Automatic quantification of defect size using normal templates: a comparative clinical study of three commercially...

Area of Science:

• Cardiology
• Nuclear Medicine
• Medical Imaging

Background:

• Myocardial infarction (MI) size is a critical prognostic indicator post-MI.
• Single-photon emission tomography (SPET) is utilized for infarct size assessment.
• Automatic quantification algorithms offer potential for standardized infarct evaluation.

Purpose of the Study:

• To compare the performance of three commercial automatic quantification algorithms for infarct size and severity.
• To evaluate the interchangeability and correlation of these algorithms with visual assessment and left ventricular ejection fraction (LVEF).

Main Methods:

• A cohort of 100 patients with remote MI underwent resting technetium-99m tetrofosmin gated SPET.
• Three algorithms were compared: Cedars-Emory quantitative analysis software (CEqual), vertical long-axis (VLAX), and 3D fitting (Perfit).
• Comparison standards included semiquantitative visual analysis and quantitative gated SPET (QGS) for LVEF.

Main Results:

• CEqual and Perfit algorithms reported significantly lower mean infarct extents compared to visual analysis and VLAX.
• Perfit yielded lower infarct extents than CEqual.
• Correlations between automatic algorithms and visual assessment were moderate, and correlations with LVEF were significantly lower than visual assessment.
• Lower correlations were observed in non-anterior infarctions and obese patients.

Conclusions:

• Automatic quantification algorithms using normal templates are not interchangeable for assessing infarct size and severity post-MI.
• These algorithms demonstrate only moderate correlation with visual analysis and LVEF.
• Clinical application requires careful consideration of algorithm choice and patient characteristics like infarct location and obesity.