Artificial Intelligence in cancer survivorship care plans: what lies beyond diagnostics?

Abstract

With the worldwide rising standards of health and survival, survivorship care experiences appear to have improved significantly. However multiple barriers are causing inequities in its formulation and delivery. Artificial intelligence based digital systems can analyze and optimize multi-disciplinary teams, treatment options, communication, prognostication, and patient outcomes, besides offering unprecedented speed, accuracy, and precision. While most of the research projects are focused on AI-based detection systems, it is important to understand the scope of AI in the life of these survivors post-diagnosis. This paper provides an understanding of the existing literature and the research efforts being made toward integrating AI in cancer survivorship care after diagnosis. Method: A literature search on the PubMed database in July 2022 using keywords (("Artificial Intelligence"[All Fields] OR "Artificial Intelligence"[MeSH Terms]) AND ("Cancer survivorship care"[All Fields] OR "cancer survivor"[MeSH Terms])) revealed 33 articles published in English. Results: Through our review we could identify three themes: ‘What are the existing loopholes in cancer survivorship care’, ‘How is AI addressing them presently’ and ‘How can AI and ML address these barriers in cancer survivorship care in the future’. We also found many loopholes in AI-assisted digital systems such as transferability, explicability, reliability, validity, data confidentiality, ownership and responsibility, noise and overdiagnosis, resources for training etc. Conclusion: In accordance with Darwin’s "survival of the fittest", AI-empowered clinicians, not AI will replace the traditional ones who refuse to be empowered. Recognition of the research results and the imminent need to advance will create a more equitable survivorship care delivery in the future.