- Email: [email protected]
Reexcision for positive margins in breast cancer: A predictive score of residual disease
Surgical Oncology xxx (2015) 1e7
Contents lists available at ScienceDirect
Surgical Oncology journal homepage: www.elsevier.com/locate/suronc
Review
Reexcision for positive margins in breast cancer: A predictive score of residual disease mie Lotersztajn d, Myriam Mimouni a, b, c, d, *, Fabrice Lecuru c, Roman Rouzier d, e, Noe Denis Heitz a, b, Julien Cohen f, Arnaud Fauconnier a, b, e, Cyrille Huchon a, b, e a
Department of Gynecology and Obstetrics, CHI Poissy-Saint-Germain, Poissy, France University Versailles Saint Quentin en Yvelines, Versailles, 10 rue du Champ Gaillard, 78300 Poissy, France ^pital Europ Department of Gynecologic and Oncologic Surgery, Ho een Georges Pompidou, AP-HP Paris, University Paris V, Ren e Descartes, 20 rue Leblanc, 75015 Paris, France d Department of Surgery and Senology, Institut Curie, Universit e Paris V, Ren e Descartes, 26 rue d’Ulm, 75005 Paris, France e EA 7285 Clinical Risks and Safety on Women's Health, University Versailles-Saint-Quentin en Yvelines, France f Medistat, Biostatistics, 10-12 rue de la Conception, 13004 Marseille, 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 30 March 2015 Received in revised form 6 August 2015 Accepted 9 August 2015
Background: Guidelines recommend re-excision if resection margins are positive in lumpectomy for breast cancer. However, residual disease (RD) is not always found. The aim of our study was to develop a score to predict RD in re-excision specimens. Materials and methods: We carried out a multicenter, retrospective study with two population groups. The ‘modeling’ group was composed of 148 patients treated in the Centre Hospitalier Poissy-St-Germain or the Georges Pompidou European Hospital and the ‘validation’ group was composed of 67 patients treated in Curie Institute. The score was built with a logistic regression model. Results: Factors independently associated with RD were: a cumulative length of all positive margins>5 mm, invasion by ductal carcinoma in situ only, a pathological tumor size>30 mm and a pathological tumor size<30 mm with a discrepancy of >50% between pathological and radiological tumor size. The 7-point score allowed the classification of patients into three risk groups for RD: low (16% of patients experienced RD), moderate (65%) and high (100%). The areas under the ROC curve of the score and the logistic model were 0.72(95%CI:0.68e0.75,p ¼ 0.60). The proportion of RD in each group of the validation population (25%, 48%, and 100% in the low, moderate and high group, respectively) confirmed the accuracy of the score in an independent population. Conclusions: This score enables the identification of patients at high risk of RD but it cannot provide guidance for the decision to undertake re-excision surgery in the low-risk group. Further studies are needed to test the score in extensive datasets and better identify low-risk patients. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Breast cancer Residual disease Re-excision Breast conserving therapy Predictive score
Contents 1. 2.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00
cologie, CHI Poissy-Saint-Germain-en* Corresponding author. Service de gyne Laye, 10 rue du Champ Gaillard, 78300 Poissy, France. E-mail addresses: [email protected] (M. Mimouni), fabrice.lecuru@ egp.aphp.fr (F. Lecuru), [email protected] (R. Rouzier), noeloter@hotmail. com (N. Lotersztajn), [email protected] (D. Heitz), [email protected] (J. Cohen), [email protected] (A. Fauconnier), [email protected] (C. Huchon). http://dx.doi.org/10.1016/j.suronc.2015.08.003 0960-7404/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
2
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
3.
4. 5.
2.3. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Development of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Validation of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Characteristics of the patients in the two groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Univariate analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Multivariate analysis and development of a score for RD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Development of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Validation of the RD score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Disclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Author contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1. Introduction
2. Material and methods
Breast-conserving therapy for the treatment of breast cancer involves two phases: partial mastectomy and radiotherapy [1e3]. Several prospective, randomized trials have shown that this treatment is equivalent to a total mastectomy in terms of overall survival and metastasis-free survival [4e6]. Breast-conserving therapy involves the resection the tumor and provides a satisfactory cosmetic result for the patient. The indications for this treatment have expanded since the development of oncoplastic techniques appropriate for resecting large tumors [7,8]. The rate of local recurrence after breast-conserving therapy is significantly lower in patients receiving systemic adjuvant therapy (hormonal therapy and chemotherapy) than in those not receiving such therapies [9]. Nevertheless, the rate of recurrence after breast-conserving therapy is higher than that after total mastectomy [10]. A positive resection margin (RM) is the main risk factor for recurrence [11], and systematically leads to subsequent revision surgery, regardless of the histological characteristics of the tumor [12]. However, the distance between the tumor and the RM that is necessary to define a negative RM has been the subject of many studies and controversies [13]. The absence of a consensus has resulted in a large variability in practice among surgeons regarding decisions to re-excise close margins [14,15]. The American Society of Clinical Oncology (ASCO), the American Society for Radiation Oncology (ASTRO) and the Society of Surgical Oncology (SSO) recently published guidelines [16] stating that the distance between the tumor and the RM does not significantly affect the rate of local recurrence for invasive carcinoma. Thus, according to these guidelines, a negative RM (no ink on cancer cells) indicates complete resection of the tumor [17]. Although the benefit of re-excision for positive margins has been shown, histological analysis of re-excisions reveals that in 30e70% of such cases, the patients show no sign of residual disease (RD) [18]. Thus, reoperation has no real benefit on the prognosis of these patients. They are nevertheless subjected to the risks associated with anesthesia and an unnecessary procedure, which are particularly dangerous for vulnerable people (elderly patients and those with comorbidities). In addition, surgical reexcision worsens the cosmetic result of the first operation, and generates additional anxiety for patients and a substantial increase in the cost of care. The objective of our study was to develop a predictive score for residual disease (RD) for positive resection margins following partial mastectomy.
This study was approved by the ethics committee for Research in Obstetrics and Gynecology (CEROG 2013-GYN-06-01-R1). 2.1. Population This multicenter study was carried out retrospectively in three French centers: the Centre Hospitalier de Poissy-Saint-Germain (January 2007eOctober 2012), Georges Pompidou European Hospital (January 2011eJanuary 2013) and the Curie Institute (August 2013eMarch 2014). All patients with breast cancer treated surgically by partial mastectomy who then underwent subsequent surgery because of a positive RM were included. Patients who underwent re-excision surgery for a close (but non-positive) RM were excluded. Clinical, radiological and pathological data from patients treated at the Centre Hospitalier de Poissy Saint Germain or at Georges Pompidou European Hospital were collected and used to develop a predictive score for RD. Data from patients treated at the Curie Institute were used to validate this score. 2.2. Data collection All variables (clinical, biological and radiological) commonly used in breast cancer care were selected from medical records and recorded in a data collection form. During the initial consultation, a complete clinical examination was performed and menopausal status, family history and the stage of cancer according to TNM classification were recorded. The imaging evaluation included bilateral mammography and bilateral breast ultrasound and in some cases MRI. This evaluation was carried out to assess the characteristics of the lesion prior to therapy, including its size, focality, nodular character, and the existence of suspicious clusters of microcalcifications. The radiological tumor size (RTS) was defined as the largest diameter of the lesion (be it a nodule or an area of microcalcification) as determined by one of the three imaging methods. If the tumor was non-palpable, it was located either before the operation by a radiologist or during the operation by the surgeon using ultrasound. During the partial mastectomy, re-excision was performed if the surgeon identified a suspected lesion on the RM or if they estimated that the palpable tumor was macroscopically too close to the RM. No surgeon routinely excised cavity shave margins. The histological report of the surgical specimen included descriptions of the following characteristics: the histological type of
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
lesion (ductal or lobular); the infiltrative nature of the lesion; the size of the tumor component in situ and that of the infiltrating component as well as the number of infiltrating satellite lesions; the Elston-Ellis (EE) histoprognostic grade; the grade of the in situ component; the presence of estrogen or progesterone hormone receptors (HR); HER2 status; the presence of lymphovascular tumor emboli; and any lymph node involvement. The pathological tumor size (PTS) was defined as the largest diameter of the tumor, regardless of its character (in situ or infiltrating). For infiltrating lesions, PTS was calculated from the size of the main nodule without taking into account satellite lesions. The four resection margins that were taken into consideration for reexcision surgery were the internal, external, superior and inferior margins (the superficial and deep margins are not relevant for reexcision surgery). The pathologist recorded the following characteristics of the resection margins: the number of positive RMs; the linear extent of the positive RM in millimeters (corresponding to the largest diameter of invasion of each margin); the cumulative length of all positive margins (corresponding to the addition of the four affected margins of the surgical specimen); the histological type involved in the positive RM and whether it was an in situ or infiltrating component of the tumor (in situ carcinoma only, infiltrating carcinoma only, or both in situ and infiltrating carcinoma); and the location of the positive RM with respect to the orientation of the specimen (internal, external, superior or inferior). Margins of less than 1 mm were classified as focal involvement. This was a retrospective, multicentric study. No standardized protocol was thus used for the histological analysis of the surgical specimen; instead, it was performed according to the practice of each pathologist and center and was described in detail in the pathology report. After revision surgery involving either a second breastconserving surgery or total mastectomy, the pathologist examined the re-excision specimen for the presence of RD. 2.3. Statistical analysis Chi-2, Fisher and ManneWhitney tests were used to verify that the characteristics of the population used to develop the score and that used to validate it were similar. 2.4. Development of the score Values missing from the database were completed by multiple imputation [19]. A logistic model was used to identify factors associated with the presence of RD. Variables with a p value of <0.20 in the univariate analysis were included in the multivariate analysis. The multivariate model was constructed by selecting the variables to maximize the area under the ROC. The HosmereLemeshow test was used to evaluate the fit of the model to the data. The RD score was obtained by rounding the b coefficients of the logistic regression to the nearest unit to build an easy to use scale. We then verified that the area under the ROC of the score and that under the ROC of the model were not significantly different. Finally, we created different risk groups by combining scores for which the proportion of patients with residual disease was not significantly different.
Table 1 Clinical, radiological, and pathological characteristics of the modeling and validation population. Characteristics
Modeling group n or median [IQR]
Number of patients Age 40 41e65 >65 Family history of breast cancer Menopause Palpable tumor Microcalcifications Radiological tumor size assessed by MRI Radiological tumor size (RTS) Pre- or intraoperative imaging exam for non-palpable tumors Intra-operative specimen radiography Histological type Ductal Lobular Invasive tumor DCIS Ellis and Elston grading I II III DCIS grading Low Intermediate High Hormone receptor negative status HER2 positive status Pathological tumor size (PTS) > 30 mm pN-stage Chemotherapy before the reexcision Discrepancy > 50% (PTS/RTS) Radical mastectomy at the reexcision Time between lumpectomy and reexcision Residual disease (RD)
Validation group p-value
% n or median [IQR]
148 55 [47e66] 8 101 39 34 96 77 66 34
5 68 26 23 65 52 45 24
10 [6e18] 71
10 [5e20] 48 38
0.82 57 0,27
37
25 21
31 0,35
124 26 109 103
84 18 74 70
57 16 55 58
85 24 82 87
24 55 25
23 8 53 28 24 12
17 58 25
28 53 13 15
30 56 14 10
0.10 43 53 4 9 0.79
16 25 [15e40] 49 60 11
17 2 24 [13e35] 36 22 41 15 7 1
5 0.06 0.49 34 0.76 22 0.01 2 0.08
70 55
60 41 37 18
65 0.49 27 0.16
40 [28e56] 94
67 53 [45e63] 7 43 14 26 42 29 29 19
%
21 26 2 6
35 [28e43] 64 32
11 67 22 39 63 43 44 28
0.29 0.32
0.02 0.76 0.24 0.93 0.46
0.81 0.28 0.21 0.01 0.66
0.12 48 0.03
Table 2 TNM classification for the modeling and validation group. Classification
T
N M
2.5. Validation of the score A ROC was constructed to confirm to the predictive power of the score in the population used for validation. The Chi-2 test was used to verify that the occurrence of RD in each risk group was not significantly different between the two populations. Statistical analysis was performed with SPSS v15.0 and Stata v12.0 software (Stata Corp., College Station, TX).
3
T0 T1 T2 T3 T4 N0 N1 M0 M1
Modeling group
Validation group
N
%
n
%
74 57 12 3 2 140 8 148 0
50 39 8 2 1 95 5 100 0
38 16 13 0 0 60 7 67 0
57 24 19 0 0 90 10 100 0
3. Results 3.1. Characteristics of the patients in the two groups This study included 2429 patients who underwent a partial
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
4
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
Table 3 Univariate analysis and multivariate analysis. Univariate analysis
Age 40 41e65 > 65 Family history of breast cancer BMI 25 Menopause Palpable tumor T0 T1, T2 T3, T4 Intra-operative specimen radiography Microcalcifications pN-stage Number of positive margins Cumulative length of all positive margins > 5 mm Invasion by DCIS only DCIS grading - High - Intermediate - Low Ellis and Elston grading - grade 1 - grade 2 - grade 3 Lympho-vascular invasion Hormone receptor negative status Time between lumpectomy and re-excision Chemotherapy before reexcision PTS <30 mm without discrepancy >50% (PTS/RTS) PTS< 30 mm with discrepancy >50% (PTS/RTS) PTS > 30 mm (With or without discrepancy >50% (PTS/RTS)
Table 4 RD-score and division of patients according to predicted risk of RD. Multivariate analysis
OR (IC 95%)
p-value ORA (IC 95%)
(ref) 0.49 [0.09e2.54] 0.85 [0.15e4.86] 1.07 (0.48e2.38)
0.39 0.85 0.87
Predictors of RD
0,36
1.63 (0.82e3.23) 0.87 (0.44e1.73) 1.68 (1.01e2.78)
0.16 0.70 0.04
2.19 (1.01e4.73)
0.04
2.37 (1.04e5.41)
2.13 (1.08e4.17)
0.03
2.56 (1.20e5.26)
1.45 (0.56e3.79) 1.34 (0.61e2.93) 1.10 (0.32e3.81)
0.45 0.47 0.88
0.59 0.68 0.63 1.09
(0.21e1.66) (0.29e1.58) (0.22e1.76) (0.43e2.77)
0.32 0.37 0.38 0.85
0.39 (0.12e1.34)
0.14
0.97 (0.93e1.01)
0.10
0.45 (0.13e1.55)
0.21
0.29 0.73 0.71 0.62 0.41
0.03
5.04 (1.81e13.98) 0.002
Validation population
2 2 3 2
7 16% (8%e24%)
25%
65% (61%e69%)
48%
100%
100%
RD: Residual Disease, RTS: Radiological Tumor Size, PTS: Pathological Tumor Size, DCIS: Ductal carcinoma In Situ.
2.57 (1.04e6.34)
0.04
The population used to develop the score (modeling population; MP) was made up of 148 patients and that used to validate it (validation population; VP) was made up of 67 patients. In the MP, 64% of patients had RD versus 48% in the VP (p ¼ 0.03). The clinical, radiological and histological characteristics of the patients are shown in Table 1. The clinical characteristics of cancer according to TNM classification are described in Table 2. Standard imaging tests, supplemented by MRI in a few cases (24% in the MP vs. 28% in the VP, p ¼ 0.46), showed the existence of microcalcifications that were comparable in the two groups (45% vs. 44%, p ¼ 0.93). Hormone receptors (HR) were expressed in most patients with an invasive lesion (10% of the MP vs. 9% of the VP were classified as HR negative, p ¼ 0.79). The protein HER2 was overexpressed in 17% of patients in the MP vs. only 5% of patients in the VP (p ¼ 0.06). The PTS was similar between the two groups, with a median of 25 mm in the MP and 24 mm in the VP (p ¼ 0.49). The RTS was also similar (10 mm in both groups, p ¼ 0.82) and there was a discrepancy of >50% between the RTS and PTS for similar proportions of patients in the two groups (60% of the MP vs. 65% of the VP, p ¼ 0.49). Revision surgery involved a total mastectomy for 37% of patients in the MP and 27% of patients in the VP. Recovery times were similar for the two groups (40 days for the MP vs. 35 days for the VP, p ¼ 0.12).
5.07 (1.71e15.01)
0.003
3.2. Univariate analysis
0.04
002
(Ref)
2.28 (1.09e4.78)
Modeling population Size of margin invasion > 5 mm Invasion by DCIS only PTS > 30 mm or PTS< 30 mm with discrepancy >50% (PTS/RTS) Total Low risk 0 Moderate risk 2e6 High risk 7
1.49 (0.72e3.08) 0.88 (0.44e1.79) 1.14 (0.58e2.22) 1.19 (0.60e2.34) (ref) 2.57 (0.27e24.25) 1,46 (0,65e3,26)
Points Predicted risk of RD % (CI95%)
p-value
Factors significantly associated with the presence of RD in the univariate analysis were: a cumulative length of all positive
RD: Residual Disease, RTS: Radiological Tumor Size, PTS: Pathological Tumor Size, DCIS: Ductal carcinoma In Situ. Bold¼factors significantly associated with presence of RD (p<0,05).
mastectomy: 866 at Poissy-Saint-Germain, 517 at Georges Pompidou European hospital and 1046 at the Curie Institute. A total of 258 patients (10.5%) underwent re-excision surgery one or more times following the finding of a positive RM upon histological examination of the surgical specimen removed during the partial mastectomy. Forty-three of the 258 patients were excluded from the study for the following reasons: 12 had clear margins (large distance between the margin and the tumor lesion or absence of tumor lesion in the surgical specimen); nine had a rare histological type of tumor; 11 received neoadjuvant chemotherapy; and 11 patients had files missing from their records.
Fig. 1. Prevalence of RD in low, moderate, and high risk groups in both modeling and validation populations.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
margins >5 mm (OR (95%CI) ¼ 2.19 (1.01e4.73)), invasion by ductal carcinoma in situ (DCIS) only (OR (95%CI) ¼ 2.13 (1.08e4.17)), a PTS > 30 mm (OR (95%CI) ¼ 5.04 (1.81e13.93)) or a PTS < 30 mm and a discrepancy of >50% between the RTS and the PTH (OR (95% CI) ¼ 2.28 (1.09e4.78)), and the number of positive margins (OR (95%CI) ¼ 1.68 (1.01e2.78)) (see Table 3). 3.3. Multivariate analysis and development of a score for RD Factors independently associated with the presence of RD and retained in the predictive model were: a cumulative length of all positive margins >5 mm (aOR (95%CI) ¼ 2.37 (1.04e5.41)), invasion by DCIS only (aOR (95%CI) ¼ 2.56 (1.20e5.26)), a PTS > 30 mm (aOR (95%CI) ¼ 5.07 (1.71e15.01)) or a PTS < 30 mm, and a discrepancy of >50% between the PTS and the RTS (aOR (95%CI) ¼ 2.57 (1.04e6.34)) (see Table 3). 3.4. Development of the score The score was constructed from the following factors: a cumulative length of all positive margins >5 mm (2 points), a PTS > 30 mm (3 points), a PTS < 30 mm associated with a discrepancy of >50% between the RTS and the PTS (2 points) and invasion by DCIS only (2 points). Note that a TTH >30 mm and a TTH <30 mm associated with a discrepancy of >50% between the TTH and the RTS are mutually exclusive factors. The score was thus calculated on a scale of 0e7. Three risk groups were defined: 9% of patients belonged to the low risk group, 82% to the moderate risk group and 9% to the high risk group. The low risk group was defined by a RD score of 0: 16% of patients belonging to the group had RD (95% CI ¼ 0%e39%). The moderate risk group was defined by a score between 2 and 6: 65% of patients belonging to the group had RD (95% CI ¼ 45%e85%). Finally, all patients classified in the high risk group defined by a score of 7 showed RD according to the analysis of the surgical specimen removed during revision surgery (see Table 4). 3.5. Validation of the RD score The model was consistent with the data according to the HosmereLemeshow test (p ¼ 0.74). The area under the ROC curve of the RD score was 0.72 (95% CI, 0.68e0.75)) and was not significantly different from that under the ROC curve of the logistic regression model (AUC-ROC ¼ 0.72 (95% CI: 0.68e0.75) p ¼ 0.60) (see Table 2).
Fig. 2. ROC of RD score.
5
The area under the ROC curve in the validation population was 0.69 (95% CI, 0.56e0.82). In the validation population, 19% of patients belonged to the low risk group, 71% to the moderate risk group and 10% to the high risk group. Among patients in the low risk group, 25% had RD (vs. 16% in the MP, p ¼ 0.62). Among patients in the moderate risk group, 48% had RD (vs. 65% in the MP, p ¼ 0.12). Finally, all patients in the high risk group had RD, consistent with findings in the MP (see Fig. 1). 4. Discussion We are the first to develop a predictive score for the presence of RD detected during revision surgery due to positive margins following partial mastectomy (Fig. 2) . Patients who underwent reexcision surgery for close (non-positive) RMs were not included in the study because such patients do not require re-excision surgery according to the latest recommendations from medical associations. This score is based on a 7 point scale and is calculated from four predictive factors: a cumulative length of all positive margins >5 mm (þ2 points), invasion by DCIS only (þ2 points), a PTS > 30 mm (þ3 points), and a PTS < 30 mm associated with a discrepancy of >50% between the PTS and the RTS (þ2 points). This score can be used to classify patients into three risk groups: patients with a score of 0 were considered as low risk, those with a score between 2 and 6 were considered as moderate risk, and those with a score of 7 were considered as high risk. We validated this scoring system in an independent population. Moreover, our score is applicable to all patients undergoing a partial mastectomy, regardless of their initial TNM classification (T0 to T4), including patients with T3/T4 disease (data not shown). The inclusion of all patients facilitates the use of this score in daily practice. Some of the risk factors that we report here have not previously been identified. Indeed, this is the first description of a significant association between the cumulative length of all positive margins in millimeters and the occurrence of RD. Most studies of RMs describe positive margins as focal or extensive. This may explain why this feature has rarely been specified in millimeters in histology reports until recently. The identification of this novel predictive factor of residual disease constitutes an important advance; however, the retrospective and multicentric nature of our study also reveals a clear lack of standardization in the measurement of the RM in millimeters. In 2011 by Meier and colleagues studied the discrepancy between the RTS and the PTS: they found a significant association between the underestimation of tumor size from imaging and the discovery of a positive RM following a partial mastectomy [20], but an association with the presence of RD was not reported. Tumor size was found to be significantly associated with the presence of RD in univariate but not in multivariate analysis in several studies investigating risk factors of RD [21] [22]. Several other factors have been previously shown to be associated with the presence of RD, including lymph node invasion [23], although we did not find evidence of this association in our study. We did not search for an association between RD and multifocality or young age [24], due to the small number of patients with these characteristics in our study. In addition, we did not analyze various other factors such as the percentage of the surgical specimen containing DCIS and the number of cancerous ducts close to the RM (<2 mm) [25]. In 2010, Margenthaler and colleagues published a predictive model of RD called the margin index. The objective of their study was to identify a factor able to predict the presence of RD with more accuracy than the RM alone. Their method was based on two variables: margin index ¼ closest margin (mm)/tumor size (mm) 100. The risk of residual disease for a margin index >5 was only 3.2%. The sensitivity of the margin index in this study was 85%,
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
6
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
the specificity was 73% and the area under the ROC was 0.88 [26]. This predictive index was tested recently in two independent cohorts [27] [28]; However, neither study found that margin index was significantly associated with RD in their population. Another descriptive model was proposed in 2009 [29]. This model is based on three variables: the ratio between the size of the surgical specimen and the tumor, the distance of carcinoma extension into surrounding tissue, and the amount of carcinoma within 3 mm of the margin. Patients were divided into three risk groups of RD: low (6%), intermediate (27%) or high (43%). However, these two predictive models are no longer applicable in light of recent recommendations for revision surgery for positive RMs, which state that the distance between the tumor and the RM is no longer an indication for revision surgery. Our study has some limitations. Notably, the area under the ROC curve indicated a modest ability to predict RD and only a small number of patients were included. Indeed, factors such as young age, multifocality, lobular histology, triple negative status (HR negative and HER2 negative) and the overexpression of HER2 protein were present in only small proportions of our patients making it difficult to search for associations with RD. In addition, data were collected retrospectively from three different centers; therefore, histological analysis was not standardized. The size of the RM is a novel predictive factor of residual disease. In the future, it would be useful to define a standardized way to measure the size of the RM during histological analysis of surgical specimens obtained from partial mastectomy, to increase the predictive value of this factor for residual disease. Overall, our score identifies patients at high risk of RD: all patients with a score of 7 had RD, both in the development and validation populations. The indication for revision surgery is essential for the therapeutic management of these patients, especially for high-risk surgery patients. However, the proportion of patients with RD in the low risk group is too high to influence the indication for revision surgery when the RM is positive. As it stands, our score therefore cannot be applied in clinical practice to advocate the use of radiotherapy without revision surgery in patients at low risk of RD. In addition, the size of our cohort is too small to enable our score to be applied to the general population. The aim of our study was to devise a valid and reproducible score of residual disease for the first time. This score should greatly facilitate subsequent studies, of which the main objective should be to identify patients at low risk of residual disease when the RM is positive. This would reduce the number of repeat surgical procedures required for these patients, despite the presence of tumor cells at resection margins. Further studies are needed to test this score in larger cohorts and confirm its validity in independent populations. Such studies would enable the exploration of other potential risk factors of RD, such as lobular histology, young age, the absence of hormone receptors and the expression of HER2. The incorporation of additional features would strengthen the score and facilitate the identification of patients at very low risk of RD. 5. Conclusion We constructed a score of RD based on four independent risk factors and validated it in an independent population. This score identifies a group at high risk of RD, for whom surgical reintervention is essential in the management of breast cancer, especially for high-risk surgical patients. However, this score should not be used to determine clinical practice regarding revision surgery for patients at low risk. Further studies are needed to test this score in larger cohorts, search for other risk factors, and better identify patients at low risk of RD.
Disclosures The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article. Author contributions M.M. contributed to data collection, analysis and interpretation, and wrote the manuscript. F.L, R.R., D.H., N.L. and A.F. contributed to the critical revision of the manuscript. J.C. contributed to the conception and design of the study and to statistical analysis. C.H. contributed to the analysis and interpretation of data and to the critical revision of the manuscript. Acknowledgments The authors wish to acknowledge the support of Medistat. References [1] M. Clarke, R. Collins, S. Darby, Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-Year survival: an overview of the randomised trials, Lancet 366 (2005) 2087e2106. [2] C. Verschraegen, Vincent Vinh-Hung, Effects of radiotherapy and surgery for early breast cancer, Lancet 367 (2006) 1654. [3] Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials, Lancet 378 (2011) 1707e1716. [4] U. Veronesi, N. Cascinelli, L. Mariani, Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer, N. Engl. J. Med. 347 (2002) 1227e1232. [5] B. Fisher, S. Anderson, J. Bryant, Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer, N. Engl. J. Med. 347 (2002) 1233e1241. €l, J.J. Mazeron, Long-term results of a randomized trial comparing [6] G. Noe breast-conserving therapy with mastectomy: European Organization for rapie J. de la Research and Treatment of Cancer 10801 trial, Cancer Radiothe te Française de Radiothe rapie Oncol. 5 (2001) 211e212. Socie [7] A.D. Fitoussi, M.G. Berry, F. Fama, Oncoplastic breast surgery for cancer: analysis of 540 consecutive cases [outcomes article], Plastic Reconstr. Surg. 125 (2010) 454e462. [8] M. Rietjens, C.A. Urban, P.C. Rey, Long-term oncological results of breast conservative treatment with oncoplastic surgery, Breastedinbg. Scotl. 16 (2007) 387e395. [9] Early Breast Cancer Trialists' Collaborative Group (EBCTCG), Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials, Lancet 365 (2005) 1687e1717. [10] I. Jatoi, M.A. Proschan, Randomized trials of breast-conserving therapy versus mastectomy for primary breast cancer: a pooled analysis of updated results, Am. J. Clin. Oncol. 28 (2005) 289e294. [11] C.C. Park, M. Mitsumori, A. Nixon, Outcome at 8 years after breast-conserving surgery and radiation therapy for invasive breast cancer: influence of margin status and systemic therapy on local recurrence, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 18 (2000) 1668e1675. fe rentiels sur le cancer du sein in situ, Institut National [12] Recommandations et re du Cancer (INCA), 2009. [13] N. Houssami, P. Macaskill, M.L. Marinovich, Meta-analysis of the impact of surgical margins on local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy, Eur. J. Cancer (Oxford, Engl. 1990) 46 (2010) 3219e3232. [14] L.E. McCahill, R.M. Single, E.J. Aiello Bowels, Variability in reexcision following breast conservation surgery, JAMA J. Am. Med. Assoc. 307 (2012) 467e475. [15] S.L. Blair, K. Thompson, J. Rococco, Attaining negative margins in breastconservation operations: is there a consensus among breast surgeons? J. Am. Coll. Surg. 209 (2009) 608e613. [16] T.A. Buchholz, M.R. Somerfield, J.J. Griegs, Margins for breast-conserving surgery with whole-breast irradiation in stage I and II invasive breast cancer: American Society of Clinical Oncology Endorsement of the Society of Surgical Oncology/American Society for Radiation Oncology Consensus Guideline, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 32 (2014) 1502e1506. [17] N. Houssami, P. Macaskill, M.L. Marinovich, The association of surgical margins and local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy: a meta-analysis, Ann. Surg. Oncol. 21 (2014) 717e730. [18] G. Houvenaeghel, E. Lambaudie, M. Buterelli, Margin status in invasive breast cancer, Bull. Du. Cancer 95 (2008) 1161e1170.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7 [19] S. Van Buuren, Flexible Imputation of Missing Data, 2012. [20] M. Meier-Meitinger, C. Rauh, B. Adamietz, Accuracy of radiological tumour size assessment and the risk for re-excision in a Cohort of Primary Breast Cancer Patients, Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 38 (2012) 44e51. [21] N. Rodriguez, L.K. Diaz, Elizabeth L. Wiley, Predictors of residual disease in repeat excisions for lumpectomies with margins less than 0.1 cm, Clin. Breast Cancer 6 (2005). [22] M.C. Smitt, K. Nowels, R.W. Karlson, Predictors of reexcision findings and recurrence after breast conservation, Int. J. Radiat. Oncol. Biol. Phys. 57 (2003) 979e985. [23] S.O. Gurdal, H. Karanlik, N. Gabioglu, Positive or close margins in breast conserving surgery: is re-excision always necessary? Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 38 (2012) 399e406. [24] M.S. Sabel, K. Rogers, K. Griffith, Residual disease after re-excision
7
lumpectomy for close margins, J. Surg. Oncol. 99 (2009) 99e103. [25] S. Wei, C.P. Kragel, K. Zhang, Factors associated with residual disease after initial breast-conserving surgery for Ductal carcinoma in Situ, Hum. Pathol. 43 (2012) 986e993. [26] J.A. Margenthaler, F. Gao, V.S. Klimberg, Margin index: a new method for prediction of residual disease after breast-conserving surgery, Ann. Surg. Oncol. 17 (2010) 2696e2701. [27] J.C. Bolger, J.G. Solon, C. Power, Analysis of margin index as a method for predicting residual disease after breast-conserving surgery in a European Cancer Center, Ann. Surg. Oncol. 19 (2012) 207e211. [28] Claire Edwards, F. Gao, G.M. Freedman, Margin index: a useful tool for the Breast Surgeon, J. Surg. Res. 190 (1) (2014) 164e169. [29] J.L. Caughran, F.A. Vicini, L.L. Kestin, Optimal use of re-excision in patients diagnosed with early-stage breast cancer by excisional biopsy treated with breast-conserving therapy, Ann. Surg. Oncol. 16 (2009) 3020e3027.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
Contents lists available at ScienceDirect
Surgical Oncology journal homepage: www.elsevier.com/locate/suronc
Review
Reexcision for positive margins in breast cancer: A predictive score of residual disease mie Lotersztajn d, Myriam Mimouni a, b, c, d, *, Fabrice Lecuru c, Roman Rouzier d, e, Noe Denis Heitz a, b, Julien Cohen f, Arnaud Fauconnier a, b, e, Cyrille Huchon a, b, e a
Department of Gynecology and Obstetrics, CHI Poissy-Saint-Germain, Poissy, France University Versailles Saint Quentin en Yvelines, Versailles, 10 rue du Champ Gaillard, 78300 Poissy, France ^pital Europ Department of Gynecologic and Oncologic Surgery, Ho een Georges Pompidou, AP-HP Paris, University Paris V, Ren e Descartes, 20 rue Leblanc, 75015 Paris, France d Department of Surgery and Senology, Institut Curie, Universit e Paris V, Ren e Descartes, 26 rue d’Ulm, 75005 Paris, France e EA 7285 Clinical Risks and Safety on Women's Health, University Versailles-Saint-Quentin en Yvelines, France f Medistat, Biostatistics, 10-12 rue de la Conception, 13004 Marseille, 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 30 March 2015 Received in revised form 6 August 2015 Accepted 9 August 2015
Background: Guidelines recommend re-excision if resection margins are positive in lumpectomy for breast cancer. However, residual disease (RD) is not always found. The aim of our study was to develop a score to predict RD in re-excision specimens. Materials and methods: We carried out a multicenter, retrospective study with two population groups. The ‘modeling’ group was composed of 148 patients treated in the Centre Hospitalier Poissy-St-Germain or the Georges Pompidou European Hospital and the ‘validation’ group was composed of 67 patients treated in Curie Institute. The score was built with a logistic regression model. Results: Factors independently associated with RD were: a cumulative length of all positive margins>5 mm, invasion by ductal carcinoma in situ only, a pathological tumor size>30 mm and a pathological tumor size<30 mm with a discrepancy of >50% between pathological and radiological tumor size. The 7-point score allowed the classification of patients into three risk groups for RD: low (16% of patients experienced RD), moderate (65%) and high (100%). The areas under the ROC curve of the score and the logistic model were 0.72(95%CI:0.68e0.75,p ¼ 0.60). The proportion of RD in each group of the validation population (25%, 48%, and 100% in the low, moderate and high group, respectively) confirmed the accuracy of the score in an independent population. Conclusions: This score enables the identification of patients at high risk of RD but it cannot provide guidance for the decision to undertake re-excision surgery in the low-risk group. Further studies are needed to test the score in extensive datasets and better identify low-risk patients. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Breast cancer Residual disease Re-excision Breast conserving therapy Predictive score
Contents 1. 2.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00
cologie, CHI Poissy-Saint-Germain-en* Corresponding author. Service de gyne Laye, 10 rue du Champ Gaillard, 78300 Poissy, France. E-mail addresses: [email protected] (M. Mimouni), fabrice.lecuru@ egp.aphp.fr (F. Lecuru), [email protected] (R. Rouzier), noeloter@hotmail. com (N. Lotersztajn), [email protected] (D. Heitz), [email protected] (J. Cohen), [email protected] (A. Fauconnier), [email protected] (C. Huchon). http://dx.doi.org/10.1016/j.suronc.2015.08.003 0960-7404/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
2
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
3.
4. 5.
2.3. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Development of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Validation of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Characteristics of the patients in the two groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Univariate analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Multivariate analysis and development of a score for RD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Development of the score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Validation of the RD score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Disclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Author contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1. Introduction
2. Material and methods
Breast-conserving therapy for the treatment of breast cancer involves two phases: partial mastectomy and radiotherapy [1e3]. Several prospective, randomized trials have shown that this treatment is equivalent to a total mastectomy in terms of overall survival and metastasis-free survival [4e6]. Breast-conserving therapy involves the resection the tumor and provides a satisfactory cosmetic result for the patient. The indications for this treatment have expanded since the development of oncoplastic techniques appropriate for resecting large tumors [7,8]. The rate of local recurrence after breast-conserving therapy is significantly lower in patients receiving systemic adjuvant therapy (hormonal therapy and chemotherapy) than in those not receiving such therapies [9]. Nevertheless, the rate of recurrence after breast-conserving therapy is higher than that after total mastectomy [10]. A positive resection margin (RM) is the main risk factor for recurrence [11], and systematically leads to subsequent revision surgery, regardless of the histological characteristics of the tumor [12]. However, the distance between the tumor and the RM that is necessary to define a negative RM has been the subject of many studies and controversies [13]. The absence of a consensus has resulted in a large variability in practice among surgeons regarding decisions to re-excise close margins [14,15]. The American Society of Clinical Oncology (ASCO), the American Society for Radiation Oncology (ASTRO) and the Society of Surgical Oncology (SSO) recently published guidelines [16] stating that the distance between the tumor and the RM does not significantly affect the rate of local recurrence for invasive carcinoma. Thus, according to these guidelines, a negative RM (no ink on cancer cells) indicates complete resection of the tumor [17]. Although the benefit of re-excision for positive margins has been shown, histological analysis of re-excisions reveals that in 30e70% of such cases, the patients show no sign of residual disease (RD) [18]. Thus, reoperation has no real benefit on the prognosis of these patients. They are nevertheless subjected to the risks associated with anesthesia and an unnecessary procedure, which are particularly dangerous for vulnerable people (elderly patients and those with comorbidities). In addition, surgical reexcision worsens the cosmetic result of the first operation, and generates additional anxiety for patients and a substantial increase in the cost of care. The objective of our study was to develop a predictive score for residual disease (RD) for positive resection margins following partial mastectomy.
This study was approved by the ethics committee for Research in Obstetrics and Gynecology (CEROG 2013-GYN-06-01-R1). 2.1. Population This multicenter study was carried out retrospectively in three French centers: the Centre Hospitalier de Poissy-Saint-Germain (January 2007eOctober 2012), Georges Pompidou European Hospital (January 2011eJanuary 2013) and the Curie Institute (August 2013eMarch 2014). All patients with breast cancer treated surgically by partial mastectomy who then underwent subsequent surgery because of a positive RM were included. Patients who underwent re-excision surgery for a close (but non-positive) RM were excluded. Clinical, radiological and pathological data from patients treated at the Centre Hospitalier de Poissy Saint Germain or at Georges Pompidou European Hospital were collected and used to develop a predictive score for RD. Data from patients treated at the Curie Institute were used to validate this score. 2.2. Data collection All variables (clinical, biological and radiological) commonly used in breast cancer care were selected from medical records and recorded in a data collection form. During the initial consultation, a complete clinical examination was performed and menopausal status, family history and the stage of cancer according to TNM classification were recorded. The imaging evaluation included bilateral mammography and bilateral breast ultrasound and in some cases MRI. This evaluation was carried out to assess the characteristics of the lesion prior to therapy, including its size, focality, nodular character, and the existence of suspicious clusters of microcalcifications. The radiological tumor size (RTS) was defined as the largest diameter of the lesion (be it a nodule or an area of microcalcification) as determined by one of the three imaging methods. If the tumor was non-palpable, it was located either before the operation by a radiologist or during the operation by the surgeon using ultrasound. During the partial mastectomy, re-excision was performed if the surgeon identified a suspected lesion on the RM or if they estimated that the palpable tumor was macroscopically too close to the RM. No surgeon routinely excised cavity shave margins. The histological report of the surgical specimen included descriptions of the following characteristics: the histological type of
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
lesion (ductal or lobular); the infiltrative nature of the lesion; the size of the tumor component in situ and that of the infiltrating component as well as the number of infiltrating satellite lesions; the Elston-Ellis (EE) histoprognostic grade; the grade of the in situ component; the presence of estrogen or progesterone hormone receptors (HR); HER2 status; the presence of lymphovascular tumor emboli; and any lymph node involvement. The pathological tumor size (PTS) was defined as the largest diameter of the tumor, regardless of its character (in situ or infiltrating). For infiltrating lesions, PTS was calculated from the size of the main nodule without taking into account satellite lesions. The four resection margins that were taken into consideration for reexcision surgery were the internal, external, superior and inferior margins (the superficial and deep margins are not relevant for reexcision surgery). The pathologist recorded the following characteristics of the resection margins: the number of positive RMs; the linear extent of the positive RM in millimeters (corresponding to the largest diameter of invasion of each margin); the cumulative length of all positive margins (corresponding to the addition of the four affected margins of the surgical specimen); the histological type involved in the positive RM and whether it was an in situ or infiltrating component of the tumor (in situ carcinoma only, infiltrating carcinoma only, or both in situ and infiltrating carcinoma); and the location of the positive RM with respect to the orientation of the specimen (internal, external, superior or inferior). Margins of less than 1 mm were classified as focal involvement. This was a retrospective, multicentric study. No standardized protocol was thus used for the histological analysis of the surgical specimen; instead, it was performed according to the practice of each pathologist and center and was described in detail in the pathology report. After revision surgery involving either a second breastconserving surgery or total mastectomy, the pathologist examined the re-excision specimen for the presence of RD. 2.3. Statistical analysis Chi-2, Fisher and ManneWhitney tests were used to verify that the characteristics of the population used to develop the score and that used to validate it were similar. 2.4. Development of the score Values missing from the database were completed by multiple imputation [19]. A logistic model was used to identify factors associated with the presence of RD. Variables with a p value of <0.20 in the univariate analysis were included in the multivariate analysis. The multivariate model was constructed by selecting the variables to maximize the area under the ROC. The HosmereLemeshow test was used to evaluate the fit of the model to the data. The RD score was obtained by rounding the b coefficients of the logistic regression to the nearest unit to build an easy to use scale. We then verified that the area under the ROC of the score and that under the ROC of the model were not significantly different. Finally, we created different risk groups by combining scores for which the proportion of patients with residual disease was not significantly different.
Table 1 Clinical, radiological, and pathological characteristics of the modeling and validation population. Characteristics
Modeling group n or median [IQR]
Number of patients Age 40 41e65 >65 Family history of breast cancer Menopause Palpable tumor Microcalcifications Radiological tumor size assessed by MRI Radiological tumor size (RTS) Pre- or intraoperative imaging exam for non-palpable tumors Intra-operative specimen radiography Histological type Ductal Lobular Invasive tumor DCIS Ellis and Elston grading I II III DCIS grading Low Intermediate High Hormone receptor negative status HER2 positive status Pathological tumor size (PTS) > 30 mm pN-stage Chemotherapy before the reexcision Discrepancy > 50% (PTS/RTS) Radical mastectomy at the reexcision Time between lumpectomy and reexcision Residual disease (RD)
Validation group p-value
% n or median [IQR]
148 55 [47e66] 8 101 39 34 96 77 66 34
5 68 26 23 65 52 45 24
10 [6e18] 71
10 [5e20] 48 38
0.82 57 0,27
37
25 21
31 0,35
124 26 109 103
84 18 74 70
57 16 55 58
85 24 82 87
24 55 25
23 8 53 28 24 12
17 58 25
28 53 13 15
30 56 14 10
0.10 43 53 4 9 0.79
16 25 [15e40] 49 60 11
17 2 24 [13e35] 36 22 41 15 7 1
5 0.06 0.49 34 0.76 22 0.01 2 0.08
70 55
60 41 37 18
65 0.49 27 0.16
40 [28e56] 94
67 53 [45e63] 7 43 14 26 42 29 29 19
%
21 26 2 6
35 [28e43] 64 32
11 67 22 39 63 43 44 28
0.29 0.32
0.02 0.76 0.24 0.93 0.46
0.81 0.28 0.21 0.01 0.66
0.12 48 0.03
Table 2 TNM classification for the modeling and validation group. Classification
T
N M
2.5. Validation of the score A ROC was constructed to confirm to the predictive power of the score in the population used for validation. The Chi-2 test was used to verify that the occurrence of RD in each risk group was not significantly different between the two populations. Statistical analysis was performed with SPSS v15.0 and Stata v12.0 software (Stata Corp., College Station, TX).
3
T0 T1 T2 T3 T4 N0 N1 M0 M1
Modeling group
Validation group
N
%
n
%
74 57 12 3 2 140 8 148 0
50 39 8 2 1 95 5 100 0
38 16 13 0 0 60 7 67 0
57 24 19 0 0 90 10 100 0
3. Results 3.1. Characteristics of the patients in the two groups This study included 2429 patients who underwent a partial
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
4
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
Table 3 Univariate analysis and multivariate analysis. Univariate analysis
Age 40 41e65 > 65 Family history of breast cancer BMI 25 Menopause Palpable tumor T0 T1, T2 T3, T4 Intra-operative specimen radiography Microcalcifications pN-stage Number of positive margins Cumulative length of all positive margins > 5 mm Invasion by DCIS only DCIS grading - High - Intermediate - Low Ellis and Elston grading - grade 1 - grade 2 - grade 3 Lympho-vascular invasion Hormone receptor negative status Time between lumpectomy and re-excision Chemotherapy before reexcision PTS <30 mm without discrepancy >50% (PTS/RTS) PTS< 30 mm with discrepancy >50% (PTS/RTS) PTS > 30 mm (With or without discrepancy >50% (PTS/RTS)
Table 4 RD-score and division of patients according to predicted risk of RD. Multivariate analysis
OR (IC 95%)
p-value ORA (IC 95%)
(ref) 0.49 [0.09e2.54] 0.85 [0.15e4.86] 1.07 (0.48e2.38)
0.39 0.85 0.87
Predictors of RD
0,36
1.63 (0.82e3.23) 0.87 (0.44e1.73) 1.68 (1.01e2.78)
0.16 0.70 0.04
2.19 (1.01e4.73)
0.04
2.37 (1.04e5.41)
2.13 (1.08e4.17)
0.03
2.56 (1.20e5.26)
1.45 (0.56e3.79) 1.34 (0.61e2.93) 1.10 (0.32e3.81)
0.45 0.47 0.88
0.59 0.68 0.63 1.09
(0.21e1.66) (0.29e1.58) (0.22e1.76) (0.43e2.77)
0.32 0.37 0.38 0.85
0.39 (0.12e1.34)
0.14
0.97 (0.93e1.01)
0.10
0.45 (0.13e1.55)
0.21
0.29 0.73 0.71 0.62 0.41
0.03
5.04 (1.81e13.98) 0.002
Validation population
2 2 3 2
7 16% (8%e24%)
25%
65% (61%e69%)
48%
100%
100%
RD: Residual Disease, RTS: Radiological Tumor Size, PTS: Pathological Tumor Size, DCIS: Ductal carcinoma In Situ.
2.57 (1.04e6.34)
0.04
The population used to develop the score (modeling population; MP) was made up of 148 patients and that used to validate it (validation population; VP) was made up of 67 patients. In the MP, 64% of patients had RD versus 48% in the VP (p ¼ 0.03). The clinical, radiological and histological characteristics of the patients are shown in Table 1. The clinical characteristics of cancer according to TNM classification are described in Table 2. Standard imaging tests, supplemented by MRI in a few cases (24% in the MP vs. 28% in the VP, p ¼ 0.46), showed the existence of microcalcifications that were comparable in the two groups (45% vs. 44%, p ¼ 0.93). Hormone receptors (HR) were expressed in most patients with an invasive lesion (10% of the MP vs. 9% of the VP were classified as HR negative, p ¼ 0.79). The protein HER2 was overexpressed in 17% of patients in the MP vs. only 5% of patients in the VP (p ¼ 0.06). The PTS was similar between the two groups, with a median of 25 mm in the MP and 24 mm in the VP (p ¼ 0.49). The RTS was also similar (10 mm in both groups, p ¼ 0.82) and there was a discrepancy of >50% between the RTS and PTS for similar proportions of patients in the two groups (60% of the MP vs. 65% of the VP, p ¼ 0.49). Revision surgery involved a total mastectomy for 37% of patients in the MP and 27% of patients in the VP. Recovery times were similar for the two groups (40 days for the MP vs. 35 days for the VP, p ¼ 0.12).
5.07 (1.71e15.01)
0.003
3.2. Univariate analysis
0.04
002
(Ref)
2.28 (1.09e4.78)
Modeling population Size of margin invasion > 5 mm Invasion by DCIS only PTS > 30 mm or PTS< 30 mm with discrepancy >50% (PTS/RTS) Total Low risk 0 Moderate risk 2e6 High risk 7
1.49 (0.72e3.08) 0.88 (0.44e1.79) 1.14 (0.58e2.22) 1.19 (0.60e2.34) (ref) 2.57 (0.27e24.25) 1,46 (0,65e3,26)
Points Predicted risk of RD % (CI95%)
p-value
Factors significantly associated with the presence of RD in the univariate analysis were: a cumulative length of all positive
RD: Residual Disease, RTS: Radiological Tumor Size, PTS: Pathological Tumor Size, DCIS: Ductal carcinoma In Situ. Bold¼factors significantly associated with presence of RD (p<0,05).
mastectomy: 866 at Poissy-Saint-Germain, 517 at Georges Pompidou European hospital and 1046 at the Curie Institute. A total of 258 patients (10.5%) underwent re-excision surgery one or more times following the finding of a positive RM upon histological examination of the surgical specimen removed during the partial mastectomy. Forty-three of the 258 patients were excluded from the study for the following reasons: 12 had clear margins (large distance between the margin and the tumor lesion or absence of tumor lesion in the surgical specimen); nine had a rare histological type of tumor; 11 received neoadjuvant chemotherapy; and 11 patients had files missing from their records.
Fig. 1. Prevalence of RD in low, moderate, and high risk groups in both modeling and validation populations.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
margins >5 mm (OR (95%CI) ¼ 2.19 (1.01e4.73)), invasion by ductal carcinoma in situ (DCIS) only (OR (95%CI) ¼ 2.13 (1.08e4.17)), a PTS > 30 mm (OR (95%CI) ¼ 5.04 (1.81e13.93)) or a PTS < 30 mm and a discrepancy of >50% between the RTS and the PTH (OR (95% CI) ¼ 2.28 (1.09e4.78)), and the number of positive margins (OR (95%CI) ¼ 1.68 (1.01e2.78)) (see Table 3). 3.3. Multivariate analysis and development of a score for RD Factors independently associated with the presence of RD and retained in the predictive model were: a cumulative length of all positive margins >5 mm (aOR (95%CI) ¼ 2.37 (1.04e5.41)), invasion by DCIS only (aOR (95%CI) ¼ 2.56 (1.20e5.26)), a PTS > 30 mm (aOR (95%CI) ¼ 5.07 (1.71e15.01)) or a PTS < 30 mm, and a discrepancy of >50% between the PTS and the RTS (aOR (95%CI) ¼ 2.57 (1.04e6.34)) (see Table 3). 3.4. Development of the score The score was constructed from the following factors: a cumulative length of all positive margins >5 mm (2 points), a PTS > 30 mm (3 points), a PTS < 30 mm associated with a discrepancy of >50% between the RTS and the PTS (2 points) and invasion by DCIS only (2 points). Note that a TTH >30 mm and a TTH <30 mm associated with a discrepancy of >50% between the TTH and the RTS are mutually exclusive factors. The score was thus calculated on a scale of 0e7. Three risk groups were defined: 9% of patients belonged to the low risk group, 82% to the moderate risk group and 9% to the high risk group. The low risk group was defined by a RD score of 0: 16% of patients belonging to the group had RD (95% CI ¼ 0%e39%). The moderate risk group was defined by a score between 2 and 6: 65% of patients belonging to the group had RD (95% CI ¼ 45%e85%). Finally, all patients classified in the high risk group defined by a score of 7 showed RD according to the analysis of the surgical specimen removed during revision surgery (see Table 4). 3.5. Validation of the RD score The model was consistent with the data according to the HosmereLemeshow test (p ¼ 0.74). The area under the ROC curve of the RD score was 0.72 (95% CI, 0.68e0.75)) and was not significantly different from that under the ROC curve of the logistic regression model (AUC-ROC ¼ 0.72 (95% CI: 0.68e0.75) p ¼ 0.60) (see Table 2).
Fig. 2. ROC of RD score.
5
The area under the ROC curve in the validation population was 0.69 (95% CI, 0.56e0.82). In the validation population, 19% of patients belonged to the low risk group, 71% to the moderate risk group and 10% to the high risk group. Among patients in the low risk group, 25% had RD (vs. 16% in the MP, p ¼ 0.62). Among patients in the moderate risk group, 48% had RD (vs. 65% in the MP, p ¼ 0.12). Finally, all patients in the high risk group had RD, consistent with findings in the MP (see Fig. 1). 4. Discussion We are the first to develop a predictive score for the presence of RD detected during revision surgery due to positive margins following partial mastectomy (Fig. 2) . Patients who underwent reexcision surgery for close (non-positive) RMs were not included in the study because such patients do not require re-excision surgery according to the latest recommendations from medical associations. This score is based on a 7 point scale and is calculated from four predictive factors: a cumulative length of all positive margins >5 mm (þ2 points), invasion by DCIS only (þ2 points), a PTS > 30 mm (þ3 points), and a PTS < 30 mm associated with a discrepancy of >50% between the PTS and the RTS (þ2 points). This score can be used to classify patients into three risk groups: patients with a score of 0 were considered as low risk, those with a score between 2 and 6 were considered as moderate risk, and those with a score of 7 were considered as high risk. We validated this scoring system in an independent population. Moreover, our score is applicable to all patients undergoing a partial mastectomy, regardless of their initial TNM classification (T0 to T4), including patients with T3/T4 disease (data not shown). The inclusion of all patients facilitates the use of this score in daily practice. Some of the risk factors that we report here have not previously been identified. Indeed, this is the first description of a significant association between the cumulative length of all positive margins in millimeters and the occurrence of RD. Most studies of RMs describe positive margins as focal or extensive. This may explain why this feature has rarely been specified in millimeters in histology reports until recently. The identification of this novel predictive factor of residual disease constitutes an important advance; however, the retrospective and multicentric nature of our study also reveals a clear lack of standardization in the measurement of the RM in millimeters. In 2011 by Meier and colleagues studied the discrepancy between the RTS and the PTS: they found a significant association between the underestimation of tumor size from imaging and the discovery of a positive RM following a partial mastectomy [20], but an association with the presence of RD was not reported. Tumor size was found to be significantly associated with the presence of RD in univariate but not in multivariate analysis in several studies investigating risk factors of RD [21] [22]. Several other factors have been previously shown to be associated with the presence of RD, including lymph node invasion [23], although we did not find evidence of this association in our study. We did not search for an association between RD and multifocality or young age [24], due to the small number of patients with these characteristics in our study. In addition, we did not analyze various other factors such as the percentage of the surgical specimen containing DCIS and the number of cancerous ducts close to the RM (<2 mm) [25]. In 2010, Margenthaler and colleagues published a predictive model of RD called the margin index. The objective of their study was to identify a factor able to predict the presence of RD with more accuracy than the RM alone. Their method was based on two variables: margin index ¼ closest margin (mm)/tumor size (mm) 100. The risk of residual disease for a margin index >5 was only 3.2%. The sensitivity of the margin index in this study was 85%,
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
6
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7
the specificity was 73% and the area under the ROC was 0.88 [26]. This predictive index was tested recently in two independent cohorts [27] [28]; However, neither study found that margin index was significantly associated with RD in their population. Another descriptive model was proposed in 2009 [29]. This model is based on three variables: the ratio between the size of the surgical specimen and the tumor, the distance of carcinoma extension into surrounding tissue, and the amount of carcinoma within 3 mm of the margin. Patients were divided into three risk groups of RD: low (6%), intermediate (27%) or high (43%). However, these two predictive models are no longer applicable in light of recent recommendations for revision surgery for positive RMs, which state that the distance between the tumor and the RM is no longer an indication for revision surgery. Our study has some limitations. Notably, the area under the ROC curve indicated a modest ability to predict RD and only a small number of patients were included. Indeed, factors such as young age, multifocality, lobular histology, triple negative status (HR negative and HER2 negative) and the overexpression of HER2 protein were present in only small proportions of our patients making it difficult to search for associations with RD. In addition, data were collected retrospectively from three different centers; therefore, histological analysis was not standardized. The size of the RM is a novel predictive factor of residual disease. In the future, it would be useful to define a standardized way to measure the size of the RM during histological analysis of surgical specimens obtained from partial mastectomy, to increase the predictive value of this factor for residual disease. Overall, our score identifies patients at high risk of RD: all patients with a score of 7 had RD, both in the development and validation populations. The indication for revision surgery is essential for the therapeutic management of these patients, especially for high-risk surgery patients. However, the proportion of patients with RD in the low risk group is too high to influence the indication for revision surgery when the RM is positive. As it stands, our score therefore cannot be applied in clinical practice to advocate the use of radiotherapy without revision surgery in patients at low risk of RD. In addition, the size of our cohort is too small to enable our score to be applied to the general population. The aim of our study was to devise a valid and reproducible score of residual disease for the first time. This score should greatly facilitate subsequent studies, of which the main objective should be to identify patients at low risk of residual disease when the RM is positive. This would reduce the number of repeat surgical procedures required for these patients, despite the presence of tumor cells at resection margins. Further studies are needed to test this score in larger cohorts and confirm its validity in independent populations. Such studies would enable the exploration of other potential risk factors of RD, such as lobular histology, young age, the absence of hormone receptors and the expression of HER2. The incorporation of additional features would strengthen the score and facilitate the identification of patients at very low risk of RD. 5. Conclusion We constructed a score of RD based on four independent risk factors and validated it in an independent population. This score identifies a group at high risk of RD, for whom surgical reintervention is essential in the management of breast cancer, especially for high-risk surgical patients. However, this score should not be used to determine clinical practice regarding revision surgery for patients at low risk. Further studies are needed to test this score in larger cohorts, search for other risk factors, and better identify patients at low risk of RD.
Disclosures The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article. Author contributions M.M. contributed to data collection, analysis and interpretation, and wrote the manuscript. F.L, R.R., D.H., N.L. and A.F. contributed to the critical revision of the manuscript. J.C. contributed to the conception and design of the study and to statistical analysis. C.H. contributed to the analysis and interpretation of data and to the critical revision of the manuscript. Acknowledgments The authors wish to acknowledge the support of Medistat. References [1] M. Clarke, R. Collins, S. Darby, Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-Year survival: an overview of the randomised trials, Lancet 366 (2005) 2087e2106. [2] C. Verschraegen, Vincent Vinh-Hung, Effects of radiotherapy and surgery for early breast cancer, Lancet 367 (2006) 1654. [3] Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials, Lancet 378 (2011) 1707e1716. [4] U. Veronesi, N. Cascinelli, L. Mariani, Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer, N. Engl. J. Med. 347 (2002) 1227e1232. [5] B. Fisher, S. Anderson, J. Bryant, Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer, N. Engl. J. Med. 347 (2002) 1233e1241. €l, J.J. Mazeron, Long-term results of a randomized trial comparing [6] G. Noe breast-conserving therapy with mastectomy: European Organization for rapie J. de la Research and Treatment of Cancer 10801 trial, Cancer Radiothe te Française de Radiothe rapie Oncol. 5 (2001) 211e212. Socie [7] A.D. Fitoussi, M.G. Berry, F. Fama, Oncoplastic breast surgery for cancer: analysis of 540 consecutive cases [outcomes article], Plastic Reconstr. Surg. 125 (2010) 454e462. [8] M. Rietjens, C.A. Urban, P.C. Rey, Long-term oncological results of breast conservative treatment with oncoplastic surgery, Breastedinbg. Scotl. 16 (2007) 387e395. [9] Early Breast Cancer Trialists' Collaborative Group (EBCTCG), Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials, Lancet 365 (2005) 1687e1717. [10] I. Jatoi, M.A. Proschan, Randomized trials of breast-conserving therapy versus mastectomy for primary breast cancer: a pooled analysis of updated results, Am. J. Clin. Oncol. 28 (2005) 289e294. [11] C.C. Park, M. Mitsumori, A. Nixon, Outcome at 8 years after breast-conserving surgery and radiation therapy for invasive breast cancer: influence of margin status and systemic therapy on local recurrence, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 18 (2000) 1668e1675. fe rentiels sur le cancer du sein in situ, Institut National [12] Recommandations et re du Cancer (INCA), 2009. [13] N. Houssami, P. Macaskill, M.L. Marinovich, Meta-analysis of the impact of surgical margins on local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy, Eur. J. Cancer (Oxford, Engl. 1990) 46 (2010) 3219e3232. [14] L.E. McCahill, R.M. Single, E.J. Aiello Bowels, Variability in reexcision following breast conservation surgery, JAMA J. Am. Med. Assoc. 307 (2012) 467e475. [15] S.L. Blair, K. Thompson, J. Rococco, Attaining negative margins in breastconservation operations: is there a consensus among breast surgeons? J. Am. Coll. Surg. 209 (2009) 608e613. [16] T.A. Buchholz, M.R. Somerfield, J.J. Griegs, Margins for breast-conserving surgery with whole-breast irradiation in stage I and II invasive breast cancer: American Society of Clinical Oncology Endorsement of the Society of Surgical Oncology/American Society for Radiation Oncology Consensus Guideline, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 32 (2014) 1502e1506. [17] N. Houssami, P. Macaskill, M.L. Marinovich, The association of surgical margins and local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy: a meta-analysis, Ann. Surg. Oncol. 21 (2014) 717e730. [18] G. Houvenaeghel, E. Lambaudie, M. Buterelli, Margin status in invasive breast cancer, Bull. Du. Cancer 95 (2008) 1161e1170.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003
M. Mimouni et al. / Surgical Oncology xxx (2015) 1e7 [19] S. Van Buuren, Flexible Imputation of Missing Data, 2012. [20] M. Meier-Meitinger, C. Rauh, B. Adamietz, Accuracy of radiological tumour size assessment and the risk for re-excision in a Cohort of Primary Breast Cancer Patients, Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 38 (2012) 44e51. [21] N. Rodriguez, L.K. Diaz, Elizabeth L. Wiley, Predictors of residual disease in repeat excisions for lumpectomies with margins less than 0.1 cm, Clin. Breast Cancer 6 (2005). [22] M.C. Smitt, K. Nowels, R.W. Karlson, Predictors of reexcision findings and recurrence after breast conservation, Int. J. Radiat. Oncol. Biol. Phys. 57 (2003) 979e985. [23] S.O. Gurdal, H. Karanlik, N. Gabioglu, Positive or close margins in breast conserving surgery: is re-excision always necessary? Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 38 (2012) 399e406. [24] M.S. Sabel, K. Rogers, K. Griffith, Residual disease after re-excision
7
lumpectomy for close margins, J. Surg. Oncol. 99 (2009) 99e103. [25] S. Wei, C.P. Kragel, K. Zhang, Factors associated with residual disease after initial breast-conserving surgery for Ductal carcinoma in Situ, Hum. Pathol. 43 (2012) 986e993. [26] J.A. Margenthaler, F. Gao, V.S. Klimberg, Margin index: a new method for prediction of residual disease after breast-conserving surgery, Ann. Surg. Oncol. 17 (2010) 2696e2701. [27] J.C. Bolger, J.G. Solon, C. Power, Analysis of margin index as a method for predicting residual disease after breast-conserving surgery in a European Cancer Center, Ann. Surg. Oncol. 19 (2012) 207e211. [28] Claire Edwards, F. Gao, G.M. Freedman, Margin index: a useful tool for the Breast Surgeon, J. Surg. Res. 190 (1) (2014) 164e169. [29] J.L. Caughran, F.A. Vicini, L.L. Kestin, Optimal use of re-excision in patients diagnosed with early-stage breast cancer by excisional biopsy treated with breast-conserving therapy, Ann. Surg. Oncol. 16 (2009) 3020e3027.
Please cite this article in press as: M. Mimouni, et al., Reexcision for positive margins in breast cancer: A predictive score of residual disease, Surgical Oncology (2015), http://dx.doi.org/10.1016/j.suronc.2015.08.003