Risk model may predict survival among African-American women with breast cancer
A new prognostic model may be able to predict survival among African-American women with breast cancer, according to a study presented at American Association for Cancer Research Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved.
“Using gene expression data, we have developed a machine learning pattern to accurately stratify African-American [patients with breast cancer] with high and low risks [for] death, which could help inform clinical decision-making,” Shristi Bhattarai, PhD candidate in the department of biology at Georgia State University, said in a press release. “As African-American women tend to have worse breast cancer outcomes, this study will help us to identify race-based differences in this cohort, which could potentially lead to specific therapeutic regimens for African-American women with breast cancer.”
Although breast cancer incidence is similar among European-American and African-American women in the United States, age-adjusted mortality rates are 40% higher among African-American women, Bhattarai said. This is because of socioeconomic inequality combined with inherently more aggressive tumor biology among women with African ancestry.
Bhattarai and colleagues used data from the Cancer Proteome Atlas to analyze protein expression levels of 224 proteins from 754 patients with breast cancer (European -American, n = 620; African-American, n = 134).
The deep learning algorithm enabled researchers to identify a combination of four proteins for optimal prognostic prediction: Bcl-2-like protein (BAX), inositol polyphosphate 4-phosphatase type II (INPP4B), X-ray repair cross-complementing protein 1 (XRCC1), and cleaved poly (ADP-ribose) polymerase (c-PARP).
The proteins individually did not have significant prognostic value among African-American patients with breast cancer (BAX: HR = 0.67; INPP4B: HR = 0.99;; XRCC1: HR = 0.26; c-PARP: HR = 0.6375). However, within a random forest model, investigators were able stratify high-risk patients with 86% accuracy (HR = 5; P < 0.001).
“Their combined effect within the machine-learning model could identify an African-American cohort that had five times increased risk [for] death,” study co-author Sergey Klimov, PhD candidate in the department of biology at Georgia State University, said in a press release.
The model retained its prognostic ability (HR = 10.74; P = 0.0006) when considering clinicopathologic variables such as stage, age and positive lymph nodes, accurately identifying women who had a nearly 11 times increased risk for death.
Investigators were not able to stratify European-American patients with breast cancer into low- and high-risk populations, suggesting this model is only prognostic for African-American patients.
Researchers acknowledged study limitations, including a lack of validation in other cohorts. Investigators say they want to make sure the model is generalizable to different methodologies.
“We are moving toward the phase of clinical research where we can identify very specific patterns for understudied demographic groups to find high-risk patients so that they can be recruited for additional therapies,” Ritu Aneja, PhD, professor in the department of biology at Georgia State University, said in the release. “We are excited that our model has the potential to inform clinicians to prioritize African-American [patients with breast cancer] for appropriate clinical trials and also help patients make decisions about enrolling in specific clinical trials.” – by John DeRosier
Bhattarai S, et al. Abstract C101. Presented at: AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; Nov. 2-5, 2018; New Orleans.
Disclosures: Grants from NCI and NIH supported this study. Aneja, Bhattarai and Klimov report no relevant financial disclosures. Please see the abstract for all other authors’ relevant financial disclosures.