Radiotherapy and Oncology, 25 (1992) 176-180 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0167-8140/92/$05.00

176 RADION 01068

Pathological features predictive of local recurrence after management by conservation of invasive breast cancer: importance of non-invasive carcinoma D. A. P a t e r s o n a, T. J. A n d e r s o n a, W. J. L. J a c k b, G. R. K e r r b, A. R o d g e r b and U. Chetty c a The Medical School, Teviot Place, Edinburgh, bDepartment at Radiation Oncology, Western General Hospital, Edinburgh and CDepartment of Surgery, Royal Infirmary of Edinburgh, Edinburgh, UK

(Received 7 October 1991, revision received 20 May 1992, accepted 2 July 1992)

Key words: Breast cancer; Breast conservation; Radiotherapy

Summary The pathological features of 236 clinical stage I and II invasive breast carcinomas treated by conservation were reviewed. On follow-up (minimum 2 years) 13 patients (6 ~o) have developed breast relapse, 10 (4 ~o) regional lymph node relapse and 26 (11 ~o)

distant metastases. Nineteen patients have died from breast carcinoma. On univariate analyses lymph node metastases, increasing amounts of non-invasive carcinoma and multiple foci of invasion were significant risk factors for breast relapse. A Cox's multivariate analysis showed the first two of these to be independently significant. The results are in agreement with other published series and confirm that assessment of non-invasive carcinoma is important. The study compares simple quantitation with the original method reported to define cases with an "extensive intraduct component".

Introduction

Conservation therapy for invasive breast cancer entailing local excision and radiotherapy is increasingly being offered to patients as an alternative to mastectomy. Several studies have analysed pathological features to identify patients at greater risk of local recurrence for whom conservation therapy may be unsuitable [3,1114,16,21]. Features identified as risk factors have included tumour size [ 13,21], histological grade [ 1113,16], lymphatic vascular invasion [13,21], lymph node metastasis [11-13], completeness of excision [16,21] and the amount of non-invasive carcinoma [3,12,13,16,21]. This latter feature has been assessed by several methods of which the Boston group's criteria for defining an "extensive intraduct component", EIC, are the best established [ 16]. This paper presents a retrospective study of patients with invasive breast carcinomas treated by conservation; its aim was to identify pathological features asso-

ciated with an increased risk of local breast recurrence. The amount of non-invasive carcinoma was assessed by both the Boston group's criteria and by a simple semi-quantitative method. Patients and methods

Patients with clinical stage I and II invasive breast carcinomas (UICC 1978) selected for conservation at the Longmore and Western General Hospitals, Edinburgh, between October 1981 and December 1986 were studied. Surgery entailed wide local excision of the gross tumour mass. The majority of patients received axillary surgery, either a standard axillary clearance or a sample of axillary nodes [8]. All patients after surgery received post-operative radiotherapy to the whole breast (45 Gy in 20 fractions over 4 weeks), the majority also receiving a boost dose by electron beam therapy (15 Gy) or interstitial iridium implant (30 Gy) to the tumour bed. Patients also un-

Address for correspondence: Dr D.A. Paterson, Lecturer in Pathology, The Medical School, Teviot Place, Edinburgh, EH8 9AG, UK.

177 derwent irradiation & t h e regional lymphatics including the ipsilateral axilla and supraclavicular region. A number of patients had additional adjuvant systemic therapy in the form of either tamoxifen or chemotherapy (cyclophosphamide, methotrexate, 5-fluorouracil - CMF) or bilateral oophorectomy. Pathological assessment The histological reports and all slides from each case were reviewed. The following pathological and histological features were assessed. Tumour size This was measured in the freshly dissected specimen, and grouped in accordance with the TNM atlas [19]. Histological type Histological type was based on a qualitative evaluation of established histological patterns [6,15 ]. They were classified into three categories [2]: (1)classical special types (ST) where 90~o or more of the tumour was of a designated pattern; (2) variant of special type (VST) with more than 50j% but less than 90~o showing a designated pattern or more than one designated pattern; and (3)not of special type (NST), those that lacked the features of the other two groups and corresponded to the category of not otherwise specified (NOS) identified by Fisher et al. [6]. The nature of the "main lesion'" Further classification of carcinomas into simple qualitative patterns of "main lesion" was undertaken to assess focality of invasion and the presence of noninvasive carcinoma within the invasive focus. More than one focus of invasion was held to be present when 1 cm of microscopically normal breast tissue was present between separate foci of invasive carcinoma. Non-invasive carcinoma was either present or absent from the invasive focus• This yielded four patterns of "main lesion", namely (1) a single invasive focus without non-invasive carcinoma, (2) multiple invasive foci without non-invasive carcinoma, (3)a single invasive focus with non-invasive carcinoma within the focus, and (4)multiple invasive loci with non-invasive carcinoma within the foci. Semi-quantitative assessment of non-invasive carcinoma in the adjacent parenchyma The adjacent parenchyma was defined as tissue lying outside a line drawn around the limits of the invasive carcinoma making up the "main lesion"• The number of units involved by non-invasive carcinoma was re-

corded and the characteristics of the adjacent parenchyma classified into three groups: (a) minimal involved units (with no non-invasive carcinoma or only a single unit involved); (b) multiple involved units (2-10); (c) widespread involvement of units ( > 10). Extensive intraduct component (EIC) Carcinomas were also assessed for the presence of EIC [4,7,16]. According to published criteria EIC was present when non-invasive carcinoma occurred as both 25 ~o or more of the tumour and was also present in the surrounding breast tissue beyond the limits of the invasive carcinoma. Carcinomas that were essentially non-invasive with foci of invasion were also considered EIC positive• Lymphatic infiltration and lymph node metastases Lymphatic vascular infiltration, was assessed as definitely present, possible, or absent• Where axillary lymph nodes were examined metastases were recorded as either present or absent and where present the number of positive lymph nodes was recorded. Completeness of excision Excision was considered complete when carcinoma, invasive or non-invasive, did not extend to the surgical resection margins. In the first 2 years of the study no specimen trimming protocol was followed, and retrospective assessment of completeness of excision was impossible• In the later 3 years margins were assessed in the majority of cases by a standard method that included inking of the resection margins [1 ]. Follow-up and statistics After completion of treatment patients were reviewed at 3-monthly intervals for 3 years and then 6-monthly. Recurrences were divided into three groups: (1)local, within the ipsilateral breast or breast skin; (2)nodal, within the regional lymph nodes; or (3) metastatic. Each pathological variable recorded was assessed by univariate and multivariate analysis for an association with an increased risk of local breast relapse. Multivariate analysis was performed by the Cox's proportional hazards model [5].

Results

Patients, treatments, follow-up and recurrence Two hundred and thirty-six patients were included in the study, 188 had an axillary node sample, 36 an axillary clearance and in 12 cases, no axillary surgery was

178 undertaken. Eighty-six patients received adjuvant systemic therapy. The duration of follow-up ranges from 2 to 7 years with a mean of 3.7 years. Since treatment, 34 patients (14 ~o) have developed recurrent carcinoma at any site. Thirteen patients in total have developed local breast relapse, including 9 with associated lymph node or distant metastases. The remaining 21 recurrences were as distant metastases, regional lymph node relapses or both. The overall percentage relapse rates were 6 % for local breast relapse, 4 ~ for regional lymph node relapse and 11 ~o for distant metastases. Twenty-six patients have died, 7 from intercurrent disease, the remaining 19 from distant metastases or locally advanced disease.

Pathological features and statistical analysis The pathological features and the distribution of local relapses between the groups together with the statistical analysis is summarised in Table I. The mean volume of excised breast tissue was 107.8 cm 3 (range 3.1686.0cm3). The mean volume of the excised carcinomas was 8.3 cm 3 (range 0.1-64.0 cm3). An average of 4.5 tissue blocks per case (range 1-25) were reviewed. There was no significant difference in the number of blocks examined between cases classified as having minimal non-invasive carcinoma and those having multiple or widespread non-invasive carcinoma in the adjacent parenchyma (Z2= 6.85, p > 0.5). Univariate analysis, considering local relapse alone, revealed that the following four factors were associated with a significantly increased risk of local relapse: (1) lymph node metastasis; (2)an extensive intraduct component; (3)multiple or widespread foci of noninvasive carcinoma in the adjacent parenchyma; (4) multiple foci of invasion (with or without noninvasive carcinoma). On multivariate analysis lymph node metastases and multiple or widespread noninvasive carcinoma in the adjacent parenchyma were revealed as significant independent factors for local relapse. None of the other variables assessed was found to be a significant risk factor including completeness of excision which could be assessed with confidence in 125 cases.

Discussion This study has confirmed that pathological features within primary excision specimens do predict the outcome of conservation therapy. Four risk factors for local recurrence have been identified namely lymph

node metastases, EIC, amount of non-invasive carcinoma in the adjacent parenchyma and multiple foci of invasion. Of the four risk factors, lymph node metastases proved to be the most significant on both univariate and multivariate analysis. Lymph node metastases have been identified as a risk factor by most series [ 11-13] and these findings highlight the need for adequate sampiing of regional lymph nodes in each case considered for conservation. Non-invasive carcinoma has also been recorded as a local recurrence risk factor by several studies [ 3,12,13,16,21 ]. Two assessments of non-invasive carcinoma are presented in this study. The first was the EIC classification that uses the histological pattern of the resected primary tumour to identify carcinomas with a greater likelihood of extension beyond the surgical excision margins. This concept is supported by studies of re-excision specimens [17] and by serial section of whole breasts [8] which show a greater likelihood of residual disease if EIC positive carcinomas are treated only by local excision. EIC has been shown to be associated with a greater likelihood of involvement of resection margins [3,9] and to be more common in younger women [3,9]. EIC has been particularly identified as a risk factor in studies with a limited excision [3,9,12,13,16,21], yet in this current series with a wide local excision EIC was still a significant risk factor. EIC was diagnosed in 32~o of cases in the initial Boston series [16] but less frequently in this current series (19~o) and in the series by Jaquemier et al. (21 ~/o) [9]. While these differences may reflect variations in the incidence between countries or selection bias it is possible they may be related to the reproducibility of the EIC classification. The second assessment of non-invasive carcinoma involved simple quantitation of units containing noninvasive ductal carcinoma in the surrounding parenchyma. This method is reproducible, avoids subjective assessments of percentage and in the present series, is not related to the number of blocks examined. Both the EIC classification and the quantitation of involved units identify a similar group of carcinomas (93 ~o of EIC positive carcinomas had multiple or widespread non-invasive carcinoma in the adjacent parenchyma). Although patients with EIC or 2 or more foci of non-invasive carcinoma in the adjacent parenchyma have a significantly greater risk of local relapse, the positive predictive value of these assessments is low, Whilst assessment of non-invasive carcinoma alone does not account for all the cases of local relapse, an evaluation of the extent within conservation specimens is clearly important. An improvement in predictive value

179 TABLE I Pathological feature

Total number n = 236

Local breast relapse n = 13

Tumour size Tta Tlb Tic T2

5 50 124 57

0 2 7 4

Tumour type Special type Variant of special type Not of special type

25 28 183

0 1 12

84 5

4 1

Nature of main lesion (1) Single invasive focus only (2) Multiple invasive foci only (3) Single invasive focus with non-invasive within (4) Multiple invasive loci with non-invasive within

135 12

Adjacent parenchyma Minimal non-invasive carcinoma Multiple loci of non-invasive carcinoma (2-10)

154 58

Widespread loci of non-invasive carcinoma (> 10)

Univariate analysis local breast relapse a

,o°/ Free of local breast relapse at 48 months for significant variables (= 95% confidence intervals)

NS

NS

NS

NS

Significant p < 0.005

98 (97-99) 67 (40-94) 97 (96-98) 81 (76-86)

Significant p < 0.005 b

24

99 (98.5-99.5) 88 (86-90) 88 (84-92)

EIC No Yes

190 46

Lymphatic infiltration Absent Possible Present

191 5 40

Lymph node metastases Absent Present Not sampled

169 55 12

5 8 0

No. of involved nodes None 1-3 >3 Not sampled

169 44 11 12

5 4 4 0

Completeness o] excision Complete Incomplete Not reported

95 30 111

7 6

Significant p

Pathological features predictive of local recurrence after management by conservation of invasive breast cancer: importance of non-invasive carcinoma.

The pathological features of 236 clinical stage I and II invasive breast carcinomas treated by conservation were reviewed. On follow-up (minimum 2 yea...
458KB Sizes 0 Downloads 0 Views