Local recurrence after partial mastectomy: Relation to initial surgical margins

The American Journal of Surgery (2011) 201, 374 –378

Midwest Surgical Association

Local recurrence after partial mastectomy: Relation to initial surgical margins Laurence E. McCahill, M.D.a,*, Richard Single, Ph.D.b, John Ratliff, J.D.c, Johanna Sheehey-Jones, R.N., B.S.N.c, Ann Grayc, Ted James, M.D.b a

Richard J. Lacks Cancer Center, Saint Mary’s Health Care, 250 Cherry Street South East, Grand Rapids, MI 49503, USA; bUniversity of Vermont, Burlington, VT, USA; cFletcher Allen Health Care, Burlington, VT, USA KEYWORDS: Breast cancer; Cancer surgery quality; Outcomes

Abstract BACKGROUND: Local recurrence (LR) after partial mastectomy (PM) has been associated with inadequate surgical margins. We assessed LR association with initial margins after PM in patients receiving postoperative radiation therapy (RT). METHODS: Initial margins, re-excision status, and ipsilateral LR were identified for all patients having initial PM from 2003 to 2008. RESULTS: Seven hundred twelve patients underwent PM as their final procedure, and 598 (84.0%) had adjuvant RT. Initial margins were positive or ⬍1-mm margins in 166 patients (27.8%). Re-excision was performed for all positive and 20.2% of patients with margins ⬍1 mm. We observed 10 LRs (1.7%) at the 3.4-year mean follow-up. For patients with initial margins ⬍1 mm, the LR rate was 4.2% (7/167) and just .7% for margins ⱖ1 mm (P ⫽ .006). CONCLUSIONS: We report lower LR rates than traditionally reported. The surgical practice of reexcision to achieve margins of 1 to 5 mm needs closer scrutiny because it may have no impact on LR. © 2011 Elsevier Inc. All rights reserved.

Controversy regarding what constitutes an appropriate surgical margin in the management of breast cancer has recently been highlighted by survey studies of breast surgeons.1,2 These surveys clearly showed wide variation in opinion among practicing breast cancer surgeons and radiation oncologists regarding the appropriate distance of surgical margins. Blair et al’s survey1 of 351 surgeons indicated 28% accept a 1-mm margin, 50% a 2-mm margin, and 12% desire a 5-mm final surgical margin. Both surveys highlight the immense potential for wide variation in the management of breast cancer margins and the use of breast re-excision. Breast surgical margins convey enough contro* Corresponding author: Tel.: ⫹1-616-685-5085; fax: ⫹1-616-685-3044. E-mail address: [email protected] Manuscript received July 24, 2010; revised manuscript September 8, 2010

0002-9610/$ - see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjsurg.2010.09.024

versy that the 2009 National Comprehensive Cancer Network Guidelines remain vague on a specific appropriate margin clearance, suggesting 0 to 10 mm for invasive cancer. These vague guidelines leave decision making regarding re-excision to individual surgeon/multidisciplinary team opinion. Because the vast majority of partial mastectomy patients will have initial margins in the 0 to 10 mm range, this leaves the majority of patients undergoing partial mastectomy open to highly variable surgical care regarding the need for re-excision. We have recently reported that mastectomy rates continue to vary widely, even among major academic medical centers, without explanation.3 We believe that the current lack of consensus regarding appropriate margin distance likely results in even wider variance in partial mastectomy re-excision rates. Recent studies have reported re-excision rates of 20% to 60%, a range that seems unacceptably broad.

L.E. McCahill et al.

Local recurrence after partial mastectomy

An acceptance of very narrow but pathologically negative margins (tumor not touching ink) may be attributed to surgeons’ active participation in North American clinical trials that helped establish the role of radiation therapy after partial mastectomy.4 For many surgeons of the 1980s and their subsequent trainees, this established the National Surgical Adjuvant Breast and Bowel Project (NSABP) definition of adequate breast cancer margins as simply tumor not touching ink. Other retrospective studies, however, have suggested that a greater microscopic distance of tumor from pathological inked margins may be important for minimizing local recurrence.5 Difficulties in establishing an appropriate margin distance through a prospective clinical trial are daunting. The anticipated differences in local recurrence based on differences in margin width would be anticipated to range from just 7% to 16%.6 The pattern of timing of local recurrence from the NSABP-06 study showed that 73% of ipsilateral local recurrences occurred within 5 years, and 18% occurred 5 to 10 years after surgery.4 Establishing an optimal margin distance with a required stratification schema of pathologically negative, ⬍1 mm, 1 to 2 mm, and so on in a prospective trial could potentially require ⬎10,000 patients. With minimal anticipated impact on overall patient survival, the likelihood of such a daunting trial moving forward is unlikely. It is possible, however, to consider assessing the impact of pathological margin distance on local recurrence rates through the analysis of wellmaintained, prospectively collected clinical databases. The University of Vermont Breast Cancer Surgical Outcomes (Vermont breast cancer surgical outcomes [VBCSO]) database was established in 2002 and has recorded detailed pathological margin status, including distance and margin direction of closest margin on initial breast cancer excisions. Additionally, we have maintained the pathological status of all re-excisions for patients undergoing their initial breast cancer surgery at our institution.7 In this study, we replaced local recurrence data from our hospital tumor registry with surgical outcomes data from the VBCSO database. Second, we sought to assess the impact of both initial breast cancer pathological margin status and final margin status (after any re-excision) with local recurrence in patients undergoing initial partial mastectomy. Because of the significant impact of radiation therapy on reducing local recurrence; we chose to focus this cooperative effectiveness study on patients undergoing partial mastectomy and completing postoperative adjuvant radiation therapy for the management of their breast cancer.

Methods We have continuously maintained a VBCSO database focusing on quality measures of breast cancer surgery for 8 years. All data elements considered pertinent to quality of breast cancer surgery were defined in 2002 and have been prospectively recorded for all consecutive breast cancer

375 operations performed at Fletcher Allen Health Care (FAHC), the hospital associated with the University of Vermont. Specific measures recorded on all breast cancer patients include the type of breast surgery performed, lymph node surgery performed, the use of neoadjuvant therapies, localization techniques, margin status of excision, anatomic direction of positive and close margins, distance to closest margin, histology at closest margin, and description of any re-excisions or subsequent mastectomy after initial partial mastectomy. In this study, we linked the surgical outcomes data with recurrence data as identified in the hospital cancer registry. The cancer registry is maintained in accordance with American College of Surgeons Guidelines, and patient status is updated every 6 months. All patients ultimately treated by partial mastectomy (PM) were then analyzed for having received adjuvant radiation therapy. Search terms for cancer recurrence included all types of recurrence including local, regional, and distant metastases. All charts with cancer registry–recorded recurrences were then reanalyzed to confirm accuracy of tumor registry coding, and corrections were made to tumor registry data as required. We limited our subsequent analysis to patients with confirmed local recurrences. For purposes of this study, we considered a local recurrence any recurrence of breast cancer occurring in the ipsilateral breast as the incident cancer. The last analysis for local recurrence was completed in June 2010. Follow-up was calculated from the date of initial surgery until the last clinical follow-up or date of death.

Results There were 1,023 female patients who underwent initial breast cancer surgery at FAHC from March 2003 to December 2008 with a mean age of 57.5 years. An initial PM was performed in 758 (74%). Linkage based on patient name, medical record number, and date of initial surgery between the VBCSO database and FAHC Tumor Registry database was 100%. The median follow-up as of June 2010 was 3.37 years (40.4 months). We subsequently excluded 21 patients with a history of a prior ipsilateral breast cancer because attributing local recurrence to the incident versus the previous ipsilateral breast cancer could not be determined. We also excluded 25 patients who underwent a total mastectomy within 60 days of initial partial mastectomy as a secondary or tertiary procedure for the management of close margins. This resulted in 712 patients with both an initial and final procedure of PM considered at risk for a subsequent local recurrence. The final diagnosis among these 712 patients was invasive ductal carcinoma 514 (72%), invasive lobular carcinoma 51 (7%), and ductal carcinoma in situ (DCIS) in 147 (20%). The distance of the closest margin and histology of the closest margin after initial PM is shown in Table 1.

376

The American Journal of Surgery, Vol 201, No 3, March 2011

Table 1

Margin status after initial PM (n ⫽ 712)

Final diagnosis

Invasive cancer (n ⫽ 565)

Histology at margin

Invasive DCIS DCIS (n ⫽ 408) (n ⫽ 157) (n ⫽ 147) Percent

Positive margin 30 ⬍1 mm 53 1.0–1.99 mm 36 2.0–4.99 mm 138 ⬎5.0 mm 151

23 50 34 36 14

DCIS (n ⫽ 147)

15 34 18 25 55

9 18 13 31 29

Among the 712 patients whose final procedure was PM, 598 (84%) underwent adjuvant radiation therapy (RT) and are the focus of our analysis for local recurrence. The radiation therapy completion rate for patients undergoing final PM in this series was 89% for patients with invasive cancer and 65% for patients with a final diagnosis of DCIS. All patients were treated with external-beam radiation therapy, and none were treated with brachytherapy or local radiation therapy techniques. A standard radiation therapy dose of 5,040 cGy was delivered, and most patients with narrow margins and/or high-risk tumor features received a tumor bed “boost,” often an additional 1,000 cGy. Among the 598 PM patients receiving adjuvant RT, their final diagnosis was DCIS in 96 (16%) and invasive cancer in 502 (84%). We evaluated the pathological margin status of the initial PM and its association with local recurrence, independent of any subsequent re-excision among 598 PM patients receiving postoperative RT (Figure 1). The initial margin status was positive in 58/598 (10%) patients, and 100% of these patients underwent re-excision. The initial margins were ⬍1

Figure 1

mm in 109/598 patients (18%). Among these patients with ⬍1-mm initial margins, 22 (20%) underwent a re-excision, 8 of 20 (40%) for patients with a final diagnosis of DCIS and 14 of 89 (16%) with a final diagnosis of invasive cancer. In cases of invasive cancer, the histology of closest margin when margins were ⬍1 mm was DCIS in 43 of 89 patients (48%). The overall local recurrence rate for the 598 PM patients receiving adjuvant radiation at a median follow-up of 3.4 years was 2%. Among 431 of these whose initial margins after PM were ⬎1 mm, we observed 3 local recurrences (1%), which was significantly lower than the 7 local recurrences (4%) observed among 167 patients with initial margins positive or ⬍1 mm (P ⫽ .006, Fig. 1). In evaluating recurrence as related to final margin status after any subsequent re-excision, we observed local recurrence in 5 of 105 patients (5%) whose final margins were ⬍1 mm compared with 5 of 493 (1%) patients whose final margins were ⬎1 mm, a difference that was statistically significant (P ⫽ .019). The overall impact of re-excision, however, cannot be determined at this time because the recurrence rate even among patients whose final margins were ⬍1 mm or positive remains quite low at ⬍5%, and many patients with ⬍1mm margins did not undergo re-excision.

Comments The precise definition of what constitutes an adequate margin for breast cancer resection has long been debated within the surgical community and represents an area of considerable variation in practice leading to disparate treatment and outcomes for breast cancer patients.1,2,6 Patient management varies widely based on what threshold surgeons accept for adequate margins and the subsequent need

Local recurrence rates following partial mastectomy with or without radiation by initial surgical margin status.

L.E. McCahill et al.

Local recurrence after partial mastectomy

for re-excision. Most often surgeons have used the presence of residual disease at re-excision as verification of the necessity of re-excision. For instance, Dillon et al8 showed a strong correlation between decreasing initial pathological margin distance for DCIS and the presence of residual DCIS at re-excision, with 62% of patients with an initial margin ⬍1 mm having residual disease compared with 17% with initial margins of 2 to 5 mm. The use of residual disease as an indicator of “appropriate” re-excision, however, is controversial. Previous prospective randomized clinical trials have shown that radiation therapy significantly diminished ipsilateral local recurrence after partial mastectomy for invasive cancer. Therefore, we can presume that subclinical microscopic disease is often present and effectively treated by radiation therapy. The pathological identification of microscopic disease on re-excision, therefore, may not necessarily indicate the necessity of re-excision.4 Identifying the impact of re-excision on long-term ipsilateral local recurrence is needed. Without clear evidence showing that reexcision for various thresholds of pathological distance improves local recurrence, the optimal management of close partial mastectomy margins remains ambiguous. The only current consensus regarding breast cancer margin management is that leaving a widely positive tumor at a resection margin is associated with increased rates of local recurrence. Re-excision should generally be performed when more than focal margin involvement is present.5 There is, however, no consensus regarding how much (measured in millimeters) margin distance beyond tumor is required. Some reports describe margin distance as a predictor of local recurrence, and other studies find that margin status does not predict local recurrence.6,9,10 We believe our study represents one of the largest series evaluating local recurrence after PM and is the first with a detailed analysis of the pathological breast excision in all directions. The Vermont Breast Care Center routinely reviews all new primary breast cancers weekly in a multidisciplinary conference, so pathology is routinely rereviewed and decisions regarding the necessity for re-excision often come about with multidisciplinary input. Our center clearly showed a propensity to tolerate invasive margins ⬍1 mm, which is consistent with an NSABP definition of tumor not touching ink, and no patient in this study with invasive margin ⬎1 mm was re-excised. Among the 12 patients with DCIS with ⬍1-mm margins who were not re-excised, for 9 of those, the direction of the closest margin was anterior or posterior, and the surgeon felt there was no residual breast tissue in that direction to warrant re-excision. Given the relatively broad practice acceptance of ⬎1-mm margins as the minimum acceptable for breast cancer shown in 2 current surveys (89% in the Azu survey and 85% in the Blair survey), the extremely low local recurrence in our study for initial margins ⬍1 mm calls into question the need to achieve margin distances of 1 to 2 mm or even 2 to 5 mm. Our dataset of patients undergoing initial

377 PM shows that 13% have an initial closest margin of 1.0 to 1.99 mm, and an additional 31% have a closest margin of 2.0 to 4.99 mm. None of these patients underwent reexcision in this study, and to date the local recurrence rate in this group remains ⬍1%. However, at other centers depending on the interpretation of appropriate margins, this large group of patients (43%) with 1.0- to 5.0-mm margins after initial PM might undergo re-excision. The total health care costs of variable surgical management of such a large percentage of breast cancer patients are immense. Based on our ipsilateral local recurrence outcomes, we would strongly question the practice of re-excision for margins ⬎1.0 mm for invasive cancer in patients receiving adjuvant RT. Our study suggests a potential association between initial margin status after PM and ipsilateral local recurrence, with local recurrences being greater (4%) in patients with ⬍1-mm margins. This is still considerably lower than previous reports of local recurrence after PM and adjuvant RT, which was 13% in the NSABP B-06 trial.4 This likely reflects an inadequate median follow-up of 3.4 years in this study, which prevent us from drawing firm conclusions. However, the low ipsilateral local recurrence rate we have observed even in patients with ⬍1 mm initial margins suggests that potential factors other than surgical margins may be contributing to improved local control. All the patients evaluated in this study were treated at a multidisciplinary breast center and received adjuvant radiation therapy. Patients with narrow margins uniformly received a radiation boost to the local tumor bed. Many patients also received systemic therapy in the form of chemotherapy and/or endocrine therapy. Recent data has also demonstrated that the use of systemic therapy influences local control; especially when treatment includes specific targeted therapies.11 These findings support the premise that local control is a multi-factorial process; and perhaps tumor biology factors are more influential in local recurrence rather than actual microscopic margin distance. It is entirely feasible that the thousands of breast re-excisions performed annually in the United States to achieve 2 mm and even 5 mm margins have little or no impact on breast cancer local recurrence. We did, however, observe a 6-fold increase in local recurrence in patients with initially close (⬍1 mm) or positive margins compared with patients with initial margins ⬎1 mm; despite our practice of routine reexcision for all cases of positive margins and the uniform use of radiation therapy in all patients. In these circumstances, the recommendation for reexcision appears appropriate for patients with initial margins ⬍1 mm, particularly radial margins. We do not yet have enough total local recurrences to distinguish a recurrence pattern between ⬍1-mm DCIS margins and ⬍1-mm invasive cancer margins. We also do not yet have enough local recurrences to identify if our practice of not re-excising ⬍1-mm anterior and posterior margins can be supported.

378 Our data support accepting margins ⱖ1 mm as adequate to achieve optimal local control in patients receiving adjuvant radiotherapy. Re-excision for margins ⱖ1 mm may be avoided; thus, limiting additional costs of treatment, the length of time to the initiation of adjuvant therapy, patient stress and anxiety, and worse cosmetic results associated with additional surgery. The strength of these findings is limited because of the short period of follow-up in this study. Other limitations of this study include the lack of separate analysis or control for patients receiving adjuvant systemic therapy, hormonal therapy, and tumor features such as grade, size, extent of intraductal pathology, nodal status, and lymphvascular invasion, all of which may potentially influence local recurrence rates. Furthermore, the application of findings in this study may be limited in that it represents a single high-volume breast center, and highvolume breast surgeons performed most procedures. The impact of close pathological margins on local recurrence in patients undergoing breast-conserving therapy in the setting of multidisciplinary care merits further investigation. Margins as narrow as 1 mm appear adequate in breast cancer patients receiving adjuvant radiation therapy. Larger studies with longer follow-up and detailed assessment of decision making and anatomic direction of closest margins will be needed to confirm whether these findings are statistically and clinically significant and may greatly influence the overall cost of treatment and quality of life for breast cancer patients.

References 1. Blair SL, Thompson K, Rococco J, et al. Attaining negative margins in breast-conservation operations: is there a consensus among breast surgeons? J Am Coll Surg 2009;209:608 –13. 2. Azu M, Abrahamse P, Katz SJ, et al. What is an adequate margin for breast-conserving surgery? Surgeon attitudes and correlates. Ann Surg Oncol 2010;17:558 – 63. 3. James TA, McCahill LE. Variant mastectomy rates: implications for quality of care in breast cancer surgery. J Clin Oncol 2010;28:e364. 4. Fisher B, Anderson S, Bryant J, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med 2002;347:1233– 41. 5. Singletary SE. Surgical margins in patients with early-stage breast cancer treated with breast conservation therapy. Am J Surg 2002;184: 383–93. 6. Morrow M. Breast conservation and negative margins: how much is enough? Breast 2009;18;Suppl 3:S84 – 6.

The American Journal of Surgery, Vol 201, No 3, March 2011 7. McCahill L, Privette A, James T, et al. Quality measures of breast cancer surgery: initial validation of feasibility and variation in outcomes among surgeons. Arch Surg 2009;144:455– 62. 8. Dillon MF, Hill AD, Quinn CM, et al. A pathologic assessment of adequate margin status in breast-conserving therapy. Ann Surg Oncol 2006;13:333–9. 9. Vargas C, Kestin L, Go N, et al. Factors associated with local recurrence and cause-specific survival in patients with ductal carcinoma in situ of the breast treated with breast-conserving therapy or mastectomy. Int J Radiat Oncol Biol Phys 2005;63:1514 –21. 10. Macdonald HR, Silverstein MJ, Lee LA, et al. Margin width as the sole determinant of local recurrence after breast conservation in patients with ductal carcinoma in situ of the breast. Am J Surg 2006;192: 420 –2. 11. Mamounas EP, Tang G, Fisher B, et al. Association between the 21-gene recurrence score assay and risk of locoregional recurrence in node-negative, estrogen receptor-positive breast cancer: results from NSABP B-14 and NSABP B-20. J Clin Oncol 2010;28:1677– 83.

Discussion Constantine Godellas, M.D. (Maywood, IL): I have 2 questions for you. Why did you primarily look at initial (vs final) margin status, and were there any patients that still had close margins after reexcision? Second, how often do use magnetic resonance imaging (MRI)? Do you think this may make a difference? Your follow-up time is relatively short, and I would like to see longer-term data. Laurence E. McCahill, M.D. (Grand Rapids, MI): The reason we looked at the initial margin status was to determine what happened to those who were not re-excised with 1- to 2-mm margins or even the 2- to 5-mm margins. Based on final margins, 5 of 100 patients with less than 1-mm final margins have recurred. We do not have the power yet to evaluate whether or not it is important to re-excise for less than 1-mm margins. We did not control for MRI use. If you look at traditional data that show local recurrence rates under 10% for partial mastectomy in patients who undergo radiation therapy, it is possible that the use of MRI may be able to decrease local recurrence rates by finding incidental or occult other cancers in the same breast. Sukumal Saha, M.D. (Flint, MI): Can you comment on how many patients had brachytherapy as opposed to wholebreast radiation with or without boost? Dr. McCahill: So during these years, almost 100% had external-beam radiation therapy, and boost was at the discretion of the radiation oncologist, but, generally, all younger women received a boost and anyone with less than 1-mm margins would have had a boost.