Artificial intelligence algorithms are already proving to be a very useful tool for preventing tumors. The applications of AI already in use include early detection, rapid diagnosis, and risk prediction.
Experts have been suggesting for some time that artificial intelligence (AI) could offer benefits to oncology, the field of medicine dedicated to researching and treating different types of cancers. In 2023, doctors at Stanford University wrote that AI could already be used as a tool to reduce diagnostic errors and improve the detection rate of malignant tumors. There are diverse tools available, including generative artificial intelligence algorithms, such as the popular GPT chatbot, which could significantly assist in the treatment of tumors.
A study conducted by Northwestern Medicine (a healthcare system comprising hospitals and research teams based in Chicago, USA) assessed the potential of the Open AI chatbot in providing appropriate answers to patient queries about radiotherapy and chemotherapy. The results were overwhelmingly positive. Consequently, this technology could potentially help to reduce physician workload and burnout. However, these results should be taken with the necessary precautions to avoid biases related to insufficient data quantity and quality.
AI applications have proven useful in different cancer-prevention areas, such as disease detection, diagnosis, and risk assessment. Apart from Chat GPT, AI already has the power to offer potentially revolutionary services. In particular, AI has demonstrated its potential as a tool for the early diagnosis of certain tumors, such as lung cancer, the most prevalent form of cancer, accounting for some 1.8 million deaths worldwide in 2020.
By combining advanced computed tomography (CT) technology with new AI applications, lung neoplasms can be identified early and treatments can be started in time, which, according to statistics, can significantly increase the chances of survival. In the USA, GE HealthCare, an industry leader in diagnostics, and Optellum, an innovative startup focusing on lung care, are already collaborating in this field.
The Optellum Virtual Nodule Clinic AI algorithm is trained on databases of thoracic CT scans and diagnostics from British, U.S., and European healthcare systems to differentiate benign from malignant lung nodules. The system uses neural network analysis to transform a standard CT scan into a lung cancer prediction score on a scale of 1 to 10, according to the risk of malignancy. This AI application enables physicians to make a rapid diagnosis at a presymptomatic stage.
There are many different generative AI algorithms used for this purpose, and researchers have already demonstrated the scientific validity of the results obtained. Sybil, the deep learning model presented in the Journal of Clinical Oncology by scientists and researchers at the Mass General Cancer Center in Boston, in collaboration with MIT (Massachusetts Institute of Technology) and CGMH (Chang Gung Memorial Hospital), can predict the risk of a person developing lung cancer within six years from a CT scan image.
Similarly, Sphinks, an algorithmic model developed by the Sylvester Comprehensive Cancer Center at the University of Miami, uses machine learning to detect tumors and, by analyzing the enormous amount of data generated for each tumor, can create more effective personalized therapies.
Meanwhile, a group of researchers, physicists, physicians, and radiologists at the University of Florence (Italy) has managed to automate the process of evaluating image quality in CT scans, using AI to reduce the amount of radiation exposure for patients.
AI is no longer limited to studies, research, insights, and hopes for the future. It is already being used in many hospitals to support all aspects of oncology diagnosis and treatment. The future is now.
