DOI: https://doi.org/10.21203/rs.3.rs-1142496/v1
Purpose: In breast cancer, the lymph node status is of prognostic importance and a decisive factor in therapy planning. This study shows the distribution of lymph nodes metastases of node-positive breast cancer patients. Risk factors for lymph node metastases are described.
Methods: 2095 patients with primary breast cancer were analyzed. Analysis included descriptive analysis (median, standard deviation, ranges) and statistical analysis (Chi², discriminant analysis).
Results: The nodal stage was positive in 39.4% of all patients and negative in 60.6%. If the nodal stage was positive, only 1 lymph node was involved in 36% of the patients; more than 1 lymph node was involved in 64% of the patients. With an increasing number of lymph node metastases in level I, the probability of an involvement in level III also increases (F 437.845, p = .000). Other indicators are evidence of hemangiosis (F 247.728, p = .000) or lymphangiosis (F 167.368, P = .000). Despite <10 affected lymph nodes, 3.4% of the patients had nodal stage N3 due to level III involvement.
Conclusion: Even with only a small number of lymph node metastases in level I higher lymphatic stations are often affected. The data cannot lead to a decision regarding an operation or radiation indication, but they can better substantiate the risk for certain therapy decisions.
In breast cancer, the lymph node status, together with the tumor biology, is of prognostic importance and a decisive factor in therapy planning (1). Data from axillary radiation therapy studies with a benefit of local control, mortality and overall survival suggest that affected lymph nodes, surgically not removed or not irradiated represent a risk of metastasis (2). To our best of knowledge it is not known yet how often the individual lymph node level I, II, III and interpectoral are affected in a positive nodal status. These data were presented in the present study. An attempt was made to determine risk factors for the affection of higher lymph node stations using discriminant analysis.
Patients: We analyzed all cases of breast cancer at the St. Josefs Hospital Offenburg who underwent conventional axillary dissection from 1997 to 2010 with 2095 patients. The lymph nodes of axillary levels I to III and interpectoral were removed and examined separately histopathologi-cally. The results were recorded in our MaDos documentation system.
Data management and statistics: All data of the tumor documentation system MaDos were exported in Microsoft Excel 2010. These data were then transferred to WinSTAT (version 2012.1) and SPSS (version 21) for further statistical analysis. The analysis included descriptive analysis (median, standard deviation, ranges) and statistical analysis (Chi², discriminant analysis)
Ethics: The consent for the project was given by the ethics committee of the University of Freiburg (application number 72/20).
In our patient collective of 2095 patients with conventional axillary dissection, the nodal status was node-negative (N -) in 1270 (61%) and node-positive (N +) in 825 patients (39%).
In order to analyze the correct lymph node distribution, it was necessary to know the operation frequency of the specific nodal stations. In patients with affected lymph nodes, the lymphadenectomy was performed in level I in 100%, in level II in 99%, in level III in 95% and in interpectoral in 71% of all patients.
Number of removed and affected lymph nodes related to the surgery in the corresponding level with N +
| level I | level II | level III | interpectoral | total number | |
|---|---|---|---|---|---|
| surgery | 825 | 816 | 785 | 582 | 825 |
| N+ | 814 | 237 | 145 | 56 | |
| % | 98.7 | 29.0 | 18.5 | 9.6 | |
| removed LN, mean | 16.5 | 6.4 | 2.7 | 1.3 | 26.5 |
| SD | 5.6 | 3.6 | 2 | 1.4 | 8.9 |
| min - max | 3 - 49 | 0 - 23 | 0 - 18 | 0 - 9 | 3 - 77 |
| affected LN, mean | 4 | 1,1 | 0,4 | 0,2 | 5.6 |
| SD | 4.6 | 2.4 | 1.2 | 0.6 | 7.6 |
| min - max | 0 - 30 | 0 - 19 | 0 - 12 | 0 - 7 | 1 - 51 |
| SD: standard deviation, LN: lymph nodes | |||||
Table 1 shows that with positive nodal stage, level I was affected in 98.7%. Skip metastases were found in 11 patients, meaning other stations than level I of the lymphatic drainage were affected. Level II was affected in 29%, level III in 18.5% and interpectoral in 9.6%. If the nodal stage was positive, a mean of a total of 26.5 lymph nodes were removed (range: 3 - 77), including 16.5 lymph nodes in level I, 6.4 lymph nodes in level II, 2.7 lymph nodes in level III and 1.3 in interpectoral. The mean was 5.6 lymph node metastases (range: 1 - 51).
| frequency at N + | ||
|---|---|---|
| Number of affected lymph nodes | n | % |
| 1 LN | 295 | 35.8 |
| 2 LN | 117 | 14.2 |
| 3 LN | 90 | 10.9 |
| 4 LN | 51 | 6.2 |
| 5 LN | 47 | 5.7 |
| 6 LN | 24 | 2.9 |
| 7 LN | 15 | 1.8 |
| 8 LN | 29 | 3.5 |
| 9 LN | 15 | 1.8 |
| ≥ 10 LN | 142 | 17.2 |
| 1 - 3 LN | 502 | 60.9 |
| 4 - 9 LN | 181 | 21.9 |
| total number | 825 | 100 |
| LN: lymph nodes, N+: nodal positive, n: number of patients | ||
Table 2 shows that of a total of 825 patients with positive node status, 295 patients (35.8%) had only one lymph node metastasis and 530 patients (64.2%) had more than one. 60.9% of the patients had 1 to 3 lymph node metastases and in 21.9% 4 to 9 lymph nodes were affected. 17.2% had more than 9 lymph node metastases. The mean was 5.6 lymph node metastases.
Frequency of axillary involvement with 0, 1, 2, 3 or more affected lymph nodes in level I.
If there were no lymph node metastases in level I but an involvement of higher lymph nodes, the distribution of lymph node metastases in a total of 825 patients with positive node status shows that level I was not affected in 11 patients (1.33%). Lymph node metastases were found in level II in 3 patients, in level III in 4 patients and in interpectoral in 4 patients.
| Affected node stations | 1 lymph node metastasis in Level I | 2 lymph node metastases in Level I | 3 lymph node metastases in Level I | ≥4 lymph node metastases in level I | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LA | n | % | LA | n | % | LA | n | % | LA | n | % | |
| level I | 312 | 312 | 100 | 120 | 120 | 100 | 92 | 92 | 100 | 290 | 290 | 100 |
| level II, III or ip | 27 | 8.7 | 21 | 17.5 | 22 | 23.9 | 199 | 68.6 | ||||
| level II | 308 | 18 | 5.8 | 120 | 13 | 10.8 | 90 | 17 | 18.9 | 287 | 186 | 64.8 |
| level III | 300 | 9 | 3.0 | 117 | 4 | 3.4 | 86 | 8 | 9.3 | 271 | 120 | 44.3 |
| ip | 221 | 6 | 2.7 | 92 | 5 | 5.4 | 64 | 5 | 7.8 | 196 | 36 | 18.4 |
LA: lymphadenectomy; n: number of patients with lymph node metastases; %: percentage of affected lymph node station in relation to the frequency of surgery; ip: interpectoral
Table 3 shows the distribution of lymph node metastasis in patients with 1, 2, 3 or more lymph node metastases in level I. Of 312 patients with 1 affected lymph node in level I, a total of 27 of the patients (8.7%) had lymph node involvement in higher lymphatic stations. In 6 patients, several lymph node stations were affected. In relation to the frequency of surgery in the higher lymphatic stations, this corresponds to a value of 5.8% for level II, 3,0% for level III and 2.7% for interpectoral. Of 120 patients with 2 affected lymph nodes in level I, a total of 21 (17.5%) of the patients had lymph node involvement in higher lymphatic stations. Several lymph node stations were affected in 1 patient. In relation to the frequency of surgery in the higher lymphatic stations, this corresponds to a value of 10.8% for level II, 3.4% for level III and 5.4% for interpectoral.
Of 92 patients with 3 affected lymph nodes in level I, a total of 22 (23.9%) of the patients had lymph node involvement in higher lymphatic stations. In 8 patients, several lymph node stations were affected. In relation to the frequency of surgery in the higher lymphatic stations, this corresponds to a value of 18.9% for level II, 9.3% for level III and 7.8% for interpectoral.
Of 290 patients with ≥4 affected lymph nodes in level I, 199 (68.6%) of the patients had lymph node involvement in higher lymphatic stations. In 143 patients, several lymph node stations were affected. In relation to the frequency of surgery in the higher lymphatic stations, this corresponds to a value of 64.8% for level II, 44.3% for level III and 18.4% for the interpectoral.
(A detailed table with more information is available upon request.)
| Number of total affected lymph nodes | Number of patients | N1 | N2 | N3 | |
|---|---|---|---|---|---|
| n | n | n | n | n | % |
| 1 | 295 | 291 | 0 | 4 | 1.4 |
| 2 | 117 | 116 | 0 | 1 | 0.9 |
| 3 | 90 | 84 | 0 | 6 | 6.7 |
| 4 - 9 | 181 | 0 | 164 | 17 | 9.4 |
| total number | 683 | 491 | 164 | 28 | 4.1 |
| ≥10 | 142 | 0 | 0 | 142 | 100 |
| N: nodal status N1 – 3 | |||||
Table 4 shows that of a total of 683 patients with lymph node involvement of <10 lymph nodes, who could not be assigned to stage N3 due to the number of affected lymph nodes, 28 of these patients (4.1%) had stage N3.
6.1 Chi-square contingency test
| Category | Lymph node metastasis in Level III | p | Chi² |
|---|---|---|---|
| Age | ns | ||
| < 53Jahre | 6.8% (30/440) | ||
| 53 bis 62 Jahre | 5.9% (23/390) | ||
| 63 bis 71 Jahre | 8.4% (36/430) | ||
| > 71 Jahre | 7.6% (30/395) | ||
| BMI | ns | ||
| ≤ 25 | 5.4% (31/576) | ||
| > 25 ≤ 30 | 6.0% (33/552) | ||
| 30 bis 34,9 | 9.3% (25/268) | ||
| ≥ 35 | 7.9% (8/101) | ||
| inside quadrant * | 0.036 | 4.382 | |
| no | 8.2% (85/1034) | ||
| yes | 5.5% (34/621) | ||
| central quadrant *** | 0.001 | 10.675 | |
| no | 6.6% (102/1540) | ||
| yes | 14.8% (17/115) | ||
| tumor stage *** | < 0.001 | 61.551 | |
| T1 | 4.4% (46/1047) | ||
| T2 | 9.7% (50/513) | ||
| T3 | 21.7% (13/47) | ||
| T4 | 21.7% (10/46) | ||
| tumor size *** | < 0.001 | 32.316 | |
| ≤ 10mm | 2% (5/252) | ||
| 11 ≤ 20mm | 5.8% (33/570) | ||
| 21≤ 35mm | 9.7% (35/361) | ||
| > 35mm | 16.9% (20/118) | ||
| Grading *** | < 0.001 | 17.269 | |
| G1 | 2.9% (3/102) | ||
| G2 | 5.8% (62/1069) | ||
| G3 | 11.2% (54/484) | ||
| Lymphangiosis *** | < 0.001 | 111.938 | |
| no | 2.0% (12/612) | ||
| yes | 23.4% (67/286) | ||
| Hemangiosis *** | < 0.001 | 132.990 | |
| no | 4.4% (34/768) | ||
| yes | 41.9% (31/74) | ||
| Number of affected lymph nodes in level I *** | < 0.001 | 515.974 | |
| 0 - 3 | 1.3% (19/1421) | ||
| > 3 | 42.7% (100/234) | ||
| hormone receptor | ns | ||
| ER and PR negative | 7.7% (14/181) | ||
| ER or PR positive | 7.1% (105/1474) | ||
| Her2 | ns | ||
| negative | 7.2% (106/1472) | ||
| positive | 7.1% (12/183) | ||
| uPA | ns | ||
| negative | 8.9% (65/729) | ||
| positive | 6.8% (841/599) | ||
| PAI-1 | ns | ||
| negative | 8.2% (57/698) | ||
| positive | 7.8% (50/639) | ||
| proliferation (Ki67/MIB) | ns | ||
| not increased (< 30/50) | 6.9% (73/1064) | ||
| increased (≥ 30/50) | 7.8% (46/591) | ||
| * = p ≤ 0.05; **= p ≤ 0.01; ***= p ≤ 0.001; ns = not significant | |||
Table 5 shows that in the univariate chi-square analysis according to Pearson (uncorrected), age, BMI and molecular biological factors (hormone receptor status, Her2 status) were not found to be significant for an association with lymph node metastases in level III. The proliferation markers (Ki67 / MIB) or invasion markers (uPA / PAI-1) were also not significant. In contrast, the tumor stage (T1-T4), tumor size, grading, lymphangiosis, hemangiosis and the number of affected lymph nodes in level I (≤3 /> 3) were found to be highly significant. The location in the inner or central quadrant was also significant for lymph node metastases in level III.
6.1 Discriminant analysis
A step-by-step discriminant analysis was performed with the significant variables from the chi-square contingency test. The group differences were found to be significant in the discriminant analysis (Wilks-Lambda .587, Chi2 = 380.489, p = .000). The most important prognostic factors were the number of positive lymph nodes in level I (≤ 3 /> 3) (F= 437.845, p = .000), hemangiosis (F= 247.728, p= .000) and lymphangiosis (F= 167.368, p= .000).
The classification results show that 91.9% (730/794) of the originally grouped cases were correctly classified. In particular, the negative predictive value was at 85.3% (677/794). However, the positive predictive value was 6.7% (53/794).
| Axillary relapse | Nodal stage N0 n =1270 | Nodal stage N+ n = 825 | Total number n = 2095 | |||
|---|---|---|---|---|---|---|
| n | % | n | % | n | % | |
| Axillary relapse only first event | 3 | 0.24 | 4 | 0.49 | 7 | 0.33 |
| Axillary relapse AND local breast relapse first event | 2 | 0.16 | 1 | 0.12 | 3 | 0.14 |
| Axillary relapse total | 5 | 0.39 | 5 | 0.61 | 10 | 0.47 |
| Axillary relapse AND later distant metastasis | 2 | 0.16 | 2 | 0.24 | 4 | 0.19 |
Table 6 shows the number of axillary relapses as first event, either alone or in combination with a local relapse. It is also shown how often a distant metastasis occurs later.
Determining the lymph node status is a diagnostic measure. It is of prognostic importance and is used for therapy planning (1). In addition, axillary lymphadenectomy also served to reduce the tumor burden. The benefit of axillary radiation therapy shows that unremoved, untreated metastases can have consequences for local control and disease-free survival (2). This suggests that any affected lymph nodes that were left were successfully treated by radiation therapy. This assumption is confirmed by the available data with the detection of lymph node metastases in higher lymphatic stations even with a low metastatic lymph node load in level I. As shown in the case of only 1, 2 or 3 affected lymph nodes in level I, level III was affected in 3.0%, 3.4% and 9.3% of the patients and the interpectoral in 2.7%, 5.4% and 7.8%. In the case of 1-3 affected lymph nodes in level I, the proportions of affected lymph nodes in level III (21/503) and in interpectoral (16/377) were both 4.2%. In the case of more than 3 affected lymph nodes in level I the proportion was 44.3% in level III and 18.4% in interpectoral. These data are similar to the results of Joshi, S. et al. (3). Their results showed a proportion of 9.4% positive lymph nodes in level III if 1-3 lymph nodes in level I and II were affected and a proportion of 53.2% in the case of 4 affected lymph nodes. In our overall results, level III was affected in 18.5% and interpectoral in 9.6% if there was a positive nodal status. Joshi et al. showed this proportion to be 27.3% and 4.7%, respectively (3). Other study groups were also able to demonstrate a level III infestation between 15% and 59% (4) (5) (6).
Altogether, in the case of only 1 affected lymph node in level I, lymph nodes metastases were detectable in the subsequent lymph node stations level II, III and interpectoral in 8% of the patients. In the case of 2 affected lymph nodes in level 1, the proportion was 15% and in the case of 3 affected lymph nodes in level I, the proportion was 20%. In the case of unaffected lymph nodes in level I, 11 of 825 patients (1.3% [based on operation frequency 1.5%]) had a positive nodal stage in level II, III or interpectoral (interpectoral). It would therefore be a false negative nodal stage. Whether a false negative rate of 1.5% justifies a conventional axillary lymphadenectomy is currently decided in the negative based on the data from the sentinel lymphadenectomy (7) (8) (9).
683 of the 825 patients with lymph node metastases had <10 affected lymph nodes, which corresponds to the nodal stage N1 or N2. The surgery of level III resulted in a change of stage from stage N2 to stage N3 in 4.1% (28 /683) of these patients due to lymph node metastases in level III (chapter 5).
A total of 170 of the 825 patients with lymph node metastases had nodal stage N3, 142 (83.5%) patients with more than 9 affected lymph nodes and 28 (16.5%) patients with <10 affected lymph nodes, which is prognostic and also possibly with therapeutic consequences.
Indicators of level III lymph node involvement in the chi-square test were tumor stage, tumor size, grading, lymphangiosis, hemangiosis and the number of affected lymph nodes in level I (≤ 3 /> 3).
The location in the inner or central quadrant was also significant for lymph node metastases in level III. This corresponds to the numbers of other published data (3). The discriminant analysis showed the number of positive lymph nodes in level I (≤3 /> 3), hemangiosis and lymphangiosis as the most important prognostic factors for lymph node metastases in level III. Exclusion of lymph node metastases in level III can be assumed at 85.3% in the absence of these factors. However, the positive predictive value in this study was 6.7%. Even if some studies showed better local control with effects on overall survival (10) (16), NSABP B-04, as the first randomized study, was not able to show a survival benefit from conventional axillary dissection (11). However, these described factors should be viewed as a particular risk for lymph node metastases and should possibly be included when considering the radiation treatment plan. Also, it should be noted, that breast cancer metastases are also known to be radio-resistant (12); and resistance rates of 8-15% must be expected (13).
This study showed that even with only a small number of lymph node metastases in level I, the higher lymphatic stations are often affected too. If the node status is positive, more than 1 lymph node is involved in 64%. The data cannot lead to a decision regarding an operation or radiation indication, but it can better substantiate the risk for certain therapy decisions.
A further improvement in the imaging of the localization of lymph node metastases and examinations for the prediction of a possible radio resistance would be desirable.
Acknowledgments
We would like to thank all those who have contributed to the establishment and maintenance of our data bank on breast cancer. This includes our tumor-documenters, namely Mrs. Andrea Heuberger, as well as all the gynecologists and general practitioners in our region who supplied us with information about the follow-up of the patients. We would especially like to thank Mr. Alexander Schattenberg, who developed the MaDos documentation system and was always ready to assist us with special queries.
Competing interests
The authors have no relevant financial or non-financial interests to disclose.
Funding
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Author Contributions
Example statement:
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Heidrun Männle, Matthias Frank, Felix Momm, Jan Willem Siebers. The first draft of the manuscript was written by Heidrun Männle and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Data Availability
The datasets generated during and analysed during the current study are not publicly available due to all data are listed in our own tumor documentation system MaDos but are available from the corresponding author on reasonable request..
Ethics approval
This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of University of Freiburg (application number 72/20).
Consent to participate
Informed consent was obtained from all individual participants included in the study.