ExplainabilityTrustworthinessSafetyeess.IV
Gastro-Intestinal Tract Segmentation Using an Explainable 3D Unet
Kai Li, Jonathan Chan
September 25, 2023
arXiv.org
1 citations
1 influential
Key Takeaways
-
Explainability is essential for clinical adoption: The paper demonstrates that "black-box" AI models lack the interpretability needed to be trusted and practical in healthcare settings, highlighting why explainable AI (XAI) isn't just a nice-to-have feature but a requirement for medical applications.
-
AI can reduce treatment delays for cancer patients: By automating the manual organ outlining process that currently prolongs radiotherapy treatment, this deep learning approach has potential to expedite care for gastrointestinal cancer patients undergoing radiation therapy.
-
Daily anatomical variation requires adaptive technology: The paper notes that tumor and intestine positions "may vary day to day," meaning AI systems must account for this biological variability rather than relying on static anatomical assumptions, making real-time imaging and segmentation critical.
-
Responsible AI in medicine requires transparency over performance alone: The authors chose to incorporate explainability into their 3D segmentation model rather than simply optimizing for accuracy, reflecting a commitment to trustworthiness and clinical viability—a principle that should guide AI development in high-stakes healthcare decisions.
Abstract
In treating gastrointestinal cancer using radiotherapy, the role of the radiation oncologist is to administer high doses of radiation, through x-ray beams, toward the tumor while avoiding the stomach and intestines. With the advent of precise radiation treatment technology such as the MR-Linac, oncologists can visualize the daily positions of the tumors and intestines, which may vary day to day. Before delivering radiation, radio oncologists must manually outline the position of the gastrointestinal organs in order to determine position and direction of the x-ray beam. This is a time consuming and labor intensive process that may substantially prolong a patient's treatment. A deep learning (DL) method can automate and expedite the process. However, many deep neural networks approaches currently in use are black-boxes which lack interpretability which render them untrustworthy and impractical in a healthcare setting. To address this, an emergent field of AI known as Explainable AI (XAI) may be incorporated to improve the transparency and viability of a model. This paper proposes a deep learning pipeline that incorporates XAI to address the challenges of organ segmentation.