Size of Extranodal Extension on Sentinel Lymph Node Dissection in the American College of Surgeons Oncology Group Z0011 Trial Era

Importance  Based on the American College of Surgeons Oncology Group Z0011 trial exclusion criteria, patients with T1N0 or T2N0 breast cancer with 1 or 2 positive sentinel lymph nodes (SLNs) are recommended to undergo axillary lymph node dissection if extranodal extension (ENE) is present.

Objective  To determine the effect of ENE size on residual axillary nodal burden, disease recurrence, and survival in patients meeting Z0011 criteria.

Design, Setting, and Participants  Retrospective cohort study between January 1, 2000, and December 31, 2012, at a single tertiary cancer center. Patients had T1 or T2 breast cancer with 1 or 2 positive SLNs. The ENE was classified as 2 mm or smaller or as larger than 2 mm.

Main Outcomes and Measures  Nodal burden, disease recurrence, and overall survival.

Results  Of 208 patients, 149 (71.6%) had no ENE, 21 (10.1%) had ENE 2 mm or smaller, and 38 (18.3%) had ENE larger than 2 mm on SLN dissection. The median follow-up time was 60 months (range, 1-158 months). The mean (SD) total number of positive lymph nodes differed significantly for the group with no ENE (1.72 [1.39]) vs the group with ENE 2 mm or smaller (3.22 [2.09]; P < .001) and vs the group with ENE larger than 2 mm (4.26 [5.01]; P < .001). Similar patterns were observed for mean (SD) nonsentinel lymph node metastases: 0.48 (1.30) for no ENE vs 1.91 (2.07) with ENE 2 mm or smaller (P = .02) and vs 2.95 (4.95) with ENE larger than 2 mm (P < .001). For the group without ENE vs the group with ENE 2 mm or smaller, there were no significant differences in recurrence (distant recurrence, 4 patients [2.7%] vs 1 patient [4.8%], respectively; P = .62) or in mortality (18 patients [12.1%] vs 4 patients [19.1%], respectively; P = .48). For the group without ENE vs the group with ENE larger than 2 mm, there were no significant differences in recurrence (distant recurrence, 4 patients [2.7%] vs 4 patients [10.5%], respectively; P = .19) or in mortality (18 patients [12.1%] vs 9 patients [23.7%], respectively; P = .07).

Conclusions and Relevance  Presence of ENE on SLN dissection is associated with N2 disease. Despite increased nodal burden, patients with 1 or 2 positive SLNs and ENE 2 mm or smaller demonstrated recurrence and survival rates similar to those of patients without ENE. Reporting of ENE size should be standardized and required.

The results of the American College of Surgeons Oncology Group (ACOSOG) Z0011 trial¹ have significantly altered management of the axilla in patients with clinically node-negative breast cancer. Whereas the previous standard of care was completion axillary lymph node (LN) dissection (ALND) for patients with even 1 positive sentinel LN (SLN), Quiz Ref IDthe ACOSOG Z0011 trial demonstrated that patients with clinical T1N0 or T2N0 breast cancer undergoing breast conservation therapy with limited disease burden on SLN dissection (SLND) can avoid the morbidity of ALND without compromising locoregional control or long-term survival.¹ Defining the subgroup of patients who can be managed by SLND alone or SLND plus radiation to axillary and medial supraclavicular fields (axillary radiotherapy [ART]) has become an increasingly important clinical question, particularly as the Z0011 study has been criticized for overrepresenting older patients with less aggressive breast cancers.²

Extranodal extension (ENE), defined as tumor breach beyond the LN capsule, is associated with both increased total number of positive LNs and increased nonsentinel LN (NSLN) burden.^3,4 The ACOSOG Z0011 trial excluded patients with gross ENE and did not analyze the effect of microscopic ENE on recurrence or survival, making management of ENE found on SLND in the post-Z0011 era unclear. In our previous work, we showed that ENE found on SLND was associated with increased risk of distant disease recurrence and decreased overall survival.⁵ The objective of this study was to determine whether stratification based on ENE size was a predictor of increased nodal burden, disease recurrence, or overall survival in patients with breast cancer meeting Z0011 criteria.

Patients with breast cancer who had T1 or T2 tumors and 1 or 2 positive SLNs on SLND at Loma Linda University Medical Center were identified from January 1, 2000, to December 31, 2012. All patients underwent ALND. Breast-specific surgical therapy included both partial and total mastectomy. Information regarding patient demographic characteristics and tumor characteristics, including age at diagnosis, tumor stage, hormone receptor status, presence of ENE, disease recurrence, vital status, and receipt of adjuvant therapies, was retrieved from our institution’s prospectively collected database. Patients with SLND performed at referring institutions whose pathology slides could not be obtained for internal review were excluded. Cases in which the SLN was an intramammary node with ENE were also excluded. Ipsilateral breast tumor recurrence was classified as local recurrence, and axillary, infraclavicular, or supraclavicular LN recurrence was classified as nodal recurrence. This study was approved by the institutional review board at Loma Linda University Medical Center. The requirement for informed consent was waived by the institutional review board as only deidentified retrospective data were used for this study.

Histopathologic review was performed on all available axillary SLN slides. Hematoxylin-eosin–stained microscopic glass slides and immunoperoxidase-stained slides for cytokeratin CAM 5.2 (if previously performed as part of our institution’s axillary SLN package) were reviewed for the presence of ENE. Extranodal extension was defined as the presence of full-thickness (ie, into and through) LN capsular invasion, seen by metastatic tumor invasion of extranodal fat with or without associated desmoplastic stromal response (ie, inflamed granulation tissue and/or fibrosis). The extent of ENE was measured as the widest diameter of the invasive front of ENE and categorized either as 2 mm or smaller or as larger than 2 mm (Figure 1). Although there is no accepted standard for ENE measurement, an ENE size of 2 mm was selected as the cutoff for the study because it is similar to the distinction between micrometastasis and macrometastasis in SLN reporting.

We used χ² tests and Fisher exact tests to compare proportions, t tests to compare quantitative variables (5% α level, assuming normal distribution and homogeneity of variance), and analysis of variance for comparisons of 3 groups of independent continuous variables. Hazard ratios (HRs) and 95% confidence intervals were calculated for each covariate in the adjusted model. Survival analyses were performed by Kaplan-Meier method. Statistical analyses were conducted using SAS version 9.4 statistical software (SAS Institute, Inc).

Quiz Ref IDOf 208 patients meeting the ACOSOG Z0011 criteria of T1 or T2 breast cancer with 1 or 2 positive SLNs, 149 (71.6%) did not have ENE, 21 (10.1%) had ENE 2 mm or smaller, and 38 (18.3%) had ENE larger than 2 mm on final pathology. There were no significant differences in age, histologic findings, grade, presence of lymphovascular invasion, first course of surgical therapy, or hormone receptor status among the 3 groups. Larger tumors were more likely to be associated with larger ENE. The mean (SD) tumor size was 1.94 (0.94) cm in the group without ENE, 2.07 (0.90) cm in the group with ENE 2 mm or smaller, and 2.18 (1.22) cm in the group with ENE larger than 2 mm (P = .003) (Table 1). In terms of adjuvant therapies, patients with ENE larger than 2 mm (30 of 38 patients [79.0%]) tended to receive systemic therapy more frequently than patients with ENE 2 mm or smaller (16 of 21 patients [76.2%]) or patients with no ENE (86 of 149 patients [57.7%]) (P = .04). Patients with ENE also tended to receive radiation therapy (ENE ≤2 mm, 17 patients [80.9%]; ENE >2 mm, 27 patients [71.1%]) more frequently than those without ENE (90 patients [60.4%]) (P = .008). There were no differences in the receipt of hormonal therapy between the 3 groups (P = .67). The median follow-up time was 60 months (range, 1-158 months).

When comparing patients without ENE with patients who had ENE 2 mm or smaller on SLND, the latter group demonstrated a significantly higher number of total positive LNs. The mean (SD) number of total positive LNs was 1.72 (1.39) in the group without ENE compared with 3.22 (2.09) in the group with ENE 2 mm or smaller (P < .001). Among the patients without ENE, 138 (92.6%) had a total of 1 or 2 positive LNs; in the patients with ENE 2 mm or smaller, only 11 (52.4%) had a total of 1 or 2 positive LNs (Table 2). Quiz Ref IDPatients with ENE 2 mm or smaller had significantly higher frequency of N2 disease (≥4 positive nodes) compared with the group without ENE (47.6% vs 2.7%, respectively; P < .001). A similar trend was observed for NSLN metastases in these 2 groups. Patients with ENE 2 mm or smaller had NSLN metastases at more than twice the frequency of patients without ENE (61.8% vs 24.2%, respectively; P < .001). The mean (SD) numbers of NSLN metastases were 0.48 (1.30) for those with no ENE and 1.91 (2.07) for those with ENE 2 mm or smaller (P = .02).

Despite harboring significantly increased numbers of total positive LNs and NSLN metastases compared with patients without ENE, patients with ENE 2 mm or smaller demonstrated similar recurrence and survival rates as patients without ENE. There were no significant differences in local (3.4% vs 4.8%), nodal (3.4% vs 0%), or distant (2.7% vs 4.8%) recurrence rates between the group without ENE and the group with ENE 2 mm or smaller (P = .62) (Table 3). In the group without ENE, 18 of 149 patients (12.1%) died during the study period, compared with 4 of 21 patients (19.1%) in the group with ENE 2 mm or smaller (P = .48). Unadjusted Kaplan-Meier analysis shows that overall survival of patients with ENE 2 mm or smaller was significantly worse than in patients without ENE (P = .03) (Figure 2). Quiz Ref IDAfter controlling for tumor size, number of positive LNs, and receipt of adjuvant therapies on multivariate analysis, ENE 2 mm or smaller was not significantly associated with increased mortality (HR = 0.51; 95% CI, 0.17-1.52; P = .23) (Table 4). Having more than 3 positive LNs was a predictor of increased mortality (HR = 3.98; 95% CI, 1.73-9.15; P = .001) (Table 4). Receipt of hormonal therapy was significantly associated with reduced mortality (HR = 0.41; 95% CI, 0.19-0.88; P = .02) (Table 4).

Comparison of patients without ENE vs patients with ENE larger than 2 mm showed significantly higher numbers of total positive LNs and NSLN metastases for the group with ENE larger than 2 mm. The mean (SD) number of total positive LNs was 4.26 (5.01) in patients with ENE larger than 2 mm compared with 1.72 (1.39) in patients without ENE (P < .001) (Table 2). Patients with ENE larger than 2 mm had significantly higher frequency of N2 disease compared with the group without ENE (42.1% vs 2.7%, respectively; P < .001), which was comparable to that of the group with ENE 2 mm or smaller. In terms of NSLN metastases, patients with ENE larger than 2 mm harbored NLSN metastases at nearly 3 times the rate of patients without ENE (68.4% vs 24.2%, respectively; P < .001). The mean (SD) numbers of NSLN metastases were 0.48 (1.30) for those with no ENE and 2.95 (4.95) for those with ENE larger than 2 mm (P < .001).

Disease recurrence and overall survival tended to be poorer for patients with ENE larger than 2 mm compared with those without ENE. Most patients in both groups did not experience recurrence. Patients with ENE larger than 2 mm had a distant recurrence rate of 10.5% compared with 2.7% in those without ENE (P = .19) (Table 3). In terms of survival, 9 of 38 patients (23.7%) with ENE larger than 2 mm died during the study period, compared with 18 of 149 patients (12.1%) in the group without ENE (P = .07). Unadjusted Kaplan-Meier analysis shows that overall survival of patients with ENE larger than 2 mm was significantly worse than in patients without ENE (P < .001) (Figure 2). However, on multivariate analysis, ENE larger than 2 mm was not significantly associated with increased mortality (HR =1.14; 95% CI, 0.58-2.24; P = .70) (Table 4). Both increasing tumor size (HR = 1.85; 95% CI, 1.12-3.07; P = .02) and having more than 3 positive LNs (HR = 3.85; 95% CI, 1.76-8.40; P < .001) were independent predictors of increased mortality (Table 4). Receipt of adjuvant hormonal therapy was protective (HR = 0.36; 95% CI, 0.19-0.68; P = .002) (Table 4).

The results of the ACOSOG Z0011 trial have significantly changed the surgical management of patients with T1 and T2 breast cancers with limited axillary nodal burden detected by SLND. The Z0011 trial demonstrated that ALND could be safely omitted in a specific subgroup of patients who met the inclusion criteria outlined by the study. With the increasing adoption of Z0011 recommendations, it is important to reconcile the management of patients who otherwise meet Z0011 inclusion criteria but have certain high-risk features.

The incidence of ENE found on SLND ranges from 24% to 40%.^3,6 Extranodal extension has been previously shown to be associated with increased incidence of NSLN metastases after initial SLND by numerous investigators.^4,6-8 It has also been linked to increased recurrence and decreased survival in other series, including our own previously published work.^5,9,10 Because patients with positive SLNs routinely underwent completion ALND regardless of ENE status in the past, the presence of ENE and its association with increased NSLN burden was of little clinical relevance. However, now that the ACOSOG Z0011 trial has shown that these selected patients still have excellent locoregional control and survival in the setting of limited residual axillary disease without further surgical intervention, do all patients with ENE really require and/or benefit from completion ALND? Additionally, the After Mapping of the Axilla: Radiotherapy or Surgery? (AMAROS) trial, a noninferiority study that assessed whether radiotherapy could provide locoregional control comparable to that of ALND in patients with SLN metastases, demonstrated very low 5-year axillary recurrence rates in both groups (1.19% with radiotherapy vs 0.43% with ALND). The AMAROS trial did not require reporting of ENE, but it is likely that a subset of the study patients had ENE. Twenty-five–year follow-up results of the NSABP-04 study showed that the locoregional recurrence rate in patients with positive nodes who underwent modified radical mastectomy or total mastectomy with radiation was not significantly different (16% vs 14%, respectively).¹¹ Results of these studies reinforce the finding that ART could achieve axillary disease control similar to that of ALND. The ACOSOG Z0011 study showed that a specific subset of patients with limited axillary nodal burden could avoid both ALND and nodal irradiation. Our objective was to investigate whether the subset could be broadened to a wider group of patients, specifically patients with microscopic ENE. Although patients with ENE 2 mm or smaller demonstrated outcomes similar to those of patients without ENE, the fact that this group still harbored N2 disease in the axilla on completion ALND at a rate similar to patients with ENE larger than 2 mm suggests that patients meeting Z0011 criteria with microscopic ENE may still warrant axillary therapy for locoregional control. Based on previously published nomograms of patients with breast cancer who had undergone ALND,^12,13 ENE was one clinicopathologic factor that was predictive of having 4 or more positive LNs, which radiation oncologists had used as a criterion for considering additional radiation fields prior to publication of the Z0011 trial results. Since the Z0011 trial, fewer patients are receiving ALND despite having positive SLNs and some radiation oncologists still recommend additional radiation fields when the risk of 4 or more positive nodes is greater than 30%.¹⁴ Our results showed that more than 40% of patients with ENE meeting Z0011 criteria with 1 or 2 positive SLNs went on to have N2 disease, making them candidates for additional radiation fields. Although the Z0011 trial does not provide us with information on the outcome of patients with ENE, the AMAROS results demonstrated that patients with positive LNs who undergo ART instead of ALND have the same recurrence and survival outcomes. Given that ART had better quality-of-life outcomes than ALND in the AMAROS trial, administration of additional radiation fields in patients with limited ENE may be an alternative option, although this certainly warrants additional study.

Only a few studies have attempted to correlate ENE size found in LNs with outcomes in patients with breast cancer. Most of these studies were performed on patients in the pre-SLND era, with all the patients receiving either radical or modified radical mastectomy.^9,10,15,16 Two published reports found that increasing ENE was correlated with increasing number of positive LNs,^10,15 but each of the studies had a different method of ENE quantification. The study by Palamba et al¹⁵ was the only one of these in which ENE size was actually measured, using a 1-mm cutoff.

A recently published study from Memorial Sloan Kettering Cancer Center examined residual axillary burden in patients with ENE on SLND who met Z0011 criteria.¹⁷ Using the same 2-mm ENE size criterion, our results corroborate their findings that increasing ENE size correlates to increasing tumor size, increasing number of positive SLNs, and increasing number of NSLN metastases at completion ALND. Although Gooch et al¹⁷ reported that 33% of patients with ENE larger than 2 mm and 9% of patients with ENE smaller than 2 mm had 4 or more additional positive LNs on completion ALND, the total number of positive LNs in the groups was not reported. Contrary to their findings, we found that approximately 30% of patients had 4 or more positive LNs on completion ALND and more than 40% had N2 disease in both the group with ENE 2 mm or smaller and the group with ENE larger than 2 mm. The recurrence and survival rates were not significantly different between patients with ENE 2 mm or smaller and patients without ENE. The overall rates of recurrence and survival were not statistically significantly different between the group with ENE larger than 2 mm and the group without ENE, and the rate of distant recurrence did not significantly differ between these groups (10.5% vs 2.7%, respectively; P = .19). In the multivariate analysis controlling for tumor size, number of positive LNs, and adjuvant therapies, both tumor size and number of positive LNs were stronger predictors of survival than ENE, suggesting that ENE may not affect survival so much as it represents a marker for poorer outcomes. Although further investigation with a larger sample is needed, our results suggest that considering the ENE size may potentially affect clinical management of these patients, as patients with ENE larger than 2 mm tended to have increased recurrence and poorer survival than patients without ENE.

Our study has several important limitations. Owing to its retrospective nature, we were limited in describing survival outcomes to overall survival, as only a subset of patients had known causes of death. Additionally, the small sample resulted in a low event rate for recurrence and survival, limiting the power of multivariate analyses. Despite these limitations, our study contributes valuable information to the small existing body of data regarding the clinical significance of ENE size in patients after SLND. Given that there is no standard for ENE reporting and that ENE reporting is not even required by the American Joint Committee on Cancer staging system or current College of American Pathologists guidelines,¹⁸ our data describe a reproducible method of ENE quantification that, taken together with the Memorial Sloan Kettering Cancer Center data, seems to provide a reasonable size cutoff that may be clinically significant.

Our study found that both ENE 2 mm or smaller and ENE larger than 2 mm demonstrated increased axillary nodal burden compared with patients without ENE on SLND in patients meeting Z0011 criteria, with ENE as a possible predictor of N2 disease. However, despite this increased nodal burden, disease recurrence and overall survival for ENE 2 mm or smaller more closely approximated that of patients without ENE, a group who would not typically receive any further axillary surgery since the Z0011 results were published. Quiz Ref IDWe recommend that reporting of ENE size detected on SLND should be incorporated into the College of American Pathologists reporting guidelines. If possible, reanalyzing cases from the Z0011 and AMAROS trials for microscopic ENE will help clarify some issues regarding clinical outcomes in patients with ENE. Validation of these findings with prospective data would help inform treatment decisions in this group of patients in an era when the Z0011 and AMAROS results are being used to better define groups of patients who can forgo further axillary surgery or be treated with ART as an alternative to ALND.

Corresponding Author: Maheswari Senthil, MD, Department of Surgery, Loma Linda University School of Medicine, 11175 Campus St, Ste 21108, Loma Linda, CA 92354 (msenthil@llu.edu).

Accepted for Publication: April 20, 2015.

Published Online: September 2, 2015. doi:10.1001/jamasurg.2015.1687.

Author Contributions: Drs Choi and Senthil had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Choi, Garberoglio, Senthil.

Acquisition, analysis, or interpretation of data: Choi, Blount, Perez, Chavez de Paz, Rodriguez, Surrusco, Lum, Senthil.

Drafting of the manuscript: Choi, Chavez de Paz, Rodriguez, Senthil.

Critical revision of the manuscript for important intellectual content: Choi, Blount, Perez, Surrusco, Garberoglio, Lum, Senthil.

Statistical analysis: Chavez de Paz, Rodriguez, Lum.

Administrative, technical, or material support: Choi, Perez, Surrusco, Lum, Senthil.

Study supervision: Garberoglio, Lum, Senthil.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This paper was presented at the 86th Annual Meeting of the Pacific Coast Surgical Association; February 22, 2015; Monterey, California.