Multimodal attention-driven network with blockchain integrity for reliable kidney stone diagnosis

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Blockchain Technology

Background:

• Renal calculi (kidney stones) obstruct the renal system, potentially leading to chronic kidney disease and renal failure.
• Early detection of renal calculi is crucial to prevent irreversible complications like scarring and renal failure.
• Ultrasound imaging for renal calculi detection faces challenges due to speckle noise, low tissue contrast, and anatomical variability.

Purpose of the Study:

• To develop a novel multimodal attention-driven network integrated with blockchain for enhanced renal calculi detection.
• To address the limitations of traditional ultrasound imaging in identifying renal calculi.
• To ensure data privacy, integrity, and auditability in clinical applications using blockchain.

Main Methods:

• Proposed a multimodal attention-driven network integrating global and local data representations.
• Developed a cross-attention fusion module with an interpretability layer for robust feature extraction.
• Utilized blockchain technology with smart contracts for secure and auditable data management.

Main Results:

• The multimodal attention-driven network achieved a diagnostic accuracy of 98.5% and an AUC of 99.1%.
• The developed model demonstrated superior performance in detecting renal calculi compared to existing methods.
• Blockchain integration ensured the privacy, integrity, and auditability of medical data.

Conclusions:

• The multimodal attention-driven network offers a secure and trustworthy solution for renal calculi detection in healthcare.
• This AI-driven approach significantly improves diagnostic performance, aiding in early intervention.
• Blockchain enhances the security and reliability of AI applications in clinical settings.