Oral Cancer Prediction Using a Probability Neural Network (PNN).

DOI:10.31557/APJCP.2023.24.9.2991 Oral Cancer Prediction RESEARCH ARTICLE Editorial Process: Submission:11/28/2022 Acceptance:09/10/2023 Oral Cancer Prediction Using a Probability Neural Network (PNN) Mahendrakan Kantharimuthu1*, Malathi M2, Sinthia P3, Aloy Anuja Mary G4, Kins Burk Sunil N5, Jalal Deen K6 Abstract Objective: In India, usually, oral cancer is mostly identified at a progressive stage of malignancy. Hence, we are motivated to identify oral cancer in its early stages, which helps to increase the lifetime of the patient, but this early detection is also more challenging. Methods: The proposed research work uses a probabilistic neural network (PNN) for the prediction of oral malignancy. The recommended work uses PNN along with the discrete wavelet transform to predict the cancer cells accurately. The classification accuracy of the PNN model is 80%, and hence this technique is best for the prediction of oral cancer. Results: Due to heterogeneity in the appearance of oral lesions, it is difficult to identify the cancer region. This research work explores the different computer vision techniques that help in the prediction of oral cancer. Conclusion: Oral screening is important in making a decision about oral lesions and also in avoiding delayed referrals, which reduces mortality rates. Keywords: Oral cancer- Discrete wavelet Transform (DWT)- Probabilistic Neural Network (PNN)- malignancy Asian Pac J Cancer Prev, 24 (9), 2991-2995 Introduction In recent days, across the world, oral cancer is one of the most affected kinds of head and neck cancer, which leads to 177,757 deaths each year. The survival rate has increased from 75 to 90 percent when it is identified by early diagnosis, so that the oral cancers must be detected at an early stage. This is due to the delays in recommendations to oral cancer experts and the lack of knowledge on oral cancer signs. The uncontrollable development of cells that attack and affect the surrounding tissue The tongue, the tissue lining the mouth and gums, beneath the tongue, at the base of the tongue, and the area of the throat at the posterior of the mouth are all places where oral cancer originates. After the age of 40, men are more likely than women to develop oral cancer. Oral cancer is affected by tobacco use, alcohol usage (or both), or infection with the human papillomavirus (HPV). It is called a squamous cell carcinoma (OSCC), since in the dental region 90% of cancers start in squamous cells. Most of the oral cancer cases are detected at a progressive stage, which leads to high mortality. To increase the life span of the patient, the researcher needs to implement early diagnosis of oral cancer. Due to the heterogeneity in the appearance of oral lesions, it is difficult to identify the cancer region. This research work explores different computer vision techniques that help in the prediction of oral cancer. Oral screening plays an important role in making decisions on oral lesions and also helps to avoid delayed referrals and reduce the mortality rate. Related work The following survey discusses the various techniques used for early oral cancer detection. 1- The survival rate of the patient can be improved with the help of oral cancer screening, which leads to an early diagnosis. But, still, a biopsy is an invasive and painful method. The proposed work uses fluorescence visualization using an optical instrument. It is one of the non-invasive techniques that provide results in real time, and investigations can be repeated. Fluorescence imaging technology was evaluated using subjective and objective evaluations. 2- Recent analysis of artificial technology shows that it has been mostly used in our daily lives and has been used in healthcare, medical science, agriculture, etc. As a part of AI, many machine learning techniques were used in medical findings and treatments using emerging medical imaging methods. The research work will perform an extensive survey of machine learning in dental, oral, and craniofacial imaging. 3- The research work recommended a fast, cost-effective Department of ECE, Hindusthan Institute of Technology, Coimbatore, India. 2Department of ECE, Rajalakshmi Institute of Technology, Chennai, India. 3Department of ECE, Saveetha Engineering College, Chennai, India. 4VelTech Rangarajan Dr.Sagunthala R&D Institute of Science and Technology, Chennai, India. 5Sethu Institute of Technology, Madurai, India. 6Solamalai College of Engineering, Madurai, India. *For Correspondence: 1 Asian Pacific Journal of Cancer Prevention, Vol 24 2991 Mahendrakan Kantharimuthu et al deep learning method to identify oral cavity squamous cell carcinoma patients using photographic images. The recommended method uses deep learning algorithms, such as cascaded convolutional neural networks, to identify OCSCC from photographic images. 4- The research paper developed a state-of-the art infrastructure approach that helps to categorize oral cancer using hyperspectral imaging. It is a noninvasive method for the classification of cancer. An accuracy of 94.75 percent has been obtained by deep Boltzmann machine (DBM) and SVM classification in hyperspectral images. 5- The recommended research article enhances the accuracy of tumor identification in the oral cavity. The research work uses CT image pre-processing, and the segmentation was done through fuzzy C-means. The accuracy of 90.11% was obtained by the proposed research using an anisotropic filter along with a fuzzy C-means algorithm. 6- Oral squamous cell carcinoma (OSCC) is a type of cancer of the oral epithelium. This OSCC-S type of cancer is diagnosed at a late stage; hence, the research article proposes an effective screening method for early diagnosis of OSCC, which in turn helps to improve the patient’s lifetime. The author recommended using deep learning technology on CLE images to perform automatic diagnosis of OSC. The recommended method provides an advantage of 88.3% over the textural feature-based machine learning technologies. 7- The survival rate of the patient is increased to 75–90% when the oral cancer is identified at an early stage. The research work uses a second-stage classifier network to classify the recognized area into three categories (benign, OPMD, and carcinoma). 8- VELSCOPE is one of the oral screening devices that uses autofluorescence. The device provides inconsistent results when utilized to discriminate between normal, premalignant, and malignant lesions. Hence, the author uses a quadrature discriminant analysis (QDA) classifier and a linear discriminant analysis (LDA) classifier for autofluorescence images. 9- The main objective of this research is to implement automatic identification of OSCC by artificial intelligence techniques. Morphological and textural features are studied from microscopic biopsy images of OSCC. The research work uses five classifiers, namely: support vector machines, logistic regression, linear discriminant, K-nearest (...truncated)