AI has been at the forefront of early cancer detection, with promising accuracy and reliable predictions. Early intervention can help patients begin treatment sooner significantly improving their chances of recovery and increasing life expectancy. Recent advancements in vision transformers and multimodal models enable the integration of data from various sensors, providing comprehensive insights that may surpass human interpretation and highlight subtle patterns that even experienced doctors could overlook. Furthermore, AI-driven models can enhance privacy and security in patient diagnosis, particularly in sensitive cases such as breast and prostate cancer, by processing data locally and minimizing exposure of personal information. AI can also aid in drug discovery that can help manage and cure more acute and novel cancer cases.
This review talks about how recent advances in AI have helped in early and rapid detection of disease prognosis prediction, and overall prevention and cure to improve life expectancy against cancer. Machine learning methods such as the “machine-determined DNA methylation” technique can classify many different forms of cancer, including prostate and colon adenocarcinoma, breast invasive carcinoma, kidney renal clearcell carcinoma (KIRC), and lung adenocarcinoma (LUAD). Vision Transformers such as ViT-Patch have proven to be highly useful in malignant detection and tumor localization. Moreover publicly available datasets can help models train on large amounts of data from actual patients and improve their performance.
https://www.tandfonline.com/doi/epdf/10.2147/JMDH.S410301?needAccess=true
Drug discovery is becoming increasingly important for treatment of novel and more aggressive forms of cancer. AI through identification of novel anticancer targets and biomarkers, can also help develop evaluate their druggability, using the ADMET properties of these drugs. This article talks about how modelling of cellular networks underlying cancer provides a quantitative framework to investigate the link between network properties and the disease by artificial intelligence biology analysis from multiomics data, thereby leading to the discovery of potential novel anticancer targets and drugs.
https://www.nature.com/articles/s41392-022-00994-0
The integration of AI with Precision oncology also has benefits in ethical considerations during cancer prognosis. An approach of classifying breast cancer through AI has been presented in this article. Through federated learning across multiple institutions having control over the fidelity of their patients data, a transfer learned model can fine tune its predictions locally to a particular institution, thereby protecting the privacy of the patients and institutions as a whole, while not compromising their performance in breast cancer detection.
https://pmc.ncbi.nlm.nih.gov/articles/PMC11300768/
https://www.sciencedirect.com/science/article/pii/S0957417424009795
https://www.nature.com/articles/s42256-023-00633-5
Public health issues significantly impact cancer outcomes, contributing to differences in cancer incidence, diagnosis, treatment, and survival rates across various populations. These Public health issues are often linked to social, economic factors, including Geographic Location, Health Literacy, Insurance Coverage, and Environmental and Occupational Exposure. Addressing these Public health issues requires targeted policies to improve access to healthcare, increase community outreach and education, and ensure impactful participation in clinical trials and innovative treatments.
This review highlights how racial minority populations face an increased burden relative to cancer interventions. Compared with Caucasians, the cancer screening rate is substantially lower among African American, Asian American, Latinx American, and American Indian/Alaska Native populations. Barriers such as low health literacy, lack of health insurance, and miscommunication between patients and providers have been identified as important factors that result in low screening rates among minority adults.
https://link.springer.com/article/10.1007/s40615-020-00763-1
Financial burden is also a very common reason for the disparity of cancer diagnosis and treatment outcomes among rural and urban populations.This is likely because of the impact of rurality on access to state-of-the-art cancer prevention, diagnosis, and treatment services, as well as higher rates of risk factors such as smoking and obesity and has been discussed in this article.
https://academic.oup.com/jnci/article/114/7/940/6527096?login=false
A lack of knowledge about gender minorities’ health needs among health care practitioners was evidenced in this review, and it represented a major hurdle to cancer prevention, care, and survivorship for transgender and gender-diverse individuals. Discrimination, discomfort caused by gender-labeled oncological services, stigma, and lack of cultural sensitivity of health care practitioners were other barriers met by transgender and gender-diverse persons in the oncology setting.
https://jamanetwork.com/journals/jamaoncology/article-abstract/2801294
AI has the potential to address public health issues in cancer care by enhancing early detection, diagnosis, treatment planning, and patient outcomes across diverse populations. By improving diagnostic accuracy, reducing bias, and increasing access to care, AI can help overcome public health challenges in cancer care—provided that models are designed and implemented with fairness, inclusiveness, and accessibility in mind.
The authors develop methods to address Public health issues in AI-based cancer prognosis such as through inclusion and exclusion critieria, literature screening, data abstraction, exploring hypothetical scenarios and using counterfactual reasoning to better understand the implications of novel AI-based decision support for cancer control in a local clinical context.
https://link.springer.com/article/10.1007/s11912-023-01376-7
In this article, the authors present an Intelligent Catchment Analysis Tool (iCAT), that employs a robust Geographic Information System (GIS) to map healthcare outcomes and disease Public health issues. It uses statistical AI and multivariate analysis. Linear regression models are used to study the linear relationship of various social and environmental variables on continuous health disparity outcomes (i.e. cancer mortality rate) while logistic regression provides the user with the ability to study important factors affecting the presence or absence of health disparity outcomes
https://www.nature.com/articles/s41598-024-57604-y
This study conducted using AI and social media revealed 3 barriers in transgender cancer care: lack of awareness, access issues, and clinical challenges. Lack of awareness issues, pronoun misuse, conversion therapy, and health care avoidance, were common during screening and diagnosis stages. Access issues included nonfinancial barriers, limited insurance coverage and cost barriers. Clinical challenges involved physical and mental health problems, highlighting that the disease burden among transgender individuals was more evident during the management stage.
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2822775