Prediction of oxygen requirement in patients with COVID-19 using a pre-trained chest radiograph xAI model: efficient development of auditable risk prediction models via a fine-tuning approach.

Risk prediction requires comprehensive integration of clinical information and concurrent radiological findings. We present an upgraded chest radiograph (CXR) explainable artificial intelligence (xAI) model, which was trained on 241,723 well-annotated CXRs obtained prior to the onset of the COVID-19 pandemic. Mean area under the receiver operating characteristic curve (AUROC) for detection of 20 radiographic features was 0.955 (95% CI 0.938-0.955) on PA view and 0.909 (95% CI 0.890-0.925) on AP view. Coexistent and correlated radiographic findings are displayed in an interpretation table, and calibrated classifier confidence is displayed on an AI scoreboard. Retrieval of similar feature patches and comparable CXRs from a Model-Derived Atlas provides justification for model predictions. To demonstrate the feasibility of a fine-tuning approach for efficient and scalable development of xAI risk prediction models, we applied our CXR xAI model, in combination with clinical information, to predict oxygen requirement in COVID-19 patients. Prediction accuracy for high flow oxygen (HFO) and mechanical ventilation (MV) was 0.953 and 0.934 at 24 h and 0.932 and 0.836 at 72 h from the time of emergency department (ED) admission, respectively. Our CXR xAI model is auditable and captures key pathophysiological manifestations of cardiorespiratory diseases and cardiothoracic comorbidities. This model can be efficiently and broadly applied via a fine-tuning approach to provide fully automated risk and outcome predictions in various clinical scenarios in real-world practice.

High-Dose Inhaled Nitric Oxide for the Treatment of Spontaneously Breathing Pregnant Patients With Severe Coronavirus Disease 2019 (COVID-19) Pneumonia.

OBJECTIVE: To evaluate whether the use of inhaled nitric oxide (iNO)200 improves respiratory function. METHODS: This retrospective cohort study used data from pregnant patients hospitalized with severe bilateral coronavirus disease 2019 (COVID-19) pneumonia at four teaching hospitals between March 2020 and December 2021. Two cohorts were identified: 1) those receiving standard of care alone (SoC cohort) and 2) those receiving iNO200 for 30 minutes twice daily in addition to standard of care alone (iNO200 cohort). Inhaled nitric oxide, as a novel therapy, was offered only at one hospital. The prespecified primary outcome was days free from any oxygen supplementation at 28 days postadmission. Secondary outcomes were hospital length of stay, rate of intubation, and intensive care unit (ICU) length of stay. The multivariable-adjusted regression analyses accounted for age, body mass index, gestational age, use of steroids, remdesivir, and the study center. RESULTS: Seventy-one pregnant patients were hospitalized for severe bilateral COVID-19 pneumonia: 51 in the SoC cohort and 20 in the iNO200 cohort. Patients receiving iNO200 had more oxygen supplementation-free days (iNO200: median [interquartile range], 24 [23-26] days vs standard of care alone: 22 [14-24] days, P=.01) compared with patients in the SoC cohort. In the multivariable-adjusted analyses, iNO200 was associated with 63.2% (95% CI 36.2-95.4%; P<.001) more days free from oxygen supplementation, 59.7% (95% CI 56.0-63.2%; P<.001) shorter ICU length of stay, and 63.6% (95% CI 55.1-70.8%; P<.001) shorter hospital length of stay. No iNO200-related adverse events were reported. CONCLUSION: In pregnant patients with severe bilateral COVID-19 pneumonia, iNO200 was associated with a reduced need for oxygen supplementation and shorter hospital stay.

Inhaled high dose nitric oxide is a safe and effective respiratory treatment in spontaneous breathing hospitalized patients with COVID-19 pneumonia.

BACKGROUND: Inhaled nitric oxide (NO) is a selective pulmonary vasodilator. In-vitro studies report that NO donors can inhibit replication of SARS-CoV-2. This multicenter study evaluated the feasibility and effects of high-dose inhaled NO in non-intubated spontaneously breathing patients with Coronavirus disease-2019 (COVID-19). METHODS: This is an interventional study to determine whether NO at 160 parts-per-million (ppm) inhaled for 30 min twice daily might be beneficial and safe in non-intubated COVID-19 patients. RESULTS: Twenty-nine COVID-19 patients received a total of 217 intermittent inhaled NO treatments for 30 min at 160 ppm between March and June 2020. Breathing NO acutely decreased the respiratory rate of tachypneic patients and improved oxygenation in hypoxemic patients. The maximum level of nitrogen dioxide delivered was 1.5 ppm. The maximum level of methemoglobin (MetHb) during the treatments was 4.7%. MetHb decreased in all patients 5 min after discontinuing NO administration. No adverse events during treatment, such as hypoxemia, hypotension, or acute kidney injury during hospitalization occurred. In our NO treated patients, one patient of 29 underwent intubation and mechanical ventilation, and none died. The median hospital length of stay was 6 days [interquartile range 4-8]. No discharged patients required hospital readmission nor developed COVID-19 related long-term sequelae within 28 days of follow-up. CONCLUSIONS: In spontaneous breathing patients with COVID-19, the administration of inhaled NO at 160 ppm for 30 min twice daily promptly improved the respiratory rate of tachypneic patients and systemic oxygenation of hypoxemic patients. No adverse events were observed. None of the subjects was readmitted or had long-term COVID-19 sequelae.