Blockchain (BC) may transform the future of medical systems through reliable storage of medical records and controlled data access, minimizing the risk of large-scale breaches and boosting trust in information exchange. On the flip side, deep learning (DL) is set to revolutionize healthcare diagnostics by quickly and accurately identifying diseases, predicting outcomes, improving health outcomes, reducing costs, and reducing clinicians' workload. An optimized encryption with a blockchain based-deep learning approach is emphasized in this manuscript for effective data sharing and classifying brain tumor in magnetic resonance (MRI) samples. First, the raw brain MRI images are collected from a freely accessible source. Then, the triple chaotic encryption with a blockchain mechanism is introduced to encrypt and store the patients MRI samples to enable access control for different stakeholders. In the encryption process, an innovative secretary bird optimizer (SBO) technique is employed for the optimal key generation process. Following this, a Kuwahara Anisotropic diffusion filtering (KADF) technique is emphasized to enhance the image quality to remove unwanted noise. Then, the Multiscale channel attention residual network (MCA-ResNet) is proposed to detect and classify various brain tumors such as glioma, meningioma, pituitary, and normal, accurately. The developed scheme is simulated via the Python platform. Here evaluates diverse computational metrics, including accuracy, computation time (CT), cost correlation (CC), false negative rate (FNR), F-score, markedness (MK), Informedness (BM), peak signal-to-noise ratio (PSNR), and precision, are evaluated and compared with traditional frameworks. The overall accuracy of 98.3%, CT of 16 s, CC of 97.5%, FNR of 1.7, BM of 0.96, MK of 0.97, and PSNR of 58.63 dB are obtained by the developed framework for securing medical data with disease classification in healthcare systems.

Large language models (LLMs) and multimodal foundation models are increasingly proposed for diagnostic decision support, symptom triage, report interpretation, and electronic health record (EHR)-grounded workflows. Because medical informatics is concerned with how information and communication technologies behave inside clinical practice, workflow evaluation and deployment readiness are central rather than optional.

Recent electrocardiogram (ECG) foundation models have extended ECG analysis from rhythm and morphology recognition to cardiac structure-function assessment, but the computational and memory requirements limit their use in resource-constrained settings. This study aimed to reduce the size and inference cost of a unified ECG foundation model while retaining its multitask performance and clinically relevant representations. We developed the ECGFM-U-Lite, a compact 1D-MobileViT student, using a three-stage compression procedure. Task outputs and intermediate representations were firstly transfered from teacher through structure-function consistent distillation, redundant channels were than removed by using physiologic attention-guided pruning while retaining anatomically informative pathways, and the structure-function-aware quantization strategy achieves adaptive precise quantification by leveraging the sensitivity discrepancies across different modules finally. The model was evaluated on six locked external cohorts. Its computational profile was examined on a laptop GPU and under Android- and Jetson-Nano-oriented ONNX Runtime simulations. ECGFM-U-Lite achieved an overall AUC of 0.862, compared with 0.918 for the teacher, and retained 93.8% and 93.5% of teacher AUC for rhythm/morphology and structure/function tasks, respectively. ECG-to-Text and Text-to-ECG R@1 values were 0.766 and 0.762 in the teacher-derived semantic consistency probe. The parameter count decreased from 86.4 M to 655.7 K, model size from 345.6 MB to 1.7 MB, and peak RAM from 320 MB to 4.4 MB. The inference process was 3.6-fold faster on the laptop GPU and approximately 3.7-fold faster in the simulated edge-oriented profiles. U-Lite substantially reduces the resource requirements of a unified ECG foundation model while preserving most of its external discrimination and teacher-derived cross-modal representation structure, providing a practical basis for subsequent evaluation on resource-constrained hardware.