Predictors of axillary lymph node metastases in patients with T1 breast carcinoma†
Presented at the Thirty-Second Annual Meeting of the American Society of Clinical Oncology, Philadelphia, Pennsylvania, May 18-21, 1996.
Axillary lymph node metastases (ALNM) are the most important predictor of survival in patients with T1 breast carcinoma. Due to a relatively low incidence of axillary metastasis in tumors ≤ 2 cm, the role of axillary lymph node dissection for these patients has been questioned. The purpose of this study was to determine the association between the incidence of ALNM and 11 clinical/pathologic factors by univariate and multivariate analysis.
The authors reviewed data from 918 patients with T1 breast carcinoma who underwent level I/II axillary dissection between 1979 and July 1995. The association between the incidence of ALNM and 11 clinical/ pathologic factors (size, lymph/vascular invasion, nuclear grade, S-phase, ploidy, palpability, age, estrogen receptor status, progesterone receptor status, HER-2/neu, and histology) was analyzed by univariate and, when significant, by multivariate analysis.
Approximately 23% of the 918 patients with T1 breast carcinoma had ALNM. Multivariate analysis identified four factors as independent predictors of ALNM: lymph/vascular invasion (P < 0.0001), tumor palpability (P < 0.0001), nuclear grade (P = 0.0004), and tumor size (P = 0.01). Among the 117 patients with nonpalpable, nonhigh grade tumors ≤ 1 cm without lymph/vascular invasion, the incidence of ALNM was only 3%. However, the 43 patients with T1c tumors with all 3 additional risk factors had a 49% incidence of ALNM.
Clinical and pathologic features of the primary tumor can be used to estimate the risk of ALNM in patients with T1 breast carcinoma. Such a risk assessment might facilitate appropriate management. Routine axillary dissection can be omitted in patients at minimal risk of ALNM, if the treatment decision is not influenced by lymph node status. Axillary lymph node dissection should be performed routinely for all patients with lesions > 1 cm. [See editorial counterpoint on pages 1856-61 and reply to counterpoint on pages 1862-4, this issue.] Cancer 1997; 79:1918-22. © 1997 American Cancer Society.
The incidence of axillary lymph node metastases (ALNM) increases as a function of primary tumor size.1 For patients with T1 tumors, the reported incidence of ALNM is 6-31%.1-6 The presence of ALNM and tumor size are the two most important prognostic factors for women with breast carcinoma.7-9 Lymph node status and pathologic features of the primary tumor are used to estimate an individual patient's prognosis and guide the selection of adjuvant therapy. Due to the relatively low incidence of ALNM in patients with small tumors, the potential morbidity and cost of the procedure, and the lack of a significant therapeutic value, the role of routine axillary staging for patients with early breast carcinoma recently has been questioned.1, 10-12 Moreover, it has been found that axillary lymph node dissection (ALND) did not influence the decision for adjuvant treatment in patients with unfavorable biopsies13 as well as older patients with T1 breast carcinomas and favorable pathologic features.4, 7, 14 Others advocate a new staging procedure (i.e., sentinel lymphadenectomy) to avoid a full ALND in the majority of T1 breast carcinoma patients.11, 15, 16
Estimation of the risk of ALNM based on clinical features and pathologic characteristics of the primary tumor is important to select the appropriate management of the axilla. However, thus far, only a few studies have attempted to identify independent factors predisposing to ALNM.
PATIENTS AND METHODS
All patients with T1 lesions initially treated at the Breast Center in Van Nuys, California from 1979 through July 1995 who underwent ALND as part of their treatment for breast carcinoma were included in this series. Tumors were categorized using the TNM system of the American Joint Committee on Cancer. Invasive carcinomas were classified using the largest dimension of the invasive component to determine size: ≤ 5 mm or less, T1a; 6-10 mm, T1b; and 11-20 mm, T1c. The lesions were measured macroscopically to the nearest mm and/or by direct measurement from the microscopic slides and by determining the number of serial sections in which the tumor, sectioned at 2-3 mm intervals, appeared. The axillary contents were dissected fresh; all identified lymph nodes were sectioned through the hilum and examined histologically. Patients were treated with breast-conserving surgery and radiation therapy (45-50 gray [gy]) whenever possible (62%; 530/918), or mastectomy (37%; 388/918), generally with immediate reconstruction. A level I/II ALND was performed for all patients. Clinical assessment of the axilla was not always recorded in the database; however, the vast majority of patients had clinically negative lymph nodes.
Pathologic characteristics of the primary tumor (size, histologic type, nuclear grade, and lymphatic/vascular invasion) were evaluated on routine hematoxylin and eosin slides. S-phase and ploidy were assessed by flow cytometry. Estrogen receptor (ER)and progesterone receptor status and HER-2/neu were determined by radioimmunoassay or immunohistochemistry.
The log rank test17 was used to perform the univariate analyses. Multivariate analysis was based on the Cox proportional hazards regression model and included any variable found to be significant in the univariate analysis.18
The characteristics of the 918 patients with T1 carcinoma of the breast treated at The Breast Center in Van Nuys between 1979 and July 1995 are summarized in Table 1. The majority (71%) had palpable tumors. The median tumor size was 13 mm. Ninety-two tumors (10%) were T1a tumors, 245 (27%) were T1b, and 581 (63%) were T1c. The majority of tumors (88%) were infiltrating ductal carcinoma, and 12% were of the infiltrating lobular type. An average of 18 axillary lymph nodes per patient were removed. Two hundred and eight patients (23%) had ALNM.
Table 1. Patients with T1 Breast Carcinoma (n = 918) Treated at The Breast Center between 1979 and July 1995
| ||Inf lobular||106||12%|
On univariate analysis tumor size, tumor palpability, lymph/vascular invasion, and nuclear grade were identified as highly significant predictors of ALNM (P < 0.0001) (Table 2). Absence of ER showed only a borderline association with ALNM (P = 0.069).
Table 2. Association between Incidence of Axillary Lymph Node Metastases and 11 Clinical/Pathologic Factors by Univariate and Multivariate Analysis
| ||Inf lobular||106||18%|
By multivariate analysis, using the Cox proportional hazards model, the four factors identified by univariate analysis proved to be independent predictors of ALNM. The presence of lymph/vascular invasion was the strongest predictor of ALNM (P < 0.0001).
Among the 117 patients with nonpalpable, nonhigh grade tumors ≤ 1 cm without lymph/vascular invasion, the incidence of ALNM was only 3%. However, the 43 patients with T1c tumors and all 3 additional risk factors had an incidence of ALNM of 49%.
Although the surgical approach to the breast in patients with breast carcinoma has become less aggressive over the years, routine ALND continues to be performed for most patients. Because lymph node status and tumor size are the two most important prognostic factors for patients with carcinoma of the breast,7-9 ALND is useful in determining prognosis and the need for adjuvant therapy. Furthermore, it lowers the risk of axillary lymph node recurrence.19, 20 However, due to the relatively low incidence of ALNM in patients with T1 breast carcinoma, the potential benefits of this procedure must be weighed against the well known potential morbidity21-23 (arm edema, nerve paresthesias and damage, and prolonged recovery time) and economic factors10 (operating room, and hospital expense due to prolonged recovery time). This has led some to question the role of routine ALND for the treatment of patients with early breast carcinoma.1, 10-12 Others advocate a new staging procedure (i.e., sentinel lymphadenectomy) avoiding full ALND in the majority of T1 breast carcinoma patients.11, 15, 16 In this context of preoperative risk assessment, predictive factors of ALNM become increasingly important, yet only a few studies have attempted to identify independent factors predisposing to ALNM.4, 24-26
Approximately 23% of the 918 patients in this study had axillary lymph node involvement. The current study data and reports from other centers2, 5, 24 show that the incidence of ALNM increases with tumor size or, more specifically, with the invasive component of the tumor.27 However, lymph/vascular invasion proved to be the most important predictor of ALNM in this study. Patients with lymph/vascular invasion had lymph node metastases in 46% of axillary specimens, compared with only 19% when this feature was absent. In a recently published report, lymph/vascular invasion also was the strongest independent predictor of ALNM,4 underscoring the importance of this tumor characteristic.28-30
In contrast to other published series, tumor palpability and nuclear grade were also identified as independent factors in the current study. Because the presence of a single independent risk factor did not necessarily help to guide further treatment, combinations of risk factors were also studied. Breast carcinoma patients who had a combination of all 4 risk factors had an incidence of ALNM of 49%. However, more important is the finding that 13% (117 patients) in the current series had nonpalpable, nonhigh grade tumors 1 cm without lymph/vascular invasion, and the incidence of ALNM was only 3%. Routine ALND can be omitted in patients at minimal risk (≤ 5%) of ALNM, especially if the decision for adjuvant treatment is not altered by the result of ALND. This might be the case in postmenopausal women with ER positive T1 breast carcinoma who are widely treated with tamoxifen, independent of their lymph node status.14, 31, 32 Alternatively, a sentinel lymph node biopsy, a minimally invasive procedure, could be performed, which allows accurate assessment of lymph node status.11, 15, 16 Confirmatory studies with this technique will establish its definitive role in the management of the axilla.
To the authors' knowledge, this is the largest series of T1 breast carcinoma patients in which clinical and pathologic characteristics of the primary lesion have been correlated with ALNM by multivariate analysis. Four independent predictors of lymph node metastases were identified. These factors can help to estimate the risk of ALNM in patients with T1 breast carcinoma. Such a risk assessment might facilitate appropriate management of the axilla.