Objective:To investigate the denoising performance of different deep learning (DL) strategies on low-dose brain 18F-FDG PET images. Methods:This retrospective methodological study was conducted on brain PET/CT images of 50 patients (35 males, 15 females, age 20-87 years) who received 3.7MBq/kg 18F-FDG at the Fifth Affiliated Hospital of Sun Yat-sen University between May 2023 and January 2024. Full-dose PET data were acquired with 2min scan. CT scans were acquired before PET scanning. Low-dose PET sinograms were generated by down-sampling the full-dose list mode data to 1/2, 1/4, and 1/20 of full-dose count level. Both full-dose and low-dose sinograms were reconstructed with random, CT-based attenuation and scatter corrections using the three-dimensional (3D) ordered-subsets expectation maximization (OSEM) algorithm (2 iterations, 20 subsets). A total of 4 DL denoising methods were established: (1) 3D conditional generative adversarial networks (GAN) using only low-dose PET as input (GAN-1); (2) 3D attention-based GAN (AttGAN) with low-dose PET input (AttGAN-1); (3) 3D AttGAN with low-dose PET and CT inputs (AttGAN-2); (4) 3D AttGAN with frequency-separation using low-dose PET and CT inputs (AttGAN-FS-2). For AttGAN-FS-2, during the frequency division process, high- and low-frequency components were extracted from the PET reconstructed images via Fourier transform, then inversed Fourier transform, denoised separately, and finally combined to produce the final denoised images. The dataset was separated into training (70%), validation (10%) and testing (20%) sets using simple random sampling without replacement with a fixed random seed. A 5-fold cross-validation scheme was then applied to test all 50 patients. Performance was evaluated against full-dose PET using normalized mean square error (NMSE), structural similarity (SSIM), peak signal-to-noise ratio (PSNR), contrast-to-noise ratio (CNR), SUV mean and SUV max bias of selected brain ROIs. Wilcoxon signed rank test was used to analyze the differences between the denoising methods. Results:AttGAN-FS-2 showed the best performance among all dose levels, with statistical difference as compared by low-dose PET and GAN-1 denoised images for NMSE, SSIM, PSNR, and CNR ( Z values: 2.92-6.15, all P<0.005). NMSE, SSIM quantitative evaluation results (median) of each model at 1/20 dose were: GAN-1: 0.08, 0.87, AttGAN-1: 0.08, 0.88, AttGAN-2: 0.07, 0.89, AttGAN-FS-2: 0.06, 0.91, respectively ( Z values: 3.24-5.77, all P<0.005). Conclusion:The DL-based method combined with multiple strategies AttGAN-FS-2 shows improved denoising performance for low-dose brain PET images.

Objective:To evaluate the imaging stability of a portable boom-type ophthalmic ultra-high-resolution optical coherence tomography (OCT) device in multiple application scenarios.Methods:The boom-type mode and handheld mode of the portable boom-type OCT and the desktop OCT were used to perform three-dimensional imaging tests on three healthy adults undergoing physical examinations at the Zhongshan Ophthalmic Center, Sun Yat-sen University as well as on OEMI-7 model eyes in a sitting position.The same two modes of the portable boom-type OCT were used to perform three-dimensional imaging on four awake non-sedated infants, two sedated infants and four healthy adults in the supine position.The obtained 3D imaging data were processed using a correlation analysis method between adjacent B-scans, and the offset of B-scan in the axial (z-axis) and the fast axis transverse (x-axis) were calculated.The procedures for in vivo human eye experiments followed the Declaration of Helsinki and were approved by the Ethics Committee of Zhongshan Ophthalmic Center, Sun Yat-sen University (No.2020 KYPJ154).All subjects and infant guardians signed the informed consent form. Results:Compared with the handheld imaging mode, the axial and fast axis lateral motion offsets of the model eye were significantly reduced in the boom-type imaging mode from (124.00±12.49)μm to (48.00±15.87)μm and from (24.00±1.00)μm to (2.67±0.57)μm, respectively ( t=2.932, 4.337; both P<0.001).In both human and model eyes, the axial and fast axis lateral motion offsets of the boom-type mode were significantly lower than in the traditional handheld operation mode (both P<0.001).The axial and lateral motion offsets between the boom-type mode and desk-top OCT imaging were comparable, without significant differences (both P>0.05).In both sedated and awake, non-sedated infants in the supine position, the axial offset of the portable boom-type OCT system was similar to that of the healthy adults, without significant difference in the overall comparison ( P=0.385), and the lateral offsets were higher than those of healthy adults, with statistically significant differences (both P=0.013).There was no significant difference in axial deviation between sedated and non-sedated infants ( P>0.05).The lateral deviation of non-sedated infants was higher than that of sedated infants, though the difference was not statistically significant ( P=0.247). Conclusions:The portable boom-type OCT system can maintain high-speed, high-resolution imaging performance while achieving imaging stability comparable to traditional desktop OCT systems.It is more suitable for bedside imaging of supine subjects, especially uncooperative infants, and has good clinical application prospects.

Objective:To evaluate the imaging stability of a portable boom-type ophthalmic ultra-high-resolution optical coherence tomography (OCT) device in multiple application scenarios.Methods:The boom-type mode and handheld mode of the portable boom-type OCT and the desktop OCT were used to perform three-dimensional imaging tests on three healthy adults undergoing physical examinations at the Zhongshan Ophthalmic Center, Sun Yat-sen University as well as on OEMI-7 model eyes in a sitting position.The same two modes of the portable boom-type OCT were used to perform three-dimensional imaging on four awake non-sedated infants, two sedated infants and four healthy adults in the supine position.The obtained 3D imaging data were processed using a correlation analysis method between adjacent B-scans, and the offset of B-scan in the axial (z-axis) and the fast axis transverse (x-axis) were calculated.The procedures for in vivo human eye experiments followed the Declaration of Helsinki and were approved by the Ethics Committee of Zhongshan Ophthalmic Center, Sun Yat-sen University (No.2020 KYPJ154).All subjects and infant guardians signed the informed consent form. Results:Compared with the handheld imaging mode, the axial and fast axis lateral motion offsets of the model eye were significantly reduced in the boom-type imaging mode from (124.00±12.49)μm to (48.00±15.87)μm and from (24.00±1.00)μm to (2.67±0.57)μm, respectively ( t=2.932, 4.337; both P<0.001).In both human and model eyes, the axial and fast axis lateral motion offsets of the boom-type mode were significantly lower than in the traditional handheld operation mode (both P<0.001).The axial and lateral motion offsets between the boom-type mode and desk-top OCT imaging were comparable, without significant differences (both P>0.05).In both sedated and awake, non-sedated infants in the supine position, the axial offset of the portable boom-type OCT system was similar to that of the healthy adults, without significant difference in the overall comparison ( P=0.385), and the lateral offsets were higher than those of healthy adults, with statistically significant differences (both P=0.013).There was no significant difference in axial deviation between sedated and non-sedated infants ( P>0.05).The lateral deviation of non-sedated infants was higher than that of sedated infants, though the difference was not statistically significant ( P=0.247). Conclusions:The portable boom-type OCT system can maintain high-speed, high-resolution imaging performance while achieving imaging stability comparable to traditional desktop OCT systems.It is more suitable for bedside imaging of supine subjects, especially uncooperative infants, and has good clinical application prospects.

Objective:To investigate the denoising performance of different deep learning (DL) strategies on low-dose brain 18F-FDG PET images. Methods:This retrospective methodological study was conducted on brain PET/CT images of 50 patients (35 males, 15 females, age 20-87 years) who received 3.7MBq/kg 18F-FDG at the Fifth Affiliated Hospital of Sun Yat-sen University between May 2023 and January 2024. Full-dose PET data were acquired with 2min scan. CT scans were acquired before PET scanning. Low-dose PET sinograms were generated by down-sampling the full-dose list mode data to 1/2, 1/4, and 1/20 of full-dose count level. Both full-dose and low-dose sinograms were reconstructed with random, CT-based attenuation and scatter corrections using the three-dimensional (3D) ordered-subsets expectation maximization (OSEM) algorithm (2 iterations, 20 subsets). A total of 4 DL denoising methods were established: (1) 3D conditional generative adversarial networks (GAN) using only low-dose PET as input (GAN-1); (2) 3D attention-based GAN (AttGAN) with low-dose PET input (AttGAN-1); (3) 3D AttGAN with low-dose PET and CT inputs (AttGAN-2); (4) 3D AttGAN with frequency-separation using low-dose PET and CT inputs (AttGAN-FS-2). For AttGAN-FS-2, during the frequency division process, high- and low-frequency components were extracted from the PET reconstructed images via Fourier transform, then inversed Fourier transform, denoised separately, and finally combined to produce the final denoised images. The dataset was separated into training (70%), validation (10%) and testing (20%) sets using simple random sampling without replacement with a fixed random seed. A 5-fold cross-validation scheme was then applied to test all 50 patients. Performance was evaluated against full-dose PET using normalized mean square error (NMSE), structural similarity (SSIM), peak signal-to-noise ratio (PSNR), contrast-to-noise ratio (CNR), SUV mean and SUV max bias of selected brain ROIs. Wilcoxon signed rank test was used to analyze the differences between the denoising methods. Results:AttGAN-FS-2 showed the best performance among all dose levels, with statistical difference as compared by low-dose PET and GAN-1 denoised images for NMSE, SSIM, PSNR, and CNR ( Z values: 2.92-6.15, all P<0.005). NMSE, SSIM quantitative evaluation results (median) of each model at 1/20 dose were: GAN-1: 0.08, 0.87, AttGAN-1: 0.08, 0.88, AttGAN-2: 0.07, 0.89, AttGAN-FS-2: 0.06, 0.91, respectively ( Z values: 3.24-5.77, all P<0.005). Conclusion:The DL-based method combined with multiple strategies AttGAN-FS-2 shows improved denoising performance for low-dose brain PET images.

Objectives:To finely divide the aortic sinus into sections and accurately localize the coronary ostium through CARTO three-dimensional mapping,and to assess the clinical effects of treating aortic sinus cusps premature ventricular contraction(ASC-PVC)and the ablation risk in the corresponding area with zero X-ray radiofrequency ablation. Methods:A total of 66 patients who underwent radiofrequency ablation for ASC-PVC from January 2020 to January 2023 were included in this analysis,patients were divided into experimental group(n=34)and conventional group(n=32).In the conventional group,the CARTO 3 system was used to create an aortic sinus model through the conventional method.The earliest stimulating target was identified by using electrical stimulation mapping(ESM).Radiofrequency ablation treatment was performed after the distance between the target and the coronary ostium was precisely measured by coronary angiography through the hollow tube of the ablation catheter or coronary angiography tube.In the experimental group,the CARTO 3 system was used to build a model of the aortic sinus and the coronary ostium and aortic sinus were divided into sections.The earliest stimulating target was identified by ESM.After localizing coronary ostium through the impedance changing pattern on the ablation catheter tips,catheter ablation was performed with zero X-ray.The data regarding the impedance of the ablation catheter in the aortic sinus were collected.The total operative time,the operative time in the aortic sinus,contrast dosage,X-ray exposure time,immediate and short-term success rates of the operation and complication rates were compared between the two groups.Besides,the distribution of successfully ablated targets and their relationship with the risk of ablation were analyzed in both groups. Results:There was no significant statistical difference in the immediate and short-term success rates between the two groups(93.8%vs.94.1%;90.6%vs.88.2%,both P>0.05).The experimental group did not receive contrast agents during the operation,and the total operation time and intra-aortic sinus operation time in the experimental group were significantly shorter than those in the conventional group([58.76±4.94]min vs.[66.91±5.94]min,P<0.001;[43.12±4.49]min vs.[50.31±5.18]min,P<0.001).During the process of moving the ablation catheter from the intra-aortic sinus to the coronary artery opening and into the coronary artery,the impedance suddenly increased,which was significantly different from the impedance in other parts of the intra-aortic sinus(all P<0.001). Conclusions:Radiofrequency ablation of ASC-PVC with zero X-ray can simplify the procedures and shorten the operative time.The steep increase in impedance at the tip of the ablation catheter can be used as a basis for localizing the coronary ostium.Dividing the aortic sinus into sections allows a detailed assessment of the risk for ablation treatment at the targets.

Objective To investigate the effect and mechanism of microRNA-10b(miR-10b)on idiopathic short stature(ISS).Methods A total of 54 children with ISS and 54 healthy children were collected.The serum expression of miR-10b was detected by RT-qPCR,and the relationship between serum miR-10b expression and clinical data of children with ISS was analyzed.miR-10b inhibitor,si-TGFBR1 and their negative control transfection C28/I2 cells were used.CCK-8 experimental detection was used to detect C28/I2 cell proliferation.Western blot assay was used to detect gnome related transcription factor 2(RUNX2),collagen type X alpha 1 chain(COL10A1),transforming growth factor beta receptor 1(TGFBR1),SMAD3 and pSMAD3 protein expression.The target of miR-10b was screened in StarBase database,and the targeting relationship between miR-10b and TGFBR1 was verified by dual luciferase reporter gene assay.Results The serum expression of miR-10b was higher in the ISS group than that of the healthy control group,and the higher the miR-10b expression,the more obvious the decrease of child height,IGF-1 and alkaline phosphatase(P＜0.05).Compared with the NC group,the cell proliferation ability and RUNX2,COL10A1,TGFBR1,and pSMAD3 protein expression were up-regulated in the miR-10b inhibitor group(P＜0.05).StarBase database suggested that miR-10b had a binding site of TGFBR1,and dual luciferase reporter gene assay confirmed that TGFBR1 interacted with miR-10b(P＜0.05).Compared with the si-NC group,the expression of TGFBR1 was down-regulated and the cell proliferation ability was decreased in the si-TGFBR1 group(P＜0.05).Conclusion miR-10b inhibits chondrocyte proliferation and hypertrophy in idiopathic short stature by targeting TGFBR1/SMAD3 pathway.

Early diagnosis of acute rejection is of significance for the protection of renal allograft function. Pathological puncture biopsy is the gold standard for the diagnosis of acute rejection of renal allografts. Nevertheless, it may provoke multiple complications, such as bleeding, infection and renal parenchymal injury, which limit its widespread application. In recent years, the sensitivity of contrast-enhanced ultrasound in the diagnosis of acute rejection has been constantly improved. Ultrasound-targeted microbubble technique has further enhanced the diagnostic specificity of contrast-enhanced ultrasound, making it possible to replace pathological puncture biopsy. Besides, in the field of acute rejection treatment, microbubble ultrasonic cavitation may promote local delivery of immunosuppressants by inducing sonoporation and exhibit anti-rejection effect. In this article, the application of contrast-enhanced ultrasound in the diagnosis and treatment of acute rejection after kidney transplantation was reviewed, aiming to provide reference for widespread application of contrast-enhanced ultrasound in kidney transplantation.

Patients infected with human immunodeficiency virus (HIV) are prone to renal insufficiency, which may progress into end-stage renal disease (ESRD). HIV infection has been previously considered as an absolute contraindication of transplantation. The lives of HIV-positive ESRD patients can only maintained through dialysis. With the development of medicine, transplantation practitioners at home and abroad attempt to perform kidney transplantation in HIV-positive patients. Whether kidney transplantation is feasible for HIV-positive patients, whether HIV-positive donor kidneys can be used for transplantation, and the efficacy of kidney transplantation for HIV-positive patients has always been hot topics in the field of transplantation. In this article, HIV-associated nephropathy, the conditions of kidney transplantation for HIV-positive patients, the efficacy of kidney transplantation for HIV-positive patients, use of HIV-positive donor kidneys, use of immune-inducing drugs and postoperative use of immunosuppressants for HIV-positive patients were illustrated, aiming to provide reference for kidney transplantation for HIV-positive patients in clinical practice, application of HIV-positive donor kidneys and postoperative management of HIV-positive patients.

Objective:To analyze the clinical characteristics of children diagnosed with systemic lupus erythematosus(SLE)complicated with thrombotic microangiopathy(TMA)for early recognition.Methods:We retrospectively reviewed the clinical records of 14 SLE patients with TMA hospitalized at Shanghai Children′s Medical Center, Shanghai Jiaotong University School of Medicine from December 2005 to October 2020.Results:The incidence of TMA was 5.65%(14/248)of the hospitalized patients with SLE and 7.87%(14/178)of the hospitalized patients with lupus nephritis.Four patients were boys while ten patients were girls.One boy was six years old and other 13 patients were from 11 to 18 years old.Their SLEDAI scores ranged from 14 to 31, and all of them were severe activity.Renal biopsy of 11 patients during TMA course all revealed lupus nephritis(type Ⅳ, n=8; type Ⅳ+ Ⅴ, n=3). These 14 SLE children were diagnosed with TMA within 3 days to 2 months after admission.At the beginning of the hospitalization, only six patients had both anemia and thrombocytopenia, while eight patients only had moderate anemia.All of the patients had obvious hypocomplementemia.Especially in the patients with first onset of SLE without treatment, their serum levels of C3 were less than 0.17 g/L and C4 were less than 0.07 g/L.Moreover, glomerular filtration rates of these patients were lower than that in normal range.The follow-up time were 0.2-11.3 years(median time was 2.6 years). After treatment, six patients obtained complete remission, and five patients obtained partial remission.One patient had sudden death during the 4th plasmapheresis, and the other two patients deteriorated. Conclusion:Children with SLE and TMA are mostly in severe disease activity, and renal pathology is type Ⅳ lupus nephritis.The SLE children with anemia should be paid special attention to the level of serum complement whether they have thrombocytopenia or not.If the level of serum complements decrease obviously, glomerular filtration rates should be monitored closely and schistocytes should be searched repeatedly in the blood smears of the peripheral blood to facilitate the early recognition of TMA.