Trustworthy AI, Clinical Validation In Breast Cancer Imaging

On June 2nd, the?Artificial Intelligence Precision Health Institute?at the University of Hawai'i Cancer Center?hosted the 7th seminar in a series of talks on AI in cancer research and clinical practice. In this presentation,?Oliver Diaz, PhD, Associate Professor at Universitat de Barcelona, explored strategies aimed at improving the trustworthiness of AI models in clinical practice such as FUTURE-AI principles. Dr. Diaz' talk was entitled Trustworthy AI and Clinical Validation In Breast Cancer Imaging.

In recent years, AI models have demonstrated remarkable effectiveness in the realms of cancer prevention, detection, and treatment planning. This success can largely be attributed to advancements in deep learning technology. Nevertheless, the integration of AI-based algorithms into clinical practices faces limitations, primarily stemming from the opaque nature of these models.

Best Practices for Trustworthy AI in Medicine

FUTURE-AI is an international, multi-stakeholder initiative for defining and maintaining concrete guidelines that will facilitate the design, development, validation and deployment of trustworthy AI solutions in medicine and healthcare. FUTURE-AI is a live, dynamic framework that is expected to continuously evolve and improve based on new developments in the field of AI in medical imaging.? Future AI is based on 6 guiding principles:

1. Fairness: Medical AI algorithms should maintain the same performance when applied to similarly situated individuals and across subgroups of individuals, including under-represented groups. Healthcare should be provided equally for all patients independently of their gender, ethnicity, income and geography.
2. Universality: Standards should be defined and applied during algorithm development, evaluation and deployment, specifically technical, clinical, ethical and regulatory standards.
3. Traceability: Medical AI algorithms should be developed with mechanisms for documenting and monitoring the whole development lifecycle as well as the functioning of the AI tools in their environment. This will increase transparency by providing detailed information such as the datasets used to train and evaluate the algorithms.
4. Usability: Medical AI solutions should be usable, acceptable, and deployable for end-users in real-world practice, such as physicians, specialists, data managers, and other end-users.
5. Robustness: Medical AI models should maintain performance and accuracy when applied under highly variable conditions in the real world, outside the controlled environment of the laboratory where the algorithm was built.?
6. Explainability: Medical AI algorithms should provide meaningful and actionable explanations to clinicians, including algorithmic mechanisms behind AI decision making processes, allowing for clinical validation and scrutinization of decisions.

Future AI created an assessment checklist composed of concrete and actionable questions to support developers, evaluators and other stakeholders in delivering medical AI tools that are trustworthy and optimized for real-world practice. Each element of the checklist provides examples to illustrate potential mitigation measures to minimize risks.

Speaker Bio

Prof. Diaz is Associate Professor at the Faculty of Mathematics and Informatics of the University of Barcelona and senior researcher at the Barcelona Artificial Intelligence in Medicine (BCN-AIM) laboratory. Dr Díaz has a PhD in Electronic Engineering from the University of Surrey, UK. He has over 15 years of experience developing and validating AI tools with biomedical data, where he has participated in more than 20 research and technology transfer projects in the field of medical imaging and medical physics. He is scientific coordinator of EU-funded project RadioVal.

Artificial Intelligence Precision Health Institute

The?AI Precision Health Institute Affinity Group?was formed to discuss current trends and applications of AI in cancer research and clinical practice. The group brings together AI researchers in a variety of fields including computer science, engineering, nutrition, epidemiology, and radiology with clinicians and advocates. Students, trainees and faculty with any or no background in AI are encouraged to attend. The goal is to foster collaborative interactions to solve problems in cancer that were thought to be unsolvable a decade ago before the broad use of deep learning and AI in medicine. The group is organized by the Artificial Intelligence Precision Health Institute at the University of Hawai'i Cancer Center.

Highlights From Past Affinity Group Webinars

AI For Ultrasound For Real-Time Breast Cancer Decision Support

Using Deep Learning To Diagnose Breast Cancer With High Accuracy

Precision Oncology: Empowering Radiologists and Oncologists With AI

Advanced Machine Learning Methods For Personalized Cancer Screening

Using AI Driven Surgical Robots To Diagnose and Treat Prostate Cancer

Subscribe and Comment

I'm interested in your feedback - please leave your comments. To subscribe please click subscribe at the top of this article.

Copyright ? 2023 Margaretta Colangelo. All Rights Reserved.

This article was written by?Margaretta Colangelo.?Margaretta is a leading AI analyst who tracks significant milestones in AI in healthcare. She consults with AI healthcare companies and writes about some of the companies she consults with. Margaretta serves on the advisory board of the AI Precision Health Institute at the University of Hawai?i?Cancer Center?@realmargaretta