Evaluation of image processing and presentation methods in thallium-201 myocardial scintigraphy

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Computer Science

Background:

• Thallium-201 myocardial imaging is crucial for diagnosing heart conditions.
• Detectability of transmural infarcts can vary, impacting diagnostic accuracy.
• Computer image processing offers potential for enhancing image quality and interpretation.

Purpose of the Study:

• To evaluate thirteen computer image processing methods for thallium-201 myocardial images.
• To determine the optimal processing technique for improving infarct detection.
• To assess the impact of image processing on diagnostic accuracy and sensitivity.

Main Methods:

• A phantom simulating LAO 45 degrees thallium-201 myocardial images with variable infarcts was used.
• Eight observers interpreted images processed by thirteen different computer methods.
• Receiver operating characteristics (ROC) curves were employed to evaluate performance.

Main Results:

• A specific method involving a nine-point smoothing routine and 30% background subtraction significantly improved interpretation accuracy (P < 0.001).
• This optimal method increased overall diagnostic sensitivity without introducing false positives.
• Grey scale cycling and color scale presentation offered effective contrast enhancement.

Conclusions:

• Computer image processing, particularly smoothing and background subtraction, significantly enhances thallium-201 myocardial image interpretation.
• The identified optimal processing method improves diagnostic sensitivity for myocardial infarcts.
• Displaying processed images with grey scale cycling or a color scale on a TV monitor provides effective image evaluation.