MétaCan
Menu
Back to cohort
Record W4416747404 · doi:10.1186/s13244-025-02132-0

Early prediction of thyroid capsule invasion in papillary microcarcinoma using ultrasound-based deep learning models: a retrospective multicenter study

2025· article· en· W4416747404 on OpenAlexaff
Sui Lin, Bojian Feng, Zhiyan Jin, Xinying Zhu, Tian Jiang, Yuqi Yan, Yahan Zhou, Chen Chen, Jincao Yao, Min Lai, Lujiao Lv, Yifan Wang, Liping Wang, Cong Li, Lina Feng, Wenwen Yue, Dejiang Yu, Kaiyuan Shi, Yang Zhang, Dong Xu

Bibliographic record

VenueInsights into Imaging · 2025
Typearticle
Languageen
FieldMedicine
TopicThyroid Cancer Diagnosis and Treatment
Canadian institutionsArtificial Intelligence in Medicine (Canada)
FundersNational Natural Science Foundation of China
KeywordsNeuroradiologyMulticenter studyRetrospective cohort studyInterventional radiologyUltrasoundDeep learningThyroid

Abstract

fetched live from OpenAlex

OBJECTIVE: Thyroid capsule invasion (TCI) predicts early progression in papillary thyroid microcarcinoma (PTMC). This study aimed to develop an integrated model that combines handcrafted peri-tumoral radiomics features with deep learning (DL)-derived intra-tumoral features for accurate early prediction of TCI, to support clinical decision-making. MATERIALS AND METHODS: Retrospective data from 964 patients with 964 pathologically confirmed PTMC lesions across three centers were collected. Radiomics features were extracted from multiple peri-tumoral regions, and the optimal peri-tumor region with the best radiomics features was selected using a support vector machine (SVM). The selected radiomics features were then combined with intra-tumoral DL features extracted from the tumors before being fed into four different DL models for training and validation. Performance was validated on the internal (n = 177) and external (n = 84) test sets. Six radiologists (senior/attending/junior) assessed TCI with/without DL assistance. RESULTS: The radiomics features, which achieved the best diagnostic performance with an AUC of 0.795 using SVM, were extracted from the peri-tumor region with 30% expansion from the original tumor. By further combining these radiomics features with intra-tumoral DL features, four different DL models were established to identify TCI in PTMC. Swin-Transformer achieved superior performance (internal AUC: 0.923; external AUC: 0.892). With DL model assistance, the AUCs of six radiologists significantly improved, for example, from 0.720 to 0.796 and from 0.725 to 0.790 for senior radiologists, and similar gains were observed for attending and junior radiologists. CONCLUSIONS: As an effective clinical assistive tool, this integrated model can provide TCI identification with high level of accuracy. With its ability to enhance radiologists' diagnostic performance, it supports early PTMC risk stratification and personalized intervention. CRITICAL RELEVANCE STATEMENT: This retrospective multicenter study establishes an integrated model for identifying TCI in PTMC. The model significantly enhances radiologists' diagnostic precision across multiple experience levels, supporting early clinical decision-making for optimized intervention strategies. KEY POINTS: Accurate prediction of TCI facilitates early assessment of PTMC progression and guides subsequent individualized clinical management. DL significantly improves the predictive performance for TCI. DL effectively assists radiologists in TCI diagnosis.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Observational · Consensus signal: Observational
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.073
Threshold uncertainty score0.835

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0010.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.019
GPT teacher head0.261
Teacher spread0.243 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designObservational
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same venueInsights into ImagingSame topicThyroid Cancer Diagnosis and TreatmentFrench-language works237,207