AI surgical tools might be injuring patients

Most Americans may not expect a robot to perform their surgery, but AI-powered surgical tools are becoming more ubiquitous in operating rooms. While these tools are only used to assist human surgeons during operations and don’t perform surgery themselves, recent investigations, along with several lawsuits, are causing some medical experts to reconsider the use of AI in hospitals.

What kind of surgical tools are powered by AI?

At least 1,357 AI-integrated medical devices are “now authorized by the FDA — double the number it had allowed through 2022,” said Reuters as part of an investigation into AI-assisted surgery. One of the most notable is the TruDi Navigation System, a device manufactured by Johnson & Johnson that uses a “machine-learning algorithm to assist ear, nose and throat specialists in surgeries.” Other AI-assisted devices are designed for surgeries on other parts of the body.

Many of these tools address the “area of vision enhancement,” said Forbes. Traditional laparoscopic surgery “presents surgeons with significant challenges: smoke obscures the surgical field, two-dimensional images make depth perception difficult and critical anatomical structures can be hard to distinguish.” AI surgical tools can eliminate these obstacles and provide surgeons with “crystal-clear views of the operative field.”

What has the result been?

There has been an influx of allegations and lawsuits against various AI tools, many of which claim these tools actively harmed patients. Several of these involve the TruDi tool, as the FDA has “received unconfirmed reports of at least 100 malfunctions and adverse events” related to the device’s AI, said Reuters. Many of the alleged errors occurred when the AI “misinformed surgeons about the location of their instruments while they were using them inside patients’ heads.”

In one case, this reportedly led to cerebrospinal fluid leaking from a patient’s nose, while in another case, a surgeon “mistakenly punctured the base of a patient’s skull,” said Reuters. Two other cases allegedly led to patients suffering strokes after major arteries were accidentally injured; in at least one of these cases, the plaintiff said the TruDi’s AI “misled” the surgeon, causing him to “injure a carotid artery, leading to a blood clot and eventually a stroke,” said Futurism.

FDA reports on malfunctioning devices “aren’t intended to determine causes of medical mishaps, so it’s not clear what role AI may have played in these events,” said Reuters. But TruDi is not the only AI-assisted medical device that allegedly has performance issues. One machine that analyzes prenatal images using AI, the Sonio Detect, has been “accused of using a faulty algorithm” that “misidentifies fetal structures and body parts,” said Tom’s Hardware. And Medtronic, a company that manufactures AI-assisted heart monitors, has faced allegations that its monitors “failed to recognize abnormal rhythms or pauses in patients.”

Overall, at least 60 AI-assisted medical devices have been linked to 182 product recalls by the FDA, according to research published in JAMA Health Forum. At least 43% of these recalls “occurred within the first 12 months” of the device’s FDA approval, said JAMA. This suggests that the FDA’s approval process “may overlook early performance failures of AI technologies.” But there is hope that the issue can be fixed, as shoring up “premarket clinical testing requirements and postmarket surveillance measures may improve identification and reduction of device errors.”