Fully automatic deep learning pulmonary hypertension diagnosis using CT pulmonary angiography
13.01 - Pulmonary hypertension, pp. 3655
Abstract
Introduction: Invasive right heart catheterisation (RHC) is the gold standard for diagnosing pulmonary hypertension (PH), where a mean pulmonary artery pressure (mPAP) of >20mmHg is diagnostic for PH. CT pulmonary angiography (CTPA) is integral in the workup of suspected PH. The aim of this study is to develop an automated tool to diagnose PH from CTPA using segmented cardiac volumes to predict mPAP. Methods: A multi-structure (see figure) CTPA deep learning segmentation model of the heart and great vessels was developed using 200 patients with suspected PH who were manually segmented by a consultant cardiothoracic radiologist. The segmentation model was applied in 1000 consecutive patients who underwent CTPA and RHC within 48 hrs, (mPAP ≤20mmHg n=72/1000) at a tertiary PH referral centre. The acquired volumetric measurements were entered in multiple logistic regression in a training cohort (n=500) to identify a diagnostic CT model to predict mPAP >20mmHg. A ROC analysis was performed in the validation cohort (n=500). Results: A diagnostic model was identified including the pulmonary artery volume, the right ventricle to left ventricle (LV) mass ratio, and the LV eccentricity. The area under the ROC was 0.86. Conclusion: A fully automatic model for the diagnosis of PH using CTPA is presented. This has wide future potential application in automated PH screening and diagnostic workup.
Authors 15
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
-
Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
Radiology Department, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
Radiology Department, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
Radiology Department, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
-
Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
3DLab, Radiology Department, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
3DLab, Radiology Department, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Rob J. van der Geest Aachen Division of Image Processing Department of Radiology, Leiden University Medical Center
Leiden University · Leiden University Medical Center
Affiliation as printed
Division of Image Processing, Department of Radiology, Leiden University Medical Centre, Leiden, Netherlands
-
Royal Hallamshire Hospital · Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
Sheffield Pulmonary Vascular Disease Unit, Royal Hallamshire Hospital, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
-
Royal Hallamshire Hospital · Sheffield Teaching Hospitals NHS Foundation Trust
Affiliation as printed
Sheffield Pulmonary Vascular Disease Unit, Royal Hallamshire Hospital, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
-
Affiliation as printed
Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom
Cited by 1 stored of 1
1 result
No patents citing this paper on Lens.org (checked 2026-10-11).