Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study

Abstract Background This study aims to develop and validate a multi-modal machine learning model for preoperative risk stratification in papillary thyroid carcinoma (PTC), addressing limitations of current systems that rely on postoperative pathological features. Methods We analyzed 974 PTC patients...

Full description

Saved in:
Bibliographic Details
Main Authors: Jia-Wei Feng, Lu Zhang, Yu-Xin Yang, Rong-Jie Qin, Shui-Qing Liu, An-Cheng Qin, Yong Jiang
Format: Article
Language:English
Published: BMC 2025-08-01
Series:Cancer Imaging
Subjects:
Online Access:https://doi.org/10.1186/s40644-025-00921-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849761290932715520
author Jia-Wei Feng
Lu Zhang
Yu-Xin Yang
Rong-Jie Qin
Shui-Qing Liu
An-Cheng Qin
Yong Jiang
author_facet Jia-Wei Feng
Lu Zhang
Yu-Xin Yang
Rong-Jie Qin
Shui-Qing Liu
An-Cheng Qin
Yong Jiang
author_sort Jia-Wei Feng
collection DOAJ
description Abstract Background This study aims to develop and validate a multi-modal machine learning model for preoperative risk stratification in papillary thyroid carcinoma (PTC), addressing limitations of current systems that rely on postoperative pathological features. Methods We analyzed 974 PTC patients from three medical centers in China using a multi-modal approach integrating: (1) clinical indicators, (2) immunological indices, (3) ultrasound radiomics features, and (4) CT radiomics features. Our methodology employed gradient boosting machine for feature selection and random forest for classification, with model interpretability provided through SHapley Additive exPlanations (SHAP) analysis. The model was validated on internal (n = 225) and two external cohorts (n = 51, n = 174). Results The final 15-feature model achieved AUCs of 0.91, 0.84, and 0.77 across validation cohorts, improving to 0.96, 0.95, and 0.89 after cohort-specific refitting. SHAP analysis revealed CT texture features, ultrasound morphological features, and immune-inflammatory markers as key predictors, with consistent patterns across validation sites despite center-specific variations. Subgroup analysis showed superior performance in tumors > 1 cm and patients without extrathyroidal extension. Conclusion Our multi-modal machine learning approach provides accurate preoperative risk stratification for PTC with robust cross-center applicability. This computational framework for integrating heterogeneous imaging and clinical data demonstrates the potential of multi-modal joint learning in healthcare imaging to transform clinical decision-making by enabling personalized treatment planning.
format Article
id doaj-art-e3690aa47bad4f278dbe83a91ae18838
institution DOAJ
issn 1470-7330
language English
publishDate 2025-08-01
publisher BMC
record_format Article
series Cancer Imaging
spelling doaj-art-e3690aa47bad4f278dbe83a91ae188382025-08-20T03:06:04ZengBMCCancer Imaging1470-73302025-08-0125111510.1186/s40644-025-00921-wDevelopment and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort studyJia-Wei Feng0Lu Zhang1Yu-Xin Yang2Rong-Jie Qin3Shui-Qing Liu4An-Cheng Qin5Yong Jiang6Department of Thyroid Surgery, The Third Affiliated Hospital of Soochow University, Changzhou First People’s HospitalDepartment of Ultrasound, Ruijin Hospital, Shanghai Jiao Tong University School of MedicineDepartment of Thyroid Surgery, The Third Affiliated Hospital of Soochow University, Changzhou First People’s HospitalThe Second Clinical Medical School of Nanjing Medical UniversityDepartment of Ultrasound, The Third Affiliated Hospital of Soochow University, Changzhou First People’s HospitalDepartment of Thyroid Surgery, Suzhou Municipal Hospital, The Affiliated Suzhou Hospital of Nanjing Medical UniversityDepartment of Thyroid Surgery, The Third Affiliated Hospital of Soochow University, Changzhou First People’s HospitalAbstract Background This study aims to develop and validate a multi-modal machine learning model for preoperative risk stratification in papillary thyroid carcinoma (PTC), addressing limitations of current systems that rely on postoperative pathological features. Methods We analyzed 974 PTC patients from three medical centers in China using a multi-modal approach integrating: (1) clinical indicators, (2) immunological indices, (3) ultrasound radiomics features, and (4) CT radiomics features. Our methodology employed gradient boosting machine for feature selection and random forest for classification, with model interpretability provided through SHapley Additive exPlanations (SHAP) analysis. The model was validated on internal (n = 225) and two external cohorts (n = 51, n = 174). Results The final 15-feature model achieved AUCs of 0.91, 0.84, and 0.77 across validation cohorts, improving to 0.96, 0.95, and 0.89 after cohort-specific refitting. SHAP analysis revealed CT texture features, ultrasound morphological features, and immune-inflammatory markers as key predictors, with consistent patterns across validation sites despite center-specific variations. Subgroup analysis showed superior performance in tumors > 1 cm and patients without extrathyroidal extension. Conclusion Our multi-modal machine learning approach provides accurate preoperative risk stratification for PTC with robust cross-center applicability. This computational framework for integrating heterogeneous imaging and clinical data demonstrates the potential of multi-modal joint learning in healthcare imaging to transform clinical decision-making by enabling personalized treatment planning.https://doi.org/10.1186/s40644-025-00921-wPapillary thyroid carcinomaRisk stratificationMachine learningRadiomicsSHAP analysis
spellingShingle Jia-Wei Feng
Lu Zhang
Yu-Xin Yang
Rong-Jie Qin
Shui-Qing Liu
An-Cheng Qin
Yong Jiang
Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
Cancer Imaging
Papillary thyroid carcinoma
Risk stratification
Machine learning
Radiomics
SHAP analysis
title Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
title_full Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
title_fullStr Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
title_full_unstemmed Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
title_short Development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma: a multicenter, retrospective cohort study
title_sort development and validation of the multidimensional machine learning model for preoperative risk stratification in papillary thyroid carcinoma a multicenter retrospective cohort study
topic Papillary thyroid carcinoma
Risk stratification
Machine learning
Radiomics
SHAP analysis
url https://doi.org/10.1186/s40644-025-00921-w
work_keys_str_mv AT jiaweifeng developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT luzhang developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT yuxinyang developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT rongjieqin developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT shuiqingliu developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT anchengqin developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy
AT yongjiang developmentandvalidationofthemultidimensionalmachinelearningmodelforpreoperativeriskstratificationinpapillarythyroidcarcinomaamulticenterretrospectivecohortstudy