Deep Learning PET Epilepsy Detection with a Novel Symmetric Loss Convolutional Autoencoder

Smith, Rodney; Chandler, Harold L.; Alsyed, Emad; Bartley, Lee; Fielding, Patrick; Marshall, C.

doi:10.1109/nss/mic42677.2020.9507980

Cited by 2 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The next step is to assess the performance of the P-HQ PET in a clinical study, ideally in a large cohort of patients with well-localized lesions (FCDs), such as seizure-free subjects after brain surgery It will be also important to evaluate performance of nuclear medicine physicians with different levels of experience: P-HQ PET should be seen as a diagnostic support to improve reader detection and confidence, allowing non-expert readers to perform closer to expert reader performance. Another interesting perspective would be to assess the improvement of an AI based anomaly detection model ( 67 ) with the P-HQ PET compared to the standard PET.…”

Section: Discussionmentioning

confidence: 99%

PET image enhancement using artificial intelligence for better characterization of epilepsy lesions

et al. 2022

View full text Add to dashboard Cite

Introduction[18F]fluorodeoxyglucose ([18F]FDG) brain PET is used clinically to detect small areas of decreased uptake associated with epileptogenic lesions, e.g., Focal Cortical Dysplasias (FCD) but its performance is limited due to spatial resolution and low contrast. We aimed to develop a deep learning-based PET image enhancement method using simulated PET to improve lesion visualization.MethodsWe created 210 numerical brain phantoms (MRI segmented into 9 regions) and assigned 10 different plausible activity values (e.g., GM/WM ratios) resulting in 2100 ground truth high quality (GT-HQ) PET phantoms. With a validated Monte-Carlo PET simulator, we then created 2100 simulated standard quality (S-SQ) [18F]FDG scans. We trained a ResNet on 80% of this dataset (10% used for validation) to learn the mapping between S-SQ and GT-HQ PET, outputting a predicted HQ (P-HQ) PET. For the remaining 10%, we assessed Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Root Mean Squared Error (RMSE) against GT-HQ PET. For GM and WM, we computed recovery coefficients (RC) and coefficient of variation (COV). We also created lesioned GT-HQ phantoms, S-SQ PET and P-HQ PET with simulated small hypometabolic lesions characteristic of FCDs. We evaluated lesion detectability on S-SQ and P-HQ PET both visually and measuring the Relative Lesion Activity (RLA, measured activity in the reduced-activity ROI over the standard-activity ROI). Lastly, we applied our previously trained ResNet on 10 clinical epilepsy PETs to predict the corresponding HQ-PET and assessed image quality and confidence metrics.ResultsCompared to S-SQ PET, P-HQ PET improved PNSR, SSIM and RMSE; significatively improved GM RCs (from 0.29 ± 0.03 to 0.79 ± 0.04) and WM RCs (from 0.49 ± 0.03 to 1 ± 0.05); mean COVs were not statistically different. Visual lesion detection improved from 38 to 75%, with average RLA decreasing from 0.83 ± 0.08 to 0.67 ± 0.14. Visual quality of P-HQ clinical PET improved as well as reader confidence.ConclusionP-HQ PET showed improved image quality compared to S-SQ PET across several objective quantitative metrics and increased detectability of simulated lesions. In addition, the model generalized to clinical data. Further evaluation is required to study generalization of our method and to assess clinical performance in larger cohorts.

show abstract

Section: Discussionmentioning

confidence: 99%

PET image enhancement using artificial intelligence for better characterization of epilepsy lesions

et al. 2022

View full text Add to dashboard Cite

show abstract

“…They are mainly based on auto-encoders, using latent space representation of the FDG normal distribution to detect out-ofdistribution areas that could correspond to the EZ. 16,27,28 Other approaches are based on image synthesis: to predict enhanced [ 18 F]FDG PET image to facilitate visual analysis 29 or to predict the FDG normal distribution from a T1w image of the patient and compare it to the actual clinical PET. Proofs of concept of clinical relevance of last approach have been published in dementia 30 and epilepsy.…”

Section: Key Pointsmentioning

confidence: 99%

“…Recently, deep learning (DL) based methods for abnormality detection on [ 18 F]FDG PET have emerged. They are mainly based on auto‐encoders, using latent space representation of the FDG normal distribution to detect out‐of‐distribution areas that could correspond to the EZ 16,27,28 . Other approaches are based on image synthesis: to predict enhanced [ 18 F]FDG PET image to facilitate visual analysis 29 or to predict the FDG normal distribution from a T1w image of the patient and compare it to the actual clinical PET.…”

Section: Introductionmentioning

confidence: 99%

Deep‐learning predicted PET can be subtracted from the true clinical fluorodeoxyglucose PET co‐registered to MRI to identify the epileptogenic zone in focal epilepsy

Flaus,

Jung,

Ostrowky‐Coste

et al. 2023

Epilepsia Open

View full text Add to dashboard Cite

ObjectiveNormal interictal [18F]FDG‐PET can be predicted from the corresponding T1w MRI with Generative Adversarial Networks (GANs). A technique we call SIPCOM (Subtraction Interictal PET Co‐registered to MRI) can then be used to compare epilepsy patients’ predicted and clinical PET. We assessed the ability of SIPCOM to identify the Resection Zone (RZ) in patients with drug‐resistant epilepsy (DRE) with reference to visual and Statistical Parametric Mapping (SPM) analysis.MethodsPatients with complete pre‐surgical work‐up and subsequent SEEG and cortectomy were included. RZ localisation, the reference region, was assigned to one of eighteen anatomical brain regions. SIPCOM was implemented using healthy controls to train a GAN. To compare, the clinical PET coregistered to MRI was visually assessed by two trained readers, and a standard SPM analysis was performed.ResultsTwenty patients aged 17 to 50 (32±7.8) years were included, 14 (70%) with temporal lobe epilepsy (TLE). Eight (40%) were MRI‐negative. After surgery, fourteen patients (70%) had a good outcome (Engel I‐II). RZ localisation rate was 60% with SIPCOM vs. 35% using SPM (p=0.015) and vs. 85% using visual analysis (p=0.54). Results were similar for Engel I‐II patients, the RZ localisation rate was 64% with SIPCOM vs 36% with SPM. With SIPCOM localisation was correct in 67% in MRI‐positive vs. 50% in MRI‐negative patients, and 64% in TLE vs. 43% in extra‐TLE. The average number of false‐positive clusters was 2.2±1.3 using SIPCOM vs. 2.3±3.1 using SPM. All RZs localised with SPM were correctly localised with SIPCOM. In one case, PET and MRI were visually reported as negative, but both SIPCOM and SPM localised the RZ.SignificanceSIPCOM performed better than the reference computer‐assisted method (SPM) for RZ detection in a group of operated DRE patients. SIPCOM's impact on epilepsy management needs to be prospectively validated.

show abstract

Deep Learning PET Epilepsy Detection with a Novel Symmetric Loss Convolutional Autoencoder

Cited by 2 publications

References 11 publications

PET image enhancement using artificial intelligence for better characterization of epilepsy lesions

PET image enhancement using artificial intelligence for better characterization of epilepsy lesions

Deep‐learning predicted PET can be subtracted from the true clinical fluorodeoxyglucose PET co‐registered to MRI to identify the epileptogenic zone in focal epilepsy

Contact Info

Product

Resources

About