Peer reviewed AI-assisted

Randomised implementation study of CT-based AI risk support for cystic renal masses (KITE-CRM)

Among 5,869 adults randomised at one Shanghai hospital to AI-assisted or standard CT reporting, benign pathology was found in 16.4% of treated lesions with AI support against 40.3% with standard care, after the model reached external AUROCs of 0.913 to 0.999 at five hospitals and raised six radiologists' AUROC from 0.825 to 0.973 in a reader study; in the subgroup of cystic masses the tool is built for, the difference was not significant.

Model
KITE-CRM (Kidney Imaging-based Triage Engine for Cystic Renal Masses), a deep-learning model on corticomedullary-phase CT
Field
Medicine
Date
2026-11-03
Human collaborators
Chengwei Chen, Yun Bian

What was found

Cystic renal masses are often overtreated. The model was developed on 570 training cases from one referral centre (2020-2021), validated internally on 235 and externally on 528 cases from five hospitals (2022-2025), with one index lesion per patient and a fixed decision threshold. In a randomised crossover reader study six radiologists read 117 external cases with and without AI, and assistance cut mean reading time from 9.7 to 5.5 minutes. From November 2025 to February 2026, 5,869 consecutive adults having contrast-enhanced abdominal CT at Changhai Hospital were randomised 1:1 to AI-assisted (2,934) or standard reporting (2,935), with the proportion of pathologically benign lesions among treated lesions as the primary clinical endpoint. In a post hoc intention-to-treat analysis benign intervention occurred in 0.95% against 3.54% of randomised participants. The trial is registered with the Chinese Clinical Trial Registry as ChiCTR2500112290 and the model code is public.

Novelty check

Europe PMC, 2026-11-16, cystic renal mass, renal cyst or Bosniak with artificial intelligence or deep learning and a randomised design: 178 records. The top results were this paper, a systematic review of AI against human interpretation in radiology, and unrelated surgical and imaging papers; no earlier randomised implementation study of AI for cystic renal masses appeared among them. Nothing in this registry covered renal imaging.

Caveats and known objections

The randomised step is single-centre and run by the developers, so nothing here is independent in the medical sense. The primary endpoint is computed among treated lesions, which conditions on a treatment decision made after randomisation; the post hoc intention-to-treat figures point the same way but were not prespecified. In the subgroup the tool is designed for, treated AI-evaluable cystic masses, the difference was 17.2% against 33.9% and not significant (p = 0.105), so much of the headline effect may come from lesions outside its intended use. One external test set had only three malignant cases. The endpoint is overtreatment over a short window, not cancer outcomes. The Chinese Clinical Trial Registry is not one of the registries this site records. Autonomy is ai-assisted: the model supplies a malignancy risk to the reporting radiologist, and clinicians made the treatment decisions.

No independent check is recorded on this entry yet.

Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.

Or on GitHub: submit a check challenge the grade send a correction or send a pull request

Entry history (1 event)
  1. AddedEntered the registry graded Peer reviewed and AI-assisted.

Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.

Graded peer reviewed for verification and ai-assisted for autonomy. What these mean.

Cite this entry

Plain text
whataifound.org. (2026). Randomised implementation study of CT-based AI risk support for cystic renal masses (KITE-CRM). whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-11-03-kite-crm-cystic-renal-masses
BibTeX
@misc{whataifound-changhaihospitalnavalmedicaluniversity-2026-masses,
  title        = {Randomised implementation study of CT-based AI risk support for cystic renal masses (KITE-CRM)},
  author       = {{whataifound.org}},
  year         = {2026},
  howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
  note         = {Result by Changhai Hospital, Naval Medical University. Verification: Peer reviewed. Autonomy: AI-assisted.},
  url          = {https://whataifound.org/finding/2026-11-03-kite-crm-cystic-renal-masses}
}

Related findings

← All medicine findings in the registry