CT-based Deep Learning Model for Automatic Segmentation and Early Predicting of Pyogenic Liver Abscess Caused by Extended-Spectrum β-lactamase-Producing Enterobacteriaceae: A Multicenter Retrospective Study (CLASS2401).
Authors
Affiliations (14)
Affiliations (14)
- Department of Radiology, Shengjing Hospital of China Medical University, Shenyang, China (Z.G., H.W., Y.G., Y.W., C.W., Z.C.).
- Department of Interventional Therapy, The First Affiliated Hospital of Dalian Medical University, Dalian, China (T.P.).
- Department of Interventional Radiology, The Fourth People's Hospital of Shenyang, China Medical University, Shenyang, China (D.Z.).
- Department of Radiology, Sir Run Run Shaw Hospital, Zhejiang University School of Medicine, Hangzhou, China (P.Y.).
- Department of Interventional Radiology, Longwan Branch of Central Hospital of Huludao, Huludao, China (Q.Z.).
- Department of Radiology, Xijing Hospital, Air Force Medical University, Xi'an, Shaanxi, China (Z.H.).
- Department of Interventional Vascular, Hongqi Hospital, Mudanjiang Medical University, Mudanjiang, China (X.Y.).
- Department of Interventional Radiology, Ansteel General Hospital, Anshan, China (Y.W.).
- Interventional Department, The Third Affiliated Hospital of Jinzhou Medical University, Jinzhou, China (J.L.).
- Department of Interventional Radiology, Zhoukou Central Hospital, Zhoukou, China (C.Y.).
- Department of Radiology, the People's Hospital of Tibet Autonomous Region, Lhasa, China (C.W.).
- School of Health Management, China Medical University, Shenyang, Liaoning, China (J.S.).
- Department of Radiology, Qilu Hospital of Shandong University Dezhou Hospital (Dezhou People's Hospital), Dezhou, China (Z.L.).
- Department of Radiology, Shengjing Hospital of China Medical University, Shenyang, China (Z.G., H.W., Y.G., Y.W., C.W., Z.C.). Electronic address: [email protected].
Abstract
Pyogenic liver abscess (PLA) caused by extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PLA) presents major antimicrobial treatment challenges and delayed pathogen identification. This study developed and validated deep learning-based models integrating clinical and CT imaging features for early prediction of ESBL-PLA prior to microbiological confirmation. This retrospective multicenter study included 442 patients from 6 centers in China, divided into training, internal test, and external test sets (n = 211, 111, and 120, respectively). An automated PLA segmentation model based on nnUNetv2 was developed using expert-annotated lesions. Clinical, radiomics, and deep learning imaging features were used to construct predictive basic models. Two combined models, the clinical-radiomics model and clinical-imaging model (CIM) were constructed. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and other metrics. The segmentation model achieved high Dice similarity coefficient values across datasets. The CIM showed strong predictive performance for ESBL-PLA, with AUCs of 0.945, 0.889, and 0.844 in the training, internal test, and external test sets, respectively. At the optimal cutoff (0.663), the CIM achieved positive predictive values of 95.2%, 86.6%, and 70.7%, and negative predictive values of 95.0%, 96.3%, and 95.7% in the training, internal test, and external test sets, respectively. And the high-risk group had a significantly higher incidence of ESBL-PLA (74.58%) than the low-risk group (6.00%, P < 0.001). Deep learning-based radiomics enables early imaging-based prediction of ESBL-PLA and may support imaging-based risk stratification for infections caused by multidrug-resistant pathogens.