Australasian Journal of Ultrasound in Medicine最新文献

Objectives

Accurate diagnosis of pathology from ultrasound images is reliant upon images of a suitable diagnostic quality being acquired. This study aimed to create a novel machine learning model to automatically assess transvaginal ultrasound (TVUS) image quality for gynaecological ultrasound.

Method

Six imaging professionals (two sonographers, two gynaecological sonologists and two radiologists) assigned a quality score to 150 TVUS images from 50 cases (50 uterus images and 100 ovary images). Images were given a score of 1–4 (1—reject/image inaccurate, 2—poor quality, 3—suboptimal quality or 4—optimal quality). As variation existed between the scores assigned by the labellers, we framed this problem as a multi-annotator noisy label problem. To address this, a new machine learning architecture was developed, combining a weighted ensemble algorithm to estimate consensus labels and a multi-axis vision transformer (MaxViT) to handle noisy labels, improving model accuracy in predicting image quality. Forty cases (120 images) were used for model training, while the remaining 10 cases (30 images) were reserved as a test set for model evaluation.

Results

The novel machine learning architecture we created was able to successfully determine image quality with a validation accuracy of 80% and a macro average recall of 77%. This significantly improved upon the 57% accuracy of the baseline machine learning method (ResNet50). The MaxViTs were able to outperform human performance in most cases, with an accuracy of 80% surpassing four of the six human labellers.

Conclusions

This novel machine learning model offers an automated method of assessing the quality of TVUS images. The tool has the potential to provide real-time feedback to those performing TVUS, reduce the need for repeated imaging, and improve the diagnosis of gynaecological pathology.

Background

Neuroblastoma, a rare embryonic tumour of neural crest origin, is one of the most common paediatric malignancies, accounting for 8%–10% of childhood cancers. It most often arises in the adrenal glands and can present in utero during the third trimester.

Key Findings

This case report presents a 29-year-old primiparous woman whose foetus was found to have bilateral adrenal masses during a routine 37-week ultrasound. The masses were initially suspected to be renal in origin, but subsequent imaging revealed bilateral adrenal involvement, leading to the diagnosis of stage 4S neuroblastoma. Postnatal ultrasound confirmed solid and cystic masses in both adrenal glands, with additional liver involvement.

Discussion

This case emphasises the importance of antenatal imaging in the early detection of neuroblastoma and highlights the role of ultrasound in identifying heterogeneous, hyperechoic suprarenal masses, which can be key indicators of the condition. The diagnostic dilemma in this case relates to the masses mimicking renal lesions, highlighting the importance of detailed ultrasound imaging of foetal abnormalities. In this case, the tumour's presence was confirmed through further imaging and biochemical testing, and the baby received specialised care at a tertiary centre. Liver metastases often assist in confirming the diagnosis of stage 4S neuroblastomas, which generally have a favourable prognosis, with many cases regressing spontaneously. Early detection, particularly via prenatal imaging, significantly improves outcomes, as it allows for careful monitoring and timely intervention. The baby in this case has responded well to treatment, underscoring the importance of early diagnosis and ongoing postnatal surveillance.

Objective

Evaluate the effectiveness of a point-of-care ultrasound (POCUS) training programme for public sector general practitioners in The Republic of Somaliland.

Methods

This retrospective, observational study analysed 500 ultrasound scans of adult patients from four public hospitals. Twelve general practitioners with no prior non-obstetric ultrasound experience completed a 12-day, multi-organ training curriculum, then performed scans in the emergency department, ICU and outpatient clinics. Participants submitted their scans with a questionnaire providing their answer to one or more (out of 21 total) specific diagnostic questions, termed ‘studies’, answered by ultrasound. Blinded experts assessed image quality and accuracy. Concordance between learner and expert was calculated for each organ system alongside interrater reliability for each of the individual studies.

Results

Overall concordance between expert and learner was 90% (n = 500). Concordance was highest for renal (95%, n = 242), lung (91%, n = 365) and hepatobiliary (91%, n = 217) imaging, whereas cardiac (87%, n = 400) and DVT (83%, n = 42) concordance was lower. ICU studies had the highest concordance (91%), followed by those conducted in the clinic (89%) and emergency department (87%). Cohen's kappa coefficient between learner and expert interpretation was > 0.8 (p value < 0.05) for 18 out of 18 studies, indicating at least substantial agreement.

Conclusion

This POCUS training programme in The Republic of Somaliland resulted in high concordance between learner and expert interpretations, indicating effective training across multiple organ systems. Several areas for improvement were identified. These findings underscore POCUS's potential to enhance diagnostic capabilities in resource-limited settings.

Objectives

Point-of-care ultrasound (POCUS) is a bedside imaging technique increasingly taught to physicians and medical students. Beginners often face challenges with probe orientation and hand–eye coordination. Extended reality (XR) can enhance POCUS education by projecting a 3D anatomical model onto a phantom or patient, providing unlimited training and direct feedback on probe orientation and image acquisition. We developed Augmedicine: Ultrasound, an XR application for training point-of-care echocardiography. This study evaluates the face and content validity of the application.

Methods

Three groups (novices, intermediates and experts in ultrasonography) were invited to complete tasks using the application. Following these tasks, all participants filled out a questionnaire to rate realism, didactic value, and usability of the application using a 5-point Likert-scale and open-ended questions.

Results

A total of 61 participants (43 novices, 10 intermediates, 8 experts) completed the evaluation. The application was recognised as a valuable tool for novice training, with 95% indicating that it is useful to enhance spatial understanding and 77% for improving hand–eye coordination. Anatomical accuracy was rated positively by 90% of participants, and 69% found the application easy to use. Experts highlighted its potential for expansion, such as inclusion of procedural skills and pathology training.

Conclusion

Our XR ultrasound application demonstrated face and content validity for echocardiography training. It holds promise as a complementary tool to traditional methods, supporting novice learners by improving spatial understanding and reducing reliance on instructors and test subjects. Future iterations will address technical optimisations, improve realism, and evaluate the application's learning effect.

Background

Extramammary Paget's disease (EMPD) is a rare skin cancer that arises in areas rich in apocrine glands, such as the genital, perineal, and perianal regions. The gold standard for diagnosis is the histopathological study of a biopsy, and the standard treatment is wide surgical excision. Once the histopathological diagnosis of EMPD has been established, high frequency ultrasound (HFUS) has shown potential as a complementary non-invasive method useful for the initial assessment and EMPD's management. HFUS provides information about lesion extent, thickness, depth of invasion and vascularity throughout, as well as lymph node metastasis. We report two cases of EMPD in which HFUS proved useful for both diagnosis and treatment monitoring.

Key Findings

Case 1: A 72-year-old woman with perianal EMPD who was not a candidate for surgery due to the risk of loss of sphincter function. Initial photodynamic therapy (PDT) resulted in partial improvement, and HFUS revealed persistent disease, leading to a switch to topical imiquimod, which was ultimately followed by CO₂ laser treatment after further HFUS-proven progression.

Case 2: A 70-year-old man with inguinal EMPD underwent HFUS imaging, which showed epidermal involvement without dermal invasion throughout the entire lesion. Surgical excision was performed, confirming the HFUS findings.

Discussion

These cases highlight the role of HFUS as a complementary tool to biopsy for the initial assessment and follow-up of EMPD, especially when surgery is contraindicated or must be carefully planned. HFUS provides a detailed assessment of lesion characteristics critical for prognosis and therapy adjustments. The non-invasive nature, cost-effectiveness and ability to monitor treatment response make it a valuable adjunct to traditional diagnostics and to guide individualised management.

Purpose

To determine the effects of acquisition parameters on measured shear wave velocity (SWV) in reference phantoms.

Methods

SWV were measured on three phantoms of certified SWV 0.794 m/s (Phantom 1), 1.711 m/s (Phantom 2) and 2.500 m/s (Phantom 3). Parameters examined were depth, location, region of interest (ROI) size, transducer type, pitch, frame rate and one-shot mode. Measurements were repeated 10 times for different levels of each parameter at 1.8 MHz and 2.5 MHz transducer frequency on both Canon Aplio i700 and i800 systems. Statistical analyses were performed using linear mixed models with Holm test for multiple comparisons.

Results

There was an interaction between acquisition depth and phantom stiffness (p < 0.001) with 1 cm of depth increasing SWV by 0.037 and 0.012 m/s in Phantoms 1 and 2 but decreasing SWV by −0.049 m/s in phantom 3. Lateral acquisition box location had increased SWV compared to the middle location (left 0.048 m/s, right 0.065 m/s, p < 0.001). ROI size (10 mm, 0.002 m/s, p = 0.846) had no statistically significant effect on SWV. Transducer type at 2.5 cm depth had no significant effect on SWV (linear 0.038 m/s, p = 0.049). Diff pitch setting of 2 had lower SWV (−0.030 m/s, p = 0.002). Frame rate had no statistically significant effect on SWV (1 frame per second increasing SWV by 0.037 m/s, p = 0.025). One-shot mode had higher SWV than continuous mode by 0.042 m/s (p < 0.001).

Conclusion

Acquisition parameters can have statistically significant effects on measured SWV and may bias measured SWV due to cumulative effects.

Background

Ultrasound is routinely used in the diagnosis of renal calculi and resulting hydronephrosis. However, its utility in the diagnosis and management of urethral calculi is less well understood. The ability to rapidly diagnose urethral calculi using point of care ultrasound (POCUS) would be useful in the emergency department (ED) to improve patient outcomes and avoid more invasive investigations.

Key Findings

This case report presents the occurrence of an obstructed urethral calculus in a 42-year-old male, where POCUS facilitated a prompt and accurate diagnosis. The ability to characterise the size, shape and location of the calculus with the urethra enabled the treating clinicians to avoid further investigations and rapidly treat the obstruction. The stone was removed using artery forceps by the urology team in the ED and POCUS was used afterwards to ensure complete removal and resolution of urinary tract obstruction. Overall, the use of POCUS contributed to the patient's safe management, avoiding radiation exposure and minimising the length of stay in the ED.

Discussion

POCUS is a useful tool for evaluating suspected urethral stones to assess for location, size and signs of urinary tract obstruction.

Background

A 27-year-old nulliparous woman with a history of subfertility presented to a routine dating scan at 13 + 1 weeks gestation with a history of left-sided abdominal pain.

Key Findings

Both standard transabdominal ultrasound and handheld point-of-care ultrasound (POCUS) identified a probable abdominal ectopic pregnancy (AEP), confirmed by MRI. Intraoperative findings revealed a live pregnancy encapsulated by the omentum with associated haemoperitoneum. Active haemorrhage necessitated conversion from laparoscopy to laparotomy, culminating in a left salpingo-oophorectomy.

Discussion

This case highlights the potential role of handheld POCUS in diagnosing a rare and life-threatening condition. The device enabled immediate, bedside imaging and facilitated rapid decision-making. To our knowledge, this is the first documented case of an abdominal ectopic pregnancy diagnosed using handheld POCUS, emphasising its potential to improve maternal outcomes through early detection and intervention.

Introduction

Point-of-care ultrasound (POCUS) can be very useful in low-resource settings, where access to radiology is often limited. The PEARLS (POCUS in Emergency and Acute care in Resource-Limited Settings) program is a scalable, innovative approach to POCUS training for clinicians in low- and middle-income countries (LMICs), where access to imaging modalities is frequently limited.

Methods

A review of the PEARLS program was conducted using participant data that were gathered from participants in Vanuatu, Tonga, and Papua New Guinea (PNG), tracking the number and type of scans uploaded and overall engagement with the training. The PEARLS program begins with a three-day in-person course focused on hands-on practice using portable ultrasound devices. Clinicians are provided with devices and guided on how to store images securely online. After the course, each participant is paired with an experienced coach who offers ongoing remote support through regular image reviews and a series of follow-up lectures. One year later, the program team revisits the site, and selected participants are transitioned into faculty roles to help train others in a train-the-trainer model.

Results

Across 40 participants, a total of 576 scans were uploaded following training. The extended FAST exam was the predominant modality, followed by critical care echocardiography and first-trimester pregnancy assessments. Although remote mentoring maintained high levels of engagement, only one participant achieved the full competency benchmark of 25 uploaded scans.

Discussion

These findings illustrate that a longitudinal training model with remote mentorship can improve POCUS proficiency in LMIC settings. However, challenges such as inconsistent internet access and equipment sharing persist, potentially limiting sustained competency.

Conclusion

The PEARLS program presents a scalable, sustainable framework for POCUS education in resource-limited environments, fostering local expertise and enhancing acute care diagnostics.

Introduction

Ulnar-sided wrist pain is a common clinical problem, which can be difficult to diagnose, manage and treat. Ultrasound imaging can be used to identify structural changes which may contribute to the pain. Knowledge of the associated relative anatomy and intertwining structures which form the triangular fibrocartilage complex (TFCC) and distal radio-ulnar joint (DRUJ), sonographic techniques to image these structures and their normal and abnormal appearances can be underappreciated and are required.

Methods

A literature search was conducted to review the current knowledge related to sonographic assessment of the TFCC and DRUJ.

Results

The intertwining components of the DRUJ and TFCC which can be demonstrated sonographically are unpacked with clear supporting figures and videos.

Discussion

The components of the TFCC, which include the disc proper, meniscus homologue, extensor carpi ulnaris (ECU) sub-sheath, palmar-sided ulno-carpal ligaments and the palmar and dorsal distal radio-ulnar ligaments, can all be individually demonstrated sonographically, in addition to structures which comprise the DRUJ. Normal sonographic images are presented.

Conclusion

Improved understanding of the sonographic anatomy, technique and normal imaging appearances of ulnar-sided wrist structures can enhance the quality of imaging and subsequently the diagnosis of structural causes of pain, which can guide patient management.