Objective To explore and address the issue of insufficient segmentation accuracy in liver tumor segmentation using the traditional U-Net algorithm,which is caused by the lack of contextual information for both the liver and tumor,as well as the large morphological variability of tumors.Methods A cascaded liver tumor segmentation algorithm,DDR-UNet++,which integrated dilated convolutions and residual modules was proposed.Firstly,CT images from the LiTS-2017 dataset were preprocessed through window width/level adjustment,histogram equalization and Gaussian filtering to reduce noise and smooth edges.Then,a cascaded liver segmentation model was employed to enhance the liver region proportion,mitigate interference from surrounding tissues and address data imbalance issue.For liver tumor segmentation,deformable dilated convolutions and residual networks were introduced to expand the receptive field and improve feature extraction capability.Results DDR-UNet++outperformed the traditional U-Net on the LiTS-2017 dataset,achieving improvements of 4.7%,1.7%,and 8.5%in Dice similarity coefficient,relative volume difference,and Jaccard index,respectively.These enhancements contribute to overcoming the inefficiency and low accuracy issues in conventional tumor segmentation,thereby improving early tumor detection rates,enhancing patient survival outcomes,and alleviating the diagnostic burden on clinicians.Conclusion The proposed method improves the feature extraction capability to some extent by enhancing the model structure and segmentation strategy,effectively increases the accuracy and robustness of liver tumor segmentation,and provides a reliable technical reference for clinical auxiliary diagnosis.

Objective To explore and address the issue of insufficient segmentation accuracy in liver tumor segmentation using the traditional U-Net algorithm,which is caused by the lack of contextual information for both the liver and tumor,as well as the large morphological variability of tumors.Methods A cascaded liver tumor segmentation algorithm,DDR-UNet++,which integrated dilated convolutions and residual modules was proposed.Firstly,CT images from the LiTS-2017 dataset were preprocessed through window width/level adjustment,histogram equalization and Gaussian filtering to reduce noise and smooth edges.Then,a cascaded liver segmentation model was employed to enhance the liver region proportion,mitigate interference from surrounding tissues and address data imbalance issue.For liver tumor segmentation,deformable dilated convolutions and residual networks were introduced to expand the receptive field and improve feature extraction capability.Results DDR-UNet++outperformed the traditional U-Net on the LiTS-2017 dataset,achieving improvements of 4.7%,1.7%,and 8.5%in Dice similarity coefficient,relative volume difference,and Jaccard index,respectively.These enhancements contribute to overcoming the inefficiency and low accuracy issues in conventional tumor segmentation,thereby improving early tumor detection rates,enhancing patient survival outcomes,and alleviating the diagnostic burden on clinicians.Conclusion The proposed method improves the feature extraction capability to some extent by enhancing the model structure and segmentation strategy,effectively increases the accuracy and robustness of liver tumor segmentation,and provides a reliable technical reference for clinical auxiliary diagnosis.

Objective To analyze the results of the final aviation medical examination of 1281 students recruited into Air Force youth aviation schools in Hunan and Hubei provinces in order to provide reference for establishing the items and standards of medical selection .Methods The data of 1281 students who participated in final aviation medical examination of Air Force youth aviation schools in 2016 were collected , who came from 28 cities in the above two provinces .The disqualification rate and related unqualified medical items were calculated , and the differences of the disqualification rate and medical geographical areas in the 28 cities were analyzed .Results According to the disqualification rate , the top five departments were ophthalmology , otolaryngology, surgery, radiology and ultrasonic departments .The top 10 unqualified items were the lack of distant vision , fundus diseases , nasal anomaly , ametropia and strabismus , spine abnormality , audition abnormality, vestibular function badness , vitreous opacity, and lens abnormalities.There was no significant difference between the 28 cities in the disqualification rate (P >0.05) or between the two provinces (P >0.05). Conclusion The results of the final aviation medical examination reflect the quality and efficiency of the initial aviation medical examination .To improve the quality of medical selection , further research is needed to set a scientific standard for the initial aviation medical examination while strengthening the scientific protection and intervention of distant vision .The efficiency of selection depends on improving the accuracy of initial aviation medical examination in nasal cavity structure, body shape ,and lens opacity .With a better understanding of the disqualification rate and abnormal items in different cities , a scientific arrangement of professional staff and technical force can make the initial aviation medical examination better, thus effectively reducing the rate of false elimination rate and misdiagnosis .

BACKGROUND:Bone marrow mesenchymal stem cell(BMSC) transplantation is one of the developmental directions in the treatment of femoral head necrosis. In recent years, the use of superparamagnetic iron oxide (SPIO) nanoparticles labeled target cells traced by MRI imaging method has become the focus of the study.
OBJECTIVE:To observe the in vivo MRI tracking and the curative effects of SPIO-labeled BMSC transplantation on rabbit femoral head necrosis.
METHODS:SPIO-labeled BMSCs, unlabeled BMSCs, and normal saline were injected in situ into the necrotic femoral head of rabbits. Fol owing MRI dectection, the image changes of transplanted BMSCs marked by SPIO were observed among the three scanning sequences of SE T2WI, FSE T2WI and GRE T2*WI. Meanwhile, the area percentage of newly formed bone trabecula in the defect samples under high power lens were observed and calculated for statistical analysis.
RESULTS AND CONCLUSION:In situ celltransplantation group showed the emerging and extinctive time of the decreased-signal region was different among the three scanning sequences of SE T2WI, FSE T2WI and GRE T2*WI. It was found that the decreased-signal region of the MRI scanning sequences was the target of the present experiment. No obvious signal change occurred in the control side. After 6 weeks of transplantation, the area percentage of newly formed bone trabecula in the defect samples showed no difference in SPIO-labeled and unlabeled BMSC transplantation groups (P>0.05), but it was higher than that in the control side (P<0.01). The SPIO-labeled BMSCs and unlabeled BMSCs are shown to have the same effects in the treatment of femoral head necrosis. The SPIO-labeled BMSCs can be observed obviously by MRI detection in vitro.