Comorbidity burden and outcomes after mechanical thrombectomy for large vessel occlusion: A retrospective analysis.

OBJECTIVES: Patients with acute ischemic stroke (AIS) often have an accumulation of pre-existing comorbidities, but its clinical impact on outcomes after mechanical thrombectomy (MT) remains unknown. Therefore, we examined whether comorbidity burden before AIS onset could predict clinical outcomes after MT. METHODS: In this retrospective cohort, we enrolled consecutive patients with community-onset AIS who underwent MT between April 2016 and December 2021. To evaluate each patient's comorbidity burden, we calculated Charlson comorbidity index (CCI), then classified the patients into the High CCI (≥ 3) and the Low CCI (< 3) groups. The primary outcome was a good neurological outcome at 90 days, defined as a modified Rankin scale 0-2 or no worse than the previous daily conditions. All-cause mortality at 90 days and hemorrhagic complications after MT were also compared between the two groups. We estimated the odds ratios and their confidence intervals using a multivariable logistic regression model. RESULTS: A total of 388 patients were enrolled, of whom 86 (22.2%) were classified into the High CCI group. Patients in the High CCI group were less likely to achieve a good neurological outcome (adjusted odds ratio of 0.26 [95% confidence interval, 0.12-0.58]). Moreover, symptomatic intracranial hemorrhage was more common in the High CCI (14.0% vs. 4.6%; adjusted odds ratio, 4.10 [95% confidence interval, 1.62-10.3]). CONCLUSIONS: Comorbidity burden assessed by CCI was associated with clinical outcomes after MT. CCI has the potential to become a simple and valuable tool for predicting neurological prognosis among patients with AIS and MT.

Pipeline embolization device dynamics: prediction of incomplete occlusion by elongation from nominal length.

OBJECTIVE: According to benchtop studies, the oversizing of a Pipeline embolization device (PED) relative to the parent artery leads to a significant increase in porosity and potentially compromises aneurysm occlusion as well as transitional zone (TZ) formation around the neck of aneurysms. However, no clinical assessment has been reported. Here this potential was studied by measuring the dynamic changes of PEDs in the clinical time course. METHODS: The authors retrospectively examined 124 anterior circulation unruptured aneurysms in 114 consecutive patients treated with a PED between July 2015 and December 2020 at their institution. The authors excluded 77 cases of 68 patients with adjunctive coil embolization or multiple stents that could affect the PED dynamics and measurements, and 47 aneurysms in 46 patients were included. Measurements were performed before, immediately after, and 6 months after treatment, and then at intervals of 6 months to 1 year after that for nonocclusion cases. RESULTS: Complete occlusion was achieved in 79.0% and incomplete occlusion in 21.0% at last follow-up. The PED length immediately after deployment was 136% nominal length. A multivariable regression analysis revealed that age (OR 1.11/year; p = 0.02) and PED elongation from nominal length (OR 1.31/mm; p = 0.012) were independently associated with a higher rate of incomplete occlusion at the last follow-up. TZ formation did not affect the occlusion rate. CONCLUSIONS: PED elongation from the nominal length is a new predictor of incomplete aneurysm occlusion. The PED showed vascular remodeling by changing its diameter and length in the clinical course. TZ formation was remodeled and did not affect the occlusion rate.

Predictive performance of the sequential organ failure assessment score for in-hospital mortality in patients with end-stage kidney disease in intensive care units: A multicenter registry in Japan.

INTRODUCTION: There is limited evidence regarding whether the performance of the Sequential Organ Failure Assessment (SOFA) score differs between patients with and without end-stage kidney disease (ESKD) in intensive care units (ICUs). METHODS: We used a multicenter registry (Japanese Intensive care Patient Database) to enroll adult ICU patients between April 2018 and March 2021. We recalibrated the SOFA score using a logistic regression model and evaluated its predictive ability in both ESKD and non-ESKD groups. The primary outcome was in-hospital mortality. RESULTS: 128 134 patients were enrolled. The AUROC of the SOFA score was lower in the ESKD group than in the non-ESKD group [0.789 (95% CI, 0.774-0.804) vs. 0.846 (95% CI, 0.841-0.850)]. The calibration plot revealed good performance in both groups. However, it overestimated in-hospital mortality in ESKD groups. CONCLUSION: The SOFA score demonstrated good predictive ability in patients with and without ESKD, but it overestimated the in-hospital mortality in ESKD patients.

Early prediction of cytokine release syndrome by measuring phosphate and magnesium levels following chimeric antigen receptor T cell therapy.

Introduction: Cytokine release syndrome (CRS) is a life-threatening side effect of chimeric antigen receptor T (CAR-T) cell therapy. This study investigated whether serum inorganic phosphate (IP) and magnesium (Mg) levels are predictive markers of CRS development. Methods: This single-center retrospective cohort study enrolled 16 consecutive patients with diffuse large B-cell lymphoma who had received CAR-T cell therapy. Logistic regression models with generalized estimating equations were used to evaluate whether changes in IP and Mg levels from their baseline values were associated with the development of CRS within 48 hours. Results: Decreased IP and Mg levels from baseline (per 10% change) were associated with an increased CRS incidence (adjusted odds ratio 2.18 [95% confidence interval (CI), 1.31-3.62], 3.18 [95% CI, 1.57-6.44], respectively). Conclusions: Changes in IP and Mg concentrations within 48 hours may be useful predictive markers of CRS onset.

Helicobacter cinaedi Infections in Emergency Departments: A Descriptive Study.

AIM: Helicobacter cinaedi, a Gram-negative spiral bacterium, is a rare cause of bacteremia in humans. Unfortunately, little is known about H. cinaedi infections in emergency departments (EDs). We aimed to describe the clinical features of H. cinaedi infections in the ED. METHODS: We conducted a descriptive study at the ED of Kobe City General Hospital (KCGH) in Japan between November 2011 and December 2020. We included all ED patients with H. cinaedi infections. We retrospectively obtained the patient data from electronic medical records and described the patient characteristics, clinical course, and management of H. cinaedi infections. RESULTS: A total of 22 patients in the ED were diagnosed with H. cinaedi infections, and all of them were detected through blood cultures. The chief complaints were vague: fever (18/22, 81.8%), chills (10/22, 45.5%), and localized pain or tenderness (8/22, 36.4%). Patients with complicated cases were also reported in the ED; three patients had vertebral osteomyelitis, two had infected aortic aneurysms, and another two had infected cysts (renal cyst and pancreatic cyst with concomitant empyema). Tetracycline (minocycline) was primarily prescribed and administered intravenously in five of 15 (33.3%) and orally in nine of 20 (45.0%) patients. Only one (4.5%) patient required surgical interventions. None of the patients died in the hospital. CONCLUSIONS: We reported the clinical features of H. cinaedi infections in the ED. Although some patients developed complicated infections, the prognosis was not poor under appropriate treatment, and most of them were successfully treated with antibiotics, primarily tetracycline.

Neurologic outcome and location of cardiac arrest in out-of-hospital cardiac arrest patients who underwent extracorporeal cardiopulmonary resuscitation: A multicentre retrospective cohort in Japan.

Aim: We examined the association between the location of cardiac arrest and outcomes of patients with out-of-hospital cardiac arrest (OHCA) who underwent extracorporeal cardiopulmonary resuscitation (ECPR). Methods: This was a secondary analysis of SAVE-J II, a multicentre retrospective registry with 36 participating institutions across Japan, which enrolled adult patients with OHCA who underwent ECPR. The outcomes of interest were favourable neurologic outcome at discharge. We compared the outcome between OHCA cases that occurred at residential and public locations, using a multilevel logistic regression model allowing for the random effect of each hospital. Results: Among 1,744 enrolled OHCAs, 809 and 935 occurred at residential (house: 603; apartment: 206) and public (street: 260; workplace: 210; others: 465) locations, respectively. The proportion of favourable neurologic outcomes was lower in OHCAs at residential locations than those at public locations (88/781 (11.3%) vs.131/891 (14.7%); adjusted odds ratio, 0.72 [95% confidence interval, 0.53-0.99]). However, subgroup analyses for patients with EMS aged <65 years call to hospital arrival within 30 minutes or during daytime revealed less difference between residential and public locations. Conclusion: When cardiac arrests occurred at residential locations, lower proportions of favourable neurologic outcomes were exhibited among patients with OHCA who underwent ECPR. However, the event's location may not affect the prognosis among appropriate and select cases when transported within a limited timeframe.

Longitudinal Association Between Social Isolation and COVID-19 Vaccine Uptake in Japanese Older Citizen: an Observational Study.

BACKGROUND: Social isolation is associated with decreased intent to receive vaccination against coronavirus disease 2019 (COVID-19); however, its association with COVID-19 vaccine uptake is unknown. OBJECTIVE: The present study aimed to determine the association between social isolation and subsequent COVID-19 vaccination using the data from a representative Japanese sample. DESIGN: Longitudinal observational study. PARTICIPANTS: Japanese citizens aged ≥ 65 years who participated in the Quality of Life in COVID-19 Era (QoLCoVE) study which began in March 2020. MAIN MEASURES: Social isolation was assessed in March 2020 using the abbreviated Lubben Social Network Scale, with a score of ≤ 11 indicating social isolation. To evaluate the relationship between social isolation at the baseline and COVID-19 vaccine uptake (including reservation for vaccine administration) between August and September 2021, we estimated adjusted risk ratios with 95% CIs using modified Poisson regression with adjustment for potential confounders. KEY RESULTS: A total of 921 participants met the inclusion criteria in the initial survey, of whom 720 (78.2%) completed the follow-up survey. Median age was 70 years (interquartile range: 66-72), and 329 (45.7%) participants were male. Twenty-one (16.5%) of 127 socially isolated participants, and 48 (8.1%) of 593 non-socially isolated participants did not receive the COVID-19 vaccine. Socially isolated participants were less likely to take COVID-19 vaccine than non-socially isolated participants (adjusted risk ratio 1.98, 95% CI: 1.18-3.32). CONCLUSIONS: Social isolation is associated with reduced COVID-19 vaccine uptake among older adults. To further promote COVID-19 vaccination in the older population, support for social isolation might be necessary.

Prehospital shock index predicts 24-h mortality in trauma patients with a normal shock index upon emergency department arrival.

BACKGROUND: The shock index (heart rate divided by systolic blood pressure) of trauma patients upon emergency department arrival predicts blood loss and death. However, some patients with normal shock indices (0.4 < shock index <0.9) upon emergency department arrival also have poor prognoses. This study aimed to determine whether abnormal prehospital shock indices in trauma patients with normal shock indices upon emergency department arrival were predictors of a high risk of mortality. METHODS: We conducted a retrospective cohort study of emergency department-admitted trauma patients from 2004 to 2017. The study included 89,495 consecutive trauma patients aged ≥16 years, with Abbreviated Injury Scale score of ≥3, who were transported to the emergency department directly from the field and had a normal shock index upon emergency department arrival. According to the prehospital shock index scores, the patients were categorized into low shock index (≤ 0.4), normal shock index, and high shock index (≥0.9) groups. Odds ratios and 95% confidence intervals were calculated using logistic regression analysis. RESULTS: The 89,495 patients had a median age of 64 (interquartile range: 43-79) years, and 55,484 (62.0%) of the patients were male. There were 1350 (1.5%) 24-h deaths in total; 176/4263 (4.1%), 1017/78,901 (1.3%), and 157/6331 (2.5%) patients were in the low, normal, and high prehospital shock index groups, respectively. The adjusted odds ratios for 24-h mortality compared with the normal shock index group were 1.63 (95% confidence interval: 1.34-1.99) in the low shock index group and 1.62 (95% confidence interval: 1.31-1.99) in the high shock index group. CONCLUSION: Trauma patients with abnormal prehospital shock indices but normal shock indices upon emergency department arrival are at a higher risk of 24-h mortality. Identifying these indices could improve triage and targeted care for patients.

Etiology and clinical characteristics of patients with severely impaired consciousness in prehospital settings: A retrospective study.

Aim: To examine the causes of patients with severely impaired consciousness and the clinical characteristics in prehospital settings that are useful for differential diagnosis, especially stroke. Methods: We retrospectively examined patients aged ≥16 years with Japan Coma Scale III-digit codes during paramedic contact and transported to our hospital between January 2018 and December 2018. Furthermore, we examined background and physical findings of patients at final diagnosis, and also examined factors associated with stroke. Results: Overall, 227 patients were included in this study. One hundred and twelve patients (49.3%) were male, and the median age was 71 years (interquartile range, 50-83 years). Stroke was the most common cause (30%). Intoxication and psychiatric disorders were significantly more common in younger patients (p < 0.01). Systolic blood pressure was the highest in patients with stroke. Mortality was the highest in stroke (55.9%). Systolic blood pressure, airway compromise, and ocular abnormalities were factors associated with stroke, with odds ratios of 1.03 (95% confidence interval [CI], 1.02-1.04), 6.88 (95% CI, 3.02-15.64), and 3.86 (95% CI, 1.61-9.27), respectively. Conclusion: Stroke was the most common cause of severely impaired consciousness. Age could be a useful indicator to consider intoxication and psychiatric disorders. Systolic blood pressure, airway compromise, and ocular abnormalities were factors associated with stroke in the prehospital setting.

Association between ICU admission during off-hours and in-hospital mortality: a multicenter registry in Japan.

BACKGROUND: The effect of ICU admission time on patient outcomes has been shown to be controversial in several studies from a number of countries. The imbalance between ICU staffing and medical resources during off-hours possibly influences the outcome for critically ill or injured patients. Here, we aimed to evaluate the association between ICU admission during off-hours and in-hospital mortality in Japan. METHODS: This study was an observational study using a multicenter registry (Japanese Intensive care PAtient Database). From the registry, we enrolled adult patients admitted to ICUs from April 2015 to March 2019. Patients with elective surgery, readmission to ICUs, or ICU admissions only for medical procedures were excluded. We compared in-hospital mortalities between ICU patients admitted during off-hours and office-hours, using a multilevel logistic regression model which allows for the random effect of each hospital. RESULTS: A total of 28,200 patients were enrolled with a median age of 71 years (interquartile range [IQR], 59 to 80). The median APACHE II score was 18 (IQR, 13 to 24) with no significant difference between patients admitted during off-hours and those admitted during office-hours. The in-hospital mortality was 3399/20,403 (16.7%) when admitted during off-hours and 1604/7797 (20.6%) when admitted during office-hours. Thus, off-hours ICU admission was associated with lower in-hospital mortality (adjusted odds ratio 0.91, [95% confidence interval, 0.84-0.99]). CONCLUSIONS: ICU admissions during off-hours were associated with lower in-hospital mortality in Japan. These results were against our expectations and raised some concerns for a possible imbalance between ICU staffing and workload during office-hours. Further studies with a sufficient dataset required for comparing with other countries are warranted in the future.

The necessity of lumbar puncture in adult emergency patients with fever-associated seizures.

OBJECTIVE: Central nervous system (CNS) infections are often suspected in adult patients with fever-associated seizures. However, it is unclear whether lumbar puncture (LP) is routinely required in patients with fever-associated seizures. This study aimed to examine the prevalence of meningitis and encephalitis in adult patients with fever-associated seizures and to evaluate whether LP is routinely required. METHODS: We retrospectively studied patients aged ≥16 years who presented to the emergency department with complaints of seizures and fever above 37.5 °C who were admitted to the hospital between January 2017 and December 2019. LP was performed when the emergency physician suspected meningitis or encephalitis. Neurologists assessed patients with normal cerebrospinal fluid (CSF) findings and those admitted without LP after hospitalization. A neurologist confirmed the diagnoses of meningitis and encephalitis. RESULTS: The study included 148 patients. Ninety-seven patients (65.5%) were male, and the median age was 60 years. LP was performed in 105 patients (70.9%), and 14 (13.4%) had CSF pleocytosis. Meningitis and encephalitis were diagnosed in nine patients (6.1%), of whom four (2.8%) had CNS infections. Patients diagnosed with meningitis and encephalitis were more likely to have Glasgow Coma Scale <13 (P = 0.03) and less likely to have a history of seizures or epilepsy (P = 0.04) and had higher C-reactive protein levels than the other patients (P = 0.02). CONCLUSION: The prevalence of meningitis or encephalitis is relatively low in adult patients with fever-associated seizures. Lumbar puncture is considered unnecessary to be performed routinely, but its indication should be carefully considered with reference to the clinical course, comorbidities, and blood tests. Further validation studies with larger sample sizes are needed to confirm the findings of this study.

Prehospital stroke notification and endovascular therapy for large vessel occlusion: a retrospective cohort study.

The impact of prehospital notification by emergency medical services (EMS) on outcomes of endovascular therapy (EVT) for large vessel occlusion (LVO) remains unclear. We therefore explored the association between prehospital notification and clinical outcomes after EVT. In this single-center retrospective study from 2016 through 2020, we identified all LVO patients who received EVT. Based on the EMS's usage of a prehospital stroke notification system, we categorized patients into two groups, Hotline and Non-hotline. The primary outcome was good neurological outcome at 90 days; other time metrics were also evaluated. Of all 312 LVO patients, the proportion of good neurological outcomes was 94/218 (43.1%) in the Hotline group and 8/34 (23.5%) in the Non-hotline group (adjusted odds ratio 2.86; 95% confidence interval 1.12 to 7.33). Time from hospital arrival to both tissue plasminogen activator and to groin puncture were shorter in the Hotline group (30 (24 to 38) min vs 48(37 to 65) min, p < 0.001; 40 (32 to 54) min vs 76 (50 to 97) min, p < 0.001), respectively. In conclusion, prehospital notification was associated with a reduction in time from hospital arrival to intervention and improved clinical outcomes in LVO patients treated with EVT.

Extracorporeal cardiopulmonary resuscitation in adult patients with out-of-hospital cardiac arrest: a retrospective large cohort multicenter study in Japan.

BACKGROUND: The prevalence of extracorporeal cardiopulmonary resuscitation (ECPR) in patients with out-of-hospital cardiac arrest (OHCA) has been increasing rapidly worldwide. However, guidelines or clinical studies do not provide sufficient data on ECPR practice. The aim of this study was to provide real-world data on ECPR for patients with OHCA, including details of complications. METHODS: We did a retrospective database analysis of observational multicenter cohort study in Japan. Adult patients with OHCA of presumed cardiac etiology who received ECPR between 2013 and 2018 were included. The primary outcome was favorable neurological outcome at hospital discharge, defined as a cerebral performance category of 1 or 2. RESULTS: A total of 1644 patients with OHCA were included in this study. The patient age was 18-93 years (median: 60 years). Shockable rhythm in the initial cardiac rhythm at the scene was 69.4%. The median estimated low flow time was 55 min (interquartile range: 45-66 min). Favorable neurological outcome at hospital discharge was observed in 14.1% of patients, and the rate of survival to hospital discharge was 27.2%. The proportions of favorable neurological outcome at hospital discharge in terms of shockable rhythm, pulseless electrical activity, and asystole were 16.7%, 9.2%, and 3.9%, respectively. Complications were observed during ECPR in 32.7% of patients, and the most common complication was bleeding, with the rates of cannulation site bleeding and other types of hemorrhage at 16.4% and 8.5%, respectively. CONCLUSIONS: In this large cohort, data on the ECPR of 1644 patients with OHCA show that the proportion of favorable neurological outcomes at hospital discharge was 14.1%, survival rate at hospital discharge was 27.2%, and complications were observed during ECPR in 32.7%.

Visceral disseminated varicella zoster infection: a rare cause of acute abdomen in a patient with well-controlled diabetes mellitus-a case report.

BACKGROUND: Visceral disseminated varicella zoster virus (VZV) infections frequently affect immunocompromised patients. Diabetes mellitus has been associated with VZV infection, and most cases of disseminated infection involve patients with poorly controlled blood glucose levels. It initially presents as severe abdominal pain, which is evaluated as an acute abdomen, however, the cause is typically unidentified due to unremarkable computed tomography (CT) findings. We report a case of visceral disseminated VZV infection in a patient with well-controlled diabetes mellitus with fat stranding around the celiac and superior mesenteric artery on CT. CASE PRESENTATION: A 61-year-old Japanese woman with well-controlled diabetes mellitus presented to the emergency department with severe abdominal pain that gradually worsened. She had stable vital signs and skin rashes, suggestive of varicella. Abdominal CT showed fat stranding around the celiac and superior mesenteric arteries. The patient tested positive for the VZV antigen and was diagnosed with a visceral disseminated VZV infection. Acyclovir was administered, and the patient was discharged on the 14th day. CONCLUSIONS: Visceral disseminated VZV infection may affect patients with well-controlled diabetes mellitus and causes acute abdomen. Periarterial fat stranding on CT is associated with abdominal pain due to visceral disseminated VZV infection.

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.

Premorbid Clinical Frailty Score and 30-day mortality among older adults in the emergency department.

OBJECTIVES: The association between frailty and short-term prognosis has not been established in critically ill older adults presenting to the emergency department. We sought to examine the association between premorbid frailty and 30-day mortality in this patient population. METHODS: This is a retrospective observational study on older adults aged over 75 who were triaged as Level 1 resuscitation with subsequent admissions to intermediate units or intensive care units (ICUs) in a single critical care center, from January to December 2019. We excluded patients with out-of-hospital cardiac arrest or those transferred from other hospitals. Frailty was evaluated by the Clinical Frailty Scale (CFS) from the patients' chart reviews. The primary outcome was 30-day mortality, and we examined the association between frailty scored on the CFS and 30-day mortality using a multivariable logistic regression model with CFS 1-4 as a reference. RESULTS: A total of 544 patients, median age: 82 years (interquartile rang 78 to 87), were included in the study. Of these, 29% were in shock and 33% were in respiratory failure. The overall 30-day mortality was 15.1%. The adjusted risk difference (95% confidence interval [CI]) in mortality for CFS 5, CFS 6, and CFS 7-9 was 6.3% (-3.4 to 15.9), 11.2% (0.4 to 22.0), and 17.7% (5.3 to 30.1), respectively; and the adjusted risk ratio (95% CI) was 1.45 (0.87 to 2.41), 1.85 (1.13 to 3.03), and 2.44 (1.50 to 3.96), respectively. CONCLUSION: The risk of 30-day mortality increased as frailty advanced in critically ill older adults. Given this high risk of short-term outcomes, ED clinicians should consider goals of care conversations carefully to avoid unwanted medical care for these patients.

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.