Cycle-consistent adversarial networks improves generalizability of radiomics model in grading meningiomas on external validation.

The heterogeneity of MRI is one of the major reasons for decreased performance of a radiomics model on external validation, limiting the model's generalizability and clinical application. We aimed to establish a generalizable radiomics model to predict meningioma grade on external validation through leveraging Cycle-Consistent Adversarial Networks (CycleGAN). In this retrospective study, 257 patients with meningioma were included in the institutional training set. Radiomic features (n = 214) were extracted from T2-weighted (T2) and contrast-enhanced T1 (T1C) images. After radiomics feature selection, extreme gradient boosting classifiers were developed. The models were validated in the external validation set consisting of 61 patients with meningiomas. To reduce the gap in generalization associated with the inter-institutional heterogeneity of MRI, the smaller image set style of the external validation was translated into the larger image set style of the institutional training set using CycleGAN. On external validation before CycleGAN application, the performance of the combined T2 and T1C models showed an area under the curve (AUC), accuracy, and F1 score of 0.77 (95% confidence interval 0.63-0.91), 70.7%, and 0.54, respectively. After applying CycleGAN, the performance of the combined T2 and T1C models increased, with an AUC, accuracy, and F1 score of 0.83 (95% confidence interval 0.70-0.97), 73.2%, and 0.59, respectively. Quantitative metrics (by Fréchet Inception Distance) showed that CycleGAN can decrease inter-institutional image heterogeneity while preserving predictive information. In conclusion, leveraging CycleGAN may be helpful to increase the generalizability of a radiomics model in differentiating meningioma grade on external validation.

Modeling Brain Volume Using Deep Learning-Based Physical Activity Features in Patients With Dementia.

There is a proven correlation between the severity of dementia and reduced brain volumes. Several studies have attempted to use activity data to estimate brain volume as a means of detecting reduction early; however, raw activity data are not directly interpretable and are unstructured, making them challenging to utilize. Furthermore, in the previous research, brain volume estimates were limited to total brain volume and the investigators were unable to detect reductions in specific regions of the brain that are typically used to characterize disease progression. We aimed to evaluate volume prediction of 116 brain regions through activity data obtained combining time-frequency domain- and unsupervised deep learning-based feature extraction methods. We developed a feature extraction model based on unsupervised deep learning using activity data from the National Health and Nutrition Examination Survey (NHANES) dataset (n = 14,482). Then, we applied the model and the time-frequency domain feature extraction method to the activity data of the Biobank Innovations for chronic Cerebrovascular disease With ALZheimer's disease Study (BICWALZS) datasets (n = 177) to extract activity features. Brain volumes were calculated from the brain magnetic resonance imaging of the BICWALZS dataset and anatomically subdivided into 116 regions. Finally, we fitted linear regression models to estimate each regional volume of the 116 brain areas based on the extracted activity features. Regression models were statistically significant for each region, with an average correlation coefficient of 0.990 ± 0.006. In all brain regions, the correlation was > 0.964. Particularly, regions of the temporal lobe that exhibit characteristic atrophy in the early stages of Alzheimer's disease showed the highest correlation (0.995). Through a combined deep learning-time-frequency domain feature extraction method, we could extract activity features based solely on the activity dataset, without including clinical variables. The findings of this study indicate the possibility of using activity data for the detection of neurological disorders such as Alzheimer's disease.

Effect of direct-acting antivirals for hepatitis C virus-related hepatocellular carcinoma recurrence and death after curative treatment.

Background/Aim: There has been a long-standing debate about the association of directacting antiviral (DAA) therapy and hepatocellular carcinoma (HCC) recurrence. This study aimed to investigate the association between DAA therapy and HCC recurrence after curative therapy. Methods: We retrospectively enrolled 1,021 patients with HCV-related (hepatitis C virus) HCC who underwent radiofrequency ablation (RFA), liver resection, or both as the first treatment modality from January 2007 to December 2016 and without a history of HCV therapy before HCC treatment from a nationwide database. The effect of HCV treatment on HCC recurrence and all-cause mortality was also investigated. Results: Among the 1,021 patients, 77 (7.5%) were treated with DAA, 14 (1.4%) were treated with interferon-based therapy, and 930 (91.1%) did not receive HCV therapy. DAA therapy was an independent prognostic factor for lower HCC recurrence rate (hazard ratio [HR], 0.04; 95% confidence interval [CI], 0.006-0.289; P=0.001 for landmarks at 6 months after HCC treatment and HR, 0.05; 95% CI, 0.007-0.354; P=0.003 for landmarks at 1 year). Furthermore, DAA therapy was associated with lower all-cause mortality (HR, 0.049; 95% CI, 0.007-0.349; P=0.003 for landmarks at 6 months and HR, 0.063; 95% CI, 0.009-0.451; P=0.006 for landmarks at 1 year). Conclusions: DAA therapy after curative HCC treatment can decrease HCC recurrence and all-cause mortality compared to interferon-based therapy or no antiviral therapy. Therefore, clinicians should consider administering DAA therapy after curative HCC treatment in patients with HCV-related HCC.

Independent Risk Factors for Hepatocellular Carcinoma Recurrence after Direct-Acting Antiviral Therapy in Patients with Chronic Hepatitis C.

Background/Aims: This study was performed to evaluate the efficacy of direct-acting antivirals (DAAs) in Korean patients with hepatitis C virus (HCV)-related hepatocellular carcinoma (HCC) and to investigate the risk factors associated with HCC recurrence. Methods: A total of 100 patients with HCV-related HCC, who were treated with DAAs between May 2015 and December 2016, were recruited from seven university hospitals in Korea. Claim data of 526 patients with HCC obtained from the Health Insurance Review and Assessment Service in South Korea were used for external validation of the results. Results: Among the 100 patients, 88% achieved a sustained virological response (SVR) 12 weeks after the end of DAA therapy (SVR12), and 37% experienced HCC recurrence after DAA therapy. Short last HCC treatment durability (<12 months) before DAA commencement was independently associated with HCC recurrence (hazard ratio [HR], 2.89; p=0.011). In the nationwide validation cohort, 20.3% of the patients experienced HCC recurrence. The last HCC treatment with a noncurative method, a short last HCC treatment durability (<12 months), and a longer total duration of HCC treatment (≥18 months) were independently related with HCC recurrence (HR 3.73, p<0.001; HR 3.34, p<0.001; and HR 1.74, p=0.006; respectively). Conclusions: DAA therapy showed an acceptable SVR12 rate in patients with HCV-related HCC. Short last HCC treatment durability (<12 months) was associated with HCC recurrence after DAA therapy. This finding suggests that the last HCC treatment durability is an important predictor of HCC recurrence after DAA therapy.

Prognostic Value of Skeletal Muscle Depletion Measured on Computed Tomography for Overall Survival in Patients with Non-Metastatic Breast Cancer.

PURPOSE: The purpose of this study was to evaluate the prognostic value of skeletal muscle depletion measured on computed tomography (CT) in patients with non-metastatic invasive breast cancer. METHODS: This retrospective study included 577 consecutive women (mean age ± standard deviation: 48.9 ± 10.2 years with breast cancer who underwent a preoperative positron-emission tomography (PET)/CT scan and curative surgery between January 2012 and August 2014. The total abdominal muscle area (TAMA), subcutaneous fat area (SFA), and visceral fat area (VFA) were measured on CT images at the L3 vertebral level. Univariate and multivariate Cox proportional-hazard regression analyses were performed to evaluate whether there was an association between sarcopenia and overall survival (OS) outcome. RESULTS: Of the 577 women, 49 (8.5%) died after a mean of 46 months. The best TAMA threshold for predicting OS was 83.7 cm2. The multivariate Cox proportional-hazard analysis revealed that sarcopenia (TAMA ≤ 83.70 cm2) was a strong prognostic biomarker (hazard ratio [HR], 1.951; 95% confidence interval [CI], 1.061-3.586), along with large tumor size, axillary lymph node metastasis, high nuclear grade, estrogen receptor status, and adjuvant radiation therapy. In the subgroup analysis of patients aged ≥ 50 years, TAMA (≤ 77.14 cm2) was a significant independent factor (HR, 2.856; 95% CI, 1.218-6.695). CONCLUSION: Skeletal muscle depletion measured on CT was associated with worse OS outcome in patients with non-metastatic breast cancer.

Effect of Pulmonary Hypertension on Clinical Outcomes in Patients With Rheumatic Mitral Stenosis.

BACKGROUND: Long-term outcomes of mitral valve replacement (MVR) recipients with pulmonary hypertension (PH) remain unclear. We determined the effect of concomitant PH on the clinical outcomes and late-onset progression of tricuspid regurgitation (TR) after MVR for rheumatic mitral stenosis. METHODS: We retrospectively reviewed 394 patients who underwent MVR between January 2000 and December 2013. PH was defined as systolic pulmonary arterial pressure (sPAP) >50 mm Hg. Changes in echocardiographic parameters (preoperative to postoperative), TR progression (grade ≥ II), and long-term survival were evaluated according to the presence of PH at MVR (non-PH, n = 322; PH, n = 72). RESULTS: The 10-year overall survival rate was significantly lower in the PH group (79.7% vs 90.7%, P = .04), whereas the rate of freedom from TR progression was similar between groups (76.9% vs 80.5%, P = .373). High preoperative sPAP and right ventricular systolic pressure (RVSP) did not affect TR progression. However, substantial postoperative reductions in sPAP and RVSP protected against TR progression (hazard ratio [95% confidence interval], 0.966 [0.942-0.991], P = .008, and 0.973 [0.960-0.986], P < .001, respectively). The 10-year rate of freedom from TR progression was significantly higher in patients with substantial sPAP and RVSP reductions (sPAP, 84.2% vs 70.6%, P = .003; RVSP, 84.9% vs 71.0%, P < .001). CONCLUSIONS: Although concomitant PH at MVR is associated with poor long-term survival, adequate sPAP and RVSP reduction can prevent TR progression even in patients with severe PH preoperatively. Therefore, we suggest careful monitoring of PAP and RVSP after MVR and should make an effort to reduce postoperative PAP.

Outcomes of transcatheter closure of ductus arteriosus in infants less than 6 months of age: a single-center experience.

PURPOSE: Transcatheter device closure of patent ductus arteriosus (PDA) is challenging in early infancy. We evaluated PDA closure in infants less than 6 months old. METHODS: We performed a retrospective review of infants less than 6 months of age who underwent attempted transcatheter device closure in our institution since 2004. To compare clinical outcomes between age groups, infants aged 6-12 months in the same study period were reviewed. RESULTS: A total of 22 patients underwent transcatheter PDA closure during the study period. Patient mean age was 3.3±1.5 months, and weight was 5.7±1.3 kg. The duct diameter at the narrowest point was 3.0±0.8 mm as measured by angiography. The most common duct type was C in the Krichenko classification. Procedural success was achieved in 19 patients (86.3%). Major complications occurred in 5 patients (22.7%), including device embolization (n=1), acquired aortic coarctation (n=2), access-related vascular injury requiring surgery (n=1), and acute deterioration requiring intubation during the procedure (n=1). Two patients had minor complications (9.1%). Twenty-four infants aged 6-12 months received transcatheter device closure. The procedural success rate was 100%, and there were no major complications. The major complication rate was significantly higher in the group less than 6 months of age (P=0.045). There was a trend toward increased major complication and procedural failure rates in the younger age group (P<0.01). CONCLUSION: A relatively higher incidence of major complications was observed in infants less than 6 months of age. The decision regarding treatment modality should be individualized.

The effect of balloon valvuloplasty for bioprosthetic valve stenosis at pulmonary positions.

BACKGROUND: Balloon dilatation of a bioprosthetic valve in the pulmonary position could be performed to delay valve replacement. We proposed to identify the long-term effectiveness of such a procedure. METHODS: We reviewed the medical records of 49 patients who underwent balloon valvuloplasty between January 2000 and December 2015. The primary goal was to determine the time interval until the following surgical or catheter intervention. RESULTS: The mean age at bioprosthetic valve insertion was 5.7 years old, and the mean age for ballooning was 11.7 years. The mean interval after pulmonary valve replacement was 71.6 months. The mean ratio of balloon size to valve size was 0.94. The pressure gradient through the pulmonary valve after balloon valvuloplasty was significantly improved (55.3 ± 18.5 mm Hg vs 33.8 ± 21.5 mm Hg, P < .001). There were no significant changes in pulmonary regurgitation and no serious adverse events. Patients had a mean freedom from re-intervention of 30.6 months after balloon valvuloplasty. The interval of freedom from re-intervention was affected only by the pressure gradient before balloon valvuloplasty and the patient age at insertion. The mean interval to re-intervention in patients with pressure gradients less than 48.5 mm Hg before ballooning was 46.0 months, which was significantly longer than for those with a higher gradient (18.7 months). CONCLUSION: The effectiveness of this process may depend on the pressure gradient before ballooning and the patient age at valve insertion. It is possible that earlier valvuloplasty at pressure gradient not over 48.5mm Hg may have a benefit to delaying re-operation.