AI in the operating room: sharp increase in complaints about the TruDi navigation system after the integration of machine learning

Since 2021, when artificial intelligence algorithms were added to the TruDi navigation system for endoscopic operations on the paranasal sinuses, the FDA database has received a solid flow of reports of failures and unwanted consequences. The system, created by Acclarent, was considered an assistant to ENT surgeons.

Before the AI ​​update, the agency recorded only seven unconfirmed cases of malfunctions and one incident with patient injury, xrust recalls. After the introduction of machine learning (TruSeg and TruPath), the number of such reports exceeded 100 (as of the end of 2025 — beginning of 2026). At least 10 patients were affected between the end of 2021 and November 2025, mainly due to alleged navigation errors: the system allegedly showed the wrong position of instruments during operations inside the head.

Described incidents and possible consequences

Reports mention serious complications: leakage of cerebrospinal fluid, perforation of the base of the skull, damage to large vessels (including the carotid artery) with subsequent strokes. Two stroke cases have led to lawsuits in Texas courts, with patients alleging that TruDi's AI component misled surgeons and contributed to their injuries. One of the statements emphasizes that the device could have been safer before the integration of artificial intelligence.

Specific examples from claims: in 2022, during sinuplasty (balloon dilatation of the sinuses), the patient’s carotid artery was damaged, which led to a blood clot and stroke; in 2023, a similar incident with another patient resulted in massive bleeding and stroke. Reuters could not independently confirm all the details of the court documents.

Manufacturer's position and official statements

Integra LifeSciences (owner of Acclarent since 2024) denies a direct link between AI technology and injuries: reports to the FDA only document the use of the system during operations with adverse outcomes, but “there is no reliable evidence of a causal relationship.” The company emphasizes that it continues to monitor signals and investigate cases.

The FDA reminds: messages in the MAUDE database are often incomplete, may be duplicated, and do not prove the fault of the device. Reports are not intended to identify the cause of errors — they serve to signal potential problems.

The boom of AI devices in medicine and challenges for the regulator

class=»notranslate»>__GTAG7__ The situation with TruDi reflects the general trend: the FDA has already approved more than 1,350 medical devices with AI elements (as of the end of 2025 — over 1,356, with an increase until 2026). The majority are in radiology (about 77%), but applications are growing in surgery, cardiology and other areas.

A study from Johns Hopkins, Georgetown, and Yale Universities (JAMA Health Forum, August 2025) found that 60 AI devices were associated with 182 recalls, 43% of which occurred less than a year after approval—nearly twice the rate of conventional devices.

FDA experts note overload: the flow of applications for AI devices is growing faster than the agency's human resources. The US Department of Health plans to strengthen expertise in this area.

Examples of problems with other AI devices

TruDi is not the only one that raises questions. The FDA reports mentioned:

• Medtronic cardiac monitors (LINQ series with AccuRhythm AI) — missing abnormal rhythms (explained by user errors);
• prenatal ultrasound Sonio Detect (Samsung Medison) — incorrect labeling of fetal body parts (without indicating harm to patients).

Manufacturers in most cases deny systemic security risks.

Risks and prospects of AI in healthcare

Artificial intelligence promises a revolution: speeding up diagnostics, helping to find drugs, increasing the accuracy of operations. However, cases like TruDi highlight the need for strict control of algorithms, especially in high-risk interventions (sinus surgery, skull base). Regulators and developers will have to balance innovation with patient safety.

Xrust AI in the operating room: a sharp increase in complaints about the TruDi navigation system after the integration of machine learning