Construction and clinical validation of a cascaded deep learning system for classification of benign and…

AI / informaticsEuropean journal of radiology· Jul 2026

Construction and clinical validation of a cascaded deep learning system for classification of benign and malignant solid small renal masses based on MRI.

MengQiu Cui, ZiLong Zeng, SiLu Chen … HaiYi Wang

Abdominal & Body ImagingAI / informaticsDiagnosisQ1IF 4.0

Deep learning model distinguished benign from malignant small renal masses with high accuracy.

Construction and clinical validation of a cascaded deep learning system for classification of benign and malignant solid small renal masses based on MRI.

MengQiu Cui … HaiYi Wang

European journal of radiology · Jul 2026 · Q1 in Radiology

Purpose

This study aims to develop a cascaded deep learning (deep learning) system based on multiparametric MRI to establish an automated pipeline for the segmentation and classification of small renal masses (small renal masses).

Methods

A retrospective collection of small renal masses patients with pathologically confirmed from three institutions was conducted.

n = 965 patients

Results

correctly told benign from malignant nearly every time, near-perfect on this 0-1 scale

Internal test set

0.936AUC

External test set

0.788AUC

More results

Institution 1 contributed 888 cases, with 712 used for training and 176 as an internal test set; Institutions 2 and 3 provided 77 cases as an external test set.

The model significantly outperformed the junior radiologist on both test sets.

More results

This finding remained consistent in the subgroup of tumors smaller than 3 cm.

“

Conclusion

The cascaded deep learning system demonstrated robust performance across multiple centers, enabling non-invasive and efficient discrimination of small renal masses malignancy, showing promise as a clinical support tool.

Construction and clinical validation of a cascaded deep learning system for classification of benign and malignant solid small renal masses based on MRI.

Construction and clinical validation of a cascaded deep learning system for classification of benign and malignant solid small renal masses based on MRI.

Related papers