Artificial Intelligence Could Enable Life-Saving Early Diagnosis and Advance the Treatment of Pulmonary Hypertension

May 30, 2023 — Thirona, a global company specialized in advanced analysis of thoracic CT images with artificial intelligence, announces its new AI-based algorithm for pulmonary artery-vein phenotyping, LungQ AVX. The promising results from multiple validation studies were presented at the ATS 2023 International Conference, 19-24 May 2023 in Washington.   

AVX is a new addition to Thirona’s AI-based lung quantification platform LungQ, allowing for objective and sensitive quantification of vascular abnormalities from non-contrast CT, with high precision. The new feature of LungQ follows many previously validated and certified solutions for quantitative assessments of various lung biomarkers that allow for an objective and precise quantifications of anatomical structures and disease patterns.  

Based on chest CT scans, AVX performs a deep characterization of the pulmonary vasculature to detect arterial and venous volume shifts in diseases such as pulmonary (arterial) hypertension (P(A)H). It separately identifies the arteries and veins from the scan and accurately quantifies vessel diameters and volumes down to the sub-voxel range, which is crucial to obtain a sensitive and precise analysis.  

Primary studies, presented the past week at the ATS 2023 International Conference in Washington, show that AVX allows for non-invasive detection of arterial and venous alterations in the lungs, potentially enabling early diagnose of P(A)H and differentiating between different vascular disease types such as PH groups (WHO classification).   

The practical implications of the potential of LungQ include diagnostic support, robust measurements of treatment outcomes, and more accurate medical and pharmaceutical research. LungQ AVX does not require a contrast enhanced CT, its sensitivity and accuracy allow for deep quantification of the pulmonary vasculature in routine CT scans without the need for invasive procedures for the patients.   

The insights provided by LungQ AVX cannot be subjectively assessed from a CT scan as the eye is not sensitive enough to pick up subtle changes in the dimensions of the pulmonary arteries and veins. AVX could potentially help PH diagnosis by detecting relevant vascular alterations way earlier than they can be picked up by current golden standard of hemodynamic measurements by right-heart catheterization.  

Leticia Gallardo Estrella, PhD., Senior Deep Learning Engineering Team Leader at Thirona: “Our expectation is that AV phenotyping will have the most noteworthy benefits in detecting pulmonary hypertension and its subtypes. PH is a rare disease with a very high underdiagnosis rate. Applying AI to help with its early detection can potentially result in slowing the disease progression in PH patients by ensuring they get the best possible treatment sooner.” Read her blog on redefining pulmonary vascular diseases to learn more. 

“It is an exciting step forward in phenotyping pulmonary vascular disease. AVX is a result of almost a decade of research and development work aimed at finding vascular phenotypes to improve the diagnosis of a multitude of vascular diseases and to identify new potential therapeutic targets. Today, we have several validation studies delivering strong evidence that AVX can potentially transform the way we diagnose and treat uncurable diseases like pulmonary (arterial) hypertension. And more studies are pending.”, says Jean-Paul Charbonnier, PhD., Chief Innovation Officer at Thirona.   

For more information: www.thirona.eu 

References 

Boomars, K. A. (2023). Can AI-based Pulmonary Vascular Phenotyping on Chest-CT Detect Volume Shifts in Pulmonary Arterial and Venous Blood Volume in Operable and Non-operable Chronic Thromboembolic Pulmonary Hypertension? In D28. HOW CAN WE DO BETTER: EMERGING DIAGNOSTICS AND THERAPEUTICS IN PULMONARY VASCULAR DISEASE (pp. A6468-A6468). American Thoracic Society.  

Maloir, Q., Gallardo Estrella, L., Ernst, B., Louis, R., Charbonnier, J. P., & Guiot, J. (2023). Artificial Intelligence-Based Analysis Differentiates PAH From PH Using Non-Contrast Chest CT Scans. In B56.-OMICS, MACHINES, AND DEVICES IN PULMONARY HYPERTENSION (pp. A3724-A3724). American Thoracic Society.  

Maloir, Q., Gallardo Estrella, L., Ernst, B., Louis, R., Charbonnier, J. P., & Guiot, J. (2023). A Step Towards Early Detection of Pulmonary Hypertension on Non-Contrast Chest CT Scans Using Artificial Intelligence. In B56.-OMICS, MACHINES, AND DEVICES IN PULMONARY HYPERTENSION (pp. A3725-A3725). American Thoracic Society.