Artificial Intelligence-based Cancer Diagnosisand Treatment: A Review

Introduction: Cancer is a fatal disease with a low survivalrate, and its treatment is expensive and time-consuming dueto high recurrence and mortality rates. Early diagnosis and accurateprognosis of cancer are very important to increase thechances of survival. Luckily Artificial intelligence seems to behelpful to aid.Materials and Methods: Due to the large number of patientsand the significant amount of data generated, much attentionis being paid to the use of artificial intelligence in cancer careby utilizing multivariate statistical analysis and artificial intelligence,such as machine learning and deep learning to predictand diagnose cancer and use it to detect, classify, grade, anddescribe tumors molecularly, predict patient outcomes and responsesto treatment, personalized treatment, automated radiotherapyworkflow, and discover new anticancer drugs.CNNs digital imaging Advancements have been utilized in convolutionalneural networks (CNNs) which can be used for:• skin cancer diagnosis and classification of malignant lesions• in polyp detection on colonoscopy images• detection of radiographic anatomical features of malignancies• identification of extranidal extension (ENE) of tumors inlymph nodes of head and neck cancer radio genomics signaturesResults: AI can also be used for predicting genetic traits throughradiographic image analysis, called radio genomics. In a studyfocused on brain MRI of low-grade glioma patients, deep learningtechniques helped neural networks accurately predict IDHmutation and MGMT methylation status very accurately.CNN vs radio genomics signatures Deep learning can predicttreatment responses based on imaging Results:• A CNN model had ۸۰% accuracy in predicting response toneoadjuvant chemotherapy.• radio genomics signatures using CT data features and an MLalgorithm were able to predict CD۸ cell tumor infiltration andresponse to immunotherapy for advanced cancers.This is crucial for patient prognosis and management, makingthe model a useful clinical decision-making tool.Liquid biopsyLung-CLiP is an AI that has been developed to detect circulatingtumor DNA in blood samples for the early detection of lungcancer.