Risk Factors for Brain Relapse in Patients with Metastatic Breast Cancer

Abstract & Commentary

Synopsis: The presence of lung metastases as the first site of relapse and a negative hormone receptor status are predictive for the occurrence of brain metastases in patients with metastatic breast cancer. A prophylactic treatment should be evaluated in these subsets of patients.

Source: Slimane K, et al. Ann Oncol. 2004;15: 1640-1644.

There is evidence suggesting that the incidence of brain metastases is rising in patients with breast cancer. One explanation is the selective destruction of non-brain metastases by new chemotherapy regimens, allowing a later development of brain metastases. It has been shown that the occurrence of brain metastases in patients with breast cancer decreases survival and alters the quality of life. The possibility of detecting early or preventing the occurrence of brain metastases could therefore lead to increased survival and better quality of life. Some approaches have been shown antitumor activity in brain metastases and could therefore be evaluated as a preventative treatment for brain metastases. This may include prophylactic cranial irradiation, high-dose methotrexate, and temozolomide. There is a need to identify the patients with metastatic breast cancer who are at risk of developing brain metastases. In the present study, predictive factors of brain metastases in patients with metastatic breast cancer were determined, in order to propose a targeted strategy aimed at screening or preventing brain relapse in these patients.

Comment by Stuart M. Lichtman, MD

This study was conducted in 2 parts. Risk factors for brain metastases were first determined in a series of metastatic breast cancer patients, and then confirmed in a second series.

Patients were included in prospective, randomized trials for the 2 series of patients. Two hundred and fifteen patients were included in the present study. These patients were selected from the database of a randomized trial that compared post-operative castration vs no castration in pre-menopausal patients with breast cancer. It included 557 patients at the Institut Gustave Roussy between 1989 and 1998. The results of this trial have been previously reported.1 Among the 557 patients, 220 patients developed a metastatic relapse of breast cancer before August 2003. The follow-up of the trial did not include any systematic brain CT scan, nor brain MRI. The clinical files of these 220 metastatic patients were reviewed in August 2003 in order to determine which patients had presented brain metastases. Five patients presenting brain metastases as first relapse were excluded from the analysis. The predictive factors for brain metastasis were therefore analyzed in the remaining 215 patients. A number of variables were evaluated. The age, tumor size, lymph node status, tumor grade, interval between primary and first metastatic relapse, first sites of metastases, loco-regional treatment and adjuvant medical treatment were extracted from the database of the clinical trial. Hormone receptor status was extracted from the clinical charts. These variables were tested for their association with the occurrence of brain metastases.

Two hundred and fifteen metastatic breast cancer patients were included in the present series. The median follow-up between diagnosis of metastatic relapse and last visit or death was 22 months (range, 0-117 months). The median follow-up for living patients was 35 months. Thirty-one patients (14%) presented a brain relapse prior to the diagnosis of the first systemic metastasis. The 2-year incidence of brain metastases was 13%. The median interval between the first metastatic relapse and the occurrence of brain metastases was 13 months. Lung metastases as the first site of relapse, negative hormone receptor status, absence of bone metastases at first relapse and short disease-free interval (> 24 months) were associated with a higher risk of developing brain metastases in the univariate analysis.

Fifteen out of 50 patients (30%) presenting with lung metastases as the first site of relapse subsequently developed brain metastases during the follow-up. Ten out of 29 patients (34%) with negative hormone receptor status subsequently presented a brain relapse. The 2-year incidence of brain metastases were 24% and 7% in patients with and without lung metastases, respectively (P = 0.0001). The 2-year incidence of brain metastases were 6% and 37% in patients with positive and negative hormone receptor status, respectively (P = 0.0001). In the multivariate analysis, lung metastases (hazard ratio, 4.3; 95% CI, 1.9-9.3; P = 0.0003) and negative hormone receptor status (hazard ratio, 4.2; 95% CI, 1.7-11; P = 0.002) were associated with and increased risk of brain metastases. The confirmatory series included 199 patients with metastatic disease. The median follow-up was 21 months (range, 0-122 months). The median follow-up for living patients was 34 months (range, 0-122 months). Lung metastases and negative hormonal receptor status were confirmed as predictive factors for the occurrence of brain metastases.

In the present study, the presence of lung metastases and negative hormone receptor status are strong predictive factors for the occurrence of a brain relapse in 2 prospective series of patients with metastatic breast cancer. Previous studies in the same population reported that negative hormonal receptor status2,3 and, more recently, Her2-neu overexpression4,5 may be associated with a higher risk of developing brain metastases. The present study confirms that negative hormone receptor status is a predictive factor for brain metastases in 2 sets of homogeneously treated, prospectively assessed metastatic breast cancer patients. Her2-neu was not performed in this series since most of the patients relapsed before the introduction of trastuzumab in daily practice. Only a few studies specifically evaluated the correlation between the first sites of metastases and the subsequent development of brain metastases. Crivellari et al reported that 11 out of 28 breast cancer patients presenting a brain relapse had lung metastases as the first site of relapse.6 The presence of lung metastases was associated with a higher risk of further developing brain metastases in this series. Miller et al4 reported that the presence of lung metastases was associated with a 2.4-fold increased risk of brain metastases (P = 0.07) in a series of 155 consecutive patients with metastatic breast cancer. In the present series, the presence of lung metastases was strongly associated with the occurrence of brain metastases. Indeed, 30% of the patients presenting lung metastases as the first site of relapse subsequently developed a brain relapse. Slimane et al raise questions regarding the management of breast cancer patients at the time of metastatic relapse: 1) Should a screening of brain metastases be performed in patients with lung metastases and/or negative hormone receptor status? and 2) Should a prophylactic treatment of brain metastases be considered? Randomized trials in these high risk populations should be considered to prevent or delay relapse and hopefully maintain or improve quality of life.

References

1. Arriagada R, et al. J Clin Oncol (Meeting Abstracts). 2003;22:14a.

2. Samaan NA, et al. Cancer. 1981;47:554-560.

3. Stewart JF, et al. Eur J Cancer. 1981;17:449-453.

4. Miller KD, et al. Ann Oncol. 2003;14:1072-1077.

5. Bendell JC, et al. Cancer. 2003;97:2972-2977.

6. Crivellari D, et al. Ann Oncol. 2001;12:353-356.

Stuart M. Lichtman, MD, FACP, Associate Professor of Medicine, NYU School of Medicine, Division of Oncology; Don Monti Division of Medical Oncology, North Shore University Hospital, Manhasset, NY, is on the Editorial Board for Clinical Oncology Alert.