TRV-2026-1235Version 3 · Revised

Saved · see the current record

Why this version exists

Voice pass: plainer wording and study attribution; no facts or figures changed.

The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-1235
version: 3
kind: revised
reason: Voice pass: plainer wording and study attribution; no facts or figures changed.
timestamp: 2026-11-05T11:20:06.380685Z
status: published
lens: trace
sector: health
headline: Artificial intelligence-assisted digital thyroid FNA cytology: Improved agreement and sensitivity for higher-risk Bethesda categories with enhanced screening efficiency
dek: Background Accurate cytologic classification of thyroid nodules is essential for clinical management, but interobserver variability and indeterminate interpretations remain persistent challenges. The clinical feasibility of AIxTHY, a disease-specific deep-learning algorithm integrated into a digital cytology platform, was evaluated for assisting thyroid fine-needle aspiration (FNA) diagnosis using whole-slide imaging (WSI). Methods Two cytopathologists and one cytologist independently reviewed 100 archival ThinP…
gain_title: In a 100-slide test with three reviewers, AI assistance raised sensitivity for higher-risk thyroid cases from 60% to 78-79% and cut slide review time by 40% to 65% compared with microscopy.
problem_title: In a 100-slide test with three reviewers, AI assistance lowered specificity because it produced more false positives, particularly in indeterminate thyroid cases.
trace_subject: AI-assisted digital thyroid FNA cytology for Bethesda risk stratification
gain_reading: In a 100-slide test with three reviewers, AI assistance raised sensitivity for higher-risk thyroid cases from 60% to 78-79% and cut slide review time by 40% to 65% compared with microscopy.
gain_evidence: AI-assisted review modalities improved overall agreement with consensus diagnoses, particularly for higher-risk categories (TBS III+: atypia of undetermined significance, follicular neoplasm, and malignant), and reduced downgrading compared to microscopy. | AI assistance increased sensitivity from 60% to 78% to 79% and accuracy from 61% to approximately 71% | AI-assisted review significantly reduced slide review time by 40% to 65% across diagnostic categories
problem_reading: In a 100-slide test with three reviewers, AI assistance lowered specificity because it produced more false positives, particularly in indeterminate thyroid cases.
problem_evidence: whereas specificity decreased because of increased false-positive classifications, particularly among indeterminate cases. | further refinement is needed to improve specificity in indeterminate lesions.
quick_read: In a 100-slide archival study, two cytopathologists and one cytologist reviewed thyroid fine-needle aspiration cases using microscopy and two AI-assisted whole-slide imaging modes powered by the AIxTHY algorithm. By the October 2026 publication date, the authors reported improved agreement with expert consensus for higher-risk Bethesda categories and a 40% to 65% reduction in review time.

The findings matter because thyroid nodule triage depends on accurate Bethesda classification, where interobserver variability and indeterminate results affect management. While sensitivity and workflow efficiency improved, specificity fell due to false positives in indeterminate lesions, leaving uncertainty about clinical impact, generalizability beyond the small reviewer group, and performance on prospective cases.
limitation: Specificity decreased with AI assistance due to false positives in indeterminate lesions, indicating need for further refinement before routine clinical use.
tag: Dual reading
key_points: Study evaluated AIxTHY, a disease-specific deep-learning algorithm integrated into a digital cytology platform using whole-slide imaging. | Two cytopathologists and one cytologist independently reviewed 100 archival ThinPrep FNA slides across microscopy, AI-assisted single-layer WSI, and AI-assisted seven-layer Z-stack WSI with 2-week washout intervals. | For binary risk stratification TBS III+ vs TBS II, sensitivity rose from 60% to 78% to 79% and accuracy from 61% to approximately 71% with AI assistance. | Review time fell by 40% to 65% across diagnostic categories with AI-assisted modalities.
rundown: Researchers tested AIxTHY on 100 archival ThinPrep thyroid FNA slides, comparing microscopy to AI-assisted single-layer and seven-layer Z-stack whole-slide imaging. Three reviewers assigned Bethesda categories with washout intervals, using expert cytologic consensus as ground truth and recording review times.

AI-assisted modalities showed higher agreement for TBS III+ categories and less downgrading versus microscopy, with sensitivity gains from 60% to 78-79% for TBS III+ versus TBS II. The trade-off was lower specificity from false positives in indeterminate cases, and authors noted feasibility as an adjunct pending specificity improvements.
sources:
- peer_reviewed | Cancer Cytopathology | https://doi.org/10.1002/cncy.70149 | 2026-10-01
prev: 829426ee57eeb93a05fe3619aa0580677ba765fec024e04d8b6ce3be6320c2f0
SHA-256 hash
5e4717378f2d0f58c764b543044b0b6f99be6dd0b1cf7af74a918327aa773811
Previous version’s hash
829426ee57eeb93a05fe3619aa0580677ba765fec024e04d8b6ce3be6320c2f0
Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-1235 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.