Purpose To analyze the value of a deep learning(DL)radiomics model based on dynamic contrast-enhanced MRI images in predicting the expression of Ki-67 status in breast cancer.Materials and Methods A retrospective analysis of 152 breast cancer patients confirmed by pathological results at Guangdong Women and Children Hospital,MRI images and clinical pathological data were reviewed,and based on postoperative immunohistochemistry results,the images of the high and low expression groups of Ki-67 were randomly sampled in a ratio of 8∶2 to form a training set of 122 cases and a validation set of 30 cases.Single-factor and multi-factor Logistic regression analyses of clinical data were performed to select independent predictors of breast cancer expressing Ki-67 status.The ResNet-18 model was used as the basic model for DL feature extraction.Hand-crafted radiomic features and DL features were extracted.Eight machine learning models were constructed based on clinical features,hand-crafted radiomic features,DL features,and their combinations.The area under the receiver operating characteristic curve was used to evaluate the predictive performance of the models,and the best model was determined as the output model.Results The progesterone receptor status(OR=0.764,P=0.040)and human epidermal growth factor receptor-2 status(OR=1.187,P=0.046)were independent clinical predictors of breast cancer expressing Ki-67 status.The combined feature models demonstrated superior performance over the individual feature models,and the support vector machine algorithm had the highest prediction performance in the validation set,with an area under the curve of 0.847.Conclusion The DL radiomics model based on dynamic contrast-enhanced MRI images can effectively predict the expression of Ki-67 status in breast cancer.The support vector machine algorithm combined with feature model is the best,which can help the clinical diagnosis and treatment of breast cancer and prognosis evaluation.

Purpose To analyze the value of a deep learning(DL)radiomics model based on dynamic contrast-enhanced MRI images in predicting the expression of Ki-67 status in breast cancer.Materials and Methods A retrospective analysis of 152 breast cancer patients confirmed by pathological results at Guangdong Women and Children Hospital,MRI images and clinical pathological data were reviewed,and based on postoperative immunohistochemistry results,the images of the high and low expression groups of Ki-67 were randomly sampled in a ratio of 8∶2 to form a training set of 122 cases and a validation set of 30 cases.Single-factor and multi-factor Logistic regression analyses of clinical data were performed to select independent predictors of breast cancer expressing Ki-67 status.The ResNet-18 model was used as the basic model for DL feature extraction.Hand-crafted radiomic features and DL features were extracted.Eight machine learning models were constructed based on clinical features,hand-crafted radiomic features,DL features,and their combinations.The area under the receiver operating characteristic curve was used to evaluate the predictive performance of the models,and the best model was determined as the output model.Results The progesterone receptor status(OR=0.764,P=0.040)and human epidermal growth factor receptor-2 status(OR=1.187,P=0.046)were independent clinical predictors of breast cancer expressing Ki-67 status.The combined feature models demonstrated superior performance over the individual feature models,and the support vector machine algorithm had the highest prediction performance in the validation set,with an area under the curve of 0.847.Conclusion The DL radiomics model based on dynamic contrast-enhanced MRI images can effectively predict the expression of Ki-67 status in breast cancer.The support vector machine algorithm combined with feature model is the best,which can help the clinical diagnosis and treatment of breast cancer and prognosis evaluation.

Objective:To summarize the key prenatal ultrasound diagnosis features of Apert syndrome, analyze the causes of missed diagnosis and misdiagnosis, and propose corresponding preventive strategies.Methods:A retrospective analysis was made on the medical records and prenatal ultrasound images of 15 fetuses (including 14 cases referred from other hospitals) who underwent prenatal ultrasound examination in Guangdong Women and Children Hospital from August 2014 to May 2022 and were eventually clinically confirmed as Apert syndrome by induction or after birth. By conducting a comparative analysis, particularly focusing on the initial and final diagnoses of referral cases, the key ultrasound diagnostic points of Apert syndrome and the causes for missed and misdiagnosis were summarized.Results:①Diagnostic accuracy: Among the 15 fetuses, 11 cases (73.3%) were correctly diagnosed by prenatal ultrasound and 4 cases (26.7%) were missed diagnosis and misdiagnosis. For the 14 referral cases, only 2 cases (14.3%) were correctly identified in the initial diagnosis at the referring hospital (14.3%), and 12 cases (85.7%) were missed missed diagnosis and misdiagnosis. ②Detection rate of ultrasound signs: In the 15 fetuses with Apert syndrome, the detection rate of " cloverleaf" skull was 13.3% (2/15), premature coronal suture was 66.7% (10/15), the " brain shadowing sign" and flat occiput were both 93.3% (14/15), prominent forehead, hypertelorism and bilateral syndactyly of hands were all 100% (15/15), and bilateral syndactyly of feet was 73.3% (11/15). ③Analysis of missed diagnosis and misdiagnosis: Among the 4 cases of missed diagnosis and misdiagnosis in our hospital, premature closure of coronal suture, " brain shadowing sign", flat occiput and hypertelorism were all not recognized. Among these, 3 cases also missed the prominent forehead, bilateral syndactyly of hands and feet. Additionally, 1 case of bilateral syndactyly of hands was misdiagnosed as partial absence of metacarpals and phalangess.Conclusions:In the prenatal ultrasound diagnosis of fetal Apert syndrome, the symmetric syndactyly of both hands serves as an extremely important diagnostic clue. The " cloverleaf" skull is not common. The premature closure of coronal suture as a direct diagnostic sign with a high detection rate, highlighting its significance in the diagnostic of Apert syndrome. Furthermore, the high detection rates of characteristic ultrasound features such as prominent forehead, flat occiput, " brain shadowing sign" and hypertelorism could help to improve the accuracy of prenatal ultrasound diagnosis for Apert syndrome and effectively reduce missed and misdiagnosis.

This article reported a case of fetal giant hepatic hemangioma with cardiomegaly managed with intrauterine treatment. At 23 weeks of gestation, the patient was referred to Guangdong Women and Children Hospital due to abnormal abdominal echogenicity of the fetus, which was suspected to be a hepatic hemangioma or a hepatic arteriovenous fistula. The prenatal ultrasound at 26 weeks of gestation revealed an enlarged fetal hepatic hemangioma of 45 mm×35 mm×42 mm and an enlarged heart (cardiothoracic area ratio of 0.50). So, with the patient's informed consent, the fetus was treated with intrauterine administration of propranolol and dexamethasone and closely monitored by ultrasound. The volume of the lump still increased at the beginning of the medication, but started to shrink in the 7th week. Besides, the fetal cardiac load was reduced and the condition was controlled. The patient delivered at 37 weeks of gestation. The baby received a CT examination on the fourth day after birth which revealed an abdominal mass of 40 mm×30 mm×44 mm requiring no treatment, and no abnormalities were reported during a one-year follow-up.

Objective To establish domestic expected fetal lung volume (FLV) regression formula according to gestational week (GW) with large sample and data distributed evenly. Methods Prenatal MRI of 457 fetuses from 18 to 39 gestational weeks with normal chest was collected. FLV was measured on the axial images of single shot fast spin echo sequence. All FLV data were used for regression analysis. The regression model was determined according to the degree of fitting and test results, and the expected FLV regression formula based on GW was obtained. Results The mean GW of 457 fetuses was (29.44±4.61) weeks. There were 43.11% (197/457) fetuses at 28 weeks and below. FLV based on MRI measurement ranged from 9.47 ml to 131.48 ml, mean (50.78±23.76) ml. The final regression formula was expected FLV=0.004g2.76 (g=GW, R2=0.845, P<0.001). Conclusion The final regression formula of expected FLV is simple, practical and applicable to most fetuses, and is helpful to the research on prenatal assessment of domestic fetuses lung.

Objective To explore the diagnostic value of prenatal MRI in fetal Joubcrt syndrome and related disorders (JSRD).Methods Prenatal MRI data of 5 JSRD fetuses were analyzed retrospectively.The ratio of AP diameters of crown to root of molar tooth sign,the ratio of AP diameters of isthmus to interpeduncular fossa,and the ratio of the transverse to AP diameters of the roof of the fourth ventricle on the axial images of the midbrain were measured.Results Molar tooth signs were observed in 5 fetuses on the axial images of midbrain.The cerebellar vermis disappeared,and cleft sign was observed.Enlarged cisterna magna was observed in 3 fetuses.The ratios of AP diameters of crown to root were 0.58-0.90,of AP diameters of isthmus to interpeduncular fossa were 0.84-1.00,and of transverse to AP diameters of the fourth ventricle roof were 0.42-0.60.Conclusion Molar tooth sign and cleft sign were characteristic findings for JSRD on MRI,therefore being helpful to prenatal diagnosis of JSRD.

Objective To study on the factors of countercurrent occured in hysterosalpingography to improve the understanding of countercurrent.Methods 180 patients who underwent hysterosalpingography due to infertility were recruited,and 63 of them who were involved in countercurrent in the process of hysterosalpingography were analyzed statistically.Results The single factor analy-sis demonstrated that such four factors of primary/secondary infertility,menstrual clean days,tubal obstruction or not,and depth of cannula were associated with countercurrent,while Logistic regression analysis indicated that the factors of menstrual clean days, tubal obstruction or not,and depth of cannula during the hysterosalpingography operation were closely related.Conclusion Counter-current are caused by the comprehensive impact of several comprehensive factors like menstrual clean day,tubal obstruction or not, and cannula operation.Therefore adequate preparation should be made before and during the operation,to reduce the occurrences of countercurrent.

This guideline was published by the American Institute of Ultrasound in Medicine,which has been developed to assist physicians performing sonographic studies of the brain in neonates and infants.Neurosonography should be performed only when there is a valid medical reason,and gain the necessary diagnostic information.Although it is not possible to detect every abnormality,adherence to the following guideline will maximize the detection of most abnormalities of the brain in neonates and infants that can be imaged with ultrasound.

Objective To investigate the ability of T2 mapping in early detection of the change of cartilage collagen by quantitative analysis.Methods 20 porcine patellae were randomly assigned into 4 groups,3 treated groups and 1 control group.The samples of 3 treated groups were respectively immersed in PBS with 100 mg/100 mL,1 50 mg/100 mL and 200 mg/100 mL typeⅡ collagenase for 4 h,whereas control group samples in PBS for 4 h.T2 relaxation times of superficial,deep and full layers of cartilage and mean densi-ty after Van Gieson stain were measured.The correlation of T2 relaxation times and mean density were statistically analyzed.Results Gross pathology showed all articular cartilages maintained intact after treatment.T2 relaxation times of superficial and full layers and mean density increased significantly in any treated group in comparison with control group(P <0.05).Significant differences in T2 relaxation times of superficial layers and mean density were found between any two treated groups,whereas no difference of T2 re-laxation time of full layers was found between 100 mg/100 mL and 1 50 mg/100 mL treated groups(P >0.05).Pearson correlation analysis showed a close correlation(r=-0.837)between T2 relaxation times of full layers and mean density.Conclusion T2 map-ping can adequately detect the change of cartilage collagen in the early stage of osteoarthritis,and shows a good clinical application prospect.

Objective To investigate the ability of the four-slice perfusion maps derived by multislice CT(MSCT)perfusion imaging (CTPI) to detect ischemic lesion in rabbit models with the middle cerebral artery occlusion(MCAO). Methods Sequential dynamic CT studies were performed at four slices in 2 control rabbits and another 8 with MCAO. The volume of critically ischemic lesions on CTPI maps at 1st,2nd and 3rd hour after operation and the volume of infarction measured by postmortem 2,3,5- triphenyltetrazolium chloride (TTC) staining were analyzed by analysis of variance with random blocks design and linear correlation. Results In 6 of 8 animals of the ischemia group, critically ischemia or infarction was visible on perfusion maps and TTC-staining,while appearing negative in 2 control rabbits. In four-slice CTPI maps, lesions were visualized in 2 of 4 slices at least. The results of analysis of variance demonstrated that the intergroup differences were discovered between 1st , 2nd and 3rd hour after operation, and between MTT map and other CTPI maps (P