Objective:To establish and evaluate a new real-time quality control method that can identify the random errors by using the backpropagation neural network (BPNN) algorithm and taking blood glucose test as an example.Methods:A total of 219 000 blood glucose results measured by Siemens advia 2 400 analytical system from January 2019 to July 2020 and derived from Laboratory Information System of Beijing Chaoyang Hospital Laboratory Department was regarded as the unbiased data of our study. Six deviations with different sizes were introduced to generate the corresponding biased data. With each biased data, BPNN and MovSD algorithms were used and tested, and then evaluated by traceability method and clinical method.Results:For BPNN algorithm, the block size was pre-set to 10 and the false-positive rate in all biases was within 0.1%. For MovSD, however, the optimal block size and exclusive limit were 150 and 10% separately and its false-positive rate in all biases was 0.38%, which was 0.28% higher than BPNN. Especially, for the least two error factors of 0.5 and 1, all the random errors were not detected by MovSD; for the error factor larger than 1, random errors could be detected by MovSD but the MNPed was higher than that of BPNN under all deviations. The difference was up to 91.67 times. 460 000 reference data were produced by traceability procedure. The uncertainty of BPNN algorithm evaluated by these reference data was only 0.078%.Conclusion:A real-time quality control method based on BPNN algorithm was successfully established to identify random errors in analytical phase, which was more efficient than MovSD method and provided a new idea and method for the identification of random errors in clinical practice.

Intelligent medical image segmentation methods have been rapidly developed and applied, while a significant challenge is domain shift. That is, the segmentation performance degrades due to distribution differences between the source domain and the target domain. This paper proposed an unsupervised end-to-end domain adaptation medical image segmentation method based on the generative adversarial network (GAN). A network training and adjustment model was designed, including segmentation and discriminant networks. In the segmentation network, the residual module was used as the basic module to increase feature reusability and reduce model optimization difficulty. Further, it learned cross-domain features at the image feature level with the help of the discriminant network and a combination of segmentation loss with adversarial loss. The discriminant network took the convolutional neural network and used the labels from the source domain, to distinguish whether the segmentation result of the generated network is from the source domain or the target domain. The whole training process was unsupervised. The proposed method was tested with experiments on a public dataset of knee magnetic resonance (MR) images and the clinical dataset from our cooperative hospital. With our method, the mean Dice similarity coefficient (DSC) of segmentation results increased by 2.52% and 6.10% to the classical feature level and image level domain adaptive method. The proposed method effectively improves the domain adaptive ability of the segmentation method, significantly improves the segmentation accuracy of the tibia and femur, and can better solve the domain transfer problem in MR image segmentation.
Humans ; Image Processing, Computer-Assisted/methods* ; Neural Networks, Computer ; Magnetic Resonance Imaging ; Knee ; Knee Joint