A Deep Learning-Based Model for Classifying Osteoporotic Lumbar Vertebral Fractures on Radiographs: A Retrospective Model Development and Validation Study.

Early diagnosis and initiation of treatment for fresh osteoporotic lumbar vertebral fractures (OLVF) are crucial. Magnetic resonance imaging (MRI) is generally performed to differentiate between fresh and old OLVF. However, MRIs can be intolerable for patients with severe back pain. Furthermore, it is difficult to perform in an emergency. MRI should therefore only be performed in appropriately selected patients with a high suspicion of fresh fractures. As radiography is the first-choice imaging examination for the diagnosis of OLVF, improving screening accuracy with radiographs will optimize the decision of whether an MRI is necessary. This study aimed to develop a method to automatically classify lumbar vertebrae (LV) conditions such as normal, old, or fresh OLVF using deep learning methods with radiography. A total of 3481 LV images for training, validation, and testing and 662 LV images for external validation were collected. Visual evaluation by two radiologists determined the ground truth of LV diagnoses. Three convolutional neural networks were ensembled. The accuracy, sensitivity, and specificity were 0.89, 0.83, and 0.92 in the test and 0.84, 0.76, and 0.89 in the external validation, respectively. The results suggest that the proposed method can contribute to the accurate automatic classification of LV conditions on radiography.

Refined ultrasonographic criteria for sinusoidal obstruction syndrome after hematopoietic stem cell transplantation.

Hepatic sinusoidal obstruction syndrome (SOS)/veno-occlusive disease is a life-threatening complication after hematopoietic stem cell transplantation (HSCT). We previously reported the efficacy of the Hokkaido Ultrasonography (US)-based scoring system (HokUS-10) for US findings. To establish easier-to-use criteria, we retrospectively evaluated US findings from 441 patients, including 30 patients with SOS using the HokUS-10 scoring system. Using logistic regression analysis, we established the novel diagnostic criteria HokUS-6. In the presence of ascites, US diagnosis was made in the presence of two of the following 6 parameters: moderate amount of ascites, the appearance of a paraumbilical vein blood flow signal, gallbladder wall thickening, portal vein dilatation, portal vein velocity decrease, and hepatic artery resistive index increase. The AUC, sensitivity, and specificity of HokUS-6 were 0.974 (95% confidence interval 0.962-0.990), 95.2%, and 96.9%, respectively. The scores were significantly higher in patients with severe SOS than in those with non-severe SOS (p = 0.013). Furthermore, the scores before HSCT were significantly higher in patients who developed SOS than in controls (p = 0.001). The HokUS-6 is an easy and useful way to diagnose and identify the risk of SOS.

Power Doppler signal calibration in the finger joint between two models of ultrasound machine: a pilot study using a phantom and joints in patients with rheumatoid arthritis.

Background Despite the advantages of ultrasound (US) in the management of rheumatoid arthritis (RA) patients, power Doppler (PD) US may be highly dependent on the type of US machine used. Purpose To present a method to calibrate the PD signal of two models of US machines by use of a flow phantom and finger joints of patients with RA. Material and Methods For the phantom study, the PD signal count was measured in the flow phantom perfusing blood mimicking fluid at various injection rates and pulse repetition frequencies (PRFs). The quantitative PD index was calculated with ImageJ. For the clinical study, the second and third metacarpophalangeal joints of five consecutive patients with RA were examined. The quantitative PD index was measured at various PRFs by use of two models of machine (the same models as the phantom study). Results For the phantom and clinical studies, negative correlations were found between the PRF and the quantitative PD index when the flow velocity was constant and positive correlations between flow velocity and the quantitative PD index at constant PRF. There was a significant difference in the depiction performance of synovial blood flow between the two models, which can be calibrated by adjusting the PRF values derived from the phantom study in each model. Conclusion Signal calibration of pannus vascularity between US machines may be possible by adjusting the PRF value according to flow phantom data. Different US machines can thus provide equivalent examination results concerning the pannus vascularity.

Power Doppler signal calibration between ultrasound machines by use of a capillary-flow phantom for pannus vascularity in rheumatoid finger joints: a basic study.

Ultrasound allows the detection and grading of inflammation in rheumatology. Despite these advantages of ultrasound in the management of rheumatoid patients, it is well known that there are significant machine-to-machine disagreements regarding signal quantification. In this study, we tried to calibrate the power Doppler (PD) signal of two models of ultrasound machines by using a capillary-flow phantom. After flow velocity analysis in the perfusion cartridge at various injection rates (0.1-0.5 ml/s), we measured the signal count in the perfusion cartridge at various injection rates and pulse repetition frequencies (PRFs) by using PD, perfusing an ultrasound micro-bubble contrast agent diluted with normal saline simulating human blood. By use of the data from two models of ultrasound machines, Aplio 500 (Toshiba) and Avius (Hitachi Aloka), the quantitative PD (QPD) index [the summation of the colored pixels in a 1 cm × 1 cm rectangular region of interest (ROI)] was calculated via Image J (internet free software). We found a positive correlation between the injection rate and the flow velocity. In Aplio 500 and Avius, we found negative correlations between the PRF and the QPD index when the flow velocity was constant, and a positive correlation between flow velocity and the QPD index at constant PRF. The equation for the relationship of the PRF between Aplio 500 and Avius was: y = 0.023x + 0.36 [y = PRF of Avius (kHz), x = PRF of Aplio 500 (kHz)]. Our results suggested that the signal calibration of various models of ultrasound machines is possible by adjustment of the PRF setting.