The Multiple Myeloma Channel is supported with funding from Sanofi (Gold) and Legend Biotech (Bronze).
VJHemOnc is an independent medical education platform. Supporters, including channel supporters, have no influence over the production of content. The levels of sponsorship listed are reflective of the amount of funding given to support the channel.
ASCO 2026 | A machine learning-based six-gene signature for risk stratification in multiple myeloma
Shahzad Raza, MD, Cleveland Clinic, Cleveland, OH, shares insight into a machine learning (ML)-based six-gene signature for risk stratification in patients with TP53-mutated multiple myeloma (MM). Dr Raza highlights that his research identified the CoxBoost model as the best-fitted model for predicting patient outcomes, suggesting that integrating machine learning with genomic and clinical data will likely be the future in myeloma care. This interview took place during the 2026 American Society of Clinical Oncology (ASCO) Meeting in Chicago, IL.
These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.
Transcript
This is a project that I work with my student and what we actually did, we looked at the TP53 mutated multiple myeloma. They have conferred a poor prognosis, poor survival. So we look at those patients first and identify the cohort who are doing the worst and have a poor survival. Then we extract the genes that actually interplay a role in those TP53-mutated multiple myelomas and their pathways, how they’ve been interconnected...
This is a project that I work with my student and what we actually did, we looked at the TP53 mutated multiple myeloma. They have conferred a poor prognosis, poor survival. So we look at those patients first and identify the cohort who are doing the worst and have a poor survival. Then we extract the genes that actually interplay a role in those TP53-mutated multiple myelomas and their pathways, how they’ve been interconnected. So we have gene expressions, we have a gene pathway. Then we use the machine learning model actually try to predict how they’re going to do over the period of years, so we use the CoxBoost analysis, LASSO analysis, and different types of machine learning algorithms, and try to predict whether the molecular profiling, and if you use the clinical profiling in that, is this a better predictor or not? And according to the scatter plot that we plotted on this machine learning model, we found that the best-fitted model is the CoxBoost model. So I think that incorporating machine learning into the genomic findings and the clinical behavior is the way to go in the future.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
Read more...
