|
Journal of Clinical Oncology, 2008 ASCO Annual Meeting Proceedings (Post-Meeting Edition).
Vol 26, No 15S (May 20 Supplement), 2008: 1040
© 2008 American Society of Clinical Oncology
A nomogram to predict subsequent brain metastasis in metastatic breast cancer (MBC) patients
N. K. Ibrahim,
B. Abdulkarim,
F. Huguet,
Z. Gabos,
L. Hsu,
O. Marpeau,
S. Uzan,
L. Pusztai,
E. A. Strom,
G. N. Hortobagyi and
R. Rouzier
University of Texas M. D. Anderson Cancer Center, Houston, TX; Cross Cancer Institute, University of Alberta, Edmonton, AB, Canada; Hopital Tenon, University of Paris VI, Paris, France
1040
Background: The development of brain metastases (BrM) is usually a terminal event in the clinical course of breast cancer patients with metastatic disease other than brain (MDOB). We hypothesized that occurrence of BrM can be predicted in patients with MDOB, thus allowing preventive treatment such as prophylactic radiation. Methods: Patients with MBC treated between 2000 and February 2007 at M. D. Anderson Cancer Center were included in this retrospective analysis. We tested 20 variables and developed a multivariate model to predict occurrence of subsequent BrM and created a nomogram that could be used for individual prediction. The model was cross-validated by bootstrapping and tested in two independent cohorts consisting of 128 patients with MDOB treated at the Cross Center Institute, Edmonton, Canada, and 91 patients with BrM treated at Hopital Tenon, France. Results: Among 2,136 patients with metastatic breast cancer, 362 subsequently developed BrM. Young age, histologic characteristics, number of metastatic lymph nodes, short disease-free survival, and number of metastatic sites were significantly and independently associated with subsequent BrM. The Nomogram will be presented at the meeting. For each parameter, points are determined on a specific scale, added, and the probability of BrM is read along the Total Points scale. The cross-validation showed that the mean error was 0.8%. In the validation set, BrM probabilities were well predicted with less than 5% difference between the predicted and observed proportions for each quartile. Conclusions: We developed a robust tool to predict subsequent BrM in patients with metastatic breast cancer. A prospective randomized study is in progress to confirm these results and test the effect of prophylactic treatment.
No significant financial relationships to disclose.
Abstract presentation from the 2008 ASCO Annual Meeting
|
