Artificial Intelligence in Oncology: Current Landscape, Challenges, and Future Directions

Abstract

Artificial intelligence (AI) in oncology is advancing beyond algorithm development to integration into clinical practice. This review describes the current state of the field, with a specific focus on clinical integration. AI applications are structured according to cancer type and clinical domain, focusing on the four most common cancers and tasks of detection, diagnosis, and treatment. These applications encompass various data modalities, including imaging, genomics, and medical records. We conclude with a summary of existing challenges, evolving solutions, and potential future directions for the field.

Significance: AI is increasingly being applied to all aspects of oncology, where several applications are maturing beyond research and development to direct clinical integration. This review summarizes the current state of the field through the lens of clinical translation along the clinical care continuum. Emerging areas are also highlighted, along with common challenges, evolving solutions, and potential future directions for the field.

Figure 1:. Stages of AI development and clinical translation: Example of AI in mammography.

We consider the development of AI for breast cancer detection in mammography as a lens to understand the broader landscape of AI in oncology. A) An artificial neural network is trained on thousands of labeled mammography images. The performance of the network in detecting breast cancer is iteratively improved, and then the network can be used to make predictions on previously unseen images. These predictions can then be used to assist radiologists in their workflow. B) AI in mammography has transitioned from research on retrospective samples to regulatory clearance, clinical integration, and real-world clinical evaluation.

Figure 2:. Overview of AI in Oncology, with specific examples highlighted.

AI is being applied across the patient care trajectory, where this review groups applications into three main categories across this trajectory. Detection applications tend to currently have the highest level of clinical maturity, where several applications have regulatory clearances and published clinical trials, which we denote as “Scaling”. Diagnosis applications tend to be less mature, but regulatory clearances exist and validation studies are underway (“Piloting”). Prognosis and treatment applications are generally furthest from maturity with much emerging research (“Developing”). The review highlights AI applications in each of the clinical categories, with a specific focus on Breast, Prostate, Lung, and Colorectal cancers.

Figure 3:. Challenges and Opportunities.

Like many other domains in medicine, AI in oncology faces significant challenges to effective development and clinical translation. These challenges pertain to curating and sharing data across institutions, guarding against biases from training through to deployment, ensuring appropriate regulation and evaluation, and integrating into clinical workflows across diverse clinical settings.

Figure 4:. Important forward-looking considerations for AI in Oncology.

Effective clinical adoption of AI in oncology will require more than technology advances, and must include critical evaluation of AI impact on patient outcomes and healthcare value. Robust processes for integration across diverse clinical settings are also required, along with ongoing monitoring to ensure patient benefit and safety.