Method of treating a subject suffering from rheumatoid arthritis with anti-TNF therapy based on a trained machine learning classifier
Inventors
Ghiassian, Susan Dina • Mellors, Theodore R. • Santolini, Marc • Ameli, Asher • Schoenbrunner, Nancy • Akmaev, Viatcheslav R. • Johnson, Keith J.
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11195595-B2
Publication Date
2021-12-07
Expiration Date
Abstract
Presented herein are systems and methods for developing classifiers useful for predicting response to particular treatments. For example, in some embodiments, the present disclosure provides a method of treating subjects suffering from an autoimmune disorder, the method comprising a step of: administering an anti-TNF therapy to subjects who have been determined to be responsive via a classifier established to distinguish between responsive and non-responsive prior subjects in a cohort who have received the anti-TNF therapy.
Core Innovation
The disclosure addresses the problem that existing biomarkers and classifiers for predicting rheumatoid arthritis response to anti-TNF therapy are inadequate, motivating improved prediction of non-response versus response. It focuses on identifying discriminatory whole-blood transcriptional features associated with rheumatoid arthritis subjects treated with anti-TNF therapy.
A cross-cohort, cross-platform, gene-expression-based machine-learning framework is provided that identifies discriminatory whole-blood transcriptional features from microarray and validates them in RNAseq. The framework evaluates rheumatoid arthritis-associated SNPs detectable in RNAseq and combines transcript, SNP, and clinical variables in a Random Forest classifier to predict non-response.
The invention provides methods of treating a subject suffering from rheumatoid arthritis by administering an anti-TNF therapy when the subject has been determined to be responsive to the anti-TNF therapy. When the subject is determined to be non-responsive, the subject is administered an alternative to anti-TNF therapy, and the trained machine learning classifier distinguishes responsive and non-responsive subjects based at least in part on analyzing expression levels of selected genes including ATRAID, COMMD5, GOLGA1, IMPDH2, SPINT2, SPON2, and STOML2.
Claims Coverage
The document includes three independent claims. Across these independent claims, the inventive features focus on determining responsiveness to anti-TNF therapy using a trained machine learning classifier, using specified gene expression levels from a defined gene set as inputs, and administering either anti-TNF therapy or an alternative therapy based on the classifier determination.
Anti-TNF therapy administration based on ML responsiveness determination
A method of treating a subject suffering from rheumatoid arthritis by administering an anti-TNF therapy, wherein the subject has been determined to be responsive to the anti-TNF therapy via a trained machine learning classifier that distinguishes between responsive and non-responsive subjects who have received the anti-TNF therapy.
Gene expression level-based classifier input using selected RA genes
Responsiveness determination is based at least in part on analyzing expression levels in the subject of one or more genes selected from ATRAID, COMMD5, GOLGA1, IMPDH2, SPINT2, SPON2, and STOML2.
Alternative therapy administration for non-responsive subjects
A method of treating a subject suffering from rheumatoid arthritis by administering an alternative to anti-TNF therapy, wherein the subject has been determined to be non-responsive to the anti-TNF therapy via a trained machine learning classifier that distinguishes between responsive and non-responsive subjects who have received the anti-TNF therapy.
Receive expression levels, determine responsiveness, and administer therapy accordingly
A method comprising receiving expression levels in the subject of one or more genes selected from ATRAID, COMMD5, GOLGA1, IMPDH2, SPINT2, SPON2, and STOML2; determining whether the subject is responsive or non-responsive to an anti-TNF therapy via a trained machine learning classifier based at least in part on analyzing expression levels; and administering an alternative to the anti-TNF therapy if non-responsive or administering the anti-TNF therapy if responsive.
Across the independent claims, the claimed methods use a trained machine learning classifier to distinguish responsive versus non-responsive rheumatoid arthritis subjects who have received anti-TNF therapy, with determination based at least in part on expression levels of genes selected from ATRAID, COMMD5, GOLGA1, IMPDH2, SPINT2, SPON2, and STOML2, followed by administering anti-TNF therapy for responsive subjects or an alternative therapy for non-responsive subjects.
Stated Advantages
Improved model performance using RNA-only versus clinical covariates.
Blinded validation of the PrismRA™ biomarker panel provides strong stratification metrics for non-response prediction, including high negative predictive value and approximately 90% accuracy for non-response prediction.
Biological interpretation is provided via interactome/network medicine and pathway enrichment, including T cell signaling.
Documented Applications
Predicting rheumatoid arthritis non-response versus response to anti-TNF therapy using a cross-cohort, cross-platform, gene-expression-based machine-learning framework and a PrismRA™ biomarker panel.
Guiding treatment decisions for rheumatoid arthritis subjects by administering anti-TNF therapy to classifier-determined responsive subjects and administering an alternative to anti-TNF therapy to classifier-determined non-responsive subjects.
Predicting rheumatoid arthritis response to anti-TNF therapy and using the prediction to decide whether to administer anti-TNF therapy or an alternative therapy.
Interested in licensing this patent?