- Email: [email protected]
Validation of a breast cancer nomogram for predicting nonsentinel node metastases after minimal sentinel node involvement: Validation of the Helsinki breast nomogram
The Breast 22 (2013) 787e792
Contents lists available at SciVerse ScienceDirect
The Breast journal homepage: www.elsevier.com/brst
Original article
Validation of a breast cancer nomogram for predicting nonsentinel node metastases after minimal sentinel node involvement: Validation of the Helsinki breast nomogram Lobna Ouldamer a, b, c, *, Flavie Arbion d, Alexis Balagny a, c, Florence Fourquet e, Henri Marret a, c, Gilles Body a, c a
Department of Gynecology, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France INSERM Unit 1069, Tours, France François Rabelais University, Tours, France d Department of Pathology, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France e Department of Biostatistics, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 September 2012 Received in revised form 28 December 2012 Accepted 6 February 2013
Background: Complete node dissection for tumor-positive sentinel lymph nodes (SLN) is becoming more controversial. Nevertheless, current practice guidelines still recommend complete axillary lymph node dissection (ALND) for breast cancer patients whose SLN contains a metastatic tumor. The Helsinki breast cancer nomogram developed by Meretoja TJ et al. aims to predict the risk of positive non-sentinel lymph nodes in patients with minimal sentinel node involvement, it uses tumor diameter and multifocality. The purpose of this study was to test the accuracy of the nomogram among patients with micrometastatic SLN-positive biopsy findings. Methods: The Helsinki nomogram was used to calculate risk of metastases for 49 consecutive patients with SLN micrometastases or isolated tumor cells (ITC) who underwent complete ALND. The nomogram was evaluated by calculating the area under the receiver-operator characteristic (ROC) curve. Results: The area under the ROC curve for the nomogram applied to all patients with micrometastases and ITC was 0.72 (range 0.60e0.85) (0.791 in the original publication). Conclusions: The Helsinki breast cancer nomogram is a useful tool for patients with minimal sentinel node involvement. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Breast cancer Sentinel lymph node Nomogram Isolated tumor cells Micrometastasis Nonsentinel node metastases
Introduction Completion axillary lymph node dissection (ALND) has been considered to be the gold standard for cases in which a metastasis is found in the sentinel lymph node (SLN).1 ALND remains the standard approach for SLN-positive patients recommended by the American Society of Clinical Oncology and the French guidelines. However, approximately 50%e70% of patients with diseasepositive SLN have no additional positive nodes, which means that it may be possible to avoid axillary dissection in selected patients.2e8 There has been discussion in the literature about whether systematic ALND for patients with positive SLNs is appropriate. In the
* Corresponding author. Department of Gynecology, CHU Bretonneau, 2 Boulevard Tonnelé, 37000 Tours, France. Tel.: þ33 02 47 47 47 47; fax: þ33 02 47 47 38 01. E-mail address: [email protected] (L. Ouldamer). 0960-9776/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.breast.2013.02.004
randomized American College of Surgeons Oncology Group (ACOSOG) Z0011 trial, patients with a positive SLN were randomized into two groups: complete ALND or axillary observation. There was similar outcome in the ALND and SN biopsy-only arms, despite a nonsentinel node involvement rate of 27% in the ALND arm suggesting that complete ALND after positive SN may not improve survival. However, the study closed prematurely due to poor recruitment, so it is unlikely that the number of subjects successfully enrolled in the study will provide adequate power to show a difference in survival, even if any such difference exists.9 Also, all patients in this trial underwent breast-conserving treatment with whole-breast radiation. So, it is unclear whether these results can be generalized to patients undergoing mastectomy without radiotherapy. Metastases are found in non-SLNs in approximately 10% of patients with isolated tumor cells (ITC) in the SLN and in 20%e35% of patients with micrometastases in the SLN10; complete ALND is routinely performed in these patients.11
788
L. Ouldamer et al. / The Breast 22 (2013) 787e792
Meretoja et al. developed a nomogram to predict the likelihood of finding additional positive nodes on ALND in patients with micrometastasis or ITC in the SLN. The nomogram predicts individual patient risk and was developed from a logistic-regression analysis and included only two statistically significant variables: tumor diameter and multifocality. The receiver-operating characteristic (ROC) curve for this tool was 0.791 indicating good prediction and discrimination.12 We therefore decided to apply this nomogram to our SLNpositive series and evaluated its usefulness for the subgroup of patients with micrometastatic SLNs. Patients and methods A total of 1283 consecutive cases of SLN biopsy from primary untreated breast cancer patients at the tertiary oncological referral center of Tours (France) were prospectively entered into a multidisciplinary team meeting register between January 2004 and December 2011. With the authorization of the institutional review board we retrospectively reviewed this prospective database. We studied a subset of 187 cases that fulfilled the following inclusion criteria: primary invasive breast carcinoma with clinically negative axillary lymph nodes and no prior systemic therapy; successful SLN biopsy and metastatic disease identified by any method (routine histopathology, serial sectioning, staining with hematoxylin and eosin [HE], immunohistochemistry [IHC]); and complete ALND. Exclusion criteria were: neoadjuvant chemotherapy, inflammatory breast cancer, ductal carcinoma in situ, failed SLN mapping, clinically suspicious axillary lymph nodes and bilateral procedures. Both blue dye (patent blue, Laboratoire Guerbet, Paris, France) and radioisotopes (99mTc-labeled colloid sulfur, Nanocoll, General Electric Healthcare SA, USA) were used for lymphatic mapping with SLN biopsies. Static lymphoscintigraphy was performed to visualize the probable localization of SLN, and a hand-held gamma detection probe (Gamma-sup/Clerad/France) was used intraoperatively to identify the most radioactive area. SLN were identified as nodes with blue dye uptake, radiotracer uptake, or both. Histopathological evaluation Fresh tissue labeled as SLN was delivered without fixation to the pathology lab. Each lymph node was bisected along its major axis. One half of the nodal tissue was frozen, sliced into 5 mm-thick sections and stained with blue methylene and with hematoxylin and eosin (HE) for intraoperative consultation. In cases with technical limitations, such as incomplete representation of the tissue or folding, additional sections were obtained at the discretion of the consultant pathologist. Upon completion of the intraoperative consultation, tissue samples were fixed in 10% buffered formalin and then embedded in paraffin. Three slides were prepared: two of sections taken at 200micron intervals were stained with HE, and a conservative section of a deeper level was studied for pancytokeratin by immunohistochemistry. Immunohistochemistry for cytokeratin AE1/AE3 was used to confirm suspicious cells in the SLN. Following the American Joint Committee on cancer recommendations, macrometastases were defined as tumors >2 mm long and micrometastases as tumors between 0.2 and 2 mm long. Tumors <0.2 mm long were regarded as isolated tumor cells. If a frozen section was positive, ALND was performed during the same surgical intervention. For patients with SLN metastases but
that were not identified in interoperative frozen sections, ALND was performed at a later date. The Helsinki breast cancer nomogram was used to predict the individual risk of non-SLN positivity from patient data. The nomogram was developed a to predict the likelihood of finding additional positive nodes on ALND in patients with micrometastasis or ITC in the SLN. First, univariate analysis was performed to determine factors associated with non SLN metastases, all variables (tumor diameter, multifocality, lymphovascular invasion, and tumor location) with p value less than 0.15 were then included into a logistic regression analysis using backward stepwise method. Variables with p value <0.05 were considered statistically significant in the multivariate analysis. Tumor diameter (p ¼ 0.002) and multifocality (p ¼ 0.039) were found to be the only statistically significant variables and were included in the final predictive model. With a 10% cutoff value for predicted probability the model had sensitivity of 38.2% and specificity of 85.7% and selects 84.1% of patients to have predicted risk of less than 10% for additional metastases showing good clinical utility.12 For nomograms, 10% or less cutoff values are considered in the literature to define the subgroup of patients with a low predicted probability of metastatic non-SLN (having a probability of metastatic non-SLN 10% predicted by nomogram). Statistical analyzes were performed by R 2.13.1 (http://www. cran.r-project.org/). For numeric data, results are reported as mean and median values standard deviation (SD). Numeric data were analyzed with Student’s t-test if normally distributed, and the ManneWhitney test if not. Categorical data were analyzed with the chi-squared test or with Fisher’s exact test. To measure discrimination by the nomogram, a receiveroperating characteristic (ROC) curve was constructed. The ROC curve assesses the relationship between the sensitivity and the false positive rate (1- specificity) of a test across all possible threshold values that define the positivity of a disease or condition. The overall accuracy of the nomogram is expressed as the area under the ROC curve (AUC). The AUC can be interpreted as the probability of a correct assignment of disease presence in random pairs of patients, one patient in each pair who has disease and one who does not. It is generally accepted that AUC values of 0.7e0.8 indicate satisfactory discrimination, and that AUC values >0.8 represent good discrimination. A calibration plot was drawn showing the actual (women who actually had positive non-SNs) versus nomogram-predicted probability. Results Between January 2004 and December 2011, metastatic disease was found in 216 patients; 29 (13 isolated tumor cells, 10 micrometastasis and 6 macrometastasis) of these patients did not undergo ALND (presence of co-morbidities or a history of controlateral ALND or patient choice) and were excluded from this analysis. Thus 187 cases met our inclusion criteria, and they included a subgroup of 49 patients with SLN micrometastases or isolated tumor cells (ITC). Descriptive clinical and histopathological data for the study population are listed in Table 1. The mean age was 57 years (range, 30e84 years). The mean pathological tumor size was 15.33 mm (range 4e 34 mm). The mean number of SLN identified was 2.37 (range 1e6), and the mean number of nodes extracted at ALND was 11 (range 1e31). The characteristics of positive-ALND patients are showed in Table 2. The nomogram was applied to the micrometastatic SLN and ITC cohort (n ¼ 49). The predicted probability of non-sentinel metastases for each patient are showed in Table 3.
L. Ouldamer et al. / The Breast 22 (2013) 787e792
calculated for each patient (Fig. 2). We found a good correspondence between the predicted probabilities and observations.
Table 1 Patient and tumors characteristics. Characteristics Patient age (years) Mean [range] Menopausal status Post menopausal Pre menopausal Histological diameter, mm Mean [range] Histology of the primary tumor Invasive ductal carcinoma Invasive lobular carcinoma Any other histology Histological grade of the tumor I II III Multifocality Yes No Lymphovascular invasion Yes No Estrogen receptor status Positive Negative Progesterone receptor status Positive Negative HER-2 status Positive Negative Number of SLN harvested on SN biopsy Mean [range] SLN ratio (positive/all SLN) <0.5 0.5e1 1 Size of the SLN metastasis Macrometastasis Micrometastasis Isolated tumor cells Detection method of the SLN metastasis Frozen section analysis Hematoxylin and eosin staining Immunohistochemistry Number of non-SLN harvested in ALND Mean (range) ALND status Positive Negative Number of positive non-SLN Mean (range)
789
Micrometastasis or ITC n ¼ 49
All patients (n ¼ 187)
57.8 [34e84]
57.12[30e84]
36 (73.5%) 13 (26.5%)
129 (69%) 58 (31%)
13.5 [4e30]
15.33[4e34]
44 (89.8%) 4 (8.2%) 1 (2%)
168 (89.9%) 13 (6.9%) 6 (3.14%)
12 (24.5%) 27 (55.1%) 10 (20.4%)
43 (23%) 105 (56.15%) 39 (21%)
5 (10.2%) 44 (89.8%)
26 (13.9%) 161 (86.1%)
13 (26.5%) 36 (73.5%)
69 (36.9%) 118 (63.1%)
44 (89.8%) 5 (10.2%)
173 (92.51%) 14 (7.49%)
40 (81.6%) 9 (18.4%)
157 (83.95%) 30 (16.04%)
4 (8.2%) 45 (91.8%)
29 (15.5%) 158 (84.5%)
2.38 [1e6]
2.37 [1e6]
16 (32.6%) 17 (34.8%) 16 (32.6%) 0 45 (91.8%) 4 (8.2%)
Application of other nomograms to the micrometastatic and ITC SN subgroup We also applied the Marseille nomogram to our minimal SLN involvement population: the AUC was 0.57 ([0.431e0.709]; 95% confidence interval). Other widely used predictive models (mskcc, tenon score) were also applied, to calculate either predicted probability for additional metastases or a score given by the model. For each of these models, we determined the AUC, the sensitivity, specificity, and clinical utility for either a cutoff value of 10% predicted probability or a cutoff threshold value given by the scoring system (Table 4). Discussion There is currently a debate about the necessity for complete ALND in every patient with metastatic SLN. Indeed, complete node dissection for a tumor-positive SLN is becoming more controversial, in view of the findings of numerous single-center studies, a large nonrandomized study and the Z0011 study by the American College of Surgeons Oncology Group (ACOSOG). The ACOSOG Z0011 study is a prospective study of patients with T1 or T2, clinically nodenegative breast cancers treated with lumpectomy, SLN biopsy, and whole-breast irradiation. Patients with one or two positive SLNs were randomized to complete ALND or no further axillary treatment. After a median follow-up of 6.3 years, the 5-year overall survival was similar in the two groups (91.8% for the ALND and 92.5% for the SLN biopsy groups). Although 27% of patients in the Z0011 trial had additional nodal metastases, complete ALND did not result in fewer locoregional recurrences or improved survival, suggesting that not all non-SLN metastases develop into clinically significant disease.9,13e20 There are several reasons for the low axillary recurrence rate, including the study excluding patients with extensive axillary disease or with 3 tumor-positive SLNs found on SN biopsy. Even for patients in whom there are residual regional lymph node metastases, the additional benefit of ALND is unclear because these patients will generally receive systemic therapy based on the presence of SLN metastases. For example, in the ACOSOG Z0011 trial, more than 95% of the patients in both arms received adjuvant therapy. Such adjuvant therapy may be one cause of the low rate of axillary recurrence. Also, all patients in this trial received whole breast radiation therapy. It is currently unclear how partial breast radiation therapy affects axillary recurrence rates in those who do not undergo complete ALND. Indeed, it is difficult to distinguish between the effects of surgery, chemotherapy, radiotherapy and antihormonal treatment on local and distant relapses. Reported axillary recurrence rates range from 0 to 2.6% with median follow-up of 2e2.5 years in patients with tumor-positive SLN who do not undergo complete ALND.9,18e28
52 (27.81%) 60 (32.1%) 75 (40.1%) 138 (73.8%) 45 (24.1%) 4 (2.1%) 99 (52.94%) 73 (39%) 15 (8.02%)
40 (81.6%) 9 (18.4%) 12.7 [5e24]
11 [1e31]
6 (12.2%) 43 (87.8%)
51 (27.27%) 136 (72.72%)
1.17 [1e2]
1 [1e18]
Application of the Helsinki nomogram to the micrometastatic and ITC SN subgroup (n ¼ 49) A ROC was calculated to assess discrimination by the Helsinki nomogram (Fig. 1). The area under the ROC was 0.72 (95% confidence interval [95% CI], 0.60e0.85) compared with 0.791 in the original study. To assess the accuracy of the Helsinki nomogram, actual probabilities were plotted against the predicted probability Table 2 Characteristics of positive-ALND patients. Patient
Age (years)
Tumor diameter (mm)
Multifocality
Number of SN
Number of positive non-SN
Size of the non-SN metastasis
1 2 3 4 5 6
48 51 41 43 50 84
19 6 15 15 30 10
Yes Yes No No No No
1 3 1 1 1 1
1 2 1 1 1 1
Micrometastasis Macro and micrometastasis Macrometastasis Macrometastasis Macrometastasis Macrometastasis
790
L. Ouldamer et al. / The Breast 22 (2013) 787e792
Table 3 The predicted probability of non-sentinel metastases for each patient. Unifocal tumor Tumor size (mm) Risk of additional metastases (%) Multifocal tumor Tumor size (mm) Risk of additional metastases (%)
4 3.6
6 3.8
6 8.3
7(n ¼ 2) 3.9
8(n ¼ 3) 4
9 4.2
10(n ¼ 7) 4.3
12(n ¼ 4) 4.5
10(n ¼ 2) 9.3
Some authors suggest that patients with micrometastases or isolated tumor cells in their SLN should not be treated by complete ALND.28e30 Until new randomized controlled trials are conducted, the results of retrospective studies, patient preference, and the statistical likelihood of non-SLN metastases, will have to be used as the basis for deciding which patients can avoid completion ALND. French guidelines and the American Society of Clinical Oncology recommend complete ALND for breast cancer patients whose SLN contain metastatic tumor. Freely available tools that can predict an individual’s clinical outcome, and have been validated internationally, are required. Nomograms for many type of cancer have become popular in recent years for treatment choices and assessments of risk, because of their simplicity and easy graphical representation.
13(n ¼ 3) 4.6
14(n ¼ 2) 4.8
15(n ¼ 9) 5
17 5.2
15 10.7
18 5.4
19(n ¼ 2) 5.5
20(n ¼ 5) 5.7
23 6.1
30 7.6
19 12
Only two of the existing predictive models for breast cancer are based on patient populations with only micrometastasis or ITC in SLN.12,31 The Marseille nomogram did not display sufficient accuracy in our study population to be applicable in clinical practice. However, the sample size in our study was small, and this may have reduced the predictive accuracy. The Helsinki model is simple, uses routinely available clinicopathologic information, and yields an AUC of 0.791 among the Helsinki patients used to build the model. Another potential strength of this model is that it does not require histological detection as a nomogram variable, thereby avoiding possible confusion arising from the classification or results of the detection method associated with differing tissue processing of SLN at different institutions.
Fig. 1. Receiver operating characteristic curve calculation for micrometastatic and ITC sentinel lymph nodes patients.
Fig. 2. Calibration plot for micrometastatic and ITC sentinel lymph nodes patients.
L. Ouldamer et al. / The Breast 22 (2013) 787e792
791
Table 4 Results of the application of 5 nomograms. Study
Year
Cut threshold
Sensitivity (%)
Specificity (%)
NPV (%)
PPV
AUC (95% IC)
Clinical utility
Marseille MSKCC Tenon Helsinki Helsinki
2008 2003 2005 2011 2011
10% 10% 3.5 10% 7%
50 50 66.67 17 50
39 57.14 88.37 98 95.35
20 11.11 5 10.64 6.82
13.79 14.29 44.44 50 60
0.57 0.61 0.657 0.725 e
40.81 56.25 85.71 87.75 89.79
(0.431e0.709) (0.473e0.747) (0.60e0.85) (0.60e0.85)
NPV: negative predictive value. PPV: positive predictive value.
Here, we report the first external validation of the Helsinki nomogram. Our overall findings indicate a satisfactory predictive accuracy of the Helsinki nomogram in our cohort with a statistically significant AUC value of 0.72 ( the AUC value is a measure of the overall discrimination of a model, and values of 0.7e0.8 are considered good, whereas values of 0.81e0.90 are considered excellent32). Based on our results, the false negative rate with The Helsinki nomogram was 10%, and it was able to correctly predict 43 (87.75%) of the patients. Using the threshold of 10%, two patients would have been operated with an ALND of which one would have a non-SLN metastasis and 47 patients would have been spared the ALND, 5 (10.6%) of those would have a non SLN metastasis. If the threshold is lowered to 7%, six patients would have been operated with an ALND of which three would have a non-SLN metastasis and 43 patients would have been spared the ALND, 3 (7%) of those would have a non-SLN metastasis. This mean that nine of every ten SLN-positive patients can be spared ALND with the use of this model; this is, in our opinion, of considerable clinical value, providing a safe and reliable tool that can be used to avoid potentially unnecessary surgical procedures. It has been reported that in 98% of node-positive patients who have multiple SN removed, metastasis is detected in the first three SLN.33 In our study, a mean of 2.37 SLN were retrieved per patient, and it could be argued that resecting additional nodes would result in less disease in the residual non-SLN. In most published large series, removal of more than one SN is strongly associated with a decreased false-negative rate.34e36 If fewer SLN are retrieved, therefore, there may be more residual disease in non-SLN and the nomogram will underestimate the risk of residual disease in these populations. In our 6 positive-ALND patients, five had only one SN. This issue requires further investigation. Our results show that the nomogram is accurate in our population of breast cancer patients. Although our study included only a small number of patients, it validates the predictive ability of the nomogram in a community breast cancer center. Predicting the risk of additional nodal metastases allows the surgeon and patient to make an individualized decision regarding the need for ALND. We report that the Helsinki nomogram provides patients and their physicians with a meaningful tool for assessing an individual’s risk of non-SLN metastases.
Conflict of interest statement
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
We declare that we have no conflict of interest. 21.
References 22. 1. Kim I, Giuliano AE, Lyman GH. Lymphatic mapping and sentinel lymph node biopsy in early-stage breast carcinoma: a metaanalysis. Cancer 2006;106:4e16. 2. Bolster MJ, Peer PGM, Bult P, Thunnissen FBJM, Schapers RFM, Meijer JWR, et al. Risk factors for non-sentinel lymph node metastases in patients with breast
23.
cancer. The outcome of a multi-institutional study. Ann. Surg. Oncol. janv 2007;14(1):181e9. Changsri C, Prakash S, Sandweiss L, Bose S. Prediction of additional axillary metastasis of breast cancer following sentinel lymph node surgey. Breast J 2004;10:392e7. Farshid G, Pradhan M, Kollias J, Gill PG. A decision aid for predicting nonsentinel node involvement in women with breast cancer and at least one positive sentinel node. Breast 2004;13:494e501. Fleming FJ, Kavanagh D, Crotty TB, Quinn CM, McDermott EW, O’Higgins N, et al. Factors affecting metastases to non-sentinel lymph nodes in breast cancer. J. Clin. Pathol. janv 2004;57(1):73e6. Fournier K, Schiller A, Perry RR, Laronga C. Micrometastatic in the sentinel lymph node of breast cancer does not mandate completion axillary dissection. Ann Surg 2004;239:859e63. Menes TS, Tartter PI, Mizrachi H. Breast cancer patients with pN0(iþ) and pN1(mi) sentinel nodes have high rate of nonsentinel nodes metastases. J Am Coll Surg 2005;200:323e7. Reynolds C, Mick R, Donohue JH, Grant CS, Farley DR, Callans LS, et al. Sentinel lymph node biopsy with metastasis: can axillary dissection be avoided in some patients with breast cancer? J. Clin. Oncol. juin 1999;17(6):1720e6. Giuliano AE, Hunt KK, Ballman KV, Beitsch PD, Whitworth PW, Blumencranz PW, et al. Axillary dissection vs no axillary dissection in women with invasive breast cancer and sentinel node metastasis: a randomized clinical trial. JAMA 9 févr 2011;305(6):569e75. Cserni G, Gregori D, Merletti F, Sapino A, Mano MP, Ponti A, et al. Meta-analysis of non-sentinel node metastases associated with micrometastatic sentinel nodes in breast cancer. Br J Surg oct 2004;91(10):1245e52. Viale G, Maiorano E, Pruneri G, Mastropasqua MG, Valentini S, Galimberti V, et al. Predicting the risk for additional axillary metastases in patients with breast carcinoma and positive sentinel lymph node biopsy. Ann. Surg févr 2005;241(2):319e25. Meretoja TJ, Strien L, Heikkila PS, Leidenius MHK. A simple nomogram to evaluate the risk of nonsentinel node metastases in breast cancer patients with minimal sentinel node involvement. Ann Surg Oncol 2012;19(2):567e76. Krag D, Weaver D, Ashikaga T, Moffat F, Klimberg VS, Shriver C, et al. The sentinel node in breast cancerea multicenter validation study. N. Engl. J. Med. 1 oct 1998;339(14):941e6. Veronesi U, Paganelli G, Galimberti V, Viale G, Zurrida S, Bedoni M, et al. Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 28 juin 1997;349(9069):1864e7. Weaver DL, Krag DN, Ashikaga T, Harlow SP, O'Connell M. Pathologic analysis of sentinel and nonsentinel lymph nodes in breast carcinoma: a multicenter study. Cancer 1 mars 2000;88(5):1099e107. Bilimoria KY, Bentrem DJ, Hansen NM, Bethke KP, Rademaker AW, Ko CY, et al. Comparison of sentinel lymph node biopsy alone and completion axillary lymph node dissection for node-positive breast cancer. J. Clin. Oncol. 20 juin 2009;27(18):2946e53. Fant JS, Grant MD, Knox SM, Livingston SA, Ridl K, Jones RC, et al. Preliminary outcome analysis in patients with breast cancer and a positive sentinel lymph node who declined axillary dissection. Ann. Surg. Oncol mars 2003;10(2): 126e30. Jeruss JS, Winchester DJ, Sener SF, Brinkmann EM, Bilimoria MM, Barrera Jr E, et al. Axillary recurrence after sentinel node biopsy. Ann. Surg. Oncol janv 2005;12(1):34e40. Naik AM, Fey J, Gemignani M, Heerdt A, Montgomery L, Petrek J, et al. The risk of axillary relapse after sentinel lymph node biopsy for breast cancer is comparable with that of axillary lymph node dissection: a follow-up study of 4008 procedures. Ann. Surg. sept 2004;240(3):462e8. discussion 468‑471. Spiguel L, Yao K, Winchester DJ, Gorchow A, Du H, Sener SF, et al. Sentinel node biopsy alone for node-positive breast cancer: 12-year experience at a single institution. J. Am. Coll. Surg juill 2011;213(1):122e8. discussion 128‑129. Chagpar A, Middleton LP, Sahin AA, Meric-Bernstam F, Kuerer HM, Feig BW, et al. Clinical outcome of patients with lymph node-negative breast carcinoma who have sentinel lymph node micrometastases detected by immunohistochemistry. Cancer 15 avr 2005;103(8):1581e6. Fan Y-G, Tan Y-Y, Wu C-T, Treseler P, Lu Y, Chan C-W, et al. The effect of sentinel node tumor burden on non-sentinel node status and recurrence rates in breast cancer. Ann. Surg. Oncol. sept 2005;12(9):705e11. Hwang RF, Gonzalez-Angulo AM, Yi M, Buchholz TA, Meric-Bernstam F, Kuerer HM, et al. Low locoregional failure rates in selected breast cancer
792
24.
25.
26.
27.
28.
29.
L. Ouldamer et al. / The Breast 22 (2013) 787e792 patients with tumor-positive sentinel lymph nodes who do not undergo completion axillary dissection. Cancer 15 août 2007;110(4):723e30. Park J, Fey JV, Naik AM, Borgen PI, Van Zee KJ, Cody 3rd HS. A declining rate of completion axillary dissection in sentinel lymph node-positive breast cancer patients is associated with the use of a multivariate nomogram. Ann. Surg mars 2007;245(3):462e8. Pejavar S, Wilson LD, Haffty BD. Regional nodal recurrence in breast cancer patients treated with conservative surgery and radiation therapy( BCSþRT). Int J Radiat Oncol Phy 2006;66:1320e7. Schulze T, Mucke J, Markwardt J, Schlag PM, Bembenek A. Long-term morbidity of patients with early breast cancer after sentinel lymph node biopsy compared to axillary lymph node dissection. J Surg Oncol. 1 févr 2006;93(2):109e19. Swenson KK, Mahipal A, Nissen MJ, Tuttle TM, Heaton K, Lally RM, et al. Axillary disease recurrence after sentinel lymph node dissection for breast carcinoma. Cancer 1 nov 2005;104(9):1834e9. Takei H, Suemasu K, Kurosumi M, Horii Y, Yoshida T, Ninomiya J, et al. Recurrence after sentinel lymph node biopsy with or without axillary lymph node dissection in patients with breast cancer. Breast Cancer 2007;14(1):16e24. Guenther JM, Hansen NM, DiFronzo LA, Giuliano AE, Collins JC, Grube BL, et al. Axillary dissection is not required for all patients with breast cancer and positive sentinel nodes. Arch Surg 2003;138:52e6.
30. Yegiyants S, Romero LM, Haigh PI, DiFronzo A. Completion axillary lymph node dissection not required for regional control in patients with breast cancer who have micrometastasis in a sentinel node. Arch Surg 2010;145:564e9. 31. Houvenaeghel G, Nos C, Giard S, Mignotte H, Esterni B, Jacquemier J, et al. A nomogram predictive of non-sentinel lymph node involvment in breast cancer patients with a sentinel lymph node micrometastasis. Eur J Surg Oncol 2009;35:690e5. 32. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: wiley; 2000. 33. McCarter MD, Yeung H, Fey J, Borgen PI, Cody 3rd HS. The breast cancer patient with multiple sentinel nodes: when to stop? J Am Coll Surg 2001;192:692e7. 34. Goyal A, Newcombe RG, Chhabra A, Mansel RE. Factors affecting failed localisation and false-negative rates of sentinel node biopsy in breast cancer- results of the ALMANAC validation phase. Breast Cancer Res Treat 2006;99:203e8. 35. Martin 2nd RCG, Chagpar A, Scoggins CR, Edwards MJ, Hagendoorn L, Stromberg AJ, et al. Clinicopathologic factors associated with false-negative sentinel lymph node biopsy in breast cancer. Ann Surg 2005;241:1005e12. discussion 1012e1015. 36. Wong SL, Edwards MJ, Chao C, Tuttle TM, Noyes RD, Carlson DJ, et al. Sentinel lymph node biopsy for breast cancer: impact of the number of sentinel nodes removed on the false-negative rate. J Am Coll Surg 2001;192:684e9. discussion 689e691.
Contents lists available at SciVerse ScienceDirect
The Breast journal homepage: www.elsevier.com/brst
Original article
Validation of a breast cancer nomogram for predicting nonsentinel node metastases after minimal sentinel node involvement: Validation of the Helsinki breast nomogram Lobna Ouldamer a, b, c, *, Flavie Arbion d, Alexis Balagny a, c, Florence Fourquet e, Henri Marret a, c, Gilles Body a, c a
Department of Gynecology, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France INSERM Unit 1069, Tours, France François Rabelais University, Tours, France d Department of Pathology, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France e Department of Biostatistics, CHRU de Tours, Hôpital Bretonneau, 2 Boulevard Tonnelé, 37044 Tours, France b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 September 2012 Received in revised form 28 December 2012 Accepted 6 February 2013
Background: Complete node dissection for tumor-positive sentinel lymph nodes (SLN) is becoming more controversial. Nevertheless, current practice guidelines still recommend complete axillary lymph node dissection (ALND) for breast cancer patients whose SLN contains a metastatic tumor. The Helsinki breast cancer nomogram developed by Meretoja TJ et al. aims to predict the risk of positive non-sentinel lymph nodes in patients with minimal sentinel node involvement, it uses tumor diameter and multifocality. The purpose of this study was to test the accuracy of the nomogram among patients with micrometastatic SLN-positive biopsy findings. Methods: The Helsinki nomogram was used to calculate risk of metastases for 49 consecutive patients with SLN micrometastases or isolated tumor cells (ITC) who underwent complete ALND. The nomogram was evaluated by calculating the area under the receiver-operator characteristic (ROC) curve. Results: The area under the ROC curve for the nomogram applied to all patients with micrometastases and ITC was 0.72 (range 0.60e0.85) (0.791 in the original publication). Conclusions: The Helsinki breast cancer nomogram is a useful tool for patients with minimal sentinel node involvement. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Breast cancer Sentinel lymph node Nomogram Isolated tumor cells Micrometastasis Nonsentinel node metastases
Introduction Completion axillary lymph node dissection (ALND) has been considered to be the gold standard for cases in which a metastasis is found in the sentinel lymph node (SLN).1 ALND remains the standard approach for SLN-positive patients recommended by the American Society of Clinical Oncology and the French guidelines. However, approximately 50%e70% of patients with diseasepositive SLN have no additional positive nodes, which means that it may be possible to avoid axillary dissection in selected patients.2e8 There has been discussion in the literature about whether systematic ALND for patients with positive SLNs is appropriate. In the
* Corresponding author. Department of Gynecology, CHU Bretonneau, 2 Boulevard Tonnelé, 37000 Tours, France. Tel.: þ33 02 47 47 47 47; fax: þ33 02 47 47 38 01. E-mail address: [email protected] (L. Ouldamer). 0960-9776/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.breast.2013.02.004
randomized American College of Surgeons Oncology Group (ACOSOG) Z0011 trial, patients with a positive SLN were randomized into two groups: complete ALND or axillary observation. There was similar outcome in the ALND and SN biopsy-only arms, despite a nonsentinel node involvement rate of 27% in the ALND arm suggesting that complete ALND after positive SN may not improve survival. However, the study closed prematurely due to poor recruitment, so it is unlikely that the number of subjects successfully enrolled in the study will provide adequate power to show a difference in survival, even if any such difference exists.9 Also, all patients in this trial underwent breast-conserving treatment with whole-breast radiation. So, it is unclear whether these results can be generalized to patients undergoing mastectomy without radiotherapy. Metastases are found in non-SLNs in approximately 10% of patients with isolated tumor cells (ITC) in the SLN and in 20%e35% of patients with micrometastases in the SLN10; complete ALND is routinely performed in these patients.11
788
L. Ouldamer et al. / The Breast 22 (2013) 787e792
Meretoja et al. developed a nomogram to predict the likelihood of finding additional positive nodes on ALND in patients with micrometastasis or ITC in the SLN. The nomogram predicts individual patient risk and was developed from a logistic-regression analysis and included only two statistically significant variables: tumor diameter and multifocality. The receiver-operating characteristic (ROC) curve for this tool was 0.791 indicating good prediction and discrimination.12 We therefore decided to apply this nomogram to our SLNpositive series and evaluated its usefulness for the subgroup of patients with micrometastatic SLNs. Patients and methods A total of 1283 consecutive cases of SLN biopsy from primary untreated breast cancer patients at the tertiary oncological referral center of Tours (France) were prospectively entered into a multidisciplinary team meeting register between January 2004 and December 2011. With the authorization of the institutional review board we retrospectively reviewed this prospective database. We studied a subset of 187 cases that fulfilled the following inclusion criteria: primary invasive breast carcinoma with clinically negative axillary lymph nodes and no prior systemic therapy; successful SLN biopsy and metastatic disease identified by any method (routine histopathology, serial sectioning, staining with hematoxylin and eosin [HE], immunohistochemistry [IHC]); and complete ALND. Exclusion criteria were: neoadjuvant chemotherapy, inflammatory breast cancer, ductal carcinoma in situ, failed SLN mapping, clinically suspicious axillary lymph nodes and bilateral procedures. Both blue dye (patent blue, Laboratoire Guerbet, Paris, France) and radioisotopes (99mTc-labeled colloid sulfur, Nanocoll, General Electric Healthcare SA, USA) were used for lymphatic mapping with SLN biopsies. Static lymphoscintigraphy was performed to visualize the probable localization of SLN, and a hand-held gamma detection probe (Gamma-sup/Clerad/France) was used intraoperatively to identify the most radioactive area. SLN were identified as nodes with blue dye uptake, radiotracer uptake, or both. Histopathological evaluation Fresh tissue labeled as SLN was delivered without fixation to the pathology lab. Each lymph node was bisected along its major axis. One half of the nodal tissue was frozen, sliced into 5 mm-thick sections and stained with blue methylene and with hematoxylin and eosin (HE) for intraoperative consultation. In cases with technical limitations, such as incomplete representation of the tissue or folding, additional sections were obtained at the discretion of the consultant pathologist. Upon completion of the intraoperative consultation, tissue samples were fixed in 10% buffered formalin and then embedded in paraffin. Three slides were prepared: two of sections taken at 200micron intervals were stained with HE, and a conservative section of a deeper level was studied for pancytokeratin by immunohistochemistry. Immunohistochemistry for cytokeratin AE1/AE3 was used to confirm suspicious cells in the SLN. Following the American Joint Committee on cancer recommendations, macrometastases were defined as tumors >2 mm long and micrometastases as tumors between 0.2 and 2 mm long. Tumors <0.2 mm long were regarded as isolated tumor cells. If a frozen section was positive, ALND was performed during the same surgical intervention. For patients with SLN metastases but
that were not identified in interoperative frozen sections, ALND was performed at a later date. The Helsinki breast cancer nomogram was used to predict the individual risk of non-SLN positivity from patient data. The nomogram was developed a to predict the likelihood of finding additional positive nodes on ALND in patients with micrometastasis or ITC in the SLN. First, univariate analysis was performed to determine factors associated with non SLN metastases, all variables (tumor diameter, multifocality, lymphovascular invasion, and tumor location) with p value less than 0.15 were then included into a logistic regression analysis using backward stepwise method. Variables with p value <0.05 were considered statistically significant in the multivariate analysis. Tumor diameter (p ¼ 0.002) and multifocality (p ¼ 0.039) were found to be the only statistically significant variables and were included in the final predictive model. With a 10% cutoff value for predicted probability the model had sensitivity of 38.2% and specificity of 85.7% and selects 84.1% of patients to have predicted risk of less than 10% for additional metastases showing good clinical utility.12 For nomograms, 10% or less cutoff values are considered in the literature to define the subgroup of patients with a low predicted probability of metastatic non-SLN (having a probability of metastatic non-SLN 10% predicted by nomogram). Statistical analyzes were performed by R 2.13.1 (http://www. cran.r-project.org/). For numeric data, results are reported as mean and median values standard deviation (SD). Numeric data were analyzed with Student’s t-test if normally distributed, and the ManneWhitney test if not. Categorical data were analyzed with the chi-squared test or with Fisher’s exact test. To measure discrimination by the nomogram, a receiveroperating characteristic (ROC) curve was constructed. The ROC curve assesses the relationship between the sensitivity and the false positive rate (1- specificity) of a test across all possible threshold values that define the positivity of a disease or condition. The overall accuracy of the nomogram is expressed as the area under the ROC curve (AUC). The AUC can be interpreted as the probability of a correct assignment of disease presence in random pairs of patients, one patient in each pair who has disease and one who does not. It is generally accepted that AUC values of 0.7e0.8 indicate satisfactory discrimination, and that AUC values >0.8 represent good discrimination. A calibration plot was drawn showing the actual (women who actually had positive non-SNs) versus nomogram-predicted probability. Results Between January 2004 and December 2011, metastatic disease was found in 216 patients; 29 (13 isolated tumor cells, 10 micrometastasis and 6 macrometastasis) of these patients did not undergo ALND (presence of co-morbidities or a history of controlateral ALND or patient choice) and were excluded from this analysis. Thus 187 cases met our inclusion criteria, and they included a subgroup of 49 patients with SLN micrometastases or isolated tumor cells (ITC). Descriptive clinical and histopathological data for the study population are listed in Table 1. The mean age was 57 years (range, 30e84 years). The mean pathological tumor size was 15.33 mm (range 4e 34 mm). The mean number of SLN identified was 2.37 (range 1e6), and the mean number of nodes extracted at ALND was 11 (range 1e31). The characteristics of positive-ALND patients are showed in Table 2. The nomogram was applied to the micrometastatic SLN and ITC cohort (n ¼ 49). The predicted probability of non-sentinel metastases for each patient are showed in Table 3.
L. Ouldamer et al. / The Breast 22 (2013) 787e792
calculated for each patient (Fig. 2). We found a good correspondence between the predicted probabilities and observations.
Table 1 Patient and tumors characteristics. Characteristics Patient age (years) Mean [range] Menopausal status Post menopausal Pre menopausal Histological diameter, mm Mean [range] Histology of the primary tumor Invasive ductal carcinoma Invasive lobular carcinoma Any other histology Histological grade of the tumor I II III Multifocality Yes No Lymphovascular invasion Yes No Estrogen receptor status Positive Negative Progesterone receptor status Positive Negative HER-2 status Positive Negative Number of SLN harvested on SN biopsy Mean [range] SLN ratio (positive/all SLN) <0.5 0.5e1 1 Size of the SLN metastasis Macrometastasis Micrometastasis Isolated tumor cells Detection method of the SLN metastasis Frozen section analysis Hematoxylin and eosin staining Immunohistochemistry Number of non-SLN harvested in ALND Mean (range) ALND status Positive Negative Number of positive non-SLN Mean (range)
789
Micrometastasis or ITC n ¼ 49
All patients (n ¼ 187)
57.8 [34e84]
57.12[30e84]
36 (73.5%) 13 (26.5%)
129 (69%) 58 (31%)
13.5 [4e30]
15.33[4e34]
44 (89.8%) 4 (8.2%) 1 (2%)
168 (89.9%) 13 (6.9%) 6 (3.14%)
12 (24.5%) 27 (55.1%) 10 (20.4%)
43 (23%) 105 (56.15%) 39 (21%)
5 (10.2%) 44 (89.8%)
26 (13.9%) 161 (86.1%)
13 (26.5%) 36 (73.5%)
69 (36.9%) 118 (63.1%)
44 (89.8%) 5 (10.2%)
173 (92.51%) 14 (7.49%)
40 (81.6%) 9 (18.4%)
157 (83.95%) 30 (16.04%)
4 (8.2%) 45 (91.8%)
29 (15.5%) 158 (84.5%)
2.38 [1e6]
2.37 [1e6]
16 (32.6%) 17 (34.8%) 16 (32.6%) 0 45 (91.8%) 4 (8.2%)
Application of other nomograms to the micrometastatic and ITC SN subgroup We also applied the Marseille nomogram to our minimal SLN involvement population: the AUC was 0.57 ([0.431e0.709]; 95% confidence interval). Other widely used predictive models (mskcc, tenon score) were also applied, to calculate either predicted probability for additional metastases or a score given by the model. For each of these models, we determined the AUC, the sensitivity, specificity, and clinical utility for either a cutoff value of 10% predicted probability or a cutoff threshold value given by the scoring system (Table 4). Discussion There is currently a debate about the necessity for complete ALND in every patient with metastatic SLN. Indeed, complete node dissection for a tumor-positive SLN is becoming more controversial, in view of the findings of numerous single-center studies, a large nonrandomized study and the Z0011 study by the American College of Surgeons Oncology Group (ACOSOG). The ACOSOG Z0011 study is a prospective study of patients with T1 or T2, clinically nodenegative breast cancers treated with lumpectomy, SLN biopsy, and whole-breast irradiation. Patients with one or two positive SLNs were randomized to complete ALND or no further axillary treatment. After a median follow-up of 6.3 years, the 5-year overall survival was similar in the two groups (91.8% for the ALND and 92.5% for the SLN biopsy groups). Although 27% of patients in the Z0011 trial had additional nodal metastases, complete ALND did not result in fewer locoregional recurrences or improved survival, suggesting that not all non-SLN metastases develop into clinically significant disease.9,13e20 There are several reasons for the low axillary recurrence rate, including the study excluding patients with extensive axillary disease or with 3 tumor-positive SLNs found on SN biopsy. Even for patients in whom there are residual regional lymph node metastases, the additional benefit of ALND is unclear because these patients will generally receive systemic therapy based on the presence of SLN metastases. For example, in the ACOSOG Z0011 trial, more than 95% of the patients in both arms received adjuvant therapy. Such adjuvant therapy may be one cause of the low rate of axillary recurrence. Also, all patients in this trial received whole breast radiation therapy. It is currently unclear how partial breast radiation therapy affects axillary recurrence rates in those who do not undergo complete ALND. Indeed, it is difficult to distinguish between the effects of surgery, chemotherapy, radiotherapy and antihormonal treatment on local and distant relapses. Reported axillary recurrence rates range from 0 to 2.6% with median follow-up of 2e2.5 years in patients with tumor-positive SLN who do not undergo complete ALND.9,18e28
52 (27.81%) 60 (32.1%) 75 (40.1%) 138 (73.8%) 45 (24.1%) 4 (2.1%) 99 (52.94%) 73 (39%) 15 (8.02%)
40 (81.6%) 9 (18.4%) 12.7 [5e24]
11 [1e31]
6 (12.2%) 43 (87.8%)
51 (27.27%) 136 (72.72%)
1.17 [1e2]
1 [1e18]
Application of the Helsinki nomogram to the micrometastatic and ITC SN subgroup (n ¼ 49) A ROC was calculated to assess discrimination by the Helsinki nomogram (Fig. 1). The area under the ROC was 0.72 (95% confidence interval [95% CI], 0.60e0.85) compared with 0.791 in the original study. To assess the accuracy of the Helsinki nomogram, actual probabilities were plotted against the predicted probability Table 2 Characteristics of positive-ALND patients. Patient
Age (years)
Tumor diameter (mm)
Multifocality
Number of SN
Number of positive non-SN
Size of the non-SN metastasis
1 2 3 4 5 6
48 51 41 43 50 84
19 6 15 15 30 10
Yes Yes No No No No
1 3 1 1 1 1
1 2 1 1 1 1
Micrometastasis Macro and micrometastasis Macrometastasis Macrometastasis Macrometastasis Macrometastasis
790
L. Ouldamer et al. / The Breast 22 (2013) 787e792
Table 3 The predicted probability of non-sentinel metastases for each patient. Unifocal tumor Tumor size (mm) Risk of additional metastases (%) Multifocal tumor Tumor size (mm) Risk of additional metastases (%)
4 3.6
6 3.8
6 8.3
7(n ¼ 2) 3.9
8(n ¼ 3) 4
9 4.2
10(n ¼ 7) 4.3
12(n ¼ 4) 4.5
10(n ¼ 2) 9.3
Some authors suggest that patients with micrometastases or isolated tumor cells in their SLN should not be treated by complete ALND.28e30 Until new randomized controlled trials are conducted, the results of retrospective studies, patient preference, and the statistical likelihood of non-SLN metastases, will have to be used as the basis for deciding which patients can avoid completion ALND. French guidelines and the American Society of Clinical Oncology recommend complete ALND for breast cancer patients whose SLN contain metastatic tumor. Freely available tools that can predict an individual’s clinical outcome, and have been validated internationally, are required. Nomograms for many type of cancer have become popular in recent years for treatment choices and assessments of risk, because of their simplicity and easy graphical representation.
13(n ¼ 3) 4.6
14(n ¼ 2) 4.8
15(n ¼ 9) 5
17 5.2
15 10.7
18 5.4
19(n ¼ 2) 5.5
20(n ¼ 5) 5.7
23 6.1
30 7.6
19 12
Only two of the existing predictive models for breast cancer are based on patient populations with only micrometastasis or ITC in SLN.12,31 The Marseille nomogram did not display sufficient accuracy in our study population to be applicable in clinical practice. However, the sample size in our study was small, and this may have reduced the predictive accuracy. The Helsinki model is simple, uses routinely available clinicopathologic information, and yields an AUC of 0.791 among the Helsinki patients used to build the model. Another potential strength of this model is that it does not require histological detection as a nomogram variable, thereby avoiding possible confusion arising from the classification or results of the detection method associated with differing tissue processing of SLN at different institutions.
Fig. 1. Receiver operating characteristic curve calculation for micrometastatic and ITC sentinel lymph nodes patients.
Fig. 2. Calibration plot for micrometastatic and ITC sentinel lymph nodes patients.
L. Ouldamer et al. / The Breast 22 (2013) 787e792
791
Table 4 Results of the application of 5 nomograms. Study
Year
Cut threshold
Sensitivity (%)
Specificity (%)
NPV (%)
PPV
AUC (95% IC)
Clinical utility
Marseille MSKCC Tenon Helsinki Helsinki
2008 2003 2005 2011 2011
10% 10% 3.5 10% 7%
50 50 66.67 17 50
39 57.14 88.37 98 95.35
20 11.11 5 10.64 6.82
13.79 14.29 44.44 50 60
0.57 0.61 0.657 0.725 e
40.81 56.25 85.71 87.75 89.79
(0.431e0.709) (0.473e0.747) (0.60e0.85) (0.60e0.85)
NPV: negative predictive value. PPV: positive predictive value.
Here, we report the first external validation of the Helsinki nomogram. Our overall findings indicate a satisfactory predictive accuracy of the Helsinki nomogram in our cohort with a statistically significant AUC value of 0.72 ( the AUC value is a measure of the overall discrimination of a model, and values of 0.7e0.8 are considered good, whereas values of 0.81e0.90 are considered excellent32). Based on our results, the false negative rate with The Helsinki nomogram was 10%, and it was able to correctly predict 43 (87.75%) of the patients. Using the threshold of 10%, two patients would have been operated with an ALND of which one would have a non-SLN metastasis and 47 patients would have been spared the ALND, 5 (10.6%) of those would have a non SLN metastasis. If the threshold is lowered to 7%, six patients would have been operated with an ALND of which three would have a non-SLN metastasis and 43 patients would have been spared the ALND, 3 (7%) of those would have a non-SLN metastasis. This mean that nine of every ten SLN-positive patients can be spared ALND with the use of this model; this is, in our opinion, of considerable clinical value, providing a safe and reliable tool that can be used to avoid potentially unnecessary surgical procedures. It has been reported that in 98% of node-positive patients who have multiple SN removed, metastasis is detected in the first three SLN.33 In our study, a mean of 2.37 SLN were retrieved per patient, and it could be argued that resecting additional nodes would result in less disease in the residual non-SLN. In most published large series, removal of more than one SN is strongly associated with a decreased false-negative rate.34e36 If fewer SLN are retrieved, therefore, there may be more residual disease in non-SLN and the nomogram will underestimate the risk of residual disease in these populations. In our 6 positive-ALND patients, five had only one SN. This issue requires further investigation. Our results show that the nomogram is accurate in our population of breast cancer patients. Although our study included only a small number of patients, it validates the predictive ability of the nomogram in a community breast cancer center. Predicting the risk of additional nodal metastases allows the surgeon and patient to make an individualized decision regarding the need for ALND. We report that the Helsinki nomogram provides patients and their physicians with a meaningful tool for assessing an individual’s risk of non-SLN metastases.
Conflict of interest statement
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
We declare that we have no conflict of interest. 21.
References 22. 1. Kim I, Giuliano AE, Lyman GH. Lymphatic mapping and sentinel lymph node biopsy in early-stage breast carcinoma: a metaanalysis. Cancer 2006;106:4e16. 2. Bolster MJ, Peer PGM, Bult P, Thunnissen FBJM, Schapers RFM, Meijer JWR, et al. Risk factors for non-sentinel lymph node metastases in patients with breast
23.
cancer. The outcome of a multi-institutional study. Ann. Surg. Oncol. janv 2007;14(1):181e9. Changsri C, Prakash S, Sandweiss L, Bose S. Prediction of additional axillary metastasis of breast cancer following sentinel lymph node surgey. Breast J 2004;10:392e7. Farshid G, Pradhan M, Kollias J, Gill PG. A decision aid for predicting nonsentinel node involvement in women with breast cancer and at least one positive sentinel node. Breast 2004;13:494e501. Fleming FJ, Kavanagh D, Crotty TB, Quinn CM, McDermott EW, O’Higgins N, et al. Factors affecting metastases to non-sentinel lymph nodes in breast cancer. J. Clin. Pathol. janv 2004;57(1):73e6. Fournier K, Schiller A, Perry RR, Laronga C. Micrometastatic in the sentinel lymph node of breast cancer does not mandate completion axillary dissection. Ann Surg 2004;239:859e63. Menes TS, Tartter PI, Mizrachi H. Breast cancer patients with pN0(iþ) and pN1(mi) sentinel nodes have high rate of nonsentinel nodes metastases. J Am Coll Surg 2005;200:323e7. Reynolds C, Mick R, Donohue JH, Grant CS, Farley DR, Callans LS, et al. Sentinel lymph node biopsy with metastasis: can axillary dissection be avoided in some patients with breast cancer? J. Clin. Oncol. juin 1999;17(6):1720e6. Giuliano AE, Hunt KK, Ballman KV, Beitsch PD, Whitworth PW, Blumencranz PW, et al. Axillary dissection vs no axillary dissection in women with invasive breast cancer and sentinel node metastasis: a randomized clinical trial. JAMA 9 févr 2011;305(6):569e75. Cserni G, Gregori D, Merletti F, Sapino A, Mano MP, Ponti A, et al. Meta-analysis of non-sentinel node metastases associated with micrometastatic sentinel nodes in breast cancer. Br J Surg oct 2004;91(10):1245e52. Viale G, Maiorano E, Pruneri G, Mastropasqua MG, Valentini S, Galimberti V, et al. Predicting the risk for additional axillary metastases in patients with breast carcinoma and positive sentinel lymph node biopsy. Ann. Surg févr 2005;241(2):319e25. Meretoja TJ, Strien L, Heikkila PS, Leidenius MHK. A simple nomogram to evaluate the risk of nonsentinel node metastases in breast cancer patients with minimal sentinel node involvement. Ann Surg Oncol 2012;19(2):567e76. Krag D, Weaver D, Ashikaga T, Moffat F, Klimberg VS, Shriver C, et al. The sentinel node in breast cancerea multicenter validation study. N. Engl. J. Med. 1 oct 1998;339(14):941e6. Veronesi U, Paganelli G, Galimberti V, Viale G, Zurrida S, Bedoni M, et al. Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 28 juin 1997;349(9069):1864e7. Weaver DL, Krag DN, Ashikaga T, Harlow SP, O'Connell M. Pathologic analysis of sentinel and nonsentinel lymph nodes in breast carcinoma: a multicenter study. Cancer 1 mars 2000;88(5):1099e107. Bilimoria KY, Bentrem DJ, Hansen NM, Bethke KP, Rademaker AW, Ko CY, et al. Comparison of sentinel lymph node biopsy alone and completion axillary lymph node dissection for node-positive breast cancer. J. Clin. Oncol. 20 juin 2009;27(18):2946e53. Fant JS, Grant MD, Knox SM, Livingston SA, Ridl K, Jones RC, et al. Preliminary outcome analysis in patients with breast cancer and a positive sentinel lymph node who declined axillary dissection. Ann. Surg. Oncol mars 2003;10(2): 126e30. Jeruss JS, Winchester DJ, Sener SF, Brinkmann EM, Bilimoria MM, Barrera Jr E, et al. Axillary recurrence after sentinel node biopsy. Ann. Surg. Oncol janv 2005;12(1):34e40. Naik AM, Fey J, Gemignani M, Heerdt A, Montgomery L, Petrek J, et al. The risk of axillary relapse after sentinel lymph node biopsy for breast cancer is comparable with that of axillary lymph node dissection: a follow-up study of 4008 procedures. Ann. Surg. sept 2004;240(3):462e8. discussion 468‑471. Spiguel L, Yao K, Winchester DJ, Gorchow A, Du H, Sener SF, et al. Sentinel node biopsy alone for node-positive breast cancer: 12-year experience at a single institution. J. Am. Coll. Surg juill 2011;213(1):122e8. discussion 128‑129. Chagpar A, Middleton LP, Sahin AA, Meric-Bernstam F, Kuerer HM, Feig BW, et al. Clinical outcome of patients with lymph node-negative breast carcinoma who have sentinel lymph node micrometastases detected by immunohistochemistry. Cancer 15 avr 2005;103(8):1581e6. Fan Y-G, Tan Y-Y, Wu C-T, Treseler P, Lu Y, Chan C-W, et al. The effect of sentinel node tumor burden on non-sentinel node status and recurrence rates in breast cancer. Ann. Surg. Oncol. sept 2005;12(9):705e11. Hwang RF, Gonzalez-Angulo AM, Yi M, Buchholz TA, Meric-Bernstam F, Kuerer HM, et al. Low locoregional failure rates in selected breast cancer
792
24.
25.
26.
27.
28.
29.
L. Ouldamer et al. / The Breast 22 (2013) 787e792 patients with tumor-positive sentinel lymph nodes who do not undergo completion axillary dissection. Cancer 15 août 2007;110(4):723e30. Park J, Fey JV, Naik AM, Borgen PI, Van Zee KJ, Cody 3rd HS. A declining rate of completion axillary dissection in sentinel lymph node-positive breast cancer patients is associated with the use of a multivariate nomogram. Ann. Surg mars 2007;245(3):462e8. Pejavar S, Wilson LD, Haffty BD. Regional nodal recurrence in breast cancer patients treated with conservative surgery and radiation therapy( BCSþRT). Int J Radiat Oncol Phy 2006;66:1320e7. Schulze T, Mucke J, Markwardt J, Schlag PM, Bembenek A. Long-term morbidity of patients with early breast cancer after sentinel lymph node biopsy compared to axillary lymph node dissection. J Surg Oncol. 1 févr 2006;93(2):109e19. Swenson KK, Mahipal A, Nissen MJ, Tuttle TM, Heaton K, Lally RM, et al. Axillary disease recurrence after sentinel lymph node dissection for breast carcinoma. Cancer 1 nov 2005;104(9):1834e9. Takei H, Suemasu K, Kurosumi M, Horii Y, Yoshida T, Ninomiya J, et al. Recurrence after sentinel lymph node biopsy with or without axillary lymph node dissection in patients with breast cancer. Breast Cancer 2007;14(1):16e24. Guenther JM, Hansen NM, DiFronzo LA, Giuliano AE, Collins JC, Grube BL, et al. Axillary dissection is not required for all patients with breast cancer and positive sentinel nodes. Arch Surg 2003;138:52e6.
30. Yegiyants S, Romero LM, Haigh PI, DiFronzo A. Completion axillary lymph node dissection not required for regional control in patients with breast cancer who have micrometastasis in a sentinel node. Arch Surg 2010;145:564e9. 31. Houvenaeghel G, Nos C, Giard S, Mignotte H, Esterni B, Jacquemier J, et al. A nomogram predictive of non-sentinel lymph node involvment in breast cancer patients with a sentinel lymph node micrometastasis. Eur J Surg Oncol 2009;35:690e5. 32. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: wiley; 2000. 33. McCarter MD, Yeung H, Fey J, Borgen PI, Cody 3rd HS. The breast cancer patient with multiple sentinel nodes: when to stop? J Am Coll Surg 2001;192:692e7. 34. Goyal A, Newcombe RG, Chhabra A, Mansel RE. Factors affecting failed localisation and false-negative rates of sentinel node biopsy in breast cancer- results of the ALMANAC validation phase. Breast Cancer Res Treat 2006;99:203e8. 35. Martin 2nd RCG, Chagpar A, Scoggins CR, Edwards MJ, Hagendoorn L, Stromberg AJ, et al. Clinicopathologic factors associated with false-negative sentinel lymph node biopsy in breast cancer. Ann Surg 2005;241:1005e12. discussion 1012e1015. 36. Wong SL, Edwards MJ, Chao C, Tuttle TM, Noyes RD, Carlson DJ, et al. Sentinel lymph node biopsy for breast cancer: impact of the number of sentinel nodes removed on the false-negative rate. J Am Coll Surg 2001;192:684e9. discussion 689e691.