Preoperative Risk Assessment of Adrenal Metastases in a Multicenter Study: Development of a Robust Federated Learning Model.
Authors
Affiliations (13)
Affiliations (13)
- Laboratory of Intelligent Detection and Information Processing, Guilin University of Aerospace Technology, Guilin 541004, China (B.F., Z.Y., Y.C., J.X., Q.H.); Jiangmen Key Laboratory of Artificial Intelligence in Medical Image Computation and Application, Jiangmen Central Hospital, Jiangmen 529030, China (B.F., W.L., E.C.). Electronic address: [email protected].
- Laboratory of Intelligent Detection and Information Processing, Guilin University of Aerospace Technology, Guilin 541004, China (B.F., Z.Y., Y.C., J.X., Q.H.). Electronic address: [email protected].
- Laboratory of Intelligent Detection and Information Processing, Guilin University of Aerospace Technology, Guilin 541004, China (B.F., Z.Y., Y.C., J.X., Q.H.). Electronic address: [email protected].
- Laboratory of Intelligent Detection and Information Processing, Guilin University of Aerospace Technology, Guilin 541004, China (B.F., Z.Y., Y.C., J.X., Q.H.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.). Electronic address: [email protected].
- Department of Radiology, The First Affiliated Hospital of Shenzhen University, Health Science Center, Shenzhen Second People's Hospital, Shenzhen 518035, China (F.L.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.). Electronic address: [email protected].
- Laboratory of Intelligent Detection and Information Processing, Guilin University of Aerospace Technology, Guilin 541004, China (B.F., Z.Y., Y.C., J.X., Q.H.). Electronic address: [email protected].
- School of Electrical Engineering, Guangxi University, Nanning 530004, China (Q.J.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.); Jiangmen Key Laboratory of Artificial Intelligence in Medical Image Computation and Application, Jiangmen Central Hospital, Jiangmen 529030, China (B.F., W.L., E.C.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.). Electronic address: [email protected].
- Department of Radiology, Jiangmen Central Hospital, Jiangmen 529030, China (Y.L., M.W., J.C., W.L., C.M., E.C.); Jiangmen Key Laboratory of Artificial Intelligence in Medical Image Computation and Application, Jiangmen Central Hospital, Jiangmen 529030, China (B.F., W.L., E.C.). Electronic address: [email protected].
Abstract
In clinical practice, the preoperative risk assessment of adrenal metastases versus benign adrenal lesions carries a substantial risk of misdiagnosis. The artificial intelligence technology holds promise for reducing misdiagnosis rates. However, due to the problem of data privacy protection and non-independent and identically distributed of multi-center data, the performance of artificial intelligence models is significantly affected. We retrospectively collected 1187 adrenal lesions from 1100 patients who underwent three phase Computed tomography (CT) scanning between January 2008 and September 2021. And then, we combined graph structure networks and mutual information to construct a robust federated learning model (RFLM) for the preoperative risk assessment of adrenal metastases and adrenal benign lesions. We conducted experiments on three-phase (pre-contrast phase (PCP), venous phase (VP) and arterial phase (AP)). In addition, we also conduct correlation analysis of the features of each center in RFLM to explore the intrinsic mechanism of the model. PCP-based RFLM outperformed AP/VP and the evaluated federated learning baselines on all four center-specific test sets, with AUCs of 0.7826, 0.8318, 0.8794, and 0.8895 for Centers 1-4, respectively. Additional analyses supported its superior performance and potential clinical utility across centers. Our results indicate that PCP-based RFLM enables effective preoperative risk assessment of adrenal metastases and benign lesions, potentially reducing reliance on enhanced CT and thus minimizing additional radiation and complications. These findings underscore RFLM as a robust and reliable diagnostic approach.