A Survey on Deep Learning Methods for Cancer Diagnosis Using Multimodal Data Fusion

The research progress in multimodal learning has grown rapidly over the last decade in several areas, especially in computer vision. The growing potential of multimodal data streams and deep learning algorithms has contributed to the increasing universality of deep multimodal learning. This involves the development of models capable of processing and analyzing the multimodal information uniformly. Unstructured real-world data can inherently take many forms, also known as modalities, often including visual and textual content. Extracting relevant patterns from this kind of data is still a motivating goal for researchers in deep learning. In this paper, we seek to improve the understanding of key concepts and algorithms of deep multimodal learning for the computer vision community by exploring how to generate deep models that consider the integration and combination of heterogeneous visual cues across sensory modalities. In particular, we summarize six perspectives from the current literature on deep multimodal learning, namely: multimodal data representation, multimodal fusion (i.e., both traditional and deep learning-based schemes), multitask learning, multimodal alignment, multimodal transfer learning, and zero-shot learning. We also survey current multimodal applications and present a collection of benchmark datasets for solving problems in various vision domains. Finally, we highlight the limitations and challenges of deep multimodal learning and provide insights and directions for future research.

This paper presents an approach for adaptive scheduling and robustness optimization in global logistics networks by integrating multimodal deep reinforcement learning with Internet of Things (IoT) technologies. We propose an integrated framework comprising a multimodal data fusion mechanism that synthesizes heterogeneous IoT sensor data, historical records, and contextual information; an adaptive deep reinforcement learning architecture that generates dynamic scheduling policies; and a multi-objective robust optimization method that balances operational efficiency with system resilience. The framework addresses key challenges in global logistics including demand volatility, transportation disruptions, and environmental uncertainties. Comprehensive experiments conducted on real-world logistics datasets demonstrate that our approach outperforms traditional methods with an 18.7% reduction in operational costs, 12.4% improvement in service levels, and significantly enhanced robustness under various disruption scenarios. The proposed method maintains 83% performance stability during complex disruptions compared to 51–72% for alternative approaches, while keeping computational requirements feasible for practical deployment. This research demonstrates potential contributions to AI-driven logistics operations management by showing improved supply chain performance through multimodal learning and robust optimization techniques.

Accurate prediction of WHO Type I and Type II pathological classifications in epithelial ovarian cancer (EOC) patients is critical for developing effective personalized treatment strategies. This study aims to develop a multimodal deep learning framework integrating enhanced CT (EC), non-contrast CT (NC), transvaginal ultrasound (US), and clinical data to predict EOC subtypes. Furthermore, we investigate the differential contributions of distinct modalities to model predictions through interpretable analysis. A retrospective analysis was conducted on 240 EOC patients who underwent postoperative histopathological subtyping alongside EC, NC and US. A total of 17 EOC patients from other centers were enrolled for external validation. The multimodal imaging data fusion deep learning model USECNC, integrating ResNet-50 with Cross-attention mechanisms, was designed to synergistically fuse multimodal image features. The integrated model USECNC + CL (USECNC+Clinical Data) combines clinical data and imaging features at the decision layer for final prediction. The USECNC + CL model demonstrated the best performance in predicting EOC pathological subtypes, achieving an AUC of 0.87 (5-fold cross-validation range: [0.82–0.92]). Baseline models based on EC, NC, US, and clinical data achieved AUC of 0.81, 0.69, 0.77 and 0.79. The optimal radiomics model achieved an AUC of 0.72. The USECNC + CL model also achieved good predictive performance in the external validation set (AUC = 0.850, ACC = 0.824). In the SHAP analysis, the contribution of imaging features is ranked from highest to lowest as follows: EC, US, NC. Meanwhile, compared to imaging fusion features, CA125 demonstrates a superior contribution. The integration of preoperative multimodal data into a deep learning model enables accurate prediction of pathological subtypes in EOC patients. Interpretability analyses confirmed the critical diagnostic contributions of EC vascular features and CA-125 levels. Accurate preoperative type I/II prediction can effectively assist in the formulation of chemotherapy regimens and surgical plans for patients.

Background/Objectives: Multimodal data fusion is increasingly applied in neuroinformatics to integrate heterogeneous sources of information. However, the optimal strategies for combining modalities with markedly different dimensionality, scale, and noise characteristics remain unclear. To our knowledge, this is among the first systematic and controlled benchmarks explicitly disentangling the effects of fusion strategy and feature scaling within a unified deep learning framework. Methods: Using data from 747 healthy participants from the Human Connectome Project, we evaluated multiple fusion paradigms-including early fusion, attention-based fusion, subspace-based fusion, and graph-based fusion-within a unified and reproducible framework. Importantly, we assessed how different feature scaling techniques (Standard, Min-Max, and Robust scaling) interact with fusion strategies and influence model performance. Biological sex was used as a controlled benchmark task to focus on methodological insights rather than task-specific optimization. Results: Early feature-level fusion consistently achieved the highest classification performance across all evaluated configurations. In particular, direct concatenation of cognitive and neuroimaging features combined with Standard Scaling yielded the best results (AUC-ROC = 0.96 (0.95-0.96)), outperforming unimodal baselines as well as intermediate and late fusion strategies. Conclusions: This systematic benchmark demonstrates that multimodal deep learning performance in neuroscience is driven primarily by the interaction between fusion strategy and feature scaling rather than by architectural complexity alone. By explicitly disentangling the effects of fusion level and preprocessing within a unified framework, this study provides practical methodological guidance for the design, evaluation, and reproducible deployment of multimodal deep learning models in neuroscience.

Learning from the multimodal brain imaging data attracts a large amount of attention in medical image analysis due to the proliferation of multimodal data collection. It is widely accepted that multimodal data can provide complementary information than mining from a single modality. However, unifying the image-based knowledge from the multimodal data is very challenging due to different image signals, resolution, data structure, etc.. In this study, we design a supervised deep model to jointly analyze brain morphometry and functional connectivity on the cortical surface and we name it deep multimodal brain network learning (DMBNL). Two graph-based kernels, i.e., geometry-aware surface kernel (GSK) and topology-aware network kernel (TNK), are proposed for processing the cortical surface morphometry and brain functional network. The vertex features on the cortical surface from GSK is pooled and feed into TNK as its initial regional features. In the end, the graph-level feature is computed for each individual and thus can be applied for classification tasks. We test our model on a large autism imaging dataset. The experimental results prove the effectiveness of our model.

Tumor mutational burden (TMB) is an indicator of the efficacy and prognosis of immune checkpoint therapy in colorectal cancer (CRC). In general, patients with higher TMB values are more likely to benefit from immunotherapy. Though whole-exome sequencing is considered the gold standard for determining TMB, it is difficult to be applied in clinical practice due to its high cost. There are also a few DNA panel-based methods to estimate TMB; however, their detection cost is also high, and the associated wet-lab experiments usually take days, which emphasize the need for faster and cheaper alternatives. In this study, we propose a multi-modal deep learning model based on a residual network (ResNet) and multi-modal compact bilinear pooling to predict TMB status (i.e. TMB high (TMB_H) or TMB low(TMB_L)) directly from histopathological images and clinical data. We applied the model to CRC data from The Cancer Genome Atlas and compared it with four other popular methods, namely, ResNet18, ResNet50, VGG19 and AlexNet. We tested different TMB thresholds, namely, percentiles of 10%, 14.3%, 15%, 16.3%, 20%, 30% and 50%, to differentiate TMB_H and TMB_L.For the percentile of 14.3% (i.e. TMB value 20) and ResNet18, our model achieved an area under the receiver operating characteristic curve of 0.817 after 5-fold cross-validation, which was better than that of other compared models. In addition, we also found that TMB values were significantly associated with the tumor stage and N and M stages. Our study shows that deep learning models can predict TMB status from histopathological images and clinical information only, which is worth clinical application.

BackgroundPathological complete response (pCR) to neoadjuvant systemic therapy (NAST) is an established prognostic marker in breast cancer (BC). Multimodal deep learning (DL), integrating diverse data sources (radiology, pathology, omics, clinical), holds promise for improving pCR prediction accuracy. This systematic review synthesizes evidence on multimodal DL for pCR prediction and compares its performance against unimodal DL.MethodsFollowing PRISMA, we searched PubMed, Embase, and Web of Science (January 2015–April 2025) for studies applying DL to predict pCR in BC patients receiving NAST, using data from radiology, digital pathology (DP), multi-omics, and/or clinical records, and reporting AUC. Data on study design, DL architectures, and performance (AUC) were extracted. A narrative synthesis was conducted due to heterogeneity.ResultsFifty-one studies, mostly retrospective (90.2%, median cohort 281), were included. Magnetic resonance imaging and DP were common primary modalities. Multimodal approaches were used in 52.9% of studies, often combining imaging with clinical data. Convolutional neural networks were the dominant architecture (88.2%). Longitudinal imaging improved prediction over baseline-only (median AUC 0.91 vs. 0.82). Overall, the median AUC across studies was 0.88, with 35.3% achieving AUC ≥ 0.90. Multimodal models showed a modest but consistent improvement over unimodal approaches (median AUC 0.88 vs. 0.83). Omics and clinical text were rarely primary DL inputs.ConclusionDL models demonstrate promising accuracy for pCR prediction, especially when integrating multiple modalities and longitudinal imaging. However, significant methodological heterogeneity, reliance on retrospective data, and limited external validation hinder clinical translation. Future research should prioritize prospective validation, integration underutilized data (multi-omics, clinical), and explainable AI to advance DL predictors to the clinical setting.

Multi-modal advanced deep learning architectures for breast cancer survival prediction

Language understanding is multimodal. During human communication, messages are conveyed not only by words in textual form, but also through speech patterns, gestures or facial emotions of the speakers. Therefore, it is crucial to fuse information from different modalities to achieve a joint comprehension. With the rapid progress in the deep learning field, neural networks have emerged as the most popular approach for addressing multimodal data fusion [1, 6, 7, 12]. While these models can effectively combine multimodal features by learning from data, they nevertheless lack an explicit exhibition of how different modalities are related to each other, due to the inherent low interpretability of neural networks [2]. In the meantime, Quantum Theory (QT) has given rise to principled approaches for incorporating interactions between textual features into a holistic textual representation [3, 5, 8, 10], where the concepts of superposition andentanglement have been universally exploited to formulate interactions. The advantages of those models in capturing complicated correlations between textual features have been observed. We hereby propose the research on quantum-inspired multimodal data fusion, claiming that the limitation of multimodal data fusion can be tackled by quantum-driven models. In particular, we propose to employ superposition to formulate intra-modal interactions while the interplay between different modalities is expected to be captured by entanglement measures. By doing so, the interactions within multimodal data may be rendered explicitly in a unified quantum formalism, increasing the performance and interpretability for concrete multimodal tasks. It will also expand the application domains of quantum theory to multimodal tasks where only preliminary efforts have been made [11]. We therefore aim at answering the following research question: RQ. Can we fuse multimodal data with quantum-inspired models? To answer this question, we propose to fuse multimodal data with complex-valued neural networks, motivated by the theoretical link between neural networks and quantum theory [4] and advances in complex-valued neural networks [9]. Our model begins with a separate complex-valued embedding learned for each unimodal data based on the existing works [5, 10] which inherently assumes superposition between intra-modal features. Then we construct a many-body system in entangled state for multimodal data, where cross-modality interactions are naturally reflected by entanglement measures. Quantum measurement operators are applied to the entanglement state to address a concrete multimodal task at hand. The whole process is instrumented by a complex-valued neural network, which is able to learn how multimodal features are combined from data, and at the same time explain the combination by means of quantum superposition and entanglement measures. We plan to examine our proposed models on CMU-MOSI [12] and CMU-MOSEI [1] which are benchmarking multimodal sentiment analysis datasets. The dataset targets at classifying sentiment into 2, 5 or 7 classes with the input of textual, visual and acoustic features. We expect to see comparable effectiveness to state-of-the-art models, and we will explore superposition and entanglement measures to better understand the inter-modal interactions.

Breast cancer is the most common cancer worldwide, and magnetic resonance imaging (MRI) constitutes a very sensitive technique for invasive cancer detection. When reviewing breast MRI examination, clinical radiologists rely on multimodal information, composed of imaging data but also information not present in the images such as clinical information. Most machine learning (ML) approaches are not well suited for multimodal data. However, attention-based architectures, such as Transformers, are flexible and therefore good candidates for integrating multimodal data. The aim of this study was to develop and evaluate a novel multimodal deep learning (DL) model combining ultrafast dynamic contrast-enhanced (UF-DCE) MRI images, lesion characteristics and clinical information for breast lesion classification. From 2019 to 2023, UF-DCE breast images and radiology reports of 240 patients were retrospectively collected from a single clinical center and annotated. Imaging data were constituted of volumes of interest (VOI) extracted around segmented lesions. Non-imaging data were constituted of both clinical (categorical) and geometrical (scalar) data. Clinical data were extracted from annotated reports and were associated to their corresponding lesions. We compared the diagnostic performances of traditional ML methods for non-imaging data, an image model based on the DL architecture, and a novel Transformer-based architecture, the Multimodal Sieve Transformer with Vision Transformer encoder (MMST-V). The final dataset included 987 lesions (280 benign, 121 malignant lesions, and 586 benign lymph nodes) and 1081 reports. For classification with non-imaging data, scalar data had a greater influence on performances of lesion classification (Area under the receiver operating characteristic curve (AUROC)=0.875±0.042) than categorical data (AUROC=0.680±0.060). MMST-V achieved better performances (AUROC=0.928±0.027) than classification based on non-imaging data (AUROC=0.900±0.045), and imaging data only (AUROC=0.863±0.025). The proposed MMST-V is an adaptative approach that can consider redundant information provided by multimodal information. It demonstrated better performances than unimodal methods. Results highlight that the combination of clinical patient data and detailed lesion information as additional clinical knowledge enhances the diagnostic performances of UF-DCE breast MRI.

Machine learning is frequently being leveraged to tackle problems in the health sector including utilization for clinical decision-support. Its use has historically been focused on single modal data. Attempts to improve prediction and mimic the multimodal nature of clinical expert decision-making has been met in the biomedical field of machine learning by fusing disparate data. This review was conducted to summarize the current studies in this field and identify topics ripe for future research. We conducted this review in accordance with the PRISMA extension for Scoping Reviews to characterize multi-modal data fusion in health. Search strings were established and used in databases: PubMed, Google Scholar, and IEEEXplore from 2011 to 2021. A final set of 128 articles were included in the analysis. The most common health areas utilizing multi-modal methods were neurology and oncology. Early fusion was the most common data merging strategy. Notably, there was an improvement in predictive performance when using data fusion. Lacking from the papers were clear clinical deployment strategies, FDA-approval, and analysis of how using multimodal approaches from diverse sub-populations may improve biases and healthcare disparities. These findings provide a summary on multimodal data fusion as applied to health diagnosis/prognosis problems. Few papers compared the outputs of a multimodal approach with a unimodal prediction. However, those that did achieved an average increase of 6.4% in predictive accuracy. Multi-modal machine learning, while more robust in its estimations over unimodal methods, has drawbacks in its scalability and the time-consuming nature of information concatenation.

In recent years, deep learning technology has made breakthroughs in computer vision. After using large-scale data training, the deep neural network represented by GAN is significantly better than previous technologies in image generation, including generating more reasonable, higher-definition, more complex, and more accurate images. With the continuous development of datasets, models, and applications, the fusion of different modal information, including fusion of natural language, semantic layouts, tags, edge maps, and other different modal information, to generate images has become a new demand and challenge. There are related reviews on image generation and multimodal deep learning. However, there has not been a review dedicated to multimodal deep-learning image generation to discuss the current status, existing problems, and challenges of this task. Therefore, this review proposes a survey on multimodal deep learning image generation. It aims to provide readers with an application scenario for multimodal deep learning image generation. Also, it provides readers with new multimodal deep learning image generation technologies, the relevant datasets, evaluation metrics used, and some results comparison. Finally, this article describes some of the challenges and future topics of multimodal deep learning image generation.

Fusion of multimodal medical data provides multifaceted, disease-relevant information for diagnosis or prognosis prediction modeling. Traditional fusion strategies such as feature concatenation often fail to learn hidden complementary and discriminative manifestations from high-dimensional multimodal data. To this end, we proposed a methodology for the integration of multimodality medical data by matching their moments in a latent space, where the hidden, shared information of multimodal data is gradually learned by optimization with multiple feature collinearity and correlation constrains. We first obtained the multimodal hidden representations by learning mappings between the original domain and shared latent space. Within this shared space, we utilized several relational regularizations, including data attribute preservation, feature collinearity and feature-task correlation, to encourage learning of the underlying associations inherent in multimodal data. The fused multimodal latent features were finally fed to a logistic regression classifier for diagnostic prediction. Extensive evaluations on three independent clinical datasets have demonstrated the effectiveness of the proposed method in fusing multimodal data for medical prediction modeling.

Multi-modal deep learning for landform recognition

Estimation of soil moisture content under high maize canopy coverage from UAV multimodal data and machine learning