Table 1 Summary of existing studies integrating BC and DL models in healthcare applications.
Authors | Techniques | Metrics | Dataset | Major findings and limitations |
|---|---|---|---|---|
Ragab et al.11 | BC, SCA, DFFNN, AFSA, Signcryption Technique | Accuracy, Precision, Recall, F1-Score | Heart Statlog, Pima Indians Diabetes, and EEG Eye State | BPEHR-SCADL outperforms existing methods in secure and accurate EHR classification |
Ali et al.12 | BC, Hybrid DL, Permissions-based Framework | Scalability, Security, Data Interoperability, Diagnostic Accuracy, Privacy Preservation | Standard Dataset | The framework improves secure, scalable healthcare data management and accurate decision-making |
Alanazi et al.13 | BCODL-SDSC, FOLS, TSO, ARO, SRNN | Accuracy, Sensitivity, Specificity, F-Score, MCC | Benchmark Medical Image Dataset | The model attained secure data sharing and classification with a peak accuracy of 99.11% |
Alamro et al.14 | BHS-ALOHDL, ALO-FSS, CNN-LSTM, FPA | Accuracy, Precision, Recall, F-Score, AUC-Score | Two Benchmark Dataset | The method enhanced intrusion detection accuracy and processing speed in IoT healthcare systems |
Mohananthini, Rajeshkumar, and Ananth15 | BGJOA-DLSMTD, GJOA, Homomorphism Encryption, BC, BOA, DBN, CapsNet-based Feature Extraction | Accuracy, Sensitivity, Specificity, F-Score, MCC | ISIC Dataset | The method attained superior performance in medical image encryption, secure transmission, and disease diagnosis |
Qu et al.16 | Quantum BC and NN, Controlled Quantum Walk Hash Function, Quantum Authentication Protocol, HQCNN | TR/TS Accuracy, Detection Stability | Standard Dataset | HQCNN attains high accuracy and stability, outperforming classical CNN with greater robustness to quantum noise |
Naveen, Dhivya, and Jenefa17 | DenseNet Architecture, BC, TL, Bottleneck and Transition Layers | Accuracy, Precision, Recall, F1-Score | Dermoscopic Image Dataset | DenseNet and BC integration improves melanoma detection accuracy and efficiency |
Kateb et al.18 | ESHCS-DLJSO, Min–Max Normalization, BFOA, CNN-LSTM-Attention, JSO | Accuracy | IoT Healthcare Security Dataset | The model attains a superior accuracy of 99.43%, enhancing disease detection and IoT healthcare security |
Rahal et al.19 | DL, BC, Model Ensembling, Secure Model Sharing | Accuracy | Breast Cancer Dataset | The BC-based ensembling approach improves diagnostic accuracy and protects patient privacy |
Hota et al.20 | CNN | Accuracy, Precision, Recall, F1-Score | Standard Dataset | The system ensures privacy and prevents data breaches while attaining 92% accuracy in diagnosis |
Rokade and Mishra21 | BC, LeNet, TCIO, DeepJoint Segmentation with Kumar-Hassebrooks Distance | Accuracy, True Positive Rate, True Negative Rate, False Negative Rate, False Positive Rate | Benchmark Dataset | The model attained high accuracy and a robust detection rate for classifying skin diseases |
Islam et al.22 | ML, Predictive Analytics, AI, DL | Diagnostic Accuracy, Disease Prediction, Resource Optimization | Commonly Used Public Datasets | ML and AI improve disease detection and healthcare efficiency, but privacy, security, and bias issues remain |
Shammi et al.23 | ML, DL, BC | Disease Diagnosis, Data Security, Patient Privacy | Public Dataset | AI improves diagnosis and monitoring, while BC secures data and privacy |
Matlo et al.24 | PoWV-XAI, BC | Delay Control, Energy Consumption, Cost Efficiency, Security Validation | Various Dataset | PoWV-XAI enhances explainability and optimizes delay, energy, cost, and security |
Khan and Jilani25 | RSA, Blowfish Encryption, BC, ANN, CNN, RNN | Accuracy, Security | Heart Disease Data | The model ensures secure, accurate heart disease prediction (0.9941) while preserving privacy in cloud healthcare systems |
Kulkarni et al.26 | BC-based IoT, Privacy Matrix Generation, AHPO, Deep Maxout Network, SpinalNet | Accuracy, True Positive Rate, True Negative Rate | Healthcare Data | The model ensures private, accurate healthcare classification with high TPR and TNR |
Gupta, Kumar, and Gupta27 | FL, BC, DenseNet-201, FedAvg Aggregation, IPFS for Storage | Accuracy, Precision, Recall, F1-Score | Heterogeneous Datasets | The FL-BC framework detects lung disease with 90% accuracy while preserving data privacy |
Ahmed et al.28 | DL, PSO-BioBERT, Named Entity Recognition, Relation Extraction | NER and RE Accuracy, NER and RE F1-Score, Accuracy Gain over Bi-LSTM and BERT | Medical Dataset | The technique improves NER and RE performance, improving smart healthcare data analysis and decision-making |
Dubey, Kapoor, and Saraswat29 | Hybrid Learning Methods, Biological Feature Selection, BC-based Data Storage, Bidirectional LSTM, Time-series Data Categorization, Alzheimer’s Diagnosis | Multi-disease Prediction, Intrusion Detection, Alzheimer’s Diagnosis Accuracy, Image Segmentation Quality | EEG Data Dataset | The method improves early disease diagnosis and risk reduction using advanced algorithms and image analysis |
Chougule et al.30 | Entropy, Correlation Coefficient | Accuracy | Healthcare Dataset | The model enhances healthcare prediction accuracy and data security using advanced techniques |
Al-Marridi, Mohamed, and Erbad31 | BC, DRL, Multi-Agent Q-Learning, MDP, Computational resource optimization, Cooperative-competitive decision-making, QoS Optimization | Decision-making speed, Resource Utilization, Latency Minimization, Cost Minimization, Security Maximization, Qos Adherence | Simulated Healthcare Data | IP-HealthChain improves decision speed, mitigates latency, optimizes resources, enhances security, and cuts costs |
Pradhan et al.32 | ML, NN, HIoT, EHR, Security Enhancements, QoS | Security Improvement, QoS Enhancement, Accuracy of ML Models | Benchmark Dataset | ML applied to HIoT can improve security and QoS in EHR systems |
