AI and the Changing Standard of Care in Radiology: What Attorneys Need to Know
Introduction
In medical malpractice litigation, the standard of care is never static. As medicine evolves, so too does the legal expectation of what a reasonably prudent physician would do in a given circumstance. One of the most significant shifts now underway is the integration of Artificial Intelligence (AI) into radiology. AI is no longer theoretical, it is being deployed in emergency rooms and imaging centers across the country. For attorneys, understanding how this technology is reshaping radiology practice is becoming critical in evaluating liability and strategy.
AI in Radiology Today
AI tools are already assisting radiologists with:
Breast cancer: Computer-aided diagnosis (CAD) has already been in real life clinical practice since the early 2000’s and is used in supplement to a human breast radiologist to assist the perception of breast calcifications, asymmetry, and nodules that could represent breast cancer.
Pulmonary nodules: algorithms can flag subtle findings on CT scans.
Rib fractures: AI helps detect injuries often overlooked on trauma imaging.
Triage alerts: AI software can autonomously review and triage imaging studies in a busy ER by scanning for potential critical findings, such as stroke, pneumothorax, or bleeding, and flag them in real time. This prioritization helps radiologists elevate the most urgent cases to the top of their worklist, ensuring life-threatening conditions are addressed without delay.
These tools don’t replace radiologists, but they increasingly function as an “extra set of eyes.” Courts may soon expect radiologists to know whether such technology was available and if it was used appropriately.
How Standard of Care Is Defined
Legally, the standard of care represents what a reasonably competent physician would do under similar circumstances. Traditionally, this has been defined by clinical training, practice guidelines, and expert testimony. As AI adoption accelerates and becomes adopted, however, attorneys must ask: does the availability of AI shift what is considered “reasonable”?
AI’s Role in Malpractice Cases
Two key questions are emerging in malpractice litigation:
Should the radiologist have used AI? If an institution had access to an FDA-cleared AI tool, a plaintiff may argue it was negligent not to use it.
How will AI data be scrutinized in discovery? Defense teams may argue that AI was consulted and supported the radiologist’s interpretation, while plaintiffs may claim AI findings were ignored.
In short: AI doesn’t remove radiology from litigation — it adds a new dimension to it.
Practical Takeaway for Attorneys
As you prepare cases, expect opposing experts to raise the issue of AI. Whether in chest CTs for lung cancer, abdominal imaging for incidentalomas, or trauma CTs for missed fractures, AI implemtentation will factor into what is “reasonable practice.” Attorneys who understand this shift will be better equipped to challenge or defend radiology testimony.
Case Example
Consider the 76-year-old lady with a 2 mm incidental pulmonary nodule present on a complicated abdomen and pelvis CT for an acute life threatening bowel perforation. The radiologist does not mention the nodule in in the report. Seventeen months later, the she develops stage 4 lung cancer at the exact site of the previous 2 mm nodule. In litigation, an AI algorithm that would have flagged the nodule could become central evidence. Was the radiologist negligent, or was the technology not yet standard? Was the radiologist practice or hospital negligent for not having adopted a software technology available on the market? The answer will hinge on expert testimony.
Conclusion & Next Steps
The standard of care in medicine is not static over time, and this becomes particularly poignant in the next 10 years as AI diagnostic software implementation goes mainstream. For malpractice cases, that means new arguments, new lines of questioning, and a higher premium on expert testimony that can bridge both medical and technological perspectives.