Abstract
Background: Brain metastases from breast cancer have poor prognosis and are a challenge to treat. Multiple treatment options are available. Descriptive and prognostic data on breast cancer brain metastases is limited. Patients and Methods: This study analyzed clinical data of patients who underwent surgical resection of one or more brain metastases. Histological and clinical characteristics, as well as treatment modalities, were analyzed. Results: Initial tumor stage or grade was found not to correlate with the median time to developing brain metastases or survival. Human epidermal growth factor receptor 2 (HER2)-positive status was not associated with shorter median time to developing brain metastases. No correlation was found between the number of brain metastases and patient outcome. Results confirm the survival benefit of surgical resection with or without irradiation. Conclusion: Data showed that patients with HER2-positive and those with triple-negative breast cancer develop brain metastases at lower stages but not earlier after diagnosis, and survival is mostly dependent on treatment modality rather than histological subtype.
Metastatic breast cancer is a major cause of cancer related deaths. Common sites of metastasis include the liver, lungs, bones, skin, and the brain (1-3). Brain metastases (BM) develop in approximately 15-25% of patients with breast cancer and lead to a poor overall prognosis. The molecular subtype of the primary tumor is correlated with the incidence of breast cancer BM (2, 4, 5). Brain metastases are less likely to develop in hormone receptor-positive (HR+) cases, as about 14% of patients with HR+ breast cancer develop BMs (3). Human epidermal growth factor receptor 2 (HER2) overexpression greatly increases the risk, about 30-35% of such patients will have BMs during the course of their disease (3, 5, 6). Patients with triple-negative breast cancer (TNBC) are also at high risk of developing BMs, about 30-40% of patients will be diagnosed with BMs (2, 3, 5).
The prognosis of BMs from breast cancer is poor. Higher performance status, solitary brain metastasis, controlled primary tumor, no extracranial metastasis and younger age (<60-65 years) are favorable prognostic factors. Without treatment, survival is only 1-2 months for all patients with BMs from breast cancer (7, 8). Whole-brain radiotherapy, which is often routinely given, increases survival to 4-6 months (7, 9). Surgical resection further improves patient outcome in patients with good performance status (7, 10). It is important to note, however, that different molecular subtypes are associated with different extents of survival. In a study, patients with HER2-overexpressing breast cancer had 16.5 months of post-progression survival (PPS), while those with HR+ and TNBC had PPS of 9.3 and 4.9 months, respectively (11).
Treatment options of BM include whole-brain radiotherapy (WBRT), surgical resection, stereotactic radiosurgery (SRS) using gamma-knife or linear accelerator through a stereotactic device, and chemotherapy. Best supportive care is only reserved for patients with poor performance status (5, 7-14). The presentation of BMs determines the treatment of choice. WBRT has been considered as the mainstay of therapy for a long time, and is still commonly used for palliation (9, 12). It has been shown that surgical resection or SRS improves survival, especially in those with solitary metastases smaller than 30 mm (7, 8, 15). As BMs have been shown to recur locally after surgery, adjuvant WBRT or SRS is recommended. In 2004, the Radiation Therapy Oncology Group 9508 randomized trial demonstrated SRS plus WBRT to be superior to WBRT alone as it improved survival for patients with a single metastasis (15). Surgical resection is preferred when histological confirmation is required, when the patient has good performance status, and when BMs cannot be treated with SRS (7, 16). The removal of large metastatic masses improves quality of life and provides prompt symptomatic relief while allowing steroid withdrawal. Chemotherapy use is limited as most agents do not penetrate into the brain because of the blood–brain barrier. A better response rate is observed when the tumors are chemo-naive, and chemosensitivity has been shown to decrease with increasing lines of chemotherapy (16). Targeted therapeutic agents (such as lapatinib) have been shown to improve response rates (11, 14, 17).
Continuous follow-up with magnetic resonance imaging is needed for all patients with all BMs from breast cancer as about half will present with new intracranial metastases over time. Then repeated surgeries or SRS have been shown to be beneficial (11, 16, 18).
Survival data for patients with BMs from breast cancer are limited in literature, as most publications focusing on BMs include multiple cancer types. Real-life data of patients with different breast cancer subtypes are needed to help clinicians assess prognosis in patients with metastatic breast cancer. Data are also limited on how various clinical parameters influence patient outcome and how additional treatments beyond surgical resection benefit patients in a real-life setting.
Patients and Methods
A total of 132 patients with BMs from breast cancer were treated between 2008 and 2018 at the Department of Oncology, University of Debrecen. In this study, we analyzed the survival data of those patients who underwent surgical resection of at least one metastatic lesion. Out of the 33 operated patients, 29 had sufficiently detailed patient files for statistical analysis. The mean age at diagnosis of breast cancer was 51.4 (range=33-77) years in the patient group. The median time to any non-brain metastasis (event-free survival 1, EFS1) was 17.0 months [95% confidence interval (CI)=6.50-27.50 months], the median time to developing brain metastasis (event-free survival 2, EFS2) was 49.0 months (95% CI=41.62-56.38 months) and the median overall survival (OS) was 75.0 months (95% CI=44.12-105.88 months). All the patients had extracranial metastases (visceral, bone or lymph nodes), and none had BMs at the time of initial diagnosis. All the patients had good performance status (Eastern Cooperative Oncology Group 0-1 or Karnofsky performance score 80 or above) at the time of their elective surgeries. All the patients received WBRT. Eight patients had SRS, out of which two had two lesions, and one had three lesions irradiated using gamma-knife. Fourteen patients received another line of systemic treatment (chemotherapy/targeted therapy) after the diagnosis of BM. Detailed patient characteristics are given in Table I.
Tumor stage, tumor grade, MIB1 (Ki-67) status, histological type, molecular subtype, HR status and HER2 overexpression, extension of breast surgery, number of metastatic lesions, as well as whether SRS was performed or not were used ad stratifying characteristics. Previous treatment modalities were not used for statistical analyses due to great variation in the regimens used.
The study aimed to identify clinical and histopathological prognostic factors in patients with BMs from breast cancer. Features of breast cancer which are routinely determined in clinical practice were used to stratify the patient cohort and identify possible differences in various survival times of patients. During the statistical analysis Kaplan–Meier estimations were used to identify differences in survival among the groups. All calculations were performed using IBM SPSS Statistics v25.0 statistical program (IBM, Armonk, NY, USA).
Clinicopathological details of breast cancer patients who developed brain metastases (BMs). Basic clinical and pathological features, as well as treatment modalities used to treat the primary or metastatic tumors.
Median survival times of patients with breast cancer who developed brain metastases at different stages. A: Patients with de novo metastatic breast cancer developed further extracranial metastasis earlier compared to those with other stages. B: The median time to the development of brain metastases, however, was not determined by the initial stage of the disease. C: Overall survival times did not differ significantly in patients who were diagnosed with BM during the course of their disease regardless of the initial extent of their disease at the time of diagnosis.
Results
Analyzing survival data of the patients showed that the stage of disease at the time of diagnosis affected median time to non-brain metastasis (EFS1), as those with metastatic breast cancer had a significantly shorter EFS1 but it did not differ significantly between stages (p<0.001, Figure 1A).
Median time to brain metastasis (EFS2), however, was not determined by the stage of the disease. EFS2 was 42.0 (95% CI=12.09-69.91) months in stage I, 49.0 (95% CI=42.07-55.93) months in stage II, 60.0 (95% CI=35.39-84.61) months in stage III, and 48.0 (95% CI=15.79-80.21) months in stage IV, respectively. The differences in median time to development of brain metastases were not significant among stages (p=0.671, Figure 1B).
Surprisingly, OS in patients with BC who underwent surgical resection of one or more intracranial metastases did not significantly differ among stages (considering that some of the patients were still alive at the time of writing). Median OS in patients with stage I disease was 46.0 (25.67-65.32) months, while it was 82.5 (95% CI=33.42-116.68) months for those with stage II, 75.0 (95% CI=36.56-95.44) months for stage III, and 103.0 (95% CI=0.13-207.51) months for stage IV (p=0.896, Figure 1C).
There was an observable difference in the median time to non-brain metastasis of patients with different tumor grades, however, these differences did not reach significance. Patients with a grade 2 tumor had a median EFS1 of 17.0 (95% CI=0.01-52.8) months, while patients with grade 3 tumors had an EFS1 of 11 (95% CI=0.431-21.57) months (p=0.061) (Figure 2). EFS2 and OS times did not significantly differ by grade. EFS2 was 53.0 (95% CI=43.29-62.72) months for those with grade 2 tumors, and 46.0 (95% CI=30.74-61.27) months for those with grade 3 tumors (p=0.577). Median OS for patients with grade 2 tumors was 90.0 (95% CI=57.51-122.49) months and was 65.0 (95% CI=54.82-75.18) months for patients with grade 3 tumors (p=0.583).
HR− status and HER2 overexpression negatively correlated with EFS1 (p=0.014), but not with EFS2 (p=0.373) and OS (p=0.245) (Figure 3). The lowest EFS1 was observed in HR-/Her2+ patients [8.0 (95% CI=1.07-14.93) months]. HR+/HER2+ patients had a slightly better EFS1 [10.0 (95% CI=0.01-43.26) months]. Patients with HR−/HER2− disease had an EFS1 of 21.0 (95% CI=12.41-29.59) months. Those with HR+/HER2− disease had the longest EFS1 at 51.0 (95% CI=19.12-82.88) months. The differences in EFS1 were found to be significant (p=0.008). EFS2 values followed the same pattern, however, the differences were not significant (p=0.373). Median OS differences were also below the level of significance (p=0.245).
Median time to non-brain and brain metastases according to tumor grade in breast cancer patients who developed brain metastases. Even though tumor grade is a determinant of the aggressiveness of tumor, no correlation was found in the median time to development of extracranial and intracranial metastases in the patient cohort.
When HR status, HER2 status and MIB1 labeling index were combined to assess the association of survival time with molecular subtype (as detected by immuno-histochemistry), we found that data not only for EFS1 (p=0.008) but also for OS (p=0.05) exhibited significant differences. EFS2, however, was not correlated with the molecular subtype of breast cancer (p=0.367) (Figure 4). EFS1 times were as follows: Luminal A-like: 90.0 months (95% CI=0.001-234.03), luminal B-like: 33.0 months (95% CI=1.293-65.07), HER2-overexpressing: 8.0 months (95% CI=1.07-14.93), and basal-like/TNBC: 21.0 months (95% CI=6.50-27.49). EFS2 times were 72.0 (95% CI=33.59-110.41) months for luminal A-like, 53.0 months (95% CI=25.99-80.01) for luminal B-like, 42.0 months (95% CI=15.67-68.34) for HER2-overexpressing, and 60.0 (95% CI=38.96-81.04) months for basal-like/TNBC. Median OS times were 218.0 (95% CI=135.70-250.30) months, 66.0 (95% CI=28.59-103.432) months, 46.0 (95% CI=19.67-72.34) months, and 84.0 (95% CI=30.10-137.90) months, respectively.
There was no significant difference in EFS1, EFS2 and OS times of patients according to MIB1 labeling index. The type of breast surgery (breast-conserving or radical mastectomy) also did not significantly affect survival times of patients with BM in this analysis. While the median EFS1 was lower for the mastectomized group (8.0 vs. 17.0 months), the difference was not significant. Median EFS2 was also lower for the mastectomized group (48.0 vs. 53 months), while median OS was higher (135.0 vs. 84 months). These differences were not statistically significant. The histological subtype of breast cancer also influenced patient survival. EFS1, EFS2 and OS were found to be significantly longer in ductal breast cancer compared to the lobular subtype (Table II).
Using stereotactic radiosurgery in addition to surgical resection also did not influence OS (p=0.304) (Table II). It was found that the number of intracranial metastases did not influence the OS in this selected group of patients (p=0.903) (Figure 5).
Median survival times according to hormonal receptor (HR) and human epidermal growth factor receptor 2 (HER2) status in patients with breast cancer who developed brain metastases. A: Median time to extracranial metastases was significantly different in patients according to HR/HER2 status. HER2 overexpression negatively influenced survival which was somewhat compensated for by HR positivity. Median time to brain metastases (B) and median overall survival (C) did not significantly differ in the patient cohort. The pattern of survival times was similar to that of median time to extracranial metastases, however the differences did not reach significance.
Discussion
BMs are more and more common in patients with breast cancer (1, 5). They are also more common in those with HER2-overexpressing and TNBC, but patients with HR-positive may also develop BMs over time (2, 3). The advancement of treatment modalities, including targeted therapies, has led to the improvement of patient survival, especially in HER2 overexpressed breast cancer (5, 6, 14, 19). Improved surveillance has also resulted in increased incidence and prevalence of BMs in breast cancer. BMs are major limitations to life expectancy and quality of life (1, 20). Treatment options for BMs include surgical resection with or without focal or WBRT, SRS and WBRT alone. Furthermore, systemic therapies have significant anticancer activities. Surgical resection is recommended in single- and oligometastatic cases when the primary disease and other metastases are well controlled, the patient has good performance status or when prompt symptomatic relief is needed (7, 11, 17, 20).
This retrospective patient cohort analysis of patients who developed BMs of breast cancer aimed to identify prognostic factors. Only those patients were included in the study who underwent surgical resection of at least one intracranial metastasis (two with oligometastatic disease had two surgeries removing two different metastases at different times) because previous research has proven that surgical resection of BMs contributes to significant survival benefit of patients. Therefore, we aimed to analyze the influence of clinical and histopathological factors in the development of BMs in breast cancer. Literature data on prognostic factors of breast cancer brain metastasis is scarce, as most previous studies have focused on either one subtype of breast cancer or on multiple cancer types.
Tumor stage has a strong influence on disease prognosis, however, we found that while it mostly influenced the median time to development of extracranial metastases (EFS1), there was little direct correlation between median time to development of BMs (EFS2) and initial tumor stage. This might be due to the different treatment approaches for different stages, as radical surgery/systemic chemotherapy is more common in patients with more advanced breast cancer (5, 6, 16, 17). Insignificant differences in OS among various initial stages is also partially explained by the above phenomenon. Furthermore, in this cohort, lower stage was associated with a higher number of HER2-overexpressing breast cancer and patients with locally advanced or metastatic breast cancer were more often HR+ than not.
Median survival times versus molecular subtypes in breast cancer patients who developed brain metastases. A: Patients with human epidermal growth factor receptor 2-positive (HER2) disease and those with triple-negative breast cancer (TNBC) had worse prognosis as these patients developed extracranial metastases earlier than those with hormone-positive breast cancer. B: There was no difference in the median time to development of brain metastases among various subtypes. C: Overall survival correlated with molecular subtypes.
Median survival times according to various clinicopathological characteristics and treatment modalities in patients with breast cancer who developed brain metastases.
Tumor grade in the primary breast cancer had a similar effect to that of tumor stage. The EFS1 was significantly shorter in cases of grade 3 tumors compared to grade 2 tumors, but no significant differences were observed in EFS2 nor OS. This is no surprise as higher grade is associated with increased invasiveness. The blood–brain barrier, however, prevents the formation of BMs, at least early in the disease (1, 11, 21). Over time, cancer cells change their characteristics, may become more aggressive and invasive, and may develop the capability to cross the blood–brain barrier. Initial tumor grade alone therefore carries little information on whether BMs will develop or not. This is consistent with results from another study where no correlation was found between tumor grade and median time to BMs (19). Another histopathological feature, the MIB1 labeling index alone had no correlation with the prognosis of breast cancer.
Overall survival according to the number of brain metastases (BMs) in patients with breast cancer who developed BMs. The number of BMs did not influence patient survival in those who underwent surgical resection of at least one intracranial metastasis. S: Single BM; O: oligometastatic disease (2 or 3 BMs); M: multiple BMs (>3).
HR and HER2 expression influenced EFS1, but not EFS2 and OS. The difference in EFS1 is well known, as HER2-overexpressing breast cancer has a worse prognosis than HR+ breast cancer (2, 5, 19). It is somewhat surprising that even though BMs are described as being more common in HER2+ breast cancer (2, 6), they do not develop much faster than in other tumor types. EFS2 was similar in all four groups, HER2 status seemingly influenced the median time to BMs as EFS2 was somewhat shorter in those with HER2+ tumors compared to HER2− tumors (42 and 46 months vs. 60 and 60 months, respectively).
Molecular subtype as detected by immunohistochemistry (which combined HR status, HER2 status and MIB1 labeling index) was found to correlate with EFS1 and OS, but not with EFS2. HER2-overexpressing breast cancer in this patient cohort was found to lead to early development of extracranial metastases despite targeted therapies, which suggests that despite recent development in the treatment of HER2+ breast cancer, clinicians still need to be cautious and look carefully for extracranial metastases in patients with HER2+ disease, especially when the tumor is HR−. TNBC also gave extracranial metastases relatively early (median EFS1 was 21 months), while HR+, HER2− breast cancer was yet again found to be a more indolent type of breast cancer, as EFS1 was 90 months. TNBC was found to be less aggressive than HER2-overexpressing breast cancer in this cohort, which might be a result of a more aggressive systemic treatment in TNBC. The molecular subtype did not show correlation with EFS2, despite the seemingly great difference in medians. HER2+ breast cancer and TNBC are traditionally considered risk factors for BMs, however, this does not mean that BMs develop earlier in patients with these subtypes of breast cancer. This underscores how little we currently understand of the molecular development of BMs in cancer. OS data were in line with literature data (5, 7, 11).
Survival times were found to be shorter in lobular breast cancer. These patients seemed to benefit less from the received chemotherapy compared to those with ductal breast cancer, as previous research has suggested (17, 22).
Our data confirm previous findings regarding breast surgery, as radical mastectomy bore no benefit in EFS1, EFS2 or OS compared to conservative breast surgeries (7, 20). We also found that gamma-knife surgery (for local recurrence after resection or for non-resectable space-occupying metastases) had no additional survival benefit, however, it helped disease control and may have provided symptomatic relief. Interestingly, our analysis suggests that the number of metastases does not influence patient survival when adequate treatment modalities are available for patients with BMs from breast cancer. It seems that stereotactic irradiation or surgical removal of larger, space occupying metastatic lesions provides local control and WBRT further stabilizes BMs even when multiplex BMs develop.
Conclusion
Brain metastases in breast cancer remain a challenge to treat despite therapeutic advancement seen in the past years. Analyzing data of patients who underwent metastatectomy of BMs from breast cancer helps us to see what histopathological or clinical factors might contribute to the development of BMs and patient outcomes from breast cancer. By including patients whose BM were surgically removed only, a somewhat homogenous patient group was created. It seems that HR status and HER2 expression determine patient outcomes when local control is reached by surgical resection. Patients with HR+ tumors have better outcomes, while HER2 overexpression results in shorter survival after BMs develop. Initial tumor stage or tumor grade, however, did not affect survival times in patients with BM from breast cancer. Interestingly, despite the fact that BMs are said to be more frequent in TNBC and HER2-overexpressing breast cancer, our data suggest that the median time to developing BMs is not shorter in these patients. More research is needed on the topic, as many questions, such as the long-term effects of HER2-blocking antibodies on the development of BMs, still remain unanswered.
Acknowledgements
László Szivos is supported by the ÚNKP-19-3 New National Excellence Program of the Ministry of Innovation and Technology.
Footnotes
Authors' Contributions
AP developed concept, summarized data and prepared the article, JT helped developing concept and contributed to article preparation. EB provided data, DT provided data, LSz collected and prepared data for statistical analysis, JN performed statistical analyses, AK provided data, JV developed concept, summarized data, and helped statistical analyses.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received January 20, 2020.
- Revision received February 9, 2020.
- Accepted February 12, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.