Machine Learning Helps ID Patients Who May Not Respond to Certain Diffuse Large B-Cell lymphoma Treatments in Hackensack University Medical Center Study
Transcriptome Analysis Gives Insight on Post-Treatment Survival Characteristics to Show Who May Not Respond Well to Traditional Drug Approaches
A recent study, published in Blood Cancer, applied machine learning to study the clinical outcomes of rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone therapy in patients with diffuse large B-cell lymphoma (DLBCL), with input from Hackensack University Medical Center researchers. Key results include:
- A novel approach for stratifying patients with (DLBCL) can help identify patients who may not respond well to certain therapies, but would otherwise benefit from alternative therapy and clinical trials.
- Before this study, DLBCL patients could be categorized biologically into subgroups but there was clinical overlap. Now, machine learning analysis presents an opportunity to stratify patients based on post-treatment survival predictions.
- The model successfully predicted four survival groups, which were validated using independent patient groups.
This approach uses data from the targeted transcriptome to predict survival. The model gauges the expression levels of 180 genes to reliably predict four survival subgroups.
Multivariate analysis showed that this patient stratification strategy includes various biological characteristics of DLBCL, with only TP53 mutations remaining as an independent prognostic biomarker.
Learn more about innovative cancer treatments at Hackensack University Medical Center.