Computer-aided diagnosis of chest X-ray for COVID-19 diagnosis in external validation study by radiologists with and without deep learning system.

To evaluate the diagnostic performance of our deep learning (DL) model of COVID-19 and investigate whether the diagnostic performance of radiologists was improved by referring to our model. Our datasets contained chest X-rays (CXRs) for the following three categories: normal (NORMAL), non-COVID-19 pneumonia (PNEUMONIA), and COVID-19 pneumonia (COVID). We used two public datasets and private dataset collected from eight hospitals for the development and external validation of our DL model (26,393 CXRs). Eight radiologists performed two reading sessions: one session was performed with reference to CXRs only, and the other was performed with reference to both CXRs and the results of the DL model. The evaluation metrics for the reading session were accuracy, sensitivity, specificity, and area under the curve (AUC). The accuracy of our DL model was 0.733, and that of the eight radiologists without DL was 0.696 ± 0.031. There was a significant difference in AUC between the radiologists with and without DL for COVID versus NORMAL or PNEUMONIA (p = 0.0038). Our DL model alone showed better diagnostic performance than that of most radiologists. In addition, our model significantly improved the diagnostic performance of radiologists for COVID versus NORMAL or PNEUMONIA.

Retrospective comparison of clinical outcomes of ultrasound-guided percutaneous cholecystostomy in patients with and without coagulopathy: a single center's experience.

PURPOSE: To compare the complication rate and clinical outcomes for percutaneous cholecystostomy (PC) in patients with or without coagulopathy. MATERIALS AND METHODS: We retrospectively reviewed electronic medical chart of patients who underwent ultrasound-guided PC with a 8.5-F drainage tube for acute cholecystitis between November 2003 and March 2017. We divided the patients into two groups: patients with coagulopathy (international normalized ratio > 1.5 or platelet count < 50 × 109/L or with a history of anticoagulant medication in preceding 5 days) and patients without coagulopathy. Duration of drainage, duration of hospital stay, 30-day mortality and complication rates were compared between these two groups. Student's t test, Chi-square test or Fisher's exact test was used for bivariate analyses. Age, age-adjusted Charlson Comorbidity Index (ACCI) and sepsis-adjusted complication rates were also compared. RESULTS: In total, 141 patients had PC (mean age was 73.3 years [SD 13.3]; range 33-96 years; 94 men and 47 women). Fifty-two patients (36.9%) had coagulopathy and 89 patients (63.1%) were without any history of coagulopathy. Hemorrhagic complication rate was 3.5% (5 out of 141 patients, including 4 with coagulopathy and 1 without). One patient with coagulopathy died due to the hemorrhage. Duration of drainage was longer in patients with coagulopathy than patients without coagulopathy (20.0 days vs. 14.8 days; P = 0.033). No significant difference was observed with regard to duration of hospital stay (32.3 days vs. 25.6 days; P = 0.103) and 30-day mortality (7.7% vs. 1.1%; P = 0.062). The overall complication rate did not significantly differ (9.6% and 11.2%; P = 0.763), nor did age, ACCI or sepsis-adjusted complications. CONCLUSION: Clinical outcomes and complications rates after PC did not statistically differ between patients with and without coagulopathy, but there was a tendency of higher risk of hemorrhage in coagulopathy patients. Therefore, the indication of this procedure should be carefully determined.

Long-Term Outcome of Percutaneous Biliary Interventions for Biliary Anastomotic Stricture in Pediatric Patients after Living Donor Liver Transplantation with Roux-en-Y Hepaticojejunostomy.

PURPOSE: To retrospectively evaluate long-term outcomes of percutaneous transhepatic biliary drainage (PTBD) followed by balloon dilation and placement of an internal drainage tube for anastomotic stricture in pediatric patients who underwent living donor liver transplantation (LDLT) with Roux-en-Y hepaticojejunostomy (RYHJ). MATERIALS AND METHODS: Fifty-two patients (23 male, 29 female; median age, 5 y) with anastomotic biliary stricture were treated with PTBD followed by balloon catheter dilation and long-term placement of an internal drainage tube, which was removed upon cholangiographic confirmation of free flow of bile into the small bowel. Clinical success, tube independence rate, risk factors of recurrent biliary stricture, and patency rates were evaluated. RESULTS: Thirty-nine patients (75%) had no stricture recurrence. Of 13 patients (25%) with recurrence, six were treated again with the same percutaneous biliary interventions and showed no further recurrence. Clinical success was noted in 43 of 52 patients (83%). Drainage tubes were removed from 49 patients (94%). Multivariate logistic regression analysis indicated that serum alanine aminotransferase level > 53 IU/L at discharge after the initial series of percutaneous biliary interventions was a significant risk factor for recurrent biliary stricture (P = .002). Kaplan-Meier analysis showed 1-, 3-, 5-, and 10-year primary and primary assisted patency rates of 75%, 70%, 70%, and 68%, and 94%, 92%, 88%, and 88%, respectively. CONCLUSIONS: PTBD followed by balloon dilation and internal drainage may be an effective treatment for anastomotic biliary stricture after pediatric LDLT with RYHJ.

Long-term outcome of percutaneous transhepatic balloon angioplasty for portal vein stenosis after pediatric living donor liver transplantation: a single institute's experience.

PURPOSE: To evaluate retrospectively the long-term outcomes of percutaneous transhepatic balloon angioplasty performed for portal vein stenosis (PVS) after pediatric living donor liver transplantation (LDLT). MATERIALS AND METHODS: Between October 1997 and December 2013, of 527 pediatric patients (age < 18 y) who underwent LDLT in a single institution, 43 patients (19 boys, 24 girls; mean age, 4.1 y ± 4.1) were confirmed to have PVS at direct portography with or without manometry and underwent percutaneous interventions, including balloon angioplasty with or without stent placement. Technical success, clinical success, laboratory findings, manometry findings, patency rates, and major complications were evaluated. Follow-up periods after initial balloon angioplasty ranged from 5-169 months (mean, 119 mo). RESULTS: Technical success was achieved in 65 of 66 sessions (98.5%) and in 42 of 43 patients (97.7%), and clinical success was achieved in 37 of 43 patients (86.0%). Platelet counts improved significantly. Of 32 patients undergoing manometry, 19 showed significant improvement of pressure gradient across the stenosis after percutaneous transhepatic balloon angioplasty. At 1, 3, 5, and 10 years after balloon angioplasty, the rates of primary patency were 83%, 78%, 76%, and 70%, and the rates of primary-assisted patency were 100%, 100%, 100%, and 96%. Two major complications subsequent to balloon angioplasty were noted: severe asthma attack and portal vein thrombosis. CONCLUSIONS: Percutaneous transhepatic balloon angioplasty is a safe and effective treatment with long-term patency for PVS after pediatric LDLT.

Long-term outcome of percutaneous interventions for hepatic venous outflow obstruction after pediatric living donor liver transplantation: experience from a single institute.

PURPOSE: To evaluate retrospectively the long-term outcome of percutaneous interventions for hepatic venous outflow obstruction (HVOO) occurring after pediatric living donor liver transplantation (LDLT). MATERIALS AND METHODS: Between October 1997 and December 2012, 48 patients (24 boys, 24 girls; median age, 6 y) who had undergone LDLT were confirmed to have HVOO using percutaneous hepatic venography and manometry. All patients underwent percutaneous interventions, including balloon angioplasty with or without stent placement. Technical success, clinical success, patency rates, stent placement, and major complications were evaluated. RESULTS: Technical success was achieved in 92 of 93 sessions (99.0%) and in 47 of 48 patients (97.9%), and clinical success was achieved in 41 of 48 patients (85.4%). During the follow-up period (range, 1-182 mo; median, 51.5 mo), 28 patients were treated with a single session of balloon angioplasty, and 20 patients who developed recurrent stenosis were treated with repeated percutaneous interventions. The rates of primary and primary-assisted patency at 1, 3, 5, and 10 years after balloon angioplasty were 64%, 57%, 57%, and 52% (primary patency) and 98%, 95%, 95%, and 95% (primary-assisted patency). Of six patients with stent placement, four had no recurrent HVOO after the stent placement, but two developed recurrent stenosis. The stent migrated to the right atrium in one patient. CONCLUSIONS: Percutaneous interventions were effective treatments for HVOO after LDLT.