The field of artificial intelligence (AI) in cancer has evolved significantly in recent years. Artificial intelligence solutions have been created to address a range of cancer-related issues.
FREMONT, CA: To present, some of the most promising work on artificial intelligence in oncology has been conducted in the field of cancer imaging, notably digital pathology, radiographic imaging, and clinical pictures. In digital pathology, AI has been used to automate time-consuming tasks and improve diagnostic accuracy by performing both low- and high-level image processing and classification tasks (e.g., tumor detection and segmentation, and predicting disease diagnosis and treatment responses based on image patterns). Numerous studies in radiology have established that AI techniques can distinguish between high- and low-risk lesions using several imaging modalities. Integration of radiographic imaging with additional data sources (e.g., clinical characteristics and genetic/biochemical markers) for risk stratifying image-detected lesions already exists and is anticipated to become more prevalent in the future. Additionally, AI is being applied to increase diagnostic accuracy and decrease diagnostic ambiguity in dermatologic and gastrointestinal cancers detected using clinical imaging.
AI can accurately predict a patient's likelihood of various outcomes, including readmission, cancer recurrence, treatment response, treatment toxicity, and mortality. Enhanced forecasts may have several potential benefits, including assisting with treatment planning, guiding population control initiatives, and allowing discussions about care goals. Predictive analytics may aid in administering oncology treatments to underserved or underrepresented populations in clinical trials and for whom evidence-based therapy is difficult to implement. For instance, among older persons, AI predictive analytics may assist oncologists in anticipating problems that are missed during a complete geriatric examination or identify risk factors for chemotherapy toxicity that are disregarded in routine treatment.
Another area in which AI is being used is precision oncology. Due to the rapid expansion of genetic tumor characterization, computational tools have been developed to aid in the interpretation of these data and to facilitate the implementation of precision oncology. For instance, AI can aid in studying tumor genomic data by highlighting potentially actionable mutations discovered during tumor next-generation sequencing experiments. One AI-based method generated data faster and more correctly than humans and aided in determining eligibility for biomarker-selected clinical trials. Additionally, machine learning (ML) can determine the type of tumor from a specific panel of DNA sequence data, assisting in selecting more appropriate therapy. Additionally, it has been demonstrated that machine learning and deep learning can improve liquid biopsies' detection capability and accuracy. As the use of liquid biopsies increases, AI-based solutions may prove invaluable to doctors responsible for correctly ordering and interpreting these complex tests.
There are also significant prospects for AI to impact primary cancer prevention positively. Around half of the malignancies could be prevented by adopting current knowledge about cancer risk reduction. Although behavioral modification is critical for cancer prevention, behavioral modification therapies are neglected. AI can assist policymakers and physicians close this gap by quickly synthesizing, interpreting, and disseminating knowledge for cancer prevention. One group of researchers is developing an ontology for behavioral change evaluations; training an automated feature extraction system to annotate evaluation reports using this ontology; developing machine learning models to predict effect sizes for combinations of behaviors, interventions, populations, and settings; and developing interfaces for querying and updating the knowledge base. In the future, this technology may assist clinicians in providing precision prevention by suggesting interventions that take into account a patient's specific biological, behavioral, and socioeconomic factors.