Artificial intelligence-based segmentation of small renal masses: a multi-center, multi-scanner, multi-sequence study.
Authors
Affiliations (8)
Affiliations (8)
- Department of Radiology, First Medical Center, Chinese PLA General Hospital, Beijing, China.
- State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, China.
- Department of Radiology, Seventh Medical Center, Chinese PLA General Hospital, Beijing, China.
- Radiology Department, Peking University First Hospital, Beijing, China.
- Department of Radiology, Beijing Friendship Hospital, Beijing, China.
- Department of Pathology, The First Medical Center, Chinese PLA General Hospital, Beijing, China.
- Department of Innovative Medical Research, Chinese PLA General Hospital, Beijing, China.
- Department of Radiology, First Medical Center, Chinese PLA General Hospital, Beijing, China. [email protected].
Abstract
This study aims to develop an artificial intelligence (AI)-based automated segmentation method for small renal masses (SRMs) using multi-center, multi-scanner, multi-sequence MRI data. MR images from 988 pathologically confirmed SRM patients from three different centers were retrospectively included. Segmentation networks were independently developed for each MRI sequence using deep learning techniques. A GE dataset of 733 patients from Center 1 was used for training and validation. A GE test set, consisting of internal (99 from Center 1) and external test sets (81 from Center 2 and 3), was created for evaluation. Furthermore, a non-GE generalization set, consisting of 75 patients from Center 2 and 3, was used to assess the generalization ability. The method's performance was evaluated in terms of detection rate and segmentation accuracy (Dice similarity coefficient [DSC]). Subgroup analysis and multiple linear regression were used for further exploration. Our method demonstrated promising results in the detection and segmentation of SRMs. All patients in the GE test set were correctly detected in at least one sequence. Our model achieved a median DSC of 0.769-0.855 across five MRI sequences and demonstrated reasonable generalization to non-GE scanners (median DSC range: 0.523-0.785). The implementation of automated segmentation achieved encouraging outcomes in both correct-detection rates and segmentation accuracy across a diverse cohort spanning multiple centers and scanners, suggesting its potential as a key component of future diagnostic pipelines for SRMs.