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 explore the clinical characteristics and genetic etiology of a child with Neurofibromatosis-Noonan syndrome (NFNS). METHODS: A child with NFNS who was treated at the Department of Endocrinology of Hangzhou Children's Hospital in January 2024 was selected as the study subject. Clinical data of the child was collected by retrospective analysis. Peripheral venous blood samples (2 mL each) were collected from the child and his parents. Genomic DNA was extracted, and trio-whole exome sequencing (Trio-WES) of the family was carried out. Sanger sequencing was used to perform family verification on the candidate variants. The identified variants were classified for pathogenicity according to the Standards and Guidelines for the Interpretation of Sequence Variants established by the American College of Medical Genetics and Genomics (ACMG) (hereafter referred to as the "ACMG guidelines"). This study has been approved by the Medical Ethics Committee of Hangzhou Children's Hospital (Ethics No. 2021-06). RESULTS: The child was a 7-year and 4-month-old male. He has short stature, numerous café-au-lait spots on the neck and trunk, and special facial features such as a full forehead, wide interpupillary distance, a low nasal bridge, and low-set ears. The results of Trio-WES showed that the he had harbored the NF1 gene c.3773G>T (p.W1258L) mutation, which was verified by Sanger sequencing to be de novo in origin. The NF1 gene was associated with NFNS, which has an autosomal dominant inheritance. According to the ACMG guidelines, this variant was judged to be a likely pathogenic variant (PS2+PM2+PP3+PP2). No pathogenic variant in genes associated with Noonan syndrome, such as PTPN11, SOS1, RAF1, RIT1, and KRAS, was found. CONCLUSION The child with NFNS has clinical features such as short stature, special facial features, and café-au-lait spots. The c.3773G>T (p.W1258L) variation in the NF1 gene may be the genetic etiology of the NFNS child in this study. The results of this study has enriched the variation spectrum of the NF1 gene.
Child ; Humans ; Male ; Exome Sequencing ; Mutation ; Neurofibromatosis 1/genetics* ; Neurofibromin 1/genetics* ; Noonan Syndrome/genetics*

AIM:To explore the causal association between the counts of five types of white blood cells—neutrophils,monocytes,eosinophils,baso-phils,and lymphocytes—and venous thromboem-bolism(VTE).METHODS:Mendelian randomization(MR)analysis method was used,with genetic vari-ants associated with the five types of white blood cells as instrumental variables,and venous throm-boembolism occurrence risk as the outcome vari-able,inverse variance-weighted(IVW)method was employed as the primary analysis method,with MR-Egger regression,weighted median(WM),sim-ple model,and weighted mode methods used as supplements,to analyze the causal association be-tween the counts of five types of white blood cells and VTE,followed by reverse MR analysis.RE-SULTS:Neutrophil and lymphocyte counts are caus-ally associated with the risk of VTE.For neutrophil count,the IVW estimate(OR=0.867,95%CI:0.761-0.981,P=0.031),MR-Egger estimate(OR=0.754,95%CI:0.571-0.996,P=0.048),weighted median es-timate(OR=0.846,95%CI:0.729-0.981,P=0.027),and weighted model estimate(OR=0.748,95%CI:0.595-0.942,P=0.014)were calculated.For lympho-cyte count,the IVW estimate(OR=0.838,95%CI:0.741-0.949,P=0.005)and weighted median esti-mate(OR=0.024,95%CI:0.718-0.977,P=0.024)were calculated.Reverse MR analysis showed a causal association between the risk of VTE and neu-trophil count,the IVW estimate(OR=0.989,95%CI:0.980-0.999,P=0.024).CONCLUSION:Neutrophil and lymphocyte counts are related to the risk of VTE,and decrease in neutrophil and lymphocyte numbers may increase the risk of VTE.VTE occur-rence risk is associated with neutrophil count,and reducing the risk of VTE occurrence may increase neutrophil count.Further research is needed to un-derstand the underlying biological mechanisms be-hind this relationship.

Objective:To explore the clinical characteristics and genetic etiology of a child with Neurofibromatosis-Noonan syndrome (NFNS).Methods:A child with NFNS who was treated at the Department of Endocrinology of Hangzhou Children′s Hospital in January 2024 was selected as the study subject. Clinical data of the child was collected by retrospective analysis method. Peripheral venous blood samples (2 mL each) were collected from the child and his parents. Genomic DNA was extracted, and trio whole exome sequencing (Trio-WES) of the family was carried out. Sanger sequencing was used to perform family verification on the candidate variants. The identified variants were classified for pathogenicity according to the Standards and Guidelines for the Interpretation of Sequence Variants established by the American College of Medical Genetics and Genomics (ACMG) (hereafter referred to as the " ACMG guidelines" ). This study has been approved by the Medical Ethics Committee of Hangzhou Children′s Hospital (Ethics No. 2021-06).Results:The child was a 7-year and 4-month-old male. He has short stature, numerous café-au-lait spots on the neck and trunk, and special facial features such as a full forehead, wide interpupillary distance, a low nasal bridge, and low-set ears.The results of Trio-WES showed that he has harbored a NF1 gene c. 3773G>T (p.W1258L) mutation, which was verified by Sanger sequencing to be de novo in origin. The NF1 gene child was associated with NFNS, which was an autosomal dominant inheritance. According to the ACMG guidelines, this variant was judged to be a likely pathogenic variant (PS2+ PM2+ PP3+ PP2). No pathogenic variant in genes associated with Noonan syndrome, such as those in PTPN11, SOS1, RAF1, RIT1, and KRAS, was found. Conclusion:The child with NFNS has clinical features such as short stature, special facial features, and café-au-lait spots. The c. 3773G>T (p.W1258L) variation in the NF1 gene may be the genetic etiology of the NFNS child in this study. The results of this study has enriched the variation spectrum of the NF1 gene.

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.

AIM:To explore the causal association between the counts of five types of white blood cells—neutrophils,monocytes,eosinophils,baso-phils,and lymphocytes—and venous thromboem-bolism(VTE).METHODS:Mendelian randomization(MR)analysis method was used,with genetic vari-ants associated with the five types of white blood cells as instrumental variables,and venous throm-boembolism occurrence risk as the outcome vari-able,inverse variance-weighted(IVW)method was employed as the primary analysis method,with MR-Egger regression,weighted median(WM),sim-ple model,and weighted mode methods used as supplements,to analyze the causal association be-tween the counts of five types of white blood cells and VTE,followed by reverse MR analysis.RE-SULTS:Neutrophil and lymphocyte counts are caus-ally associated with the risk of VTE.For neutrophil count,the IVW estimate(OR=0.867,95%CI:0.761-0.981,P=0.031),MR-Egger estimate(OR=0.754,95%CI:0.571-0.996,P=0.048),weighted median es-timate(OR=0.846,95%CI:0.729-0.981,P=0.027),and weighted model estimate(OR=0.748,95%CI:0.595-0.942,P=0.014)were calculated.For lympho-cyte count,the IVW estimate(OR=0.838,95%CI:0.741-0.949,P=0.005)and weighted median esti-mate(OR=0.024,95%CI:0.718-0.977,P=0.024)were calculated.Reverse MR analysis showed a causal association between the risk of VTE and neu-trophil count,the IVW estimate(OR=0.989,95%CI:0.980-0.999,P=0.024).CONCLUSION:Neutrophil and lymphocyte counts are related to the risk of VTE,and decrease in neutrophil and lymphocyte numbers may increase the risk of VTE.VTE occur-rence risk is associated with neutrophil count,and reducing the risk of VTE occurrence may increase neutrophil count.Further research is needed to un-derstand the underlying biological mechanisms be-hind this relationship.

Objective:To explore the clinical characteristics and genetic etiology of a child with Neurofibromatosis-Noonan syndrome (NFNS).Methods:A child with NFNS who was treated at the Department of Endocrinology of Hangzhou Children′s Hospital in January 2024 was selected as the study subject. Clinical data of the child was collected by retrospective analysis method. Peripheral venous blood samples (2 mL each) were collected from the child and his parents. Genomic DNA was extracted, and trio whole exome sequencing (Trio-WES) of the family was carried out. Sanger sequencing was used to perform family verification on the candidate variants. The identified variants were classified for pathogenicity according to the Standards and Guidelines for the Interpretation of Sequence Variants established by the American College of Medical Genetics and Genomics (ACMG) (hereafter referred to as the " ACMG guidelines" ). This study has been approved by the Medical Ethics Committee of Hangzhou Children′s Hospital (Ethics No. 2021-06).Results:The child was a 7-year and 4-month-old male. He has short stature, numerous café-au-lait spots on the neck and trunk, and special facial features such as a full forehead, wide interpupillary distance, a low nasal bridge, and low-set ears.The results of Trio-WES showed that he has harbored a NF1 gene c. 3773G>T (p.W1258L) mutation, which was verified by Sanger sequencing to be de novo in origin. The NF1 gene child was associated with NFNS, which was an autosomal dominant inheritance. According to the ACMG guidelines, this variant was judged to be a likely pathogenic variant (PS2+ PM2+ PP3+ PP2). No pathogenic variant in genes associated with Noonan syndrome, such as those in PTPN11, SOS1, RAF1, RIT1, and KRAS, was found. Conclusion:The child with NFNS has clinical features such as short stature, special facial features, and café-au-lait spots. The c. 3773G>T (p.W1258L) variation in the NF1 gene may be the genetic etiology of the NFNS child in this study. The results of this study has enriched the variation spectrum of the NF1 gene.

Objective To explore the application value of deep learning image reconstruction(DLIR)with coronary computed tomography angiography(CCTA)using 100 kV.Methods Sixty patients who underwent CCTA were selected.The tube voltage of 100 kV,noise index of 24 were applied.The 60％adaptive statistical iterative reconstruction-Veo(ASIR-V)and DLIR-low(DLIR-L),DLIR-medium(DLIR-M)and DLIR-high(DL1R-H)were reconstructed.The CT values and standard deviation(SD)values of the aortic root,left main artery,left anterior descending artery,left circumflex artery,right coronary artery and pericardial fat of the four groups of images were measured,and the signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)were calculated.Two radiologists with five-year working experience subjectively evaluated the image quality using a five-point method double-blindly.Results The differences in noise(SD values),SNR values and CNR values among the four groups of images were statistically significant(P＜0.001).As the reconstruction gradually changed of 60％ASIR-V,DLIR-L,DLIR-M and DLIR-H,the coronary SD values gradually decreased,while the SNR values and CNR values gradually increased,among which the DLIR-H had the lowest SD values and the highest SNR values and CNR values.The subjective scores of the four groups of images by two radiologists showed good consistency(Kappa value=0.929,P＜0.001),and the subjective scores were all above 3 points which met the clinical diagnosis criteria.The subjective scores of DLIR-L,DLIR-M and DLIR-H were significantly higher than those of 60％ASIR-V(P＜0.001),with the DLIR-H achieving the highest subjective score.Conclusion DLIR can significantly reduce image SD value and improve image quality of CCTA with 100 kV,among which DLIR-H has the best effect on improving CCTA image quality.

Objective To investigate the difference of facial soft tissue changes in patients with different vertical bone facial types after orthodontic treatment.Methods A total of 137 female patients with class Ⅱ malocclusion aged 18 to 30 years old were selected for retrospective analysis using facial soft tissue 3D model data.According to the mandibular plane angle(FH-MP)angle,they were divided into high angle group,average angle group and low angle group.The EinScan Pro 2X 2020 handheld high-precision 3D scanner was used to capture facial soft tissue images of patients before treatment(T0)and at 6 months during treatment(T1)and after treatment(T2).The patients'facial images were overlapped using reverse engineering software Geomagic Wrap 2021,and the differ-ences within and between groups were statistically analyzed using SPSS 26.0 statistical software.Results Before and after orthodontic treatment,the average overall facial changes in the high angle group were(-3.25±0.22)mm,in the average angle group was(-3.28±0.30)mm,and in the average low angle group was(-3.69±0.36)mm.Compared with the other two groups,the changes in the low angle group decreased more,and the difference was statistically significant(P＜0.05).The mandibular angle area and temporal area decreased the most in the low angle group,which were(-2.78±0.18)mm and(-2.27±0.35)mm,respectively,and the differ-ence was statistically significant compared with the other two groups(P＜0.05),while there was no statistically significant difference among the other groups(P＞0.05).Conclusion The whole face and all facial regions of the three groups had some negative changes,but the collapse in the mandibular angle area and the temporal muscle ar-ea of the low angle group was more obvious than that of the other two groups.

Objective:To evaluate the accuracy of glomerular filtration rate (GFR) assessed from the renal dynamic imaging method (Gates method) with 99Tc m-diethylene triamine pentoacetic acid (DTPA) in the heart transplant population. Methods:From September 2017 to June 2018, 34 patients with advanced heart failure who were prepared for surgery (30 males, 4 females; age: (45±14) years; heart transplant group) and 41 patients with normal heart function (19 males, 22 females; age: (50±17) years; control group) in Fuwai Hospital were respectively enrolled. GFRs of all patients were measured using Gates method (gGFR) and dual plasma sample method (DPSM; dGFR) with 99Tc m-DTPA. The accuracy of Gates method for detecting GRF was verified by using DPSM as the reference. Seventeen patients in heart transplant group underwent 99Tc m-DTPA renal dynamic imaging for Gates and DPSM results repeatedly after the surgery. The single kidney (left and right) functions (dGFRL and dGFRR) of DPSM were obtained according to the results of Gates method. Pearson correlation analysis and paired t test were used to analyze the data. Results:The gGFR in heart transplant group was higher than dGFR ((66.49±15.66) vs (49.16±13.24) ml·min -1·1.73 m -2; t=6.728, P<0.01), and there was a moderate correlation between them ( r=0.467, P<0.01). No difference between gGFR and dGFR in control group was observed ((65.35±26.28) vs (62.22±21.37) ml·min -1·1.73 m -2; t=1.268, P=0.212), and there was a good correlation between them ( r=0.799, P<0.01). The difference between 2 correlation coefficients was statistically significant ( z=-2.44, P<0.05). Serum creatinine decreased, while dGFR, dGFRL and dGFRR increased after the surgery, suggesting the improved renal function. Conclusions:The renal dynamic imaging method (Gates method) with 99Tc m-DTPA has less accuracy in the heart transplant patients. Combination of DPSM and Gates method can provide the precise total GFR and assess single kidney GFR, and may serve as a tool to monitor the renal function for the heart transplant patients in clinic.