Systems and methods of diagnosing idiopathic pulmonary fibrosis
Inventors
Kennedy, Giulia C. • Diggans, James • Huang, Jing • CHOI, Yoonha • KIM, Su Yeon • Pankratz, Daniel • Pagan, Moraima
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12534763-B2
Publication Date
2026-01-27
Expiration Date
Abstract
The present invention provides methods for sequencing and analysis of nucleic acids and determining that a subject is positive for a non-usual interstitial pneumonia subtype.
Core Innovation
The invention provides a method for generating a classification of a biological sample as being positive for usual interstitial pneumonia (UIP) or non-UIP. A subject suspected of having interstitial lung disease is identified using an imaging test, a pulmonary function test, or a lung tissue analysis to produce a test result, and responsive to the test result, a biological sample is obtained from the subject as a transbronchial biopsy sample or a bronchoalveolar lavage sample.
Nucleic acid molecules derived from the biological sample are assayed to determine a level of expression of a plurality of genes. The determined gene expression levels are processed with a classifier to generate the classification of the biological sample as positive for UIP or non-UIP. The classifier differentiates between UIP and a plurality of non-UIP subtypes comprising hypersensitivity pneumonitis (HP), non-specific interstitial pneumonia (NSIP), sarcoidosis, respiratory bronchiolitis (RB), bronchiolitis, diffuse alveolar damage (DAD), and organizing pneumonia (OP).
The disclosure also includes using expert pathology diagnoses as truth data for training a classifier based on molecular signatures, and addressing smoking-status bias via covariates or a tiered smoker/non-smoker classifier. Further disclosed aspects include optional determination of sequence variants and splice variants, gene fusions, loss of heterozygosity (LOH) features, and biological pathway effect features, together with differential expression information for UIP and non-UIP samples within a smoking-only subgroup.
Claims Coverage
The independent claim and related claims cover an end-to-end classification method that combines clinical suspicion of interstitial lung disease, collection of a transbronchial biopsy or bronchoalveolar lavage sample, gene expression assaying, and classifier-based differentiation of UIP versus multiple named non-UIP subtypes. The inventive features center on the subtype-differentiating classifier and the linked workflow from suspicion to molecular classification.
Suspected interstitial lung disease screening producing a test result
Performing an imaging test, a pulmonary function test, or a lung tissue analysis on a subject to produce a test result indicating that the subject is suspected of having interstitial lung disease.
Obtaining a transbronchial biopsy or bronchoalveolar lavage sample
Responsive to the test result, obtaining a biological sample from the subject, wherein the biological sample comprises a transbronchial biopsy sample or a bronchoalveolar lavage sample.
Assaying nucleic acids to determine levels of expression of a plurality of genes
Assaying nucleic acid molecules derived from the biological sample to determine a level of expression of a plurality of genes.
Classifier processing that distinguishes UIP from multiple non-UIP subtypes
Processing the level of expression of the plurality of genes with a classifier to generate a classification of the biological sample as being positive for UIP or non-UIP, wherein the classifier differentiates between UIP and a plurality of non-UIP subtypes comprising hypersensitivity pneumonitis (HP), non-specific interstitial pneumonia (NSIP), sarcoidosis, respiratory bronchiolitis (RB), bronchiolitis, diffuse alveolar damage (DAD), and organizing pneumonia (OP).
Classifier trained using molecular signature features
Using expert pathology diagnoses as truth data for training a classifier based on molecular signatures, with optional features including sequence variants, splice variants, gene fusions, loss of heterozygosity (LOH), and biological pathway effect features.
Machine-learning-based recognition of non-smokers versus smokers using gene signatures
Using a trained machine-learning algorithm to recognize gene signatures that distinguish non-smokers from smokers, with smoking-status bias addressed via covariates or a tiered smoker/non-smoker classifier.
Assaying by reverse transcription PCR or DNA microarray hybridization
Performing the nucleic acid assaying by reverse transcription polymerase chain reaction (RT-PCR) or deoxyribonucleic acid (DNA) microarray hybridization.
The claims coverage centers on clinical suspicion of interstitial lung disease, molecular assaying from transbronchial biopsy or bronchoalveolar lavage samples, and classifier processing that distinguishes UIP from multiple named non-UIP subtypes. Dependent refinements specify training and feature types, smoker/non-smoker recognition, and assay modalities such as RT-PCR and DNA microarray hybridization.
Stated Advantages
Provides measurable classification performance distinguishing UIP versus non-UIP, including reported AUC values for microarray and RNASeq workflows.
Differentiates UIP from multiple non-UIP subtypes including hypersensitivity pneumonitis, NSIP, sarcoidosis, respiratory bronchiolitis, bronchiolitis, diffuse alveolar damage, and organizing pneumonia.
Supports binary UIP versus non-UIP labeling using ATS/2011-2013-aligned interstitial lung disease pathology terminology.
Supports differentiation using pathway enrichment and accounts for spatial heterogeneity via sample-level lobe pathology labels.
Documented Applications
Generating a classification of a biological sample as UIP-positive or non-UIP in subjects suspected of interstitial lung disease, using a transbronchial biopsy or bronchoalveolar lavage sample and gene-expression assaying.
Using the UIP-positive classification to support administration of a therapeutic intervention to treat UIP.
Interested in licensing this patent?