Abstract
Background
Neoadjuvant chemotherapy (NAC) is considered disruptive for axillary nodal staging in breast cancer patients, by altering the total number of excised nodes, and the number of positive and negative nodes. We aimed to identify the axillary variable having the strongest association with survival, and also to compare the prognostic efficacy of the pathological node (pN) staging versus the lymph node ratio (LNR) systems in post-NAC cases.
Methods
In this retrospective single-center study of 169 post-neoadjuvant chemotherapy breast cancer patients, three absolute axillary nodal counts (positive [P], negative [N], and total excised nodes [T]) and six derived ratios (P/T, N/T, P/N, N/P, T/P, and T/N) were evaluated. Associations with survival and relative prognostic strength were assessed using univariate, bivariate, and multivariate Cox regression models. The prognostic performance of pathological node (pN)-staging and lymph node ratio (LNR) systems was evaluated using Kaplan–Meier survival analysis.
Results
The N, T, and T/P and N/P values had no significant associations with survival, while the P value had the strongest association with survival (p < 0.0001). The ranking of axillary variables based on their strength of association with survival was: P > P/N (or T/N) > P/T (or N/T). Both the pN-staging and LNR systems and also their respective bases (the P and P/T values) were significantly associated with survival (respectively: p = 0.005, p = 0.041, p < 0.0001, p = 0.006).
Conclusion
Both the pN- and LNR staging systems can significantly classify post-NAC patients into survival groups, while the pN-staging system showed superior credentials to identify high-risk cases.
Trial registration
Not applicable, as this is a retrospective observational study and not a clinical trial.
Keywords: Post neo-adjuvant chemotherapy, Axillary staging, Pathological node staging, PN-staging, Lymph node ratio, LNR, Node factors
Introduction
The two axillary staging systems currently used for breast cancer (BC) are ‘pathological node staging’ (pN-staging), part of the global TNM system1, and the ‘lymph node ratio’ (LNR) system [1, 2]. The pN-staging system is based on the “absolute number of positive (involved) lymph nodes (LNs)” from the excised axillary specimen, by which the patients are categorized into 4 stages: patients with no axillary involvement are staged as pN0, those with 1 to 3 positive LNs as pN1, 4 to 9 as pN2, and ≥ 10 as pN3 [3]. The LNR system on the other hand is based on the “ratio of positive to total excised LNs” which produces a decimal quantity [2]. In this system, patients with the ratio equal to zero are staged as LNR0, those with 0 < LNR ≤ 0.2 as low-risk, 0.2 < LNR ≤ 0.65 as intermediate-risk, and LNR > 0.65 as high-risk [4]. The two abovementioned fundamental quantities (or bases) are not the only extractable axillary metrics, as several other measures are also extractable from the axilla that might have significant associations with the post-NAC survival. Axillary staging systems such as pN-staging and the lymph node ratio (LNR) have been widely studied for their prognostic significance in breast cancer patients after neoadjuvant chemotherapy. In parallel, recent studies have highlighted the role of molecular and immunological markers in refining prognostic assessment, such as miR-584-5p regulation of MSMO1 [5] and CD1B expression in triple-negative breast cancer [6]. These findings underscore the broader effort to integrate clinical staging systems with emerging biological markers.
Theoretically, 9 axillary variables including 3 absolute numbers and 6 ratios can be drawn out of the crude axillary pathological results, and of course, all are alterable by the effect of Neo-Adjuvant Chemotherapy (NAC), but we are to see which maintains its association with survival in post-NAC conditions. They are: 1.the Total number of excised LNs ( the T value); 2. The absolute number of Positive LNs (the P value) 2; 3. The absolute number of Negative LNs (the N value); 4. The Positive to Total excised LN ratio (the P/T value); 5. The Negative to Total LN ratio (the N/T value); 6. The Positive to Negative LN ratio (the P/N value); 7. The Negative to Positive LN ratio (the N/P value); 8. The Total to Positive LN ratio (the T/P value); and 9. The Total to Negative LN ratio (the T/N value).
The primary rationale behind the use of NAC in breast cancer maybe is to shrink the mass size of the primary tumor to operable conditions [7]. Therefore, the NAC reduces the rate of mastectomies since breast conserving procedures become feasible for more patients [7]. Despite the many therapeutic assets of the NAC, when facing with the staging of the disease, it is broadly believed to be highly disruptive, as it directly targets the tumor cells, not only in the primary tumor, but in the axillary LNs, and even in the metastases [8–22]. Accordingly, the NAC may potentially alter the pre-NAC tumor [23], node, and metastasis stages of the disease, and thus the predictability of survival functions by staging systems might be questioned.
This study aimed to evaluate the association between nine axillary nodal variables and post-neoadjuvant chemotherapy breast cancer survival, and to identify the nodal metric with the strongest prognostic value. Based on these analyses, the prognostic performance of the pathological node (pN) staging and lymph node ratio (LNR) systems was compared to determine their relative ability to stratify post-NAC patients according to survival risk.
Patients and methods
Study design
This was a retrospective cohort study using data from the registry of the Kaviani Breast Disease Institute (KBDI), a single-center database in Tehran, Iran. Patients were accrued between 2008 (start) and 2018 (end). No separate development and validation datasets were used, as the study focused on validating existing prognostic models (pN-staging and LNR systems) in a single cohort of post-neoadjuvant chemotherapy breast cancer patients. No formal sample size calculation was performed; the sample size was determined by the number of eligible patients available in the registry during the study period. From a total of 1,497 registered breast cancer patients at the KBDI in that period, 235 had received neoadjuvant chemotherapy. Of these, 31 were excluded due to missing survival data, 21 due to sentinel lymph node biopsy alone, and 14 due to missing data on positive or total excised lymph nodes. The final study cohort therefore comprised 169 patients who met all inclusion criteria.
The included patients were known cases with breast carcinomas who were indicated for breast surgery in addition to axillary evaluation and had pathological reports on the excised axillary specimen. The primary tumor was approached using various surgical techniques, including radical mastectomies to diverse subsets of breast‑conserving and oncoplastic surgeries. Patients with clinically positive axillary LNs underwent the classic axillary LN dissection (ALND), while those with clinically negative LNs had an initial sentinel LN biopsy (SLNB), and if the sentinels were involved, then an additional ALND was also performed. The excluded patients from the series were those exempted from NAC, those to whom the Z‑0011 trial findings were applied (patients with less than two involved sentinels who had no further ALND), patients having metastases at diagnosis, cases having inflammatory breast disease or a second primary cancer, and finally, those with deficient follow‑up data. All the study populations were operated on by the senior author of this article, Prof. Kaviani, in either affiliate hospitals of the Tehran University of Medical Sciences or other private centers in Tehran, Iran.
Data collection
We collected the data on patient’s age, tumor size, pT-stage, histology of cancer, nuclear and histologic grades, state of lymph vascular invasion (LVI), state of estrogen and progesterone receptors expression (ER and PR, respectively), state of HER2neu receptor expression, p53 status, and Ki67 levels (both as continuous and categorical: ≤14% and > 14%) as baseline clinicopathological characteristics (CPC). The treatment features (TF) considered were the status and number of NAC delivery sessions, and adjuvant chemo-, radio-, and hormone-therapy. For neoadjuvant chemotherapy, we recorded whether NAC was administered and the number of sessions delivered for each patient. Detailed regimen composition was not consistently available across the study period, and this limitation has been acknowledged.
Start-point and end-point
For survival analyses, the date of surgery admission was set as the start‑point. The primary outcome predicted by the nodal staging models and variables was Breast Cancer-Free Survival (BCFS), defined as the time from surgery to death from any cause, local recurrence, regional recurrence, or distant metastasis. Local recurrence was defined as tumor reappearance in the ipsilateral breast or chest wall, regional recurrence as involvement of ipsilateral axillary, supraclavicular, or internal mammary lymph nodes, and distant metastasis as spread to organs beyond locoregional sites. Patients who did not experience events during the follow‑up period were censored at the last known alive date. Outcome events (death, local/regional recurrence, or distant metastasis) were assessed retrospectively from clinical records and follow-up data without blinding of assessors to the predictors (axillary nodal metrics and staging), as this was an observational study using existing registry data.
Statistical analyses
Theoretically, 9 node-derived variables (NDVs) may be extracted from the excised axillary nodal specimen, two of which are the “absolute number of positive nodes” and the “ratio of positive to total LNs” that respectively are foundations of pN-staging and LNR systems. As the introduction section explained these variables are the T, P, N, P/T, N/T, P/N, N/P, T/P, and T/N values, which make distinct quantities as numbers and ratios with different associations with BCFS. On the other hand, linear mathematical dependencies are expected between 3 pairs of the ratios because of the following equations: N/T = 1-P/T, T/N = 1 + P/N, and T/P = 1 + N/P. This makes them interchangeable in Cox regression models. The effects of the 9 NDVs and the clinicopathological and treatment factors on BCFS were individually assessed using univariate Cox models, and p-values ≤ 0.05 were considered to be significant. Because numerous univariate Cox models were performed across node-derived, clinicopathological, and therapeutic variables, these analyses were considered exploratory. In addition, a Benjamini–Hochberg false discovery rate (FDR) adjustment was applied to the nine node-derived variables to evaluate the robustness of their prognostic significance.
To compare the effects of NDVs on BCFS, a multivariate Cox model was made by including the variables having significant effects on survival, which resulted from the univariate analyses. In this model, the variable(s) remaining significantly associated with survival was considered independent and superior to its covariates which lost their significance. To determine the hierarchical order of NDVs based on the strength of their association with survival, we made multiple bivariate Cox models, in each of which a pair of NDVs were included. Again, the one remaining significant was considered superior to its covariate, and the one losing significance was regarded as inferior. The impact of CPCs and TFs on BCFS was also investigated using univariate Cox models, and their strength of association and level of independence were compared in a multivariate Cox model. Finally, to identify the variable having the strongest and most independent association with survival, a multivariate Cox model was made when all NDVs, CPCs, and TFs with significant associations with survival were included as covariates.
The pN-staging and LNR systems are node staging systems, respectively originating from the P and P/T values. These systems, actually convert the P and P/T values from continuous to categorical variables by defining thresholds in their quantities. We used Kaplan-Meier (KM) survival estimation analysis to compare systems’ prediction capabilities where the accuracy of patients’ categorizations into survival groups was evaluated by log rank test, and the 5 year survival estimation for the categories was calculated.
No imputation methods were applied in this study. Core prognostic variables, including axillary nodal metrics and staging system categorizations (pathological node staging and lymph node ratio), were complete for all included cases. For other clinicopathological variables with occasional missing values (e.g., Ki-67), analyses were performed using available data only, according to the default case-wise handling of missing data in SPSS.
Results
Descriptive results
A total of 169 women with pathologically confirmed breast carcinomas, all of whom received neoadjuvant chemotherapy (median 6 sessions, range: 2–16), were enrolled in the study. Patients’ mean age was 47.65yrs (Std. D.3: 10.11) ranging from 28 to 74yrs. Forty-four cases (26%) underwent breast conserving surgeries and the rest 125 (74%) underwent mastectomies. For the axillary workups, 19 (11.2%) underwent SLNB alone (those whose SLNB results came out negative), 26 (15.4%) underwent an initial SLNB proceeded by a completion ALND (those with positive SLNBs), and 124 cases (73.4) had ALND alone (those with clinically positive axillary involvement). The median total number of excised LNs by the axillary dissections was 8 (range: 1–29), while the median absolute number of positive nodes was 1 (range: 0–23). The mean diameter of the primary tumor was 27.9 mm (Std. D.: 17.67, range: 1-100). Post-surgically, the patients were followed for a median 513 days (range 23-2934), during which 44 (26%) experienced events. The baseline clinicopathological characteristics of the disease and also treatment details are presented in Table 1.
Table 1.
Clinicopathological and treatment characteristics
| Characteristic | Frequency (%) | Characteristic | Frequency (%) |
|---|---|---|---|
| Histopathology of cancer | P53 | ||
| IDC | 153 (90.5) | Negative | 24 (14.2) |
| ILC | 5 (3.0) | Positive | 28 (16.6) |
| IDC + ILC | 4 (2.4) | N/A | 117 (69.2) |
| IDC + Paget | 1 (0.6) | ||
| Other | 6 (3.6) | ||
| TNM Tumor Sizing (T stage) | Ki67 | ||
| T1 (≤ 20 mm) | 60 (35.5) | ≤ 14% | 39 (23.1) |
| T2 (> 20, ≤ 50 mm) | 66 (39.1%) | > 14% | 63 (37.3) |
| T3 (> 50 mm) | 9 (5.3) | N/A | 67 (39.6) |
| N/A | 34 (20.1) | Median | 19.5% |
| Range | 2–90% | ||
| Histologic Grade | Neo-Adjuvant Chemotherapy | ||
| Grade I | 13 (7.7) | Positive | 169 (100) |
| Grade II | 91 (53.8) | Mode (sessions) | 6 |
| Grade III | 26 (15.4) | Range | 2–16 |
| N/A | 39 (23.1) | Negative | 0 |
| N/A | 0 | ||
| Nuclear Grade | Adjuvant Chemotherapy | ||
| Grade I | 7 (4.1) | Positive | 64 (37.9) |
| Grade II | 84 (49.7) | Mode (sessions) | 4 |
| Grade III | 30 (17.8) | Range | 1–12 |
| N/A | 48 (28.4) | N/A | 105 (62.1) |
| Lympho-Vascular Invasion | Adjuvant Radiotherapy | ||
| Negative | 51 (30.2) | Positive | 143 (84.6) |
| Positive | 84 (49.7) | Mode (sessions) | 25 |
| N/A | 34 (20.1) | Range | 16–35 |
| N/A | 26 (15.4) | ||
| Estrogen Receptor | Hormone Therapy | ||
| Negative | 48 (28.4) | Positive | 94 (55.6) |
| Positive | 109 (64.5) | Tamoxifen | 64 (37.9) |
| N/A | 12 (7.1) | Letrozole | 16 (9.5) |
| Progesterone Receptor | Tamoxifen + Dipherline | 5 (3.0) | |
| Negative | 63 (37.3) | Tamoxifen + Letrozole | 4 (2.4) |
| Positive | 94 (55.6) | Exemestane | 4 (2.4) |
| N/A | 12 (7.1) | Tamoxifen + Letrozole + Dipherline | 1 (0.6) |
| Negative | 2 (1.2) | ||
| N/A | 73 (43.2) | ||
| Her2-neu Receptor | Targeted Therapy (Trastuzumab) | ||
| Negative (0 and 1+) | 90 (53.3) | Positive | 39 (23.1) |
| Positive (2 + and 3+) | 48 (28.4) | Negative | 1 (0.6) |
| N/A | 31 (18.3) | N/A | 129 (76.3) |
Univariate results
The status of axillary nodal involvement in post-NAC patients (node positivity versus negativity), as disclosed by the univariate Cox analysis, was significantly and directly associated with survival (β=+0.905, HR = 2.472, 95% CI = 1.148–5.326, p = 0.021). The effect of node-derived metrics and also clinicopathological factors on the post-NAC patients’ survival are addressed in Table 2. Among the NDVs, the P, P/T, N/T (linearly dependent with P/T regarding the N/T = 1-P/T equation), P/N, and T/N (linearly dependent with P/N respecting the T/N = 1 + P/N equation) had significant associations with event occurrences; while among the CPCs, the tumor diameter, the pT-stage, and the ER status, and among the TFs, the number of radiotherapy sessions, and the status of hormone-therapy had significant associations with BCFS (Table 2). To assess the stability of the NDV associations, a Benjamini–Hochberg FDR adjustment was applied to the nine node-derived variables. The key predictors (P, P/N, T/N, P/T, and N/T) remained statistically significant after correction.
Table 2.
Impact of node-derived, clinicopathological, and therapeutic factors on post-NAC patients’ survival (BCFS)
| Factor | β Coefficient | Hazard Ratio (95% CI) | p-value | |
|---|---|---|---|---|
| Node-derived Factors | Total number of excised LNs (T) | + 0.058 | 1.060 (0.999–1.124) | 0.053 |
| Absolute number of positive LNs (P) | + 0.112 | 1.119 (1.057–1.184) | < 0.000 | |
| Absolute number of negative LNs (N) | -0.040 | 0.961 (0.895–1.032) | 0.276 | |
| Positive to total LN ratio (P/T) | + 1.166 | 3.208 (1.401–7.345) | 0.006 | |
| Negative to total LN ratio (N/T) | -1.166 | 0.312 (0.136–0.714) | 0.006 | |
| Positive to negative LN ratio (P/N) | + 0.141 | 1.151 (1.071–1.238) | < 0.000 | |
| Negative to positive LN ratio (N/P) | -0.062 | 0.940 (0.833–1.060) | 0.313 | |
| Total to positive LN ratio (T/P) | -0.062 | 0.940 (0.833–1.060) | 0.313 | |
| Total to negative LN ratio (T/N) | + 0.141 | 1.151 (1.071–1.238) | < 0.000 | |
| Clinicopathological Factors | Age (as continuous) | -0.012 | 0.988 (0.958–1.019) | 0.449 |
| Type of breast surgery | ||||
| BCS vs. Mastectomy | + 0.054 | 1.055 (0.504–2.210) | 0.886 | |
| Type of axillary surgery | 0.805 (overall) | |||
| SLNB vs. ALND | + 0.227 | 1.255 (0.381–4.136) | 0.709 | |
| SLNB vs. SLNB + ALND | + 0.469 | 1.598 (0.381-6.712) | 0.522 | |
| Tumor Histology | 0.998 (overall) | |||
| IDC vs. IDC + ILC | + 0.158 | 1.171 (0.160–8.558) | 0.876 | |
| IDC vs. ILC | + 0.141 | 1.151 (0.157–8.461) | 0.890 | |
| IDC vs. IDC + Paget | -9.002 | 0.000 (0.000) | 0.981 | |
| Tumor Size (as continuous) | + 0.015 | 1.016 (1.001–1.030) | 0.039 | |
| pT-stage (TNM tumor sizing) | 0.020 (overall) | |||
| T1 vs. T2 | + 1.127 | 3.566 (1.466–8.674) | 0.005 | |
| T1 vs. T3 | + 1.043 | 2.837 (0.790-10.183) | 0.110 | |
| Histologic Grade | 0.168 (overall) | |||
| Grade I vs. II | + 0.519 | 1.681 (0.395–7.147) | 0.482 | |
| Grade I vs. III | + 1.095 | 2.990 (0.668–13.391) | 0.152 | |
| Nuclear Grade | 0.957 (overall) | |||
| Grade I vs. II | + 0.047 | 1.048 (0.246–4.470) | 0.949 | |
| Grade I vs. III | + 0.153 | 1.165 (0.252–5.381) | 0.845 | |
| Lympho-Vascular Invasion | ||||
| Negative vs. Positive | + 0.669 | 1.952 (0.943–4.041) | 0.072 | |
| Estrogen Receptor Status | ||||
| Negative vs. Positive | -0.794 | 0.452 (0.244–0.836) | 0.011 | |
| Progesterone Receptor Status | ||||
| Negative vs. Positive | -0.064 | 0.938 (0.503–1.749) | 0.841 | |
| Her2-neu Amplification | ||||
| Negative vs. Positive | − 0.0324 | 0.723 (0.359–1.459) | 0.365 | |
| p53 | ||||
| Negative vs. Positive | -0.099 | 0.906 (0.332–2.471) | 0.847 | |
| Ki67 (quantity in % as continuous) | 0.000 | 1.000 (0.981–1.020) | 0.988 | |
| Ki67 categories | ||||
| ≤ 14% vs. >14% | + 0.662 | 1.983 (0.709–5.297) | 0.197 | |
| Therapeutic Factors | Neo-adjuvant chemotherapy (No. of sessions as continuous) | + 0.019 | 1.019 (0.879–1.181) | 0.804 |
| Adjuvant chemotherapy (No. of Sessions as continuous) | -0.117 | 0.890 (0.718–1.102) | 0.286 | |
| Adjuvant Radiotherapy (No. of Sessions as continuous) | -0.165 | 0.848 (0.769–0.936) | 0.001 | |
| Status of Hormone therapy | ||||
| Negative vs. Positive | -2.266 | 0.104 (0.023–0.475) | 0.004 |
Identification of the NDV with the strongest association with BCFS
When the 3 P, P/T, and P/N variables which were significantly associated with survival were put in a multivariate Cox model as covariates (the N/T and T/N ratios were not included because of their linear dependencies), they all lost their significance (respectively: p = 0.443, p = 0.970, and p = 0.341), and the multiple regression analysis could not identify the superior NDV over the others. Therefore, we ran 3 pair-wise regressions (or bivariate Cox models) in order to identify the NDV with the strongest and most independent association with survival, and also to determine their relative superiority in prognostication. In the model including the P and P/T variables as covariates, the P value remained significantly (p = 0.028) associated with BCFS and took precedence over the P/T ratio, while the P/T ratio lost its significance (p = 0.713). In the model including P/N and P/T ratios as covariates, again P/T ratio lost its significance (p = 0.496) and P/N ratio remained significantly associated with BCFS (p = 0.031). In the third model including the P value and the P/N ratio as covariates, they both lost their significances (p = 0.291 and p = 0.328, respectively), which conveys both quantities have approximately the same impacts on BCFS (although the P value had a smaller p-value) while both took precedence over the P/T ratio for post-NAC patients.
Multivariate analysis
When the CPCs with significant effects on survival (tumor size, pT-stage, and ER) were included in a Cox multiple regression, they all lost their significance (respectively, p = 0.728, 0.118, and 0.084). But when the same model was made including CPCs in addition to NDVs having significant impacts on survival, the tumor size (p = 0.988), pT-stage (p = 0.066), P value (p = 0.718), P/T value (p = 0.431), and P/N value (p = 0.868) lost their significance while the ER status (p = 0.046) remained significantly and independently associated with BCFS.
Comparison of pN-staging and LNR systems in terms of prognostication in post-NAC BC patients
As unveiled by the aforesaid results, the P value (as the foundation of the pN-staging system), had a stronger association with BCFS than the P/T ratio (as the base of the LNR system); accordingly, the pN-staging system may literally be considered superior than the LNR system in post-NAC prognostication for its dominant foundation. Patients’ distribution among the stages of the pN-staging and the LNR systems are depicted in Table 3. The KM analyses estimated the 5-year breast cancer-free survival for the pN-stages as 72.2%, 58.0%, 24.2%, and 0.0% respectively for pN0 to pN3 stages (log rank test, χ2 = 13.03, p = 0.005), and 72.2%, 50%, 31.7%, and 26.7% respectively for LNR0, low-, intermediate-, and high-risk LNR categories (log rank test, χ2 = 8.23, p = 0.041), (Fig. 1 and Fig. 2). The proportionality of hazards between the stages of the pN-staging system was 1 versus 1.796, 2.817, and 5.553 for pN0 (as reference) versus respective higher pN-stages (p = 0.009), and 1 versus 4.042, 2.112, and 3.348 for LNR0 (as reference) versus respective higher LNR stages (p = 0.055). This conveys that the proportionality of hazard is significant between the stages of pN-staging system but not among the stages of the LNR system. The stage progression from pN0 to pN3 for the pN-staging system, and from LNR0 to high-risk for the LNR system (as continuous variables) were both significantly associated with BCFS (p = 0.001 and p = 0.007 respectively) which means patients in the higher stages of both systems have higher risks of event experiences. But when the two pN-stage-progression and LNR-stage-progression variables were put in a bivariate Cox model (as continuous variables), neither remained significant (p = 0.067 and p = 0.865, respectively for pN-staging and LNR systems). This also may be translated as: stage-progressions in both systems have approximately the same effects on survival, and neither is superior over the other.
Table 3.
Patients’ distribution among the pN-staging and the LNR survival groups
| LNR Staging Categories, Count (%) | |||||
|---|---|---|---|---|---|
|
LNR0
N = 59 (34.9) |
Low-Risk
N = 29 (17.1) |
Intermediate-Risk
N = 44 (26.0) |
High-Risk
N = 37 (21.8) |
||
| pN-Staging Categories, Count (%) |
pN0 N = 59 (34.9) |
59 (34.9) | 0 | 0 | 0 |
|
pN1 N = 55 (32.5) |
0 | 29 (17.1) | 20 (11.8) | 6 (3.5) | |
|
pN2 N = 44 (26.0) |
0 | 0 | 24 (14.2) | 20 (11.8) | |
|
pN3 N = 11 (6.5) |
0 | 0 | 0 | 11 (6.5) | |
Fig. 1.
The survival function and the 5-year survival estimation for post-NAC patients by the pN-staging system
Fig. 2.
The survival function and the 5-year survival estimation for post-NAC patients by the LNR staging system
Discussion
The NAC, theoretically, has the potential to change the presurgical axillary nodal status by various and maybe unpredictable manners. It may reduce the total number of nodes in the excised axillary specimen (the T value) [10, 13, 24, 25], or decrease the number of positive nodes by eliminating the nodal cancer cells (the P value), or even may change the number of negative nodes (the N value) regarding the T = P + N equation [8–22]. This, furtherly influences on the ratios made by the T, P, and N values, so that, not only the axillary nodal metrics are altered by the NAC, but also the two current node staging systems that are founded on these measures are inconsistently altered. Therefore, the majority of clinicians consider the post-NAC axillary staging unreliable [8–22]. Notwithstanding, some recent studies have concluded that the two pN-staging and LNR systems maintain their significant associations with survival, and can prognosticate the patients based on the post-NAC axillary nodal climate [24, 26, 27]. Even the sentinels’ samplings in post-NAC patients have surprisingly been claimed to be prognosticative and reliable [28]. The accuracy, reliability, and precision of the post-NAC axillary staging was the main focus of our study, and our findings confirmed that, in spite of all the counter effects of neoadjuvant chemotherapy on the axillary nodes, both the pN-staging and LNR systems, and also their bases (the P and the P/T values) remain highly and significantly prognosticative in post-NAC patients.
The NDVs (as continuous variables) are broadly overlooked in the literature as prognosticators, while Clinicians prefer to use staging “systems” (categorical variables) to “absolute numbers” or “ratios” because of their easier clinical application. The associations of the NDVs (as continuous variables) with survival were evaluated by univariate Cox models and any specific change in the quantity of an NDV would be associated with a specific change in survival. Accordingly, for interpretation of such results, we need an individualized coefficient (the hazard ratio of the variable) which is given by the Cox analysis, that shows the extent of variable’s impact on survival; this seems to be somehow complex for clinical use. That is while the two pN-staging and LNR systems are categorical variables, respectively based on two continuous variables: the P and the P/T values. What these two systems actually do, is changing the P and P/T values from continuous to categorical variables by defining thresholds in the quantities of them, which consequently makes 4 groups (categories) with claimed distinct survival functions. The way that the NDVs (as continuous variables) and the staging systems (as categorical variables) prognosticate BC patients may clearly be exposed by this example: consider a post-NAC patient with 4 positive nodes out of 10 excised LNs in our study; such a patient would be classified as pN2 by the pN-staging system and as intermediate-risk by the LNR system, and accordingly, 58% and 50% survival rates are respectively estimated for her by KM analysis at 5 years. While the P value, rather than estimation, is directly and significantly associated with event-occurrences, and any unit increase in the positive nodes is associated with 11.9% increase in event experiences (Table 2), and thus this patient with 4 positive nodes is 47.6% more likely to experience events than the patients with no axillary involvement. So, the impact of NDVs on survival, which is given by the Cox analyses, may be more comprehendible but harder to interpret, while the stages by the staging systems are easier to interpret but harder to comprehend; since it is difficult to understand how a patient with five positive nodes has the same likelihood of an event as one with nine positive LNs (approximately 60% vs. 100%). The way the cut-points or thresholds are set in the quantities of the P, and the P/T values [28] and how the categories have evolved are matters of debate for both staging systems. Despite that Vinh-Hung et al. described a bootstrapping procedure to identify the thresholds for the LNR staging system [27], there is limited evidence explaining how the pN-staging system evolved [4]; it seems to be more empirically evolved than being mathematically calculated. On the other hand, the cut-points theoretically can be put at any point in the continuum of the P and the P/T quantities [29] because by identification of any threshold, smaller quantities (lower stages) would have better survival functions while greater quantities (higher stages) would have worse. Accordingly (and hypothetically), many other classification themes may be applicable for patient categorization and they all might be statistically significant.
Likewise, the NDVs other that the P and the P/T values are totally overlooked in the literature for their probable prognostic values, and we have been so fortunate to address them here for the post-NAC BC patients. The P value in our study appeared to be the variable with the strongest and most independent association with survival for the post-NAC BC patients. The fact that the P value remained the top quantity associated with survival in regression analyses (Cox bivariate and multivariate regressions) denotes that the positivity of the nodes after neoadjuvant chemotherapy is the most important indicator of survival and the key predictor of event occurrences. The positivity of the nodes after neoadjuvant chemotherapy may implicate an advanced disease, or insufficiency of neoadjuvant-treatment, or a residual nodal component, or a chemo-resistant cancer that might lower the patients’ survival. Therefore, the neoadjuvant chemotherapy perhaps should be aimed to pre-surgically clear the axilla from positive nodes.
The ranking of the NDVs based on their strength of association with BCFS (prognosticative capabilities) in the post-NAC patients revealed to be: P > P/N (or T/N) > P/T (or N/T). Thereupon, the staging system founded on the P value best merits to be employed for the staging of post-NAC BC patients. Here are the reasons why both the pN-staging and the LNR systems are effectively prognosticative: as previously noted, both the P and the P/T values (respectively the bases of the pN-staging and the LNR systems) are significantly associated with survival, so they are founded on reliable bases. The results of KM analyses confirmed that both the pN-staging and the LNR systems can significantly stratify the post-NAC BC patients into distinct survival groups with claimed different survival functions. The stage-progression variable (continuous) of both systems was significantly associated with worse survivals, this means patients at higher stages in both systems have lower survival rates.
However, we think the pN-staging system is the superior classification system for patients undergoing neoadjuvant chemotherapy, because: First, regarding the bases of the systems, both the P and P/T value were significantly associated with survival, but the P value had a higher significance level the P/T ratio (p < 0.000 vs. p = 0.006). Second, when the bases of the two systems (i.e., the P and the P/T values) were included in a bivariate Cox model, the P value appeared to be more strongly and independently associated with survival and took precedence over the P/T ratio. In other words, the P value maintained its statistical significance in association with survival while the P/T ratio lost (p = 0.028 vs. p = 0.713). Third, regardless of the bases, the two systems themselves were significantly associated with survival as categorical variables revealed by the KM analyses; here again, the pN-staging system had a stronger association with survival than the LNR system (p = 0.005 vs. p = 0.041). Fourth, a univariate categorical Cox model showed that the proportionality of hazard between the stages of the pN-staging system was statistically significant and followed a stepwise increase from pN0 to pN3 stages (p = 0.009). While for the LNR system, not only the proportionality of hazard was not stepwise, the results also were not statistically significant (p = 0.055). The statistical significance of KM results for the LNR system, and by the way the lack of statistical significance for the proportionality of hazard between its stages, reveals the fact that the significant results of log rank test in KM analysis does not necessarily mean that all the stages within the evaluated system have significantly different survival functions. In other words, the results of KM analysis will turn statistically significant if at least two of the categories within that system have significantly different survival functions [4].
With all these, by looking at the graphs of patients’ stratifications (Fig. 1, A and B), survival groups are not well separated and multiple crosses are observable in both of the graphs. The pN-staging system seems to be more successful in discriminating high-risk cases as the pN3 group has the least crosses with other lower stages, while the LNR system appeared to be more effective in isolating the low-risk cases, only. The poor patient discrimination by the two systems (regarding multiple crosses in the graphs) probably is because of systems’ defects in identification of appropriate thresholds in the quantities of the P and the P/T values, and that necessitates revisions to redetermine new cut-points for post-NAC patients staging. Accordingly, neither of the two current staging systems showed up to be flawless in this study, and none can be confidently claimed to be superior, but in the current situations, the pN-staging system seems to be a better choice founded on a more reliable cornerstone for post-NAC BC patients. Even newer systems based on the P/N ratio may be defined in the near future for post-NAC cases since this NDV was even superior than the P/T ratio as the base of the LNR system. In practical terms, this means that for patients undergoing neoadjuvant chemotherapy, the absolute state of residual nodal involvement (pN‑staging) should be prioritized in everyday clinical decision‑making, as it provides more reliable prognostic guidance than the lymph node ratio. This reinforces the importance of meticulous pathological evaluation of axillary nodes after NAC when planning adjuvant therapy and follow‑up.
Study limitations
This study has several limitations that should be acknowledged. First, the retrospective, single‑center, single‑surgeon design restricts generalizability, and therapeutic practices during the study period (2008–2018) may not fully reflect current standards. We emphasize that our conclusions should be interpreted with caution and not overgeneralized to all NAC-treated breast cancer patients. In addition, the relatively high mastectomy rate observed in our cohort reflects the referral nature of the Kaviani Breast Disease Institute (KBDI) and cultural preferences in Tehran during that period; therefore, this surgical pattern should be interpreted in context and may not represent the broader breast cancer population. Second, the relatively small sample size, particularly in the pN3 subgroup (n = 11), reduces statistical power and limits the reliability of subgroup analyses. Third, although all patients received standard NAC regimens according to institutional protocols, detailed neoadjuvant regimen composition was not consistently documented, precluding stratification by regimen type or intensity and introducing potential confounding. Fourth, our performance assessment relied on Kaplan-Meier separation and Cox hazard ratios. We did not calculate the concordance index (C-index) due to the limited sample size and the primary aim of comparing established categorical staging systems rather than developing a new continuous prediction model. Fifth, Residual Cancer Burden (RCB), a validated composite measure of residual disease burden, was not routinely measured in our institution and could not be incorporated; our study instead focused on the prognostic value of the nodal environment (pN‑staging and LNR), which is conceptually distinct from RCB. Future studies integrating RCB scoring with nodal staging systems will be necessary to determine their complementary roles in prognostic assessment. Sixth, the relatively short median follow‑up (513 days) may affect the reliability of long‑term survival estimates, and although Kaplan–Meier methods allow extrapolation, these results should be interpreted with caution. Longer follow‑up in future studies will be required to confirm the prognostic trends observed here. Seventh, although ER status was analyzed and shown to remain independently significant in multivariate models, the interaction between biological markers and nodal classifiers warrants further exploration in larger, prospective studies. Future multi‑center investigations with longer follow‑up and more comprehensive datasets will be required to validate and extend these findings. Finally, this study applied established LNR thresholds, which have been validated in large cohorts, to evaluate their competence in the post‑NAC setting. Derivation of new cohort‑specific cutoffs was not feasible given the limited sample size and was beyond the scope of the study. Future multi‑center studies with larger datasets will be required to determine whether revalidated or cohort‑specific LNR thresholds can improve prognostic accuracy.
Conclusion
Both the pN-staging and the LNR systems remain significant prognostic classification systems in post-NAC BC patients, while the pN-staging system seems to be superior than the LNR system in these patients.
Acknowledgements
The present study was made possible by the support of the Tehran University of Medical Sciences and Kaviani Breast Disease Institute. The authors would like to thank all 169 patients who participated in this study and all staff of the Kaviani Breast Disease Institute who contributed. No AI tools were used in the preparation of this manuscript.
Abbreviations
- LNR
Lymph Node Ratio
- P value
the absolute number of positive axillary lymph nodes
- N value
the absolute number of negative axillary lymph nodes
- T value
the total number of excised lymph nodes
- P/T value
the ratio of positive to total excised lymph nodes
- N/T value
the ratio of negative to total excised lymph nodes
- P/N value
the ratio of positive to negative lymph nodes
- N/P value
the ratio of negative to positive lymph nodes
- T/P value
the ratio of total to positive lymph nodes
- T/N value
the ratio of total to negative lymph nodes
- NDFs
Node-Derived Factors
- CPVs
Clinicopathological Variables
- CPCs
Clinicopathological Characteristics
- TFs
Treatment Factors
- pN-stage
pathological Node stage
- pT-stage
pathological Tumor stage
- pM-stage
pathological Metastasis stage
- LN
Lymph Node
- TNM
the global Tumor-Node-Metastasis staging system
- BC
Breast Cancer
- ALND
Axillary Lymph Node Dissection
- SLNB
Sentinel Lymph Node Biopsy
- ER
Estrogen Receptor
- LVI
lympho-vascular invasion
- PR
progesterone receptor
- BCFS
Breast Cancer-free Survival
- HR
Hazard Ratio
- KM
Kaplan-Meier
- Std. D.
Standard Deviation
- NAC
Neo-adjuvant Chemotherapy
- AJCC
American Joint Committee on Cancer
- SPSS
Statistical Package for Social Sciences
Authors’ contributions
AS and NM conceived and designed the study. AS proposed the lymph node numbers and ratios, designed the methodological approach, analyzed the data, prepared the tables, and drafted the initial manuscript. NM acquired the data by reviewing patients’ files and collecting all relevant information, and conducted the literature review. AtD and AnD provided critical revisions to the manuscript for important intellectual content. AK supervised the project and performed scientific editing. All authors contributed substantially to the interpretation of data, critically revised the manuscript for important intellectual content, approved the final version to be published, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No sponsor was involved in the study design; collection, analysis, or interpretation of data; writing of the report; or the decision to submit the report for publication. All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to patient privacy restrictions under institutional ethics guidelines but are available from the corresponding author (Amin Safavi, [aminsafavi.md@gmail.com]) on reasonable request, subject to approval by the ethics board of Tehran University of Medical Sciences. All data supporting the findings are included in the manuscript and its supplementary files (e.g., tables and figures).
Declarations
Ethics approval and consent to participate
This retrospective study was approved by the ethics board of Tehran University of Medical Sciences. Written informed consent was obtained from all patients at the time of admission for participation in research and use of their anonymized data, in accordance with institutional protocols. All procedures were performed in compliance with the ethical standards of the institutional and national research committees and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Consent for publication
Not applicable, as this study does not contain any individual person’s data in any form (e.g., images, videos, or identifiable details). The ethics board of Tehran University of Medical Sciences confirmed that consent for publication was not required due to the anonymized and retrospective nature of the data.
Competing interests
The authors declare no competing interests.
Footnotes
The global Tumor-Node-Metastasis (TNM) cancer staging system.
The P value in bold capital in this article stands for the absolute number of positive nodes, and should not be mistaken for the statistical p-value which in this article, is in italics proceeded by a dash.
Standard Deviation.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Wang MS, Wang MZ, Wang Z, Song Y, Gao P, Wang P, et al. Comparison of three lymph node staging methods for predicting outcome in breast cancer patients with mastectomy. Ann Transl Med. 2021;9(4):300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Özkavruk Eliyatkın N, Başkır İ, İşlek A, Zengel B. New lymph node parameters and a comparison with the American joint committee on cancer N-Stages in breast cancer. Cyprus J Med Sci. 2023;8(4):276–86. 10.4274/cjms.2023.2023-34.
- 3.Giuliano AE, Connolly JL, Edge SB, Mittendorf EA, Rugo HS, Solin LJ, et al. Breast cancer-Major changes in the American joint committee on cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67(4):290–303. [DOI] [PubMed] [Google Scholar]
- 4.Safavi A, Kaviani A, Mohammadzadeh N, Zand S, Elahi A, Krag D. Breast cancer prognostication by pathologic node staging (pN-staging) system versus lymph node ratio (LNR): A critical review of conflicts with number of nodes, Z-0011 Trial, staging Cut-points, Neo-adjuvant Therapy, and survival Estimation. Archives Breast Cancer. 2017;4(4):110–23. [Google Scholar]
- 5.Li X, Liu J, He L, Tian M, Xu Y, Peng B. miR-584-5p regulates MSMO1 to modulate the AKT/PI3K pathway and inhibit breast cancer progression. Protein Pept Lett. 2025;32(3):171–82. [DOI] [PubMed] [Google Scholar]
- 6.Jin H, Wan M, Pan S, Wang Z, Wang W, Zhang J, et al. CD1B expression in triple-negative breast cancer: its implications for prognosis and immunotherapy outcomes. Curr Gene Ther. 2025. Epub ahead of print. [DOI] [PubMed]
- 7.Chen Y, Qi Y, Wang K. Neoadjuvant chemotherapy for breast cancer: an evaluation of its efficacy and research progress. Front Oncol. 2023;13:1169010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ahn SH, Kim HJ, Lee JW, Gong GY, Noh DY, Yang JH, et al. Lymph node ratio and pN staging in patients with node-positive breast cancer: a report from the Korean breast cancer society. Breast Cancer Res Treat. 2011;130(2):507–15. [DOI] [PubMed] [Google Scholar]
- 9.Ataseven B, Kümmel S, Weikel W, Heitz F, Holtschmidt J, Lorenz-Salehi F, et al. Additional prognostic value of lymph node ratio over pN staging in different breast cancer subtypes based on the results of 1,656 patients. Arch Gynecol Obstet. 2015;291(5):1153–66. [DOI] [PubMed] [Google Scholar]
- 10.Chen S, Liu Y, Huang L, Chen CM, Wu J, Shao ZM. Lymph node counts and ratio in axillary dissections following neoadjuvant chemotherapy for breast cancer: a better alternative to traditional pN staging. Ann Surg Oncol. 2014;21(1):42–50. [DOI] [PubMed] [Google Scholar]
- 11.Chen YL, Wang CY, Wu CC, Lee MS, Hung SK, Chen WC, et al. Prognostic influences of lymph node ratio in major cancers of taiwan: a longitudinal study from a single cancer center. J Cancer Res Clin Oncol. 2015;141(2):333–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dings PJ, Elferink MA, Strobbe LJ, de Wilt JH. The prognostic value of lymph node ratio in node-positive breast cancer: a Dutch nationwide population-based study. Ann Surg Oncol. 2013;20(8):2607–14. [DOI] [PubMed] [Google Scholar]
- 13.Duraker N, Bati B, Çaynak ZC, Demir D. Lymph node ratio May be supplementary to TNM nodal classification in node-positive breast carcinoma based on the results of 2,151 patients. World J Surg. 2013;37(6):1241–8. [DOI] [PubMed] [Google Scholar]
- 14.Giuliano AE, McCall L, Beitsch P, Whitworth PW, Blumencranz P, Leitch AM, et al. Locoregional recurrence after Sentinel lymph node dissection with or without axillary dissection in patients with Sentinel lymph node metastases: the American college of surgeons oncology group Z0011 randomized trial. Ann Surg. 2010;252(3):426–32. discussion 32 – 3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hatoum HA, Jamali FR, El-Saghir NS, Musallam KM, Seoud M, Dimassi H, et al. Ratio between positive lymph nodes and total excised axillary lymph nodes as an independent prognostic factor for overall survival in patients with nonmetastatic lymph node-positive breast cancer. Indian J Surg Oncol. 2010;1(4):305–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ibrahim EM, Elkhodary TR, Zekri JM, Bahadur Y, El-Sayed ME, Al-Gahmi AM, et al. Prognostic value of lymph node ratio in poor prognosis node-positive breast cancer patients in Saudi Arabia. Asia Pac J Clin Oncol. 2010;6(2):130–7. [DOI] [PubMed] [Google Scholar]
- 17.Inal A, Akman T, Yaman S, Ozturk SC, Geredeli C, Bilici M, et al. Is lymph node ratio prognostic factor for survival in elderly patients with node positive breast cancer? The Anatolian society of medical oncology. Ann Ital Chir. 2013;84(2):143–8. [PubMed] [Google Scholar]
- 18.Kim SH, Jung KH, Kim TY, Im SA, Choi IS, Chae YS, et al. Prognostic value of axillary nodal ratio after neoadjuvant chemotherapy of Doxorubicin/Cyclophosphamide followed by docetaxel in breast cancer: A multicenter retrospective cohort study. Cancer Res Treat. 2016;48(4):1373–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kim SI, Cho SH, Lee JS, Moon HG, Noh WC, Youn HJ, et al. Clinical relevance of lymph node ratio in breast cancer patients with one to three positive lymph nodes. Br J Cancer. 2013;109(5):1165–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Saxena N, Hartman M, Yip CH, Bhoo-Pathy N, Khin LW, Taib NA, et al. Does the axillary lymph node ratio have any added prognostic value over pN staging for South East Asian breast cancer patients? PLoS ONE. 2012;7(9):e45809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang F, He W, Qiu H, Wang X, Guo G, Chen X, et al. Lymph node ratio and pN staging show different superiority as prognostic predictors depending on the number of lymph nodes dissected in Chinese patients with luminal A breast cancer. Clin Breast Cancer. 2012;12(6):404–11. [DOI] [PubMed] [Google Scholar]
- 22.Xiao XS, Tang HL, Xie XH, Li LS, Kong YN, Wu MQ, et al. Metastatic axillary lymph node ratio (LNR) is prognostically superior to pN staging in patients with breast cancer–results for 804 Chinese patients from a single institution. Asian Pac J Cancer Prev. 2013;14(9):5219–23. [DOI] [PubMed] [Google Scholar]
- 23.Zhu K, Jin H, Li Z, Gao Y, Zhang Q, Liu X, et al. The prognostic value of lymph node ratio after neoadjuvant chemotherapy in patients with locally advanced gastric adenocarcinoma. J Gastric Cancer. 2021;21(1):49–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Saxena N, Hartman M, Aziz R, Rapiti E, Bhoo Pathy N, Lim SE, et al. Prognostic value of axillary lymph node status after neoadjuvant chemotherapy. Results from a multicentre study. Eur J Cancer. 2011;47(8):1186–92. [DOI] [PubMed] [Google Scholar]
- 25.Erbes T, Orlowska-Volk M, Zur Hausen A, Rücker G, Mayer S, Voigt M, et al. Neoadjuvant chemotherapy in breast cancer significantly reduces number of yielded lymph nodes by axillary dissection. BMC Cancer. 2014;14:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tsai J, Bertoni D, Hernandez-Boussard T, Telli ML, Wapnir IL. Lymph node ratio analysis after neoadjuvant chemotherapy is prognostic in hormone Receptor-Positive and Triple-Negative breast cancer. Ann Surg Oncol. 2016;23(10):3310–6. [DOI] [PubMed] [Google Scholar]
- 27.Wu SG, Li Q, Zhou J, Sun JY, Li FY, Lin Q, et al. Using the lymph node ratio to evaluate the prognosis of stage II/III breast cancer patients who received neoadjuvant chemotherapy and mastectomy. Cancer Res Treat. 2015;47(4):757–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Pilewskie M, Morrow M. Axillary nodal management following neoadjuvant chemotherapy: A review. JAMA Oncol. 2017;3(4):549–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Liu D, Chen Y, Deng M, Xie G, Wang J, Zhang L, et al. Lymph node ratio and breast cancer prognosis: a meta-analysis. Breast Cancer. 2014;21(1):1–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to patient privacy restrictions under institutional ethics guidelines but are available from the corresponding author (Amin Safavi, [aminsafavi.md@gmail.com]) on reasonable request, subject to approval by the ethics board of Tehran University of Medical Sciences. All data supporting the findings are included in the manuscript and its supplementary files (e.g., tables and figures).