Artificial intelligence for nuclear medicine in oncology

Gastric Emptying Scintigraphy Protocol Optimization Using Machine Learning for the Detection of Delayed Gastric Emptying

PURPOSE

The purpose of this study is to examine the feasibility of a machine learning (ML) system for optimizing a gastric emptying scintigraphy (GES) protocol for the detection of delayed gastric emptying (GE), which is considered a primary indication for the diagnosis of gastroparesis.

METHODS

An ML model was developed using the JADBio AutoML artificial intelligence (AI) platform. This model employs the percent GE at various imaging time points following the ingestion of a standardized radiolabeled meal to predict normal versus delayed GE at the conclusion of the 4 h GES study. The model was trained and tested on a cohort of 1002 patients who underwent GES using a 70/30 stratified split ratio for training vs. testing. The ML software automated the generation of optimal predictive models by employing a combination of data preprocessing, appropriate feature selection, and predictive modeling analysis algorithms.

RESULTS

The area under the curve (AUC) of the receiver operating characteristic (ROC) curve was employed to evaluate the predictive modeling performance. Several models were developed using different combinations of imaging time points as input features and methodologies to achieve optimal output. By using GE values at time points 0.5 h, 1 h, 1.5 h, 2 h, and 2.5 h as input predictors of the 4 h outcome, the analysis produced an AUC of 90.7% and a balanced accuracy (BA) of 80.0% on the test set. This performance was comparable to the training set results (AUC = 91.5%, BA = 84.7%) within the 95% confidence interval (CI), demonstrating a robust predictive capability. Through feature selection, it was discovered that the 2.5 h GE value alone was statistically significant enough to predict the 4 h outcome independently, with a slightly increased test set performance (AUC = 92.4%, BA = 83.3%), thus emphasizing its dominance as the primary predictor for delayed GE. ROC analysis was also performed for single time imaging points at 1 h and 2 h to assess their independent predictiveness of the 4 h outcome. Furthermore, the ML model was tested for its ability to predict "flipping" cases with normal GE at 1 h and 2 h that became abnormal with delayed GE at 4 h.

CONCLUSIONS

An AI/ML model was designed and trained for predicting delayed GE using a limited number of imaging time points in a 4 h GES clinical protocol. This study demonstrates the feasibility of employing ML for GES optimization in the detection of delayed GE and potentially shortening the protocol's time length without compromising diagnostic power.

Applications of artificial intelligence in dementia

The recent evolution of artificial intelligence (AI) can be considered life-changing. In particular, there is great interest in emerging hot topics in AI such as image classification and natural language processing. Our world has been revolutionized by convolutional neural networks and transformer for image classification and natural language processing, respectively. Moreover, these techniques can be used in the field of dementia. We introduce some applications of AI systems for treating and diagnosing dementia, including image-classification AI for recognizing facial features associated with dementia, image-classification AI for classifying leukoaraiosis in MRI images, object-detection AI for detecting microbleeding in MRI images, object-detection AI for support care, natural language-processing AI for detecting dementia within conversations, and natural language-processing AI for chatbots. Such AI technologies can significantly transform the future of dementia diagnosis and treatment. Geriatr Gerontol Int 2024; 24: 25-30.

Clinical application of 18F-fluorodeoxyglucose positron emission tomography/computed tomography radiomics-based machine learning analyses in the field of oncology

Machine learning (ML) analyses using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomics features have been applied in the field of oncology. The current review aimed to summarize the current clinical articles about 18F-FDG PET/CT radiomics-based ML analyses to solve issues in classifying or constructing prediction models for several types of tumors. In these studies, lung and mediastinal tumors were the most commonly evaluated lesions, followed by lymphatic, abdominal, head and neck, breast, gynecological, and other types of tumors. Previous studies have commonly shown that 18F-FDG PET radiomics-based ML analysis has good performance in differentiating benign from malignant tumors, predicting tumor characteristics and stage, therapeutic response, and prognosis by examining significant differences in the area under the receiver operating characteristic curves, accuracies, or concordance indices (> 0.70). However, these studies have reported several ML algorithms. Moreover, different ML models have been applied for the same purpose. Thus, various procedures were used in 18F-FDG PET/CT radiomics-based ML analysis in oncology, and 18F-FDG PET/CT radiomics-based ML models, which are easy and universally applied in clinical practice, would be expected to be established.

The OCDA-Net: a 3D convolutional neural network-based system for classification and staging of ovarian cancer patients using [18F]FDG PET/CT examinations

OBJECTIVE

To create the 3D convolutional neural network (CNN)-based system that can use whole-body [18F]FDG PET for recurrence/post-therapy surveillance in ovarian cancer (OC).

METHODS

In this study, 1224 image sets from OC patients who underwent whole-body [18F]FDG PET/CT at Kowsar Hospital between April 2019 and May 2022 were investigated. For recurrence/post-therapy surveillance, diagnostic classification as cancerous, and non-cancerous and staging as stage III, and stage IV were determined by pathological diagnosis and specialists' interpretation. New deep neural network algorithms, the OCDAc-Net, and the OCDAs-Net were developed for diagnostic classification and staging of OC patients using [18F]FDG PET/CT images. Examinations were divided into independent training (75%), validation (10%), and testing (15%) subsets.

RESULTS

This study included 37 women (mean age 56.3 years; age range 36-83 years). Data augmentation techniques were applied to the images in two phases. There were 1224 image sets for diagnostic classification and staging. For the test set, 170 image sets were considered for diagnostic classification and staging. The OCDAc-Net areas under the receiver operating characteristic curve (AUCs) and overall accuracy for diagnostic classification were 0.990 and 0.92, respectively. The OCDAs-Net achieved areas under the receiver operating characteristic curve (AUCs) of 0.995 and overall accuracy of 0.94 for staging.

CONCLUSIONS

The proposed 3D CNN-based models provide potential tools for recurrence/post-therapy surveillance in OC. The OCDAc-Net and the OCDAs-Net model provide a new prognostic analysis method that can utilize PET images without pathological findings for diagnostic classification and staging.

From FDG and beyond: the evolving potential of nuclear medicine

The radiopharmaceutical 2-[fluorine-18]fluoro-2-deoxy-D-glucose (FDG) has been dominantly used in positron emission tomography (PET) scans for over 20 years, and due to its vast utility its applications have expanded and are continuing to expand into oncology, neurology, cardiology, and infectious/inflammatory diseases. More recently, the addition of artificial intelligence (AI) has enhanced nuclear medicine diagnosis and imaging with FDG-PET, and new radiopharmaceuticals such as prostate-specific membrane antigen (PSMA) and fibroblast activation protein inhibitor (FAPI) have emerged. Nuclear medicine therapy using agents such as [177Lu]-dotatate surpasses conventional treatments in terms of efficacy and side effects. This article reviews recently established evidence of FDG and non-FDG drugs and anticipates the future trajectory of nuclear medicine.

J-ACCESS investigation and nuclear cardiology in Japan: implications for heart failure

While coronary heart disease remains a global cause of mortality, the prevalence of heart failure (HF) is increasing in developed countries including Japan. The continuously increasing aging population and the relatively low incidence of ischemic origins are features of the HF background in Japan. Information about nuclear cardiology practice and prognosis has accumulated, thanks to the multicenter prognostic J-ACCESS investigations (Series 1‒4) over two decades in Japan. Although the rate of hard cardiac events is lower in Japan than in the USA and Europe, similar predictors have been identified as causes of major adverse cardiac events. The highest proportion (50-75%) of major events among patients indicated for nuclear cardiology examinations in the J-ACCESS registries is severe HF requiring hospitalization. Therefore, the background and the possible reasons for the higher proportion of severe HF events in Japan require clarification. Combinations of age, myocardial perfusion defects, left ventricular dysfunction, and comorbid diabetes and chronic kidney disease are major predictors of cardiovascular events including severe HF. Although the Japanese Circulation Society has updated its clinical guidelines to incorporate non-invasive imaging modalities for diagnosing chronic coronary artery disease, the importance of risk-based approaches to optimal medical therapy and coronary revascularization is emphasized herein.

Proposal to Improve the Image Quality of Short-Acquisition Time-Dedicated Breast Positron Emission Tomography Using the Pix2pix Generative Adversarial Network

This study aimed to evaluate the ability of the pix2pix generative adversarial network (GAN) to improve the image quality of low-count dedicated breast positron emission tomography (dbPET). Pairs of full- and low-count dbPET images were collected from 49 breasts. An image synthesis model was constructed using pix2pix GAN for each acquisition time with training (3776 pairs from 16 breasts) and validation data (1652 pairs from 7 breasts). Test data included dbPET images synthesized by our model from 26 breasts with short acquisition times. Two breast radiologists visually compared the overall image quality of the original and synthesized images derived from the short-acquisition time data (scores of 1−5). Further quantitative evaluation was performed using a peak signal-to-noise ratio (PSNR) and structural similarity (SSIM). In the visual evaluation, both readers revealed an average score of >3 for all images. The quantitative evaluation revealed significantly higher SSIM (p < 0.01) and PSNR (p < 0.01) for 26 s synthetic images and higher PSNR for 52 s images (p < 0.01) than for the original images. Our model improved the quality of low-count time dbPET synthetic images, with a more significant effect on images with lower counts.

Diagnosis of Parkinson syndrome and Lewy-body disease using 123I-ioflupane images and a model with image features based on machine learning

OBJECTIVES

¹²³I-ioflupane has been clinically applied to dopamine transporter imaging and visual interpretation assisted by region-of-interest (ROI)-based parameters. We aimed to build a multivariable model incorporating machine learning (ML) that could accurately differentiate abnormal profiles on ¹²³I-ioflupane images and diagnose Parkinson syndrome or disease and dementia with Lewy bodies (PS/PD/DLB).

METHODS

We assessed ¹²³I-ioflupane images from 239 patients with suspected neurodegenerative diseases or dementia and classified them as having PS/PD/DLB or non-PS/PD/DLB. The image features of high or low uptake (F1), symmetry or asymmetry (F2), and comma- or dot-like patterns of caudate and putamen uptake (F3) were analyzed on 137 images from one hospital for training. Direct judgement of normal or abnormal profiles (F4) was also examined. Machine learning methods included logistic regression (LR), k-nearest neighbors (kNNs), and gradient boosted trees (GBTs) that were assessed using fourfold cross-validation. We generated the following multivariable models for the test database (n = 102 from another hospital): Model 1, ROI-based measurements of specific binding ratios and asymmetry indices; Model 2, ML-based judgement of abnormalities (F4); and Model 3, features F1, F2 and F3, plus patient age. Diagnostic accuracy was compared using areas under receiver-operating characteristics curves (AUC).

RESULTS

The AUC was high with all ML methods (0.92-0.96) for high or low uptake. The AUC was the highest for symmetry or asymmetry with the kNN method (AUC 0.75) and the comma-dot feature with the GBT method (AUC 0.94). Based on the test data set, the diagnostic accuracy for a diagnosis of PS/PD/DLB was 0.86 ± 0.04 (SE), 0.87 ± 0.04, and 0.93 ± 0.02 for Models 1, 2 and 3, respectively. The AUC was optimal for Model 3, and significantly differed between Models 3 and 1 (p = 0.027), and 3 and 2 (p = 0.029).

CONCLUSIONS

Image features such as high or low uptake, symmetry or asymmetry, and comma- or dot-like profiles can be determined using ML. The diagnostic accuracy of differentiating PS/PD/DLB was the highest for the multivariate model with three features and age compared with the conventional ROI-based method.

Medical Radiation Exposure Reduction in PET via Super-Resolution Deep Learning Model

In positron emission tomography (PET) imaging, image quality correlates with the injected [18F]-fluorodeoxyglucose (FDG) dose and acquisition time. If image quality improves from short-acquisition PET images via the super-resolution (SR) deep learning technique, it is possible to reduce the injected FDG dose. Therefore, the aim of this study was to clarify whether the SR deep learning technique could improve the image quality of the 50%-acquisition-time image to the level of that of the 100%-acquisition-time image. One-hundred-and-eight adult patients were enrolled in this retrospective observational study. The supervised data were divided into nine subsets for nested cross-validation. The mean peak signal-to-noise ratio and structural similarity in the SR-PET image were 31.3 dB and 0.931, respectively. The mean opinion scores of the 50% PET image, SR-PET image, and 100% PET image were 3.41, 3.96, and 4.23 for the lung level, 3.31, 3.80, and 4.27 for the liver level, and 3.08, 3.67, and 3.94 for the bowel level, respectively. Thus, the SR-PET image was more similar to the 100% PET image and subjectively improved the image quality, as compared to the 50% PET image. The use of the SR deep-learning technique can reduce the injected FDG dose and thus lower radiation exposure.