Objective:To develop and validate a cascaded deep learning algorithm for kidney stone detection on plain CT images.Methods:Plain CT images of the patients with kidney stones were retrospectively archived from January 2018 to July 2018 in Peking University First Hospital. The cases were divided into two datasets according to the date of the CT scanning: training dataset （ n=30) and held-out test dataset （ n=29). The development of the kidney stone detection method consisted of three steps. First, a U-Net model was trained on the training dataset for kidney segmentation, and the model′s performance was estimated with the dice coefficient. Second, the thresholding and region growing methods were used to detect the stones in the renal region predicted by the trained U-Net model. Third, the stones′ lengths （maximal, middle and minimal length) and CT attenuation were calculated and integrated into a structured report automatically. Using the radiologist′s labels and measurements (maximal, middle, minimal length and CT attenuation) as ground truth, the stone detection algorithm performance was evaluated with sensitivity, specificity and precision, and the stone measurement algorithm performance was evaluated with Bland-Altman plots. Results:The held-out test dataset consisted of 29 cases, containing 58 kidneys and 11 358 CT slices. The 38 kidneys containing 56 stones and 20 kidneys did not contain stones. The U-Net model showed good performance, with a mean dice coefficient of 0.96. And 10 945 of 11 358 CT slices had a dice coefficient no less than 0.90. The sensitivity, precision, and specificity of stone detection were 100% （38/38), 100% （38/38) and 100% （20/20) in the organ-level. The sensitivity and precision of stone detection were 100% （56/56) and 96.6% （56/58) in the lesion-level.Conclusion:A cascaded algorithm is constructed and can be used to detect kidney stones in plain CT images. The algorithm can improve efficiency with results automatically integrated into the structured report in clinical practice.

Objective To construct a 3D ResNet deep learning model based on multi-parametric prostate MRI(mpMRI),and to observe its value for automatically identifying the main MR sequences.Methods Totally 1 153 sets pre-biopsy prostate mpMRI data of 1 086 patients who underwent ultrasound-guided prostate biopsy in 3 hospitals were collected and divided into different image datasets,i.e.T2WI,diffusion-weighted imaging(DWI)and apparent diffusion coefficient(ADC)maps with a total of 5 151 images.Then the images were categorized into non-fat-suppressed T2WI(T2WI_nan,n=1 000),fat-suppressed T2WI(T2WI_fs,n=1 188),high b-value DWI(DWI_High,b-value≥500 s/mm2,n=1 045),low b-value DWI(DWI_Low,b-value＜500 s/mm2,n=1 012)or ADC map(n=906),also divided into training set(n=4 122),verification set(n=513)and test set(n=516)at the ratio of 8∶1∶1.After preprocessing and augmentation,a 3D ResNet model for automatically identifying image categories was trained and optimized in the training and verification sets,and its classification efficiency was evaluated in the test set.Results The identifying accuracy,sensitivity,specificity,positive predictive value,negative predictive value,F1 score and Kappa value of the obtained model for automatically identifying categories of images in the test set was 0.995-1.000,0.990-1.000,0.998-1.000,0.990-1.000,0.998-1.000,0.995-1.000 and 0.994-1.000,respectively.Conclusion The obtained 3D ResNet deep learning model could effectively and automatically identify the main sequences of prostate mpMRI.

Objective:To investigate the characteristics of nasal flora and the pathogenic role of differential microbiome in patients with allergic rhinitis (AR) and non-allergic rhinitis (nAR).Methods:Thirty-five patients with AR who attended the rhinology outpatient clinic of the Second Hospital of Harbin Medical University from February to July 2022 were selected. A total of 35 nAR patients were selected as the test group, and 20 cases of healthy people with physical examination at the same period were selected as the control group, including 39 males and 51 females, aged 8 to 55 years. 16SrDNA High-throughput sequencing was used to analyze the relative abundance from nasal flora in the three groups of subjects. Alpha diversity index analysis was conducted with R software, and differences between groups were analyzed with LEfSe, Metastats, and t tests. At the same time, the role of microbiome and its relationship with environmental factors were analyzed with R software. Results:There was a significant difference in the bacterial composition of the samples from the three groups, with the relative abundance of Staphylococcus aureus ( P=0.032) and Corynebacterium proinquum ( P=0.032) within the AR group being significantly higher than that of the nAR group, and that of Lactobacillus murinus, Lactobacillus kunkeei, and Alcaligenes faecalis ( P value was 0.016, 0.005, and 0.001, respectively) being significantly lower than that of the nAR group. The relative abundance of Ackermannia muciniphila within the nAR group was higher than that of the control group ( P=0.009). Correlation analysis of environmental factors showed a negative correlation between Lactobacillus kunkeei and IgE ( P=0.044), and a positive correlation between Lactobacillus murinus and age ( P=0.019). AR and nAR random forest prediction models were constructed for the five genera, respectively, and the area under the curve (AUC) of the models of Streptococcus-SP-FF10, Pseudoalteromonas luteoviolacea, Pseudomonas parafulva, Acinetobacter ursingii, and Azotobacter chroococcum in the AR group was 100% (95%CI: 100% to 100%). The AUC for the Pseudomonas parafulva, Azotobacter chroococcum, Closoridium baratii, Turicibacter-SP-H121, and Streptococcus lutetiensis models in the nAR group was 98.4% (95%CI: 94.9% to 100%). Conclusions:The distribution of nasal flora in AR patients, nAR patients and healthy subjects is significantly different, and the changes of bacterial flora abundance are significantly related to the occurrence of AR and nAR. Combined detection of microbiota has the potential to diagnose AR and nAR patients.

Objective To develop a 3D V-Net deep learning segmentation model based on multi-center data,and to evaluate its value for automatic segmentation of prostate on T2WI and apparent diffusion coefficient(ADC)maps.Methods Totally 2 894 sets of multi-parametric MRI data of 2 673 patients with clinically suspected prostate cancer from 3 medical centers within 1 month before biopsy were retrospectively collected.Finally 5 974 sets axial images were enrolled,including 3 654 sets of T2WI and 2 320 sets of ADC maps.Prostate contours were manually annotated layer by layer on axial T2WI and ADC maps,and the left-to-right,anterior-to-posterior,superior-to-inferior diameters and volume of prostate were measured and taken as reference standards.The images were divided into training set(n=4 780,including 2 907 sets of T2WI and 1 873 sets of ADC map),verification set(n=601,including 384 sets of T2WI and 217 sets of ADC map)and test set(n=593,including 363 sets of T2WI and 230 sets of ADC map)at the ratio of 8:1:1.After preprocessing and augmentation,3D V-Net was used to construct and train the segmentation model based on training and verification sets,and the segmentation performance of the model was evaluated in test set using Dice similarity coefficient(DSC),Jaccard coefficient(JACARD)and volume similarity(VS),respectively.The parameters measured with the model were compared with the reference standards,and the correlations were explored.Results Compared with the corresponding ADC maps,DSC,JACARD and VS of the model for automatic segmentation of prostate on T2WI in test set were all higher(all P＜0.001).The left-to-right,anterior-to-posterior and superior-to-inferior diameters of prostate measured with the model on both T2WI and ADC maps were all larger than the reference standards(all P＜0.001),while no significant difference of the volume was found(both P＞0.05).All parameters measured with the model on T2WI and ADC maps were positively correlated with reference standards(rs=0.794-0.985).Conclusion 3D V-Net deep learning model could automatically segment prostate on T2WI and ADC maps with high accuracy,and its efficiency based on T2WI was better than that based on ADC maps.

OBJECTIVES: Pelvic lymph node metastasis (PLNM) is an important factor that affects the stage and prognosis of prostate cancer. Invasive extended pelvic lymph node dissection (ePLND) is the most effective method for clinically diagnosing PLNM. Accurate preoperative prediction of PLNM can reduce unnecessary ePLND. This study aims to investigate the clinical value of radiomics nomogram in predicting PLNM of prostate cancer based on T₂-weighted imaging (T₂WI). METHODS: Magnetic resonance (MR) data of 71 patients with prostate cancer who underwent ePLND from January 2017 to June 2021 in Peking University First Hospital were collected retrospectively. All patients were assigned into a training set (January 2017 to December 2020, n=56, containing 186 lymph nodes) and a test set (January 2021 to June 2021, n=15, containing 45 lymph nodes) according to the examination time of multiparametric magnetic resonance imaging (mpMRI). Two radiologists matched the dissected lymph nodes on MRI images, and manually annotated the region of interest (ROI). Based on the outlined ROI, 3 metastatic lymph node prediction models were established: Model 1 (only image features of T₂WI), Model 2 (radiomics features based on random forest), and Model 3 (combination of the image and radiomics features). A nomogram was also established. The clinicopathologic characteristics of the patients were obtained from the medical records, including age, the Gleason score, the level of prostate-specific antigen (PSA), and clinical and pathological T stage. The preoperative radiological features of the pelvic lymph nodes (LNs) include size of LNs (the short and long diameters) and volume of LNs. Receiver operating characteristic (ROC) curve was used to evaluate the diagnostic efficacy of the 3 models and decision curve analysis (DCA) was used to evaluate the clinical benefits of the models. RESULTS: No significant differences were found between the training set and test set regarding age, Gleason scores, PSA level, and clinical and pathological T stage (all P>0.05). The differences in volume, short diameter and long diameter between metastatic and non-metastatic LNs were statistically significant in both training set and test set (all P<0.05). In multivariate regression analysis, the short diameter and marginal status of LNs were included in Model 1. Eighteen omics features were selected to construct Model 2. The signal distribution of LNs and Rad score were the significant risk factors for predicting metastasis of pelvic LNs in Model 3. The C-index of nomogram based on Model 3 reached 0.964, and the calibration curve showed that the model had high calibration degree. In the test set, the area under the curves of Model 1, 2, and 3 were 0.78, 0.93, and 0.96 respectively, Model 2 and Model 3 showed significantly higher diagnostic efficiency than Model 1 (Model 1 vs Model 2, P=0.019; Model 1 vs Model 3, P=0.020). There was no significant difference in the area under the curve between Model 2 and Model 3 (P=0.649). The DCA results of the 3 models showed that all models obtained higher net benefits than the PLNM-all or PLNM-none protocol in different ranges of threshold probabilities and Model 3 had the highest clinical benefit. CONCLUSIONS The radiomics nomogram based on T₂WI shows a good predictive efficacy for preoperative PLNM in patients with prostate cancer, which could be served as an imaging biomarker to optimize decision-making and adjust adjuvant treatments.
Humans ; Lymph Nodes/pathology* ; Lymphatic Metastasis ; Male ; Prostate-Specific Antigen ; Prostatic Neoplasms/pathology* ; Retrospective Studies

Objective:To investigate the epidemiological characteristics, diagnosis, treat-ment and prognosis of gallbladder cancer in China from 2010 to 2017.Methods:The single disease retrospective registration cohort study was conducted. Based on the concept of the real world study, the clinicopathological data, from multicenter retrospective clinical data database of gallbladder cancer of Chinese Research Group of Gallbladder Cancer (CRGGC), of 6 159 patients with gallbladder cancer who were admitted to 42 hospitals from January 2010 to December 2017 were collected. Observation indicators: (1) case resources; (2) age and sex distribution; (3) diagnosis; (4) surgical treatment and prognosis; (5) multimodality therapy and prognosis. The follow-up data of the 42 hospitals were collected and analyzed by the CRGGC. The main outcome indicator was the overall survival time from date of operation for surgical patients or date of diagnosis for non-surgical patients to the end of outcome event or the last follow-up. Measurement data with normal distribu-tion were represented as Mean±SD, and comparison between groups was conducted using the t test. Measurement data with skewed distribution were represented as M( Q1, Q3) or M(range), and com-parison between groups was conducted using the U test. Count data were described as absolute numbers or percentages, and comparison between groups was conducted using the chi-square test. Univariate analysis was performed using the Logistic forced regression model, and variables with P<0.1 in the univariate analysis were included for multivariate analysis. Multivariate analysis was performed using the Logistic stepwise regression model. The life table method was used to calculate survival rates and the Kaplan-Meier method was used to draw survival curves. Log-rank test was used for survival analysis. Results:(1) Case resources: of the 42 hospitals, there were 35 class A of tertiary hospitals and 7 class B of tertiary hospitals, 16 hospitals with high admission of gallbladder cancer and 26 hospitals with low admission of gallbladder cancer, respectively. Geographical distribution of the 42 hospitals: there were 9 hospitals in central China, 5 hospitals in northeast China, 22 hospitals in eastern China and 6 hospitals in western China. Geographical distribution of the 6 159 patients: there were 2 154 cases(34.973%) from central China, 705 cases(11.447%) from northeast China, 1 969 cases(31.969%) from eastern China and 1 331 cases(21.611%) from western China. The total average number of cases undergoing diagnosis and treatment in hospitals of the 6 159 patients was 18.3±4.5 per year, in which the average number of cases undergoing diagnosis and treatment in hospitals of 4 974 patients(80.760%) from hospitals with high admission of gallbladder cancer was 38.8±8.9 per year and the average number of cases undergoing diagnosis and treatment in hospitals of 1 185 patients(19.240%) from hospitals with low admission of gallbladder cancer was 5.7±1.9 per year. (2) Age and sex distribution: the age of 6 159 patients diagnosed as gallbladder cancer was 64(56,71) years, in which the age of 2 247 male patients(36.483%) diagnosed as gallbladder cancer was 64(58,71)years and the age of 3 912 female patients(63.517%) diagnosed as gallbladder cancer was 63(55,71)years. The sex ratio of female to male was 1.74:1. Of 6 159 patients, 3 886 cases(63.095%) were diagnosed as gallbladder cancer at 56 to 75 years old. There was a significant difference on age at diagnosis between male and female patients ( Z=-3.99, P<0.001). (3) Diagnosis: of 6 159 patients, 2 503 cases(40.640%) were initially diagnosed as gallbladder cancer and 3 656 cases(59.360%) were initially diagnosed as non-gallbladder cancer. There were 2 110 patients(34.259%) not undergoing surgical treatment, of which 200 cases(9.479%) were initially diagnosed as gallbladder cancer and 1 910 cases(90.521%) were initially diagnosed as non-gallbladder cancer. There were 4 049 patients(65.741%) undergoing surgical treatment, of which 2 303 cases(56.878%) were initially diagnosed as gallbladder cancer and 1 746 cases(43.122%) were initial diagnosed as non-gallbladder cancer. Of the 1 746 patients who were initially diagnosed as non-gallbladder cancer, there were 774 cases(19.116%) diagnosed as gallbladder cancer during operation and 972 cases(24.006%) diagnosed as gallbladder cancer after operation. Of 6 159 patients, there were 2 521 cases(40.932%), 2 335 cases(37.912%) and 1 114 cases(18.087%) undergoing ultrasound, computed tomography (CT) or magnetic resonance imaging (MRI) examination before initial diagnosis, respec-tively, and there were 3 259 cases(52.914%), 3 172 cases(51.502%) and 4 016 cases(65.205%) undergoing serum carcinoembryonic antigen, CA19-9 or CA125 examination before initially diagnosis, respectively. One patient may underwent multiple examinations. Results of univariate analysis showed that geographical distribution of hospitals (eastern China or western China), age ≥72 years, gallbladder cancer annual admission of hospitals, whether undergoing ultrasound, CT, MRI, serum carcinoembryonic antigen, CA19-9 or CA125 examination before initially diagnosis were related factors influencing initial diagnosis of gallbladder cancer patients ( odds ratio=1.45, 1.98, 0.69, 0.68, 2.43, 0.41, 1.63, 0.41, 0.39, 0.42, 95% confidence interval as 1.21-1.74, 1.64-2.40, 0.59-0.80, 0.60-0.78, 2.19-2.70, 0.37-0.45, 1.43-1.86, 0.37-0.45, 0.35-0.43, 0.38-0.47, P<0.05). Results of multivariate analysis showed that geographical distribution of hospitals (eastern China or western China), sex, age ≥72 years, gallbladder cancer annual admission of hospitals and cases undergoing ultrasound, CT, serum CA19-9 examination before initially diagnosis were indepen-dent influencing factors influencing initial diagnosis of gallbladder cancer patients ( odds ratio=1.36, 1.42, 0.89, 0.67, 1.85, 1.56, 1.57, 0.39, 95% confidence interval as 1.13-1.64, 1.16-1.73, 0.79-0.99, 0.57-0.78, 1.60-2.14, 1.38-1.77, 1.38-1.79, 0.35-0.43, P<0.05). (4) Surgical treatment and prognosis. Of the 4 049 patients undergoing surgical treatment, there were 2 447 cases(60.435%) with complete pathological staging data and follow-up data. Cases with pathological staging as stage 0, stage Ⅰ, stage Ⅱ, stage Ⅲa, stage Ⅲb, stage Ⅳa and stage Ⅳb were 85(3.474%), 201(8.214%), 71(2.902%), 890(36.371%), 382(15.611%), 33(1.348%) and 785(32.080%), respectively. The median follow-up time and median postoperative overall survival time of the 2 447 cases were 55.75 months (95% confidence interval as 52.78-58.35) and 23.46 months (95% confidence interval as 21.23-25.71), respectively. There was a significant difference in the overall survival between cases with pathological staging as stage 0, stage Ⅰ, stage Ⅱ, stage Ⅲa, stage Ⅲb, stage Ⅳa and stage Ⅳb ( χ2=512.47, P<0.001). Of the 4 049 patients undergoing surgical treatment, there were 2 988 cases(73.796%) with resectable tumor, 177 cases(4.371%) with unresectable tumor and 884 cases(21.833%) with tumor unassessable for resectabi-lity. Of the 2 988 cases with resectable tumor, there were 2 036 cases(68.139%) undergoing radical resection, 504 cases(16.867%) undergoing non-radical resection and 448 cases(14.994%) with operation unassessable for curative effect. Of the 2 447 cases with complete pathological staging data and follow-up data who underwent surgical treatment, there were 53 cases(2.166%) with unresectable tumor, 300 cases(12.260%) with resectable tumor and receiving non-radical resection, 1 441 cases(58.888%) with resectable tumor and receiving radical resection, 653 cases(26.686%) with resectable tumor and receiving operation unassessable for curative effect. There were 733 cases not undergoing surgical treatment with complete pathological staging data and follow-up data. There was a significant difference in the overall survival between cases not undergoing surgical treatment, cases undergoing surgical treatment for unresectable tumor, cases undergoing non-radical resection for resectable tumor and cases undergoing radical resection for resectable tumor ( χ2=121.04, P<0.001). (5) Multimodality therapy and prognosis: of 6 159 patients, there were 541 cases(8.784%) under-going postoperative adjuvant chemotherapy and advanced chemotherapy, 76 cases(1.234%) under-going radiotherapy. There were 1 170 advanced gallbladder cancer (pathological staging ≥stage Ⅲa) patients undergoing radical resection, including 126 cases(10.769%) with post-operative adjuvant chemotherapy and 1 044 cases(89.231%) without postoperative adjuvant chemo-therapy. There was no significant difference in the overall survival between cases with post-operative adjuvant chemotherapy and cases without postoperative adjuvant chemotherapy ( χ2=0.23, P=0.629). There were 658 patients with pathological staging as stage Ⅲa who underwent radical resection, including 66 cases(10.030%) with postoperative adjuvant chemotherapy and 592 cases(89.970%) without postoperative adjuvant chemotherapy. There was no significant difference in the overall survival between cases with postoperative adjuvant chemotherapy and cases without postoperative adjuvant chemotherapy ( χ2=0.05, P=0.817). There were 512 patients with pathological staging ≥stage Ⅲb who underwent radical resection, including 60 cases(11.719%) with postoperative adjuvant chemotherapy and 452 cases(88.281%) without postoperative adjuvant chemotherapy. There was no significant difference in the overall survival between cases with postoperative adjuvant chemo-therapy and cases without post-operative adjuvant chemo-therapy ( χ2=1.50, P=0.220). Conclusions:There are more women than men with gallbladder cancer in China and more than half of patients are diagnosed at the age of 56 to 75 years. Cases undergoing ultrasound, CT, serum CA19-9 examination before initial diagnosis are independent influencing factors influencing initial diagnosis of gallbladder cancer patients. Preoperative resectability evaluation can improve the therapy strategy and patient prognosis. Adjuvant chemotherapy for gallbladder cancer is not standardized and in low proportion in China.