Dementia is a degenerative disease that is increasingly prevalent in an aging society. Alzheimer’s disease (AD), the most common type of dementia, is best mitigated via early detection and management. Deep learning is an artificial intelligence technique that has been used to diagnose and predict diseases by extracting meaningful features from medical images. The convolutional neural network (CNN) is a representative application of deep learning, serving as a powerful tool for the diagnosis of AD. Recently, vision transformers (ViT) have yielded classification performance exceeding that of CNN in some diagnostic image classifications. Because the brain is a very complex network with interrelated regions, ViT, which captures direct relationships between images, may be more effective for brain image analysis than CNN. Therefore, we propose a method for classifying dementia images by applying 18F-Florbetaben positron emission tomography (PET) images to ViT. Data were evaluated via binary (normal control and abnormal) and ternary (healthy control, mild cognitive impairment, and AD) classification. In a performance comparison with the CNN, VGG19 was selected as the comparison model. Consequently, ViT yielded more effective performance than VGG19 in binary classification. However, in ternary classification, the performance of ViT cannot be considered excellent. These results show that it is hard to argue that the ViT model is better at AD classification than the CNN model.

¹⁸F-florbetaben (FBB) positron emission tomography is a representative imaging test that observes amyloid deposition in the brain. Compared to delay-phase FBB (dFBB), early-phase FBB shows patterns related to glucose metabolism in ¹⁸F-fluorodeoxyglucose perfusion images. The purpose of this study is to prove that classification accuracy is higher when using dual-phase FBB (dual FBB) versus dFBB quantitative analysis by using machine learning and to find an optimal machine learning model suitable for dual FBB quantitative analysis data. The key features of our method are (1) a feature ranking method for each phase of FBB with a cross-validated F1 score and (2) a quantitative diagnostic model based on machine learning methods. We compared four classification models: support vector machine, naïve Bayes, logistic regression, and random forest (RF). In composite standardized uptake value ratio, RF achieved the best performance (F1: 78.06%) with dual FBB, which was 4.83% higher than the result with dFBB. In conclusion, regardless of the two quantitative analysis methods, using the dual FBB has a higher classification accuracy than using the dFBB. The RF model is the machine learning model that best classifies a dual FBB. The regions that have the greatest influence on the classification of dual FBB are the frontal and temporal lobes.

Background: Parkinson disease (PD) presents diagnostic challenges due to its heterogeneous motor and nonmotor manifestations. Traditional machine learning (ML) approaches have been evaluated on structured clinical variables. However, the diagnostic utility of large language models (LLMs) using natural language representations of structured clinical data remains underexplored. Objective: This study aimed to evaluate the diagnostic classification performance of multiple LLMs using natural language prompts derived from structured clinical data and to compare their performance with traditional ML baselines. Methods: We reformatted structured clinical variables from the Parkinson’s Progression Markers Initiative (PPMI) dataset into natural language prompts and used them as inputs for several LLMs. Variables with high multicollinearity were removed, and the top 10 features were selected using Shapley additive explanations (SHAP)–based feature ranking. LLM performance was examined across few-shot prompting, dual-output prompting that additionally generated post hoc explanatory text as an exploratory component, and supervised fine-tuning. Logistic regression (LR) and support vector machine (SVM) classifiers served as ML baselines. Model performance was evaluated using F1-scores on both the test set and a temporally independent validation set (temporal validation set) of limited size, and repeated output generation was carried out to assess stability. Results: On the test set of 122 participants, LR and SVM trained on the 10 SHAP-selected clinical variables each achieved a macro-averaged F1-score of 0.960 (accuracy 0.975). LLMs receiving natural language prompts derived from the same variables reached comparable performance, with the best few-shot configurations achieving macro-averaged F1-scores of 0.987 (accuracy 0.992). In the temporal validation set of 31 participants, LR maintained a macro-averaged F1-score of 0.903, whereas SVM showed substantial performance degradation. In contrast, multiple LLMs sustained high diagnostic performance, reaching macro-averaged F1-scores up to 0.968 and high recall for PD. Repeated output generation across LLM conditions produced generally stable predictions, with rare variability observed across runs. Under dual-output prompting, diagnostic performance showed a reduction relative to few-shot prompting while remaining generally stable. Supervised fine-tuning of lightweight models improved stability and enabled GPT-4o-mini to achieve a macro-averaged F1-score of 0.987 on the test set, with uniformly correct predictions observed in the small temporal validation set, which should be interpreted cautiously given the limited sample size and exploratory nature of the evaluation. Conclusions: This study provides an exploratory benchmark of how modern LLMs process structured clinical variables in natural language form. While several models achieved diagnostic performance comparable to LR across both the test and temporal validation datasets, their outputs were sensitive to prompting formats, model choice, and class distributions. Occasional variability across repeated output generations reflected the stochastic nature of LLMs, and lightweight models required supervised fine-tuning for stable generalization. These findings highlight the capabilities and limitations of current LLMs in handling tabular clinical information and underscore the need for cautious application and further investigation.

BACKGROUND Parkinson disease (PD) presents diagnostic challenges due to its heterogeneous motor and nonmotor manifestations. Traditional machine learning (ML) approaches have been evaluated on structured clinical variables. However, the diagnostic utility of large language models (LLMs) using natural language representations of structured clinical data remains underexplored. OBJECTIVE This study aimed to evaluate the diagnostic classification performance of multiple LLMs using natural language prompts derived from structured clinical data and to compare their performance with traditional ML baselines. METHODS We reformatted structured clinical variables from the Parkinson’s Progression Markers Initiative (PPMI) dataset into natural language prompts and used them as inputs for several LLMs. Variables with high multicollinearity were removed, and the top 10 features were selected using Shapley additive explanations (SHAP)–based feature ranking. LLM performance was examined across few-shot prompting, dual-output prompting that additionally generated post hoc explanatory text as an exploratory component, and supervised fine-tuning. Logistic regression (LR) and support vector machine (SVM) classifiers served as ML baselines. Model performance was evaluated using F1-scores on both the test set and a temporally independent validation set (temporal validation set) of limited size, and repeated output generation was carried out to assess stability. RESULTS On the test set of 122 participants, LR and SVM trained on the 10 SHAP-selected clinical variables each achieved a macro-averaged F1-score of 0.960 (accuracy 0.975). LLMs receiving natural language prompts derived from the same variables reached comparable performance, with the best few-shot configurations achieving macro-averaged F1-scores of 0.987 (accuracy 0.992). In the temporal validation set of 31 participants, LR maintained a macro-averaged F1-score of 0.903, whereas SVM showed substantial performance degradation. In contrast, multiple LLMs sustained high diagnostic performance, reaching macro-averaged F1-scores up to 0.968 and high recall for PD. Repeated output generation across LLM conditions produced generally stable predictions, with rare variability observed across runs. Under dual-output prompting, diagnostic performance showed a reduction relative to few-shot prompting while remaining generally stable. Supervised fine-tuning of lightweight models improved stability and enabled GPT-4o-mini to achieve a macro-averaged F1-score of 0.987 on the test set, with uniformly correct predictions observed in the small temporal validation set, which should be interpreted cautiously given the limited sample size and exploratory nature of the evaluation. CONCLUSIONS This study provides an exploratory benchmark of how modern LLMs process structured clinical variables in natural language form. While several models achieved diagnostic performance comparable to LR across both the test and temporal validation datasets, their outputs were sensitive to prompting formats, model choice, and class distributions. Occasional variability across repeated output generations reflected the stochastic nature of LLMs, and lightweight models required supervised fine-tuning for stable generalization. These findings highlight the capabilities and limitations of current LLMs in handling tabular clinical information and underscore the need for cautious application and further investigation.