Journal of Medical Ultrasonics最新文献

Purpose: Sonazoid contrast-enhanced ultrasound (CEUS) offers valuable diagnostic information on hepatic lesions, but it is time-consuming. In this study, we investigated a novel composite machine-learning framework that integrates heuristic knowledge and model-specific classification to differentiate liver lesions using only the first 2 min of CEUS imaging.

Methods: CEUS images from 232 patients with 232 focal liver lesions (benign: 61, hepatocellular carcinoma [HCC]: 104, non-HCC malignancies [ML]: 67) were analyzed. For each case, six frames from injection to peak enhancement and static images at 1 and 2 min post-injection were used. Two deep learning models were developed: Model 1 classified heterogeneous enhancement patterns into "benign," "HCC," "ML," or "Uniform" (homogeneous). Model 2 further classified "Uniform" cases into three diagnostic categories. Lesion brightness values were incorporated as input features. The artificial intelligence (AI) mode was also evaluated by observer study of three hepatologists using the area under the receiver operating characteristic curve (AUC).

Results: The composite model was evaluated on 58 independent test cases, achieving classification accuracy of 81.8% for benign, 93.5% for HCC, and 68.8% for ML, with an overall accuracy of 84.5%. Binary classification (benign vs. malignant) yielded 97.9% sensitivity, 94.8% specificity, and 94.8% overall accuracy. For discrimination between benign and malignant, the mean AUC for the three observers was significantly improved with AI output, where the difference in AUC (95% confidence interval) was 0.095 (0.0197, 0.1703) (P = 0.013).

Conclusions: The proposed AI-based framework enables accurate liver lesion classification using early phase CEUS, eliminating the need for Kupffer-phase imaging in many cases.

Purpose: Attenuation techniques have been developed to quantify liver fat, improving diagnostic accuracy. However, the standardization of measurement methods remains undefined, limiting clinical application. This study evaluated the effect of different region of interest (ROI) settings on the attenuation coefficient (AC) and coefficient of determination (R²) in attenuation imaging (ATI).

Methods: We analyzed 5212 ultrasound examinations performed at our institution. Two periods were compared: early [ROI positioned at twice the subcutaneous thickness; double positioning (DP), 2 × 3 cm] and late [ROI placed beneath the multireflection layer; color positioning (CP), 2 × 4 cm]. Additionally, AC and R² were analyzed in 331 patients with stored raw data across four conditions: DP and CP for positioning, and 2 × 3 cm and 2 × 4 cm for size, respectively, using Bland-Altman analysis.

Results: The early and late periods included 2881 and 2331 examinations, respectively. Patient characteristics were comparable. Significant differences were observed in the AC (median, 0.58 vs. 0.56 dB/cm/MHz), steatotic liver diagnostic rates based on AC, R² (median, 0.87 vs. 0.93), and proportion of R² ≥ 0.90. Raw data analysis showed slight differences in AC and R² according to ROI position; however, there were significant differences in ROI size (AC: 2 × 3 > 2 × 4; R²: 2 × 4 > 2 × 3).

Conclusion: In ATI, the larger ROI improved R², whereas the smaller ROI yielded higher AC values, suggesting that ROI size affects diagnostic reliability. Standardizing ATI protocols is essential, with uniformity in ROI size being a priority.

Purpose: To evaluate the dynamics of the talus-lateral malleolus distance during landing and compare the results between individuals with chronic ankle instability (CAI) and healthy adults.

Methods: Participants were divided into the healthy (12 adults, 12 feet) and CAI (12 adults, 12 feet) groups. Participants performed forward single-leg drop landing from a 30-cm-high box. Through the synchronization of ultrasound with a motion capture system, measurements of the talus-lateral malleolus distance and ankle joint angles were obtained during forward single-leg drop landing. The analysis interval was from 200 ms before to 200 ms after initial contact; the time was normalized to 100%. Statistical parametric mapping was employed to investigate differences in the temporal changes of the talus-lateral malleolus distances and ankle joint angles during landing between groups.

Results: A significantly longer talus-lateral malleolus distance was observed before initial contact in the CAI group (1-49%) as compared with the healthy group (p = 0.000). In addition, the ankle adduction angle before initial contact (12-61%) was larger in the CAI group than in the healthy group (p = 0.005).

Conclusion: Our findings suggest that the talus-lateral malleolus distance increases along with the ankle adduction movement in individuals with CAI, especially before initial contact, potentially indicating high mechanical ankle instability. This study established an innovative measurement system for quantitative evaluation of the dynamics of the talus and lateral malleolus during landing. The findings of this study may contribute to a more detailed understanding of pathology and features of movement in patients with CAI.

Purpose: We previously reported a manipulation technique for frozen shoulder using ultrasound-guided cervical nerve root block. Although the results were excellent, we considered that the range of external rotation could be improved. In the present study, we aimed to confirm the effectiveness of our modified manipulation technique for further improving the range of external rotation at the shoulder joint.

Methods: We retrospectively searched our electronic medical records to identify patients who underwent manipulation for frozen shoulder at our institution between January 2024 and June 2024. A total of 173 shoulders of 169 patients (mean age 53.8 ± 6.9 years) were included. Three changes were made to our previous method: a change in the method for intra-articular injection of the glenohumeral joint, addition of C7 to the nerve block, and changes in the manual manipulation technique. Range of motion (ROM) at the shoulder was measured before manipulation and at 1 week, 1 month, 2 months, and 3 months after the procedure. The results were compared with those in our previous report.

Results: Mean range of external rotation (95% confidence interval) was 11.5° (9.5-13.5, p = 0.51) before manipulation, 56.2° (52.9-59.5, p < 0.001) at 1 week, and 58.0° (53.9-62.7; p < 0.001) at 3 months. The results for range of external rotation were significantly better than in the previous report. There was no significant difference in ROM in terms of flexion and abduction between the two studies.

Conclusion: Range of external rotation at the shoulder joint was significantly improved by use of the revised manipulation technique.

Purpose: Epidural anesthesia under real-time ultrasound guidance has been studied to enhance puncture safety. However, in the thoracic spine, where the anatomy is complex and the interlaminar space is narrow, improved image quality is necessary. Because imaging quality is degraded in bones that are greatly tilted relative to the probe, we aimed to improve the imaging quality by steering the transmitted ultrasonic beam.

Methods: We investigated a method to improve bone visibility by selecting appropriate transmit steering angles and receiving aperture ranges for each of the two bones forming the interlaminar space, and by compensating for the directivity of the receiving elements.

Results: The proposed method emphasized the delineation of a bone phantom near the gap corresponding to the puncture site. Furthermore, when comparing the contrast ratio in the region of interest of the bone and gap areas, the proposed method showed an improvement of 24.2 dB on the left side and 4.2 dB on the right side compared to the conventional B-mode image.

Conclusion: The proposed method was confirmed to emphasize the bones and is expected to improve the image quality of the thoracic spine gap.

Ultrasound echo signals from organs contain various physical information and numerous studies have been conducted to extract new diagnostic insights from these signals. This article reviews the research and development I have undertaken in diagnosing liver tissue characteristics, reflecting on its historical progression. Initially, texture analysis aimed to identify features of liver fibrosis, but it was found to be more effective for quantifying fatty liver. Consequently, the analysis algorithm and product specifications were refined to address this need. For fatty liver, the importance of early detection of non-alcoholic steatohepatitis (NASH) has grown, leading to investigations into quantifying the liver's attenuation coefficient. After several prototype trials, a real-time quantitative measurement function was achieved.

Purpose: Posterior interosseous nerve (PIN) syndrome is an uncommon neuropathy that may mimic lateral epicondylitis. High-resolution ultrasound aids its diagnosis by revealing nerve swelling, while ultrasound-guided injection is increasingly employed for decompression. Nevertheless, the influence of injection axis on injectate distribution remains unclear. This study aimed to compare the spread characteristics and accuracy of short-axis vs. long-axis ultrasound-guided injections at the proximal and distal PIN using cadaveric validation.Kindly check and confirm the city name is correctly identified in affiliation [6].It is correctly identified.  METHODS: Ten cadaveric limbs were randomized to short-axis (n = 5) or long-axis (n = 5) injections. Ultrasound parameters, including fascicle count, cross-sectional area, and surrounding muscle or tendon thickness, were recorded. Proximal and distal injections were performed separately with 5 mL of injectate. Subsequent dissection assessed target infiltration and spread dimensions (length, width).

Results: Ultrasound findings were comparable between the groups with no significant differences in cross-sectional area or muscle/tendon thickness. On dissection, all injections achieved successful proximal and distal PIN infiltration. At the proximal level, short-axis injections produced significantly greater infiltration width (43.6 ± 6.2 mm vs. 24.3 ± 13.2 mm, p = 0.032), while infiltration length was not significantly different (88.8 ± 17.4 mm vs 77.1 ± 37.1 mm, p = 0.690). At the distal level, both infiltration length (81.0 ± 24.4 mm vs. 67.7 ± 37.4 mm, p = 0.548) and width (28.2 ± 10.9 mm vs. 22.5 ± 8.8 mm, p = 0.548) were numerically greater with short-axis injections, although the differences did not reach statistical significance.

Conclusion: Both short- and long-axis ultrasound-guided injections consistently achieved target infiltration of the PIN. The short-axis approach provided broader proximal spread, which may enhance circumferential perineural coverage, whereas the distal injections showed no significant differences between techniques. These findings support the reliability of both approaches, while highlighting a potential advantage of short-axis guidance at the proximal forearm.