Novel Algorithm for the Automatic Detection and Quantification of Pulmonary Fibrosis on CT Scans Download HRPP SOPs Clinical Trials Find a Researcher NJH ID: #16-09 Background Idiopathic pulmonary fibrosis (IPF) is a progressive and incurable disease characterized by scarring of the lung parenchyma. Disease progression varies between patients with some experiencing periods of relative stability, while others presenting acute progression and rapid decline. Accurate diagnosis and prediction of disease severity are essential in clinical decisions to establish effective therapeutic approaches. Pulmonary physiology such as forced vital capacity (FVC) is the standard for monitoring disease progression. However, it is an indirect measure of disease activity, lacking sensitivity to detect subtle changes in disease, and presenting a myriad of variable results depending on technique and patient’s effort. High resolution computed tomography (HRCT) is the preferred imaging modality to assess lung abnormalities noninvasively, but traditional visual assessment is hampered by inter-observer variation and is insufficiently precise for longitudinal evaluation. Difficulties in assessing lung function and disease progression has prompted interest in establishing standardized quantitative measures of fibrosis that could be applicable in clinical settings. Technology Drs. Humphries and Lynch have developed a computer algorithm, based on machine learning principles, that is capable of automatically detecting and quantifying lung fibrosis on CT scans. This artificial intelligence approach employs a data-driven texture analysis (DTA) paradigm and can produce quantitive scores for extent of fibrotic interstitial lung disease. They have shown that objective scores for fibrosis extent are associated with patient status on cross-sectional analysis and that change in extent on sequential HRCT scans can detect clinically significant change. They have also demonstrated that extent of fibrosis on baseline scans is associated with risk of progression. Further, they have shown that this technique is capable of detecting clinically meaningful early disease. Potential Applications Pulmonary specialty clinics treating patients with IPF or other forms of progressive pulmonary fibrosis could use the DTA as a standardized quantitative measurement of lung fibrosis extent and for monitoring change over time. Researchers and pharmaceutical companies pursuing therapeutics for ILDs could also use objective CT image analysis to establish drug effectiveness against lung fibrosis and disease control. State of Development Quantitative HRCT using DTA scores has been well validated in research and clinical trial cohorts. The investigators are currently confirming the applicability of this method in individual patient care and assessing the translation of DTA into clinical practice. Publications Walsh, SFL. et al. Lancet Respir. Med. 2020 Nov;8(11):1144-1153. PMID: 32109428 Wu, X. et al Am J Respir. Crit. Care Med. 2019 Jan 1;199(1):12-21. PMID: 29986154 Humphries, SM. et al. European Respiratory Journal. 2018 Sep 1;52(3): 180138. PMID: 30168753 Humphries, SM. et al. Radiology. 2017 Oct;285(1):270-8. PMID: 28493789 Patent Status US patent #10,706,533 Inventors Stephen Humphries, PhD and David Lynch, MD Licensing Status This technology is available for licensing. For Further Information, Contact: Emmanuel Hilaire, PhD Director Technology Transfer Office National Jewish Health 1400 Jackson Street, Room M206b Denver, CO 80206 Voice: 303.398.1262 HilaireE@njhealth.org