Published Paper Details:

: Oral cancer remains a critical health issue globally, with more than 360,000 new cases diagnosed annually. Despite advances in medical treatments, survival rates are significantly hampered by late-stage diagnoses. Early detection is paramount, as it can greatly enhance patient survival and reduce treatment costs. This study introduces an innovative framework that integrates Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks for the detection of oral cancer using multimodal data fusion. CNNs are employed to extract spatial features from clinical images, while LSTMs analyze the temporal dependencies present in medical imaging data, particularly in the case of longitudinal scans or patient records over time. By fusing spatial and temporal information, the model can detect early-stage lesions, subtle changes, and abnormalities often overlooked by traditional diagnostic methods. To assess the effectiveness of this framework, we conducted extensive experiments on a diverse oral cancer dataset, demonstrating the model's exceptional sensitivity, specificity, and area under the curve (AUC). These results emphasize the robustness and generalization capability of the model across various patient demographics and imaging conditions. The proposed deep learning model offers significant promise for clinical implementation, providing healthcare professionals with a powerful tool for early screening, diagnosis, and improving patient care outcomes. The successful application of CNNs and LSTMs in oral cancer detection underscores the transformative role of deep ABSTRACT: Oral cancer remains a critical health issue globally, with more than 360,000 new cases diagnosed annually. Despite advances in medical treatments, survival rates are significantly hampered by late-stage diagnoses. Early detection is paramount, as it can greatly enhance patient survival and reduce treatment costs. This study introduces an innovative framework that integrates Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks for the detection of oral cancer using multimodal data fusion. CNNs are employed to extract spatial features from clinical images, while LSTMs analyze the temporal dependencies present in medical imaging data, particularly in the case of longitudinal scans or patient records over time. By fusing spatial and temporal information, the model can detect early-stage lesions, subtle changes, and abnormalities often overlooked by traditional diagnostic methods. To assess the effectiveness of this framework, we conducted extensive experiments on a diverse oral cancer dataset, demonstrating the model's exceptional sensitivity, specificity, and area under the curve (AUC). These results emphasize the robustness and generalization capability of the model across various patient demographics and imaging conditions. The proposed deep learning model offers significant promise for clinical implementation, providing healthcare professionals with a powerful tool for early screening, diagnosis, and improving patient care outcomes. The successful application of CNNs and LSTMs in oral cancer detection underscores the transformative role of deep learning in advancing medical image analysis and healthcare diagnostics. Keywords: Oral Cancer Detection, CNN, LSTM, Multimodal Data Fusion, Deep Learning, Medical Imaging, Early Diagnosis.