Objective . Neuroimaging uncovers important information about disease in the brain. Yet in Alzheimer’s disease (AD), there remains a clear clinical need for reliable tools to extract diagnoses from neuroimages. Significant work has been done to develop deep learning (DL) networks using neuroimaging for AD diagnosis. However, no particular model has emerged as optimal. Due to a lack of direct comparisons and evaluations on independent data, there is no consensus on which modality is best for diagnostic models or whether longitudinal information enhances performance. The purpose of this work was (1) to develop a generalizable DL model to distinguish neuroimaging scans of AD patients from controls and (2) to evaluate the influence of imaging modality and longitudinal data on performance. Approach . We trained a 2-class convolutional neural network (CNN) with and without a cascaded recurrent neural network (RNN). We used datasets of 772 ( N _AD = 364, N _control = 408) 3D ¹⁸ F-FDG PET scans and 780 ( N _AD = 280, N _control = 500) T1-weighted volumetric-3D MR images (containing 131 and 144 patients with multiple timepoints) from the Alzheimer’s Disease Neuroimaging Initiative, plus an independent set of 104 ( N _AD = 63, N _NC = 41) ¹⁸ F-FDG PET scans (one per patient) for validation. Main Results . ROC analysis showed that PET-trained models outperformed MRI-trained, achieving maximum AUC with the CNN + RNN model of 0.93 ± 0.08, with accuracy 82.5 ± 8.9%. Adding longitudinal information offered significant improvement to performance on ¹⁸ F-FDG PET, but not on T1-MRI. CNN model validation with an independent ¹⁸ F-FDG PET dataset achieved AUC of 0.99. Layer-wise relevance propagation heatmaps added CNN interpretability. Significance . The development of a high-performing tool for AD diagnosis, with the direct evaluation of key influences, reveals the advantage of using ¹⁸ F-FDG PET and longitudinal data over MRI and single timepoint analysis. This has significant implications for the potential of neuroimaging for future research on AD diagnosis and clinical management of suspected AD patients.

(1) Background: The CT-based attenuation correction of SPECT images is essential for obtaining accurate quantitative images in cardiovascular imaging. However, there are still many SPECT cameras without associated CT scanners throughout the world, especially in developing countries. Performing additional CT scans implies troublesome planning logistics and larger radiation doses for patients, making it a suboptimal solution. Deep learning (DL) offers a revolutionary way to generate complementary images for individual patients at a large scale. Hence, we aimed to generate linear attenuation coefficient maps from SPECT emission images reconstructed without attenuation correction using deep learning. (2) Methods: A total of 384 SPECT myocardial perfusion studies that used ^99mTc-sestamibi were included. A DL model based on a 2D U-Net architecture was trained using information from 312 patients. The quality of the generated synthetic attenuation correction maps (ACMs) and reconstructed emission values were evaluated using three metrics and compared to standard-of-care data using Bland–Altman plots. Finally, a quantitative evaluation of myocardial uptake was performed, followed by a semi-quantitative evaluation of myocardial perfusion. (3) Results: In a test set of 66 test patients, the ACM quality metrics were MSSIM = 0.97 ± 0.001 and NMAE = 3.08 ± 1.26 (%), and the reconstructed emission quality metrics were MSSIM = 0.99 ± 0.003 and NMAE = 0.23 ± 0.13 (%). The 95% limits of agreement (LoAs) at the voxel level for reconstructed SPECT images were: [−9.04; 9.00]%, and for the segment level, they were [−11; 10]%. The 95% LoAs for the Summed Stress Score values between the images reconstructed were [−2.8, 3.0]. When global perfusion scores were assessed, only 2 out of 66 patients showed changes in perfusion categories. (4) Conclusion: Deep learning can generate accurate attenuation correction maps from non-attenuation-corrected cardiac SPECT images. These high-quality attenuation maps are suitable for attenuation correction in myocardial perfusion SPECT imaging and could obviate the need for additional imaging in standalone SPECT scanners.

Positron emission tomography (PET) imaging allows for measurement of activity concentrations of a given radiotracer in vivo . The quantitative capabilities of PET imaging are particularly important in the context of monitoring response to treatment, where quantitative changes in tracer uptake could be used as a biomarker of treatment response. Reconstruction algorithms and settings have a significant impact on PET quantification. In this work we introduce a novel harmonization methodology requiring only a simple cylindrical phantom and show that it can match the performance of more complex harmonization approaches based on phantoms with spherical inserts. Resolution and noise measurements from cylindrical phantoms are used to simulate the spherical inserts from NEMA image quality phantoms. An optimization algorithm was used to find the optimal smoothing filters for the simulated NEMA phantom images to identify those that best harmonized the PET scanners. Our methodology was tested on seven different PET models from two manufacturers installed at five institutions. Our methodology is able to predict contrast recovery coefficients (CRCs) from NEMA phantoms with errors within ±5.2% for CRCmax and ±3.7% for CRCmean (limits of agreement = 95%). After applying the proposed harmonization protocol, all the CRC values were within the tolerances from EANM. Quantitative harmonization in compliance with the EARL FDG-PET/CT accreditation program is achieved in a simpler way, without the need of NEMA phantoms. This may lead to simplified scanner harmonization workflows more accessible to smaller institutions.

Purpose. To investigate image intensity histograms as a potential source of useful imaging biomarkers in both a clinical example of detecting immune-related colitis (irColitis) in ¹⁸ F-FDG PET/CT images of immunotherapy patients and an idealized case of classifying digital reference objects (DRO). Methods. Retrospective analysis of bowel ¹⁸ F-FDG uptake in N = 40 patients receiving immune checkpoint inhibitors was conducted. A CNN trained to segment the bowel was used to generate the histogram of bowel ¹⁸ F-FDG uptake, and percentiles of the histogram were considered as potential metrics for detecting inflammation associated with irColitis. A model of the colon was also considered using cylindrical DRO. Classification of DRO with different intensity distributions was undertaken under varying geometry and noise settings. Results. The most predictive biomarker of irColitis was the 95th percentile of the bowel SUV histogram (SUV _95% ). Patients later diagnosed with irColitis had a significantly higher increase in SUV _95% from baseline to first on-treatment PET than patients who did not experience irColitis (p = 0.02). An increase in SUV _95% > + 40% separated pre-irColitis change from normal variability with a sensitivity of 75% and specificity of 88%. Furthermore, histogram percentiles were ideal metrics for classifying ‘hot center’ and ‘cold center’ DRO, and were robust to varying DRO geometry and noise, and to the presence of spoiler volumes unrelated to the detection task. Conclusions. The 95th percentile of the bowel SUV histogram was the optimal metric for detecting irColitis on ¹⁸ F-FDG PET/CT. Image intensity histograms are a promising source of imaging biomarkers for clinical tasks.

Image quality in positron emission tomography (PET) is limited by the number of detected photons. Heavier patients present higher photon attenuation levels, thus increasing image noise. In this work, we propose a new method that uses the combined patient attenuation/system matrix together with a tracer uptake prediction model to optimize scan times for different bed positions in whole body scans. Our main goal is to achieve consistent noise levels across patients and anatomical regions. We propose to optimize scan times for individual bed positions, for patients of any size, based on the scanner sensitivity and patient-specific attenuation. Variable scan times for every bed position were determined by combining the system matrix, derived from the computed tomography (CT) and the scanner-specific geometric sensitivity profiles, and estimations of the global tracer uptake for each patient. The method was validated with anthropomorphic phantoms and whole-body patient ¹⁸ F-FDG PET/CT scans, where variable and fixed times were compared. Phantom experiments showed that the proposed method was successful in keeping noise level constant for different attenuation setups. In real patients, image noise variability was reduced to less than one-half compared with conventional fixed-time scans at the expense of a four-fold increase in scan times between the biggest and smallest patients. Our method can homogenize image quality not only across patients of different sizes but also across different bed positions of the same patient.