- Browse by Subject
Browsing by Subject "Regression"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Association of Histologic Disease Activity With Progression of Nonalcoholic Fatty Liver Disease(JAMA Network, 2019-10-02) Kleiner, David E.; Brunt, Elizabeth M.; Wilson, Laura A.; Behling, Cynthia; Guy, Cynthia; Contos, Melissa; Cummings, Oscar; Yeh, Matthew; Gill, Ryan; Chalasani, Naga; Neuschwander-Tetri, Brent A.; Diehl, Anna Mae; Dasarathy, Srinivasan; Terrault, Norah; Kowdley, Kris; Loomba, Rohit; Belt, Patricia; Tonascia, James; Lavine, Joel E.; Sanyal, Arun J.; Nonalcoholic Steatohepatitis Clinical Research Network; Medicine, School of MedicineImportance: The histologic evolution of the full spectrum of nonalcoholic fatty liver disease (NAFLD) and factors associated with progression or regression remain to be definitively established. Objective: To evaluate the histologic evolution of NAFLD and the factors associated with changes in disease severity over time. Design, Setting, and Participants: A prospective cohort substudy from the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) NAFLD Database study, a noninterventional registry, was performed at 8 university medical research centers. Masked assessment of liver histologic specimens was performed, using a prespecified protocol to score individual biopsies. Participants included 446 adults with NAFLD enrolled in the NASH CRN Database studies between October 27, 2004, and September 13, 2013, who underwent 2 liver biopsies 1 or more year apart. Data analysis was performed from October 2016 to October 2018. Main Outcomes and Measures: Progression and regression of fibrosis stage, using clinical, laboratory, and histologic findings, including the NAFLD activity score (NAS) (sum of scores for steatosis, lobular inflammation, and ballooning; range, 0-8, with 8 indicating more severe disease). Results: A total of 446 adults (mean [SD] age, 47 [11] years; 294 [65.9%] women) with NAFLD (NAFL, 86 [19.3%]), borderline NASH (84 [18.8%]), and definite NASH (276 [61.9%]) were studied. Over a mean (SD) interval of 4.9 (2.8) years between biopsies, NAFL resolved in 11 patients (12.8%) and progressed to steatohepatitis in 36 patients (41.9%). Steatohepatitis resolved in 24 (28.6%) of the patients with borderline NASH and 61 (22.1%) of those with definite NASH. Fibrosis progression or regression by at least 1 stage occurred in 132 (30%) and 151 [34%] participants, respectively. Metabolic syndrome (20 [95%] vs 108 [72%]; P = .03), baseline NAS (mean [SD], 5.0 [1.4] vs 4.3 [1.6]; P = .005), and smaller reduction in NAS (-0.2 [2] vs -0.9 [2]; P < .001) were associated with progression to advanced (stage 3-4) fibrosis vs those without progression to stage 3 to 4 fibrosis. Fibrosis regression was associated with lower baseline insulin level (20 vs 33 μU/mL; P = .02) and decrease in all NAS components (steatosis grade -0.8 [0.1] vs -0.3 [0.9]; P < .001; lobular inflammation -0.5 [0.8] vs -0.2 [0.9]; P < .001; ballooning -0.7 [1.1] vs -0.1 [0.9]; P < .001). Only baseline aspartate aminotransferase (AST) levels were associated with fibrosis regression vs no change and progression vs no change on multivariable regression: baseline AST (regression: conditional odds ratio [cOR], 0.6 per 10 U/L AST; 95% CI, 0.4-0.7; P < .001; progression: cOR, 1.3; 95% CI, 1.1-1.5; P = .002). Changes in the AST level, alanine aminotransferase (ALT) level, and NAS were also associated with fibrosis regression and progression (ΔAST level: regression, cOR, 0.9; 95% CI, 0.6-1.2; P = .47; progression, cOR, 1.3; 95% CI, 1.0-1.6; P = .02; ΔALT level: regression, cOR, 0.7 per 10 U/L AST; 95% CI, 0.5-0.9; P = .002; progression, cOR, 1.0 per 10 U/L AST; 95% CI, 0.9-1.2; P = .93; ΔNAS: regression, cOR, 0.7; 95% CI, 0.6-0.9; P = .001; progression, cOR, 1.3; 95% CI, 1.1-1.5; P = .01). Conclusions and Relevance: Improvement or worsening of disease activity may be associated with fibrosis regression or progression, respectively, in NAFLD.Item Harmonization of Newborn Screening Results for Pompe Disease and Mucopolysaccharidosis Type I(MDPI, 2023-02-27) Dorley, M. Christine; Dizikes, George J.; Pickens, Charles Austin; Cuthbert, Carla; Basheeruddin, Khaja; Gulamali-Majid, Fizza; Hetterich, Paul; Hietala, Amy; Kelsey, Ashley; Klug, Tracy; Lesko, Barbara; Mills, Michelle; Moloney, Shawn; Neogi, Partha; Orsini, Joseph; Singer, Douglas; Petritis, Konstantinos; Pathology and Laboratory Medicine, School of MedicineIn newborn screening, false-negative results can be disastrous, leading to disability and death, while false-positive results contribute to parental anxiety and unnecessary follow-ups. Cutoffs are set conservatively to prevent missed cases for Pompe and MPS I, resulting in increased falsepositive results and lower positive predictive values. Harmonization has been proposed as a way to minimize false-negative and false-positive results and correct for method differences, so we harmonized enzyme activities for Pompe and MPS I across laboratories and testing methods (Tandem Mass Spectrometry (MS/MS) or Digital Microfluidics (DMF)). Participating states analyzed proofof- concept calibrators, blanks, and contrived specimens and reported enzyme activities, cutoffs, and other testing parameters to Tennessee. Regression and multiples of the median were used to harmonize the data. We observed varied cutoffs and results. Six of seven MS/MS labs reported enzyme activities for one specimen for MPS I marginally above their respective cutoffs with results classified as negative, whereas all DMF labs reported this specimen’s enzyme activity below their respective cutoffs with results classified as positive. Reasonable agreement in enzyme activities and cutoffs was achieved with harmonization; however, harmonization does not change how a value would be reported as this is dependent on the placement of cutoffs.Item Regression analysis of big count data via a-optimal subsampling(2018-07-19) Zhao, Xiaofeng; Tan, Fei; Peng, HanxiangThere are two computational bottlenecks for Big Data analysis: (1) the data is too large for a desktop to store, and (2) the computing task takes too long waiting time to finish. While the Divide-and-Conquer approach easily breaks the first bottleneck, the Subsampling approach simultaneously beat both of them. The uniform sampling and the nonuniform sampling--the Leverage Scores sampling-- are frequently used in the recent development of fast randomized algorithms. However, both approaches, as Peng and Tan (2018) have demonstrated, are not effective in extracting important information from data. In this thesis, we conduct regression analysis for big count data via A-optimal subsampling. We derive A-optimal sampling distributions by minimizing the trace of certain dispersion matrices in general estimating equations (GEE). We point out that the A-optimal distributions have the same running times as the full data M-estimator. To fast compute the distributions, we propose the A-optimal Scoring Algorithm, which is implementable by parallel computing and sequentially updatable for stream data, and has faster running time than that of the full data M-estimator. We present asymptotic normality for the estimates in GEE's and in generalized count regression. A data truncation method is introduced. We conduct extensive simulations to evaluate the numerical performance of the proposed sampling distributions. We apply the proposed A-optimal subsampling method to analyze two real count data sets, the Bike Sharing data and the Blog Feedback data. Our results in both simulations and real data sets indicated that the A-optimal distributions substantially outperformed the uniform distribution, and have faster running times than the full data M-estimators.Item Sufficient principal component regression for pattern discovery in transcriptomic data(Oxford University Press, 2022-05-14) Ding, Lei; Zentner, Gabriel E.; McDonald, Daniel J.; Biology, School of ScienceMotivation: Methods for the global measurement of transcript abundance such as microarrays and RNA-Seq generate datasets in which the number of measured features far exceeds the number of observations. Extracting biologically meaningful and experimentally tractable insights from such data therefore requires high-dimensional prediction. Existing sparse linear approaches to this challenge have been stunningly successful, but some important issues remain. These methods can fail to select the correct features, predict poorly relative to non-sparse alternatives or ignore any unknown grouping structures for the features. Results: We propose a method called SuffPCR that yields improved predictions in high-dimensional tasks including regression and classification, especially in the typical context of omics with correlated features. SuffPCR first estimates sparse principal components and then estimates a linear model on the recovered subspace. Because the estimated subspace is sparse in the features, the resulting predictions will depend on only a small subset of genes. SuffPCR works well on a variety of simulated and experimental transcriptomic data, performing nearly optimally when the model assumptions are satisfied. We also demonstrate near-optimal theoretical guarantees. Availability and implementation: Code and raw data are freely available at https://github.com/dajmcdon/suffpcr. Package documentation may be viewed at https://dajmcdon.github.io/suffpcr.