Photo credit: www.sciencedaily.com

New Guidelines for AI in Neuro-Oncology Released

An international coalition of distinguished neuro-oncology experts has introduced fresh guidelines aimed at enhancing clinical practice, specifically concerning the deployment of artificial intelligence (AI) in the diagnosis, monitoring, and treatment of brain cancer patients. These guidelines are designed to bolster the reliability of clinical trial outcomes while safeguarding patient welfare.

The recommendations were recently published in two related policy reviews in The Lancet Oncology. This effort is part of the Response Assessment in Neuro-Oncology (RANO) cooperative group, an esteemed collaboration of global authorities that aims to create consistent standards for evaluating treatment responses in brain cancer clinical trials.

Spyridon Bakas from Indiana University School of Medicine leads the second policy review, which outlines the necessary protocols for the standardization and validation of AI applications in neuro-oncology. Bakas emphasizes that the new guidelines represent a crucial upgrade to the existing standard of care, which currently relies on individual radiologists to assess tumor dimensions, a method that is prone to subjectivity. This variability can lead to inconsistent treatment plans based on differing interpretations of imaging scans.

“AI presents the opportunity to analyze tumor images with greater objectivity,” stated Bakas, who is also the Joshua Edwards Associate Professor in Pathology and Laboratory Medicine and directs the Division of Computational Pathology at the IU School of Medicine. “These AI systems can rapidly assess the type and subtype of the tumor as well as its grade, in addition to monitoring the lesion’s evolution throughout the treatment process.”

The research team noted the emergence of predictive, prognostic, and diagnostic AI models, which are becoming accessible to healthcare professionals. However, the implementation of these tools often diverges significantly between institutions.

“With advancements in technology, AI can effectively assist in determining whether a tumor is progressing or remaining stable,” explained Raymond Y. Huang, associate professor at Harvard Medical School and chief of neuroradiology at Brigham and Women’s Hospital in Boston. “Nevertheless, it is imperative to establish standardized methods for utilizing AI in the accurate diagnosis and treatment of patients.”

The authors conducted a thorough review of current literature and studies on AI advancements in neuro-oncology to form these guidelines, which were shared during prominent medical conferences, including the American Society of Clinical Oncology meeting in Chicago and the European Association for Neuro-Oncology’s annual gathering in Glasgow. Further discussions are slated for the Society for Neuro-Oncology’s meeting in Houston this coming November.

Among the key recommendations put forth by the authors are:

• Utilizing software developed from extensive and diverse patient datasets.
• Ensuring that AI models align with World Health Organization standards for tumor classification.
• Scrutinizing the processes through which tumor images are obtained, processed, and analyzed before evaluation.

“These guidelines are essential for guaranteeing that AI tools developed in the U.K. and elsewhere adhere to strict standards, ultimately leading to improved patient outcomes,” remarked Thomas Booth from King’s College London, a co-author of the guidelines. “With these recommendations in place, we can aim for more consistent and accurate AI applications that serve both clinicians and patients globally.”

As AI technology remains in its early stages, these recommendations mark some of the first formal guidelines for its application in cancer treatment, highlighting the necessity for ongoing research.

“Continued exploration of these AI models on large, diverse patient populations is crucial for deepening our understanding of disease and refining their use,” Bakas added.

The research included contributions from numerous experts, including Philipp Vollmuth and Norbert Galldiks of German institutions, among others, collaborating to advance neuro-oncological practices internationally.

This significant research initiative has received backing from various esteemed organizations, including the National Institutes of Health, the U.S. Department of Veterans Affairs, and the Dana Foundation’s David Mahoney Neuroimaging Program, reflecting a broad commitment to enhancing cancer care through modern technological approaches.

Source
www.sciencedaily.com