In 2013, Dr. Susan Halabi authored the article “Prognostic Model Predicting Metastatic Castration-Resistant Prostate Cancer Survival in Men Treated With Second-Line Chemotherapy” in the Journal of National Cancer Institute (JNCI). The JNCI contacted Dr. Halabi to share the news of the large reach and impact this article has had on other scientific research. It was originally covered by 8 news outlets and currently has 33 citations, and a Web of Science (citation indexing service) score of 27.
In the study, Halabi and colleagues developed a new prognostic tool using data from a pivotal phase III trial of prostate cancer patients whose cancer returned after they had undergone a regimen of docetaxel, the standard first-round chemotherapy that is used after hormone treatments have been ineffective. Halabi used penalized regression (adaptive LASSO) as a variable selection method and it was challenging to estimate the standard error for the important selected covariates. She and her post-doctoral associate (Chen-Yen Lin) developed a method to estimate the standard error for the important predictors for adaptive LASSO. This is in press. Furthermore, the model was validated using an external dataset. The model is available online.
Using this tool, doctors can predict the survival of individual patients with advanced prostate cancer, enabling them to better and rapidly assess whether to try additional rounds of treatment or seek clinical trials. "For patients with metastatic prostate cancer who are appropriate candidates for second-line chemotherapy, this model can be helpful for guiding care. It could also be used in randomized phase II or phase III clinical trials to assign patients in risk groups" said Susan Halabi, Ph.D., professor of biostatistics and bioinformatics at Duke and lead author of the study. This work was funded by a R01 that the first author received.