Abstract
At present, the risk factors and prognosis of sentinel lymph node metastasis (SLNM) are analyzed based on the study of axillary lymph node metastasis, but whether there is a difference between the two is unclear. Therefore, an accurate and appropriate predictive model needs to be proposed to evaluate patients undergoing sentinel lymph node biopsy (SLNB) for breast cancer. We selected 16983 women with breast cancer from the Surveillance Epidemiology and End Results (SEER) database. They were randomly assigned to two cohorts, one for development (n = 11891) and one for validation (n = 5092). multi-factor logistics regression was used to distinguish risk factors affecting SLNM. The potential prognostic factors were identified using the COX regression analysis. The hazard ratio (HR) and 95% confidence interval (95%CI) were calculated for all results. Multiple Cox models are included in the nomogram, with a critical P value of .05. In order to evaluate the model’s performance, Concordance index and receiver operating characteristic curves were used. Six independent risk factors affecting SLNM were screened out from the Logistic regression, including tumor location, number of regional lymph nodes (2-5), ER positive, PR positive, tumor size (T2-3), and histological grade (Grade II-III) are independent risk factors for SLNM in patients (P < .05). Eight prognostic factors were screened out in the multivariate COX regression analysis (P < .05): Age: Age 60 to 79 years, Age ≥ 80 years; Race; Histological grading: Grade II, Grade III; No radiotherapy; Tumor size: T2, T3; ER positive:, sentinel lymph node positive, married. Histological grade, tumor location, T stage, ER status, PR status and the number of SLNB are significantly correlated with axillary SLNM. Age, ethnicity, histological grade, radiotherapy, tumor size, ER status, SLN status, and marital status were independent risk factors for Breast cancer specific survival (BCSS). Moreover, the survival rate of patients with 3 positive SLNs was not significantly different from that with one or two positive SLNs, We concluded that patients with stage N1 breast cancer were exempt from axillary lymph node dissection, which is worthy of further study.
Keywords: breast cancer, factor prognosis, prognostic, risk model
1. Introduction
Sentinel lymph node (SLN) was found during dorsal phallic lymphangiography by Cabanas in 1977,[1]and was the first lymph node stop for primary tumor metastasis. When SLN did not have metastasis, there was a very low probability of non-sentinel lymph node metastasis (SLNM).[2] sentinel lymph node biopsy (SLNB) provides prognostic information and guides adjuvant treatment. NSABP-B32 study confirmed that negative axillary sentinel lymph node biopsy exempted axillary lymph node dissection.[3] Multiple prospective clinical studies[4–6] (such as ACOSOG Z0011, AMAROS, IBCSG 23-01, etc.) have shown that For breast cancer patients with sentinel lymph node metastasis (1–2 SLNM, early breast cancer with cN0 SLN positive, sentinel lymph node micro-metastases), whether or not ALND has no significant effect on postoperative recurrence and survival. Now SLNB is recommended for clinically negative axillary patients after neoadjuvant chemotherapy.[7] At present, breast-conserving surgery and SLNB have been widely recognized as radical surgical methods for patients.
With improvement of the lymph node technique and research, SLNB has been widely used in the diagnosis and treatment of cancers, especially for breast cancer, Axillary lymph node dissection (ALND) has gradually been replaced by this procedure. SLNB does not affect the diagnostic accuracy and prognostic information, and has become the standard surgical procedure for clinical lymph node negative (cN0) breast cancer patients. A number of trials have confirmed that the incidence of upper limb edema, numbness, pain, paresthesia, shoulder joint mobility disorder and other complications in SLNB patients is significantly lower than that of ALND, which improves the quality of life of patients.[8,9] In the 8th edition of the American Joint Committee on Cancer (AJCC)[10] breast cancer staging, the number of SLNs was less than 6, and SLN footnote “sn” could not be used if there were more than 6. But the minimum number of SLNs is not specified. So it’s theoretically possible to take one. Studies have shown that the false negative rate of sentinel lymph nodes is less than 10%.[11] But surgeons tend to take more sentinel nodes during surgery to reduce the false-negative rate. There is evidence that removing fewer nodes will increase the false negative rate.[12] It has been reported that the average number of SLN taken during surgery is 2.35,[13] and it is believed that the number of SLN should be greater than 1. Therefore, There is an urgent need to establish an accurate model for evaluating the factors and prognosis of SLNM. Based on the Surveillance Epidemiology and End Results (SEER) database, this study collected all the information of patients with breast conserving breast cancer from 2014 to 2015, and studied the clinicopathological features, prognosis and risk factors of SLNM through the big data level. To evaluate the specific survival rate, we are working on establishing a prognostic nomogram based on important factors.
2. Patients and method
2.1. Patient data and sources
The patients were drawn from SEER database and included the all available information. In terms of demographics, these registries represent the United States’ general population.[14] The inclusion criteria were as follows: diagnosed in 2014 to 2015; Confirmed surgical methods: breast conserving surgery + SLNB (code 20, 22, 23, 24), and the number of sentinel lymph nodes was 1 to 5; Clinical stage: T1-3. The chemotherapy we studied was adjuvant setting. Exclusion criteria: M1 and Mx staging; and Incomplete information.
2.2. Data characteristics and endpoints
The variables included: Patients’ demographics, treatment course, and tumor specific information. The primary end points were: Risk factors for SLNM. BCSS.
We convert age were transformed into categorical variables: 20 to39 years, 40 to 59 years, 60 to 79 years, ≥80years. And defined marital status as married, separated, divorced or widowed (SDW), and single. The remaining variables remain the same According to the seventh edition AJCC staging, this study documented accurate information about the TMN system.
2.3. Statistical analysis
We used analysis software to randomize patients into a 7 to 3 ratio between the development cohort and the validation cohort. The descriptive analysis used t test and Chi-square test to explore the baseline characteristics of patients in both groups. In the development cohort, multi-factor logistics regression was used to identify risk factors affecting SLNM. The potential prognostic factors were identified using the univariate COX regression analysis. The multivariate COX proportional risk regression model was used when the P value < .05. The hazard ratio (HR) and 95% confidence interval (95% CI) were calculated for all results. Multiple Cox models are included in the nomogram, with a critical P value of .05. The nomogram was created to visually predict survival probabilities in the developmental cohort. In order to evaluate the model’s performance, the Harrell’s concordance index (C-index) and receiver operating characteristic curves were used. The higher the value (close to 1), the more accurate the prediction of prognosis. Statistical analysis using R Version 4.2.2 (https://cran.r-project.org/bin/windows/base/).
3. Results
3.1. Population characteristics
Our study included 16,983 patients with breast cancer, 11,891 patients (70%) were assigned to the developmental cohort, and 5092 patients (30%) were assigned to the validation cohort. Table 1 shows the demographic and clinicopathological characteristics of the patients. The 2 groups were no statistical difference in terms of variables. For the general, developmental, and validation cohorts, the median follow-up time was 57 months. The baseline demographic and clinicopathological characteristics of positive and negative sentinel nodes in the developmental cohort are shown in Table 2. Among them, age, tumor site, histological grade, tumor size, estrogen receptor (ER) status, progesterone receptor (PR) status, the number of SLNs, and whether or not chemotherapy had statistical significance (P < .05).
Table 1.
Baseline demographical and clinicopathological characteristics of patients.
| Characteristics | Total cohort, N (%) | Development cohort, N (%) | Validation cohort, N (%) | P value |
|---|---|---|---|---|
| Number of patients | N = 16983 | N = 11891 (70) | N = 5092 (30) | |
| Age | ||||
| 20–39 yr | 358 (2.11) | 257 (2.16) | 101 (1.98) | .887 |
| 40–59 yr | 6179 (36.4) | 4326 (36.4) | 1853 (36.4) | |
| 60–79 yr | 9332 (54.9) | 6533 (54.9) | 2799 (55.0) | |
| ≥80 yr | 1114 (6.56) | 775 (6.52) | 339 (6.66) | |
| Race | ||||
| White | 13920 (82.0) | 9731 (81.8) | 4189 (82.3) | .71 |
| Black | 1530 (9.01) | 1087 (9.14) | 443 (8.70) | |
| AI | 112 (0.66) | 75 (0.63) | 37 (0.73) | |
| API | 1421 (8.37) | 998 (8.39) | 423 (8.31) | |
| Site | ||||
| Nipple | 66 (0.39) | 41 (0.34) | 25 (0.49) | .261 |
| Upper-inner quadrant | 3816 (22.5) | 2687 (22.6) | 1129 (22.2) | |
| Lower-inner quadrant | 1559 (9.18) | 1117 (9.39) | 442 (8.68) | |
| Upper-outer quadrant | 9497 (55.9) | 6634 (55.8) | 2863 (56.2) | |
| Lower-outer quadrant | 2045 (12.0) | 1412 (11.9) | 633 (12.4) | |
| Histology | ||||
| Others | 3729 (22.0) | 2574 (21.6) | 1155 (22.7) | .135 |
| No special type | 13254 (78.0) | 9317 (78.4) | 3937 (77.3) | |
| Grade | ||||
| Grade I | 5243 (30.9) | 3654 (30.7) | 1589 (31.2) | .861 |
| Grade II | 7614 (44.8) | 5346 (45.0) | 2268 (44.5) | |
| Grade III | 4126 (24.3) | 2891 (24.3) | 1235 (24.3) | |
| Laterality | ||||
| Left | 8629 (50.8) | 6028 (50.7) | 2601 (51.1) | .645 |
| Right | 8354 (49.2) | 5863 (49.3) | 2491 (48.9) | |
| T | ||||
| T1 | 12920 (76.1) | 9057 (76.2) | 3863 (75.9) | .292 |
| T2 | 3845 (22.6) | 2672 (22.5) | 1173 (23.0) | |
| T3 | 218 (1.28) | 162 (1.36) | 56 (1.10) | |
| Radiation | ||||
| Yes | 14222 (83.7) | 9966 (83.8) | 4256 (83.6) | .711 |
| No | 2761 (16.3) | 1925 (16.2) | 836 (16.4) | |
| Chemotherapy | ||||
| Yes | 5202 (30.6) | 3655 (30.7) | 1547 (30.4) | .644 |
| No | 11781 (69.4) | 8236 (69.3) | 3545 (69.6) | |
| ER | ||||
| Negative | 14943 (88.0) | 10452 (87.9) | 4491 (88.2) | .583 |
| Positive | 2040 (12.0) | 1439 (12.1) | 601 (11.8) | |
| PR | ||||
| Negative | 13400 (78.9) | 9371 (78.8) | 4029 (79.1) | .643 |
| Positive | 3583 (21.1) | 2520 (21.2) | 1063 (20.9) | |
| HER2 | ||||
| Negative | 1894 (11.2) | 1344 (11.3) | 550 (10.8) | .342 |
| Positive | 15089 (88.8) | 10547 (88.7) | 4542 (89.2) | |
| Regional nodes examined | ||||
| 1 | 5254 (30.9) | 3668 (30.8) | 1586 (31.1) | .938 |
| 2 | 4977 (29.3) | 3487 (29.3) | 1490 (29.3) | |
| 3 | 3525 (20.8) | 2485 (20.9) | 1040 (20.4) | |
| 4 | 1998 (11.8) | 1388 (11.7) | 610 (12.0) | |
| 5 | 1229 (7.24) | 863 (7.26) | 366 (7.19) | |
| Positive lymph nodes | ||||
| No | 13631 (80.3) | 9553 (80.3) | 4078 (80.1) | .706 |
| Yes | 3352 (19.7) | 2338 (19.7) | 1014 (19.9) | |
| Marital status | ||||
| Single | 2324 (13.7) | 1626 (13.7) | 698 (13.7) | .025 |
| Married | 10180 (59.9) | 7198 (60.5) | 2982 (58.6) | |
| SDW | 4479 (26.4) | 3067 (25.8) | 1412 (27.7) | |
AI = American Indian/Alaska Native, API = Asian or Pacific Islander, SDW = separated, divorced or widowed.
Table 2.
Clinicopathological parameters of included patients and association with SLN status.
| Negative (N = 9553) | Positive (N = 2338) | P value | |
|---|---|---|---|
| Age | |||
| 20–39 yr | 190 (2.0%) | 67 (2.9%) | <.001 |
| 40–59 yr | 3398 (35.6%) | 928 (39.7%) | |
| 60–79 yr | 5341 (55.9%) | 1192 (51.0%) | |
| ≥80 years | 624 (6.5%) | 151 (6.5%) | |
| Race | |||
| W | 7831 (82.0%) | 1900 (81.3%) | .0517 |
| B | 842 (8.8%) | 245 (10.5%) | |
| AI | 61 (0.6%) | 14 (0.6%) | |
| API | 819 (8.6%) | 179 (7.7%) | |
| Site | |||
| Nipple | 26 (0.3%) | 15 (0.6%) | <.001 |
| Upper-inner quadrant | 2314 (24.2%) | 373 (16.0%) | |
| Lower-inner quadrant | 920 (9.6%) | 197 (8.4%) | |
| Upper-outer quadrant | 5188 (54.3%) | 1446 (61.8%) | |
| Lower-outer quadrant | 1105 (11.6%) | 307 (13.1%) | |
| Histologic | |||
| Others | 2082 (21.8%) | 492 (21.0%) | .446 |
| No special type | 7471 (78.2%) | 1846 (79.0%) | |
| Grade | |||
| Grade I | 3143 (32.9%) | 511 (21.9%) | <.001 |
| Grade II | 4171 (43.7%) | 1175 (50.3%) | |
| Grade III | 2239 (23.4%) | 652 (27.9%) | |
| Laterality | |||
| Left | 4859 (50.9%) | 1169 (50.0%) | .468 |
| Right | 4694 (49.1%) | 1169 (50.0%) | |
| T | |||
| T1 | 7675 (80.3%) | 1382 (59.1%) | <.001 |
| T2 | 1779 (18.6%) | 893 (38.2%) | |
| T3 | 99 (1.0%) | 63 (2.7%) | |
| Radiation | |||
| Yes | 7999 (83.7%) | 1967 (84.1%) | .661 |
| No | 1554 (16.3%) | 371 (15.9%) | |
| Chemotherapy | |||
| Yes | 2423 (25.4%) | 1232 (52.7%) | <.001 |
| No | 7130 (74.6%) | 1106 (47.3%) | |
| ER | |||
| Negative | 8350 (87.4%) | 2102 (89.9%) | .001 |
| Positive | 1203 (12.6%) | 236 (10.1%) | |
| PR | |||
| Negative | 7445 (77.9%) | 1926 (82.4%) | <.001 |
| Positive | 2108 (22.1%) | 412 (17.6%) | |
| HER2 | |||
| Negative | 1068 (11.2%) | 276 (11.8%) | .413 |
| Positive | 8485 (88.8%) | 2062 (88.2%) | |
| Examined | |||
| 1 | 3117 (32.6%) | 551 (23.6%) | <.001 |
| 2 | 2837 (29.7%) | 650 (27.8%) | |
| 3 | 1942 (20.3%) | 543 (23.2%) | |
| 4 | 1044 (10.9%) | 344 (14.7%) | |
| 5 | 613 (6.4%) | 250 (10.7%) | |
| Marital | |||
| Single | 1314 (13.8%) | 312 (13.3%) | .154 |
| Married | 5743 (60.1%) | 1455 (62.2%) | |
| SDW | 2496 (26.1%) | 571 (24.4%) | |
AI = American Indian/Alaska Native, API = Asian or Pacific Islander, SLN = sentinel lymph node.
3.2. Analysis of factors influencing SLNM in a developmental cohort
In Table 2, there were 7 statistically significant factors: age, tumor site, histological grade, tumor size, ER status, PR status, and the number of SLNs. It has been previously reported that the expression of human epidermal growth factor receptor-2[15] may also affect axillary metastasis in patients. These 8 factors were used as independent variables and SLNM was used as the dependent variable in a multivariate binary logistic regression analysis. Finally, 6 independent risk factors affecting SLNM were screened out from the multivariate Logistic regression analysis, including tumor location (upper inner quadrant, lower inner quadrant, upper outer quadrant, Outer and lower quadrant), number of regional lymph nodes examined (2-5), ER positive, PR positive, tumor size (T2-3), and histological grade (Grade II-III) are independent risk factors for SLNM in patients. (Table 3).
Table 3.
Multivariate analysis of SLNM.
| Pr(> z ) | OR | 2.50% | 97.50% | |
|---|---|---|---|---|
| Age 40–59 yr | 0.599 | 0.922 | 0.686 | 1.25 |
| Age 60–79 yr | 0.239 | 0.835 | 0.621 | 1.13 |
| Age ≥ 80 yr | 0.344 | 0.846 | 0.599 | 1.2 |
| Site Upper-inner quadrant | <0.01 | 0.281 | 0.145 | 0.563 |
| Site Lower-inner quadrant | <0.01 | 0.4 | 0.204 | 0.808 |
| Site Upper-outer quadrant | 0.044 | 0.504 | 0.262 | 1 |
| Site Lower-outer quadrant | 0.047 | 0.504 | 0.259 | 1.01 |
| Node examined 2 | <0.01 | 1.27 | 1.11 | 1.44 |
| Node examined 3 | <0.01 | 1.52 | 1.33 | 1.74 |
| Node examined 4 | <0.01 | 1.77 | 1.51 | 2.07 |
| Node examined 5 | <0.01 | 2.22 | 1.86 | 2.66 |
| ER Positive | <0.01 | 0.704 | 0.571 | 0.866 |
| PR Positive | <0.01 | 0.668 | 0.568 | 0.783 |
| HER2 Positive | 0.071 | 1.15 | 0.989 | 1.33 |
| T2 | <0.01 | 2.68 | 2.41 | 2.9 |
| T3 | <0.01 | 3.49 | 2.5 | 4.85 |
| Grade II | <0.01 | 1.56 | 1.38 | 1.75 |
| Grade III | <0.01 | 1.65 | 1.42 | 1.92 |
SLNM = sentinel lymph node metastasis.
3.3. Prognostic factors of patients in development cohort
In multivariate COX analysis, 8 prognostic factors were selected, including age, race, histological grade, radiotherapy, tumor size, ER status, SLN positive status and marital status. The 8 independent factors were identified (P ≤ .05): Age: Age 60 to 79 years, Age ≥ 80 years; Race: Black, American Indian/Alaska Native (AI), Asian or Pacific Islander (API); Histological grading: Grade II, Grade III; No radiotherapy; Tumor size: T2, T3; ER positive: sentinel lymph node positive, married (Table 4).
Table 4.
Univariate and multivariate regression analyses for BCSS.
| Characteristics | Univariate analysis HR (95% CI) | P value | Multivariate analysis HR (95% CI) | P value |
|---|---|---|---|---|
| Age | ||||
| Age 20–39 yr | Ref. | Ref. | ||
| Age 40–59 yr | 0.888 (0.482–1.634) | .702 | 1.167 (0.632–2.155) | .595 |
| Age 60–79 yr | 1.806 (0.994–3.282) | .052 | 2.489 (1.357–4.564) | .002 * |
| Age ≥ 80 yr | 6.357 (3.463–11.670) | <.001* | 6.536 (3.487–12.250) | <.001* |
| Race | ||||
| White | Ref. | Ref. | ||
| Black | 1.647 (1.363–1.989) | <.001* | 1.454 (1.194–1.770) | <.001* |
| AI | 1.776 (0.921–3.428) | .087 | 2.061 (1.065–3.986) | .035* |
| API | 0.650 (0.483–0.872) | .004* | 0.738 (0.548–0.993) | .041* |
| Site | ||||
| Nipple | Ref. | |||
| Upper-inner quadrant | 0.531 (0.219–1.292) | .163 | ||
| Lower-inner quadrant | 0.639 (0.260–1.573) | .33 | ||
| Upper-outer quadrant | 0.611 (0.253–1.475) | .273 | ||
| Lower-outer quadrant | 0.560 (0.228–1.376) | .206 | ||
| Histologic type | ||||
| Others | Ref. | |||
| No special type | 1.148 (0.973–1.355) | .103 | ||
| Grade | ||||
| Grade I | Ref. | Ref. | ||
| Grade II | 1.311 (1.100–1.562) | .002* | 1.209 (1.011–1.445) | .0344 * |
| Grade III | 2.329 (1.951–2.782) | <.001* | 2.031 (1.647–2.504) | <.001* |
| Laterality | ||||
| Left | Ref. | |||
| Right | 0.975 (0.855 –1.112) | .701 | ||
| Radiation | ||||
| Yes | Ref. | Ref. | ||
| No | 2.362 (2.044–2.728) | <.001* | 1.812 (1.560–2.104) | <.001* |
| Chemotherapy | ||||
| Yes | Ref. | |||
| No | 0.947 (0.822–1.090) | .448 | ||
| T | ||||
| T1 | Ref. | Ref. | ||
| T2 | 1.833 (1.593–2.109) | <.001* | 1.498 (1.287–1.744) | <.001* |
| T3 | 2.519 (1.674–3.789) | <.001* | 2.277 (1.492–3.475) | <.001* |
| ER | ||||
| Negative | Ref. | Ref. | ||
| Positive | 1.917 (1.627–2.259) | <.001* | 1.387 (1.085–1.773) | .009* |
| PR | ||||
| Negative | Ref. | Ref. | ||
| Positive | 1.67 (1.447–1.927) | <.001* | 1.170 (0.953–1.436) | .162 |
| HER2 | ||||
| Negative | Ref. | |||
| Positive | 1.01 (0.819–1.244) | .927 | ||
| Node Examined | ||||
| Ref. | ||||
| 2 | 0.994 (0.840–1.177) | .947 | ||
| 3 | 0.944 (0.782–1.139) | .549 | ||
| 4 | 1.075 (0.863–1.339) | .518 | ||
| 5 | 0.864 (0.650–1.149) | .316 | ||
| Node positive | ||||
| Negative | Ref. | Ref. | ||
| Positive | 1.502 (1.294–1.743) | <.001* | 1.4981 (1.281–1.752) | <.001* |
| Marital | ||||
| Single | Ref. | Ref. | ||
| Married | 0.669 (0.549–0.816) | <.001* | 0.719 (0.586–0.881) | .001* |
| SDW | 1.531 (1.254–1.870) | <.001* | 1.149 (0.932–1.415) | .193 |
AI = American Indian/Alaska Native, API = Asian or Pacific Islander, SDW = separated, divorced or widowed.
3.4. Prognostic nomogram for CSS
According to the COX regression analysis, nomogram predicted 3-year and 5-year CSS for breast cancer patients (Fig. 1). As a result of the contribution to the nomogram, a corresponding score is assigned to all variables in the nomogram, ranging from 0 to 100. Patients can get an overall score by adding the scores for each subgroup.
Figure 1.
Nomogram predicting 3-year and 5-year cancer-specific survival probability for BCSS.
3.5. Feasibility of the nomogram
The C index was 0.807 (0.782–0.832) In the development cohort. At the same time, receiver operating characteristic curve was used to evaluate the discriminant ability of the model. AUC values were significantly higher for both 3-year (0.816) and 5-year (0.806) forecasts (Fig. 2A and B). According to the calibrated graphs for both the 3-year and 5-year development cohorts, actual observations and predictions are in good agreement (Fig. 3A and B), As a result, we can conclude that our model has relatively good performance. In addition, A validation queue was used to evaluate the nomogram’s applicability. The C index is 0.826 (0.790–0.863), AUC values were significantly higher for both 3-year (0.816) and 5-year (0.806) forecasts (Fig. 2C and D). The calibration curves of the verification queue predict the results well and are in good agreement with the actual results. Results from internal validation indicated that the diagram was of satisfactory applicability to patients with SLNB (Fig. 3C and D).
Figure 2.
ROC curves of the nomogram predicting 3-year (A) and 5-year (B) BCSS in the development cohort; 3-year (C) and 5-year (D) BCSS in the validation cohort. ROC = receiver operating characteristic.
Figure 3.
C-index of the nomogram predicting 3-year (A) and 5-year (B) BCSS in the development cohort; 3-year (C) and 5-year (D) BCSS in the validation cohort.
3.6. Survival curve for nomogram
We followed the patients for a median of 57 months (0–71 months). We observed that 889 patients died, including 304 breast cancer specific deaths. The 3-year BCSS rates were 98.5%, and the 5-year BCSS rates were 97.6%. In order to observe the survival of different numbers of SLNM, BCSS curves of different positive SLNs were calculated using the Kaplan–Meier curve (Fig. 4). We observed significant differences between different amounts of SLNM (P < .001)
Figure 4.
BCSS curves of different positive SLNs were calculated using the Kaplan–Meier cure.
4. Discussion
Based on the large clinical studies represented by SENTINA and ACOSOG Z0011, breast conserving surgery combined with SLNB has become the mainstream operation at present, which has been widely recognized by clinicians. The latest guidelines recommend that patients with one or two SLNM who have received breast-conserving therapy and postoperative radiation therapy should avoided ALND.[16] At present, the risk factors and prognosis of SLNM are analyzed based on the study of axillary lymph node metastasis, But it is unclear whether there is difference. Therefore, an accurate and appropriate predictive model needs to be proposed for the evaluation of breast cancer patients performing SLNB.
Comprehensive analysis of all the available factors in the SEER database was conducted. The rate of SLNM was 19.7% in the study. Among the samples with SLNM, the proportion of biopsy was 23.6% for one SLN, 27.8% for 2 SLNs, 23.2% for SLNs, 14.7% for SLNs, and 10.7% for 5 SLNs. The results showed that histological grade, tumor location, T stage, ER status, PR status, and the number of SLNs were significantly correlated with SLNM. Except for the number of SLN, the results are similar to those of domestic and foreign studies on axillary lymph node metastasis.[17,18] Although the false negative rate of one SLN biopsy during SLNB operation was less than 10%,[11,19] in this study, The number of 2 to 5 SLNs is an independent risk factor for positive SLN by logistic regression, and according to the NCCN(National Comprehensive Cancer Network) guidelines, breast conserving breast cancer patients with one to two SLNM can be exempted from ALND surgery. Therefore, when performing SLNB surgery, we should biopsy two or more SLNs to accurately predict whether patients have non-SLNM and whether ALND surgery can be exempted.
The latest prospective trial[7]showed that after neoadjuvant chemotherapy, The false negative rate of SLN was < 10% when there were 3 or more SLNs. This also has reference value for the results of this study. The combined tracer technique is the gold standard for SLNB and can achieve a detection rate of more than 95% and a false negative rate of less than 10%.[20] Therefore, the use of dual tracer technology can improve the accuracy of our experiment.
We constructed nomogram using variables from the multifactor COX model and used it to predict BCSS. With this approach, an accurate tool was produced that accurately included only variables associated with survival. the survival nomogram was successfully constructed with relatively good predictability. another advantage of the nomogram compared with the multiple regression is that it provides the probability of individual survival outcomes at a specific point in time, rather than the relative risk concept. Meanwhile, compared with the traditional COX regression model,[21,22] nomogram’s accuracy can also be evaluated using Harrell’s C-index.
In the multivariate COX model, age, ethnicity, histological grade, radiotherapy, tumor size, ER status, SLNM status and marital status were independent prognostic factors for BCSS. Although the ACOSOG Z0011 trial confirmed that there was no significant difference in 10-year survival rate between SLNB and ALND for patients with 1 to 2 SLNM, N1 included 1 to 3 SLNM patients according to the eighth edition of AJCC breast cancer Staging Guidelines.[10] Therefore, We also studied the effect of more than two SLNs positive on survival and prognosis of patients. Interestingly, the risk score of 3 SLNs was slightly lower than that of one SLN in nomogram, and the Kaplan–Meier method obtained different BCSS curves with positive SLNs. Kaplan–Meier method was used to compare the survival curves of 1 to 2 SLNM patients with 3 SLNM patients, 3 SLNM patients with 4 to 5 SLNM patients, respectively. During the follow-up period, there was no significant difference in the survival rate between patients with 3 SLNM and those with 1 to 2 SLNM (P > .05) (Fig. 4A), while there was a significant difference between patients with 3 SLNMs and those with 4 to 5 SLNMs (P < .05) (Fig. 4B). According to the diagram, The survival rate of patients who had 3 SLNM were not significantly different from that with one or two SLNM. So are N1 breast cancer patients exempt from ALND surgery?C. Bonneau et all found that patients with T1T2 invasive breast cancer with 3 lymph node metastases did not benefit from ALND after SLNB, and ALND was limited to staging.[23] Yun Fu et al[24] suggested that radiotherapy after SLNB could replace ALND in patients with N1 breast cancer. The results showed that patients with stage N1 could be exempted from ALND after SLNB. The difference between the results of other studies and the ACOSOG Z0011 trial may be due to the fact that patients with 1-2 SLNM were included in the ACOSOG Z0011 trial, and patients with 3 SLNM and only SLNB were not included in the study. However, ACOSOG Z0011 is a prospective study. Other studies are retrospective and have certain limitations. At the same time, the AMAROS trial confirmed that radiotherapy can achieve the same control effect as ALND in SLNM cT1-2 breast cancer patients.[25] However, we generally believe that the more lymph node metastases, the worse the prognosis.[26] Therefore, we should conduct prospective studies for further verification.
This study uses SEER database to provide a large sample for analysis, but it still has drawbacks. First, the results are inevitably affected by selection bias. For example, the SEER database collects a large number of patients information from multiple regions and hospitals. Doctors have certain differences in the treatment methods of patients, such as the dosage of therapeutic drugs and radiotherapy may be different. Lastly, even though internal validation was performed, As a result of using the same database for both development and validation, the results were not perfect. For external validation, a large prospective clinical trial is required.
5. Conclusion
This SEER database-based study revealed demographic, clinicopathological and therapeutic characteristics that were significantly associated with specific survival of breast cancer patients undergoing sentinel node biopsy. We constructed and validated prognostic nomogram to predict individualized probabilities of 3-year and 5-year specific survival in breast cancer patients. Nomogram facilitates patient consultation, follow-up planning and treatment selection. We also concluded that patients with stage N1 breast cancer were exempt from axillary lymph node dissection. However, Whether the clinical operation can be downgraded needs to be validated prospectively.
Author contribution
Data curation: Ruihao Liu, Yulong Liao.
Formal analysis: Jian Chen, Ting Li, Yulong Liao.
Investigation: Ting Li.
Methodology: Wei Cao, Yingliang Li.
Software: Ruihao Liu, Wei Cao.
Supervision: Yingliang Li.
Visualization: Jian Chen.
Writing – original draft: Ruihao Liu.
Writing – review & editing: Yingliang Li.
Abbreviations:
- CI
- confidence interval
- C-index
- concordance index
- COX
- proportional hazards model
- ER
- estrogen receptor
- HR
- hazard ratio
- PR
- progesterone receptor
- SEER
- the Surveillance Epidemiology and End Results
- SLN
- sentinel lymph node
- SLNB
- sentinel lymph node biopsy
- SLNM
- sentinel lymph node metastasis
This study used anonymized and de-identified data previously collected from the SEER database. Therefore, no additional ethical approval or consent is required.
The authors have no conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are publicly available.
How to cite this article: Liu R, Chen J, Cao W, Li T, Liao Y, Li Y. Risk factors and prognosis of sentinel lymph node metastasis in breast-conserving breast cancer: A retrospective study based on the SEER database. Medicine 2024;103:9(e37263).
Contributor Information
Ruihao Liu, Email: 18779105145@163.com.
Jian Chen, Email: chenjiancj2023@163.com.
Wei Cao, Email: drcaoweidh@163.com.
Ting Li, Email: xionghwgj@163.com.
Yulong Liao, Email: drliuzgdh@163.com.
References
- [1].Cabanas RM. An approach for the treatment of penile carcinoma. Cancer. 1977;39:456–66. [DOI] [PubMed] [Google Scholar]
- [2].Habal N, Giuliano AE, Morton DL. The use of sentinel lymphadenectomy to identify candidates for postoperative adjuvant therapy of melanoma and breast cancer. Semin Oncol. 2001;28:41–52. [DOI] [PubMed] [Google Scholar]
- [3].Harlow SP, Krag DN, Julian TB, et al. Prerandomization surgical training for the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-32 trial: a randomized phase III clinical trial to compare sentinel node resection to conventional axillary dissection in clinically node-negative breast cancer. Ann Surg. 2005;241:48–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Donker M, Van Tienhoven G, Straver ME, et al. Radiotherapy or surgery of the axilla after a positive sentinel node in breast cancer (EORTC 10981-22023 AMAROS): a randomised, multicentre, open-label, phase 3 non-inferiority trial. Lancet Oncol. 2014;15:1303–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Galimberti V, Cole BF, Viale G, et al.; International Breast Cancer Study Group Trial 23-01. Axillary dissection versus no axillary dissection in patients with breast cancer and sentinel-node micrometastases (IBCSG 23-01): 10-year follow-up of a randomised, controlled phase 3 trial. Lancet Oncol. 2018;19:1385–93. [DOI] [PubMed] [Google Scholar]
- [6].Giuliano AE, Ballman KV, Mccall L, et al. Effect of axillary dissection vs no axillary dissection on 10-year overall survival among women with invasive breast cancer and sentinel node metastasis: the ACOSOG Z0011 (Alliance) randomized clinical trial. JAMA. 2017;318:918–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Barrio AV, Montagna G, Mamtani A, et al. Nodal recurrence in patients with node-positive breast cancer treated with sentinel node biopsy alone after neoadjuvant chemotherapy-a rare event. JAMA Oncol. 2021;7:1851–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Collins M, O’donoghue C, Sun W, et al. Use of axillary lymph node dissection (ALND) in patients with micrometastatic breast cancer. J Surg Res. 2017;215:55–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Mansel RE, Fallowfield L, Kissin M, et al. Randomized multicenter trial of sentinel node biopsy versus standard axillary treatment in operable breast cancer: the ALMANAC Trial. J Natl Cancer Inst. 2006;98:599–609. [DOI] [PubMed] [Google Scholar]
- [10].Amin MB, Greene FL, Edge SB, et al. The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin. 2017;67:93–9. [DOI] [PubMed] [Google Scholar]
- [11].Kuehn T, Bauerfeind I, Fehm T, et al. Sentinel-lymph-node biopsy in patients with breast cancer before and after neoadjuvant chemotherapy (SENTINA): a prospective, multicentre cohort study. Lancet Oncol. 2013;14:609–18. [DOI] [PubMed] [Google Scholar]
- [12].Krag DN, Anderson SJ, Julian TB, et al.; National Surgical Adjuvant Breast and Bowel Project. Technical outcomes of sentinel-lymph-node resection and conventional axillary-lymph-node dissection in patients with clinically node-negative breast cancer: results from the NSABP B-32 randomised phase III trial. Lancet Oncol. 2007;8:881–8. [DOI] [PubMed] [Google Scholar]
- [13].Dixon JM, Grewar J, Twelves D, et al. Factors affecting the number of sentinel lymph nodes removed in patients having surgery for breast cancer. Breast Cancer Res Treat. 2020;184:335–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Nattinger AB, Mcauliffe TL, Schapira MM. Generalizability of the surveillance, epidemiology, and end results registry population: factors relevant to epidemiologic and health care research. J Clin Epidemiol. 1997;50:939–45. [DOI] [PubMed] [Google Scholar]
- [15].Ahmed AR. HER2 expression is a strong independent predictor of nodal metastasis in breast cancer. J Egypt Natl Canc Inst. 2016;28:219–27. [DOI] [PubMed] [Google Scholar]
- [16].Gradishar WJ, Anderson BO, Abraham J, et al. Breast cancer, version 3.2020, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2020;18:452–78. [DOI] [PubMed] [Google Scholar]
- [17].Kasangian AA, Gherardi G, Biagioli E, et al. The prognostic role of tumor size in early breast cancer in the era of molecular biology. PLoS One. 2017;12:e0189127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Yun SJ, Sohn YM, Seo M. Risk stratification for axillary lymph node metastases in breast cancer patients: what clinicopathological and radiological factors of primary breast cancer can predict preoperatively axillary lymph node metastases? Ultrasound Q. 2017;33:15–22. [DOI] [PubMed] [Google Scholar]
- [19].Zetterlund L, Celebioglu F, Axelsson R, et al. Swedish prospective multicenter trial on the accuracy and clinical relevance of sentinel lymph node biopsy before neoadjuvant systemic therapy in breast cancer. Breast Cancer Res Treat. 2017;163:93–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Peek MC, Charalampoudis P, Anninga B, et al. Blue dye for identification of sentinel nodes in breast cancer and malignant melanoma: a systematic review and meta-analysis. Future Oncol. 2017;13:455–67. [DOI] [PubMed] [Google Scholar]
- [21].Wu P, Zhao K, Liang Y, et al. Validation of breast cancer models for predicting the nonsentinel lymph node metastasis after a positive sentinel lymph node biopsy in a Chinese population. Technol Cancer Res Treat. 2018;17:1533033818785032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Shen K, Yao L, Wei J, et al. Worse characteristics can predict survival effectively in bilateral primary breast cancer: a competing risk nomogram using the SEER database. Cancer Med. 2019;8:7890–902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Bonneau C, Hequet D, Estevez JP, et al. Impact of axillary dissection in women with invasive breast cancer who do not fit the Z0011 ACOSOG trial because of three or more metastatic sentinel lymph nodes. Eur J Surg Oncol. 2015;41:998–1004. [DOI] [PubMed] [Google Scholar]
- [24].Hunt KK, Yi M, Mittendorf EA, et al. Sentinel lymph node surgery after neoadjuvant chemotherapy is accurate and reduces the need for axillary dissection in breast cancer patients. Ann Surg. 2009;250:558–66. [DOI] [PubMed] [Google Scholar]
- [25].Yan M, Abdi MA, Falkson C. Axillary management in breast cancer patients: a comprehensive review of the key trials. Clin Breast Cancer. 2018;18:e1251–9. [DOI] [PubMed] [Google Scholar]
- [26].Uehiro N, Horii R, Iwase T, et al. Validation study of the UICC TNM classification of malignant tumors, seventh edition, in breast cancer. Breast Cancer. 2014;21:748–53. [DOI] [PubMed] [Google Scholar]