Dr Shreve on the Potential Impact and Utility of AI in Cancer Care

Jacob Shreve, MD, MS, hematology/oncology fellow, Mayo Clinic, discusses the potential impact of implementing artificial intelligence (AI) into cancer care, as well as highlights the need for improved external validation of AI models in ongoing research.

In an upcoming presentation at the City of Hope Annual Advances and Innovations in Endoscopic Oncology and Multidisciplinary Gastrointestinal (GI) Cancer Care Meeting, Shreve will highlight the potential impact of AI models in medicine and clinical practice. The aim of this presentation is to educate and promote AI in a way that acknowledges its promise while maintaining a grounded perspective and avoiding sensationalism, Shreve explains.

Although AI has shown immense promise across various industries, including medicine, its integration into clinical practice has been hindered by a lack of reproducibility in research studies, Shreve asserts. However, there is optimism that this is changing, with advancements in AI technology and growing interest in leveraging AI for healthcare applications, he says.

Additionally, Shreve states that the future of AI in medicine is multimodal, where different types of data are combined to create cohesive models that can enhance patient care. Specific examples of this approach include as a disease-agnostic software pipeline that integrates electronic medical record (EMR) data, computer vision interpretations of radiology scans, and genetic traits to develop prognostic models.

Shreve will also present a review article evaluating various AI models currently available for application in gastrointestinal (GI) cancers. This includes models for liver fibrosis diagnosis, detection, and overall GI health. Studies have shown promise in modifying clinical practice, but they often lack external validation and quality assessment, which are essential for clinician acceptance and adoption, Shreve says. Oncologists are known for their cautious approach to adopting new technologies, and they will require rigorous demonstration of validation and reliability before incorporating AI models as a standard in clinical practice, Shreve notes.

With recent advancements in computer science, hardware, data availability, and societal acceptance, the stage is set for AI to become an integral part of clinical decision-making, Shreve emphasizes. Ultimately, AI has the potential to revolutionize patient care and improve healthcare systems globally, and its integration into clinical practice is imminent with the right approach and validation measures, he concludes.