2.2.2. Treatment

BCT included wide local excision of the tumour and appropriate axillary management followed by irradiation of the whole breast, mostly including a boost to the primary tumour bed. Further details about this cohort have been published before.^13 and 14 Some major changes in the management of breast cancer in the Netherlands took place during the period covered by our study. The first major change was in 1998, with the publication of the new Dutch guideline²⁷ according to which lymph node-negative patients with high-risk features, depending on the size, grade and hormone receptor status of their tumour, were advised to receive adjuvant systemic therapy. The second major change was an update of the guideline in 2005, where women with HER2-positive breast cancer were recommended to receive trastuzumab in conjunction with adjuvant chemotherapy. Based on these changes in the guidelines for administering systemic therapy, we decided to divide the study population into three subgroups; 1988–1998, 1999–2005 and 2006–2010.

2.3.3. Definition of end-points

The end-points of this study were LR, distant metastasis and death. For each end-point, the time to event was defined as the interval between the pathologic diagnosis of the primary tumour and occurrence of the event of interest. In the absence of the event of interest, observation time was censored at the last contact date of the patient. LR was defined as any reappearance of tumour growth in the initially treated breast or overlying skin. LR’s that were diagnosed after distant metastases or with synchronous distant disease were not counted as events and these patients were censored at the date of diagnosis of distant disease.

2.4.4. Follow-up

The median follow-up time for all 1143 patients was 6.8 (range 0.1–24.6) years, and 8.5 years for the patients still alive. The median duration of follow-up was 11.5 years for the patients treated in 1988–1998 (14.4 years for those still alive), 8.1 years for the patients treated in 1999–2005 (8.5 years for those still alive) and 3.6 years for the patients treated in 2006–2010 (3.7 years for those still alive). On 1st January 2010, 78 patients (6.8%) were lost to follow-up, which meant that their last contact date was before 1st January 2010. The median follow-up for these 78 patients was 9.6 years.

2.5.5. Statistical analysis

The 5-, 10- and 15-year actuarial rates of LR and the curves for LR-free survival were generated using the Kaplan–Meier method, with comparisons made using the log-rank test. Multivariate analyses were carried out using Cox proportional hazards models to identify factors that were associated with an increased risk of LR and factors that were associated with the occurrence of distant disease. A backward elimination model omitted covariates that had a p-value greater than 0.05. The multiple imputation method was used to generate possible values for missing values of the oestrogen receptor status, tumour grade, tumour size, nodal status and the use of (neo)adjuvant systemic therapy. Multiple datasets were generated using a model that included all variables used in the Cox regression analysis. To examine the impact of LR on the subsequent occurrence of distant disease, we used the Kaplan–Meier method and the Cox proportional hazards model. All tests were two-sided, and a p-value <0.05 was considered statistically significant. All data were analysed using SPSS version 20.

3. Results

3.1.1. Patient, tumours and treatment characteristics

The median age at diagnosis was 37 years (range, 21–40 years). No differences were detected between the three subgroups (1988–1998, 1999–2005 and 2006–2010) with respect to age, nodal status, tumour type, oestrogen and HER2 receptor status (Table 1).

All patients, except for nine, underwent axillary staging, either by axillary dissection, sentinel node biopsy or both. The proportion of patients receiving (neo)adjuvant systemic treatment increased from 32% in the period 1988–1998 to 71% in the period 1999–2005 and to 83% in the period 2006–2010 (Table 1). The increase was much larger for patients with negative axillary lymph nodes (from 6% to 75%) than for patients with positive lymph nodes (from 84% to 98%). Forty-nine (96%) of the 51 patients with an HER2 positive tumour treated in the period 2006–2010, received trastuzumab in combination with chemotherapy (Table 1).

3.2.2. Local tumour control

During follow-up, a LR was diagnosed in 176 patients without evidence of metastatic disease at a mean of 7.9 (median 6.8) years after BCT (range 0.58–24.1). Twenty-six additional LR’s, diagnosed either after or at the same time as distant disease, were not included in the current analyses. Characteristics of the 176 LR’s are presented in Table 2.

The 5-, 10- and 15-year actuarial LR-rates were 7.5% (95% CI: 5.7–9.3), 16.3% (95% CI 13.5–19.0) and 24.6% (95% CI 20.6–28.5), respectively. Univariate analyses showed that a positive axillary lymph node status (P < 0.0001) and the use of (neo)adjuvant systemic therapy (P < 0.0001) were associated with a significantly lower risk of LR (Table 3 and Fig. 1a). Furthermore, we observed a significant trend towards improving local control over the study period (Table 3 and Fig. 1b). The 5-year LR-rate decreased from 9.8% in the period 1988–1998 to 3.3% in the period 2006–2010 (P = 0.006). A stratified analysis according to axillary nodal status showed that this improvement only occurred in patients with negative lymph nodes ( Fig. 1c and Fig. 1d).

The multivariate analysis was complicated by the strong concordance between axillary lymph node status and use of systemic treatment and the sharp increase in the use of systemic treatment over time. Therefore, we constructed two separate multivariate models (Table 4). The first model, in which systemic treatment was not included as a co-variate, showed that patients with a positive lymph node status had a 55% (HR 0.45; 95% CI 0.23–0.60) lower risk of developing a LR, compared to those with negative lymph nodes, and that the risk was also significantly lower for patients treated more recently as compared to the patients treated in the period 1988–1998. In the second model, in which the period of diagnosis was not taken into account, the use of (neo)adjuvant systemic treatment was associated with a reduction of the LR-risk almost 60% (HR 0.42; 95% CI 0.29–0.60) and nodal status was no longer an independent prognostic factor for LR. Separate multivariate analyses for patient with and without (neo)adjuvant systemic treatment also demonstrated that lymph node status was not associated with the risk of LR (data not shown).

In both multivariate models the presence of a positive oestrogen receptor was the only other significant predictor of improved local control and age at diagnosis, tumour size and tumour grade were not associated with the risk of LR (Table 4).

3.3.3. Distant relapse-free survival and other end-points

During follow-up, 34 of the 1143 patients developed a regional (nodal) recurrence, not taking into account the regional recurrences that were diagnosed after distant metastases or with synchronous distant disease. The 5-year regional recurrence rate decreased from 3.1% (95% CI 1.5–4.7) in the period 1988–1998 to 0.4% (95% CI 0.0–1.2) in the period 2006–2010 (P = 0.057).

Ninety-nine patients developed contralateral breast cancer, 287 patients distant metastases and 273 patients died. The 5-year distant relapse-free survival improved from 74.9% (95% CI 71.4–78.4) in the period 1988–1998 to 89.4% (95% CI 86.1–92.7) in the period 1999–2005, and to 92.7% (95% CI 88.8–96.6) in the period 2006–2010 (P < 0.0001).

The 10-year distant relapse-free survival rate was 78.6% (95% CI 72.3–84.9) for patients with and 76.3% (95% CI 73.2–79.4) for patients without a LR (P = 0.351). Also the multivariate analysis showed no statistically significant difference in distant relapse-free survival between patients with and without a LR. When only the early-LR’s, occurring within 5 years after BCT, were taken into account we did find a difference in distant relapse-free survival rates. Patients with an early-LR (n = 70) had a worse distant relapse-free survival compared to patients without an early-LR (n = 1073) (HR 1.83; 95% CI 1.27–2.64; P = 0.001) (Fig 2a). No independent prognostic effect of LR on distant relapse-free survival was observed in patients with a disease-free interval of more than 5 years after BCT (n = 702) (HR 1.24; 95% CI 0.74–2.08; P = 0.407) (Fig. 2b).

4. Discussion

The current study is the largest study so far with a long-term follow-up (up to 20 years) focusing on local control after BCT in women up to 40 years of age with early-stage breast cancer. Our results indicate a significant improvement in local control over time. A stratified analysis according to axillary nodal status showed that the improvement in local control did only occur in patients with negative lymph nodes. We also noted that early LR’s had a negative impact on distant relapse-free survival.

In our series, patients who received (neo)adjuvant systemic therapy had an almost 60% lower risk of developing a LR, underlining the importance of the use of adjuvant systemic treatment in this patient group. Several studies have confirmed the possible benefit of adjuvant systemic therapy on local control after BCT.^{1, 7 and 12} The lower risk of LR in node-positive patients can be explained by the use of adjuvant systemic treatment, as is illustrated by our finding that lymph node status was no longer an independent prognostic factor in the subgroups of patients who did and did not receive adjuvant systemic treatment.

Recently, studies of tumour biology by gene expression micro-array technology have demonstrated patterns correlating with breast cancer in young women²² and even LR’s in young women.²⁸ Young women appear to have a higher proportion of more aggressive tumours and are, therefore, expected to benefit more from systemic therapy. Additional studies about tumour biology may be helpful to determine which patients are most likely to benefit from systemic treatment and to further optimise therapeutic options for breast cancer in young women.

Several studies, presented in 2005, showed that the use of trastuzumab in combination with chemotherapy reduced the locoregional and disease-free survival rate with approximately 50%^20 and 21 and the mortality rate with 33%²¹ among women with surgically removed HER2-positive breast cancer. As a result of these trials, HER2-testing has become routine and trastuzumab in combination with chemotherapy is now standard treatment for most patients with HER2-positive breast cancer. After 2005, all patients in our study were tested for HER2 status, of whom 20% presented with HER2-positive tumours (Table 1). As described previously, young women appear to have more aggressive tumours, including a higher incidence of HER2 overexpression compared to older women.^{7, 18 and 22} The length of follow-up in our study is too short and the number of patients with known HER2 positivity is too small to draw definitive conclusions about the impact of trastuzumab on local control in this group of patients. However, a recent study revealed significant improved local control after BCT in patients with small node-negative HER2-positive breast cancer who received trastuzumab.²³

Minimising the risk of LR remains an important clinical issue since many studies on BCT have shown that a LR significantly correlates with subsequent distant metastases.^{1, 2, 3 and 29} In our study women with a LR occurring within 5 years after BCT were nearly twice as likely to experience distant disease compared to women without a LR. However, no worse distant relapse-free survival was observed for women with a LR occurring more than 5 years after BCT. Late-occurring LR’s are probably more often new primary tumours and are generally thought to have a more favourable prognosis than the early-occurring true recurrences. Our results seem to confirm this opinion, although we could not verify this based on histological comparison of the primary with the new/recurrent tumours. The finding that a LR is an unfavourable prognostic factor, especially the ones occurring shortly after BCT (⩽2–5 years) is in line with other studies.^{1, 2, 3 and 5}

Collecting long-term follow-up information for this group of young women was challenging. Nevertheless, we managed to limit the amount of missing disease-specific follow-up information after 1st January 2010 to only 78 patients (6.8%). Most of them were treated in the group 1988–1998 and still had a median follow-up duration of 9.6 years. Another limitation of this study are the large number of missing data with respect to oestrogen receptor status and tumour grade for those patients treated before 1999. As these missing data were largely due to incomplete pathology reports, we assumed that the data were missing at random and therefore used the multiple imputation method to generate possible values.

5. Conclusion

Our study demonstrates that early-occurring LR’s are associated with distant metastases. We found a substantial decline in the incidence of LR’s after BCT in women aged ⩽40 years which seems to be largely attributable to the increased use and effectiveness of (neo)adjuvant systemic therapy during the study period. Therefore, we support the choice of BCT over mastectomy also for young women, especially when they will receive systemic treatment as well.

Conflict of interest statement

None declared.

References

Copyright © 2013 Elsevier Ltd. All rights reserved.