Pre Loader

Developing a Prototype-Oral Cancer Identification tool using AI Technology

The Cochin Cancer Research Centre (CCRC), an apex cancer care and research centre established in 2016 by the Government of Kerala started its functioning in the temporary building of Kalamassery Medical College for the diagnosis, treatment and awareness of cancer patients in Central Kerala. In 2023-24, the centre is moving to its own newly constructed building. The Cochin Cancer Research Centre, which is growing at the national level, is spearheading several research projects for technology-based diagnosis.

The incidence of cancer in Kerala is much higher than the national average. In Kerala, Lung cancer is most common among men. It is estimated to occur in thirty-three people among one lakh every year. Lung cancer is mainly caused by the use of tobacco products. Oral cancer is the second most among men and is more common in economically disadvantaged people of the lower strata of society. It can be understood that changes can be detected in earlier stages prior to the stage of cancer.

If the disease is detected in earlier stages, it is cent percent treatable and curable. Various early stage tests using the application are successful and help in detecting oral cancer accurately. Oral cancer and many pre-cancerous changes can be detected accurately and quickly with this application.

The ability to recognize and identify oral cavity lesions is one of the major challenges to health care professionals including doctors. Although most of the oral lesions may be benign, it is not uncommon for more serious lesions to present themselves. Awareness of the pathologic entities and consistent clinical examinations are critical for effective management as well as improved outcomes for all patients. The common oral conditions encountered regularly in our patients are leukoplakia, erythroplakia and invasive cancers. Each patient case should be assessed for a differential diagnosis that is the most fitting for his or her unique presentation.

Early detection and treatment of oral cancer are essential for giving better health conditions and reducing the negative impacts of the disease. A mobile application for early detection of oral cancer using artificial intelligence has been designed by Cochin Cancer Research Center with the help of Computer Applications Department of Cochin University of Science and Technology. In this work, we are trying to make an efficient oral cancer early detection method using deep learning and artificial intelligence. This study highlights the importance of incorporating advances in Al, such as deep learning into clinical practice to
improve patient outcomes. The development of a diagnostic tool will assist doctors to improve the lives of millions of people affected by this disease. Deep learning has shown significant improvement in image classification in recent years. Using state-of-the-art models achieve near-human-level performance on many practical applications in fields such as healthcare, autonomous driving, and security. This work provides a set of deep learning pretrained models VGG16, VGG19, VIT_b16, ResNet50, and EfficientNet for feature extraction. The extracted features are fused and passed through RandomForestClassifier which selects the important features. These features go through a five-layered Deep neural network which classifies the image into oral cancer or normal.

Various early stage tests using the application are successful and help in detecting oral cancer accurately. Oral cancer and many pre-cancerous changes can be detected accurately and quickly with this application. Adding more photos and diagnostic reports in the future will improve the accuracy of this mobile app. Oral cancer diagnosis using artificial intelligence is very limited in India itself. It is used by millions of people in India. Apart from this, the mobile application is very useful for diagnosis of people in interior and hilly areas.