Improved performance AI-assisted cancer detection using piezoelectric sensor signals: A synergistic optimization approach using GRU-LSTM

Cancer remains a leading cause of death worldwide, creating a strong demand for accurate, rapid, and noninvasive diagnostic technologies for early-stage detection. Conventional diagnostic methods, including biopsies, MRI, CT scans, mammography, and histopathological analysis, are often costly, time-consuming, and dependent on expert interpretation. Recent advances in biosensing and Artificial Intelligence (AI) have enabled intelligent cancer diagnosis using bioelectrical and biomechanical signals. Piezoelectric sensors have emerged as highly sensitive tools capable of detecting subtle physiological variations, cellular interactions, and biomechanical abnormalities associated with cancer progression. However, analyzing piezoelectric sensor signals remains challenging due to noise, nonlinear biological behavior, signal complexity, and temporal dependencies. To address these limitations, this study proposes an AI-driven cancer detection framework integrating piezoelectric sensor signals with a synergistic optimization strategy based on Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM) networks. The hybrid GRU-LSTM model combines GRU computational efficiency with LSTM long-term memory capabilities to improve classification accuracy, convergence stability, and early-stage cancer detection performance. The framework also incorporates advanced preprocessing, feature selection, and explainable AI techniques to enhance reliability and interpretability. Experimental results demonstrate superior diagnostic accuracy, sensitivity, specificity, and reduced error rates compared with conventional machine learning and standalone deep learning methods.