Machine Learning Models for ASCVD Risk Prediction in an Asian Population - How to Validate the Model is Important.

Introduction: Atherosclerotic cardiovascular disease (ASCVD) is prevalent worldwide including Taiwan, however widely accepted tools to assess the risk of ASCVD are lacking in Taiwan. Machine learning models are potentially useful for risk evaluation. In this study we used two cohorts to test the feasibility of machine learning with transfer learning for developing an ASCVD risk prediction model in Taiwan. Methods: Two multi-center observational registry cohorts, T-SPARCLE and T-PPARCLE were used in this study. The variables selected were based on European, U.S. and Asian guidelines. Both registries recorded the ASCVD outcomes of the patients. Ten-fold validation and temporal validation methods were used to evaluate the performance of the binary classification analysis [prediction of major adverse cardiovascular (CV) events in one year]. Time-to-event analyses were also performed. Results: In the binary classification analysis, eXtreme Gradient Boosting (XGBoost) and random forest had the best performance, with areas under the receiver operating characteristic curve (AUC-ROC) of 0.72 (0.68-0.76) and 0.73 (0.69-0.77), respectively, although it was not significantly better than other models. Temporal validation was also performed, and the data showed significant differences in the distribution of various features and event rate. The AUC-ROC of XGBoost dropped to 0.66 (0.59-0.73), while that of random forest dropped to 0.69 (0.62-0.76) in the temporal validation method, and the performance also became numerically worse than that of the logistic regression model. In the time-to-event analysis, most models had a concordance index of around 0.70. Conclusions: Machine learning models with appropriate transfer learning may be a useful tool for the development of CV risk prediction models and may help improve patient care in the future.

Large-Scale Structural Covariance Networks Predict Age in Middle-to-Late Adulthood: A Novel Brain Aging Biomarker.

The aging process is accompanied by changes in the brain's cortex at many levels. There is growing interest in summarizing these complex brain-aging profiles into a single, quantitative index that could serve as a biomarker both for characterizing individual brain health and for identifying neurodegenerative and neuropsychiatric diseases. Using a large-scale structural covariance network (SCN)-based framework with machine learning algorithms, we demonstrate this framework's ability to predict individual brain age in a large sample of middle-to-late age adults, and highlight its clinical specificity for several disease populations from a network perspective. A proposed estimator with 40 SCNs could predict individual brain age, balancing between model complexity and prediction accuracy. Notably, we found that the most significant SCN for predicting brain age included the caudate nucleus, putamen, hippocampus, amygdala, and cerebellar regions. Furthermore, our data indicate a larger brain age disparity in patients with schizophrenia and Alzheimer's disease than in healthy controls, while this metric did not differ significantly in patients with major depressive disorder. These findings provide empirical evidence supporting the estimation of brain age from a brain network perspective, and demonstrate the clinical feasibility of evaluating neurological diseases hypothesized to be associated with accelerated brain aging.

The pathophysiology of episodic cluster headache: Insights from recent neuroimaging research.

Background Cluster headache is a disorder characterized by intermittent, severe unilateral head pain accompanied by cranial autonomic symptoms. Most cases of CH are episodic, manifesting as "in-bout" periods of frequent headache separated by month-to-year-long "out-of-bout" periods of remission. Previous imaging studies have implicated the hypothalamus and pain matrix in the pathogenesis of episodic CH. However, the pathophysiology driving the transition between in- and out-of-bout periods remains unclear. Methods The present study provides a narrative review of previous neuroimaging studies on the pathophysiology of episodic CH, addressing alterations in brain structures, metabolism, and structural and functional connectivity occurring between bout periods. Results Although the precise brain structures responsible for episodic CH are unknown, major roles are indicated for the posterior hypothalamus (especially in acute attacks), the pain neuromatrix with an emphasis on central descending pain modulation, and non-traditional pain processing networks including the occipital, cerebellar, and salience networks. These areas are potentially related to dynamic transitioning between in- and out-of-bout periods. Conclusion Recent progress in magnetic resonance imaging of episodic CH has provided additional insights into dynamic bout-associated structural and functional connectivity changes in the brain, especially in non-traditional pain processing network areas. These areas warrant future investigations as targets for neuromodulation in patients with CH.

Bout-associated intrinsic functional network changes in cluster headache: A longitudinal resting-state functional MRI study.

Background Previous imaging studies on the pathogenesis of cluster headache (CH) have implicated the hypothalamus and multiple brain networks. However, very little is known regarding dynamic bout-associated, large-scale resting state functional network changes related to CH. Methods Resting-state functional magnetic resonance imaging data were obtained from CH patients and matched controls. Data were analyzed using independent component analysis for exploratory assessment of the changes in intrinsic brain networks and their relationship between in-bout and out-of-bout periods, as well as correlations with clinical observations. Results Compared to healthy controls, CH patients had functional connectivity (FC) changes in the temporal, frontal, salience, default mode, somatosensory, dorsal attention, and visual networks, independent of bout period. Compared to out-of-bout scans, in-bout scans showed altered FC in the frontal and dorsal attention networks. Lower frontal network FC correlated with longer duration of CH. Conclusions The present findings suggest that episodic CH with dynamic bout period shifts may involve bout-associated FC changes in multiple discrete cortical areas within networks outside traditional pain processing areas. Dynamic changes in FC in frontal and dorsal attention networks between bout periods could be important for understanding episodic CH pathophysiology.

Multiscale brain age prediction reveals region-specific accelerated brain aging in Parkinson's disease.

Brain biological age, which measures the aging process in the brain using neuroimaging data, has been used to assess advanced brain aging in neurodegenerative diseases, including Parkinson disease (PD). However, assuming that whole brain degeneration is uniform may not be sufficient for assessing the complex neurodegenerative processes in PD. In this study we constructed a multiscale brain age prediction models based on structural MRI of 1240 healthy participants. To assess the brain aging patterns using the brain age prediction model, 93 PD patients and 91 healthy controls matching for sex and age were included. We found increased global and regional brain age in PD patients. The advanced aging regions were predominantly noted in the frontal and temporal cortices, limbic system, basal ganglia, thalamus, and cerebellum. Furthermore, region-level rather than global brain age in PD patients was associated with disease severity. Our multiscale brain age prediction model could aid in the development of objective image-based biomarkers to detect advanced brain aging in neurodegenerative diseases.

Advanced brain age in community-dwelling population with combined physical and cognitive impairments.

We investigated whether advanced brain biological age is associated with accelerated age-related physical and/or cognitive functional decline: mobility impairment no disability (MIND), cognitive impairment no dementia (CIND), and physio-cognitive decline syndrome (PCDS). We constructed a brain age prediction model using gray matter features from the magnetic resonance imaging of 1482 healthy individuals (aged 18-92 years). Predicted and chronological age differences were obtained (brain age gap [BAG]) and analyzed in another 1193 community-dwelling population aged ≥50 years. Among the 1193 participants, there were 501, 346, 148, and 198 in the robust, CIND, MIND, and PCDS groups, respectively. Participants with PCDS had significantly larger BAG (BAG = 2.99 ± 8.97) than the robust (BAG = -0.49 ± 9.27, p = 0.002; η2 = 0.014), CIND (BAG = 0.47 ± 9.16, p = 0.02; η2 = 0.01), and MIND (BAG = 0.36 ± 9.69, p = 0.036; η2 = 0.013) groups. Advanced brain aging is involved in the pathophysiology of the co-occurrence of physical and cognitive decline in the older people. The PCDS may be a clinical phenotype reflective of accelerated biological age in community-dwelling older individuals.

Shared alterations in hippocampal structural covariance in subjective cognitive decline and migraine.

Introduction: Subjective cognitive decline (SCD) and migraine are often comorbid. Hippocampal structural abnormalities have been observed in individuals with both SCD and migraine. Given the known structural and functional heterogeneity along the long axis (anterior to posterior) of the hippocampus, we aimed to identify altered patterns of structural covariance within hippocampal subdivisions associated with SCD and migraine comorbidities. Methods: A seed-based structural covariance network analysis was applied to examine large-scale anatomical network changes of the anterior and posterior hippocampus in individuals with SCD, migraine and healthy controls. Conjunction analyses were used to identify shared network-level alterations in the hippocampal subdivisions in individuals with both SCD and migraine. Results: Altered structural covariance integrity of the anterior and posterior hippocampus was observed in the temporal, frontal, occipital, cingulate, precentral, and postcentral areas in individuals with SCD and migraine compared with healthy controls. Conjunction analysis revealed that, in both SCD and migraine, altered structural covariance integrity was shared between the anterior hippocampus and inferior temporal gyri and between the posterior hippocampus and precentral gyrus. Additionally, the structural covariance integrity of the posterior hippocampus-cerebellum axis was associated with the duration of SCD. Conclusion: This study highlighted the specific role of hippocampal subdivisions and specific structural covariance alterations within these subdivisions in the pathophysiology of SCD and migraine. These network-level changes in structural covariance may serve as potential imaging signatures for individuals who have both SCD and migraine.

Regional rather than global brain age mediates cognitive function in cerebral small vessel disease.

The factors and mechanisms underlying the heterogeneous cognitive outcomes of cerebral small vessel disease are largely unknown. Brain biological age can be estimated by machine learning algorithms that use large brain MRI data sets to integrate and compute neuroimaging-derived age-related features. Predicted and chronological ages difference (brain-age gap) reflects advanced or delayed brain aging in an individual. The present study firstly reports the brain aging status of cerebral small vessel disease. In addition, we investigated whether global or certain regional brain age could mediate the cognitive functions in cerebral small vessel disease. Global and regional (400 cortical, 14 subcortical and 28 cerebellum regions of interest) brain-age prediction models were constructed using grey matter features from MRI of 1482 healthy individuals (age: 18-92 years). Predicted and chronological ages differences were obtained and then applied to non-stroke, non-demented individuals, aged ≥50 years, from another community-dwelling population (I-Lan Longitudinal Aging Study cohort). Among the 734 participants from the I-Lan Longitudinal Aging Study cohort, 124 were classified into the cerebral small vessel disease group. The cerebral small vessel disease group demonstrated significantly poorer performances in global cognitive, verbal memory and executive functions than that of non-cerebral small vessel disease group. Global brain-age gap was significantly higher in the cerebral small vessel disease (3.71 ± 7.60 years) than that in non-cerebral small vessel disease (-0.43 ± 9.47 years) group (P = 0.003, η2 = 0.012). There were 82 cerebral cortical, 3 subcortical and 4 cerebellar regions showing significantly different brain-age gap between the cerebral small vessel disease and non-cerebral small vessel disease groups. Global brain-age gap failed to mediate the relationship between cerebral small vessel disease and any of the cognitive domains. In 89 regions with increased brain-age gap in the cerebral small vessel disease group, seven regional brain-age gaps were able to show significant mediation effects in cerebral small vessel disease-related cognitive impairment (we set the statistical significance P < 0.05 uncorrected in 89 mediation models). Of these, the left thalamus and left hippocampus brain-age gap explained poorer global cognitive performance in cerebral small vessel disease. We demonstrated the interconnections between cerebral small vessel disease and brain age. Strategic brain aging, i.e. advanced brain aging in critical regions, may be involved in the pathophysiology of cerebral small vessel disease-related cognitive impairment. Regional rather than global brain-age gap could potentially serve as a biomarker for predicting heterogeneous cognitive outcomes in patients with cerebral small vessel disease.

Improving Individual Brain Age Prediction Using an Ensemble Deep Learning Framework.

Brain age is an imaging-based biomarker with excellent feasibility for characterizing individual brain health and may serve as a single quantitative index for clinical and domain-specific usage. Brain age has been successfully estimated using extensive neuroimaging data from healthy participants with various feature extraction and conventional machine learning (ML) approaches. Recently, several end-to-end deep learning (DL) analytical frameworks have been proposed as alternative approaches to predict individual brain age with higher accuracy. However, the optimal approach to select and assemble appropriate input feature sets for DL analytical frameworks remains to be determined. In the Predictive Analytics Competition 2019, we proposed a hierarchical analytical framework which first used ML algorithms to investigate the potential contribution of different input features for predicting individual brain age. The obtained information then served as a priori knowledge for determining the input feature sets of the final ensemble DL prediction model. Systematic evaluation revealed that ML approaches with multiple concurrent input features, including tissue volume and density, achieved higher prediction accuracy when compared with approaches with a single input feature set [Ridge regression: mean absolute error (MAE) = 4.51 years, R 2 = 0.88; support vector regression, MAE = 4.42 years, R 2 = 0.88]. Based on this evaluation, a final ensemble DL brain age prediction model integrating multiple feature sets was constructed with reasonable computation capacity and achieved higher prediction accuracy when compared with ML approaches in the training dataset (MAE = 3.77 years; R 2 = 0.90). Furthermore, the proposed ensemble DL brain age prediction model also demonstrated sufficient generalizability in the testing dataset (MAE = 3.33 years). In summary, this study provides initial evidence of how-to efficiency for integrating ML and advanced DL approaches into a unified analytical framework for predicting individual brain age with higher accuracy. With the increase in large open multiple-modality neuroimaging datasets, ensemble DL strategies with appropriate input feature sets serve as a candidate approach for predicting individual brain age in the future.

Shared Patterns of Brain Functional Connectivity for the Comorbidity between Migraine and Insomnia.

Migraine is commonly comorbid with insomnia; both disorders are linked to functional disturbance of the default mode network (DMN). Evidence suggests that DMN could be segregated into multiple subnetworks with specific roles that underline different cognitive processes. However, the relative contributions of DMN subnetworks in the comorbidity of migraine and insomnia remain largely unknown. This study sought to identify altered functional connectivity (FC) profiles of DMN subnetworks in the comorbidity of migraine and insomnia. Direct group comparisons with healthy controls, followed by conjunction analyses, were used to identify shared FC alterations of DMN subnetworks. The shared FC changes of the DMN subnetworks in the migraine and insomnia groups were identified in the dorsomedial prefrontal and posteromedial cortex subnetworks. These shared FC changes were primarily associated with motor and somatosensory systems, and consistently found in patients with comorbid migraine and insomnia. Additionally, the magnitude of FC between the posteromedial cortex and postcentral gyrus correlated with insomnia duration in patients with comorbid migraine and insomnia. Our findings point to specific FC alterations of the DMN subnetwork in migraine and insomnia. The shared patterns of FC disturbance may be associated with the underlying mechanisms of the comorbidity of the two disorders.

Restenotic hepaticojejunostomy secondary to laparoscopic cholecistectomy bile duct injury treated with self- modified Gianturco-Rosh stents.

BACKGROUND/AIMS: Laparoscopic cholecystectomy is considered as a standard procedure for symptomatic gallstones. However, the incidence of iatrogenic bile duct injury is higher that the conventional cholecystectomy. In the present study was analyzed the results in 6 patients with iatrogenic bile duct injury during laparoscopic cholecystectomy with restenotic hepaticojejunostomy treated with self-modified Gianturco-Rosch stents. METHODOLOGY: Data were collected retrospectively on May 2000 to October 2008 on six patients with major bile duct injury secondary to cholecystectomy. All patients underwent surgical reconstruction with a Roux-en-Y hepaticojejunostomy and presented clinically as obstructive jaundice. Percutaneous transhepatic and/or endoscopic retrograde cholangiography, cholangioplasty by balloon dilation and biliary catheter placement were done in each patient prior to stents placement. Modified Gianturco-Rosch stents with 3cm length and 10mm diameter were used. Follow-up was obtained with direct patients contact or hospital records. RESULTS: Metallic stents were successfully implanted in all 6 patients and the mean patency rate was 46.5 months (range = 14-101 months). One patient required percutaneous recanalization procedure for recurrent cholangitis and obstruction. CONCLUSIONS: Gianturco-Rosch stents placement should be considered in patient with post-hepaticojejunostomy restenosis that repeat surgery is not feasible.

Synthesis and antitumor activity of cis-dichloroplatinum(II) complexes of 1-(2-aminophenyl)-1,2,3,4-tetrahydroisoquinolines.

Fifteen cis-dichloroplatinum complexes (5a-5o) were synthesized by treatment of 1-(2-aminophenyl)-1,2,3,4-THIQs (4a-4o) with K(2)PtCl(4). The antitumor activity of these compounds was examined against four different human tumor cell lines. Their structure-activity relationships for antitumor activity are reported. All of these compounds exhibited activity against MCF-7 cell line and showed good activity against WiDr cell line except 5c and 5f. On the other hand, compounds 5j and 5o are more active than the other compounds against Hepa59T/VGH cell line. The electron-donating group at the 6-position of isoquinoline ring seems to decrease the antitumor activity and the chloro substituent at the C-4 position of the aniline ring shown the highest potency. The "trans influence" dominates the control of the stability of [1-(2-aminophenyl)-1,2,3,4-THIQ]dichloroplatinums(II).

Synthesis and antitumor activity of cis-dichloridoplatinum(II) complexes of 1,1'-biisoquinolines.

A simple approach to 1,1',2,2',3,3',4,4'-octahydro-1,1'-biisoquinolines is described. Reaction of phenethylamines with oxalyl chloride led to N,N'-bis(phenethyl) oxamides (1). Cyclization of oxamides by using Bischler-Napieralski conditions gave 3,3',4,4'-tetrahydro-1,1'-biisoquinoline (3) and unusual products 2, 4, 5. Reduction of 3,3',4,4'-tetrahydro-1,1'-biisoquinolines with sodium boron hydride gave both rac-1,1',2,2',3,3',4,4'-octahydro-1,1'-biisoquinolines (6) and meso-1,1',2,2',3,3',4,4' -octahydro-1,1'-biisoquinolines (7). Compound 6 was resolved to (1S, 1S') (8) and (1R, 1R') (9) furtherly. By treating all the biisoquinolines with K2PtCl4 afforded their cis-dichloridoplatinum (II) complexes (12-18). The antitumor activity of these complexes was evaluated.

The synthesis of N-phenoxyethyl-1-substituted-1,2,3,4-tetrahydroisoquinolines and their alpha1-adrenoceptor blocking activity.

A series of phenoxyisoquinolines, N-phenoxyethyl-1-(2-nitrophenyl)-1,2,3,4-THIQs 3a-3d, N-phenoxyethyl-1-benzyl-1,2,3,4-THIQ 3e, N-phenoxyethyl-1-(2-aminophenyl)-1,2,3,4-THIQs 5f-5i, N-phenoxyethyl-1-(2-phenoxyethylaminophenyl)-1,2,3,4-THIQs 5f'-5i', have been synthesized and tested in isolated rat vas deferens alpha-adrenoreceptors. Comparison of pA2 values for these compounds in the presence of phenylephrine confirms that alpha(1)-adrenoceptor blocking activity of 3a-3d (-NO(2) series) is more active than 6a-6c (-NH(2) series) in the aortic rings isolated from SD rats. On the other hand, the electron-donating group at the 6-position of isoquinoline ring either increases or decreases the alpha(1)-adrenoceptor blocking activity.