What question did this study set out to answer?

To develop and evaluate a deep learning model for cervical precancer and cancer detection using liquid-based cytology whole slide images across screening and triage workflows.

April 13, 2025Nature CommunicationsOpen Access

Deep learning enabled liquid-based cytology model for cervical precancer and cancer detection

View Full Paper

Ask AI

Bookmark

Authors

PXPeng XueHarbin Engineering University LDLe DangPingdingshan University LKLinghua KongChinese Academy of Medical Sciences & Peking Union Medical College

Discussion

Member takes

Implication

Multicenter diagnostic study demonstrates enhanced screening accuracy and faster reading times in women undergoing cervical cytology, highlighting deep learning's utility as a clinical assistance...

Key Points

To develop and evaluate a deep learning model for cervical precancer and cancer detection using liquid-based cytology whole slide images across screening and triage workflows.
Trained a deep learning model on whole cytology slides from 17,397 women and tested it on an independent cohort of 10,826 cases across nine hospitals.
Assessed performance across a three-stage evaluation framework, including a multi-reader multi-case study, community-based organized screening, and hospital-based opportunistic screening with HPV triage.
The model cut slide review time from 218 seconds to 30 seconds (p < 0.0001) and surpassed cytopathologists' baseline sensitivity by 9% in multi-reader testing.
In community-based screening, the model achieved sensitivity comparable to senior cytopathologists (0.878 vs 0.854; p > 0.999) with lower specificity (0.831 vs 0.901; p < 0.0001).
Assisted junior cytopathologists in opportunistic screening achieved higher sensitivity (0.857 vs 0.657; p < 0.0001) and specificity (0.840 vs 0.737; p < 0.0001) compared to unassisted reading, showing superior triage in HPV-positive cases.

Cite This Study

Xue et al. (2025) studied this question.

synapsesocial.com/papers/6a03be713a92041adea9c9de https://doi.org/10.1038/s41467-025-58883-3

View Full Paper

Ask AI

Bookmark

Implication

Multicenter diagnostic study demonstrates enhanced screening accuracy and faster reading times in women undergoing cervical cytology, highlighting deep learning's utility as a clinical assistance...

Deep learning enabled liquid-based cytology model for cervical precancer and cancer detection

Discussion

Member takes

Implication

Key Points

Cite This Study

Deep learning enabled liquid-based cytology model for cervical precancer and cancer detection

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider