To assess the value of a deep learning-based nomogram in predicting axillary lymph node (ALN) status following neoadjuvant chemotherapy (NAC) in breast cancer patients.
Methods
Four hundred and fourteen ALN-positive breast cancer patients who received NAC between March 2020 and June 2023 were enrolled in this retrospective study and divided into a training set and an external test set.The training set consisted of 257 patients from the First Affiliated Hospital of Nanjing Medical University, while the external test set included 157 patients from the General Hospital of Eastern Theater Command.All patients were divided into pathologically complete response (pCR) and non-pCR(npCR) groups based on the pathology results of ALN surgery post-NAC.A deep learning model based on the ResNet50 architecture was trained and established using pre-NAC ultrasound images of breast tumors.Univariate and multivariate logistic regression analyses were performed on the training set to identify independent risk factors for post-NAC ALN status.These independent risk factors were then used to construct a clinical model.A deep learning-based nomogram was constructed by combining independent risk factors and deep learning predictive probabilities.The performance of the models was evaluated using receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis, and clinical impact curve.Two radiologists with different experience levels independently predicted ALN status in the external test set based on ultrasound images and pre-NAC immunohistochemical results, and performed a second prediction with the assistance of the deep learning-based nomogram.The two prediction results were compared.
Results
Estrogen receptor (ER) and human epidermal growth factor receptor 2 (Her-2) were identifled as independent risk factors for predicting post-NAC ALN status.The area under the curve (AUC) values of the clinical model, deep learning model, and deep learning-based nomogram were 0.724, 0.872, and 0.878 in the training set, and 0.698, 0.831, and 0.859 in the external test set, respectively.The deep learning-based nomogram outperformed the clinical model (both P<0.001 in training and external test sets) and showed superior performance to the deep learning model in the external test set (P=0.024).The AUC values of radiologist 1 (low-experience) and radiologist 2 (high-experience) for independent judgment were 0.570 and 0.606, respectively, both signiflcantly lower than those of the deep learning model and the deep learningbased nomogram (all P<0.001).With the assistance of the deep learning-based nomogram, the AUC values of radiologist 1 and radiologist 2 improved to 0.796 and 0.807, respectively, showing statistically signiflcant differences compared to independent judgment (both P<0.001).
Conclusion
The deep learning-based nomogram based on pre-NAC ultrasound images can effectively predict the pathological status of ALN in breast cancer patients after NAC treatment, providing more evidence for the development of personalized treatment plans.
To analyze the breast cancer screening results based on ultrasound in the rural areas of Weibei, Shaanxi Province from 2021 to 2023, and to explore the screening effectiveness and related influencing factors of the ultrasound-based breast cancer screening model.
Methods
A total of 139 622 women aged 35-64 years who participated in the free breast cancer screening program in the rural areas of Weibei, Shaanxi Province from January 2021 to December 2023 were selected.The data were derived from the breast cancer screening result reports submitted by the counties participating in the screening program in this region.All the women included in the study underwent clinical examination and breast ultrasound.Ultrasound reports with BI-RADS categories 1 and 2 were considered negative screening results,while those with BI-RADS categories 0, 3, 4, and 5 were considered positive screening results.Women with BI-RADS category 0 or 3 underwent supplemental breast mammography, while those with BI-RADS category 4 or 5 underwent biopsy.The survey questionnaires returned by the participants and the screening results were statistically analyzed to calculate screening-related indicators, including recall rate, biopsy rate, positive predictive value of ultrasound screening, breast cancer detection rate, and early diagnosis rate of breast cancer.Multivariate binary logistic regression analysis was used to identify the factors influencing positive ultrasound screening results and breast cancer detection.
Results
The initial ultrasound screening results showed a recall rate of 4.18% (5832/139 622) and a positive predictive value of ultrasound screening of 1.06% (62/5832).The biopsy rate in the study sample was 0.25% (356/139 622), with a total of 63 breast cancer cases detected.The breast cancer detection rate was 0.45‰ (45.12 per 100 000), and the early diagnosis rate was 80.95%(51/63).Among these, 62 cases were detected by ultrasound, while 1 case was not detected by ultrasound.The analysis of factors influencing positive ultrasound screening results showed that age 45-54 years [OR (95%CI):1.16 (1.08-1.25)], high school education [OR (95%CI):1.46 (1.37-1.56)], college or higher education [OR(95%CI):1.30 (1.16-1.45)], and premenopausal status [OR (95%CI):1.68 (1.56-1.81)] were independent risk factors for positive ultrasound screening results (all P<0.05).The analysis of the influencing factors for breast cancer detection revealed that having an educational level of college degree or above [OR (95%CI):4.77 (1.34-16.95), P<0.05] was an independent risk factor for breast cancer detection.
Conclusion
The detection and early diagnosis of breast cancer in the breast cancer screening program in the rural areas of Weinan, Shaanxi Province meet the relevant quality control requirements, indicating that the ultrasound-based breast cancer screening model is feasible and effective in this region.
To investigate the incidence, associated characteristics, and postnatal follow-up results of ventricular septal defect (VSD) detected in a national ultrasound screening program for fetal congenital heart defects (CHDs) in China.
Methods
From January 1, 2011 to December 31,2013, a ultrasound screening for CHDs was investigated in 92 hospitals across 31 provinces/municipalities/autonomous regions in China.A total of 759 physicians in these hospitals, who had undergone standardized prenatal ultrasound training and obtained qualiflcations for prenatal diagnosis, participated in this research project.Fetal ultrasound screening was performed in accordance with relevant practice guidelines and hierarchical referral criteria.Data on fetuses with CHDs and their mothers were collected and uploaded using a “data investigation software” system developed by the research group.The incidence, associated characteristics, and follow-up outcomes of VSD was analyzed based on the aggregated data from 92 hospitals during the fetal cardiac screening.
Results
A total of 18 171 fetal CHD cases were identified from 2 452 249 pregnancies.Among these cases, 4345 fetuses were diagnosed with ventricular septal defect(VSD), accounting for 23.9% of all CHD cases (4345/18 171).The incidence of VSD was 17.72 per 10 000(4345/2 452 249; 95% confldence interval:17.20-18.25), which was the highest among the 36 subtypes of CHD.The 31 provinces, municipalities, and autonomous regions in China were divided into six geographical regions.The highest incidence of fetal VSD was found in the eastern region (20.59 per 10 000; 854/414 815),while the lowest incidence occurred in the northwestern region (8.85 per 10 000; 109/123 189).There was a statistically signiflcant difference in incidence among the six geographical regions (χ² = 166.71, P<0.01).Among the 4345 VSD cases, isolated VSD was identifled in 3171 cases (72.98%), while 1015 cases (23.36%)were associated with extracardiac malformations.The top three extracardiac malformations were central nervous system anomalies (427 cases), single umbilical artery and umbilical cord cysts (241 cases), and craniofacial anomalies (219 cases).A total of 534 VSD fetuses underwent genetic study for chromosomal abnormality, with 142 cases (27%) showing chromosomal abnormalities.The most common abnormalities were trisomy 21 (31 cases) and trisomy 18 (59 cases).Postnatal follow-up was carried out for the fetal VSD cases, with 2835 fetuses born alive, 674 cases terminated, and 836 cases lost to follow-up.The concordance rate between prenatal screening results and postnatal veriflcation was 89.60% (3144/3509).
Conclusion
This study provides the incidence of VSD in China, including the regional incidence in the six geographical regions, as well as types of extracardiac malformations associated with VSD, the incidence of chromosomal abnormalities, and the concordance rate between prenatal screening results and postnatal veriflcation.These flndings offer valuable foundational data for further research on fetal VSD.
To assess the value of fetal heart quantiflcation (fetal HQ) in evaluating heart morphology and function in fetuses with simple complete transposition of the great arteries (D-TGA).
Methods
Thirty-one fetuses with simple D-TGA (D-TGA group) diagnosed by fetal echocardiography who visited Sir Run Run Shaw Hospital, Zhejiang University School of Medicine from May 2020 to December 2023 were retrospectively analyzed, and 171 normal fetuses with matched gestational weeks were used as controls (control group).Fetal standard 4-chamber views (4CV) were collected, and fetal HQ software was applied to measure end diastolic basal apical length, transverse width, and area, and calculate the global spherical index (GSI).Measurements of left and right ventricular functional changes were also performed,including left ventricular ejection fraction (EF), and left and right ventricular fraction area change (FAC).Left and right ventricular end-diastolic areas were analyzed by speckle tracking imaging (STI), and longitudinal diameter and 24-segment end-diastolic diameter (ED), spherical index (SI), fractional shortening (FS),and global longitudinal strain (GLS) of the left and right ventricles were analyzed by STI.The great artery parameters of the fetuses in both groups were analyzed, including aortic diameter (AO), pulmonary artery diameter (PA), main pulmonary artery diameter (MPA), and their Z-scores.
Results
The differences in GSI, left and right ventricular FAC, left and right ventricular GLS, left ventricular end-diastolic area, and left ventricular length were signiflcant between the D-TGA group and the control group (all P<0.05).There were statistically signiflcant differences in ED of left ventricular segments 20-24 ED and right ventricular segments 8-24 (all P<0.05), in FS of left ventricular segments 20-24 and right ventricular segments (all P<0.05).and in SI of left ventricular segments 1-4, 10-17, and 19, and right ventricular segments 1-9 (all P<0.05)between the two groups.The difference between fetal AO Z-score and PA Z-score in the D-TGA group was statistically significant (1.32±0.70 vs -0.22±1.09, P<0.05).
Conclusion
Quantitative assessment of fetal cardiac morphology and function using the fetal HQ technique provides a reference basis for analyzing ventricular remodeling in D-TGA fetuses.
To evaluate the current situation, clinical value, and limitations of the involvement of ultrasound in the multi-disciplinary treatment (MDT) of inflammatory bowel disease (IBD)in China.
Methods
This questionnaire-based survey was conducted on an online questionnaire platform(Questionnaire Star) from November 1, 2024 to December 1, 2024.The survey examined the frequency,methods, personnel, working procedures, clinical value, and major limitations of ultrasound involvement in IBD-MDT across various centers in China.
Results
A total of 90 medical centers conducting IBD-MDT and 201 respondents participated in the survey.Ultrasound radiologists were involved in 47.8% (43/90)of the IBD-MDT centers, including 3 non-tertiary hospitals.The participation of ultrasound radiologists was associated with hospital level, the duration of IBD-MDT implementation, the frequency of IBDMDT meetings, the number of patients discussed in a single session, the proportion of patients receiving intestinal ultrasound, and the professionalism of ultrasound radiologists (P<0.05).In 69.8% (30/43) of the centers, the participation frequency of ultrasound radiologists was sometimes, occasional, or rare, but 74.4% (32/43) of the centers reported in-depth discussions during the meetings.In 76.9% (30/39) of the centers, 1-2 fixed ultrasound radiologists participated in IBD-MDT, and 71.4% (30/42) of the centers had ultrasound radiologists specialized in gastrointestinal ultrasound.In 69.5% (16/23), 78.3% (18/23), and 78.3% (18/23) of the centers, ultrasound radiologists always or usually prepared in advance by understanding the objectives of the discussions, reviewing medical records, and examining ultrasound images.In 91.3%(21/23) of the centers, ultrasound radiologists prepared PPT or relevant documents in advance or presented images and engaged in oral discussions during the meetings.The clinical value of intestinal ultrasound was widely recognized in IBD-MDT, especially in assessing disease activity, complications, perianal lesions,and drug efficacy (Median=4, IQR=1).The main limitations to the implementation of intestinal ultrasound included a lack of competent intestinal ultrasound radiologists (128/201, 63.7%), insufficient consistency among operators (117/201, 58.2%), variability in report quality due to insufficient standardization (102/201,50.8%), and insufficient awareness of intestinal ultrasound among clinical departments (111/201, 55.2%).
Conclusion
The value of ultrasound in IBD-MDT is widely acknowledged.However, variations exist in participation frequency, depth, personnel specialization, and working procedures.Strengthening training in IBD intestinal ultrasound to improve examination quality, and promoting the involvement of ultrasound radiologists in IBD-MDT, which will help accumulate valuable experience, are important for enhancing the role of ultrasound in the comprehensive management of IBD.
To analyze the causes of missed diagnosis of constrictive pericarditis by echocardiography, and investigate the safety of pericardiectomy in patients with atypical constrictive pericarditis.
Methods
A total of 92 patients with constrictive pericarditis who visited the Third People's Hospital of Chengdu from January 2020 to December 2024 were retrospectively included.All patients underwent echocardiography and CT examination.Based on the first echocardiographic examination, the patients were divided into an accurate diagnosis group(n=82)and a missed diagnosis group(n=10),and the clinical and imaging characteristics of the two groups were compared and analyzed.
Results
Compared to the accurate diagnosis group, the missed diagnosis group had a shorter disease course [9.00(11.00)months vs 2.50(7.90)months, P<0.05] and milder liver function impairment.The accurate diagnosis group exhibited higher proportions of hepatic vein dilation, ascites, and pleural effusion (all P<0.05).The missed diagnosis group demonstrated milder cardiac morphological changes, including the changes of left atrial size [41.00(8.30)mm vs 37.50(4.30)mm, P<0.05], left ventricular size [40.50(4.00)mm vs 45.50(4.30)mm, P<0.05], right atrial size [41.00(7.00)mm vs 36.50(9.00)mm,P<0.05], right atrioventricular angle [(80.71±30.55)° vs(106.50±35.87)°, P<0.05], apical sphericity index [69.00(29.00) vs 84.00(25.00), P<0.05], and left ventricular mass index [(69.79±15.91)g/m2 vs(84.51±13.53)g/m2, P<0.05].Both groups had thickened pericardium, but the missed diagnosis group did not have calcifled pericardium.Postoperative mortality [3.7% (3/82) vs 0 (0/10)] and adverse event rates showed no signiflcant difference between the two groups (P>0.05).
Conclusion
Echocardiography exhibits characteristic flndings for constrictive pericarditis, but misdiagnosis occurs in clinical practice.Missed cases are often early-stage patients with pericardial thickening but no calciflcation, subtle cardiac morphological changes, and milder symptoms.Multimodal imaging evaluation combined with clinical history can improve diagnostic accuracy compared with reliance on echocardiography alone.
To evaluate the long-term dynamic changes in right heart morphology and function in children with repaired tetralogy of Fallot (rTOF) by using echocardiography for exploring the application value of right ventricular longitudinal strain.
Methods
A total of 22 rTOF children who had undergone surgery at Beijing Children's Hospital from July 2023 to September 2024 were included in this study.The average age was 11.9±2.5 years, and the median follow-up time was 11 years.The diameters of heart chambers, pulmonary valve regurgitation, and parameters of right ventricular function including tricuspid annular plane systolic excursion (TAPSE), fractional area change (FAC), free wall longitudinal strain (FWLS), septal longitudinal strain (SLS), and global longitudinal strain (GLS) were obtained.The echocardiographic data of rTOF patients obtained 5 years ago were reviewed (the median follow-up time was 5 years).The data comparison between the 11-year follow-up and 5-year follow-up was carried out.Healthy children of similar body surface area and the same gender were included as the control group.The right heart morphology and function parameters of rTOF patients and the control group at two follow-up time points were analyzed.
Results
Compared with the control group at the two follow-up time points, rTOF patients had enlarged right ventricles and increased pulmonary artery flow velocity.The parameters of TAPSE, FAC,and longitudinal strain (FWLS, SLS, and GLS) were all lower than those of the control groups (P<0.05).Among the 22 patients at the two follow-up time points, 50.0% (11/22, median follow-up time of 5 years)and 59.1% (13/22, median follow-up time of 11 years) had moderate to severe pulmonary regurgitation,respectively.Comparing the data of rTOF patients at the two follow-up time points (median follow-up time of 11 years vs median follow-up time of 5 years), TAPSE [(0.23±0.04) vs (0.25±0.05)], FAC [(27.46±5.75)%vs (28.03±6.61)%], and FWLS [(-19.00±4.23)% vs (-20.47±5.49)%] did not show a progressive decline.In contrast, SLS [(-17.79±3.46)% vs (-20.82±4.30)%, P<0.05] and GLS [(-18.39±3.68)% vs(-20.65±4.61)%, P<0.05] showed a progressive decrease.
Conclusion
In the long-term follow-up of rTOF patients, echocardiography indicates right ventricular volume overload due to the persistent pulmonary valve regurgitation, and right heart function parameters are lower than those of healthy children.Certain right ventricular longitudinal strain parameters can be used to evaluate dynamic right ventricular function.
To develop a dose prediction model for ultrasound-guided thrombin treatment of pseudoaneurysms using the Light Gradient Boosting Machine (LightGBM) algorithm.
Methods
A retrospective analysis was conducted on 84 patients diagnosed with femoral artery pseudoaneurysms via ultrasound and treated with ultrasound-guided thrombin injection at the First Affiliated Hospital of Nanjing Medical University between January 2018 and December 2024.Patients were categorized into three groups based on thrombin dosage:low-dose (<500 IU, 30 cases), mediumdose (≥500 IU and <1000 IU, 36 cases), and high-dose (≥1000 IU, 18 cases).The cohort was randomly divided into a training set (67 cases) and a validation set (17 cases) at an 8:2 ratio.Feature variables were screened using ordinal logistic regression analysis to construct a logistic-based thrombin dose prediction model.Additionally, LightGBM contribution-based feature selection was applied to build a LightGBM-based dose prediction model.Model performance was evaluated using overall accuracy,micro-average area under the curve (AUC), recall, F1-score, and receiver operating characteristic(ROC) curve analysis.
Results
In the training set, the logistic regression model demonstrated an overall accuracy of 0.677 and a micro-average AUC of 0.744 (95% confidence interval [CI]:0.674-0.815); in the validation set, the corresponding values were 0.686 and 0.758 (95%CI:0.624-0.891).The LightGBM-based model exhibited superior performance, with a training set overall accuracy of 0.930,micro-average AUC of 0.975 (95%CI:0.955-0.995), micro-average recall of 92.5%, and micro-average F1-score of 0.899.In the validation set, it achieved an overall accuracy of 0.804, micro-average AUC of 0.872 (95%CI:0.766-0.978), micro-average recall of 76.5%, and micro-average F1-score of 0.722.
Conclusion
The LightGBM-based thrombin dose prediction model effectively forecasts thrombin dosage requirements, offering a valuable reference for achieving precision and individualized treatment in ultrasound-guided thrombin injection therapy.