Tabular LLMs for Interpretable Few-Shot Alzheimer's Disease Prediction with Multimodal Biomedical Data
Authors: Sophie Kearney, Shu Yang, Zixuan Wen, Weimin Lyu, Bojian Hou, Duy Duong-Tran, Tianlong Chen, Jason H. Moore, Marylyn D. Ritchie, Chao Chen, Li Shen
Organizations: University of Pennsylvania, Philadelphia, PA, USA · Stony Brook University, Stony Brook, NY, USA · University of North Carolina at Chapel Hill, Chapel Hill, NC, USA · Cedars-Sinai Medical Center, West Hollywood, CA, USA · Medical University of South Carolina, Charleston, SC, USA
Accurate diagnosis of Alzheimer's disease (AD) requires handling tabular biomarker data, yet such data are often small and incomplete, where deep learning models frequently fail to outperform classical methods. Pretrained large language models (LLMs) offer few-shot generalization, structured reasoning, and interpretable outputs, providing a powerful paradigm shift for clinical prediction. We propose TAP-GPT Tabular Alzheimer's Prediction GPT, a domain-adapted tabular LLM framework built on TableGPT2 and fine-tuned for few-shot AD classification using tabular prompts rather than plain texts. We evaluate TAP-GPT across four ADNI-derived datasets, including QT-PAD biomarkers and region-level structural MRI, amyloid PET, and tau PET for binary AD classification. Across multimodal and unimodal settings, TAP-GPT improves upon its backbone models and outperforms traditional machine learning baselines in the few-shot setting while remaining competitive with state-of-the-art general-purpose LLMs. We show that feature selection mitigates degradation in high-dimensional inputs and that TAP-GPT maintains stable performance under simulated and real-world missingness without imputation. Additionally, TAP-GPT produces structured, modality-aware reasoning aligned with established AD biology and shows greater stability under self-reflection, supporting its use in iterative multi-agent systems. To our knowledge, this is the first systematic application of a tabular-specialized LLM to multimodal biomarker-based AD prediction, demonstrating that such pretrained models can effectively address structured clinical prediction tasks and laying the foundation for tabular LLM-driven multi-agent clinical decision-support systems. The source code is publicly available on GitHub: https://github.com/sophie-kearney/TAP-GPT.
Figures & tables
Fig. 1: Overview of the TAP-GPT framework. For each dataset, we split the subjects into training, testing, and pools for in-context examples, with an optional feature selection step. We construct tables for finetuning and evaluating TAP-GPT for the task of Alzheimer’s disease prediction.
Fig. 2: The four prompt formats used in our experiments, with tabular prompts shown in green and serialized prompts in blue. In each case, the model is asked to predict Alzheimer’s disease status for the same held-out patient. All values shown are synthetic with abbreviated prompts.
Fig. 3: QT-PAD mean F1 across models in zero-shot and few-shot ( k=8 ) contexts. LLMs use tabular (green) and serialized (blue) prompts with error bars for standard deviation; TabPFN and traditional ML (yellow) operate directly on structured data.
Fig. 4: Imaging ROI mean F1 across models in the zero- and few-shot ( k=4 ) contexts. Tabular and Serialized prompts are given to LLMs, with traditional ML and TabPFN models directly using the data. Model performance is evaluated across three imaging modalities, Amyloid PET, Tau PET, and Structural MRI.
Fig. 5: Number of ICL examples ( k ) ablation analysis on QT-PAD data across TableGPT2, TabPFN, and TAP-GPT. TabPFN performance steadily improved with larger k, TableGPT2 improved up to k=6 and declined thereafter, and TAP-GPT peaked at k=8 .
Fig. 6: Number of ICL examples ( k ) and number of features ( p ) ablation analysis on imaging data for TAP-GPT. Mean F1 score (± standard deviation across three random seeds) is shown for each ( k , p ) combination.
Fig. 7: TAP-GPT Performance using all 72 ROIs compared to selecting the top 16 features for each modality.
Fig. 8: Example interpretable outputs from TAP-GPT across all datasets. For each modality, the model correctly predicts the patient’s diagnosis and returns a structured JSON containing the prediction, confidence score, and reasoning.
Fig. 9: Regional feature importance as ranked by GPT-4.1-mini across modalities. Cortical and subcortical regions are colored according to their normalized importance scores (0–1), derived from the LLM-generated ranking of input features. Labeled regions reflect the top 10 ROIs chosen for each modality.
Fig. 10: Effect of self-reflection on GPT-4.1-mini and TAP-GPT across modalities. We compare standard prompting to a self-reflection setting in which each model reviews and potentially revises its initial prediction.
Fig. 11: Performance of multiple models on simulated missingness in the QT-PAD dataset at inference. TAP-GPT retains high performance until high missingness at 50%.
Fig. 12: Performance across real-world target missingness levels. est samples were grouped by the proportion of missing features in the target patient only, while the in-context learning (ICL) pool retained its natural missingness distribution across all patients with missing data (mean missingness = 0.247).
Alzheimer's disease (AD) is a growing global health challenge as populations age, and timely, accurate diagnosis is essential to reduce individual and societal burden. However, real-world AD assessment is hampered by incomplete, heterogeneous multimodal data and variability across sites and patient demographics. Although large language models (LLMs) have shown promise in biomedicine, their use in AD has largely been confined to answering narrow, disease-specific questions rather than generating comprehensive diagnostic reports that support clinical decision-making. Here we expand LLM capabilities for clinical decision support by introducing AD-CARE, a modality-agnostic agent that performs guideline-grounded diagnostic assessment from incomplete, heterogeneous inputs without imputing missing modalities. By dynamically orchestrating specialized diagnostic tools and embedding clinical guidelines into LLM-driven reasoning, AD-CARE generates transparent, report-style outputs aligned with real-world clinical workflows. Across six cohorts comprising 10,303 cases, AD-CARE achieved 84.9% diagnostic accuracy, delivering 4.2%-13.7% relative improvements over baseline methods. Despite cohort-level differences, dataset-specific accuracies remain robust (80.4%-98.8%), and the agent consistently outperforms all baselines. AD-CARE reduced performance disparities across racial and age subgroups, decreasing the average dispersion of four metrics by 21%-68% and 28%-51%, respectively. In a controlled reader study, the agent improved neurologist and radiologist accuracy by 6%-11% and more than halved decision time. The framework yielded 2.29%-10.66% absolute gains over eight backbone LLMs and converges their performance. These results show that AD-CARE is a scalable, practically deployable framework that can be integrated into routine clinical workflows for multimodal decision support in AD.
Wenlong Hou, Sheng Bi, Guangqian Yang +16
Department of Biomedical Engineering, The Hong Kong Polytechnic University, Hong Kong SAR, China. · Department of Radiology and Nuclear Medicine, Xuanwu Hospital, Capital Medical University, Beijing, China. · Amazon, USA. +10
Clinical decision making heavily relies on predicting the disease progression trajectory by seeking to understand patient's health status which is characterised by multimodal medical data. AI holds great potential for learning useful representations from multimodal medical data to predict disease progression and aid clinical decision making. However, development of predictive AI models is constrained by missing modalities and incomplete tabular data frequently occurring in medical datasets. In addition, disease labels alone may only provide limited supervisory signals for learning representations from high-dimensional multimodal data. Here, we present MMAP, a novel Multimodal Missing-aware Alignment Pretraining method for learning image-tabular representations from incomplete data. An image encoder is pretrained with efficient sigmoid contrastive learning combined with generative reconstruction. A tabular encoder is built upon a tabular foundation model. A missing token generator enables the two encoders to take incomplete data as input, enabling the model to be robust against missing modalities, either with missing images or missing tabular data. We evaluate the clinical usefulness of the learnt multimodal representations on two challenging longitudinal clinical tasks for Alzheimer's disease: predicting disease stage conversion and predicting amyloid status. The proposed method outperforms strong multimodal and unimodal baselines.
Fiona Kekwick, Matthew Baugh, Bernhard Kainz +2
Department of Computing, Imperial College London, London, UK · FAU Erlangen-Nürnberg, Erlangen, Germany · Department of Brain Sciences, Imperial College London, London, UK +1
Accurate prediction of conversion from Mild Cognitive Impairment (MCI) to Alzheimers Diseases (AD) is essential for early intervention, however, developing reliable conversion predictive models is difficult to develop due to limited longitudinal data availability We evaluate TabPFN (Tabular Pre-Trained Foundation Network) against traditional machine learning methods for predicting 3 year MCI to AD conversion using the TADPOLE dataset derived from ADNI. Using multimodal biomarker features extracted from demographics, APOE4, MRI volumes, CSF markers, and PET imaging, we conducted an experimental comparison across varying training set sizes (N=50 to 1000) and models including XGBoost, Random Forest, LightGBM, and Logistic Regression. TabPFN achieved one the highest performance (AUC=0.892), outperforming LightGBM (AUC=0.860) and demonstrating advantages in low data settings. At N=50 training samples, TabPFN maintained strong AUC while the traditional machine learning models struggles at small training samples. These findings demonstrate that foundation models are promising for disease prediction in data limited scenarios, such as Alzheimers diseases.
Brad Ye, Bulent Soykan, Gulsah Hancerliogullari Koksalmis +2
Department of Medicine, University of Central Florida College of Medicine, Orlando, FL, USA · NIMBUS Lab, University of Central Florida · Department of Mechanical, Industrial & Manufacturing Eng., The University of Toledo, Toledo, OH, USA +3