Annals of Surgical Oncology

, Volume 20, Issue 8, pp 2556–2561

The Exportability of the ACOSOG Z0011 Criteria for Omitting Axillary Lymph Node Dissection After Positive Sentinel Lymph Node Biopsy Findings: A Multicenter Study

Authors

    • Department of Breast Medical OncologyThe University of Texas MD Anderson Cancer Center
    • Department of Gynecology and ObstetricsLariboisiere Hospital/AP-HP, The University Denis Diderot
  • Alexandre Bricou
    • Department of Gynecology and ObstetricsJean Verdier Hospital/AP-HP, The University Paris Nord
  • Ruben Lousquy
    • Department of Gynecology and ObstetricsLariboisiere Hospital/AP-HP, The University Denis Diderot
  • Delphine Hudry
    • Department of Surgical OncologyGeorges Francois Leclerc Cancer Center, The University of Bourgogne
  • Clémentine Jankowski
    • Department of Surgical OncologyGeorges Francois Leclerc Cancer Center, The University of Bourgogne
  • Claire Willecocq
    • Department of Gynecology and ObstetricsJean Verdier Hospital/AP-HP, The University Paris Nord
  • Anne Thoury
    • Department of Gynecology and ObstetricsLariboisiere Hospital/AP-HP, The University Denis Diderot
  • Catherine Loustalot
    • Department of Surgical OncologyGeorges Francois Leclerc Cancer Center, The University of Bourgogne
  • Charles Coutant
    • Department of Surgical OncologyGeorges Francois Leclerc Cancer Center, The University of Bourgogne
  • Emmanuel Barranger
    • Department of Gynecology and ObstetricsLariboisiere Hospital/AP-HP, The University Denis Diderot
Breast Oncology

DOI: 10.1245/s10434-013-2917-6

Cite this article as:
Delpech, Y., Bricou, A., Lousquy, R. et al. Ann Surg Oncol (2013) 20: 2556. doi:10.1245/s10434-013-2917-6

Abstract

Purpose

To determine the exportability of the criteria defined by the American College of Surgeons Oncology Group Z0011 trial for selecting patients who are eligible for omitting completion axillary lymph node dissection (cALND) after a positive sentinel lymph node (SLN) biopsy result and to investigate whether not following the Z0011 criteria might affect patient outcomes.

Methods

From a multicenter database, we selected 188 patients with positive SLNs and then excluded patients with positive SLNs on immunohistochemistry only. We retrospectively applied the Z0011 criteria and grouped the patients as eligible or ineligible for omitting cALND. The eligible group was compared with the cohort included in the Z0011 trial and with the ineligible group. Kaplan–Meier survival curves were calculated for each group, and univariate analyses assessed associations between the groups and clinicopathological variables.

Results

The final analysis involved 125 patients with positive SLNs. Eighty-seven patients (69.6 %) were potentially eligible for omitting cALND. The estrogen receptor status, T stage, grade, and number of positive non-SLNs were not statistically different between the eligible group and the Z0011 cohort. The ineligible group had significantly more positive non-SLNs (P = 0.01) and a lower 5-year overall survival rate than the eligible group (P < 0.001).

Conclusions

The similarity of clinical characteristics between the Z0011 trial cohort and our eligible group confirms the exportability of these criteria to another population. The worse prognosis of patients who did not meet the Z0011 criteria suggests prudence before disregarding or enlarging broadening the indications for omitting cALND.

The management of the axilla in breast cancer patients has undergone unprecedented changes over the past decade. In addition to omitting completion axillary lymph node dissection (cALND) in patients with negative sentinel lymph node (SLN) biopsy findings, we are now also omitting cALND in patients with a low volume of disease in the SLNs. This new standard, now incorporated into the National Comprehensive Cancer Network’s (NCCN) guidelines, was supported by the American College of Surgeons Oncology Group (ACOSOG) Z0011 trial.1 The Z0011 trial demonstrated that patients with two or fewer positive SLNs did not benefit from cALND in terms of overall survival, disease-free survival, or locoregional recurrence.2,3 However, the trial concluded that only a subgroup of breast cancer patients is eligible for omitting cALND. This subgroup comprises patients who meet the inclusion criteria and the treatment protocol of the Z0011 trial: clinical stage T1 or T2, N0, M0 breast cancer; no palpable adenopathy; one or two positive SLNs; and treatment with breast-conserving therapy, whole breast irradiation, and adjuvant systemic therapy (chemotherapy and/or endocrine therapy). Patients who received neoadjuvant systemic therapy or additional radiation fields are not eligible to forgo cALND.

Several groups in the United States have already modified their practice based on the Z0011 criteria.4 However, practices in Europe have been slower to incorporate the Z0011 criteria into their standard of care for several reasons. Questions have been raised about whether data analyzing a US cohort are exportable to other populations and whether the adoption of these criteria might result in hazardous undertreatment.5 In addition to this debate, there are questions about the effect of omitting cALND without regard for the Z0011 criteria. A recent survey conducted by researchers at the University of Texas MD Anderson Cancer Center revealed that many physicians omit cALND in patients who do not meet all the criteria defined in the Z0011 trial.6 However, to date there are no data on omitting cALND in such patients, which might result in hazardous undertreatment.

To evaluate the exportability of the Z0011 criteria, we examined whether following the Z0011 criteria would select patients with different characteristics from those of the cohort included in the Z0011 trial. To establish the possible effects of omitting cALND without regard for the Z0011 criteria, we assessed the clinical characteristics and prognosis of patients ineligible for omitting cALND.

Patients and Methods

Patient Population

One hundred eighty-eight patients were identified for this study from prospectively maintained clinical databases at three different institutions: 46 patients from the department of gynecology at Lariboisiere Hospital (Paris, France; November 2007–January 2012), 42 patients from the department of gynecology at Jean Verdier Hospital (Bondy, France; February 2004 to January 2012), and 100 patients from the department of surgical oncology at George Francois Leclerc Cancer Center (Dijon, France; January 2000–December 2005). Patients were included if they had breast cancer without palpable adenopathy before surgery, had one or more positive SLNs on histological examination, and had not received any neoadjuvant systemic therapy. In order to define positive SLNs as they were defined in the Z0011 trial, 63 patients were excluded because their SLNs were found to be positive only after immunohistochemical staining. 125patients were included in the final analysis. After this selection, the patients were divided into two groups for statistical analysis. The eligible group comprised patients who met the Z0011 criteria for omitting cALND, and the ineligible group comprised patients who did not meet these criteria (Fig. 1). The institutional review board of the French College of Gynecology and Obstetrics approved this study.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-013-2917-6/MediaObjects/10434_2013_2917_Fig1_HTML.gif
Fig. 1

CONSORT diagram. JVR Jean Verdier Hospital, GFLCC George Francois Leclerc Cancer Center, LRB Lariboisiere Hospital, SLN sentinel lymph node, ALND axillary lymph node dissection, IHC immunohistochemistry, HES hematoxylin–eosin staining. *Six patients had 2 ineligibility criteria

Surgical Management and Pathological Assessment

In patients treated before 2009, SLN biopsy was performed if no previous axillary or breast surgery had been done, after breast carcinoma was histologically proven by preoperative core needle biopsy, and if the tumor was unifocal with a preoperative diameter <2 cm. After 2009, in accordance with “French Saint Paul de Vence referential,” SLN biopsy was performed on some patients who had had prior breast surgery and on patients with tumors up to 3 cm in diameter. All SLN biopsies were performed with radioisotope and/or blue dye injection.

All patients included in this study underwent cALND. cALND was performed during the breast surgery if imprint cytology or frozen section analysis revealed SLN metastases (91 out of 125 patients had immediate axillary lymph node dissection). cALND was performed as a separate procedure after the breast surgery if postoperative hematoxylin and eosin staining revealed SLN metastases (34 out of 125 patients had delayed axillary lymph node dissection). For patients undergoing lumpectomy, a second operation was performed to achieve clear margins if invasive or in situ carcinoma was observed within 2 mm of the surgical margin on microscopic examination (i.e., positive or close tumor margins). The decision whether to recommend mastectomy was made by the treating physician.

The pretreatment clinical tumor stage was determined by physical examination and mammography. Tumor size was determined by the final pathological report. Estrogen receptor (ER) positivity was defined as nuclear staining in at least 10 % of tumor cells. Histological grade was assessed by the modified Scarff-Bloom-Richardson (SBR) system. Non-SLNs were defined as nodes removed during cALND.

Comparison and Statistical Analysis

The eligible group was compared with the cohort included in the “ALND” arm of the Z0011 trial (n = 420, in intention to treat). The χ² test (or Fisher’s exact test when the sample size was small) was used to evaluate associations between the origin of the patients (eligible group vs. cohort from Z0011 trial) and the following categorical variables: clinical T stage, ER status, tumor grade, histological subtype, size of the SLN metastases, number of positive nodes, non-SLN status (positive vs. negative).

The eligible group was also compared to the ineligible group. The Wilcoxon rank sum test was used for comparing the following continuous variables: total number of nodes removed (SLNs plus non-SLNs), the total number of positive nodes, and the number of positive non-SLNs.

Overall survival was measured from the date of diagnosis of primary cancer to the date of death from any cause. Patients who were alive at last follow-up were censored in the analysis. Disease-free survival was calculated from the date of diagnosis to the date of local or distant recurrence. Survival curves for the eligible and ineligible groups were plotted using the Kaplan–Meier method and were compared using the log-rank test. Date of surgery of all ineligible patients had a homogenous repartition during the time period of the study.

All analyses were performed using the R package with survival and rms libraries (Comprehensive R Archive Network, http://lib.stat.cmu.edu//R/CRAN/). A P value of <0.05 was considered significant.

Results

After applying the Z0011 criteria for omitting axillary lymph node dissection to the 125 patients included in this analysis, we assigned 87 patients (69.6 %) to the eligible group and 38 patients (30.4 %) to the ineligible group (Fig. 1).

Grade 1 tumors were observed in a higher percentage of the eligible group than in the cohort included in the Z0011 trial (Table 1). The clinical stage, ER status, and histological subtype were not statistically significant between these cohorts (Table 1). Regarding the extent of the disease in the nodes, patients in the eligible group were less likely to have micrometastases (e.g., <2 mm) (P = 0.01) than were those included in the Z0011 trial (Table 2). However, the total number of positive nodes and the number of patients with positive non-SLNs were not statistically different between the two cohorts (Table 2).
Table 1

Demographic and clinical characteristics of patients eligible for omitting ALND according to ACOSOG Z0011 criteria in the current study and in the ALND arm of the Z0011 trial

Characteristic

Current study eligible group (n = 87)

Z0011 ALND arm (n = 420)a

Pb

Age, years, median (range)

56 (27–87)

56 (26–92)

Clinical T stage

 T1

65 (74.7 %)

284 (67.6 %)

0.2

 T2

22 (25.3 %)

134 (31.9 %)

 

 Missing

2

2

 

 Tumor size, cm, median (range)

1.6 (0.5–4.8)

1.7 (0.4–7.0)

Estrogen receptor status

 Positive

78 (89.7 %)

317 (75.5 %)

0.1

 Negative

8 (9.2 %)

66 (15.7 %)

 

 Missing

1 (1.1 %)

37 (8.8 %)

 

Tumor grade

 1

32 (36.8 %)

71 (16.9 %)

0.01

 2

35 (40.2 %)

158 (37.6 %)

 

 3

19 (21.8 %)

94 (22.4 %)

 

 Missing

1 (1.1 %)

97 (23.1 %)

 

Histological subtype

 Ductal carcinoma

73 (83.9 %)

344 (81.9 %)

0.06

 Lobular carcinoma

10 (11.5 %)

27 (6.4 %)

 

 Other

4 (4.6 %)

45 (10.7 %)

 

 Missing

0 (0 %)

4 (0.1 %)

 

ALND axillary lymph node dissection, ACOSOG American College of Surgeons Oncology Group

aPatients in the ALND cohort, intention to treat

bP value evaluable only for categorical covariates (χ² test)

Table 2

Extent of disease in SLNs and non-SLNs for patients eligible for omitting ALND according to ACOSOG Z0011 criteria in the current study and in the ALND arm of the Z0011 trial

Characteristic

Current study eligible group (n = 87)

Z0011 ALND arm (n = 420)a

P

Size of SLN metastasis

 Micro

19 (22 %)

137 (33 %)

0.01

 Macro

65 (75 %)

228 (54 %)

 

 Unknown

3 (3 %)

55 (13 %)

 

No. of positive nodesb

 0

0 (0 %)

4 (1 %)

0.7

 1

48 (55 %)

199 (47 %)

 

 2

20 (23 %)

68 (16 %)

 

 ≥3

16 (18 %)

72 (17 %)

 

 Unknown

3 (3 %)

77 (18 %)

 

Non-SLN status

 Positive

25 (29 %)

97 (23 %)

0.4

 Negative

59 (68 %)

291 (69 %)

 

 Unknown

3 (3 %)

32 (8 %)

 

SLN sentinel lymph node, ALND axillary lymph node dissection, ACOSOG American College of Surgeons Oncology Group

aPatients in the ALND cohort, intention to treat

bTotal number of positive nodes: SLN and non-SLN

The ineligible group had a significantly higher median total positive nodes than did the eligible group as well as a higher percentage of patients with three or more positive non-SLNs (26 vs. 7 %, respectively) (Table 3).
Table 3

Number and extent of disease of lymph nodes by eligibility for omitting ALND according to Z0011 criteria

Demographic or clinical characteristic

Eligible group

Ineligible group

P

(n = 87)

(n = 38)

Total no. nodes removed, median (range)

12 (2–32)

12 (3–29)

0.6

Total no. positive nodes, median (range)

1 (1–14)

2.5 (1–15)

<0.001

No. positive non-SLNs, n (%)

 0

59 (68 %)

22 (58 %)

0.01

 1

14 (16 %)

3 (8 %)

 

 2

5 (6 %)

1 (3 %)

 

 ≥3

6 (7 %)

10 (26 %)

 

 Unknown

3 (3 %)

2 (5 %)

 

ALND axillary lymph node dissection, SLN sentinel lymph node

The median follow-up time was 72 months (range, 1–131 months). Fourteen patients died, and 15 experienced disease recurrence. The 5-year overall survival rates were 96 % (95 % confidence interval [CI] 91–100) and 81 % (95 % CI 64–90), respectively, for the eligible and ineligible groups (P < 0.0001; Fig. 2). The 5-year disease-free survival rates were 91 % (95 % CI 83–98) and 69 % (95 % CI 52–92), respectively, for the eligible and ineligible groups (P = 0.006).
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-013-2917-6/MediaObjects/10434_2013_2917_Fig2_HTML.gif
FIG. 2

a Overall and b disease-free survival by Z0011 eligibility criteria for omitting completion axillary lymph node dissection in patients with positive sentinel lymph nodes

Discussion

We found no significant difference in clinical characteristics between our group of patients who were potentially eligible for omitting cALND according to the Z0011 criteria and the population reported in the Z0011 trial. This finding supports the safe exportability of the Z0011 criteria. In our comparison of the characteristics and prognosis of the patients who were not eligible for omitting cALND with those who were eligible, the ineligible group had a worse prognosis and a significantly higher percentage of patients with advanced disease in the axilla. This result confirms the importance of strict adherence to the Z0011 criteria, and we can infer that omitting cALND in patients deemed ineligible for its omission by the Z0011 criteria may result in hazardous undertreatment.

Very few authors have reported the exact proportion of patients eligible for avoiding cALND after incorporating the Z0011 criteria into their standard of care.7 Our study found that 69 % (87 of 125) of patients with positive SLNs were eligible for omitting cALND. This high number of patients illustrates the importance of Z0011 as a practice-changing trial.8 While many authors have addressed these changes in practice, we think that the most important changes will be for the patients. It is well documented that SLN biopsy is associated with less morbidity than cALND.6,912 SLN biopsy has improved patients’ quality of life and markedly reduced the incidence of lymphedema.1315 Thus, we could postulate that the benefit of omitting cALND in 69 % of patients with positive SLNs will be markedly reduced morbidity and improved quality of life after breast-conserving therapy.

One criticism of the Z0011 trial was a potential bias related to the enrollment of only patients with a good prognosis.16 The 5-year overall survival rates reported were 92.5 % (95 % CI 90.0–95.1) in the SLN group and 91.8 % (95 % CI 89.1–94.5) in the cALND group. However, we reported a very similar overall survival rate after applying the Z0011 criteria to a different population of patients with positive SLNs. The similar survival rates between our eligible group and the Z0011 cohort are clearly consistent with our finding no significant differences in the ER status, tumor grade, and T stage, which are very important prognostic factors.17 Moreover, we compared the proportions of patients with positive non-SLNs, which can be considered an estimation of the false negative rate of the Z0011 criteria, and found a similar proportion of patients with positive non-SLNs in our cohort. These results support the robustness of the Z0011 criteria. However the very low number of event in the eligible group confirms that a “Z0011” trial should include a very high number of patients in order to determine the true benefit of cALND in patients with subclinical nodal metastases.

To define positive SLNs as they were defined in the Z0011 trial, we chose to exclude 63 patients with SLN metastases diagnosed after immunohistochemical staining. Most of these patients had micrometastatic disease. The optimal management for patients with micrometastatic disease will probably be answered by the IBCSG trial, which was closed in February 2010.18 However, it is reasonable to presume that these patients would not benefit from a cALND. Micrometastasis represents a small volume of disease in the axilla, and the rate of non-SLN involvement is as low as 20 %.19 As a result, many clinicians have already omitted cALND in patients found to have micrometastatic disease by SLN biopsy. A 2008 Surveillance, Epidemiology, and End Results database study by Rescigno et al. 20 reported that 58 % of such patients do not undergo cALND. By contrast, some retrospective studies suggest that some subgroups of patients with micrometastatic disease in the SLNs may benefit from cALND.21,22 As noted in patients included in the Z0011 trial, adjuvant systemic and/or radiotherapy may benefit patients with micrometastatic disease in the SLNs in whom cALND is omitted. Further research should address this important point.

Our findings that the ineligible group had a much worse prognosis than the eligible group and had a high volume of residual disease in the axilla after cALND are critical because the next logical step in the evolution of axillary staging is to ask whether all SLN-positive patients require cALND. Our results suggest that the omission of cALND cannot be extended to all patients with positive SLNs. However, not all of the Z0011 criteria may have the same impact on patient prognosis. Some reports have already supported omitting cALND in patients undergoing mastectomy or partial breast irradiation. For example, Setton et al. 23 recently analyzed the radiation field design in 326 patients with positive SLNs who did not undergo cALND. The rate of regional control was high even in patients who received partial breast irradiation only. Additionally, Milgrom et al. reported on 210 patients who underwent total mastectomy and had a positive SLN but did not undergo cALND. The 4-year local, regional, and distant failure rates were 1.7, 1.2, and 0.7 %, respectively.24 Supporting the idea of extending of the indications for omitting cALND is our finding that 58 % of the ineligible patients had no positive non-SLNs. Nomograms based on clinical covariates have demonstrated their power for predicting which patients with positive SLNs will have fewer positive non-SLNs and should be investigated.25,26 Further studies could more clearly establish to which patients omitting cALND might be extended.

This work has some limitations. Because this was a retrospective analysis with small dataset, we could not prevent some bias linked to patient selection. The fact that we selected consecutive patients from three different centers may reduce this bias. One could argue that we have included patients with good prognosis because of the very strict criteria applied in France for performing SLN biopsies. However, we think that these strict criteria reflect everyday practice. We found also a high proportion of patients eligible for omitting cALND in part because of the low rate of mastectomy in France compared with other countries.27 It is possible that the rate of patients who are eligible for omitting cALND fluctuates in accordance with the local rate of mastectomy.

In conclusion, this work has demonstrated the exportability of the Z0011 criteria to another population of patients with positive SLN biopsy findings. We also reported the worse prognosis of patients who did not meet the Z0011 criteria, which might result in hazardous undertreatment if the criteria are not followed. The Z0011 criteria have changed everyday practice by demonstrating that cALND does not play an important role in the treatment of patients with a small volume of metastases in the SLN. Because many patients are eligible for and could benefit from avoiding cALND, the incorporation of the Z0011 criteria should be considered by all clinicians. The exact impact of adjuvant systemic and/or radiotherapy on remaining node metastases in the axilla is now addressed. A better understanding of these interactions can help in the prudent extension of the indications for omitting cALND in patients with positive SLNs.

Acknowledgment

Conflict of Interest

The authors declare no conflict of interest.

Copyright information

© Society of Surgical Oncology 2013