AI-driven diffusion weighted imaging-based non-contrast protocol for breast cancer diagnosis - a multicentre, multidimensional validation study

Q: What problem does the study address in breast MRI diagnosis?

It targets limitations of contrast-enhanced breast MRI, including complexity, long acquisition times, and reliance on gadolinium-based contrast agents. It tests whether a DWI-only approach can replace contrast for accurate diagnosis.

Q: How was the DWI-based deep learning model evaluated?

The study developed a DWI-DL model using one hospital cohort, validated and tested it internally, and then assessed performance on external cohorts plus a prospective cohort retrospectively analyzed. It compared results to an abbreviated enhanced-based model and two expert radiologists.

Q: What were the key outcomes of using DWI-DL and an AI-guided selective protocol?

DWI-DL showed diagnostic performance comparable to the abbreviated enhanced-based model across cohorts and outperformed expert radiologists when interpreting DWI alone. In multireader multicase validation, the AI-guided selective contrast-enhanced workflow was non-inferior to the full protocol while reducing interpretation time by 55.5%.

Sep 29, 2026

Uploaded by Terk

Contrast-enhanced breast MRI is sensitive but constrained by workflow complexity, long acquisition time, and dependence on gadolinium-based contrast agents. This multicentre validation evaluates whether deep learning can enable accurate breast cancer diagnosis using only diffusion-weighted imaging (DWI). A four-centre dataset of 2493 pathologically confirmed patients trains, validates, and tests a DWI-only model and compares it with an abbreviated enhanced model and expert radiologists across external and prospective cohorts.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Summary

Background

Methods

Study design and cohorts

Comparators and validation

Evaluation metrics

Findings

Interpretation

Funding

Research in context

Articles AI-driven diffusion weighted imaging-based non-contrast protocol for breast cancer diagnosis: a multicentre, multidimensional validation study Yulu Liu,a,b,k Haoquan Chen,a,k Jiaqi Zhao,c,k Lieji Wang,d Zimei Gao,e Yuan Peng,f Wenhui Ren,g Kaining Shi,d Sen Zhang,a Xiufeng Chen,h Chenxin Pang,i Manle Yan,j Caizhen Feng,a Fan Chai,a Jin Gao,b Xiaoxuan Jia,a,∗∗∗∗ Shu Wang,f,∗∗∗ Nan Hong,a,∗∗ and Yi Wanga,∗ aDepartment of Radiology, Peking University People’s Hospital, Beijing, 100044, China bDepartment of Radiology, The Third People’s Hospital of Chengdu/ the Affiliated Hospital of Southwest Jiaotong University, Chengdu, 610031, China cDepartment of Radiology, Jiangmen Central Hospital, Jiangmen, 529030, China dGE Healthcare, Guangzhou, 510623, China eDepartment of Radiology, Shijiazhuang People’s Hospital, Shijiazhuang, 050031, China fDepartment of Breast Surgery, Peking University People’s Hospital, Beijing, 100044, China gDepartment of Clinical Epidemiology and Biostatistics, Peking University People’s Hospital, Beijing, 100044, China hDepartment of General Surgery, Beijing Aerospace General Hospital, Beijing, 100049, China iSchool of Healthcare Technology, Chengdu Neusoft University, Chengdu, 611844, China jSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang, 621010, China Summary Background Contrast-enhanced breast MRI, while highly sensitive, faces limitations including complexity, long acquisition times, and reliance on gadolinium-based contrast agents. Noncontrast diffusion-weighted imaging (DWI) offers an ultrafast alternative, but its diagnostic accuracy has been insufficient for standalone use. We investigated whether a deep learning (DL) model could enable accurate breast cancer diagnosis using only DWI. Methods This four-centre study included 2493 patients with pathologically confirmed breast lesions. A DWI-based model (DWI-DL) was developed using data from 1286 patients at Peking University People’s Hospital (PKUPH; January 2015 to July 2021), who were randomly divided into three cohorts at a ratio of 6:2:2 (train n = 774, validation n = 256, test n = 256) . Independent testing used three external cohorts (n = 661) and a prospective cohort from PKUPH that was retrospectively analysed (n = 546; August 2021 to September 2022) . The diagnostic performance of DWI-DL model was compared to that of an abbreviated enhanced-based (AE-DL) model and two expert radiologists. Subgroup analysis assessed performance across external and prospective cohorts. Amultireader multicase validation, using DW-DL model and selective contrast-enhanced workflow, was conducted by 12 radiologists from ten institutions to evaluate the clinical utility of DWI-DL in assisting diagnosis. Performance was assessed using the area under the curve (AUC) of receiver operating characteristics. The multireader multicase study was registered with the China Clinical Trial Registry (ChiCTR2500095953) . Findings DWI-DL model demonstrated diagnostic performance across all cohorts that was comparable to the AE-DL model (AUC: 0.771–0.912 vs. 0.780–0.898, p = 0.36–0.98). DWI-DL outperformed two expert radiologists interpreting DWI alone (AUC: 0.781–0.858 vs. 0.714–0.770, p range <0.0001–0.023) . In the multireader multicase validation, the AI-guided selective sequence protocol was non-inferior to the full protocol (Protocol E vs. C, AUC: 0.834 [95% CI: 0.785, 0.883] vs. 0.835 [95% CI: 0.789, 0.881], difference: −0.001 [95% CI: −0.029, 0.027], p = 0.94), while simultaneously reduced interpretation time by 55.5%(Protocol E vs. C, mean time: 59.6 [43.5] vs. 134.0 [72.4] seconds, p < 0.0001) . Interpretation A DL model using only non-contrast DWI can accurately diagnose breast cancer. The DWI-DL model demonstrated robust performance, comparable to the AE-DL model and surpassing human experts in DWI *Corresponding author. **Corresponding author. ***Corresponding author. ****Cor

FAQs

What problem does the study address in breast MRI diagnosis?

It targets limitations of contrast-enhanced breast MRI, including complexity, long acquisition times, and reliance on gadolinium-based contrast agents. It tests whether a DWI-only approach can replace contrast for accurate diagnosis.

How was the DWI-based deep learning model evaluated?

The study developed a DWI-DL model using one hospital cohort, validated and tested it internally, and then assessed performance on external cohorts plus a prospective cohort retrospectively analyzed. It compared results to an abbreviated enhanced-based model and two expert radiologists.

What were the key outcomes of using DWI-DL and an AI-guided selective protocol?

DWI-DL showed diagnostic performance comparable to the abbreviated enhanced-based model across cohorts and outperformed expert radiologists when interpreting DWI alone. In multireader multicase validation, the AI-guided selective contrast-enhanced workflow was non-inferior to the full protocol while reducing interpretation time by 55.5%.

Sep 29, 2026

Uploaded by Terk