Research progress on medical image dataset expansion methods / 生物医学工程学杂志

Computer-aided diagnosis (CAD) systems play a very important role in modern medical diagnosis and treatment systems, but their performance is limited by training samples. However, the training samples are affected by factors such as imaging cost, labeling cost and involving patient privacy, resulting in insufficient diversity of training images and difficulty in data obtaining. Therefore, how to efficiently and cost-effectively augment existing medical image datasets has become a research hotspot. In this paper, the research progress on medical image dataset expansion methods is reviewed based on relevant literatures at home and abroad. First, the expansion methods based on geometric transformation and generative adversarial networks are compared and analyzed, and then improvement of the augmentation methods based on generative adversarial networks are emphasized. Finally, some urgent problems in the field of medical image dataset expansion are discussed and the future development trend is prospected.

A survey of loss function of medical image segmentation algorithms / 生物医学工程学杂志

Medical image segmentation based on deep learning has become a powerful tool in the field of medical image processing. Due to the special nature of medical images, image segmentation algorithms based on deep learning face problems such as sample imbalance, edge blur, false positive, false negative, etc. In view of these problems, researchers mostly improve the network structure, but rarely improve from the unstructured aspect. The loss function is an important part of the segmentation method based on deep learning. The improvement of the loss function can improve the segmentation effect of the network from the root, and the loss function is independent of the network structure, which can be used in various network models and segmentation tasks in plug and play. Starting from the difficulties in medical image segmentation, this paper first introduces the loss function and improvement strategies to solve the problems of sample imbalance, edge blur, false positive and false negative. Then the difficulties encountered in the improvement of the current loss function are analyzed. Finally, the future research directions are prospected. This paper provides a reference for the reasonable selection, improvement or innovation of loss function, and guides the direction for the follow-up research of loss function.

MRI features of primary lymphoma in central nervous system / 实用放射学杂志

Objective To study the MRI features of primary central nervous system lymphoma (PCNSL)to improve its diagnosis. Methods The MRI data of 1 3 patients with PCNSL confirmed by biopsy or surgical pathology were analyzed retrospectively.All patients underwent nonenhanced and contrast-enhanced MRI examinations.Results Multiple lesions in 8 cases and solitary one in 5 cases were found.Five leisons (38.5%)were located in the cerebral hemispheres.Eight(61.5%)were located in deep structures of the brain including periventricular white matter,basal ganglia,thalami and corpus callosum,6 (46.2%)of whom involved in the ependyma including 1 involved in the ependyma of the 4 th ventricle.One leison(7.7%)was located in the cerebellar hemisphere.The tumor showed isointense or mildly hypointense signal on T1WI,mildly hyperintense signal on T2WI and hyperintense signal on DWI.Mild to moderate perilesional edema was noted.The tumors in 13 cases showed remarkable homogeneous enhancement after the contrast agent was injected with the signs of characteristic notch,butterfly and satellite.Focal necrosis occurred in 1 lesions.Conclusion PCNSL shouldbe first considered when a lesion located in the deep central structeres of the brain,multiple and involve ependymal membrane,showed isointensity on T1WI and T2WI,highintensity on DWI in tumoral parenchyma demonstrated,significantly enhancement in the solid part of the tumor.

MRI fingings of central neuroeytoma / 实用放射学杂志

Objective To analyze the MRI features of central neurocytoma (CNC),and to improve the diagnostic accuracy of this lesion.Methods The MRI findings of 12 patients with CNC confirmed by surgery and pathology were evaluated retrospectively.All patients underwent contrast-enhanced and uncontrast-enhanced MR scan,and 3 patients also received plain CT scan.Results The lesions of 12 cases were all located in the lateral ventricles of septum pellucidum,which were close to the Monro's foramen.5 lesions located in the right lateral ventricle,5 lesions located in both lateral ventricles with 1 case reached the third ventricle,2 lesions located in the left lateral ventricle with 1 case reached the third ventricle.The mean diameter of the CNC lesions was 4.8 cm (ranged from 2.2 to 7.8 cm).11cases (92％) had vavious degrees of cystic change,8 cases (66.7％) had a spongy or soap bubble appearance and 1 case(8.3％) had completely cystic change.3 cases (25％) had a scalloping appearance on sagittal T1-weighted MRI.The solid part of the tumor was similar to the gray matter in signal intensity,with slightly high signal intensity on T2FLAIR.Among 9 cases with DWI,7 cases showed high signal intensity,1 case showed equal signal intensity,and 1 case showed slightly low signal intensity.On enhanced MRI,the solid part of tumors showed slight to medium (83％) or obvious (17％) enhancement.Conclusion CNC should be first considered when a young adult has a lesion located in the lateral ventricles of septum pellucidum,accompainied by varying degrees of cysts,showed slight to medium enhancement in the solid part of the tumor.