Validation of a Machine Learning Predictive Model to Distinguish Post-capillary Pulmonary Hypertension (Optiek 2-0)

Optimizing the Pulmonary Hypertension Diagnostic Network in Belgium: Validation of a Machine Learning Predictive Model to Distinguish Post-capillary Pulmonary Hypertension

In this study the diagnostic accuracy of a diagnostic tool for the diagnosis of post-capillary pulmonary hypertension will be investigated.

The diagnostic tool was designed based on artificial intelligence, using machine learning on a database of 344 patients with group 1 or group 2 pulmonary hypertension. The tool uses 20 non-invasive parameters which are derived from laboratory results, ECG, echocardiography and spirometry. Based on these parameters, the predictive tool estimates the probability of group 2 pulmonary hypertension.

During this clinical study, patients with an intermediate or high suspicion of pulmonary hypertension, with an indication for a diagnostic right heart catheterization, will be included. Patients with risk factors for group 3, 4 or 5 pulmonary hypertension will be excluded.

The necessary parameters to run the predictive model will be extracted from the patients medical file. Patients will undergo a standard of care right heart catheterization (gold standard). Afterwards the results of the predictive model will be compared to those of the right heart catheterization, to allow the assessment of the sensitivity, specificity, positive and negative predictive value of the predictive tool.

Study Overview

• Status: Recruiting
• Conditions: Pulmonary Arterial Hypertension; Pulmonary Hypertension; HFpEF - Heart Failure With Preserved Ejection Fraction
• Intervention / Treatment: Diagnostic test: Optiek predictive model

• Study Type: Observational
• Enrollment (Estimated): 100

Contacts and Locations

• Study Contact: Name: Laura Hardy, MD; Phone Number: +3216338917; Email: laura.hardy@uzleuven.be

Study Locations

• Belgium
  • Limburg
    • Genk, Limburg, Belgium, 3600
      • Recruiting
      • Ziekenhuis Oost-Limburg
      • Contact: David Ruttens, MD, PhD; Phone Number: +32477827441; Email: david.ruttens@zol.be
    • Hasselt, Limburg, Belgium, 3500
      • Recruiting
      • Jessa Hospital
      • Contact: Guido Claessens, MD, PhD; Phone Number: 0032 494690168; Email: guido.claessen@jessazh.be
  • Vlaams Brabant
    • Leuven, Vlaams Brabant, Belgium, 3000
      • Recruiting
      • UZ Leuven
      • Principal Investigator: Marion Delcroix, MD, PhD
      • Contact: Laura Hardy, MD; Phone Number: +3216338917; Email: laura.hardy@uzleuven.be
      • Contact: Marion Delcroix, MD, PhD; Phone Number: +3216346833; Email: marion.delcroix@uzleuven.be
  • West-Vlaanderen
    • Kortrijk, West-Vlaanderen, Belgium, 8500
      • Recruiting
      • AZ Groeninge
      • Contact: Mathias Leys, MD; Phone Number: +32486421591; Email: mathias.leys@azgroeninge.be
      • Principal Investigator: Mathias Leys, MD

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Probability Sample

Study Population

Patients with an intermediate to high probability of pulmonary hypertension based on an echocardiography. Patients with a probability of group 3, 4 or 5 pulmonary hypertension will be excluded.

Description

Inclusion Criteria:

1. Voluntary written informed consent of the participant or their legally authorized representative has been obtained prior to any screening procedures
2. Male or female patients of at least 18 years old.

3. Availability of the results of a basic work-up:

   1. Medical history, demographic information and clinical information (including BMI)
   2. Laboratory tests including hemoglobin, hematocrit and uric acid
   3. ECG
   4. Pulmonary function tests
   5. Echocardiography
4. Intermediate to high probability of PH based on echocardiography according to the 2022 ESC/ERS guidelines (see Figure 2 and Table 2). (1)
5. Indication for RHC according to ESC/ERS 2022 guidelines. (1)

Exclusion Criteria:

1. Evidence of significant pulmonary comorbidity based on abnormal pulmonary function tests (FEV1 below 60%) or aberrant lung parenchyma more than mild on radiological imaging.
2. Perfusion defects and ventilation mismatch on a recent V/Q scan.
3. Arterial perfusion defects on a recent thoracic CT angiography.

4. The following comorbidities associated with group 1 PH:

   1. Connective tissue disease
   2. HIV infection
   3. Portal hypertension
   4. Congenital heart disease

5. The following comorbidities associated with group 5 PH:

   1. Hematological disorders such as chronic hemolytic anemia or myeloproliferative disorders.
   2. Systemic and metabolic disorders such as pulmonary Langerhans cell histiocytosis, Gaucher disease, glycogen storage diseases, neurofibromatosis or sarcoidosis.
   3. Chronic renal failure (eGFR below 30 ml/min) with or without hemodialysis
   4. Fibrosing mediastinitis

Study Plan

How is the study designed?

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Diagnostic accuracy of the Optiek model	sensitivity, specificity, positive and negative predictive value of the model compared to a right heart catheterization (gold standard)	18 months

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Number of right heart catheterizations which could have been avoided	The number of avoidable rght heart catheterizations will be derived from the sensitivity of the model.	18 months
Right heart catheterization related adverse events	All subsequent adverse events will be registered in the eCRF.	18 months
Feasibility of the implementation of the Optiek model in clinical practice	The feasibility will be assessed using a questionnaire adressed to the physicians using the model.	18 months

Collaborators and Investigators

• Sponsor: KU Leuven
• Collaborators: Research Foundation Flanders

Study record dates

Study Major Dates

• Study Start (Actual): February 21, 2024
• Primary Completion (Estimated): August 1, 2025
• Study Completion (Estimated): February 1, 2026

Study Registration Dates

• First Submitted: April 16, 2024
• First Submitted That Met QC Criteria: May 7, 2024
• First Posted (Actual): May 8, 2024

Study Record Updates

• Last Update Posted (Actual): May 8, 2024
• Last Update Submitted That Met QC Criteria: May 7, 2024
• Last Verified: May 1, 2024

More Information

Terms related to this study

• Keywords: machine learning; artificial intelligence; pulmonary hypertension; pulmonary arterial hypertension; HFpEF; non-invasive diagnosis; diagnostic tool; post-capillary pulmonary hypertension
• Additional Relevant MeSH Terms: Cardiovascular Diseases; Vascular Diseases; Respiratory Tract Diseases; Lung Diseases; Hypertension; Pulmonary Arterial Hypertension; Hypertension, Pulmonary
• Other Study ID Numbers: S68038; CIV-23-11-044814 (Other Identifier: Eudamed)

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No