Detection of COVID-19 Using Deep Learning Algorithms on Chest Radiographs.

PURPOSE: To evaluate the performance of a deep learning (DL) algorithm for the detection of COVID-19 on chest radiographs (CXR). MATERIALS AND METHODS: In this retrospective study, a DL model was trained on 112,120 CXR images with 14 labeled classifiers (ChestX-ray14) and fine-tuned using initial CXR on hospital admission of 509 patients, who had undergone COVID-19 reverse transcriptase-polymerase chain reaction (RT-PCR). The test set consisted of a CXR on presentation of 248 individuals suspected of COVID-19 pneumonia between February 16 and March 3, 2020 from 4 centers (72 RT-PCR positives and 176 RT-PCR negatives). The CXR were independently reviewed by 3 radiologists and using the DL algorithm. Diagnostic performance was compared with radiologists' performance and was assessed by area under the receiver operating characteristics (AUC). RESULTS: The median age of the subjects in the test set was 61 (interquartile range: 39 to 79) years (51% male). The DL algorithm achieved an AUC of 0.81, sensitivity of 0.85, and specificity of 0.72 in detecting COVID-19 using RT-PCR as the reference standard. On subgroup analyses, the model achieved an AUC of 0.79, sensitivity of 0.80, and specificity of 0.74 in detecting COVID-19 in patients presented with fever or respiratory systems and an AUC of 0.87, sensitivity of 0.85, and specificity of 0.81 in distinguishing COVID-19 from other forms of pneumonia. The algorithm significantly outperforms human readers (P<0.001 using DeLong test) with higher sensitivity (P=0.01 using McNemar test). CONCLUSIONS: A DL algorithm (COV19NET) for the detection of COVID-19 on chest radiographs can potentially be an effective tool in triaging patients, particularly in resource-stretched health-care systems.

Ultrasound-MRI Fusion for Targeted Biopsy of Myopathies.

OBJECTIVE. The purpose of this article is to describe the use of ultrasound-MRI fusion imaging to guide precise and targeted muscle biopsy in patients with suspected myopathies. CONCLUSION. Ultrasound-MRI fusion-guided muscle biopsy allows targeted sampling of tissues with active inflammatory changes and facilitates diagnosis of myopathies.

Opportunistic screening for osteoporosis in abdominal computed tomography for Chinese population.

This study assessed the possibility of diagnosing and excluding osteoporosis with routine abdominal CT scans in a Chinese population who underwent both DXA and CT for unrelated reasons. Statistical correlation was made between the HU measured of the spine on CT and various parameters on DXA. Diagnostic cutoff points in terms of HU were established for the diagnosis (≤ 136 HU) and exclusion (≥ 175 HU) of osteoporosis on sagittal reformatted images. There was excellent positive and negative predictive value for the DXA-defined diagnostic subgroups and were also comparable with previous studies in Caucasian populations. The authors exhort radiologists to report these incidental findings to facilitate early detection and treatment of osteoporosis in unsuspecting patients to prevent fractures and related complications. PURPOSE: The suspicion for osteoporosis can be raised in diagnostic computed tomography of the abdomen performed for other indications. We derived cutoff thresholds for the attenuation value of the lumbar spinal vertebrae (L1-5) in Hounsfield units (HU) in a Chinese patient population to facilitate implementation of opportunistic screening in radiologists. METHODS: We included 109 Chinese patients who concomitantly underwent abdominal CT and dual X-ray absorptiometry (DXA) within 6 months between July 2014 and July 2017 at a university hospital in Hong Kong. Images were retrospectively reviewed on sagittal reformats, and region-of-interest (ROI) markers were placed on the anterior portion of each of the L1-L5 vertebra to measure the HU. The mean values of CT HU were then compared with the bone mineral density (BMD) and T-score obtained by DXA. Receiver operator characteristic (ROC) curves were generated to determine diagnostic cutoff thresholds and their sensitivity and specificity values. RESULTS: The mean CT HU differed significantly (p < 0.01) for the three DXA-defined BMD categories of osteoporosis (97 HU), of osteopenia (135 HU), and of normal individuals (230 HU). There was good correlation between the mean CT HU and BMD and T-score (Pearson coefficient of 0.62 and 0.61, respectively, p < 0.001). The optimal cutoff point for exclusion of osteoporosis or osteopenia was HU ≥ 175 with negative predictive value as 98.9% and with area under curve (AUC) of ROC curve as 0.97. The optimal cutoff point for diagnosis of osteoporosis was HU ≤ 136 with positive predictive value as 81.2% and with AUC of ROC curve as 0.86. CONCLUSION: This is the first study on osteoporosis diagnosis with routine CT abdominal scans in Chinese population. The cutoff values were comparable with previous studies in Caucasian populations suggesting generalizability. Radiologists should consider routinely reporting these opportunistic findings to facilitate early detection and treatment of osteoporosis to prevent fractures and related complications.