Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-2258
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-26T06:45:18.576328Z
status: published
lens: g_space
sector: health
headline: [THE ROLE OF AI TECHNOLOGY IN MAMMOGRAPHIC SCREENING OF BREAST CANCER WITH EVALUATION OF EFFICIENCY AND POTENTIAL FOR PRACTICAL HEALTH CARE]
dek: In conditions of increasing workload on organized mammographic screening programs and deficiency of qualified personnel, implementation of auxiliary technologies on the basis of artificial intelligence (AI) acquires special actuality. The results of retrospective cohort study included 100 histologically verified cases of breast cancer (BI-RADS category 6) permitted to evaluate diagnostic efficiency of AI algorithm as compared to independent conclusions of three certified radiologists with length of service more…
gain_title: In 100 histologically verified breast cancer cases, the AI algorithm reached 86%-96% sensitivity and flagged suspicious findings that experts missed in 1%-3% of cases without adding groundless examinations.
problem_title: (none)
trace_subject: (none)
gain_reading: In 100 histologically verified breast cancer cases, the AI algorithm reached 86%-96% sensitivity and flagged suspicious findings that experts missed in 1%-3% of cases without adding groundless examinations.
gain_evidence: The sensitivity of technology of AI amounted 86%-96% | In 1%-3% of cases, the AI algorithm identified suspicious findings missed by experts
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers ran a retrospective cohort study of 100 histologically verified breast cancer cases (BI-RADS category 6). They compared an AI algorithm with the independent conclusions of three certified radiologists, each with more than 10 years of experience.

The work was motivated by heavier workloads in organized screening programs and a shortage of qualified staff. The authors conclude the tool could standardize and relieve screening work and lower false-negative risk, though the sample consisted only of confirmed cancers, so its performance on routine screening populations is not shown.
limitation: The study was retrospective and included only 100 histologically verified cancer cases, so it cannot show how the algorithm performs across a full screening population including healthy women.
tag: Evidence-backed gain
key_points: Sensitivity of the AI algorithm ranged from 86% to 96%, depending on the dichotomization model. | Total coincidence with expert appraisal varied between 87% and 91%. | Weighted Cohen's kappa was 0.76 (95% CI 0.68-0.83), described as good concordance. | In 1%-3% of cases the AI found suspicious findings that experts missed. | Disputed cases were conservatively classed as BI-RADS 3 with dynamic observation, avoiding more groundless examinations.
rundown: The study sample was 100 breast cancer cases confirmed by histology and assigned BI-RADS category 6. The AI algorithm was benchmarked against three radiologists who each gave independent conclusions.

Agreement and sensitivity were measured in several ways. Sensitivity depended on how results were dichotomized, and total coincidence with experts was 87-91%, while weighted kappa indicated good concordance.

The authors report the AI classified the most disputable cases as BI-RADS 3, which orders dynamic observation instead of immediate further workup. They say this avoided an increase in groundless examinations and fits within established clinical protocols.
sources:
- peer_reviewed | Problems of Social Hygiene, Public Health and History of Medicine | https://doi.org/10.32687/0869-866x-2026-34-4-851-856 | 2026-12-15
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

4b3efe93d06a240d7b151b32c6ff5e8afd43a42c1819316db4081c9cafb792d4

previous

0000000000000000000000000000000000000000000000000000000000000000