A multivariate analysis on prognostic factors for lobular carcinoma of the breast

ABSTRACT CONTEXT AND OBJECTIVE: Lobular carcinoma is the second most common type of breast neoplasia and has unique clinical and pathological features. Our aim was to evaluate prognostic factors for this type of breast cancer. DESIGN AND SETTING: Retrospective study at a tertiary oncological institution. METHODS: 162 patients diagnosed and treated between January 1985 and January 2002 were included. The inclusion criteria were: absence of previous treatment, histological diagnosis of lobular carcinoma, no previous history of breast cancer and minimum follow-up of 36 months. RESULTS: In univariate analysis, the following factors were statistically significant: clinical stage T (P = 0.0005), clinical stage N (P = 0.0014), neoadjuvant chemotherapy (P = 0.0008), primary tumor size (P < 0.0001), vascular invasion (P < 0.0001), lymphatic invasion (P = 0.0004), neural invasion (P = 0.0004), skin invasion (P < 0.0001), capsular transposition (P = 0.0008), lymph node ratio (P < 0.0001), estrogen receptor expression (P = 0.0186), progesterone receptor expression (P = 0.0286), pathological stage T (P < 0.0001), pathological stage N (P < 0.0001), adjuvant chemotherapy (P < 0.0001) and postoperative hormone therapy (P = 0.0367). After grouping the variables, multivariate analysis was performed. Presence of lymph node metastases, capsular transposition, lymph node ratio and postoperative hormone therapy remained significant. CONCLUSION: In this series, the most important prognostic factors for lobular carcinoma of the breast seemed to relate to lymph node status and presence of capsular transposition. Factors relating to axillary involvement, capsular transposition and hormone therapy were significant for survival.


INTRODUCTION
Invasive lobular carcinoma (ILC) is the second most common type of breast cancer after invasive ductal carcinoma (IDC), accounting for 5-15% of all breast cancer cases. 1 Its incidence rates increased from 1987 to 1999, and predominantly in postmenopausal women, in contrast to IDC rates, which remained largely constant throughout this period.The proportion of breast cancers with a lobular component increased from 9.5% in 1987 to 15.6% in 1999. 2 The morphological features of lobular carcinoma differ from those of ductal carcinoma.ILC is characterized by small, round cells that are bland in appearance and have sparse cytoplasm.These cells infiltrate the stroma in single file and surround benign breast tissues in a targeted manner.Infiltration typically does not destroy anatomical structures or induce a substantial connective tissue response.By virtue of their distinctive growth pattern and biology, lobular carcinomas often fail to form distinct masses that can easily be diagnosed by palpation or mammography. 3,4here are clinical and pathological differences between ILC and IDC.A large retrospective study compared the clinical and biological features of 4,140 patients with ILC with those of 45,169 patients with IDC. 5 Compared with IDC, ILC occurred significantly more frequently in older patients, was larger in size and was frequently more estrogen receptor and progesterone receptor-positive.Moreover, ILC had a lower S-phase fraction and tended to be diploid and HER-2, p53 and epidermal growth factor receptor-negative. [5][6][7] Multivariate analysis has not identified any prognostic differences associated with ILC and IDC.The same standard prognostic factors (tumor size, axillary nodal status, hormone receptors, S-phase and age) are applicable both to lobular and to ductal carcinoma. 5

OBJECTIVE
The purpose of this paper was to report on a multivariate analysis of ILC using a single institution's experience, and to attempt to define prognostic factors for ILC that are not based on comparisons with IDC.

PATIENTS AND METHODS
Patients with a histological diagnosis of ILC of the breast who were treated between January 1985 and January 2002 were selected for this study.All these patients were enrolled at a single private tertiary cancer center.The inclusion criteria were absence of previous treatment, confirmed histology with review at our institution, no previous history of breast cancer and minimum follow-up of 36 months after conclusion of treatment, except for patients who died due to treatment complications or other causes.The histological diagnosis included optical microscopy examination of samples from all the patients, while immunohistochemical studies were indicated at the pathologist's discretion.Patients with distant metastasis at diagnosis were excluded from this series.From these criteria, 162 patients were identified and included in this study.
The patients were staged in accordance with the 2002 TNM classification. 8The following data were collected: age, race, symptoms, previous gynecological history, clinical and pathological staging, treatment, pathological characteristics of the tumor, hormone receptor status of the primary tumor, presence of c-erb-B2, recurrence (if present) and status at last follow-up.
The statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) package, version 11 for MacOS X. P values greater than 0.05 were considered non-significant throughout the study.Initially, a univariate analysis was performed and the factors identified as significant were included in a multivariate analysis.This was performed using chunkwise testing methods with a backward elimination procedure.A set of logically related predictors of equal importance constituted the first chunk and backward elimination was applied.Then, respectively, new factors were added to the chunks and the backward elimination was reapplied, while significant factors were kept throughout the steps. 9Survival curves were constructed and compared using the Kaplan-Meier and Cox methods.

RESULTS
All the patients were female.The age at diagnosis ranged from 28 to 87 years (mean of 55.89 years and median of 55.50 years).For statistical purposes, the patients were divided according to a cutoff point of 45 years: 36 patients (22.2%) were younger than 45 years and 126 (77.8%) were above the cutoff point.According to race, 134 were white and 28 were nonwhite.
In 142 patients (87.7%), clinical symptoms were present at the time of diagnosis, while in 20 (12.3%),only radiological findings were indicative of breast disease.Forty-five patients (27.8%) reported previous use of exogenous hormones as contraceptives (32 patients, 19.8%) or hormone replacement (13 patients, 8.0%).Thirty patients (18.5%) had a family history of breast cancer.
The treatment was based on the protocols used at the time of diagnosis.Preoperative chemotherapy was indicated for 13 patients (8.0%).Mastectomy was performed on 124 patients (76.5%), while breast-conserving surgery was performed on 38 patients (23.5%).In the postoperative setting, chemotherapy was administered for 87 patients (53.7%), radiotherapy for 100 patients (61.7%) and hormone therapy for 74 patients (45.7%).At the time of this study, trastuzumab was not used as routine chemotherapy, and no patients received this drug.
The duration of follow-up ranged from 1.45 to 147.78 months (mean, 75.87 and median, 56.53 months).There were six cases of local recurrence (3.7%), four of regional recurrence (2.5%) and 50 of distant recurrence (30.9%).At the last contact, 105 patients (64.8%) were alive and without evidence of active disease, one patient (0.6%) had active disease, 50 patients (30.9%) had died due to disease progression and six patients (6.7%) had died from other, unrelated causes.
After grouping the variables in steps, multivariate analysis was performed.Table 1 details each step and shows which variables were evaluated and eliminated over the course of the analysis.After the final step, presence of metastatic lymph nodes (Figure 1), capsular transposition (Figure 2), lymph node ratio and hormone therapy (Figure 3) were found to be significant in the multivariate analysis.

DISCUSSION
In this report, we analyzed the prognostic factors for ILC in a manner not performed before in the literature, with multivariate analysis performed through chunkwise testing methods with a backward elimination procedure.It allowed us to diminish the number of variables analyzed at each step.One major limitation of this study was its retrospective nature, but lobular carcinomas of the breast are not as common as ductal carcinomas and prospective studies are difficult to conduct because of the length of time involved in amassing a significant number of patients.
The prognostic factors for ILC have already been assessed in other reports.A previous study analyzed 390 patients and demonstrated that age (P = 0.0039), stage T (T1/2 versus T3, P = 0.0099), lymph node status (N0/1 versus N2, P = 0.0009), and grade (I/II versus III, P = 0.0128) affected the prognosis, in a multivariate model. 10The role of lymph node status, capsular transposition and the size of the metastatic deposits have also been demonstrated in other reports. 11Histological grading has also been shown to be a significant prognostic factor.In a report from the Danish Breast Cancer Cooperative Group, patients with grade III tumors showed a higher proportion of non-classical subtypes and significantly worse prognosis for overall and disease-free survival, in comparison with patients with grade I/II carcinomas. 12In another report, histological grading was correlated with other factors for poor prognosis, such as increasing tumor size, positive lymph nodes, vascular invasion and estrogen receptor negativity.Multivariate analysis has shown that histological grading is a significant factor indicating shorter disease-specific and disease-free survival. 13The role of bilateral disease as a prognostic factor has also been assessed.In a previous report, it was shown that patients with synchronous, bilateral ILC had a significant worse prognosis than did patients with unilateral or metachronous bilateral disease (P < 0.02). 14nformation on prognostic factors for ILC may be also found in reports that had the original intent of comparing this histological type with IDC.In a report originally intended to compare the prognostic significance of lobular histology, the following factors were found to be significantly linked to overall survival in cases of ILC: node status, age, primary tumor size, ER status and S-phase. 5Another study also compared the prognosis using histological analysis and concluded that for ILC, the following factors were significant in multivariate analysis: age at diagnosis, tumor size, lymph node status and pathological stage.In another report analyzing the prognostic factors for 217 patients with ILC, the following factors were found to be significant: tumor size (P < 0.0001), axillary nodal metastasis (P < 0.0001), absence of tumor necrosis (P < 0.0001), low mitotic count (P < 0.0001), low histological grading (P = 0.001) and S-phase fraction lower than the median (P = 0.005). 15ge at diagnosis, tumor size, pathological stage and lymph node status were shown to be independent prognostic indicators for 10-year survival, with no significant prognostic difference between ILC and IDC. 16he lymph node ratio has been shown to be an important predictor for survival among patients with stage I and II breast cancer. 17n our analysis, the number of factors found to be significant was smaller than in these other series.Age at diagnosis was found to be nonsignificant at different cutoff points and as a continuous variable, and tumor size was only significant in the univariate analysis, and thus was excluded from the multivariate analysis.No significance was found in the multivariate analysis for ER status or histological grading, which were significant in previous reports. 5On the other hand, postoperative hormone therapy was significant for survival.

CONCLUSIONS
From the data in this report, we conclude that the most important prognostic factors for lobular carcinoma of the breast seem to relate to lymph node status and presence of capsular transposition.The presence of estrogen receptors and the inclusion of postoperative hormone therapy was also associated with a significant improvement in survival.However, further prospective studies would be needed to confirm this evidence.

Figure 3 .
Figure 3. Overall survival rates for patients with and without postoperative hormone therapy.

Figure 2 .
Figure 2. Overall survival rates for patients with extracapsular spread.

Figure 1 .
Figure 1.Overall survival rates for patients with and without metastatic axillary nodes.pN = pathological stage N.

Table 1 .
Multivariate analysis on prognostic factors