Brain structure ages-A new biomarker for multi-disease classification.

Age is an important variable to describe the expected brain's anatomy status across the normal aging trajectory. The deviation from that normative aging trajectory may provide some insights into neurological diseases. In neuroimaging, predicted brain age is widely used to analyze different diseases. However, using only the brain age gap information (i.e., the difference between the chronological age and the estimated age) can be not enough informative for disease classification problems. In this paper, we propose to extend the notion of global brain age by estimating brain structure ages using structural magnetic resonance imaging. To this end, an ensemble of deep learning models is first used to estimate a 3D aging map (i.e., voxel-wise age estimation). Then, a 3D segmentation mask is used to obtain the final brain structure ages. This biomarker can be used in several situations. First, it enables to accurately estimate the brain age for the purpose of anomaly detection at the population level. In this situation, our approach outperforms several state-of-the-art methods. Second, brain structure ages can be used to compute the deviation from the normal aging process of each brain structure. This feature can be used in a multi-disease classification task for an accurate differential diagnosis at the subject level. Finally, the brain structure age deviations of individuals can be visualized, providing some insights about brain abnormality and helping clinicians in real medical contexts.

Deep grading for MRI-based differential diagnosis of Alzheimer's disease and Frontotemporal dementia.

Alzheimer's disease and Frontotemporal dementia are common forms of neurodegenerative dementia. Behavioral alterations and cognitive impairments are found in the clinical courses of both diseases, and their differential diagnosis can sometimes pose challenges for physicians. Therefore, an accurate tool dedicated to this diagnostic challenge can be valuable in clinical practice. However, current structural imaging methods mainly focus on the detection of each disease but rarely on their differential diagnosis. In this paper, we propose a deep learning-based approach for both disease detection and differential diagnosis. We suggest utilizing two types of biomarkers for this application: structure grading and structure atrophy. First, we propose to train a large ensemble of 3D U-Nets to locally determine the anatomical patterns of healthy people, patients with Alzheimer's disease and patients with Frontotemporal dementia using structural MRI as input. The output of the ensemble is a 2-channel disease's coordinate map, which can be transformed into a 3D grading map that is easily interpretable for clinicians. This 2-channel disease's coordinate map is coupled with a multi-layer perceptron classifier for different classification tasks. Second, we propose to combine our deep learning framework with a traditional machine learning strategy based on volume to improve the model discriminative capacity and robustness. After both cross-validation and external validation, our experiments, based on 3319 MRIs, demonstrated that our method produces competitive results compared to state-of-the-art methods for both disease detection and differential diagnosis.

Towards better interpretable and generalizable AD detection using collective artificial intelligence.

Alzheimer's Disease is the most common cause of dementia. Accurate diagnosis and prognosis of this disease are essential to design an appropriate treatment plan, increasing the life expectancy of the patient. Intense research has been conducted on the use of machine learning to identify Alzheimer's Disease from neuroimaging data, such as structural magnetic resonance imaging. In recent years, advances of deep learning in computer vision suggest a new research direction for this problem. Current deep learning-based approaches in this field, however, have a number of drawbacks, including the interpretability of model decisions, a lack of generalizability information and a lower performance compared to traditional machine learning techniques. In this paper, we design a two-stage framework to overcome these limitations. In the first stage, an ensemble of 125 U-Nets is used to grade the input image, producing a 3D map that reflects the disease severity at voxel-level. This map can help to localize abnormal brain areas caused by the disease. In the second stage, we model a graph per individual using the generated grading map and other information about the subject. We propose to use a graph convolutional neural network classifier for the final classification. As a result, our framework demonstrates comparative performance to the state-of-the-art methods in different datasets for both diagnosis and prognosis. We also demonstrate that the use of a large ensemble of U-Nets offers a better generalization capacity for our framework.

Recurrence of tuberculosis among patients following treatment completion in eight provinces of Vietnam: A nested case-control study.

BACKGROUND: Patients completing treatment for tuberculosis (TB) in high-prevalence settings face a risk of developing recurrent disease. This has important consequences for public health, given its association with drug resistance and a poor prognosis. Previous research has implicated individual factors such as smoking, alcohol use, HIV, poor treatment adherence, and drug resistant disease as risk factors for recurrence. However, little is known about how these factors co-act to produce recurrent disease. Furthermore, perhaps factors related to the index disease means higher burden/low resource settings may be more prone to recurrent disease that could be preventable. METHODS: We conducted a case-control study nested within a cohort of consecutively enrolled adults who were being treated for smear positive pulmonary TB in 70 randomly selected district clinics in Vietnam. Cases were patients with recurrent TB, identified by follow-up from the parent cohort study. Controls were selected from the cohort by random sampling. Information on demographic, clinical and disease-related characteristics was obtained by interview. Treatment information was extracted from clinic registries. Logistic regression, with stepwise selection, was used to develop a fully adjusted model for the odds of recurrence of TB. RESULTS: We recruited 10,964 patients between October 2010 and July 2013. Median follow-up was 988 days. At the end of follow-up, 505 patients (4.7%) with recurrence were identified as cases and 630 other patients were randomly selected as controls. Predictors of recurrence included multidrug-resistant (MDR)-TB (adjusted odds ratio 79.6; 95% CI: 25.1-252.0), self-reported prior TB therapy (aOR=2.5; 95% CI: 1.7-3.5), and incomplete adherence (aOR=1.9; 95% CI 1.1-3.1). CONCLUSIONS: Index disease treatment history is a leading determinant of relapse among patients with TB in Vietnam. Further research is required to identify interventions that will reduce the risk of recurrent disease and enhance its early detection within high-risk populations.

Evaluation of microscopic observation drug susceptibility assay for diagnosis of multidrug-resistant tuberculosis in Viet Nam.

BACKGROUND: Early diagnosis of tuberculosis (TB) and multidrug resistant tuberculosis (MDR TB) is important for the elimination of TB. We evaluated the microscopic observation drug susceptibility (MODS) assay as a direct rapid drug susceptibility testing (DST) method for MDR-TB screening in sputum samples METHODS: All adult TB suspects, who were newly presenting to Pham Ngoc Thach Hospital from August to November 2008 were enrolled into the study. Processed sputum samples were used for DST by MODS (DST-MODS) (Rifampicin (RIF) 1 µg/ml and Isoniazid (INH) 0.4 µg/ml), MGIT culture (Mycobacterial Growth Indicator Tube) and Lowenstein Jensen (LJ) culture. Cultures positive by either MGIT or LJ were used for proportional DST (DST-LJ) (RIF 40 µg/ml and INH 0.2 µg/ml). DST profiles on MODS and LJ were compared. Discrepant results were resolved by multiplex allele specific PCR (MAS-PCR). RESULTS: Seven hundred and nine TB suspects/samples were enrolled into the study, of which 300 samples with DST profiles available from both MODS and DST-LJ were analyzed. Cording in MODS was unable to correctly identify 3 Mycobacteria Other Than Tuberculosis (MOTT) isolates, resulting in 3 false positive TB diagnoses. None of these isolates were identified as MDR-TB by MODS. The sensitivity and specificity of MODS were 72.6% (95%CI: 59.8, 83.1) and 97.9% (95%CI: 95.2, 99.3), respectively for detection of INH resistant isolates, 72.7% (95%CI: 30.9, 93.7) and 99.7% (95%CI: 98.1, 99.9), respectively for detecting RIF resistant isolates and 77.8% (95%CI: 39.9, 97.1) and 99.7% (95%CI: 98.1, 99.9), respectively for detecting MDR isolates. The positive and negative predictive values (PPV and NPV) of DST-MODS were 87.5% (95%CI: 47.3, 99.6) and 99.3% (95%CI: 97.5, 99.9) for detection of MDR isolates; and the agreement between MODS and DST-LJ was 99.0% (kappa: 0.8, P < 0.001) for MDR diagnosis. The low sensitivity of MODS for drug resistance detection was probably due to low bacterial load samples and the high INH concentration (0.4 µg/ml). The low PPV of DST-MODS may be due to the low MDR-TB rate in the study population (3.8%). The turnaround time of DST-MODS was 9 days and 53 days for DST-LJ. CONCLUSION: The DST-MODS technique is rapid with low contamination rates. However, the sensitivity of DST-MODS for detection of INH and RIF resistance in this study was lower than reported from other settings.

Dexamethasone and long-term outcome of tuberculous meningitis in Vietnamese adults and adolescents.

BACKGROUND: Dexamethasone has been shown to reduce mortality in patients with tuberculous meningitis but the long-term outcome of the disease is unknown. METHODS: Vietnamese adults and adolescents with tuberculous meningitis recruited to a randomised, double-blind, placebo-controlled trial of adjunctive dexamethasone were followed-up at five years, to determine the effect of dexamethasone on long-term survival and neurological disability. RESULTS: 545 patients were randomised to receive either dexamethasone (274 patients) or placebo (271 patients). 50 patients (9.2%) were lost to follow-up at five years. In all patients two-year survival, probabilities tended to be higher in the dexamethasone arm (0.63 versus 0.55; p = 0.07) but five-year survival rates were similar (0.54 versus 0.51, p = 0.51) in both groups. In patients with grade 1 TBM, but not with grade 2 or grade 3 TBM, the benefit of dexamethasone treatment tended to persist over time (five-year survival probabilities 0.69 versus 0.55, p = 0.07) but there was no conclusive evidence of treatment effect heterogeneity by TBM grade (p = 0.36). The dexamethasone group had a similar proportion of severely disabled patients among survivors at five years as the placebo group (17/128, 13.2% vs. 17/116, 14.7%) and there was no significant association between dexamethasone treatment and disability status at five years (p = 0.32). CONCLUSIONS: Adjunctive dexamethasone appears to improve the probability of survival in patients with TBM, until at least two years of follow-up. We could not demonstrate a five-year survival benefit of dexamethasone treatment which may be confined to patients with grade 1 TBM. TRIAL REGISTRATION: ClinicalTrials.gov NCT01317654.

An algorithm for tuberculosis screening and diagnosis in people with HIV.

BACKGROUND: Tuberculosis screening is recommended for people with human immunodeficiency virus (HIV) infection to facilitate early diagnosis and safe initiation of antiretroviral therapy and isoniazid preventive therapy. No internationally accepted, evidence-based guideline addresses the optimal means of conducting such screening, although screening for chronic cough is common. METHODS: We consecutively enrolled people with HIV infection from eight outpatient clinics in Cambodia, Thailand, and Vietnam. For each patient, three samples of sputum and one each of urine, stool, blood, and lymph-node aspirate (for patients with lymphadenopathy) were obtained for mycobacterial culture. We compared the characteristics of patients who received a diagnosis of tuberculosis (on the basis of having one or more specimens that were culture-positive) with those of patients who did not have tuberculosis to derive an algorithm for screening and diagnosis. RESULTS: Tuberculosis was diagnosed in 267 (15%) of 1748 patients (median CD4+ T-lymphocyte count, 242 per cubic millimeter; interquartile range, 82 to 396). The presence of a cough for 2 or 3 weeks or more during the preceding 4 weeks had a sensitivity of 22 to 33% for detecting tuberculosis. The presence of cough of any duration, fever of any duration, or night sweats lasting 3 or more weeks in the preceding 4 weeks was 93% sensitive and 36% specific for tuberculosis. In the 1199 patients with any of these symptoms, a combination of two negative sputum smears, a normal chest radiograph, and a CD4+ cell count of 350 or more per cubic millimeter helped to rule out a diagnosis of tuberculosis, whereas a positive diagnosis could be made only for the 113 patients (9%) with one or more positive sputum smears; mycobacterial culture was required for most other patients. CONCLUSIONS: In persons with HIV infection, screening for tuberculosis should include asking questions about a combination of symptoms rather than only about chronic cough. It is likely that antiretroviral therapy and isoniazid preventive therapy can be started safely in people whose screening for all three symptoms is negative, whereas diagnosis in most others will require mycobacterial culture.

Beijing genotype of Mycobacterium tuberculosis is significantly associated with high-level fluoroquinolone resistance in Vietnam.

Consecutive fluoroquinolone (FQ)-resistant isolates (n = 109) identified at the Pham Ngoc Thach Hospital for Tuberculosis, Ho Chi Minh City, Vietnam, were sequenced in the quinolone resistance-determining regions of the gyrA and gyrB genes and typed by large sequence polymorphism typing and spoligotyping to identify the Beijing genotype of Mycobacterium tuberculosis. Beijing genotype prevalence was compared with 109 consecutive isolates from newly presenting patients with pulmonary tuberculosis from the hospital outpatient department. Overall, 82.6% (n = 90/109) of isolates had mutations in gyrAB. Nine novel mutations were identified in gyrB (S486F, N538T, T539P, D500A, D500H, D500N, G509A, E540V, and E540D). The influence of these novel gyrB mutations on FQ resistance is not proven. The Beijing genotype was significantly associated with FQ resistance (odds ratio [OR], 2.39 [95% confidence interval {CI}, 1.34 to 4.25]; P = 0.003). Furthermore, Beijing genotype FQ-resistant isolates were significantly more likely than FQ-resistant isolates of other genotypes to have gyrA mutations (OR, 7.75 [95% CI, 2.84 to 21.15]; P = 0.0001) and high-level (>8 microg/ml) FQ resistance (OR, 11.0 [95% CI, 2.6 to 47.0]; P = 0.001). The underlying mechanism of the association of the Beijing genotype with high-level FQ resistance in this setting remains to be determined. The association of the Beijing genotype with relatively high-level FQ resistance conferred by specific gyrA mutations reported here is of grave concern given the epidemic spread of the Beijing genotype and the current hopes for shorter first-line treatment regimens based on FQs.

Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults.

BACKGROUND: Tuberculous meningitis kills or disables more than half of those affected with the disease. Previous studies have been too small to determine whether adjunctive treatment with corticosteroids can reduce the risk of disability or death among adults with tuberculous meningitis, and the effect of coinfection with the human immunodeficiency virus (HIV) is unclear. METHODS: We performed a randomized, double-blind, placebo-controlled trial in Vietnam in patients over 14 years of age who had tuberculous meningitis, with or without HIV infection, to determine whether adjunctive treatment with dexamethasone reduced the risk of death or severe disability after nine months of follow-up. We conducted prespecified subgroup analyses and intention-to-treat analyses. RESULTS: A total of 545 patients were randomly assigned to groups that received either dexamethasone (274 patients) or placebo (271 patients). Only 10 patients (1.8 percent) had been lost to follow-up at nine months of treatment. Treatment with dexamethasone was associated with a reduced risk of death (relative risk, 0.69; 95 percent confidence interval, 0.52 to 0.92; P=0.01). It was not associated with a significant reduction in the proportion of severely disabled patients (34 of 187 patients [18.2 percent] among survivors in the dexamethasone group vs. 22 of 159 patients [13.8 percent] in the placebo group, P=0.27) or in the proportion of patients who had either died or were severely disabled after nine months (odds ratio, 0.81; 95 percent confidence interval, 0.58 to 1.13; P=0.22). The treatment effect was consistent across subgroups that were defined by disease-severity grade (stratified relative risk of death, 0.68; 95 percent confidence interval, 0.52 to 0.91; P=0.007) and by HIV status (stratified relative risk of death, 0.78; 95 percent confidence interval, 0.59 to 1.04; P=0.08). Significantly fewer serious adverse events occurred in the dexamethasone group than in the placebo group (26 of 274 patients vs. 45 of 271 patients, P=0.02). CONCLUSIONS: Adjunctive treatment with dexamethasone improves survival in patients over 14 years of age with tuberculous meningitis but probably does not prevent severe disability.

Synthesis, SAR, crystal structure, and biological evaluation of benzoquinoliziniums as activators of wild-type and mutant cystic fibrosis transmembrane conductance regulator channels.

Chloride channels play important roles in homeostasis and regulate cell volume, transepithelial transport, and electrical excitability. Despite recent progress made in the genetic and molecular aspect of chloride channels, their pharmacology is still poorly understood. The cystic fibrosis transmembrane conductance regulator (CFTR) is a cAMP-regulated epithelial chloride channel for which mutations cause cystic fibrosis. Here we have synthesized benzo[c]quinolizinium and benzo[f]indolo[2,3-a]quinolizinium salts (MPB) and performed a SAR to identify the structural basis for activation of the CFTR chloride channel. Synthesized compounds were evaluated on wild-type CFTR and on CFTR having the glycine-to-aspartic acid missense mutation at codon 551 (G551D-CFTR), using a robot and cell-based assay. The presence of an hydroxyl group at position 6 of the benzo[c]quinolizinium skeleton associated with a chlorine atom at position 10 or 7 and an alkyl chain at position 5 determined the highest activity. The most potent product is 5-butyl-7-chloro-6-hydroxybenzo[c]quinolizinium chloride (8u, MPB-104). 8u is 100 times more potent than the parent compound 8a (MPB-07).