Objective:To develop a deep learning model based on super-resolution endorectal ultrasound（ERUS）images for the preoperative prediction of perineural invasion（PNI）in patients with rectal cancer，thereby providing a reference for risk stratification and individualized treatment planning.Methods:A retrospective analysis was conducted on 382 patients with rectal cancer who underwent total mesorectal excision at Fujian Medical University Union Hospital between June 2019 and February 2024. Patients were randomly divided into a training set（ n=305）and a test set（ n=77）at a ratio of 8∶2，and further grouped into PNI-negative group and PNI-positive group subgroups based on pathological results. Super-resolution ultrasound images were generated from original ERUS images using a generative adversarial network（GAN）. Deep convolutional neural networks were developed based on features from intratumoral and peritumoral regions to identify the optimal region of interest（ROI）. The dSR5_ResNet18 and dSR5_ResNet50 models were constructed using the super-resolution images with a 5-pixel peritumoral extension. Representative clinical features were selected for subgroup analysis based on sample size and intergroup statistical differences between PNI-positive and PNI-negative patients. Forest plots were used to evaluate model applicability and robustness across subgroups. Results:The dSR5_ResNet18 model，built using super-resolution images of the tumor combined with a 5-pixel peritumoral region，achieved the best predictive performance，with an AUC of 0.867（95% CI=0.782 - 0.952）in the test set. Decision curve analysis demonstrated that the dSR5_ResNet18 model provided the greatest net clinical benefit. Forest plot analysis indicated strong generalizability of the models across subgroups such as pathological N stage，maximum lesion length，and lymph node enlargement，though relatively weaker performance was observed in the carcinoembryonic antigen（CEA）subgroup. Among all models，dSR5_ResNet18 exhibited the most consistent performance across subgroups，with the narrowest confidence intervals and highest robustness. Conclusions:The deep learning model incorporating ERUS-based super-resolution reconstruction demonstrated excellent performance in the preoperative prediction of PNI in rectal cancer. It offers significant advantages in image quality and generalizability，and may serve as a valuable tool to assist clinicians in formulating personalized treatment strategies.

Objective:To develop a deep learning model based on super-resolution endorectal ultrasound（ERUS）images for the preoperative prediction of perineural invasion（PNI）in patients with rectal cancer，thereby providing a reference for risk stratification and individualized treatment planning.Methods:A retrospective analysis was conducted on 382 patients with rectal cancer who underwent total mesorectal excision at Fujian Medical University Union Hospital between June 2019 and February 2024. Patients were randomly divided into a training set（ n=305）and a test set（ n=77）at a ratio of 8∶2，and further grouped into PNI-negative group and PNI-positive group subgroups based on pathological results. Super-resolution ultrasound images were generated from original ERUS images using a generative adversarial network（GAN）. Deep convolutional neural networks were developed based on features from intratumoral and peritumoral regions to identify the optimal region of interest（ROI）. The dSR5_ResNet18 and dSR5_ResNet50 models were constructed using the super-resolution images with a 5-pixel peritumoral extension. Representative clinical features were selected for subgroup analysis based on sample size and intergroup statistical differences between PNI-positive and PNI-negative patients. Forest plots were used to evaluate model applicability and robustness across subgroups. Results:The dSR5_ResNet18 model，built using super-resolution images of the tumor combined with a 5-pixel peritumoral region，achieved the best predictive performance，with an AUC of 0.867（95% CI=0.782 - 0.952）in the test set. Decision curve analysis demonstrated that the dSR5_ResNet18 model provided the greatest net clinical benefit. Forest plot analysis indicated strong generalizability of the models across subgroups such as pathological N stage，maximum lesion length，and lymph node enlargement，though relatively weaker performance was observed in the carcinoembryonic antigen（CEA）subgroup. Among all models，dSR5_ResNet18 exhibited the most consistent performance across subgroups，with the narrowest confidence intervals and highest robustness. Conclusions:The deep learning model incorporating ERUS-based super-resolution reconstruction demonstrated excellent performance in the preoperative prediction of PNI in rectal cancer. It offers significant advantages in image quality and generalizability，and may serve as a valuable tool to assist clinicians in formulating personalized treatment strategies.

Objective:To investigate the application value of shear wave elastography (SWE) in the evaluation of T re-staging after neoadjuvant chemoradiotherapy (nCRT) for locally advanced rectal cancer.Methods:Clinical, endorectal ultrasound (ERUS) and SWE data of 271 patients with locally advanced rectal cancer who underwent nCRT and total mesorectal excision in Fujian Medical University Union Hospital from October 2021 to March 2023 were prospectively collected. The independent predictors for low T staging were analyzed and screened, and the Logistic regression model was constructed. An independent test set was used to validate the prediction performance of the models and compare them with the diagnostic results of sonographers.Results:Binary multivariate Logistic regression analysis showed that Emean of the mesentery around the lesion, thickness, and enlarged lymph nodes around the rectum were the independent predictors for low T staging, and the odds ratios were 1.089, 1.214, 0.183, respectively. The Logistic regression model A established by Emean, thickness and enlarged lymph nodes around the lesion and the Logistic regression model B established by Emean around the lesion had high diagnostic efficiencies (area under the ROC curve were 0.931, 0.918, respectively, the accuracy were 0.888 and 0.887, respectively). There was no significant difference in diagnostic accuracy between the two models ( P=1.000), and both models were significantly higher than that of sonographers (all P<0.001). Conclusions:SWE can effectively predict whether the tumor is of low T staging after nCRT in locally advanced rectal cancer, and can be used as an important supplement to ERUS in evaluating the T re-staging of rectal cancer after nCRT.

Objective:To construct preliminarily the malignant risk classification system for the cervical lymph node.Methods:A total of 301 patients with cervical lymphadenopathy were collected in this prospective study from Union Hospital, Fujian Medical University from July 2018 to December 2019. The ultrasonographic features(including the short diameter, ratio of long to short diameters(L/S), margin, border, matting, echogenic hilum, echogenicity, gross necrosis, microcalcification, hyperechoic area, flow type, vascular distribution), history of malignancy, inflammation performance of the neck and history of tuberculosis were analyzed. A score was assigned for each significant index related to benign/malignant lymph nodes by a Logistic regression analysis. The classification of the malignant risk was determined on the basis of the scores.Results:The factors significantly associated with the malignant lymph nodes were enlargement of the short diameter, L/S<2, microcalcification, hyperechoic area, irregular margin, matting, abnormal flow pattern, malignancy history. While the factors related to the benign were the inflammation performance and the fuzzy boundary. The risk of malignancy increased as the score of lymph node increased. The malignant risk of lymph node according to the classification system was as follows: category 1, 7.30%; category 2a, 35.00%; category 2b, 69.30%; category 2c, 91.50%; and category 3, 99.05%. The area under the ROC curve of the system was 0.913.Conclusions:The system has great potential of clinical application to assess the risk of malignancy in cervical lymph nodes.

Objective: To investigate the value of transrectal ultrasonography(TRUS) in the preoperative diagnosis and tumor T-staging of anorectal malignant melanoma(ARMM). Methods: The clinical and ultrasound data of 19 patients(22 lesions) with ARMM confirmed by pathology from February 2008 to Apirl 2019 in Fujian Medical University Affiliated Union Hospital were collected. TRUS was performed within one week before pathological examination. The sonographic features were summarized, ultrasonic typing was performed according to the lesion size and growth pattern, and the accuracies of preoperative ultrasound diagnosis and tumor staging were analyzed. Results: Among the 22 lesions, 63.64%(14/22) were with a distance of ≤3 cm between the anal verge and the lower margin of tumor, 68.18%(15/22) with regular shape, 81.82%(18/22) with clear boundary, and 63.64%(14/22) with Adler blood flow grading from Ⅱ to Ⅲ. The preoperative TRUS diagnostic coincidence rate was 73.68%(14/19), and the tumor T-staging accuracy was 75%(12/16), respectively. Conclusions TRUS exhibits certain sonographic characteristics, with high preoperative diagnosis rate and tumor T-staging accuracy in anorectal malignant melanoma, which is of great guiding significance for clinical diagnosis and treatment.

Objective: To explore the sonographic features of primary testicular lymphoma(PLT). Methods: The sonographic features of 16 PLT patients confirmed by surgery and pathology were retrospectively analyzed. Results: Sixteen PLT patients were typed into 3 ultrasonic patterns: ①The diffuse homogeneous echo type: 2 cases, the ultrasonic features were homogeneous sieve-like hypoechoic echo, the blood flow signal was rich and radially distributed, with interstisial fibrosis inapparently under-microscope. ②The diffuse uneven homogeneous echo type: 9 cases, the ultrasonic features lobulated or map-like heterogeneous hypoechoic features, the blood flow signal was rich in linear distribution, and interstisial fibrosis proliferated to form fibrous septa under-microscope. ③The focal nodular pattern echo type: 5 cases, the ultrasonic features were homogeneous hypoechoic nodules, the blood flow signal was few, interstisial fibrosis hyperplasia surrounding the tumor under-microscope. Conclusions Ultrasonic appearances of primary testicular lymphoma are closely related to its pathologic characteristics, the recognization of their relations contributes to improve the diagnostic value of ultrasound.

Objective To explore the relationship between ultrasonic appearances of epididymal spermatic granuloma and its pathologic characteristics. Methods The sonographic features of 32 cases of epididymal spermatic granuloma confirmed by pathology were retrospectively analyzed. Results A total of 34 epididymal spermatic granuloma nodules in 32 cases included 4 located in caput epididymidis and 30 located in cauda epididymidis,and maximum diameter of nodules were about 0.5-2.3 cm.Sonographic and pathologic appearances of nodules were: ① mostly low echo type,21 nodules mostly showed the obscure boundary and quite abundant blood flow signals,early granulomas formed with interstitial fibrosis inapparently under-microscope; ② inhomogeneous echo type,12 nodules mostly showed the barely clear boundary and blood flow signals from small amount to abundant amount,granulomas of different maturites distributed dispersedly with interstitial fibrosis of varying degrees under-microscope or single giant lesions formed;③mostly high echo type,1 nodule showed the obscure boundary and inapparent blood flow signals, center of the nodules distributed by a large amount of dense fibrous tissue under-microscope. Anechoic ethmoid sinus seen in 29 nodules was section of dilated epididymal ductus.Dotted strong echoes seen in 6 nodules were calcification of sperms in dilated epididymal ductus and necrotic tissue. Conclusions Ultrasonic appearances of epididymal spermatic granuloma are closely related to its pathologic characteristics.To recognize both relations contributes to improve the diagnostic value of ultrasound.

Objective To explore novel lipid microbubbles with matrix metalloproteinase-9(TMB-9) targeting in the radiography evaluation of the protective effects of doxycycline on testicular ischemia-reperfusion injury by contrast-enhanced ultrasound(CEUS). Methods Thirty-six rabbits were randomly divided into control groups (group S1 and group S2),ischemia reperfusion groups (group R1 and group R2) and doxy intervention groups (group D1 and group D2).In the control groups,threading without ligation was performed after the exposure of the spermatic cord.The establishment of ischemia-reperfusion model was performed in ischemia reperfusion groups and doxy intervention groups,followed by 1-day feeding for group R1 and D1 as well as 4-day feeding for group R2 and D2.Group D1,group D2 and group S1 were given intraperitoneal injection of doxy 20 mg· kg -1· d-1.Group R1,group R2 and group S2 were given intraperitoneal injection of equivalent normal saline.MB and TMB-9 CEUS were undertook preoperative and postoperative respectively in each group.And the changes of the contrast parameters were analyzed, including PI,TP,Slope,MTT,DT/2 and AUC.The testes were taken for pathological examination after operation.Results TMB-9 contrast parameters of the group D1 including PI and AUC were significantly decreased,MTT and DT/2 decreased( P < 0.05),compared with Group R1.While compared with MB, TMB-9 contrast parameters of group D1 including PI and AUC were increased,MTT and DT/2 prolonged (P<0.05).TMB-9 contrast parameters of Group D2 including PI and AUC were significantly decreased, MTT and DT/2 decreased( P <0.01),compared with Group R2.TMB-9 contrast parameters of Group D2 only PI increased( P <0.01),while compared with MB.Compared with the ischemia reperfusion group,the expression of MMP-9 in the vascular basement membrane of doxy intervention groups was significantly decreased ( P < 0.01).Conclusions TMB-9 can sensitively and objectively evaluate testicular ischemia-reperfusion injury of testicular microvasculature at different reperfusion times before and after doxy intervention.

Objective: To evaluate the value of ultrasound-guided fine needle aspiration(US-FNA) combined with detection of BRAF V600E and thyroid imaging reporting and data system(TI-RADS) in diagnosis of benign and malignant thyroid nodules. Methods: In this study, 123 operative thyroid nodules from 114 patients who underwent US-FNA and detection of BRAF V600E were enrolled. TI-RADS was apply for the classification of each nodule before surgery. Specimens from each nodule were subjected for hematoxylin and eosin (HE) staining and cytological diagnosis and detection of BRAF V600E mutation. Results: ①BRAF V600E mutation was found in 71 (71/123) nodules with histologic confirmation of papillary-thyroid carcinoma, 58 of which were cytologically diagnosed as carcinoma and 13 were indeterminate. Compared with the postoperative pathological results, US-FNA combined with BRAF V600E could improve the sensitivity and accuracy of diagnosis to thyroid nodules compared with individual US-FNA, and the difference was statistically significant(P<0.001). ②The mutation rate of BRAF V600E was associated with thyroid capsular invasion(χ²=8.44, P=0.004), and combined with TI-RADS could indicate the high-risk of this invasion. ③Among 123 operative nodules, 18 nodules were BRAF V600E negative and cytologically diagnosed as indetermination, 10 of which were TI-RADS 3b or above. After thyroidectomy, 6 nodules were confirmed as papillary-thyroid carcinoma, 1 nodule was thyroid follicular carcinoma, and 3 nodules were benign ones. Conclusions US-FNA combined with detection of BRAF V600E and TI-RADS can improve the diagnostic accuracy and decrease the misdiagnosis in indeterminate nodules.

Objective To analyze the changes of benign and malignant thyroid nodules during followup and to increase the diagnostic accuracy.Methods The ultrasonographic changes of 161 thyroid nodules confirmed by pathology were retrospectively analyzed,including size,internal components,echogenicity,margin,microcalcification,anteroposterior to transverse dimension ratio(A/T) and the relationships with thyroid capsule,and the change of cervical lymph nodes.Results Ninety-three benign thyroid nodules and 68 malignant thyroid nodules were enrolled in this study,the changes in size,internal components,microcalcification,the relationships with glandular capsule,and cervical lymph nodes were significantly different (P ＜ 0.05),while the echogenicity,margin,and A/T were not significantly different,benign nodules changed more easily in size and internal components,while malignant nodules changed more easily in microcalcification.Both the nodules broken the thyroid capsule or cervical lymph nodes exhibited malignant signs suggest probable malignancy.Conclusions Analyzing the ultrasonographic changes during follow-up contributes to identify benign and malignant thyroid nodules.