumor size of >30 - 40 mm, an approximate 1:1 ratio exists between BCT and mastectomy.

https://doi.org/10.1371/journal.pone.0084535.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Frequency of BCT Failure by Tumor Size

As tumor size increases, the rate of breast conservation therapy (BCT) failure increases. Frequencies are indicated above each bar.

https://doi.org/10.1371/journal.pone.0084535.g002

Discussion

Clinical trials have demonstrated the benefit of NAC for increasing breast conservation rates without compromising overall survival compared to adjuvant chemotherapy [16-18,24-28]. We found that NAC use varied significantly across four large institutions, suggesting non-uniform practice patterns among the participating institutions, and potentially throughout the country. The data presented here also suggest that NAC may be relatively underutilized as a means of tumor down-staging as evidenced by BCT failure. Interestingly, the percentage of patients undergoing TM as a first procedure increased dramatically for tumor sizes >3.0 cm, rather than the 5.0 cm size traditionally evaluated in clinical trials examining the effectiveness of NAC in tumor downstaging [1,18]. A similar finding was reported in a previous single institution study [29]. The present study demonstrates that a marked increase in initial TM for patients with tumors >3.0 cm exists across multiple institutions. Our findings suggest that in addition to larger tumor size, factors such as younger age and positive nodal status are also associated with increased NAC use. NAC was used infrequently in patients aged 65 or older. Lower use in the elderly may reflect higher co-morbidity or less patient concern with breast preservation.

Common practice would suggest that NAC use would increase with increasing tumor size, as smaller tumors are more likely to be amenable to proceeding with BCT without requiring NAC. However, we did not observe a proportional increase in utilization of NAC with larger pre-operative tumor size, especially in tumors >4 cm in size (Table 1). When compared to the frequency of TM as the initial surgery, utilization of NAC is variable among the institutions, even for larger tumors. The rates of NAC use in tumors >5 cm were 3.9%, 0%, 43.5%, and 20% for the four institutions, despite frequencies of initial TM as the initial surgery of 88.5%, 81.8%, 78.3%, and 60%, respectively (Table 3). Data from these four institutions, all of which show a predilection for BCT in patients with smaller tumors, demonstrate potentially concerning variation in the utilization of NAC. At one institution, NAC use was extremely infrequent (1.1%). Two of the institutions had slightly more prevalent use (8.4% and 6.7%), but one of these surprisingly had lower utilization for tumor size >5 cm.

It is possible that many women who underwent TM as the initial procedure may have chosen TM as their treatment of choice, but we are unable to evaluate patient decision-making or education regarding potential NAC use in our study. Nevertheless, given that the majority of women with smaller breast tumors in this study chose initial BCT, it is not unreasonable to assume that women with larger tumors would also desire BCT if feasible. The large disparity between NAC use and BCT failure (compare Table 1 with Figure 2) suggests a potential opportunity to increase the rate of breast conservation through the greater use of NAC.

Variation in NAC use among various treatment centers deserves further investigation into the possible causes. Legitimate variation could result from inherent differences in the patient population or presentation of disease. However, we sought to minimize these differences by careful selection of study inclusion and exclusion criteria. Despite strict selection criteria, variation in NAC use persisted across institutions. There are several potential driving factors that may account for this finding including discrepancies in available resources such as multidisciplinary breast clinics, timing of multidisciplinary tumor conferences relative to time of initial surgery, level of training, and local culture [30].

A major limitation of the present study is the retrospective nature of data collection, which limited the determination of other treatment decision-making factors. Individual patient or surgeon preference was not available for review. Women with comorbidities precluding chemotherapy, strong family history or genetic risk factors such as BRCA status, or tumors with an extensive intraductal component may account for some of the omission of NAC [7]. Patient awareness of NAC as an option before surgery and evaluation of how it may impact the success of their initial surgical therapy could not be evaluated. Similarly, prospective involvement of medical oncologists or utilization of multidisciplinary tumor conferences before treatment initiation were not evaluated as part of this study. Gene expression profile testing was not widely implemented during the timeframe of this study, and therefore, should not have impacted the results. HER2-neu status was not available, as it was not consistently tested during the study time period. Type of NAC used was not available, and specific information regarding accessibility and enrollment in clinical trials was not available. Despite these limitations, the findings of the present study suggest further investigation may be useful for optimizing patient treatment preferences and initial treatment strategy. Strategies to reduce clinical variation and match best initial treatment selection with patient desired outcomes are essential steps in quality improvement.

Management of invasive breast cancer requires an effective multi-disciplinary program. Early involvement of medical oncologists in treatment discussions may impact the utilization of NAC for newly diagnosed patients who desire BCT. Institutional variation in the use of NAC may represent an opportunity to enhance quality in overall breast cancer care. Understanding the factors that influence this variation and potential underuse of NAC may advance efforts to improve treatment outcomes and optimize patient treatment choices and patient satisfaction in breast cancer care. Further investigation into improving both patient and clinician education is crucial.