Machine learning methods offer great promise for fast and accurate detection and prognostication of coronavirus disease 2019 (COVID-19) from standard-of-care chest radiographs (CXR) and chest computed tomography (CT) images. Many articles have been published in 2020 describing new machine learning-based models for both of these tasks, but it is unclear which are of potential clinical utility. In this systematic review, we consider all published papers and preprints, for the period from 1 January 2020 to 3 October 2020, which describe new machine learning models for the diagnosis or prognosis of COVID-19 from CXR or CT images. All manuscripts uploaded to bioRxiv, medRxiv and arXiv along with all entries in EMBASE and MEDLINE in this timeframe are considered. Our search identified 2,212 studies, of which 415 were included after initial screening and, after quality screening, 62 studies were included in this systematic review. Our review finds that none of the models identified are of potential clinical use due to methodological flaws and/or underlying biases. This is a major weakness, given the urgency with which validated COVID-19 models are needed. To address this, we give many recommendations which, if followed, will solve these issues and lead to higher-quality model development and well-documented manuscripts.

The objective of this study is to estimate the value of diffusion-weighted MRI (DWI) in the detection of abnormalities in sacroiliac joints in the patients with early ankylosing spondylitis (AS) and investigate the feasibility of whole-body DWI (WB-DWI) in systemic evaluation of AS. Sixteen patients with early AS, 18 patients with simple low back pain (LBP), and 18 healthy volunteers were involved in this study. All subjects underwent conventional MRI and DWI. Apparent diffusion coefficient (ADC) in subchondral bone marrows of sacroiliac joints was measured. Independent-sample t test was used to statistically analyze the difference of ADC values between groups. WB-DWI was performed in additional 12 patients with clinically confirmed AS. The image results were analyzed by multiple post-processing techniques, as compared to conventional MRI. In AS patients, mean ADC values were (0.494 ± 0.170) × 10(-3) mm(2)/s in sacrum and (0.513 ± 0.129) × 10(-3) mm(2)/s in ilium, which were significantly higher than those of LBP ((0.306 ± 0.057) × 10(-3) mm(2)/s in sacrum and (0.323 ± 0.083) × 10(-3) mm(2)/s in ilium) and healthy volunteers ((0.315 ± 0.009) × 10(-3) mm(2)/s in sacrum and (0.319 ± 0.012) × 10(-3) mm(2)/s in ilium). No statistical differences were found between simple LBP and healthy volunteers. Mean ADC value of multiple lesions in AS was (0.932 ± 0.299) × 10(-3) mm(2)/s. By WB-DWI, abnormal signals of sacroiliac joints and extra-sacroiliac joint lesions were demonstrated in the locations corresponding with clinical findings. Mean ADC values of multiple lesions were (1.298 ± 0.323) × 10(-3) mm(2)/s in sacrum and (1.216 ± 0.311) × 10(-3) mm(2)/s in ilium. DWI and WB-DWI were shown to be effective in differentiation and systemic evaluation of early AS. Both techniques are likely to play an importance role in the early diagnosis of AS and assessment of treatment response.

The clinical application of orbital magnetic resonance (MR) T2-mapping imaging in detecting the disease activity of Graves' ophthalmopathy (GO), and the predictive values of therapy response to intravenous glucocorticoid (ivGC) were investigated. Approved by the local institutional review board (IRB), 106 consecutive patients with GO were included in this prospective study. All subjects were divided into two groups according to the patients' clinical activity score (CAS): the CAS positive group (CAS ≥3) or the CAS negative group (CAS <3). T2 relaxation time of extraocular muscles (T2RT; ms) and the areas of four extra-ocular muscles (AEOMs; mm) were measured by 3D T2-mapping MR sequence before and after methylprednisolone treatment, so as the CAS and some ophthalmic examinations including visual acuity, intra-ocular pressure, eyeball movement, diplopia and proptosis. In addition, 24 healthy volunteers were recruited as the control group. The mean T2RT and AEOMs in CAS positive group were higher than those in CAS negative group. Both CAS positive and negative groups had significantly higher mean T2RT and AEOMs than the control group (P<0.01). There was a positive correlation between T2RT and AEOMs values in GO patients, both of them had a positive correlation with CAS and the ophthalmic examinations. It was concluded that to evaluate the activity of GO, CAS was mostly related to inflammation symptoms of ocular surface, more than that, T2RT and AEOMs were also related to abnormal findings of the ophthalmic examinations including high ocular pressure, impaired eyeball movement, diplopia and proptosis. T2RT and AEOMs can reflex the inflammation state of ocular muscles better. CAS combined with 3D T2-mapping MR imaging could improve the sensitivity of detection of active GO so as the prediction and evaluation of the response to methylprednisolone treatment.