Objective:To develop PixelSIFT-UNet, a novel semantic segmentation model that integrates deep learning with scale-invariant feature transform (SIFT) algorithm to improve the segmentation accuracy of oral mucosal lesions.Methods:This investigation utilized 838 standard clinical white light images of oral mucosal diseases acquired from January 2020 to December 2022 at the Stomatology Hospital Zhejiang University School of Medicine. Randomization was achieved through Python′s random.seed function implementation. The random sample function was subsequently applied for sampling distribution. The dataset was stratified into three subsets with a 6∶2∶2 ratio: training ( n=506), validation ( n=166), and testing ( n=166). Lesion boundaries were annotated using Labelme software, and a PixelSIFT-UNet-based deep learning model was developed with VGG-16 and ResNet-50 backbone networks. Model parameters were optimized using the validation set, and performance metrics [including Dice coefficient, mean intersection over union (mIoU), mean pixel accuracy (mPA), and Precision] were assessed on the test set. The model′s performance was benchmarked against conventional semantic segmentation frameworks (U-Net and PSPNet). Results:The developed PixelSIFT-UNet model could achieve precise segmentation of three common oral mucosal lesions: oral lichen planus, oral leukoplakia, and oral submucous fibrosis. Utilizing VGG-16 as the backbone network, the model achieved Dice coefficient, mIoU, mPA, and Precision values of 0.642, 0.699, 0.836, and 0.792, respectively. Implementation with ResNet-50 backbone network yielded metrics of 0.668, 0.733, 0.872 and 0.817, demonstrating significant improvements across all performance indicators compared to conventional U-Net model (relevant metrics: 0.662, 0.717, 0.861 and 0.809) and PSPNet model (relevant metrics: 0.671, 0.721, 0.858 and 0.813).Conclusions:The proposed PixelSIFT-UNet architecture demonstrates superior performance in oral mucosal lesion segmentation tasks, surpassing conventional semantic segmentation models and providing robust quantitative improvements in segmentation accuracy.

In recent decades,the incidence of human papillomavirus(HPV)-associated oropharyngeal squamous cell carcinoma(OPSCC)has shown a marked increase.Significant changes have also occurred in the OPSCC diagnosis and treatment paradigm.Deter-mining HPV status prior to treatment is now essential,and radiotherapy/chemotherapy,immunotherapy,and minimally invasive surgical techniques have progressively emerged as key modalities for managing OPSCC.However,alongside these paradigm shifts,a comprehen-sive technical consensus guiding the entire diagnostic and therapeutic process for OPSCC patients is currently lacking.Given China's large population base and the rising incidence of OPSCC,an expert panel convened to develop a clinical technical consensus on OPSCC diagno-sis and management tailored to China's specific context.This consensus aims to further enhance and standardize understanding of OPSCC management techniques among relevant healthcare professionals.

Objective:To develop PixelSIFT-UNet, a novel semantic segmentation model that integrates deep learning with scale-invariant feature transform (SIFT) algorithm to improve the segmentation accuracy of oral mucosal lesions.Methods:This investigation utilized 838 standard clinical white light images of oral mucosal diseases acquired from January 2020 to December 2022 at the Stomatology Hospital Zhejiang University School of Medicine. Randomization was achieved through Python′s random.seed function implementation. The random sample function was subsequently applied for sampling distribution. The dataset was stratified into three subsets with a 6∶2∶2 ratio: training ( n=506), validation ( n=166), and testing ( n=166). Lesion boundaries were annotated using Labelme software, and a PixelSIFT-UNet-based deep learning model was developed with VGG-16 and ResNet-50 backbone networks. Model parameters were optimized using the validation set, and performance metrics [including Dice coefficient, mean intersection over union (mIoU), mean pixel accuracy (mPA), and Precision] were assessed on the test set. The model′s performance was benchmarked against conventional semantic segmentation frameworks (U-Net and PSPNet). Results:The developed PixelSIFT-UNet model could achieve precise segmentation of three common oral mucosal lesions: oral lichen planus, oral leukoplakia, and oral submucous fibrosis. Utilizing VGG-16 as the backbone network, the model achieved Dice coefficient, mIoU, mPA, and Precision values of 0.642, 0.699, 0.836, and 0.792, respectively. Implementation with ResNet-50 backbone network yielded metrics of 0.668, 0.733, 0.872 and 0.817, demonstrating significant improvements across all performance indicators compared to conventional U-Net model (relevant metrics: 0.662, 0.717, 0.861 and 0.809) and PSPNet model (relevant metrics: 0.671, 0.721, 0.858 and 0.813).Conclusions:The proposed PixelSIFT-UNet architecture demonstrates superior performance in oral mucosal lesion segmentation tasks, surpassing conventional semantic segmentation models and providing robust quantitative improvements in segmentation accuracy.

In recent decades,the incidence of human papillomavirus(HPV)-associated oropharyngeal squamous cell carcinoma(OPSCC)has shown a marked increase.Significant changes have also occurred in the OPSCC diagnosis and treatment paradigm.Deter-mining HPV status prior to treatment is now essential,and radiotherapy/chemotherapy,immunotherapy,and minimally invasive surgical techniques have progressively emerged as key modalities for managing OPSCC.However,alongside these paradigm shifts,a comprehen-sive technical consensus guiding the entire diagnostic and therapeutic process for OPSCC patients is currently lacking.Given China's large population base and the rising incidence of OPSCC,an expert panel convened to develop a clinical technical consensus on OPSCC diagno-sis and management tailored to China's specific context.This consensus aims to further enhance and standardize understanding of OPSCC management techniques among relevant healthcare professionals.

Objective:To establish and evaluate a rapid nucleic acid detection method for SARS-CoV-2 based on COYOTE ? Flash20 real-time fluorescent quantitative PCR instrument. Methods:A rapid reaction system was constructed by using specific primer and probe sets targeting ORF1ab and N gene of SARS-CoV-2, and the sensitivity and specificity of the system were verified. At the same time, 108 clinical samples of COVID-19 were used to evaluate the application of this method.Results:The detection method did not require nucleic acid extraction, and the manual operation time was only one minute. After the sample was sent to the system, the test could be completed in 30 minutes. The detection limit of this method was 4×10 2 copies/ml. It had no cross-reactivity with other human coronaviruses (including HCoV-229E, HCoV-NL63, HCoV-OC43, HCoV-HKU1, SARS-CoV and MERS-CoV) and other respiratory viruses. The evaluation of clinical sample application showed that the total coincidence rate with the conventional RT-qPCR which required nucleic acid extraction was 98.15%. Conclusions:Through the application evaluation of the rapid fluorescent quantitative PCR method of SARS-CoV-2, it was found that the method was simple, fast, specific and sensitive, and it was suitable for real-time and rapid detection needs in varieties of situations.

Objective To investigate the relationship between Epstein-Barr virus (EBV) DNA and EBV serology markers and evaluate the clinical application values in different diseases.Methods Plasma samples from 397 diagnosed EBV infection-associated patients and 120 health donors from May 2014 to November 2015 in Wuhan Tongji Hospital were collected.Real-time fluorescent quantitative PCR was performed to detect the levels of EBV-DNA in peripheral blood mononuclear cell and plasma.ELISA was used to detect VCA IgA,VCA IgM,VCA IgG,EA(D) IgG and EBNA IgG antibodies in plasma.The positive rate of EBV-DNA and EBV antibodies were counted in each group according to the detection threshold.Kappa statistic and Spearman sank correlation test were used to analysis the correlation and uniformity between EBV-DNA and EBV serology indicators.Results The positive rate of VCA IgG in patient and health control was 94.2% (374/397) and 93.3% (112/120) respectively (χ2=0.125,P=0.67);The positive rate of EBNA IgG in patient and health control was 95.4% (379/397) and 95.0% (114/120) respectively (χ2=0.045,P=0.807);but the positive rate of VCA IgM was 5.5% (22/397) and 0% (0/120) respectively (χ2=6.9,P<0.01);The positive rate of VCA IgA was 43.3% (172/397) and 9.2% (10/120) respectively (χ2=49.5,P<0.01);The positive rate of EA(D) IgG was 42.0% (167/397) and 7.5% (9/120) respectively (χ2=49,P<0.01).The positive rate of EBV-DNA was 65.5% (260/397) and 16.7% (20/120) respectively (χ2=88.5,P<0.01);The positive rate of EBV-DNA in plasma was 45.8% (182/397) and 5.0% (6/120) respectively (χ2=66.4,P<0.01).Furthermore,the uniformity and Spearman correlation analysis showed that there was no significant correlation between EBV-DNA and EBV serology indicators.The correlation analysis between PBMC EBV-DNA and VCA IgM,VCA IgA,EA(D) IgG showed the Kappa was 0.073,0.147,0.073,respectively;the correlation analysis between plasma EBV-DNA and VCA IgM,VCA IgA,EA(D) IgG showed the Kappa was 0.144,0.369,0.288,respectively.Thus,the patients were divided into different groups according to the discharge diagnosis,it was observed that the positive rates of EBV-DNA is more than 90% in extra-nodal NK/T cells lymphoma,EBV-associated hemophagocytic lymphoid tissue hyperplasia,chronic active EBV infection and infectious mononucleosis.In nasopharyngeal carcinoma patients,the positive rate of EBV antibodies (VCA IgA and EA(D) IgG) were higher than the detection of EBV-DNA.Conclusions There was no significant correlation between EBV-DNA and EBV serology markers for the same sample.The clinical application values of EBV DNA and EBV serology markers were not identical in nasopharyngeal carcinoma,extra-nodal NK/T cells lymphoma,infectious mononucleosis and EBV-associated hemophagocytic lymphoid tissue hyperplasia.

Objective To inspect the correlation of synthetic effect criteria to sensitivity,reliability of late stage lung cancer curative effect verdict and median survival time. Method Using size change of tumor,symptom change related to tumor,Karnofsky score,body weight etc.as synthetic effect criteria and nstituting clinical synthetic effect criteria,to assess the curative effect of late stage lung cancer patients who received different treatment in department of tumor of Xiyuan Hospital between March 1997 and March 2000, to analysis medium survival time, and being compared with RR of tumor.Result 52 patients entered this experiment, in which 42 patients were assessable (follow up), 13 patients were effective(30.95％ ),17 were patients stable(40.48％ ),12 patients deteriorated (28.57％ ).The result of tumor objective remission rate: CR is 0,PR is 16.67％ (7 patients),NC is 64.29％ (27 patients) and PD is 19.05％ (8 patients).There was apparent difference between two effect criteria(P< 0.05).The mean survival time of this group is (9.3± 1.6)months,median survival time is 10.5 months(2～ 8months),survival rate of one year is 33.33％ (14 patients).In the patients who live longer than median survival time, most of them survive associated with tumor existence who are in NC of tumor objective remission rate, but also in the effective column of synthetic effect criteria. Compared with tumor objective remission rate,the synthetic effect criteria have more apparent correlation to survival time. Conclusion In the assessment of late stage lung cancer,synthetic effect criteria has better sensitivity and reliability,and emphasize survival quality,at same time it reflect the reaction of tumor and host to treatment,it reflect prognosis of tumor preferably and have better correlation to the survival time. It can be used as clinical effect criteria of late stage lung cancer after further consummation.