Postoperative delirium affects 10–60% of elderly surgical patients and is linked to longer hospital stays, cognitive decline, and increased mortality. Although machine learning models have been developed to predict this condition using perioperative data, most rely on point predictions that fail to express uncertainty, limiting their clinical reliability in high-stakes surgical decision-making. These models often report a single risk estimate without indicating whether predictions are supported by strong or sparse evidence, which can lead to overconfidence and potential patient harm in vulnerable populations with heterogeneous frailty and comorbidity profiles. We argue that Bayesian deep learning is essential for postoperative delirium prediction because it provides distributional outputs and uncertainty estimates that allow clinicians to assess prediction reliability. Incorporating uncertainty quantification can transform these models from opaque tools into clinically trustworthy decision aids. We recommend that uncertainty reporting be required in all predictive models for postoperative delirium and that regulatory and publication standards enforce the use of Bayesian approaches. Overall, replacing point estimates with distributional predictions is necessary to improve safety and clinical utility in perioperative care of elderly patients.
Acute ischemic stroke prediction from electronic health record time series data holds significant potential for enabling early intervention and reducing long-term disability. LSTMs have been widely used to model clinical sequences such as vital signs and laboratory trends, showing strong performance in stroke-related prediction tasks from 2018–2022. However, their sequential nature limits scalability and long-range dependency modeling in large EHR datasets. Transformers, despite transforming sequence modeling in other domains since 2017, remain underused in stroke prediction compared to LSTMs. Although early healthcare studies suggest potential benefits of attention-based models, robust validation in acute ischemic stroke contexts is still limited. Transformers offer advantages in parallel processing, long-range dependency modeling, and interpretability, but require more data and computational resources. They are likely to complement rather than replace LSTMs, with hybrid architectures providing a balanced solution for clinical time series analysis. Key themes include long-range dependency capture, parallel computation, interpretability, and data efficiency trade-offs between LSTMs and transformers. Hybrid LSTM–transformer models may offer improved performance and practicality for stroke prediction, with model selection depending on data scale and clinical constraints. Further benchmarking is needed to determine when transformers or hybrid models outperform LSTMs, guiding the development of more effective stroke prediction systems.
Pancreatic cancer is highly lethal, and surgical resection is the only curative option. Preoperative assessment using contrast-enhanced CT is essential for determining tumor resectability based on involvement of key vessels such as the superior mesenteric artery, celiac trunk, and portal vein. Accurate pancreatic tumor segmentation is difficult due to unclear boundaries, low contrast with surrounding tissue, and proximity to major vessels. Manual segmentation is slow, subjective, and inconsistent, especially in borderline cases, while tumor-associated fibrosis further obscures lesion margins. We propose a deep learning-based framework using an attention-enhanced U-Net with multi-scale feature fusion and deep supervision for tumor segmentation and resectability assessment. The model incorporates attention gates, atrous spatial pyramid pooling, and auxiliary losses at multiple decoder levels to improve feature learning and gradient flow. A pre-trained encoder extracts hierarchical features refined by attention mechanisms in skip connections. A multi-scale decoder reconstructs segmentation maps, supported by deep supervision at different resolutions. A parallel branch models tumor–vessel spatial relationships using distance maps to improve resectability classification. This framework enables automated pancreatic tumor segmentation and resectability evaluation from CT scans, improving accuracy, interpretability, and clinical utility. Validation on datasets such as Pancreas-CT and Medical Segmentation Decathlon is recommended.
Oncology drug development is an expensive and high-failure process, with costs exceeding two billion dollars per approved drug and success rates below 10%. Deep learning has recently been explored as a strategy to improve efficiency across the drug discovery pipeline. This systematic review evaluates its application in target identification, compound screening and de novo drug design, and clinical trial optimization. Following PRISMA 2020 guidelines, multiple databases were searched and studies were screened using predefined inclusion criteria, with risk of bias assessed via established tools. The literature shows that graph neural networks and transformer-based models are the most widely used architectures, particularly in early-stage discovery tasks. Although many studies report strong in silico performance, often with AUC values above 0.80, only a small proportion demonstrate experimental or clinical validation. Overall, deep learning significantly advances computational drug discovery in oncology, but translation into clinically validated therapies remains limited, especially in trial optimization, highlighting the need for stronger prospective and experimental validation frameworks.
Sleep disorders, including obstructive sleep apnea, insomnia, restless legs syndrome, narcolepsy, and central sleep apnea, represent a major public health burden. Polysomnography is the diagnostic gold standard but is resource-intensive, leading to increasing use of home sleep apnea testing and wearable devices to improve accessibility. This systematic review evaluates deep learning models in sleep medicine across polysomnography, home sleep apnea testing, and wearable data, focusing on architectures, signal types, validation approaches, diagnostic tasks, and clinical readiness. A PRISMA 2020–compliant search was conducted in PubMed, IEEE Xplore, Scopus, and Web of Science for studies published from 2017 to 2025, including those applying deep learning for sleep staging, apnea/hypopnea detection, or sleep disorder diagnosis using PSG, HSAT, or wearable-derived signals. Twenty-nine studies were included. Convolutional neural networks were the most widely used architecture, often combined with recurrent or hybrid models for temporal dependencies, while transformer-based models have recently emerged for long-sequence sleep analysis. Deep learning methods demonstrate strong performance in sleep staging and respiratory event detection, especially using polysomnography data. However, limited external validation, heterogeneous datasets, and a lack of prospective clinical deployment remain major barriers to clinical translation.
Rare pediatric tumors like sarcomas, neuroblastoma, medulloblastoma, and retinoblastoma pose a challenge for developing deep learning models due to the limited availability of histopathology images, which are distributed across multiple institutions. This scarcity is compounded by privacy concerns, as whole-slide images often contain sensitive clinical and genomic data, and generative adversarial networks (GANs) risk memorizing and leaking training samples. To address this, a differentially private GAN framework is proposed for synthesizing high-resolution histopathology patches of rare pediatric cancers. The framework incorporates a generator for image synthesis, a discriminator for realism assessment, per-sample gradient clipping, Gaussian noise injection, and a privacy accountant, ensuring provable privacy guarantees during the training process. The synthetic images generated can aid in data augmentation, model pre-training, and benchmarking without exposing identifiable pathology data, offering a privacy-preserving solution for dataset augmentation while emphasizing the importance of clinical validation.
Anticoagulation management requires balancing multiple factors such as bleeding risk, thromboembolic risk, drug interactions, and renal function. Deep learning can assist in risk prediction, but its effectiveness relies on clinicians' ability to understand and verify the recommendations. Black-box models may recommend actions without providing clear explanations. In contrast, clinical guidelines are rule-based but not directly executable by neural models. This article introduces a neuro-symbolic XAI framework that combines deep learning predictions with explicit clinical guidelines. It includes a neural prediction module, a symbolic reasoning engine, and an integration layer for traceable justifications. The neuro-symbolic approach connects data-driven predictions to clinical rules, improving auditability and trustworthiness in decision support. This framework aims to enhance anticoagulation management by providing verifiable, clinician-understandable decision support, focusing on explainability-by-design.
Inpatient falls in medical-surgical units remain frequent, clinically serious, and difficult to prevent using periodic risk assessment alone. Static scales can support bedside awareness but may miss rapidly changing patient conditions. Existing approaches often fail to integrate nursing narratives, medication burden, mobility scores, bed-exit alarm activity, and room-level environmental hazards. These signals are usually documented in separate systems and are not continuously synthesized into fall risk estimates. This article proposes a multimodal deep learning model to predict the probability of an inpatient fall within the next 24 hours. The model is designed for medical-surgical units and uses both structured and unstructured clinical inputs. The proposed architecture uses a late-fusion design with a clinical text encoder for nursing progress notes and a structured-feature subnetwork for medication burden, mobility assessment scores, alarm logs, and environmental indicators. A final risk-scoring layer would generate a dynamic probability estimate suitable for clinical decision support. Conceptually, the model would produce an updated fall risk score that reflects subtle language cues, recent medication changes, impaired mobility, repeated bed-exit activity, and modifiable room hazards. The score would support continuous surveillance rather than replacing nursing judgment. A multimodal deep learning model could help shift inpatient fall prevention from episodic screening toward continuous, data-driven monitoring. Silent validation and careful workflow integration would be essential before clinical activation.
High-risk transitions of care remain a major source of preventable readmissions, emergency department use, medication harm, and post-discharge deterioration. Existing risk tools often simplify the transition period into structured clinical variables and may not fully represent clinical complexity, social vulnerability, medication safety, or home support. Discharge planning commonly depends on generic risk scores, clinician judgment, and incomplete structured fields. These approaches may overlook risk signals embedded in discharge summaries, medication reconciliation records, follow-up plans, case management documentation, and social risk screening. This manuscript proposes a conceptual deep learning model for predicting high-risk transitions of care. The model is designed to fuse discharge summary narratives, medication reconciliation data, follow-up appointment status, home support indicators, and social risk variables into a unified transition risk score. The proposed architecture combines a clinical language encoder for discharge summaries with a structured-data network for medication, follow-up, home support, and social risk features. These representations are integrated through a multimodal fusion layer that would generate a patient-level risk estimate at the point of discharge. Conceptually, the model could identify patients at elevated risk of readmission, emergency department use, or adverse post-discharge outcomes who might be missed by traditional scores. It would also be expected to highlight clinically interpretable risk drivers that could support targeted transitional care planning. A multimodal deep learning model could strengthen precision transitional care by integrating clinical, logistical, medication-related, and social dimensions of risk. Such a model could help discharge teams prioritize intensified interventions for patients most vulnerable to unsafe care transitions.
Clinical service line margins are increasingly pressured by value-based reimbursement, payer-specific contracting, and operational constraints that alter the relationship between clinical activity and financial performance. Traditional forecasting remains largely anchored in static budgets and delayed variance reports rather than continuously updated clinical and operational signals. Current financial planning tools often cannot dynamically incorporate evolving case mix, payer contracts, length of stay, resource consumption, and throughput patterns before deviations appear in the general ledger. As a result, service line leaders may recognize margin deterioration only after financial corrective action is already delayed. This article proposes a deep learning model for predicting clinical service line financial performance, including revenue, cost, and contribution margin. The model is designed to integrate clinical, operational, resource utilization, payer, and volume-based predictors into a unified forecasting framework. The proposed approach uses a temporal deep learning architecture that fuses static service line characteristics with dynamic monthly features. It would generate forward-looking financial forecasts with uncertainty-aware outputs and interpretable drivers for service line executives. Conceptually, the model would identify a pending margin shortfall driven by a combination of rising patient acuity, unfavorable payer contract exposure, higher resource consumption, longer length of stay, and declining procedure volume. Such forecasts would support earlier operational review and targeted cost-management actions before formal budget variance escalation. A deep learning model for service line financial forecasting could enable continuous surveillance of revenue, expense, and margin risk. By linking clinical activity, payer dynamics, and operational throughput, the approach could support proactive decision-making by service line directors and health system finance leaders.