Philips puts AI auto-measurement and on-cart Koios on its general-imaging ultrasound

A 510(k) clearance brings automated abdominal measurements and breast/thyroid decision support to EPIQ Elite and Affiniti carts — but the headline accuracy figure rests on a vendor's retrospective n=150 analysis, not a peer-reviewed trial.

Philips said it has received FDA 510(k) clearance for Elevate Plus, a software upgrade that brings AI auto-measurement and on-cart decision support to its EPIQ Elite and Affiniti general-imaging ultrasound systems. The package also carries a CE Mark.

The headline addition is Auto Measure Abdomen, which automates routine abdominal measurement steps. Philips reports it delivers “over 93% accuracy compared to manual measurements by clinical experts” and frames the tool as a way to reduce operator variability. The second pillar is Koios AI Decision Support, which Philips has now moved on-cart after previously offering it only off-cart. Koios classifies breast lesions (BI-RADS) and thyroid nodules (TI-RADS); the company says the thyroid model leverages “more than 350,000 pathology-proven cases” and that its breast read returns in “under 2 seconds.” Two imaging enhancements, XRes Pro+ and Super Res MVI Pro, round out the release.

Read the numbers carefully

The supporting figures here come from the vendor, not from a published trial. Philips does disclose the basis for the 93% figure in a footnote: it was “obtained from a retrospective data analysis study involving data from 150 subjects,” with annotations made using the MD.AI tool and three clinical experts as the comparison standard. That is a stated study type and sample size — but it is a vendor-internal retrospective analysis (n=150), not peer-reviewed, and no confidence interval is reported. A separate “up to 30%” reduction in scanning time is attributed to a single customer — an ultrasound manager at Boston Medical Center — and is not drawn from any study.

A 510(k) clears a device as substantially equivalent to a predicate. It is not a finding of clinical superiority, and none of the supporting figures here are peer-reviewed.

For clinicians, the practical shift is consolidation: measurements and lesion classification that previously required separate steps or off-cart tools now sit inside the scanning workflow. Whether the reported accuracy and time savings hold up in routine practice remains to be demonstrated outside Philips’s own materials.

Trials Today

Phase 3 TROPiCS-04 (sacituzumab govitecan, urothelial cancer) — Negative confirmatory readout: primary OS endpoint missed, HR 0.86 (95% CI 0.73-1.02), p=0.0870; results posted to registry.

Phase 3 TRIUMPH-1 (retatrutide, obesity) — Status changed to COMPLETED (last update 2026-06-03); enrollment 2,335; no efficacy results posted yet (hasResults false).

Phase 3 LoTam (low-dose tamoxifen, early breast cancer) — SUSPENDED for protocol amendment to increase sample size (NCI/Alliance); reason stated as operational, not safety.

Phase 3 RASolute 305 (zoldonrasib, KRAS G12D pancreatic cancer) — Newly registered and recruiting first-line trial; co-primary PFS and OS; planned enrollment 670.

Phase 3 Sibeprenlimab (IgA nephropathy) — Status COMPLETED (last update 2026-06-03; primary completion 2026-05-13); full efficacy results not yet posted.

At the Agencies

Zaynich (cefepime/zidebactam) approved for complicated UTI — Novel beta-lactam/beta-lactamase-inhibitor combo, NDA 220787; first new chemical entity fully developed by an Indian company to win FDA approval.

Cingulate CTx-1301 (ADHD) receives Complete Response Letter — Per company, CRL cited only CMC/manufacturing deficiencies with no current safety or efficacy concerns; resubmission planned.

FDA draft guidance on platform knowledge for genome-editing therapies — Non-binding draft (June 2, 2026) on reusing platform CMC, nonclinical and clinical data; comment period before finalization.