Search

OBJECTIVES : The authors explored the development and validation of machine-learning models for augmenting the echocardiographic grading of aortic stenosis (AS) severity. BACKGROUND : In AS, symptoms and adverse events develop secondarily to valvular obstruction and left ventricular decompensation. The current echocardiographic grading of AS severity focuses on the valve and is limited by diagnostic uncertainty. METHODS : Using echocardiography (ECHO) measurements (ECHO cohort, n ¼ 1,052), we performed patient similarity analysis to derive high-severity and low-severity phenogroups of AS. We subsequently developed a supervised machine learning classifier and validated its performance with independent markers of disease severity obtained using computed tomography (CT) (CT cohort, n ¼ 752) and cardiovascular magnetic resonance (CMR) imaging (CMR cohort, n ¼ 160). The classifier’s prognostic value was further validated using clinical outcomes (aortic valve replacement [AVR] and death) observed in the ECHO and CMR cohorts. RESULTS : In 1,964 patients from the 3 multi-institutional cohorts, 1,346 (68%) subjects had either nonsevere or discordant AS severity. Machine learning identified 1,117 (57%) patients as having high-severity and 847 (43%) as having low-severity AS. High-severity patients in CT and CMR cohorts had higher valve calcium scores and left ventricular mass and fibrosis, respectively than the low-severity group. In the ECHO cohort, progression to AVR and progression to death in patients who did not receive AVR was faster in the high-severity group. Compared with the conventional classification of disease severity, machine-learning–based severity classification improved discrimination (integrated discrimination improvement: 0.07; 95% confidence interval: 0.02 to 0.12) and reclassification (net reclassification improvement: 0.17; 95% confidence interval: 0.11 to 0.23) for the outcome of AVR at 5 years. For both ECHO and CMR cohorts, we observed prognostic val

OBJECTIVES : The authors explored the development and validation of machine-learning models for augmenting the echocardiographic grading of aortic stenosis (AS) severity. BACKGROUND : In AS, symptoms and adverse events develop secondarily to valvular obstruction and left ventricular decompensation. The current echocardiographic grading of AS severity focuses on the valve and is limited by diagnostic uncertainty. METHODS : Using echocardiography (ECHO) measurements (ECHO cohort, n ¼ 1,052), we performed patient similarity analysis to derive high-severity and low-severity phenogroups of AS. We subsequently developed a supervised machine learning classifier and validated its performance with independent markers of disease severity obtained using computed tomography (CT) (CT cohort, n ¼ 752) and cardiovascular magnetic resonance (CMR) imaging (CMR cohort, n ¼ 160). The classifier’s prognostic value was further validated using clinical outcomes (aortic valve replacement [AVR] and death) observed in the ECHO and CMR cohorts. RESULTS : In 1,964 patients from the 3 multi-institutional cohorts, 1,346 (68%) subjects had either nonsevere or discordant AS severity. Machine learning identified 1,117 (57%) patients as having high-severity and 847 (43%) as having low-severity AS. High-severity patients in CT and CMR cohorts had higher valve calcium scores and left ventricular mass and fibrosis, respectively than the low-severity group. In the ECHO cohort, progression to AVR and progression to death in patients who did not receive AVR was faster in the high-severity group. Compared with the conventional classification of disease severity, machine-learning–based severity classification improved discrimination (integrated discrimination improvement: 0.07; 95% confidence interval: 0.02 to 0.12) and reclassification (net reclassification improvement: 0.17; 95% confidence interval: 0.11 to 0.23) for the outcome of AVR at 5 years. For both ECHO and CMR cohorts, we observed prognostic val

Background Imatinib mesylate is currently the drug of choice to treat chronic myeloid leukemia. However, patient resistance and cytotoxicity make secondary lines of treatment, such as omacetaxine mepesuccinate, a necessity. Given that drug cytotoxicity represents a major problem during treatment, it is essential to understand the biological pathways affected to better predict poor drug response and prioritize a treatment regime. Methods We conducted cell viability and gene expression assays to determine heritability and gene expression changes associated with imatinib and omacetaxine treatment of 55 non-cancerous lymphoblastoid cell lines, derived from 17 pedigrees. In total, 48,803 transcripts derived from Illumina Human WG-6 BeadChips were analyzed for each sample using SOLAR, whilst correcting for kinship structure. Results Cytotoxicity within cell lines was highly heritable following imatinib treatment (h2 = 0.60-0.73), but not omacetaxine treatment. Cell lines treated with an IC20 dose of imatinib or omacetaxine showed differential gene expression for 956 (1.96%) and 3,892 transcripts (7.97%), respectively; 395 of these (0.8%) were significantly influenced by both imatinib and omacetaxine treatment. k-means clustering and DAVID functional annotation showed expression changes in genes related to kinase binding and vacuole-related functions following imatinib treatment, whilst expression changes in genes related to cell division and apoptosis were evident following treatment with omacetaxine. The enrichment scores for these ontologies were very high (mostly >10). Conclusions Induction of gene expression changes related to different pathways following imatinib and omacetaxine treatment suggests that the cytotoxicity of such drugs may be differentially tolerated by individuals based on their genetic background.

Background: The impact of viral load (VL) decay and cumulative VL on CD4 recovery and AIDS after highly-active antiretroviral therapy (HAART) is unknown. Methods and Findings: Three virologic kinetic parameters (first year and overall exponential VL decay constants, and first year VL slope) and cumulative VL during HAART were estimated for 2,278 patients who initiated HAART in the U.S. Military HIV Natural History Study. CD4 and VL trajectories were computed using linear and nonlinear Generalized Estimating Equations models. Multivariate Poisson and linear regression models were used to determine associations of VL parameters with CD4 recovery, adjusted for factors known to correlate with immune recovery. Cumulative VL higher than the sample median was independently associated with an increased risk of AIDS (relative risk 2.38, 95% confidence interval 1.56-3.62 p<0.001). Among patients with VL suppression, first year VL decay and slope were independent predictors of early CD4 recovery (p = 0.001) and overall gain (p<0.05). Despite VL suppression, those with slow decay during the first year of HAART as well as during the entire therapy period (overall), in general, gained less CD4 cells compared to the other subjects (133 vs. 195.4 cells/microL; p = 0.001) even after adjusting for potential confounders. Conclusions: In a cohort with free access to healthcare, independent of established predictors of AIDS and CD4 recovery during HAART, cumulative VL and virologic decay patterns were associated with AIDS and distinct aspects of CD4 reconstitution., Published in PLoS ONE, v6 n5 article e17956, May 2011.

Polymorphisms in CCR5, the major coreceptor for HIV, and CCL3L1, a potent CCR5 ligand and HIV-suppressive chemokine, are determinants of HIV-AIDS susceptibility. Here, we mathematically modeled the potential impact of these genetic factors on the epidemic spread of HIV, as well as on its prevention. Methods and Results: Ro, the basic reproductive number, is a fundamental concept in explaining the emergence and persistence of epidemics. By modeling sexual transmission among HIV+/HIV2 partner pairs, we find that Ro estimates, and concordantly, the temporal and spatial patterns of HIV outgrowth are highly dependent on the infecting partners? CCL3L1- CCR5 genotype. Ro was least and highest when the infected partner possessed protective and detrimental CCL3L1-CCR5 genotypes, respectively. The modeling data indicate that in populations such as Pygmies with a high CCL3L1 gene dose and protective CCR5 genotypes, the spread of HIV might be minimal. Additionally, Pc, the critical vaccination proportion, an estimate of the fraction of the population that must be vaccinated successfully to eradicate an epidemic was ,1 only when the infected partner had a protective CCL3L1-CCR5 genotype. Since in practice Pc cannot be .1, to prevent epidemic spread, population groups defined by specific CCL3L1-CCR5 genotypes might require repeated vaccination, or as our models suggest, a vaccine with an efficacy of .70%. Further, failure to account for CCL3L1-CCR5-based genetic risk might confound estimates of vaccine efficacy. For example, in a modeled trial of 500 subjects, misallocation of CCL3L1-CCR5 genotype of only 25 (5%) subjects between placebo and vaccine arms results in a relative error of ,12% from the true vaccine efficacy. Conclusions: CCL3L1-CCR5 genotypes may impact on the dynamics of the HIV epidemic and, consequently, the observed heterogeneous global distrib, Pub. in PLoS one, v3 n11, p1-9, Nov 2008.

Background: Whether vexing clinical decision-making dilemmas can be partly addressed by recent advances in genomics is unclear. For example, when to initiate highly active antiretroviral therapy (HAART) during HIV-1 infection remains a clinical dilemma. This decision relies heavily on assessing AIDS risk based on the CD4+ T cell count and plasma viral load. However, the trajectories of these two laboratory markers are influenced, in part, by polymorphisms in CCR5, the major HIV coreceptor, and the gene copy number of CCL3L1, a potent CCR5 ligand and HIV-suppressive chemokine. Therefore, we determined whether accounting for both genetic and laboratory markers provided an improved means of assessing AIDS risk. Methods and Findings: In a prospective, single-site, ethnically-mixed cohort of 1,132 HIV-positive subjects, we determined the AIDS risk conveyed by the laboratory and genetic markers separately and in combination. Subjects were assigned to a low, moderate or high genetic risk group (GRG) based on variations in CCL3L1 and CCR5. The predictive value of the CCL3L1- CCR5 GRGs, as estimated by likelihood ratios, was equivalent to that of the laboratory markers. GRG status also predicted AIDS development when the laboratory markers conveyed a contrary risk. Additionally, in two separate and large groups of HIV+ subjects from a natural history cohort, the results from additive risk-scoring systems and classification and regression tree (CART) analysis revealed that the laboratory and CCL3L1-CCR5 genetic markers together provided more prognostic information than either marker alone. Furthermore, GRGs independently predicted the time interval from seroconversion to CD4+ cell count thresholds used to guide HAART initiation. Conclusions: The combination of the laboratory and genetic markers captures a broader spectrum of AIDS risk than either marker alone. By track, Pub. in PloS one, v3 n9, p1-14, Sep 2008.