Median age of the cohort was 55 years, ranging between 24–88 years. There were 162 premenopausal (36.3%) and 284 postmenopausal (63.7%) BC patients. Large proportion of the patients were from urban background (80–85%), 12% were illiterate and 76% belonged to home maker category. Around 13% patients had family history of BC and 15% patients had family history of any other cancer in the first-degree relatives. There was an overlap of 3% cases having family history of BC and any other cancer in the first-degree relatives.
Duration of the lump felt was ranging from a week to few years. In majority of the cases duration of the lump felt was between > 1 month to 12 months (55%). There were three cases that were detected at screening. Of the total patients 52% had tumor in the right breast, 45.5% had in the left breast and 2.5% had bilateral BC. Most common mode of detection was mammography (54%) followed by fine needle aspiration cytology (FNAC) (27%), ultrasound, biopsy/surgery, computed tomography (CT) scan and positron emission tomography (PET) CT. All patients received age and stage appropriate standard of care treatment. Majority of the patients (> 90%) received anthracycline or taxane based chemotherapy and all HR + patients received antihormonal therapy with tamoxifen or aromatase inhibitor. Most common site of metastasis was bone, liver, lung, followed by skeletal muscle, brain and others.
Association of reproductive features between pre and postmenopausal breast cancer
The association of reproductive features like, age at menarche, age at FCB and LCB and parity with the risk of the BC development/ progression were assessed between pre and postmenopausal BC groups. Based on median cut-off of age at menarche, FCB and LCB and parity, BC patients were divided into two categories. Results are represented in Table1.
Table 1
Association of reproductive features between pre and postmenopausal breast cancer
Variable | Subcategory (N = 446) | Premenopausal Group (N = 162) n (%) | Postmenopausal group (N = 284) n (%) | P value |
Age at Menarche (years) | < 14 (n = 193) | 66 (45.2) | 127 (48.7) | 0.503 |
≥ 14 (n = 214) | 80 (54.8) | 134 (51.3) |
Age at FCB (years) | < 23 (n = 169) | 55 (42.6) | 114 (48.1) | 0.316 |
≥ 23 (n = 197) | 74 (57.4) | 123 (51.9) | |
Age at LCB (years) | < 30 (n = 138) | 53 (56.4) | 85(42.7) | 0.029 |
≥ 30 (n = 155) | 41 (43.6) | 114 (57.3) | |
Parity | < 2 (n = 54) | 29 (21.1) | 25 (10) | 0.003 |
≥ 2 (n = 330) | 108 (78.9) | 222 (90) | |
p < 0.05 - Statistically significant (represented in bold). FCB-First childbirth, LCB-Last childbirth |
We observed significantly higher proportion of postmenopausal BC women with older age at LCB (≥ 30 years) which was further supported by increased parity in this group. Age at menarche and FCB did not differ with menopausal status. We also did not find any association of family history of BC or any other cancer with pre and postmenopausal groups.
Further, the association of these reproductive features towards the risk of progression of BC was assessed independently in the entire cohort, in pre and postmenopausal BC. None of these factors were associated with the risk of disease progression (data not shown).
Association of clinicopathological features with disease progression within pre and postmenopausal tumors
The risk associated with each tumor characteristics towards the progression of the disease was evaluated by univariate Cox proportional hazard model in the entire cohort. Grade II and III tumors were categorised into high grade with grade I as the reference (low grade) for examination of association with survival. High grade (HR = 4.3 95% (1.06–17.5), larger tumor size (> 3cm) (HR = 2.2, 95% CI = 1.45–3.2), LN positivity (HR = 2.9, 95% CI = 1.9–4.5), HER2 amplification (HR = 2.07 95% CI = 1.33–3.3) and TNBC subtype (HR = 1.62, 95% CI = 1.03–2.56) were significantly associated with increased hazard (p < 0.05), whereas presence of the TILs infiltration (HR = 0.62, 95% CI = 0.00-0.93), had significantly decreased hazard for the disease progression (p = 0.02). Pre and postmenopausal groups did not show any association with risk of the disease progression in our cohort. We further confirmed the association of these pivotal factors with progression of the disease by multivariate analysis. Tumor size, LN positivity, HER2 amplification and TNBC subtype were associated with significantly increased hazard whereas presence of TILs was associated with lower risk of disease progression as mentioned in Table 3.
Next, we evaluated the association of these factors independently in the pre and postmenopausal groups. Among the postmenopausal tumors, association of tumor size, LN positivity, HER2 amplification and TNBC subtype showed the similar trends of association as in the entire cohort. Presence of TILs was associated with decreased hazard and LN positivity was associated with increased hazard of disease progression in premenopausal BC.
Table 3
Multivariate analysis of clinicopathological tumor characteristics in entire cohort, pre and postmenopausal groups for disease free survival
| | | Entire Cohort (N = 314) | Premenopausal groups (N = 106) | Postmenopausal groups (N = 208) |
| Reference | Variable | HR (95%CI) | P value | HR (95%CI) | P value | HR (95%CI) | P value |
T size (cm) | > 3 | ≤ 3 | 1.7 (1.08–2.67) | 0.020 | 1.06 (0.5–2.22) | 0.89 | 2.2 (1.23–3.83) | 0.008 |
LN status | Negative | Positive | 3.2 (1.75–5.57) | 0.0001 | 2.82 (1.1–7.5) | 0.03 | 3.45 (1.67–7.3) | 0.001 |
TILs | Absent | Present | 0.55 (0.00-0.88) | 0.011 | 0.42 (0.00–0.88) | 0.02 | 0.7 (0.4–1.23) | 0.218 |
Subtype | HR+ | HER2 amplified | 1.75 (1.05–2.96) | 0.033 | 1.44 (0.64–3.25) | 0.38 | 1.89 (0.97–3.7) | 0.06 |
| HR+ | TNBC | 1.64 (0.93–2.89) | 0.088 | 0.59 (0.2–1.7) | 0.32 | 2.92 (1.47 − 5.78) | 0.002 |
p < 0.05-statistically significant (represented in bold), HR- hazard ratio, CI- confidence interval, LN status- lymph node status, TILs- tumour infiltrating lymphocytes, HR+- Hormone Receptor positive, HER2- human epithelial growth factor receptor, TNBC- triple negative breast cancer |
Though TNBC subtype was associated with higher hazard in the entire cohort and within postmenopausal subgroup, TNBC tumors within premenopausal group did not show any association with the hazard in multivariate analysis (Table 3). To verify these findings further, we evaluated the hazard associated with menopausal status within TNBC tumors alone. Both univariate (HR = 0.28 (0.0-0.67), p = 0.004) and multivariate analysis (HR = 0.20 (0-0-0.56), p = 0.002) showed within TNBC subtype, premenopausal BC patients had lower hazard of disease progression.
Kaplan Meier survival analysis confirmed the results of multivariate analysis and showed no difference in the DFS between women with pre and postmenopausal BC (mean survival time 77.1 vs 77.9 months) (p = 0.77). We next examined the influence of subtypes on DFS independently within pre and postmenopausal BC. In postmenopausal BC, patients with TNBC and HER2 amplified tumors were associated with poor outcome as expected and there was no difference in the survival between patients with HER2 amplified and TNBC tumors (mean survival, HR + vs TNBC, 83.8 vs 54 months, p < 0.0001, HR + vs HER2 amplified tumor 83.8 vs 63.3, p = 0.014, HER2 amplified vs TNBC 63.3 vs 54 months, p = 0.328).
However, comparison of DFS among different subtypes within premenopausal patients showed, no difference in the survival between HR + and TNBC breast cancer patients and both subtypes were associated with better outcome compared to patients with HER2 amplified tumors (Fig. 1A, mean survival between TNBC and HER2 amplified BC 79.2 vs 61 months, p = 0.006. Mean survival HR + vs HER2 amplified BC, 74.4 vs 61 months, p-0.058. Mean survival TNBC vs HR + tumors 79.2 vs 74.4 months, p = 0.228).
Next, we examined the DFS and OS independently in each of the molecular subtypes that is HR+, HER2 amplified and TNBC between pre and postmenopausal tumors. There was no difference in the survival between the two groups in the HR + tumors (mean survival pre vs post 74.4 vs 83.8 months, p = 0.374) and in HER2 amplified tumors (mean survival pre vs post 61 vs 63.3 months, p = 0.690). No difference was observed in OS either.
Interestingly in the TNBC subtype, premenopausal patients were associated with better outcome compared to postmenopausal BC patients (Fig. 1B, mean survival pre vs post 79.2 vs 54 months, p = 0.002) which was contrary to the notion that TNBC tumors are always associated with aggressive behaviour. Similar trends were noted with overall survival (mean survival pre vs post, 72.5 vs 49.5 months, p = 0.002). To rule out if these results could be due to the bias within our cohort, we validated our findings in two publicly available larger databases, SCAN-B and METABRIC with different population. OS analysis in these cohorts confirmed our findings on association of premenopausal BC patients with better prognosis within TNBC subtype (Fig. 1C and D, In SCAN-B database, mean survival pre vs post 94.2 vs 84.4 months, p = 0.008 and in METABRIC database, mean survival pre vs post 186.8 vs 157.2 months p = 0.033). We also performed the COX proportional hazard ratio model in these two cohorts to verify similar trends with pre and postmenopausal groups within TNBC subtype tumors. In both SCAN-B and METABRIC databases, TNBC tumors with premenopausal BC were associated with decreased hazard on univariate analysis [SCAN-B, HR = 0.42, 95% CI = 0.00-0.82), p = 0.010, METABRIC, HR = 0.68, 95% CI 0.00-0.98), p = 0.034] and multivariate analysis (SCAN-B HR = 0.45, 95% CI = 0.00-0.85), p = 0.014, METABRIC HR = 0.56 (0.00-0.82), p = 0.003) compared to postmenopausal tumors.