This work proposes an uncertainty-aware multi-view dynamic decision framework for efficient and trustworthy disease classification that provides a scalable and practical solution for balancing diagnostic accuracy and economic cost, facilitating the deployment of multi-omics models in real-world clinical settings.
Abstract
Multi-omics data integration holds great promise for precision medicine, yet its clinical adoption is hindered by high acquisition costs and the complexity of heterogeneous data representations. To address these challenges, we propose an uncertainty-aware multi-view dynamic decision framework for efficient and trustworthy disease classification. Unlike conventional static fusion strategies, our approach leverages evidential deep learning grounded in Dempster-Shafer theory to explicitly disentangle predictive confidence from epistemic uncertainty, enabling cost-sensitive and progressive inference. Specifically, omics modalities are introduced adaptively, such that additional data are only acquired when the current evidence is insufficient to support a reliable decision. At the core of the UMCA-Net, a Transformer-based multi-stream architecture with global joint cross-attention captures rich cross-modal interactions and produces Dirichlet-based evidential representations. This design allows principled uncertainty quantification and supports dynamic decision-making. We evaluate the proposed method on four benchmark multi-omics datasets (ROSMAP, LGG, BRCA, and KIPAN). Experimental results demonstrate that our model achieves state-of-the-art performance while significantly reducing data acquisition requirements. Notably, in certain cohorts, over 90% of samples can be confidently classified using only low-cost initial modalities without compromising accuracy. Overall, this work provides a scalable and practical solution for balancing diagnostic accuracy and economic cost, facilitating the deployment of multi-omics models in real-world clinical settings. Our code is available to the public at github.com/chenzhao2023/UMCA-Net.
Background: Reliability-aware integration of heterogeneous data sources remains a fundamental challenge in multimodal deep learning: prevailing fusion strategies assume uniform reliability across sources and instances, limiting their responsiveness to data-dependent trustworthiness. Methods: We introduce REM-Fuse (Reliability-aware Evidential Multimodal Fusion), an evidential deep learning (EDL) framework in which per-source Dirichlet uncertainty adaptively weights each source through dual-channel weighting, asymmetric cross-scale enhancement, and Dempster–Shafer-inspired evidence accumulation. As a case study for cancer prognosis, REM-Fuse integrates multi-scale histopathology (10×, 20×) and RNA-seq on TCGA-BRCA (n = 831) via five-fold cross-validation with subtype- and stage-stratified analyses. Results: REM-Fuse attained a concordance index of 0.715 and a 60-month time-dependent AUC of 0.729, indicating moderate discrimination and significant risk separation (log-rank p < 0.001). Adaptive source weights and per-patient uncertainty varied significantly across molecular subtypes (Kruskal–Wallis p = 0.010 and p = 0.007), indicating patient-specific rather than fixed multimodal integration. Conclusions: REM-Fuse provides a compact reliability-aware fusion strategy for cancer prognosis, although external validation is needed before broader clinical or cross-cohort generalization.
Yalu Huang, Yushuai Yuan, Wenbin Ye et al.· Mathematics· 0 citations
Background: Tumor heterogeneity arises from complex interactions among diverse biological factors, posing a major challenge for the development of robust multi-omics data integration methods. While the existing Tumor Multi-Omics pre-trained Network (TMO-Net) enables the fusion of multi-omics features into unified representations, its practical utility is constrained by issues such as missing modalities, incomplete within-omics data, and high-dimensional noise. To overcome these limitations, we propose TMO-Net+, an enhanced architecture specifically designed to improve the robustness and reliability of multi-omics modeling. Methods: TMO-Net+ introduces several coordinated architectural enhancements. First, a feature attention encoder is applied to each omics data type to reduce the influence of modality-dependent input variation. Second, we combine a gated Mixture-of-Experts (MoE) module with a Product-of Experts (PoE) mechanism to capture sample-specific contributions and enable robust inference even when partial omics data are available. Additionally, a supervised deep classification head with a tailored loss function is incorporated to enhance the separability of learned embeddings in the latent space. Results: Extensive experiments on pan-cancer datasets demonstrate that TMO-Net+ consistently outperforms the original TMO-Net, as measured by LogME scores. Furthermore, in various downstream tasks (e.g., pan-cancer classification, primary/metastatic site prediction, and prognostic modeling), TMO-Net+ achieves superior performance under partial-omics settings, which proves that it enhances the robustness and cross-cancer transferability of the multi-omics representations. Conclusions: The proposed TMO-Net+ improves the robustness and cross-cancer transferability of multi-omics representations within the evaluated TCGA cohorts. Biological interpretability analyses further show that TMO-Net+ prioritizes established cancer-driver genes, preserves cancer-dependent molecular-state information, and adaptively redistributes relative modality contributions across molecular states. By addressing modality-level missingness and modality-dependent input variation, it offers a reliable framework for integrative tumor analysis within the evaluated TCGA cohorts.
Wei Liu, Xuan Liu, Shu-Yu Zhou et al.· Genes· 0 citations
Adaptive Confidence-weighted Expansion (ACE), a novel framework to enhance the trustworthiness of multimodal fusion models, provides a more stable and robust data fusion method that facilitates the use of multimodal learning in addressing high-stakes problems.
Mohammad Raahemi, Ali Sekhavati, Alireza Maleki et al.· International Conference on...· 0 citations
Recent developments in multimodal deep learning have brought great progress to early disease detection; yet, wide-scale implementation of such models in clinics is hindered by the inherently inscrutable reasoning of existing methods. Current frameworks often employ post-hoc explanations that are not cross-modal consistent and are unable to disambiguate between causality and correlation, compromising both clinician trust and patient safety. In order to resolve these key issues, we introduce IMPACT-X, a novel Causally-Grounded Interpretable Multimodal Deep Learning Framework. IMPACT-X fuses mul-tiple heterogeneous modalities—medical imaging with Vision Transformers, medical records with Tabular Transformers, and genetic sequences with Graph Neural Networks—into a single and interpretable model.
Our framework includes a novel Causal Multimodal Fusion Layer (CMFL) which leverages cross-modal attention alignment in order to align the representation in a dynamic manner. Fur-thermore, an SCM module with DAG learning capabilities helps identify latent confounders and ensures the causally-consistent nature of the predictions. An uncertainty-aware decision-making layer estimates epistemic uncertainty through Monte Carlo Dropout in order to produce confidence scores. A unique cross-modal interpretability alignment loss function ensures coherent explanations across multiple modalities. The experimental results show that IMPACT-X achieves an SOTA performance with AUC-ROC score of 0.94, beating the best black-box baseline by 5.2%. Quantitative evaluation shows that IMPACT-X is 40% better in terms of faithfulness than traditional attention mechanism-based explanation approaches. A qualitative study with practicing medical professionals shows the benefits of causality-grounded predictions by increasing the level of physician trust in the system output. With its combination of high prediction accuracy and causal interpretability, IMPACT-X can pave the way for the development of a regulatory compliant and interpretable paradigm of medical AI that can safely be implemented in clinics, while enabling more accurate personalized medicine practices.Index Terms—Multimodal Deep Learning; Causal Inference; Interpretability; Early Disease Detection; Clinical Decision Sup-port; Genomic Integration
Tunan Shikder, Shreyanjan Neogi, Addita Rani Dash et al.· International Journal of Lat...· 0 citations
Explainability is increasingly seen as a crucial requirement in AI-based medical diagnosis, particularly in safety-critical clinical decision-making. Most existing explainability methods in healthcare operate in a post-hoc manner and are predominantly designed for unimodal data, which limits their applicability in increasingly prevalent multimodal diagnostic settings. This paper addresses the problem of self-explainable multimodal diagnosis by formulating it within the information bottleneck (IB) framework. We propose a unified learning paradigm that jointly optimizes predictive performance and modality-specific explainability by identifying the most informative elements inside each modality that contribute to diagnostic decisions. To enable tractable and stable optimization, we employ a matrix-based Renyi's $\alpha$-order entropy functional under the assumption of sufficiently expressive encoders. Extensive experiments on representative medical datasets spanning heterogeneous modalities demonstrate that the proposed method consistently achieves strong diagnostic performance, including an absolute accuracy improvement of 9.1 percentage points on the iCTCF dataset. Moreover, the learned explanations provide transparent and modality-aware insights into feature relevance, thereby improving both the explainability and generalization.
Yu-Qing Yang, Alexander Schmatz, Zhao-Zhao Ma et al.· 0 citations
The growing availability of heterogeneous clinical data has provided new prospects of data-driven clinical decision support, but at the same time, brought about challenges to do with multimodal integration, uncertainty management, and interpretability. Current AI systems in clinical settings are also likely to be based on fixed fusion methods and single-mode thinking, which restricts their resilience and reliability in practice in healthcare settings. In order to overcome these shortcomings, this paper presents a new confidence-aware hybrid design, CARE-LLM-GRAPH, which combines large language models (LLMs) to perform clinical reasoning, multimodal deep learning to analyze medical images, and population-aware graph intelligence to provide cohort-level information. The new framework clearly formulates modality-specific uncertainty and also uses a confidencesensitive adaptive fusion process to combine dynamically the text, visual and graph-based evidence. In addition, the iterative refinement process that is organized by an LLM allows an adaptive reasoning in situations when clinical conditions are ambiguous or incomplete. Most experiments performed out of publicly available multimodal clinical datasets show that CARE-LLM-GRAPH has been demonstrating consistently better discriminative performance, recommendation ranking quality, calibration, and performance under missing data conditions compared to state-of-the-art unimodal and multimodal baselines. The framework also produces interpretable evidence-based explanations, which increase the level of clinical transparency and trust. These findings underscore how CARE-LLMGRAPH can be a valid and explicable clinical decision support framework to serve practical purposes in healthcare environments.
Unknown authors· European Journal of Prosthod...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
