Machine Learning May Improve MF Evaluation in MPN Trials

By Andrew Moreno - Last Updated: July 10, 2024

Within a multicenter, phase II, clinical trial involving patients with myeloproliferative neoplasms (MPN), investigators assessed the utility of artificial intelligence-driven quantitative analysis for evaluating myelofibrosis (MF). Based on their findings, such analysis may offer improved detection, histological assessment, and comparison of MF within MPN trials. The investigators’ assessment was published in HemaSphere.

Advertisement

Patients in the trial had MPN with primary or secondary MF and were administered zinpentraxin alfa for fibrosis management. The trial then evaluated bone marrow response and any effects on MPN-associated anemia, thrombocytopenia, or symptoms.

During stage 2 of the trial, investigators assessed Continuous Indexing of Fibrosis (CIF) quantitative analysis of MF. They trained this machine learning system on images of the reticulin‐stained bone marrow trephine (BMT) sample slides used for the trial. The analysis produced a set of CIF scores of fibrosis severity for each slide.

From 50 patients in the trial, 157 BMT samples were collected at three timepoints. CIF was used to evaluate 142 samples. The investigators compared CIF scoring of the samples against manually performed evaluations of the same samples.

The comparison showed how the subjectivity of manual MF assessments can lead to issues and inconsistencies in MF evaluation.

“Although CIF‐based analysis is not designed to specify an MF grade, it provides an objective measure of fibrosis severity and heterogeneity within BMTs, which is beyond conventional manual grading criteria,” wrote the investigators.

They highlighted correlations between improvements in CIF score and better overall response in patients. This association “warrants further evaluation of quantitative fibrosis analysis as a surrogate for clinical response in MPN trials aiming to stabilize or reverse marrow fibrosis,” the investigators concluded.

Reference

Ryou H, Sirinukunwattana K, Wood R, et al. Quantitative analysis of bone marrow fibrosis highlights heterogeneity in myelofibrosis and augments histological assessment: an insight from a phase II clinical study of zinpentraxin alfa. Hemasphere. 2024. doi:10.1002/hem3.105

Advertisement
Advertisement
Advertisement