Veracyte Test IDs IPF from Other Lung-Scarring Diseases Without the Need for Surgery

Veracyte Test IDs IPF from Other Lung-Scarring Diseases Without the Need for Surgery
Illustration of bronchus and alveolus covered in inorganic particles, fibroblasts,and growth factor obstructing the internal lining [Dorling Kindersley/Getty Images]

Idiopathic pulmonary fibrosis (IPF) is a deadly lung disease that causes irreversible fibrotic damage to lung tissue, which impairs breathing. Timely diagnosis is critical to slow the progression of disease before irreversible damage is done, but current technologies leave patients and doctors lacking adequate diagnostic tools. Veracyte Inc. fills the diagnostic gap with a new machine-learning derived diagnostic tool, which identifies IPF-specific lesion patterns via non-invasive bronchoscopy. This advanced technology offers vast improvements in the accuracy and confidence of IPF diagnosis, according to clinical validation and utility studies published today in The Lancet Respiratory Medicine.

“IPF is often challenging to distinguish from other interstitial lung diseases (ILDs), but timely and accurate diagnosis is critical so that patients with IPF can access therapies that may slow progression of the disease, while avoiding potentially harmful treatments,” said Ganesh Raghu, M.D., director, Center for Interstitial Lung Diseases, and professor of medicine at the University of Washington and lead author of the paper.

Yet misdiagnosis is a major problem. A study by the Pulmonary Fibrosis Foundation reports that 55% of patients are misdiagnosed at least once, with misdiagnoses delaying treatment for at least 3 years in 20%. This is in part because doctors rely on high-resolution CT (HRCT) imaging to examine lung tissue for characteristic IPF lesion patterns, but the technology is only able to identify such patterns in 43% of cases.  For the remaining 57% of patients, ambiguous or negative results requiring invasive surgery to reach a more definitive diagnosis.  Patients who are too frail for surgery may never be accurately diagnosed.

The Envisia Genomic Classifier promises to change the landscape of IPF diagnosis, by improving the sensitivity of detection from 43% with HRCT, to 70%. Its high accuracy and minimal false-positive rate (with 88% specificity) is achieved through machine learning, by training the algorithm with clinical samples derived from patients with and without IPF. In this clinical validation study, the algorithm was tested on 49 patients as part of an ongoing, 29-site, blinded BRAVE (Bronchial Sample Collection for a Novel Genomic Test) study.

Not only does the Envisia classifier offer greater accuracy and confidence than HRCT, but also than subsequent surgical histopathology. In the clinical utility arm of the study, the authors examined the cases of 94 patients and found that histopathology results only provided 56% confidence in IPF diagnosis, compared to 89% for the Envisia classifier.

“Our results with molecular classification through machine learning (the Envisia classifier) are promising—physicians may be able to utilize the molecular classification as a diagnostic tool to make a more informed and confident diagnoses through multidisciplinary discussions, along with clinical information and radiological features in HRCT imaging,” said Raghu.

Notably, the Envisia classifier offers all patients access to an accurate diagnosis, by providing these diagnostic improvements through a non-invasive and routine bronchoscopy procedure. The test is widely accessible to the estimated 100,000 people who suffer from ILDs like IPF in the United States, being commercially available and recently approved by Medicare. “This new paper, combined with the Medicare coverage policy for the Envisia classifier issued recently, will fuel our efforts to make the classifier more widely available to the patients across the country who can benefit from it,” said Bonnie Anderson, chairman and chief executive officer of Veracyte.

Veracyte has already commercialized similar tests to aid in the diagnosis of thyroid and lung cancers, which combine advanced genomic technology, machine learning and clinical science to inform diagnostic and treatment decisions for doctors and patients.