-
PaperImproving 18F-FDG PET Quantification Through a Spatial Normalization Method
Background: Quantification of 18F-FDG PET images is useful for accurate diagnosis and evaluation of various brain diseases, including brain tumors, epilepsy, dementia, and Parkinson disease. However, accurate quantification of 18F-FDG PET images requires matched 3-dimensional T1 MRI scans of the same individuals to provide detailed information on brain anatomy. In this paper, we propose a transfer learning approach to adapt a pretrained deep neural network model from amyloid PET to spatially normalize 18F-FDG PET images without the need for 3-dimensional MRI. Methods: The proposed method is based on a deep learning model for automatic spatial normalization of 18F-FDG brain PET images, which was developed by fine-tuning a pretrained model for amyloid PET using only 103 18F-FDG PET and MR images. After training, the algorithm was tested on 65 internal and 78 external test sets. All T1 MR images with a 1-mm isotropic voxel size were processed with FreeSurfer software to provide cortical segmentation maps used to extract a ground-truth regional SUV ratio using cerebellar gray matter as a reference region. These values were compared with those from spatial normalization-based quantification methods using the proposed method and statistical parametric mapping software. Results: The proposed method showed superior spatial normalization compared with statistical parametric mapping, as evidenced by increased normalized mutual information and better size and shape matching in PET images. Quantitative evaluation revealed a consistently higher SUV ratio correlation and intraclass correlation coefficients for the proposed method across various brain regions in both internal and external datasets. The remarkably good correlation and intraclass correlation coefficient values of the proposed method for the external dataset are noteworthy, considering the dataset’s different ethnic distribution and the use of different PET scanners and image reconstruction algorithms. Conclusion: This study successfully applied transfer learning to a deep neural network for 18F-FDG PET spatial normalization, demonstrating its resource efficiency and improved performance. This highlights the efficacy of transfer learning, which requires a smaller number of datasets than does the original network training, thus increasing the potential for broader use of deep learning–based brain PET spatial normalization techniques for various clinical and research radiotracers.Keywords: brain PET, quantification, spatial normalization, glucose metabolismJournal of Nuclear Medicine August 2024, jnumed.123.267360; DOI: https://doi.org/10.2967/jnumed.123.267360Link: https://jnm.snmjournals.org/content/early/2024/08/29/jnumed.123.267360
-
PaperImpact of shortening time on diagnosis of 18F-florbetaben PET
Background: 18F-Florbetaben amyloid positron emission tomography (PET) scan is crucial for diagnosing Alzheimer’s disease, typically involving a 20 min acquisition. However, maintaining such prolonged scans can be challenging in some cases. This study explores the diagnostic impact and feasibility of reducing scan durations by comparing quantitative measures between shortened and standard scans. Additionally, we identified the optimal Centiloid threshold to distinguish between positive and negative amyloid results.Results: We analyzed 307 PET scans from our memory clinic, each followed up for a minimum of two years. The scans, conducted 90 to 110 min after approximately 300 MBq of 18F-Florbetaben injection, were categorized into four sets of 5 min durations: 5, 10, 15, and 20 min. Nuclear medicine physicians validated and rated each scan as either amyloid-positive or negative. For quantitative assessments, we employed the standardized uptake value ratio (SUVR) and Centiloid scales, comparing total SUVR and Centiloid values across five subregions (global, frontal, posterior cingulate-precuneus, lateral temporal, and parietal) using Bland–Altman analysis. Receiver operator characteristic (ROC) curves were utilized to develop optimal Centiloid thresholds. Comparing the images at 5, 10, 15, and 20 min images, SUVR and Centiloid values gradually increased with prolonged scan times. The mean SUVR difference between 5 and 20 min was 0.03 for the amyloid-positive and 0.01 for the amyloid-negative groups; Centiloid differences were 4.60 and 2.38, respectively. Additionally, no significant variation was observed in total SUVR and Centiloid values among the durations across all subregions in positive and negative groups (all p > 0.1). ROC analysis indicated that a Centiloid threshold of 21.86 at 5 min provided optimal agreement with visual assessments (AUC = 0.985, sensitivity = 0.950, specificity = 0.972), especially using the global area.Conclusions: This study demonstrated that 5 min image scans with an optimal threshold of CL = 21.86 exhibited minimal bias in SUVR and Centiloid values compared to longer scans (10, 15, and 20 min). Our findings suggest that shorter scan times are a viable and effective option for brain amyloid PET imaging in clinical settings.Keywords: Alzheimer's disease, PET, Florbetaben, Shortening time, Centiloid threshold, AmyloidEJNMMI Res 14, 114 (2024). DOI: 10.1186/s13550-024-01181-8Link: Impact of shortening time on diagnosis of 18F-florbetaben PET
-
PaperEvaluation of Fibroblast Activation Protein Expression Using 68Ga-FAPI46 PET in Hypertension-Induced Tissue Changes
Abstract: Chronic hypertension leads to injury and fibrosis in major organs. Fibroblast activation protein (FAP) is one of key molecules in tissue fibrosis, and 68Ga-labeled FAP inhibitor-46 (FAPI46) PET is a recently developed method for evaluating FAP. The aim of this study was to evaluate FAP expression and fibrosis in a hypertension model and to test the feasibility of 68Ga-FAPI46 PET in hypertension.Methods: Hypertension was induced in mice by angiotensin II infusion for 4 wk. 68Ga-FAPI46 biodistribution studies and PET scanning were conducted at 1, 2, and 4 wk after hypertension modeling, and uptake in the major organs was measured. The FAP expression and fibrosis formation of the heart and kidney tissues were analyzed and compared with 68Ga-FAPI46 uptake. Subgroups of the hypertension model underwent angiotensin receptor blocker administration and high-dose FAPI46 blocking, for comparison. As a preliminary human study, 68Ga-FAPI46 PET images of lung cancer patients were analyzed and compared between hypertension and control groups.Results: Uptake of 68Ga-FAPI46 in the heart and kidneys was significantly higher in the hypertension group than in the sham group as early as week 1 and decreased after week 2. The uptake was specifically blocked in the high-dose blocking study. Immunohistochemistry also revealed FAP expression in both heart and kidney tissues. However, overt fibrosis was observed in the heart, whereas it was absent from the kidneys. The angiotensin receptor blocker–treated group showed lower uptakein the heart and kidneys than did the hypertension group. In the pilot human study, renal uptake of 68Ga-FAPI46 significantly differed between the hypertension and control groups.Conclusions: In hypertension, FAP expression is increased in the heart and kidneys from the early phases and decreases over time. FAP expression appears to represent fibrosis activity preceding or underlying fibrotic tissue formation. 68Ga-FAPI46 PET has potential as an effective imaging method for evaluating FAP expression in progressive fibrosis by hypertension.Keywords: hypertension; FAP; 68Ga-FAPI46; PETJ Nucl Med 2024; 65:1776–1781DOI: 10.2967/jnumed.124.267489Link: Evaluation of Fibroblast Activation Protein Expression Using 68Ga-FAPI46 PET in Hypertension-Induced Tissue Changes
-
PaperFast and Accurate Amyloid Brain PET Quantification Without MRI Using Deep Neural Networks
Fast and Accurate Amyloid Brain PET Quantification Without MRI Using Deep Neural NetworksAbstractThis paper proposes a novel method for automatic quantification of amyloid PET using deep learning-based spatial normalization (SN) of PET images, which does not require MRI or CT images of the same patient. The accuracy of the method was evaluated for 3 different amyloid PET radiotracers compared with MRI-parcellation-based PET quantification using FreeSurfer. Methods: A deep neural network model used for the SN of amyloid PET images was trained using 994 multicenter amyloid PET images (367 18F-flutemetamol and 627 18F-florbetaben) and the corresponding 3-dimensional MR images of subjects who had Alzheimer disease or mild cognitive impairment or were cognitively normal. For comparison, PET SN was also conducted using version 12 of the Statistical Parametric Mapping program (SPM-based SN). The accuracy of deep learning-based and SPM-based SN and SUV ratio quantification relative to the FreeSurfer-based estimation in individual brain spaces was evaluated using 148 other amyloid PET images (64 18F-flutemetamol and 84 18F-florbetaben). Additional external validation was performed using an unseen independent external dataset (30 18F-flutemetamol, 67 18F-florbetaben, and 39 18F-florbetapir). Results: Quantification results using the proposed deep learning-based method showed stronger correlations with the FreeSurfer estimates than SPM-based SN using MRI did. For example, the slope, y-intercept, and R 2 values between SPM and FreeSurfer for the global cortex were 0.869, 0.113, and 0.946, respectively. In contrast, the slope, y-intercept, and R 2 values between the proposed deep learning-based method and FreeSurfer were 1.019, -0.016, and 0.986, respectively. The external validation study also demonstrated better performance for the proposed method without MR images than for SPM with MRI. In most brain regions, the proposed method outperformed SPM SN in terms of linear regression parameters and intraclass correlation coefficients. Conclusion: We evaluated a novel deep learning-based SN method that allows quantitative analysis of amyloid brain PET images without structural MRI. The quantification results using the proposed method showed a strong correlation with MRI-parcellation-based quantification using FreeSurfer for all clinical amyloid radiotracers. Therefore, the proposed method will be useful for investigating Alzheimer disease and related brain disorders using amyloid PET scans.Keywords: amyloid PET; deep learning; quantification; spatial normalizationJ Nucl Med. 2023 Apr;64(4):659-666. doi: 10.2967/jnumed.122.264414.Link: Fast and Accurate Amyloid Brain PET Quantification Without MRI Using Deep Neural Networks - PubMed (nih.gov)
-
PaperAccurate Automated Quantification of Dopamine Transporter PET Without MRI Using Deep Learning-based Spatial Normalization
Accurate Automated Quantification of Dopamine Transporter PET Without MRI Using Deep Learning-based Spatial NormalizationPurposeDopamine transporter imaging is crucial for assessing presynaptic dopaminergic neurons in Parkinson’s disease (PD) and related parkinsonian disorders. While 18F-FP-CIT PET offers advantages in spatial resolution and sensitivity over 123I-β-CIT or 123I-FP-CIT SPECT imaging, accurate quantification remains essential. This study presents a novel automatic quantification method for 18F-FP-CIT PET images, utilizing an artificial intelligence (AI)-based robust PET spatial normalization (SN) technology that eliminates the need for anatomical images.MethodsThe proposed SN engine consists of convolutional neural networks, trained using 213 paired datasets of 18F-FP-CIT PET and 3D structural MRI. Remarkably, only PET images are required as input during inference. A cyclic training strategy enables backward deformation from template to individual space. An additional 89 paired 18F-FP-CIT PET and 3D MRI datasets were used to evaluate the accuracy of striatal activity quantification. MRI-based PET quantification using FIRST software was also conducted for comparison. The proposed method was also validated using 135 external datasets.ResultsThe proposed AI-based method successfully generated spatially normalized 18F-FP-CIT PET images, obviating the need for CT or MRI. The striatal PET activity determined by proposed PET-only method and MRI-based PET quantification using FIRST algorithm were highly correlated, with R2 and slope ranging 0.96–0.99 and 0.98–1.02 in both internal and external datasets.ConclusionOur AI-based SN method enables accurate automatic quantification of striatal activity in 18F-FP-CIT brain PET images without MRI support. This approach holds promise for evaluating presynaptic dopaminergic function in PD and related parkinsonian disorders.Nucl Med Mol Imaging (2024). https://doi.org/10.1007/s13139-024-00869-yLink: Accurate Automated Quantification of Dopamine Transporter PET Without MRI Using Deep Learning-based Spatial Normalization | Nuclear Medicine and Molecular Imaging (springer.com)
-
PaperAge and gender effects on striatal dopamine transporter density and cerebral perfusion in individuals with non-degenerative parkinsonism: a dual-ph...
Age and gender effects on striatal dopamine transporter density and cerebral perfusion in individuals with non-degenerative parkinsonism: a dual-phase 18F-FP-CIT PET studyBackgroundDual-phase fluorine-18 labeled N-3-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl) nortropane (18F-FP-CIT) positron emission tomography (PET) scans could be used to support disorders like Parkinson’s disease (PD). Dopamine transporter (DAT) binding and cerebral perfusion are associated with ageing and gender. We investigated the effects of age and gender on non-degenerative parkinsonism, using automated quantification in striatum: specific binding ratios (SBRs) for DAT binding in delayed phase PET (dCIT) and standardized-uptake-value ratios (SUVRs) for cerebral perfusion in early phase PET (eCIT). We also examined the correlations between SBR and SUVR.MethodsThis retrospective study analyzed subjects with dual-phase 18F-FP-CIT PET scans. The eCIT images were acquired immediately post-injection, and dCIT images were taken 120 min later. With Brightonix software, automated quantification of SBRs for dCIT and SUVRs for eCIT were acquired from visually normal scans. The effects of aging and gender were assessed by regressing SBRs and SUVRs on age for both genders. The correlations between SUVRs and SBRs were evaluated.ResultsWe studied 79 subjects (34 males and 45 females). An age-related reduction in SBRs was observed in the dorsal striatum, ventral striatum, caudate nucleus, and putamen for both genders. SUVRs were found to negatively correlate with age in the dorsal striatum, ventral striatum, caudate nucleus, and putamen for males and in the dorsal striatum and caudate nucleus for females. Positive correlations between SBRs and SUVRs in the dorsal striatum, ventral striatum, caudate nucleus, and putamen for male and in the dorsal striatum, caudate nucleus, and putamen for females.ConclusionsUsing quantified values from dual-phase 18F-FP-CIT PET with a single injection, we demonstrate a negative impact of age on SBRs (DAT binding) in the striatum for both genders and SUVRs (cerebral perfusion) in the dorsal striatum and caudate nucleus for both genders and in the ventral striatum and putamen for males. Additionally, we found positive associations between SBR and SUVR values in the dorsal striatum, caudate nucleus, and putamen for both genders and in the ventral striatum for males.EJNMMI Research volume 14, Article number: 65 (2024)Link: Age and gender effects on striatal dopamine transporter density and cerebral perfusion in individuals with non-degenerative parkinsonism: a dual-phase 18F-FP-CIT PET study | EJNMMI Research | Full Text (springeropen.com)