IoT-driven cancer prediction: Leveraging AI for early detection of protein structure variations

Abstract

Cancer and tumour development can be predicted by analyzing altered protein structures via misfolding and accumulations.The sensitivity and specificity required to diagnose early stages of cancer by existing biomarker testing are not always fulfilled. This causes either false positives (resulting in unneeded stress and procedures) or false negatives (not detecting cancer in its early stages).Earlystructural change detection is performedbased on degradation, synthesis, and regular apoptosis. The Internet of Things (IoT) is often applied to providing personal and technology-based healthcare solutions, such as predicting early changes in protein structures. The cellular functions are imported from the clinical data to verify the modular changes and theirThis article introduces a pre-mature structure variation analysis method (PSVAM) to identifymisfolding cellular functions due to tumour/cancer influence. The analysis model is supported by a convolution neural network (CNN) for handling misfolding variations from the daily cellular functionsprogressions. CNN managesthehighest cellular variations in correlation to the provided clinical data. Clinical and observed data are communicated using IoT infrastructures to the medical/ diagnosis centre; the definite clinical data are analyzed. The protein variations are reported amid the completion of the CNN process to identifytheprogression of such changes. Based on the intensity of progress, the tumour/cancer is predicted. This method can make more accurate cancer predictionsbyexploiting the IoT computational intelligence and clinical data compared to existing methods.The experimental results show that the proposed method achieves a high accuracy of 0.9, error of 0.046, detection time of 1.637 m, and variation detection of 87.516 % compared to other methods.

Keywords

• Cancer prediction
• CNN
• IoT
• Protein structure
• Variation detection