Related Experiment Videos
Pulmonary perfusion patterns and pulmonary arterial pressure
1Division of Nuclear Medicine, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 55 Fruit St, Boston, MA 02114, USA.
Radiology
|July 31, 2002
Summary
Artificial intelligence accurately estimates pulmonary arterial (PA) pressure using radionuclide perfusion images. This noninvasive method reveals crucial hemodynamic information, aiding in diagnosing pulmonary hypertension.
Area of Science:
- Radiology
- Artificial Intelligence
- Cardiovascular Imaging
Background:
- Pulmonary arterial (PA) pressure measurement is crucial for diagnosing cardiovascular conditions.
- Traditional PA pressure measurement requires invasive angiography.
- Noninvasive methods for estimating PA pressure are highly desirable.
Purpose of the Study:
- To develop and validate an artificial intelligence (AI) model for estimating PA pressure from radionuclide perfusion images.
- To determine if quantitative perfusion image parameters can predict invasively measured PA pressure.
Main Methods:
- Radionuclide perfusion images from 120 patients with normal chest radiographs were analyzed.
- An artificial neural network (ANN) was trained using statistical and boundary image parameters.
- Leave-one-out cross-validation was employed to predict PA systolic pressure.
Main Results:
- ANN predictions showed a strong correlation with measured PA systolic pressures (r = 0.846, P <.001).
- Prediction accuracy was unaffected by pulmonary embolism.
- The model effectively identified patients with and without pulmonary hypertension based on predicted PA pressures.
Conclusions:
- Quantitative analysis of radionuclide perfusion images, combined with AI, provides meaningful PA pressure estimation.
- This noninvasive approach can reveal important physiological information not evident on visual inspection.
- AI-driven image analysis offers a promising tool for noninvasive hemodynamic assessment.