Integrated model for COVID-19 diagnosis based on computed tomography artificial intelligence, and clinical features: a multicenter cohort study.

Background: We developed and validated a machine learning diagnostic model for the novel coronavirus (COVID-19) disease, integrating artificial-intelligence-based computed tomography (CT) imaging and clinical features. Methods: We conducted a retrospective cohort study in 11 Japanese tertiary care facilities that treated COVID-19 patients. Participants were tested using both real-time reverse transcription polymerase chain reaction (RT-PCR) and chest CTs between January 1 and May 30, 2020. We chronologically split the dataset in each hospital into training and test sets, containing patients in a 7:3 ratio. A Light Gradient Boosting Machine model was used for the analysis. Results: A total of 703 patients were included, and two models-the full model and the A-blood model-were developed for their diagnosis. The A-blood model included eight variables (the Ali-M3 confidence, along with seven clinical features of blood counts and biochemistry markers). The areas under the receiver-operator curve of both models [0.91, 95% confidence interval (CI): 0.86 to 0.95 for the full model and 0.90, 95% CI: 0.86 to 0.94 for the A-blood model] were better than that of the Ali-M3 confidence (0.78, 95% CI: 0.71 to 0.83) in the test set. Conclusions: The A-blood model, a COVID-19 diagnostic model developed in this study, combines machine-learning and CT evaluation with blood test data and performs better than the Ali-M3 framework existing for this purpose. This would significantly aid physicians in making a quicker diagnosis of COVID-19.

Intensive care burden of COVID-19 in tertiary care hospitals during the first year of outbreak in Kawasaki City, Japan: A retrospective cohort study.

INTRODUCTION: This study aimed to describe the changes in the intensive care burden of coronavirus disease 2019 (COVID-19) during the first year of outbreak in Japan. METHODS: This retrospective cohort study included COVID-19 patients who received mechanical ventilation (MV) support in two designated hospitals for critical patients in Kawasaki City. We compared the lengths of MV and stay in the intensive care unit (ICU) or high care unit (HCU) according to the three epidemic waves. We calculated in-hospital mortality rates in patients with or without MV. RESULTS: The median age of the sample was 65.0 years, and 22.7% were women. There were 37, 29, and 62 patients in the first (W1), second (W2), and third waves (W3), respectively. Systemic steroids, remdesivir, and prone positioning were more frequent in W2 and W3. The median length of MV decreased from 18.0 days in W1 to 13.0 days in W3 (P = 0.019), and that of ICU/HCU stay decreased from 22.0 days in W1 to 15.5 days in W3 (P = 0.027). The peak daily number of patients receiving MV support was higher at 18 patients in W1, compared to 8 and 15 patients in W2 and W3, respectively. The mortality rate was 23.4%, which did not significantly change (P = 0.467). CONCLUSIONS: The lengths of MV and ICU/HCU stay per patient decreased over time. Despite an increase in the number of COVID-19 patients who received MV in W3, this study may indicate that the intensive care burden during the study period did not substantially increase.

Efficacy of awake prone positioning for severe illness coronavirus disease 2019 patients: a propensity score-adjusted cohort study.

AIM: Awake prone positioning (PP) in patients with coronavirus disease 2019 (COVID-19) can improve oxygenation. However, evidence showing that it can prevent intubation is lacking. This study investigated the efficacy of awake PP in patients with COVID-19 who received remdesivir, dexamethasone, and anticoagulant therapy. METHODS: This was a two-center cohort study. Patients admitted to the severe COVID-19 patient unit were included. The primary outcome was the intubation rate and secondary outcome was length of stay in the severe COVID-19 unit. After propensity score adjustment, we undertook multivariable regression to calculate the estimates of outcomes between patients who received awake PP and those who did not. RESULTS: Overall, 108 patients were included (54 [50.0%] patients each who did and did not undergo awake PP), of whom 25 (23.2%) were intubated (with awake PP, 5 [9.3%] vs. without awake PP, 20 [37.0%]; P < 0.01). The median length of stay in the severe COVID-19 unit did not significantly differ (with awake PP, 5 days vs. without awake PP, 5.5 days; P = 0.68). After propensity score adjustment, those who received awake PP had a lower intubation rate than those who did not (odds ratio, 0.22; 95% confidence interval, 0.06-0.85; P = 0.03). Length of stay in the severe COVID-19 patient unit did not differ significantly (adjusted percentage difference, -24.4%; 95% confidence interval, -56.3% to 30.8%; P = 0.32). CONCLUSION: Awake PP could be correlated with intubation rate in patients with COVID-19 who are receiving remdesivir, dexamethasone, and anticoagulant therapy.

A fatal case of extensive gastrointestinal necrosis due to portal and mesenteric vein thrombosis in the post-acute phase of COVID-19.

Portal vein thrombosis (PVT) is considered a relatively rare thrombotic complication in coronavirus disease 2019 (COVID-19). Most reported cases of PVT develop within 2 weeks from COVID-19 onset. We report a fatal case of extensive gastrointestinal necrosis due to portal and mesenteric vein thrombosis approximately 6 weeks after the onset of critical COVID-19. Excessive elevation of his plasma D-dimer level had continued for weeks during the hospitalization contrary with improvement of respiratory failure. Thrombotic complication should be cautiously paid attention even in the post-acute phase of COVID-19, especially in patients with persistent elevation of plasma D-dimer level.

Development and external validation of a deep learning-based computed tomography classification system for COVID-19.

BACKGROUND: We aimed to develop and externally validate a novel machine learning model that can classify CT image findings as positive or negative for SARS-CoV-2 reverse transcription polymerase chain reaction (RT-PCR). METHODS: We used 2,928 images from a wide variety of case-control type data sources for the development and internal validation of the machine learning model. A total of 633 COVID-19 cases and 2,295 non-COVID-19 cases were included in the study. We randomly divided cases into training and tuning sets at a ratio of 8:2. For external validation, we used 893 images from 740 consecutive patients at 11 acute care hospitals suspected of having COVID-19 at the time of diagnosis. The dataset included 343 COVID-19 patients. The reference standard was RT-PCR. RESULTS: In external validation, the sensitivity and specificity of the model were 0.869 and 0.432, at the low-level cutoff, 0.724 and 0.721, at the high-level cutoff. Area under the receiver operating characteristic was 0.76. CONCLUSIONS: Our machine learning model exhibited a high sensitivity in external validation datasets and may assist physicians to rule out COVID-19 diagnosis in a timely manner at emergency departments. Further studies are warranted to improve model specificity.

Accuracy of deep learning-based computed tomography diagnostic system for COVID-19: A consecutive sampling external validation cohort study.

Ali-M3, an artificial intelligence program, analyzes chest computed tomography (CT) and detects the likelihood of coronavirus disease (COVID-19) based on scores ranging from 0 to 1. However, Ali-M3 has not been externally validated. Our aim was to evaluate the accuracy of Ali-M3 for detecting COVID-19 and discuss its clinical value. We evaluated the external validity of Ali-M3 using sequential Japanese sampling data. In this retrospective cohort study, COVID-19 infection probabilities for 617 symptomatic patients were determined using Ali-M3. In 11 Japanese tertiary care facilities, these patients underwent reverse transcription-polymerase chain reaction (RT-PCR) testing. They also underwent chest CT to confirm a diagnosis of COVID-19. Of the 617 patients, 289 (46.8%) were RT-PCR-positive. The area under the curve (AUC) of Ali-M3 for predicting a COVID-19 diagnosis was 0.797 (95% confidence interval: 0.762‒0.833) and the goodness-of-fit was P = 0.156. With a cut-off probability of a diagnosis of COVID-19 by Ali-M3 set at 0.5, the sensitivity and specificity were 80.6% and 68.3%, respectively. A cut-off of 0.2 yielded a sensitivity and specificity of 89.2% and 43.2%, respectively. Among the 223 patients who required oxygen, the AUC was 0.825. Sensitivity at a cut-off of 0.5% and 0.2% was 88.7% and 97.9%, respectively. Although the sensitivity was lower when the days from symptom onset were fewer, the sensitivity increased for both cut-off values after 5 days. We evaluated Ali-M3 using external validation with symptomatic patient data from Japanese tertiary care facilities. As Ali-M3 showed sufficient sensitivity performance, despite a lower specificity performance, Ali-M3 could be useful in excluding a diagnosis of COVID-19.

Functional decline in hospitalized older patients with coronavirus disease 2019: a retrospective cohort study.

BACKGROUND: This study aimed to determine the frequency of functional decline and to identify the factors related to a greater risk of functional decline among hospitalized older patients with coronavirus disease 2019 (COVID-19). METHODS: We reviewed the medical records of patients aged over 65 years who were admitted to a tertiary care hospital for COVID-19 over 1 year from February 2020. We evaluated the proportion of functional decline, which was defined as a decrease in the Barthel Index score from before the onset of COVID-19 to discharge. Multivariable logistic regression analyses were performed to evaluate the associations between the demographic and clinical characteristics of patients at admission and a greater risk of functional decline. Two sensitivity analyses with different inclusion criteria were performed: one in patients without very severe functional decline before the onset of COVID-19 (i.e., limited to those with Barthel Index score ≥ 25), and the other with a composite outcome of functional decline and death at discharge. RESULTS: The study included 132 patients with COVID-19; of these, 72 (54.5%) developed functional decline. The severity of COVID-19 did not differ between patients with functional decline and those without (P = 0.698). Factors associated with a greater risk of functional decline included female sex (adjusted odds ratio [aOR], 3.14; 95% confidence interval [CI], 1.25 to 7.94), Barthel Index score < 100 before the onset of COVID-19 (aOR, 13.73; 95% CI, 3.29 to 57.25), and elevation of plasma D-dimer level on admission (aOR, 3.19; 95% CI, 1.12 to 9.07). The sensitivity analyses yielded similar results to those of the main analysis. CONCLUSIONS: Over half of the older patients who recovered from COVID-19 developed functional decline at discharge from a tertiary care hospital in Japan. Baseline activities of daily living impairment, female sex, and elevated plasma D-dimer levels at admission were associated with a greater risk of functional decline.

Cerebral toxoplasmosis with multiple hemorrhage lesions in an HIV infected patient: A case report and literature review.

A 47-year-old man was admitted to a hospital for disturbance of consciousness. He was diagnosed with multiple hemorrhagic brain abscesses in bilateral hemispheres with human immunodeficiency virus (HIV) infection, and was transferred to our hospital for further examination and treatment. On admission, although he could respond to pain stimuli, he could not talk or communicate. His laboratory data on admission revealed CD4-positive T cell count of 67 cells/µL, and HIV1-RNA viral load of 5.6 × 105 copies/mL. Both the serum IgG Toxoplasma gondii antibody and the cerebrospinal fluid polymerase chain reaction for Toxoplasma gondii DNA were positive. He was diagnosed with cerebral toxoplasmosis and HIV infection. His level of consciousness worsened, and the number of hemorrhagic lesions had increased in both hemispheres and the left thalamus on the computed tomography scan following two weeks of antitoxoplasma therapy. These newly discovered hemorrhagic lesions revealed in the CT had been found as the high intensity signal regions of initial fluid-attenuated inversion recovery magnetic resonance imaging. After five weeks of treatment, the hemorrhagic lesions gradually improved along with the patient's consciousness. Antiretroviral therapy was initiated six weeks following antitoxoplama therapy with reassurance that immune reconstitution inflammatory syndrome did not occur. After approximately four months of antitoxoplasma therapy, the patient was discharged into a group home with residual left hemiparesis on maintenance antitoxoplasma and antiretroviral therapy. Clinicians should recognize the delay of clinical and radiological improvement for hemorrhagic cerebral toxoplasmosis and patiently continue the antitoxoplasma therapy.

COVID-19 and Fatal Sepsis Caused by Hypervirulent Klebsiella pneumoniae, Japan, 2020.

A patient in Japan with coronavirus disease and hypervirulent Klebsiella pneumoniae K2 sequence type 86 infection died of respiratory failure. Bacterial and fungal co-infections caused by region-endemic pathogens, including hypervirulent K. pneumoniae in eastern Asia, should be included in the differential diagnosis of coronavirus disease patients with acutely deteriorating condition.

Epidemiology and prognosis of ommaya reservoir-related bacterial meningitis in adult patients with leptomeningeal metastases from solid tumors: A 10-year retrospective single-center study in Japan.

INTRODUCTION: Leptomeningeal metastases (LM) from solid tumors have poor prognosis. Intrathecal chemotherapy through the Ommaya reservoir (OR) is one of the options for treating LM; however, OR-related bacterial meningitis (ORRBM) is a severe complication in patients who underwent OR placement. Little is known about the incidence rate and prognosis of ORRBM among patients with LM from solid tumors. METHODS: We retrospectively reviewed the records of patients who underwent OR placement to treat LM from solid tumors at Kawasaki Municipal Kawasaki Hospital between January 2009, and December 2018. RESULTS: Among 136 patients with OR placement (median age of 64.5 years) including 30,320 Ommaya-days, 18 (13.2%) developed ORRBM (5.9 infections per 10,000 Ommaya-days). The major primary diseases were lung cancer (65.4%). The median times from OR placement and from last OR puncture to ORRBM onset were 20 days and 4.5 days, respectively. Major clinical symptoms were fever (83.3%), headache (50.0%), disturbance of consciousness (50.0%), and nausea (38.9%). Seventeen of 18 patients underwent an OR removal operation. One patient died from ORRBM, and another patient died from heart failure during ORRBM treatment. The median duration of treatment with antibiotics was 16.5 days. The median survival period from the day of OR placement was 146.5 days among patients who developed ORRBM and 142.5 days among patients who did not develop ORRBM. CONCLUSIONS: The rate of ORRBM among patients with LM from solid tumors in our hospital was 13.2%. ORRBM may not shorten the patients' survival period with adequate management including removal of the device.

Clinical course of 2019 novel coronavirus disease (COVID-19) in individuals present during the outbreak on the Diamond Princess cruise ship.

We investigated the clinical course of individuals with 2019 novel coronavirus disease (COVID-19) who were transferred from the Diamond Princess cruise ship to 12 local hospitals. The conditions and clinical courses of patients with pneumonia were compared with those of patients without pneumonia. Among 70 patients (median age: 67 years) analyzed, the major symptoms were fever (64.3%), cough (54.3%), and general fatigue (24.3%). Forty-three patients (61.4%) had pneumonia. Higher body temperature, heart rate, and respiratory rate as well as higher of lactate dehydrogenase (LDH), aspartate aminotransferase (AST), and C-reactive protein (CRP) levels and lower serum albumin level and lymphocyte count were associated with the presence of pneumonia. Ground-glass opacity was found in 97.7% of the patients with pneumonia. Patients were administered neuraminidase inhibitors (20%), lopinavir/ritonavir (32.9%), and ciclesonide inhalation (11.4%). Mechanical ventilation and veno-venous extracorporeal membrane oxygenation was performed on 14 (20%) and 2 (2.9%) patients, respectively; two patients died. The median duration of intubation was 12 days. The patients with COVID-19 transferred to local hospitals during the outbreak had severe conditions and needed close monitoring. The severity of COVID-19 depends on the presence of pneumonia. High serum LDH, AST and CRP levels and low serum albumin level and lymphocyte count were found to be predictors of pneumonia. It was challenging for local hospitals to admit and treat these patients during the outbreak of COVID-19. Assessment of severity was crucial to manage a large number of patients.

An HIV-infected Patient with No Serious Adverse Events after Overdosing on Raltegravir.

Patients with HIV infection represent a high-risk group for medication overdose because of the high frequency of complicating psychiatric disorders. Raltegravir is well-known for its low frequency of adverse effects. We herein report a 42-year-old Japanese man with HIV infection who was hospitalized 6 hours after overdosing with 24,000 mg of raltegravir in a suicide attempt. No serious adverse events occurred, although the plasma concentration of raltegravir at 18 hours after the overdose was 79,871.1 ng/mL. Raltegravir may be well-indicated for HIV patients at risk of overdosing.

Septic Meningitis and Liver Abscess due to Hypermucoviscous Klebsiella pneumoniae Complicated with Chronic Strongyloidiasis in a Human T-lymphotropic Virus 1 Carrier.

Recently, severe cases of infection due to hypermucoviscous Klebsiella pneumonia (hmKP) have been reported in Japan. The Amami Islands in Japan are also endemic regions for Strongyloides stercoralis. Disseminated strongyloidiasis strain often causes severe enterobacteria infection; however, whether or not chronic strongyloidiasis induces it remains unclear. We herein report a 71-year-old man who developed meningitis and liver abscess due to hmKP complicated with chronic strongyloidiasis. He died on the seventh hospital day. Strongyloides stercoralis were only found around the polyp in the cecum. Chronic strongyloidiasis can also induce severe infection due to enterobacteria, especially hypervirulent pathogens like hmKP, through the induction of mucosal rupture.

Splenic infarction complicated with immune reconstitution inflammatory syndrome due to disseminated Mycobacterium genavense infection in a patient infected with human immunodeficiency virus.

Mycobacterium genavense (M. genavense) is one of the most fastidious, difficult to culture Mycobacterium species. Patients infected with human immunodeficiency virus (HIV) may develop immune reconstitution inflammatory syndrome (IRIS) due to disseminated M. genavense infection as well as disseminated M. avium and intracellulare complex infection. Consensus regarding treatment of IRIS due to disseminated mycobacterium infection has not yet been obtained, although systemic steroid therapy has been recommended in recent guidelines. Here we report the case of a 48-year-old Japanese man diagnosed with HIV and disseminated M. genavense infection. His initial CD4-positive T cell count was 3/µL, and his HIV1-RNA viral load was 13,000 copies/mL. He developed IRIS due to disseminated M. genavense infection after two weeks of receiving antiretroviral agents. The patient's serum alkaline phosphatase level, as a barometer of disseminated M. genavense infection in this case, was difficult to control with several anti-mycobacterial agents, although his fever was improved by non-steroidal anti-inflammatory drugs. About five weeks after the onset of IRIS, the patient developed acute left upper quadrant pain and was diagnosed with splenic infarction by contrast-enhanced computed tomography. After the splenic infarction, the patient's serum alkaline phosphatase level decreased without systemic steroid therapy or anticoagulant agents, and his left upper quadrant pain improved naturally within a few days. This case suggests that IRIS due to disseminated M. genavense infection can complicate splenic infarction in patients with HIV, and splenic infarction could improve the IRIS due to disseminated M. genavense infection.

[A Case of Leptospirosis in which the Causative Pathogen was Detected Using Cerebrospinal Fluid PCR Eight Days after Onset].

We report a patient with leptospirosis caused by infection with Leptospira interrogans serovar Rachmati. A 30-year-old Japanese man took part in a survival camp on Iriomote Island, Okinawa, from July 9 to July 15, 2014. During the camp, he swam in the river and kayaked. He developed a high fever and fatigue 7 days after completing his trip and was admitted to our hospital on July 22. On admission, he complained of a posterior cervical pain and a loss of appetite. Laboratory findings revealed granulocytosis, mildly elevated AST and ALT levels, elevated BUN and Cr levels, and a significantly elevated CRP level. No pathogenic bacteria were isolated from blood, urine, or cerebrospinal fluid cultures. We included leptospirosis in the differential diagnosis because of the patient's history of participating in a survival camp on Iriomote Island. Minocycline 200 mg, p.o. showed an excellent efficacy. The Leptospira flagellar gene FlaB was detected using a cerebrospinal fluid PCR. A microscopic agglutination test (MAT) during the convalescent stage demonstrated significant increases in antibodies against L. interrogans serovar Rachmati, confirming the diagnosis of leptospirosis. A medical history including occupation and recent travel history, and an adequate specimen sampling are crucial for the accurate and early diagnosis of leptospirosis.