Over the past 20 years screening mammography has become utilized increasingly and the quality of mammography has improved progressively.1, 2 The proportion of invasive breast carcinomas measuring < 1 cm was 5.4% of all cases identified in the National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) Program between 1977 and 1982.3 In the subsequent 5-year time period (1983-1987), the proportion of carcinomas measuring ≤ 1 cm increased to 14.4%.4 More recently, Cady et al. reported that breast tumors measuring ≤ 1.0 cm accounted for 28% and 29%, respectively, of invasive breast malignancies diagnosed at two New England hospitals between 1989 and 1993.5 It is anticipated that the proportion of small invasive breast carcinomas will continue to rise, and the median greatest dimension of invasive breast carcinoma cases will be 1.0 cm in the near future.6
In women with invasive breast carcinomas measuring ≤ 1.0 cm, the detection of axillary lymph node metastases is the most important factor in determining whether adjuvant systemic therapy is recommended.7, 8 The reported frequency of axillary lymph node metastases associated with invasive breast malignancies measuring ≤ 1.0 cm ranges from 0-57% for T1a breast carcinomas and 7-37% for T1b breast carcinomas.3, 4, 9–15 In several recent series reporting a low frequency of axillary lymph node metastases, the authors have suggested that routine axillary lymph node dissection may be avoided in selected subsets of patients with small invasive breast carcinomas.5, 16 The avoidance of axillary lymph node dissections in selected patients has potential benefits, including significant health care savings and the prevention of morbidity associated with this procedure. The desire to perform axillary lymph node dissections selectively is balanced by the necessity of identifying those patients with lymph node metastases who would benefit from adjuvant systemic therapy.
As the frequency of small invasive breast carcinoma increases, the need to establish optimal clinical management guidelines is becoming increasingly important. The purpose of this study was to examine the effect of common clinical and histologic factors on the frequency of lymph node metastases associated with small invasive breast carcinomas.
- Top of page
Cases for the study cohort were identified from data collected by nine population-based cancer registries that are part of the SEER Program of the NCI. The SEER Program collects data on cancer incidence from registries in various areas, including the states of Connecticut, Iowa, New Mexico, Utah, and Hawaii and the metropolitan areas of Detroit, San Francisco, Seattle-Puget Sound, and Atlanta.
The SEER registries reported a total of 105,234 in situ and invasive female breast carcinoma cases during the inclusive period of January 1, 1988 to December 31, 1993. These cases included 26,314 cases in which the tumors were ≤ 1 cm in size. The following cases were excluded from the analysis: 6474 cases diagnosed with in situ lesions, 20 cases that were not microscopically confirmed, 10 noncarcinoma cases, 6 inflammatory carcinomas, 523 cases with extension upgrading the primary lesion to a pathologic T4 classification, 2931 cases in which no lymph nodes were examined, 3111 cases in which 1-9 lymph nodes were examined, and 289 cases in which the number of lymph nodes examined was unknown. After these exclusions, the study cohort was comprised of 12,950 invasive breast carcinoma cases.
The affect of clinical and histologic factors including primary tumor size, histologic type, histologic grade, tumor location, and patient age on the frequency of axillary lymph node metastases was reviewed. The recorded tumor size was determined utilizing abstracting and coding guidelines for cancer registrars. Tumor size was determined in priority order from the pathology report, surgical report, physical examination, and mammographic examination. In cases with tumor in more than one specimen the composite size was recorded only if indicated by a pathologist. Tumor size was not recorded in cases treated with preoperative radiation or systemic therapy. Tumors recorded as having a microscopic focus or foci of malignancy or those measuring ≤ 0.5 cm were classified as T1a malignancies. Tumors measuring 0.6-1.0 cm were classified as T1b malignancies.17 Histologic diagnoses were recorded in the SEER registry utilizing the International Classification of Diseases-Oncology (ICD-O) classification system. Histologic grading of tumors was performed with “well,” “moderately,” and “poorly” differentiated tumors approximating to Grades 1, 2, and 3. Undifferentiated or anaplastic tumors corresponded to a histologic Grade 4. Tumors located in the upper inner and lower inner quadrants were grouped as the medial breast, and tumors located in the upper outer and lower outer quadrants were grouped as the lateral breast. Tumors located in the subareolar region were recorded as the central breast. Cases were grouped by age as 0-39 years, 40-49 years, 50-59 years, 60-69 years, and ≥ 70 years.
Comparisons of the frequency of axillary lymph node metastases between different subgroups of the cohort were evaluated with a chi-square test. All reported P values were two-sided. In analyses examining grade and tumor size, the Cochran-Armitage test was used to test for trend. Cases with unknown grade were not included in analyses evaluating tumor grade.
Logistic regression was used to test for the significance of clinical and pathologic variables on lymph node metastases. All analysis was performed using SAS statistical software (SAS Institute, Inc., Cary, NC).
- Top of page
The relation between the number of axillary lymph nodes examined and the frequency of lymph node metastases is shown in Table 1. Overall, the frequency of axillary lymph node metastases increased 24.7% for T1a carcinomas and 20.2% for T1b carcinomas when cases with 1-5 examined lymph nodes were compared with cases in which ≥ 10 lymph nodes were examined. The proportion of cases with detected lymph node metastases varied significantly based on the number of lymph nodes examined (P = 0.004). Cases in which fewer than ten lymph nodes were examined histologically were excluded from the remaining analyses.
Table 1. Proportion of Cases with Axillary Lymph Node Metastases as a Function of the Number of Examined Lymph Nodes
|Size (cm)||Proportion with positive lymph nodes (%)|
|1–5 (n = 759)||6–9 (n = 2352)||≥10 (n = 12,950)|
The relation between tumor size and the frequency of lymph node metastases is shown in Table 2. Over the size range of ≤ 0.3-1.0 cm, the proportion of carcinomas with any lymph node metastases gradually increased from 7.4% to 17.4%. In cases involving a single microscopic focus or multiple foci of malignancy, the proportion of carcinomas associated with any lymph node metastases was 10.6%. The frequency of lymph node metastases associated with T1a tumors was less than that of T1b tumors (9.6% vs. 14.3%; P < 0.001). As a proportion of all cases, those associated with four or more lymph node metastases progressively increased with increasing primary tumor size.
Table 2. Proportion of Cases with Axillary Lymph Node Metastases as a Function of Tumor Size
|Lymph node metastases||Micro||Tumor size (cm)||P value|
The relation between histology and the frequency of lymph node metastases is shown in Table 3. The frequency of axillary lymph node metastases associated with T1a infiltrating ductal carcinomas and infiltrating lobular carcinomas was 9.9% and 14.8%, respectively. The difference in the frequency of lymph node metastases is attributable in large part to differences among the subgroup of microscopically measured tumors. Among T1a carcinomas, microscopic foci of malignancy were associated more frequently with infiltrating lobular carcinoma than with infiltrating ductal carcinoma (38.2% vs. 29.1%). In cases with microscopic foci of malignancy the frequency of lymph node metastases was 18.7% for infiltrating lobular carcinoma and 11.6% for infiltrating ductal carcinoma. In T1b primary tumors the frequency of associated lymph node metastases was 15.1% and 15.2% for infiltrating ductal carcinoma and infiltrating lobular carcinoma, respectively. Tumors with favorable histology (mucinous, papillary, and tubular carcinomas) had a lower frequency of lymph node metastases than all other histologic types (3.9% vs. 13.9%; P < 0.001).
Table 3. Proportion of Cases with Lymph Node Metastases as a Function of Tumor Size and Histology
|Size, cm||Histology||P value|
|≤0.5||2189 (9.9%)||196 (14.8%)||11 (18.2%)||58 (0.0%)||24 (0.0%)||119 (5.0%)||251 (7.6%)||67 (9.0%)||—|
|0.6–1.0||8137 (15.1%)||625 (15.2%)||73 (17.8%)||256 (3.5%)||74 (6.8%)||289 (4.2%)||419 (14.1%)||162 (11.1%)||—|
|Total||10326 (14.0%)||821 (15.1%)||84 (17.9%)||314 (2.9%)||98 (5.1%)||408 (4.4%)||670 (11.6%)||229 (10.5%)||< 0.001|
The frequency of lymph node metastases was related to the histologic grade as shown in Table 4. A large proportion of tumors (45.2%) included in the SEER registry were ungraded. When T1a and T1b tumors were compared the proportion of Grade 1 tumors decreased with increasing tumor size (P < 0.001) and the proportion of Grade 2 and Grade 3 tumors increased with increasing tumor size (P < 0.001). The proportion of Grade 4 tumors was not significantly different over the size range studied (P = 0.208). Increasing histologic grade was associated with an increased risk of lymph node metastases ranging from 7.8% in Grade 1 tumors to 21.0% in Grade 4 tumors (P < 0.001).
Table 4. Proportion of Cases with Axillary Lymph Node Metastases as a Function of Tumor Size and Histologic Grade
|Size (cm)||Histologic grade||P value|
|≤0.5||365 (3.0%)||587 (9.4%)||289 (15.2%)||47 (12.8%)||1627 (10.0%)||—|
|0.6–1.0||1349 (9.0%)||2759 (14.6%)||1518 (20.0%)||182 (23.1%)||4227 (13.3%)||—|
|Total||1714 (7.8%)||3346 (13.7%)||1807 (19.3%)||229 (21.0%)||5854 (12.4%)||< 0.001|
The frequency of lymph node metastases varied with the breast primary location as shown in Table 5. Carcinomas of the medial and central breast had a frequency of lymph node metastases that was lower than that in carcinomas of the lateral breast. In cases in which the primary tumor was located in overlapping or in unspecified regions, the frequency of lymph node metastases was similar to the frequency of lymph node metastases in the breast as a whole. Variation in the frequency of lymph node metastases among the primary tumor locations in the breast (medial, central, lateral) was significant (P < 0.001).
Table 5. Proportion of Cases with Axillary Lymph Node Metastases as a Function of Tumor Size and Location
|Size (cm) (%)||Primary site||P value|
|Upper outer quadrant||Lower outer quadrant||Central||Upper inner quadrant||Lower inner quadrant||Othera|
|≤0.5||1171 (10.8%)||144 (8.3%)||183 (7.1%)||239 (6.3%)||122 (9.8%)||1056 (9.5%)||—|
|0.6–1.0||4056 (16.0%)||579 (17.6%)||605 (14.4%)||963 (9.9%)||615 (12.2%)||3217 (13.3%)||—|
|Total||5227 (14.8%)||723 (15.8%)||788 (12.7%)||1202 (9.2%)||737 (11.8%)||4273 (12.4%)||< 0.001|
The frequency of associated lymph node metastases progressively decreased with increasing age as shown in Table 6 (P < 0.001). Despite a lower frequency of lymph node metastases with increasing age, the frequency of axillary lymph node metastases associated with T1a and T1b tumors was 7.0% and 11.0%, respectively, in women age ≥ 70 years. The maximal risk of axillary lymph node metastases (25.9%) was detected in women age < 40 years with T1b tumors.
Table 6. Proportion of Cases with Axillary Lymph Node Metastases as a Function of Tumor Size and Age
|Size (cm)||Age (yrs)||P value|
|≤0.5||157 (13.4%)||493 (12.4%)||638 (9.2%)||841 (9.9%)||786 (7.0%)||—|
|0.6–1.0||444 (25.9%)||1449 (17.3%)||1974 (15.7%)||3122 (13.5%)||3046 (11.0%)||—|
|Total||601 (22.6%)||1942 (16.3%)||2612 (14.1%)||3963 (12.7%)||3832 (10.2%)||< 0.001|
All variables that were found to be statistically significant based on the univariate analyses were included in a multivariate modeling of the data. Variables were entered into the model in a stepwise building process and interactions were evaluated at each stage of the modeling. None of the interactions evaluated was statistically significant, so the final model included only main effects. In the final model (Table 7 ), tumor size, histologic grade, location, and age at diagnosis all remained as significant independent predictors of the frequency of lymph node metastases.
Table 7. Association Between the Frequency of Axillary Lymph Node Metastases and Clinical/Histologic Characteristicsa
|Variable||Odds ratio||P value|
|Tumor size, T1b vs. T1a||1.78||< 0.0001|
|Histology, unfavorable vs. favorableb||2.35||< 0.0012|
|Grade, 2 vs. 1||1.70||< 0.0001|
|Grade, 3 vs. 1||2.38||< 0.0001|
|Grade, 4 vs. 1||2.74||< 0.0001|
|Location, central vs. medial||1.61||< 0.0056|
|Location, lateral vs. medial||1.83||< 0.0001|
|Location, other vs. medial||1.32||< 0.0205|
|Age, 0–39 yrs vs. ≥ 70 yrs||2.56||< 0.0001|
|Age, 40–49 yrs vs. ≥ 70 yrs||1.71||< 0.0001|
|Age, 50–59 yrs vs. ≥ 70 yrs||1.40||< 0.0014|
|Age, 60–69 yrs vs. ≥ 70 yrs||1.24||< 0.0254|
Carcinomas with an associated frequency of axillary lymph node metastases ≤ 5% were T1a and T1b mucinous and tubular carcinomas, T1a papillary carcinomas, and T1a Grade 1 carcinomas. Invasive breast carcinomas measuring ≤ 1 cm with clinical and histologic features associated with a frequency of lymph node metastases ≤ 5% accounted for 8.2% of the case cohort.
- Top of page
In cases of invasive breast carcinoma both the frequency of any axillary lymph node metastases and the number of lymph nodes detected with metastatic disease increase as the number of examined lymph nodes increases, up to a plateau at ten examined lymph nodes.18–20 Based on these observations, the evaluation of a minimum of ten lymph nodes was chosen for this study. The exclusion of cases in which fewer than ten axillary lymph nodes were examined may indicate to some that there was a selection bias to include only cases with unfavorable primary tumor characteristics or cases with clinically palpable axillary lymph nodes. Under these circumstances the frequency of axillary lymph node metastases would be elevated artificially.
The influence of the number of lymph nodes examined on the frequency of lymph node metastases was explored separately by the evaluation of cases in which one to five and six to nine lymph nodes were examined. The frequency of lymph node metastases associated with T1a and T1b carcinomas, in cases for which at least 1 lymph node was histologically examined, was 9.5% and 13.9%, respectively. These frequencies did not differ markedly from the frequencies of lymph node metastases associated with cases in which ten or more lymph nodes were examined. Based on the low proportion of cases with one to nine examined lymph nodes and the frequency of lymph node metastases in this group, the data from this study suggest that there was not a significant selection bias in the performance of an axillary lymph node dissection rather than an axillary lymph node sampling in cases with unfavorable characteristics. We conclude that the proportion of lymph node metastases noted in this large patient cohort with ten or more histologically examined lymph nodes represents the frequency of lymph node metastases when an adequate axillary lymph node dissection is performed.
This study has three limitations related to a lack of central histologic review. First, the method of tumor size determination was not recorded. As a result the proportion of lymph node metastases associated with the most precise measurement, microscopic measurement of the invasive tumor size, could not be determined. Second, the definition of microscopic invasion varied among the reporting institutions. It is anticipated that acceptance of a uniform definition of microscopic invasion (now restricted to invasive carcinomas measuring ≤ 1 mm beyond the basement membrane) will result in future studies demonstrating a lower frequency of lymph node metastases than noted in this series.21 Third, cases with multicentric and multifocal involvement are not well defined in the SEER database. In the case of T1a infiltrating lobular carcinoma, a higher proportion of cases with microscopic foci of malignancy and a higher frequency of lymph node metastases was noted when compared with infiltrating ductal carcinoma. We attribute these differences to an increased frequency of multicentricity, multifocality, and indistinct tumor margins among infiltrating lobular carcinomas.22–24 Among infiltrating lobular carcinomas identified as limited to microscopic foci of malignancy the factors mentioned earlier likely resulted in an underestimation of the extent of disease in an unknown percentage of cases. Despite these limitations, the frequency of lymph node metastases associated with microscopically invasive carcinomas and the progressive increase in the frequency of lymph node metastases with increasing tumor size demonstrate that invasive carcinomas measuring ≤ 1 cm are associated with a significant frequency of lymph node metastases.
The necessity of performing routine axillary lymph node dissections recently has been questioned based on a failure to demonstrate a statistically significant survival advantage among any of the treatment arms in NSABP (National Surgical Adjuvant Breast and Bowel Project) B-04.25 Selective avoidance of axillary lymph node dissection has been suggested for cases of invasive breast carcinoma that will require adjuvant systemic therapy based on the characteristics of the primary tumor alone, patient age > 70 years, or for those cases with a low likelihood of having lymph node metastases.5, 11, 13, 16, 26–29 Recently, the threshold value below which the frequency of associated lymph node metastases is “low enough” to avoid an axillary lymph node dissection has been suggested to be ≤ 5%.30 It generally is accepted that pure tubular carcinomas measuring ≤ 1 cm, which have a frequency of associated axillary lymph node metastases < 5%, do not require an axillary lymph node dissection.31–33
In the future, it may be possible to perform therapeutic axillary lymph node dissections selectively. The accuracy of the sentinel lymph node biopsy technique at specialized centers indicates that it is promising technique in the management of breast malignancies. However, this technique can be technically challenging with significant variation in success rates among surgeons.34 Confirmation of reproducibility and a low false-negative rate is required prior to recommending that this technique be applied in the general medical community.35 Currently, the performance of an axillary lymph node dissection is the only method by which the presence or absence of lymph node metastases can be determined reliably. Due to the substantial change in the management of breast carcinoma with lymph node metastases, a Level I and Level II axillary lymph node dissection is recommended to stage the majority of cases with invasive breast carcinomas measuring ≤ 1 cm in greatest dimension.
A variety of histologic and clinical factors previously have been noted to affect the frequency of axillary lymph node metastases in breast malignancies.3, 11, 25, 31, 32 In this current study (limited to T1a and T1b breast carcinomas) lower size, favorable histology (mucinous, papillary, and tubular carcinomas), lower histologic grade, and increasing age were associated significantly with a lower frequency of axillary lymph node metastases. Statistically significant variation also was noted in the frequency of lymph node metastases among tumor location sites in the breast. The clinical relevance of variation in the frequency of lymph node metastases based on tumor location is questionable. Future analyses including combinations of factors utilized in this current study will permit the frequency of axillary lymph node metastases to be more precisely determined. Irrespective of other considerations, such as a limited life expectancy, an axillary lymph node dissection reasonably can be avoided in cases that lack palpable axillary adenopathy and have an associated frequency of lymph node metastases similar to that of tubular carcinomas (≤ 5%).30 Based on the data presented in this study it is unnecessary to perform a routine axillary lymph node dissection in cases of T1a and T1b mucinous and tubular carcinomas, T1a papillary carcinomas, and T1a Grade 1 carcinomas unless palpable axillary adenopathy is present.