A Comparative Analysis of Data Synthesis Techniques to Improve Classification Accuracy of Raman Spectroscopy Data.

Raman spectra are examples of high dimensional data that can often be limited in the number of samples. This is a primary concern when Deep Learning frameworks are developed for tasks such as chemical species identification, quantification, and diagnostics. Open-source data are difficult to obtain and often sparse; furthermore, the collecting and curating of new spectra require expertise and resources. Deep generative modeling utilizes Deep Learning architectures to approximate high dimensional distributions and aims to generate realistic synthetic data. The evaluation of the data and the performance of the deep models is usually conducted on a per-task basis and provides no indication of an increase to robustness, or generalization, on a wider scale. In this study, we compare the benefits and limitations of a standard statistical approach to data synthesis (weighted blending) with a popular deep generative model, the Variational Autoencoder. Two binary data sets are divided into 3-fold to simulate small, limited samples. Synthetic data distributions are created per fold using the two methods and then augmented into the training of two Deep Learning algorithms, a Convolutional Neural Network and a Fully-Connected Neural Network. The goal of this study is to observe the trends in learning as synthetic data are continually augmented to the training data in increasing batches. To determine the impact of each synthetic method, Principal Component Analysis and the discrete Fréchet distance are implemented to visualize and measure the distance between the source and synthetic distributions along with the Machine Learning metric balanced accuracy for evaluating performance on imbalanced data.

Interplay of Enzyme Therapy and Dietary Management of Murine Homocystinuria.

Albeit effective, methionine/protein restriction in the management of classical homocystinuria (HCU) is suboptimal and hard to follow. To address unmet need, we developed an enzyme therapy (OT-58), which effectively corrected disease symptoms in various mouse models of HCU in the absence of methionine restriction. Here we evaluated short- and long-term efficacy of OT-58 on the background of current dietary management of HCU. Methionine restriction resulted in the lowering of total homocysteine (tHcy) by 38-63% directly proportional to a decreased methionine intake (50-12.5% of normal). Supplemental betaine resulted in additional lowering of tHcy. OT-58 successfully competed with betaine and normalized tHcy on the background of reduced methionine intake, while substantially lowering tHcy in mice on normal methionine intake. Betaine was less effective in lowering tHcy on the background of normal or increased methionine intake, while exacerbating hypermethioninemia. OT-58 markedly reduced both hyperhomocysteinemia and hypermethioninemia caused by the diets and betaine in HCU mice. Withdrawal of betaine did not affect improved metabolic balance, which was established and solely maintained by OT-58 during periods of fluctuating dietary methionine intake. Taken together, OT-58 may represent novel, highly effective enzyme therapy for HCU performing optimally in the presence or absence of dietary management of HCU.

Robust Classification of High-Dimensional Spectroscopy Data Using Deep Learning and Data Synthesis.

This paper presents a new approach to classification of high-dimensional spectroscopy data and demonstrates that it outperforms other current state-of-the art approaches. The specific task we consider is identifying whether samples contain chlorinated solvents or not, based on their Raman spectra. We also examine robustness to classification of outlier samples that are not represented in the training set (negative outliers). A novel application of a locally connected neural network (NN) for the binary classification of spectroscopy data is proposed and demonstrated to yield improved accuracy over traditionally popular algorithms. Additionally, we present the ability to further increase the accuracy of the locally connected NN algorithm through the use of synthetic training spectra, and we investigate the use of autoencoder based one-class classifiers and outlier detectors. Finally, a two-step classification process is presented as an alternative to the binary and one-class classification paradigms. This process combines the locally connected NN classifier, the use of synthetic training data, and an autoencoder based outlier detector to produce a model which is shown to both produce high classification accuracy and be robust in the presence of negative outliers.

Prevalence, characteristics, and costs of diagnosed homocystinuria, elevated homocysteine, and phenylketonuria in the United States: a retrospective claims-based comparison.

BACKGROUND: Classical homocystinuria (HCU), an inborn error of homocysteine metabolism, has previously been estimated to affect approximately 1 in 100,000-200,000 people in the United States (US). HCU is poorly detected by newborn screening, resulting in underestimates of its prevalence. This study compared characteristics, healthcare use and costs, and projected prevalence between patients with diagnosed HCU, elevated total homocysteine (tHcy), and diagnosed phenylketonuria (PKU). METHODS: Patients in the MarketScan® Research Databases were identified with strictly-defined HCU (> 2 diagnoses, including 1 ICD-10), broadly-defined HCU (> 1 ICD-10), elevated tHcy (> 20 µmol/L) without an HCU diagnosis, or > 1 ICD-9/ICD-10 PKU diagnosis during 1/1/2010-12/31/2016 (first qualifying claim = index). Demographics and healthcare utilization and costs per patient per month (PPPM) were compared between all cohorts, frequencies of comorbidities and medications were compared between HCU and elevated tHcy patients, and healthcare provider types were assessed among HCU patients. The prevalence of patients meeting each cohort definition was projected to the United States (US) population. RESULTS: Patients with strictly-defined (N = 2450) and broadly-defined (N = 6613) HCU, and with elevated tHcy (N = 2017), were significantly older than PKU patients (N = 5120) (57 vs. 56 vs. 53 vs. 18 years; p < 0.05). Vitamin D deficiency, hyperlipidemia, folic acid/B vitamins, and lipid-lowering medications, among others, were more common among diagnosed HCU patients vs. those with elevated tHcy (all p < 0.05). Rates of healthcare utilization were generally higher among HCU and elevated tHcy patients, compared to PKU, though total healthcare costs were similar between groups. Most HCU patients (~ 38%) received their index diagnosis from a primary care physician; very few (~ 1%) had any claim from a geneticist during their enrollment. The age-adjusted national prevalence of HCU was projected at 31,162 (95% CI: 30,411 - 31,913; ~ 1 in 10,000 of the US population) using the broad definition. CONCLUSIONS: The actual prevalence of HCU may be > 10 times prior estimates, at 1 in 10,000 in the US, and this study suggests that HCU is not being diagnosed until later in life. Improvements to newborn screening, detection in young children, and physician education regarding HCU among patients may be necessary to alleviate the burden of this genetic disease.

Estimated prevalence of moderate to severely elevated total homocysteine levels in the United States: A missed opportunity for diagnosis of homocystinuria?

Classical homocystinuria (HCU) is a genetic disorder caused by mutations in the cystathionine beta synthase gene, which results in impaired metabolism of the sulfur-bearing amino acid homocysteine and its accumulation in blood and tissues. Classical HCU can be detected via newborn screening in the United States, but the test is widely acknowledged to miss many patients. While severely elevated homocysteine levels (>100 µmol /L) frequently lead to a classical HCU diagnosis, intermediate levels (>30 to 100 µmol /L), though linked to many of the known complications of HCU, are not always recognized as associated with HCU. We aimed to identify and describe potentially undiagnosed classical HCU patients using a nationally-representative database of administrative claims and laboratory results. We estimated the national prevalence of patients with homocysteine >30 µmol /L, and compared their demographic and clinical characteristics to those of patients with homocysteine levels ≤30 µmol/L. Among 57,580 patients with a homocysteine test result, 1.8% had a value >30 µmol /L. Patients with homocysteine >30 µmol /L were more frequently diagnosed with hypothyroidism (39.2% vs. 20.7%, p < .001) and renal disease (9.7% vs. 5.5%, p < .001), and were more likely to have a prescription for an anxiolytic/antidepressant (44.5% vs. 38.9%), opioid (58.4% vs. 53.1%), steroid (46.4% vs. 42.5%), or thyroid hormone (38.8% vs. 18.8%), compared to patients with homocysteine ≤30 µmol /L (all p < .05). Both groups were equally likely to have a diagnosis of homocystinuria or another disorder of sulfur-bearing amino acid metabolism (3.8% vs. 4.0%, p = .752). The age-adjusted national prevalence of homocysteine >30 µmol /L was estimated at 33,068 (95% CI: 1033 - 35,104). These findings suggest that thousands of people in the US may be living with intermediate to severely elevated homocysteine levels and may require further evaluation for the presence of classical HCU.

Effect of Antihuman T Lymphocyte Globulin on Immune Recovery after Myeloablative Allogeneic Stem Cell Transplantation with Matched Unrelated Donors: Analysis of Immune Reconstitution in a Double-Blind Randomized Controlled Trial.

We recently conducted a randomized double-blind study in which we demonstrated that moderate/severe chronic graft-versus-host disease (cGVHD) but not cGVHD-free survival was reduced in patients receiving anti-T lymphocyte globulin (ATLG) versus placebo. In a companion study we performed immunophenotypic analysis to determine the impact of ATLG on immune reconstitution (IR) and to correlate IR with clinical outcomes. The randomized study (n = 254) included patients (aged 18 to 65 years) who underwent myeloablative transplants for acute myeloid leukemia, myelodysplastic syndrome, or acute lymphoblastic leukemia from HLA-matched unrelated donors. Ninety-one patients consented for the companion IR study (ATLG = 44, placebo = 47). Blood samples were collected on days 30, 100, 180, and 360 after hematopoietic cell transplantation (HCT), and multiparameter flow cytometry was performed in a blinded fashion. Reconstitution of CD3+ and CD4+ T cells was delayed up to 6 months post-HCT in the ATLG arm, whereas absolute regulatory T cell (Treg) (CD4+25+127-) numbers were lower only in the first 100 days. Analysis of the CD4+ Treg and conventional T cells (Tconv) (CD4+25-127+) compartments showed a profound absence of naive Tregs and Tconv in the first 100 days post-HCT, with very slow recovery for 1 year. B cell and natural killer cell recovery were similar in each arm. Higher absolute counts of CD3+, CD4+, CD8+ T, Tregs, and Tconv were associated with improved overall survival, progression-free survival, and nonrelapse mortality but not moderate/severe cGVHD. Although ATLG delays CD3+ and CD4+ T cell recovery post-transplant, it has a relative Treg sparing effect after the early post-HCT period, with possible implications for protection from cGVHD. ATLG severely compromises the generation of naive CD4+ cells (Treg and Tconv), potentially affecting the diversity of the TCR repertoire and T cell responses against malignancy and infection.

Pharmacokinetics and pharmacodynamics of PEGylated truncated human cystathionine beta-synthase for treatment of homocystinuria.

AIMS: PEGylated human truncated cystathionine beta-synthase, lacking the C-terminal regulatory domain (PEG-CBS), is a promising preclinical candidate for enzyme replacement therapy in homocystinuria (HCU). It was designed to function as a metabolic sink to decrease the severely elevated plasma and tissue homocysteine concentrations. In this communication, we evaluated pharmacokinetics (PK), pharmacodynamics (PD) and sub-chronic toxicity of PEG-CBS in homocystinuric mice, wild type rats and monkeys to estimate the minimum human efficacious dose for clinical trials. MAIN METHODS: Animal models received single or multiple doses of PEG-CBS. Activity of PEG-CBS and sulfur amino acid metabolites were determined in plasma and used to determine PK and PD. KEY FINDINGS: The plasma half-lives of PEG-CBS after a single subcutaneous (SC) injection were approximately 20, 44 and 73 h in mouse, rat and monkey, respectively. The SC administration of PEG-CBS resulted in a significant improvement or full correction of metabolic imbalance in both blood and tissues of homocystinuric mice. The PD of PEG-CBS in mouse was dose-dependent, but less than dose-proportional, with the maximal efficacy achieved at 8 mg/kg. PEG-CBS was well-tolerated in mice and monkeys, but resulted in dose-dependent minimal-to-moderate inflammation at the injection sites and vacuolated macrophages in rats. Allometric scaling of animal data was linear and the estimated human efficacious dose was determined as 0.66 mg/kg administered once a week. SIGNIFICANCE: These results provide critical preclinical data for the design of first-in-human PEG-CBS clinical trial.

Prospective, Randomized, Double-Blind, Phase III Clinical Trial of Anti-T-Lymphocyte Globulin to Assess Impact on Chronic Graft-Versus-Host Disease-Free Survival in Patients Undergoing HLA-Matched Unrelated Myeloablative Hematopoietic Cell Transplantation.

Purpose Several open-label randomized studies have suggested that in vivo T-cell depletion with anti-T-lymphocyte globulin (ATLG; formerly antithymocyte globulin-Fresenius) reduces chronic graft-versus-host disease (cGVHD) without compromising survival. We report a prospective, double-blind phase III trial to investigate the effect of ATLG (Neovii Biotech, Lexington, MA) on cGVHD-free survival. Patients and Methods Two hundred fifty-four patients 18 to 65 years of age with acute leukemia or myelodysplastic syndrome who underwent myeloablative HLA-matched unrelated hematopoietic cell transplantation (HCT) were randomly assigned one to one to placebo (n =128 placebo) or ATLG (n = 126) treatment at 27 sites. Patients received either ATLG or placebo 20 mg/kg per day on days -3, -2, -1 in addition to tacrolimus and methotrexate as GVHD prophylaxis. The primary study end point was moderate-severe cGVHD-free survival. Results Despite a reduction in grade 2 to 4 acute GVHD (23% v 40%; P = .004) and moderate-severe cGVHD (12% v 33%; P < .001) in ATLG recipients, no difference in moderate-severe cGVHD-free survival between ATLG and placebo was found (2-year estimate: 48% v 44%, respectively; P = .47). Both progression-free survival (PFS) and overall survival (OS) were lower with ATLG (2-year estimate: 47% v 65% [ P = .04] and 59% v 74% [ P = .034], respectively). Multivariable analysis confirmed that ATLG was associated with inferior PFS (hazard ratio, 1.55; 95% CI, 1.05 to 2.28; P = .026) and OS (hazard ratio, 1.74; 95% CI, 1.12 to 2.71; P = .01). Conclusion In this prospective, randomized, double-blind trial of ATLG in unrelated myeloablative HCT, the incorporation of ATLG did not improve moderate-severe cGVHD-free survival. Moderate-severe cGVHD was significantly lower with ATLG, but PFS and OS also were lower. Additional analyses are needed to understand the appropriate role for ATLG in HCT.