Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1693
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-17T10:42:41.090016Z
status: published
lens: p_space
sector: health
headline: Surgical Artificial Intelligence Innovation Trends and Regulatory Insights: A Systematic Analysis of Food and Drug Administration-Approved Devices
dek: Objective To characterise the landscape of surgical artificial intelligence (AI) devices approved by the US Food and Drug Administration (FDA). Background AI has the potential to improve surgical pathways, offering capability across screening, diagnosis, decision support, and intraoperative care. Despite this, real-world adoption remains limited. Regulatory approval serves as a key step in clinical translation but is influenced by challenges in evidence generation, safety, bias, transparency, and implementation.…
gain_title: (none)
problem_title: Few cleared surgical AI devices have strong prospective validation, and many cleared through 510(k) show high predicate-creep risk, which weakens confidence that they are equivalent to earlier devices.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: Few cleared surgical AI devices have strong prospective validation, and many cleared through 510(k) show high predicate-creep risk, which weakens confidence that they are equivalent to earlier devices.
problem_evidence: 37.5% demonstrated a high risk of predicate creep
quick_read: Researchers screened FDA-cleared AI and machine learning medical devices up to July 2025 and found that 314, about a quarter, had surgical applications. Most were imaging-based tools for general, orthopaedic, neurosurgical or cardiac surgery, and nearly all were descriptive or diagnostic rather than predictive or prescriptive.

The findings matter because regulatory clearance is a key step toward clinical use, yet the supporting evidence is mostly retrospective or bench-based and nearly all approvals came through a pathway that relies on equivalence to earlier devices. What remains unclear is how the shift toward predictive and interventional AI will be evaluated and whether current clearance practices will keep pace.
limitation: The analysis covers devices cleared up to July 2025 and describes evidence quality and predicate-creep risk, not demonstrated patient outcomes, so the real-world effect of these gaps is not established.
tag: Evidence-backed problem
key_points: 314 devices, 25.1% of the FDA-cleared AI/ML devices screened, had surgical applications. | Devices mapped mainly to general surgery (24.8%), orthopaedic surgery (18.2%), neurosurgery (17.2%) and cardiac surgery (10.8%). | Only 1.3% of devices were predictive and 7.6% prescriptive; the rest were descriptive (50.6%) or diagnostic (40.4%). | The 510(k) pathway accounted for 95.2% of approvals, and 37.5% of those showed high predicate-creep risk. | Authors call for better evidence transparency and robust assessment of device equivalence as AI moves into predictive and interventional roles.
rundown: The study systematically screened FDA-cleared AI/ML devices for surgical relevance, then extracted each eligible device's clinical application, underlying AI technology, evidence quality and regulatory pathway. Devices were sorted with an adapted descriptive, diagnostic, predictive and prescriptive framework, and analyses were run in R.

Results show the market is dominated by imaging-based diagnostic tools. Clinical validation was mostly retrospective or bench studies, and only 5.1% of devices had higher-quality prospective evidence, which limits what can be said about performance in real surgical pathways.

Regulatory findings centre on the 510(k) premarket notification route, used for 95.2% of approvals. Predicate networks were evaluated, and 37.5% of these approvals showed high predicate-creep risk, meaning substantial differences in intended use or technology compared with predicate devices.
sources:
- peer_reviewed | Annals of Surgery | https://doi.org/10.1097/sla.0000000000007229 | 2026-10-08
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

3bf6b69fc4c35ef934d57863e0282b047a5acc08f1679f26466b4ee882c483ab

previous

0000000000000000000000000000000000000000000000000000000000000000