Clinical Ultrasound Images Research Articles

YOLO-based thyroid ultrasound for assisted diagnosis

Abstract Real-time intelligent diagnosis of thyroid ultrasound images using deep learning techniques can mitigate the impact of human subjective factors on diagnostic outcomes. Three thousand and fifty-five thyroid ultrasound images were acquired from 205 selected patients aged 10 to 77 years who underwent ultrasound examinations between 2019 and 2023. Each image contained at least one area of thyroid nodules, totaling 3, 088 nodules, including 1, 752 benign and 1, 336 malignant nodules. YOLOV8 is the baseline model for developing an end-to-end architecture for thyroid nodule detection. This architecture automatically identifies nodule lesions in ultrasound images, classifies them as benign or malignant, and enables real-time detection in video frames. The experimental results on 611 clinical thyroid ultrasound images demonstrate that our method can accurately diagnose benign and malignant nodal lesions in thyroid ultrasound with 88.1% and 84.6% accuracy, respectively. The mean average accuracy is as high as 91.1%, indicating a 5.1% improvement compared to the baseline model.

The Study of Echocardiography of Left Ventricle Segmentation Combining Transformer and Convolutional Neural Networks.

Accurate prediction of echocardiographic parameters is essential for diagnosis and treatment of cardiac disease, especially for segmentation of the left ventricle to obtain measurements such as left ventricular ejection fraction and volume. However, manually outlining left ventricle on echocardiographic images is a time-consuming and physician experience-dependent task. Therefore, it is crucial to develop an accurate and efficient automatic segmentation tool. Therefore, we aimed to explore a model to perform echocardiography of left ventricle segmentation by combining transformer and convolutional neural networks (CNN).ResNet-50 was used in CNN branch. The encoder-decoder architecture was used for transformer branch, which was fused to the corresponding feature maps of the CNN branches. Fusion module was used to effectively combine feature information from the CNN and transformer. Bridge attention used to increase sensitivity and prediction accuracy of model. The entire network was trained end-to-end using the binary cross-entropy with logits loss L.In this work, we propose an automatic left ventricular (LV) segmentation model based on Transformer and CNN that efficiently captures global dependencies and spatial details and create a fusion module using CBAM that fuses Transformer and CNN features. In addition, attention is also computed using multi-level fusion features to obtain the final attention segmentation map. The model was trained and evaluated on a large cardiac image dataset, EchoNet-Dynamic, with test dice coefficient of 92.4%.The results show that our model can better segment left ventricle. We also tested our model on clinical patient ultrasound images, and visualization results proved effectiveness of the model.

Application and diagnostic points of carotid artery ultrasound in the detection of internal carotid artery absence: Analysis of 16 cases and literature review.

To summarize the ultrasound diagnostic key points and imaging characteristics of absence of the internal carotid artery (ICA), and to explore the application value of carotid artery ultrasound in the diagnosis of ICA absence. Sixteen patients diagnosed with ICA absence at our hospital were retrospectively analyzed. Clinical data, ultrasound images, and other relevant imaging data were reviewed. The ultrasound diagnostic key points and related imaging characteristics of ICA absence were summarized. Combined with a literature review, the application of carotid artery ultrasound in the diagnosis of ICA absence was discussed. Among the 16 patients, there were 10 males and 6 females ages ranging from 13 to 80 years. Among them, 5 patients had a history of cerebral infarction, 1 patient had a history of subarachnoid hemorrhage, and 10 patients had no history of stroke. Among the 16 patients, there were 10 left-sided lesions, 5 right-sided lesions, and 1 bilateral lesion. Three patients had concomitant aneurysms. Carotid ultrasound examinations in all 16 ICA absent patients revealed that a significantly slender common carotid artery (CCA) diameter compared to the contralateral (3.91 ± 0.73 vs. 6.92 ± 1.42, p < 0.001). On the side of ICA absence, the bifurcation of the carotid artery was not detected, with the CCA directly continuing as a single artery (external carotid artery, ECA). The temporal superficial artery percussion test was positive in all cases, and the blood flow spectrum of the single artery (ECA) on the side of the absent ICA showed a serrated pattern during diastole. On the side where the ICA was absent, the blood flow spectrum morphology of the CCA and ECA was consistent. Based on other relevant imaging examinations, the patients were classified into Lie types: A (10 cases), B (2 cases), C (1 case), D (1 case), and 2 cases were unclassified. Carotid artery ultrasound has the advantages of easy operation, affordability, non-invasiveness, and capable of dynamic and high repeatability. Compared to other imaging examinations, it can provide more evidence for the diagnosis of ICA absence, and is suitable for widespread application and promotion.

Ovarian-adnexal reporting and data system ultrasound evaluation and pathological characteristics of ovarian collision tumor.

Collision tumor are neoplasms, including two histologically distinct tumors that coexist in the same mass without histological admixture. The incidence of collision tumor is low and is rare clinically. To investigate ultrasound images and application of ovarian-adnexal reporting and data system (O-RADS) to evaluate the risk and pathological characteristics of ovarian collision tumor. This study retrospectively analyzed 17 cases of ovarian collision tumor diagnosed pathologically from January 2020 to December 2023. All clinical features, ultrasound images and histopathological features were collected and analyzed. The O-RADS score was used for classification. The O-RADS score was determined by two senior doctors in the gynecological ultrasound group. Lesions with O-RADS score of 1-3 were classified as benign tumors, and lesions with O-RADS score of 4 or 5 were classified as malignant tumors. There were 17 collision tumors detected in 16 of 6274 patients who underwent gynecological surgery. The average age of 17 women with ovarian collision tumor was 36.7 years (range 20-68 years), in whom, one occurred bilaterally and the rest occurred unilaterally. The average tumor diameter was 10 cm, of which three were 2-5 cm, 11 were 5-10 cm, and three were > 10 cm. Five (29.4%) tumors with O-RADS score 3 were endometriotic cysts with fibroma/serous cystadenoma, and unilocular or multilocular cysts contained a small number of parenchymal components. Eleven (64.7%) tumors had an O-RADS score of 4, including two in category 4A, six in category 4B, and three in category 4C; all of which were multilocular cystic tumors with solid components or multiple papillary components. One (5.9%) tumor had an O-RADS score of 5. This case was a solid mass, and a small amount of pelvic effusion was detected under ultrasound. The pathology was high-grade serous cystic cancer combined with cystic mature teratoma. There were nine (52.9%) tumors with elevated serum carbohydrate antigen (CA)125 and two (11.8%) with elevated serum CA19-9. Histological and pathological results showed that epithelial-cell-derived tumors combined with other tumors were the most common, which was different from previous results. The ultrasound images of ovarian collision tumor have certain specificity, but diagnosis by preoperative ultrasound is difficult. The combination of epithelial and mesenchymal cell tumors is one of the most common types of ovarian collision tumor. The O-RADS score of ovarian collision tumor is mostly ≥ 4, which can sensitively detect malignant tumors.

Proposal of a classification of cannulation damage in vascular access grafts based on clinical, ultrasound, and microscopic observations.

Arteriovenous grafts (AVGs) serve as an alternative to native arteriovenous fistulas (AVFs) in the context of hemodialysis patient life planning. AVGs are more susceptible to developing outflow stenosis (due to intimal hyperplasia), thrombosis, and infections. However, an often overlooked contributor to AVG failure is cannulation damage. The objective of this paper is to assess the impact of cannulations on AVGs. We aim to establish a classification of AVG damage by comparing clinical data and ultrasound images with microscopic morphological findings obtained from explanted grafts. This study is conducted at a single center. We included all patients who underwent AVG creation between 2011 and 2019. Comprehensive data on clinical history, follow-up, and complications were collected and reviewed. Duplex ultrasound (DUS) characteristics were documented, and all grafts explanted during the analysis period underwent optical microscopy evaluation. Finally, clinical data, along with DUS and microscopic findings, were integrated to derive a damage classification. During the study period, 247 patients underwent 334 early cannulation AVGs. The median follow-up duration was 714 days (IQR 392, 1195). One hundred eleven (33%) grafts were explanted. Clinical data and DUS findings were utilized to formulate a four-grade classification system indicating increasing damage. Cannulation damage alone does not solely account for AVG failure. It results from a biological host-mediated process that promotes the growth of intimal hyperplasia at the cannulation sites. This process is not clinically significant within the initial 2 years after AVG creation.

Development and validation of a clinical and ultrasound features-based nomogram for preoperative differentiation of renal urothelial carcinoma and central renal cell carcinoma.

This study aimed to develop and validate an ultrasound (US)-based nomogram for the preoperative differentiation of renal urothelial carcinoma (rUC) from central renal cell carcinoma (c-RCC). Clinical data and US images of 655 patients with 655 histologically confirmed malignant renal tumors (521 c-RCCs and 134 rUCs) were collected and divided into training (n = 455) and validation (n = 200) cohorts according to examination dates. Conventional US and contrast-enhanced US (CEUS) tumor features were analyzed to determine those that could discriminate rUC from c-RCC. Least absolute shrinkage and selection operator regression was applied to screen clinical and US features for the differentiation of rUC from c-RCC. Using multivariate logistic regression analysis, a diagnostic model of rUC was constructed and visualized as a nomogram. The diagnostic model's performance was assessed in the training and validation cohorts by calculating the area under the receiver operating characteristic curve (AUC) and calibration plot. Decision curve analysis (DCA) was used to assess the clinical usefulness of the US-based nomogram. Seven features of both clinical features and ultrasound imaging were selected to build the diagnostic model. The nomogram achieved favorable discrimination in the training (AUC = 0.996, 95% CI: 0.993-0.999) and validation (AUC = 0.995, 95% CI: 0.974, 1.000) cohorts, and good calibration (Brier scores: 0.019 and 0.016, respectively). DCA demonstrated the clinical usefulness of the US-based nomogram. A noninvasive clinical and US-based nomogram combining conventional US and CEUS features possesses good predictive value for differentiating rUC from c-RCC.

Assessment of thermal index compliance in clinical ultrasound examinations.

The thermal index (TI) is critical for monitoring potential heating risks in ultrasound examinations. This study aims to analyse adherence to TI guidelines in clinical ultrasound exams, focusing on the thermal index for soft tissue (TIS) and thermal index for bone at the focus (TIB). Clinical ultrasound images over a five working day period from a single NHS University Hospital Trust were analysed using a custom MATLAB script. TI values were extracted from DICOM headers, supplemented by optical character recognition for missing values. Time at each TI value (image interval) was measured assuming a constant TI between consecutive images. TI values and corresponding image intervals were compared to TI guidelines provided by the British Medical Ultrasound Society (BMUS). Analysis included 627 obstetric and 1023 non-obstetric examinations, comprising 4643 obstetric and 19,228 non-obstetric image intervals. TIS values across all examinations and image intervals adhered to recommended BMUS limits; however, breaches of the guidelines were observed for TIB. In obstetrics, 44 image intervals across 47 examinations exceeded the TIB limits (1.01% of TIB image intervals and 7.02% of examinations). For non-obstetrics, two image intervals across two examinations exceeded the TIB limits (0.02% of intervals and 0.44% of examinations). The majority of ultrasound examinations considered for this analysis demonstrated compliance with BMUS guidelines for TI, except for a small fraction which showed breaches in the obstetric thermal index for bone at the focus limits. All breaches were associated with the use of pulsed-wave Doppler.

Identification method of thyroid nodule ultrasonography based on self-supervised learning dual-branch attention learning framework.

Thyroid ultrasound is a widely used diagnostic technique for thyroid nodules in clinical practice. However, due to the characteristics of ultrasonic imaging, such as low image contrast, high noise levels, and heterogeneous features, detecting and identifying nodules remains challenging. In addition, high-quality labeled medical imaging datasets are rare, and thyroid ultrasound images are no exception, posing a significant challenge for machine learning applications in medical image analysis. In this study, we propose a Dual-branch Attention Learning (DBAL) convolutional neural network framework to enhance thyroid nodule detection by capturing contextual information. Leveraging jigsaw puzzles as a pretext task during network training, we improve the network's generalization ability with limited data. Our framework effectively captures intrinsic features in a global-to-local manner. Experimental results involve self-supervised pre-training on unlabeled ultrasound images and fine-tuning using 1216 clinical ultrasound images from a collaborating hospital. DBAL achieves accurate discrimination of thyroid nodules, with a 88.5% correct diagnosis rate for malignant and benign nodules and a 93.7% area under the ROC curve. This novel approach demonstrates promising potential in clinical applications for its accuracy and efficiency.

Application of Machine-learning based on Radiomics Features in Differential Diagnosis of Superficial Lymphadenopathy.

The accurate diagnosis of superficial lymphadenopathy is challenging. We aim to explore a non-invasive and accurate machine-learning method for distinguishing benign lymph nodes, lymphoma, and metastatic lymph nodes. The clinical data and ultrasound images of 160 patients with superficial lymphadenopathy (58 benign lymph nodes, 62 lymphoma, 40 metastatic lymph nodes) admitted to our hospital from January 2020 to November 2022 were retrospectively studied. Patients were randomly divided into a training set and test set according to the ratio of 6:4. Firstly, the radiomics features of each lymph node were extracted, and then a series of statistical methods were used to avoid over-fitting. Then, the gradient boosting machine(GBM) was used to build the model. The area under receiver(AUC) operating characteristic curve, precision, recall rate and F1 value were calculated to evaluate the effectiveness of the model. Ten robust features were selected to build the model. The AUC values of benign lymph nodes, lymphoma and metastatic lymph nodes in the training set were 1.00, 0.98 and 0.99, and the AUC values of the test set were 0.96, 0.84 and 0.90, respectively. It was a reliable and non-invasive method for the differential diagnosis of lymphadenopathy based on the model constructed by machine learning.

Multimodal Machine Learning-Based Ductal Carcinoma in situ Prediction from Breast Fibromatosis.

To develop a clinical-radiomics model using a multimodal machine learning method for distinguishing ductal carcinoma in situ (DCIS) from breast fibromatosis. The clinical factors, ultrasound features, and related ultrasound images of 306 patients (198 DCIS patients) were retrospectively collected. Patients in the development and validation cohort were 184 and 122, respectively. The independent clinical and ultrasound factors identified by the multivariable logistic regression analysis were used for the clinical-ultrasound model construction. Then, the region of interest of breast lesions was delineated and radiomics features were extracted. Six machine learning algorithms were trained to develop a radiomics model. The algorithm with higher and more stable prediction ability was chosen to convert the output of the results into the Radscore. Further, the independent clinical predictors and Radscore were enrolled into the logistic regression analysis to generate a combined clinical-radiomics model. The receiver operating characteristic curve analysis, DeLong test, and decision curve analysis were adopted to compare the prediction ability and clinical efficacy of three different models. Among the six classifiers, logistic regression model was selected as the final radiomics model. Besides, the combined clinical-radiomics model exhibited a superior ability in distinguishing DCIS from breast fibromatosis to the clinical-ultrasound model and the radiomics model. The combined model by integrating clinical-ultrasound factors and radiomics features performed well in predicting DCIS, which might promote prompt interventions to improve the early diagnosis and prognosis of the patients.

Comparative Evaluation of Filters for Speckle Noise Reduction in a Clinical Liver Ultrasound Image

Comparative Evaluation of Filters for Speckle Noise Reduction in a Clinical Liver Ultrasound Image

CLINICAL AND PARACLINICAL CHARACTERISTICS OF FASCIOLIASIS PATIENTS IN THE PERIOD OF 2019 - 2023

Objectives: To describe the clinical andparaclinical characteristics of Fasciola infected patients .Subjets and methods: Cross-sectional descriptive study on 128 patients diagnosed with fascioliasis from January 2019 to June 2023 treated at National Hospital for Tropical Diseases and National Institute of Malariology, Parasitology and Entomology.Results: Clinical symptoms: Right upper quadrant pain accounted for 74.2%, followed by epigastric pain (44.5%), fatigue and loss of appetite (43.8%), weight loss (42.2%), fever (39.8%) and hepatomegaly (7%). Paraclinical findings: Hypereosinophilia accounted for 75%, elevation of AST and ALT accounted for12.5% and 24.2%, respectively. 91% of patients were positive with anti-fasciola antibodies by ELISA testing, 24% of them had fasciola’s eggs in stool. 100% of patients had liver lesions on ultrasound, of which 62.5% were found in the right liver. The lesions sized from 3 - 5,9 cm in 35.9% and from 6 - 8,9 cm in 38.3% of patients.Conclusions: Clinical and paraclinical characteristics of patients infected with fascioliasis are very diverse. Diagnosis should be based on the combination of clinical features, hypereosinophilia, fasciola antibodies and ultrasound images.

A Multiscale Deep Encoder–Decoder with Phase Congruency Algorithm Based on Deep Learning for Improving Diagnostic Ultrasound Image Quality

Ultrasound imaging is widely used as a noninvasive lesion detection method in diagnostic medicine. Improving the quality of these ultrasound images is very important for accurate diagnosis, and deep learning-based algorithms have gained significant attention. This study proposes a multiscale deep encoder–decoder with phase congruency (MSDEPC) algorithm based on deep learning to improve the quality of diagnostic ultrasound images. The MSDEPC algorithm included low-resolution (LR) images and edges as inputs and constructed a multiscale convolution and deconvolution network. Simulations were conducted using the Field 2 program, and data from real experimental research were obtained using five clinical datasets containing images of the carotid artery, liver hemangiomas, breast malignancy, thyroid carcinomas, and obstetric nuchal translucency. LR images, bicubic interpolation, and super-resolution convolutional neural networks (SRCNNs) were modeled as comparison groups. Through visual assessment, the image processed using the MSDEPC was the clearest, and the lesions were clearly distinguished. The structural similarity index metric (SSIM) value of the simulated ultrasound image using the MSDEPC algorithm improved by approximately 38.84% compared to LR. In addition, the peak signal-to-noise ratio (PSNR) and SSIM values of clinical ultrasound images using the MSDEPC algorithm improved by approximately 2.33 times and 88.58%, respectively, compared to LR. In conclusion, the MSDEPC algorithm is expected to significantly improve the spatial resolution of ultrasound images.

Management of inflammatory bowel disease and lymphangiectasia in a dog with octreotide and tranexamic acid

In dogs, IBD is categorized as a rare form of intestinal disease. Most cases are associated with lymphangiectasia. A 2-year-old, American Staffordshire Terrier male dog, presented with progressive diarrhea for more than 3 months. Clinical findings, laboratory findings, and ultrasound images were compatible with IBD and lymphangiectasia. Also, histopathological evaluation of punch biopsy samples from the intestines were compatible with IBD and lymphangiectasia. The treatment of the dog was started with the treatment protocol containing octreotide (10 µg/dog per day, BID, sc.) and tranexamic acid (10 mg/kg, BID, im.) which was used for the first time in dogs with IBD and lymphangiectasia. Fecal consistency and defecation frequency returned to normal on the 15th day, also at the end of the first month of therapy hypoalbuminemia began to normalize, ascites completely disappeared and the thickness of the lamina muscularis layer began to normalize, and the patient began to gain weight. In the controls performed at the end of the second month, it was determined that the patient's clinical signs and all laboratory results improved. So these results suggest that the combination of octreotide and tranexamic acid can effectively and without any side effects be used for the treatment of IBD in dogs. To our knowledge, this is the first case of using these two medications as a treatment protocol in dogs with this disease. We think that veterinarians should be aware of possibility to use octreotide and tranexamic acid as a treatment protocol in canine patients with IBD and lymphangiectasia.

Review on ultrasonographic diagnosis of thyroid diseases based on deep learning

In recent years, the incidence of thyroid diseases has increased significantly and ultrasound examination is the first choice for the diagnosis of thyroid diseases. At the same time, the level of medical image analysis based on deep learning has been rapidly improved. Ultrasonic image analysis has made a series of milestone breakthroughs, and deep learning algorithms have shown strong performance in the field of medical image segmentation and classification. This article first elaborates on the application of deep learning algorithms in thyroid ultrasound image segmentation, feature extraction, and classification differentiation. Secondly, it summarizes the algorithms for deep learning processing multimodal ultrasound images. Finally, it points out the problems in thyroid ultrasound image diagnosis at the current stage and looks forward to future development directions. This study can promote the application of deep learning in clinical ultrasound image diagnosis of thyroid, and provide reference for doctors to diagnose thyroid disease.

MS-CFNet: a multi-scale clinical studying-based and feature extraction-guided network for breast fibroadenoma segmentation in ultrasonography.

Segmenting breast tumors in ultrasonography is challenging due to the low image quality and presence of artifacts. Radiologists' studying and diagnosis skills are integrated with artificial intelligence to establish a clinical learning-based deep learning network in order to robustly extract and delineate features of breast fibroadenoma. The spatial local feature contrast (SLFC) module captures overall tumor contours, while the channel recursive gated attention (CRGA) module enhances edge perception through high-dimensional information interaction. Additionally, full-scale feature fusion and enhanced deep supervision are applied to improve model stability and performance. To achieve smoother boundaries, we introduce a new loss function (cosh-smooth) that penalizes and finely tunes tumor edges. Our dataset comprises 1016 clinical ultrasound images of breast fibroadenoma with labeled masks, alongside a publicly available dataset of 246 ones. Segmentation performance is evaluated using the Dice similarity coefficient (DSC) and mean intersection over union (MIOU). Extensive experiments demonstrate that our proposed MS-CFNet outperforms state-of-the-art methods. Compared to TransUNet as a baseline model, MS-CFNet improves by 1.47% in DSC and 2.56% in MIOU. The promising result of MS-CFNet is attributed to the integration of radiologists' clinical diagnosis procedure and the bionic mindset, enhancing the network's ability to recognize and segment breast fibroadenomas effectively.

The application of ultrasound in the diagnosis of small bowel volvulus in adults

Objective: To analyze the ultrasound characteristics of small bowel volvulus among adults and to investigate the value of ultrasound in the diagnosis of small bowel volvulus. Methods: Totally 34 adults with small bowel volvulus confirmed by clinical diagnosis or surgery and who underwent ultrasound examination at Peking Union Medical College Hospital from August 2017 to October 2022 were enrolled, including 19 males and 15 females, aged (55.0±21.8) years (range: 19 to 94 years). The clinical characteristics, CT images and ultrasound images of the patients were retrospectively reviewed, and the ultra, sound features of small bowel volvulus and its diagnostic efficacy were analyzed. Results: Abdominal pain was the typical clinical symptom of all patients. Other symptoms included 21 cases of abdominal distension, 19 cases of nausea and vomiting, and 13 cases of cessation of passage of stool or flatus. Eight patients had signs of peritonitis and 22 patients had abnormal bowel sounds. Twenty patients had a history of abdominal surgery. Twenty-seven patients underwent surgery for intestinal obstruction, and the remaining 7 patients improved after conservative treatment. All cases were evaluated by ultrasound, 11 cases showed a "whirl sign" and were diagnosed as small bowel volvulus, the diagnostic accuracy rate was 32.4% (11/34), ultrasound simultaneously diagnosed intestinal obstruction in 21 cases, 17 cases of abdominal effusion, 4 cases of intestinal wall thickening, 2 cases of abdominal mass, 1 case of intussusception, 1 case of right sided inguinal hernia. CT and ultrasound had a consistent positive discovery in 88.2% (30/34) of all the patients. Conclusion: Ultrasound is valuable in the diagnosis of small bowel volvulus and the evaluation of complications.

Factors associated with postoperative muscle reconnection in children's congenital muscular torticollis.

In prior studies, there has been no report of clinical observation of postoperative reconnection of the sternocleidomastoid muscle (SCM) in children with congenital muscular torticollis (CMT). Therefore, the objective of this study is to investigate the factors associated with postoperative reconnection of the SCM in children with CMT, and to provide clinical evidence. A retrospective study was conducted, wherein 83 CMT children without any missing data were followed up from November 2019 to June 2021. The age at the time of surgery, sex, preoperative and postoperative follow-up duration, laterality, neck mass history, preoperative physical therapy history, and severity type were recorded. The severity classification of CMT was based on clinical features and ultrasound images of SCM. The postoperative reconnection of SCM was measured. Out of 83 patients, ten had postoperative reconnection. The rate of postoperative reconnection of SCM in children with CMT who had undergone unipolar SCM release surgery was 18.994 times higher than in patients who had not undergone such surgery. This difference was statistically significant [odds ratio (OR) =18.994, 95% confidence interval (CI): 1.583 to 227.897, P=0.020]. The history of SCM release surgery in CMT children can predict the postoperative reconnection of SCM, which will aid in determining the optimal surgical approach for recurrent CMT patients.

Clinical and ultrasound features of pseudomyxoma peritonei and its histopathological subtypes among women seen at a Philippine tertiary hospital: A 10-year review

OBJECTIVES: Pseudomyxoma peritonei (PMP) in women arises from an appendiceal or ovarian pathology and presents either of two histological subtypes of differing prognosis, disseminated peritoneal adenomucinosis (DPAM), or peritoneal mucinous carcinomatosis (PMCA). This study aimed to evaluate the demographic, clinical, and sonographic features among affected women and the differences between the two histological subtypes. METHODS: A retrospective study was conducted involving 36 women with histopathological diagnosis of PMP who had preoperative ultrasound and underwent surgery at the department of obstetrics and gynecology in a tertiary hospital. Demographic and clinical data, ultrasound images and reports, and final histopathology were reviewed. To compare the subtypes, one-way analysis of variance for continuous data and Chi-square/Fisher exact test for categorical data were used, with P &lt; 0.05 indicating statistical significance. RESULTS: Patients were mostly &gt;50 years of age, multigravid, and presented with abdominal distention. Ultrasound examinations consistently showed amorphous, mixed echo or echogenic ascites, peritoneal thickening, and omental caking. Adnexal/ovarian masses were detected in 66.7% of cases. Omental caking was significantly more prevalent in PMCA (83.3%; P = 0.0002), whereas larger ovarian tumors (&gt;20 cm) and papillarities were more common in DPAM (both 92.9%; P = 0.0005). Most patients underwent gynecologic surgery (n = 31; 86.1%), and 14 (38.9%) required readmission due to recurrence. The final histopathology revealed largest tumor involvement of the appendix (n = 13; 36.1%), the ovaries (n = 11; 30.5%), or undetermined (n = 12; 33.3%). CONCLUSIONS: Preoperative diagnosis of PMP is possible based on its clinical and ultrasound features. Although the subtypes are similar in most of these features, certain ultrasound findings may aid in distinguishing them.

Deep-learning based segmentation of ultrasound adipose image for liposuction.

To develop an automatic and reliable ultrasonic visual system for robot- or computer-assisted liposuction, we examined the use of deep learning for the segmentation of adipose ultrasound images in clinical and educational settings. To segment adipose layers, it is proposed to use an Attention Skip-Convolutions ResU-Net (Attention SCResU-Net) consisting of SC residual blocks, attention gates and U-Net architecture. Transfer learning is utilised to compensate for the deficiency of clinical data. The Bama pig and clinical human adipose ultrasound image datasets are utilized, respectively. The final model obtains a Dice of 99.06±0.95% and an ASD of 0.19±0.18mm on clinical datasets, outperforming other methods. By fine-tuning the eight deepest layers, accurate and stable segmentation results are obtained. The new deep-learning method achieves the accurate and automatic segmentation of adipose ultrasound images in real-time, thereby enhancing the safety of liposuction and enabling novice surgeons to better control the cannula.