Added value of quantitative [¹⁸F]FDG-PET analysis in MRI-negative epilepsy: A simulation-based study using realistic...

Area of Science:

• Neuroimaging
• Epilepsy Research
• Medical Imaging Analysis

Background:

• Accurate localization of epileptic foci is crucial for non-lesional epilepsy management.
• Quantitative analysis of [¹⁸F]FDG-PET images holds promise for improving focus localization.
• Lack of gold standards has hindered comprehensive evaluation of quantitative [¹⁸F]FDG-PET approaches.

Purpose of the Study:

• To evaluate the diagnostic improvements offered by quantitative analysis of [¹⁸F]FDG-PET images.
• To assess the impact of quantitative results on physicians' diagnostic confidence and accuracy.
• To introduce a novel dataset of realistic simulated MRI-negative epilepsy [¹⁸F]FDG-PET studies.

Main Methods:

• Generation of 125 realistic simulated [¹⁸F]FDG-PET studies (100 with synthetic hypometabolic foci, 25 controls).
• Visual rating (VR) and quantitative results (QR) assessment by eight nuclear physicians.
• Analysis of changes in interpretation, success rate (SR), diagnostic confidence (DC), and inter-rater agreement before and after QR.

Main Results:

• Physicians modified interpretations in 31.3% of cases after QR, increasing SR from 16.3% to 61.0%.
• Overall SR improved from 49.5% (VR) to 63.5% (QR), with significant gains in challenging cases (+71.8%).
• Inter-rater agreement improved (0.273 to 0.475, p<0.001) and diagnostic confidence significantly increased (8.1% to 38.5% 'High', p<0.001).

Conclusions:

• Quantitative analysis of [¹⁸F]FDG-PET images significantly enhances diagnostic accuracy, confidence, and inter-rater agreement in epilepsy.
• Improvements are particularly notable in challenging cases, aiding the localization of hypometabolic foci.
• The study presents a valuable methodological approach using simulated PET images for quantification research in epilepsy.