Predicting Mechanical Ventilation and Mortality in COVID-19 Using Radiomics and Deep Learning on Chest Radiographs: A Multi-Institutional Study.

In this study, we aimed to predict mechanical ventilation requirement and mortality using computational modeling of chest radiographs (CXRs) for coronavirus disease 2019 (COVID-19) patients. This two-center, retrospective study analyzed 530 deidentified CXRs from 515 COVID-19 patients treated at Stony Brook University Hospital and Newark Beth Israel Medical Center between March and August 2020. Linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and random forest (RF) machine learning classifiers to predict mechanical ventilation requirement and mortality were trained and evaluated using radiomic features extracted from patients' CXRs. Deep learning (DL) approaches were also explored for the clinical outcome prediction task and a novel radiomic embedding framework was introduced. All results are compared against radiologist grading of CXRs (zone-wise expert severity scores). Radiomic classification models had mean area under the receiver operating characteristic curve (mAUCs) of 0.78 ± 0.05 (sensitivity = 0.72 ± 0.07, specificity = 0.72 ± 0.06) and 0.78 ± 0.06 (sensitivity = 0.70 ± 0.09, specificity = 0.73 ± 0.09), compared with expert scores mAUCs of 0.75 ± 0.02 (sensitivity = 0.67 ± 0.08, specificity = 0.69 ± 0.07) and 0.79 ± 0.05 (sensitivity = 0.69 ± 0.08, specificity = 0.76 ± 0.08) for mechanical ventilation requirement and mortality prediction, respectively. Classifiers using both expert severity scores and radiomic features for mechanical ventilation (mAUC = 0.79 ± 0.04, sensitivity = 0.71 ± 0.06, specificity = 0.71 ± 0.08) and mortality (mAUC = 0.83 ± 0.04, sensitivity = 0.79 ± 0.07, specificity = 0.74 ± 0.09) demonstrated improvement over either artificial intelligence or radiologist interpretation alone. Our results also suggest instances in which the inclusion of radiomic features in DL improves model predictions over DL alone. The models proposed in this study and the prognostic information they provide might aid physician decision making and efficient resource allocation during the COVID-19 pandemic.

Seamless Integration of Artificial Intelligence Into the Clinical Environment: Our Experience With a Novel Pneumothorax Detection Artificial Intelligence Algorithm.

Although interest in artificial intelligence (AI) has exploded in recent years and led to the development of numerous commercial and noncommercial algorithms, the process of implementing such tools into day-to-day clinical practice is rarely described in the burgeoning AI literature. In this report, we describe our experience with the successful integration of an AI-enabled mobile x-ray scanner with an FDA-approved algorithm for detecting pneumothoraces into an end-to-end solution capable of extracting, delivering, and prioritizing positive studies within our thoracic radiology clinical workflow. We also detail several sample cases from our AI algorithm and associated PACS workflow in action to highlight key insights from our experience. We hope this report can help inform other radiology enterprises seeking to evaluate and implement AI-related workflow solutions into daily clinical practice.

COVID-19 infection complicated by acute ST-elevation myocardial infarction.

Clinicians should be aware of the potential for cardiovascular involvement in COVID-19 infection. Coronavirus disease-2019 (COVID-19) is a viral illness caused by severe acute respiratory syndrome-coronavirus-2. While it primarily causes a respiratory illness, a number of important cardiovascular implications have been reported. We describe a patient presenting with COVID-19 whose hospital course was complicated by ST elevation myocardial infarction requiring percutaneous coronary intervention. The goal is to help clinicians gain awareness of the possibility of cardiovascular disease in COVID-19 infection, and maintain a high index of suspicion particularly for patients with risk factors or a prior history of cardiovascular disease.

Considerations in imaging interpretations for colitis in critically ill patients during the COVID-19 era.

OBJECTIVE: The study aims to demonstrate risk factors for colitis in intensive care unit patients with and without coronavirus disease 2019 (COVID-19). METHODS: Retrospective review was performed to identify intensive care unit (ICU) patients with the diagnosis of COVID-19 with computed tomography (CT) between March 20 and December 31, 2020. ICU patients without COVID-19 diagnosis with CT between March 20 and May 10, 2020 were also identified. CT image findings of colitis or terminal ileitis as well as supportive treatment including ventilator, vasopressors, or extracorporeal membrane oxygenation (ECMO) were recorded. Statistical analysis was performed to determine if clinical factors differed in patients with and without positive CT finding. RESULTS: Total 61 ICU patients were selected, including 32 (52%) COVID-19-positive patients and 29 (48%) non-COVID-19 patients. CT findings of colitis or terminal ileitis were identified in 27 patients (44%). Seventy-four percent of the patients with positive CT findings (20/27) received supportive therapies prior to CT, while 56% of the patients without abnormal CT findings (19/34) received supportive therapies. Vasopressor treatment was significantly associated with development of colitis and/or terminal ileitis (p = 0.04) and COVID-19 status was not significantly different between these groups (p = 0.07). CONCLUSIONS: In our study, there was significant correlation between prior vasopressor therapy and imaging findings of colitis or terminal ileitis in ICU patients, independent of COVID-19 status. Our observation raises a possibility that the reported COVID-19-related severe gastrointestinal complications and potential poor outcome could have been confounded by underlying severe critically ill status, and warrants a caution in diagnosis of gastrointestinal complication.

Are asymptomatic gastrointestinal findings on imaging more common in COVID-19 infection? Study to determine frequency of abdominal findings of COVID-19 infection in patients with and without abdominal symptoms and in patients with chest-only CT scans.

PURPOSE: To identify incidence of abdominal findings in COVID-19 patients with and without abdominal symptoms on various imaging modalities including chest-only CT scans and to correlate them with clinical, laboratory and chest CT findings. MATERIALS AND METHODS: In this retrospective study, we searched our clinical database between March 1st, 2020 and May 22nd, 2020 to identify patients who had positive real-time reverse transcriptase polymerase chain reaction (RT-PCR) on throat swabs for COVID-19, had availability of clinical, laboratory information and had availability of CT scan of chest or abdominal radiograph, abdominal ultrasound or CT scan within 2 weeks of the diagnosis. Abdominal imaging findings on all imaging modalities were documented. Chest CT severity score (CT-SS) was assessed in all patients. Clinical and laboratory findings were recorded from the electronic medical record. Statistical analysis was performed to determine correlation of abdominal findings with CT-SS, clinical and laboratory findings. RESULTS: Out of 264 patients with positive RT-PCR, 73 patients (38 males and 35 females; 35 African American) with mean age of 62.2 (range 21-94) years were included. The median CTSS was 13.5 (IQR 75-25 18-8). Most common finding in the abdomen on CT scans (n = 72) were in the gastrointestinal system in 13/72 patients (18.1%) with fluid-filled colon without wall thickening or pericolonic stranding (n = 12) being the most common finding. Chest-only CT (n = 49) found bowel findings in 3 patients. CTSS did not differ in terms of age, sex, race or number of comorbidities but was associated with longer duration of hospitalization (p = 0.0.0256), longer intensive care unit stay (p = 0.0263), more frequent serum lactate dehydrogenase elevation (p = 0.0120) and serum C-reactive protein elevation (p = 0.0402). No statistically significant correlation of occurrence of bowel abnormalities with CTSS, clinical or laboratory features. Deep venous thrombosis was seen in 7/72 patients (9.8%) with three patients developing pulmonary embolism CONCLUSION: Abnormal bowel is the most common finding in the abdomen in patients with COVID-19 infection, is often without abdominal symptoms and occurs independent of severity of pulmonary involvement, other clinical and laboratory features.

Quantifying the decrease in emergency department imaging utilization during the COVID-19 pandemic at a multicenter healthcare system in Ohio.

PURPOSE: To illustrate the change in emergency department (ED) imaging utilization at a multicenter health system in the state of Ohio during the COVID-19 pandemic. METHODS: A retrospective observational study was conducted assessing ED imaging volumes between March 1, 2020, and May 11, 2020, during the COVID-19 crisis. A rolling 7-day total value was used for volume tracking and comparison. Total imaging utilization in the ED was compared with new COVID-19 cases in our region. Utilization was first categorized by modality and then by plain films and computed tomography (CT) scans grouped by body part. CT imaging of the chest was specifically investigated by assessing both CT chest only exams and CT chest, abdomen, and pelvis (C/A/P) exams. Ultimately, matching pair-wise statistical analysis of exam volumes was performed to assess significance of volume change. RESULTS: Our multicenter health system experienced a 46% drop in imaging utilization (p < 0.0001) during the pandemic. Matching pair-wise analysis showed a statistically significant volume decrease by each modality and body part. The exceptions were non-contrast chest CT, which increased (p = 0.0053), and non-trauma C/A/P CT, which did not show a statistically significant volume change (p = 0.0633). CONCLUSION: ED imaging utilization trends revealed through actual health system data will help inform evidence-based decisions for more accurate volume predictions and therefore institutional preparedness for current and future pandemics.