Modified Mixture of Experts for the Diagnosis of Perfusion Magnetic Resonance Imaging Measures in Locally Rectal Cancer Patients / 대한의료정보학회지

OBJECTIVES: This study demonstrates the feasibility of using a modified mixture of experts (ME) model with repeated measured tumoural Ktrans value to perform an automatic diagnosis of responder based on perfusion magnetic resonance imaging (MRI) of rectal cancer. METHODS: The data used in this study was obtained from 39 patients with primary rectal carcinoma who were scheduled for preoperative chemoradiotherapy. The modified ME model is a joint modeling of the ME model via the linear mixed effect model. First, we considered two local experts and a gating network, and the modified expert network as a liner mixed effect model. Afterward, the finding estimates were obtained via the expectation-maximization algorithm. All computation was performed by R-2.15.2. RESULTS: We found that two experts have different patterns. The feature of expert 1 (n = 10) had a higher baseline value and a lower slope than expert 2 (n = 29). A comparison of the estimated experts and responder/non-responder groups according to T-downstaging criteria showed that expert 1 had a more effect treatment responder than expert 2. CONCLUSIONS: A novel feature of this study is that it is an extension of classical ME models in case of repeatedly measured data. The proposed model has the advantages of flexibility and adaptability for identifying distinct subgroups with various time patterns, and it can be applied to biomedical data which is measured repeatedly, such as time-course microarray data or cohort data. This method can assist physicians as important diagnostic decision making mechanism.

A Mixture of Experts Model for the Diagnosis of Liver Cirrhosis by Measuring the Liver Stiffness / 대한의료정보학회지

OBJECTIVES: The mixture-of-experts (ME) network uses a modular type of neural network architecture optimized for supervised learning. This model has been applied to a variety of areas related to pattern classification and regression. In this research, we applied a ME model to classify hidden subgroups and test its significance by measuring the stiffness of the liver as associated with the development of liver cirrhosis. METHODS: The data used in this study was based on transient elastography (Fibroscan) by Kim et al. We enrolled 228 HBsAg-positive patients whose liver stiffness was measured by the Fibroscan system during six months. Statistical analysis was performed by R-2.13.0. RESULTS: A classical logistic regression model together with an expert model was used to describe and classify hidden subgroups. The performance of the proposed model was evaluated in terms of the classification accuracy, and the results confirmed that the proposed ME model has some potential in detecting liver cirrhosis. CONCLUSIONS: This method can be used as an important diagnostic decision support mechanism to assist physicians in the diagnosis of liver cirrhosis in patients.