Toru Higaki PhD, Presenter: Nothing to Disclose

Daisuke Komoto MD, Abstract Co-Author: Nothing to Disclose

Yoko Kaichi, Abstract Co-Author: Nothing to Disclose

Koji Iida, Abstract Co-Author: Nothing to Disclose

Yutaka Hirokawa, Abstract Co-Author: Nothing to Disclose

Kazuo Awai MD, Abstract Co-Author: Research Grant, Toshiba Corporation Research Grant, Hitachi Ltd Research Grant, Bayer AG Research Consultant, DAIICHI SANKYO Group Research Grant, Eisai Co, Ltd

On 18F-FDG-PET scans epileptic foci are demonstrated as a defect in FDG uptake. However, if the lesions are bilateral or subtle their detection can be difficult. We developed a computer-aided diagnosis (CAD) system that compares "normal-" and patient FDG-PET images to identify epileptic foci.

We enrolled 16 volunteers and 6 patients undergoing 18F-FDG-PET scanning on a Discovery ST16 PET/CT scanner (GE). We administered 3 MBq/kg of 18F-FDG to all patients and volunteers. All normal- and patient FDG-PET images were normalized based on the statistical parametric mapping (SPM) PET template and a "normal" database was generated. We applied the voxel-based z-test to all images, extracted voxels with significantly decreased uptake (p<0.05), and defined clusters of neighboring extracted voxels as "lesion candidates". For each lesion candidate we calculated the likelihood value by multiplying the mean z-value with the volume. Finally, we ranked the lesion candidates according to the likelihood value and displayed them on MR images. Definitive diagnosis of the epileptic foci was by a neurosurgeon specializing in epileptology based on MRI-, FDG-PET-, magnetoencephalogram-, and clinical findings.

In 5 patients the first- or second-ranked lesion candidates identified by our CAD system coincided with epileptic foci identified by the neurosurgeon (1st candidate in 4 patients, 2nd candidate in 1). In the other patient the 12th-ranked candidate coincided with the epileptic focus; in this patient uptake in the epileptic focus was not significantly decreased.

Our CAD system accurately detected epileptic foci in which the 18F-FDG uptake was significantly decreased.

Our new CAD system may help to detect epileptic foci before surgery for epilepsy.

Higaki, T, Komoto, D, Kaichi, Y, Iida, K, Hirokawa, Y, Awai, K, Computer-aided Detection of Epileptic Foci on 18F-FDG-PET Images.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14005606.html