Improving Prediction of Risk of Hospital Admission in Chronic Obstructive Pulmonary Disease: Application of Machine Learning to Telemonitoring Data.

BACKGROUND: Telemonitoring of symptoms and physiological signs has been suggested as a means of early detection of chronic obstructive pulmonary disease (COPD) exacerbations, with a view to instituting timely treatment. However, algorithms to identify exacerbations result in frequent false-positive results and increased workload. Machine learning, when applied to predictive modelling, can determine patterns of risk factors useful for improving prediction quality. OBJECTIVE: Our objectives were to (1) establish whether machine learning techniques applied to telemonitoring datasets improve prediction of hospital admissions and decisions to start corticosteroids, and (2) determine whether the addition of weather data further improves such predictions. METHODS: We used daily symptoms, physiological measures, and medication data, with baseline demography, COPD severity, quality of life, and hospital admissions from a pilot and large randomized controlled trial of telemonitoring in COPD. We linked weather data from the United Kingdom meteorological service. We used feature selection and extraction techniques for time series to construct up to 153 predictive patterns (features) from symptom, medication, and physiological measurements. We used the resulting variables to construct predictive models fitted to training sets of patients and compared them with common symptom-counting algorithms. RESULTS: We had a mean 363 days of telemonitoring data from 135 patients. The two most practical traditional score-counting algorithms, restricted to cases with complete data, resulted in area under the receiver operating characteristic curve (AUC) estimates of 0.60 (95% CI 0.51-0.69) and 0.58 (95% CI 0.50-0.67) for predicting admissions based on a single day's readings. However, in a real-world scenario allowing for missing data, with greater numbers of patient daily data and hospitalizations (N=57,150, N+=55, respectively), the performance of all the traditional algorithms fell, including those based on 2 days' data. One of the most frequently used algorithms performed no better than chance. All considered machine learning models demonstrated significant improvements; the best machine learning algorithm based on 57,150 episodes resulted in an aggregated AUC of 0.74 (95% CI 0.67-0.80). Adding weather data measurements did not improve the predictive performance of the best model (AUC 0.74, 95% CI 0.69-0.79). To achieve an 80% true-positive rate (sensitivity), the traditional algorithms were associated with an 80% false-positive rate: our algorithm halved this rate to approximately 40% (specificity approximately 60%). The machine learning algorithm was moderately superior to the best symptom-counting algorithm (AUC 0.77, 95% CI 0.74-0.79 vs AUC 0.66, 95% CI 0.63-0.68) at predicting the need for corticosteroids. CONCLUSIONS: Early detection and management of COPD remains an important goal given its huge personal and economic costs. Machine learning approaches, which can be tailored to an individual's baseline profile and can learn from experience of the individual patient, are superior to existing predictive algorithms and show promise in achieving this goal. TRIAL REGISTRATION: International Standard Randomized Controlled Trial Number ISRCTN96634935; http://www.isrctn.com/ISRCTN96634935 (Archived by WebCite at http://www.webcitation.org/722YkuhAz).

N-Glycan Profile and Kidney Disease in Type 1 Diabetes.

OBJECTIVE: Poorer glycemic control in type 1 diabetes may alter N-glycosylation patterns on circulating glycoproteins, and these alterations may be linked with diabetic kidney disease (DKD). We investigated associations between N-glycans and glycemic control and renal function in type 1 diabetes. RESEARCH DESIGN AND METHODS: Using serum samples from 818 adults who were considered to have extreme annual loss in estimated glomerular filtration rate (eGFR; i.e., slope) based on retrospective clinical records, from among 6,127 adults in the Scottish Diabetes Research Network Type 1 Bioresource Study, we measured total and IgG-specific N-glycan profiles. This yielded a relative abundance of 39 total (GP) and 24 IgG (IGP) N-glycans. Linear regression models were used to investigate associations between N-glycan structures and HbA1c, albumin-to-creatinine ratio (ACR), and eGFR slope. Models were adjusted for age, sex, duration of type 1 diabetes, and total serum IgG. RESULTS: Higher HbA1c was associated with a lower relative abundance of simple biantennary N-glycans and a higher relative abundance of more complex structures with more branching, galactosylation, and sialylation (GP12, 26, 31, 32, and 34, and IGP19 and 23; all P < 3.79 × 10-4). Similar patterns were seen for ACR and greater mean annual loss of eGFR, which were also associated with fewer of the simpler N-glycans (all P < 3.79 × 10-4). CONCLUSIONS: Higher HbA1c in type 1 diabetes is associated with changes in the serum N-glycome that have elsewhere been shown to regulate the epidermal growth factor receptor and transforming growth factor-ß pathways that are implicated in DKD. Furthermore, N-glycans are associated with ACR and eGFR slope. These data suggest that the role of altered N-glycans in DKD warrants further investigation.

Automated Integration of a UPLC Glycomic Profile.

Ultra-performance liquid chromatography (UPLC) is the established technology for accurate analysis of IgG Fc N-glycosylation due to its superior sensitivity, resolution, speed, and its capability to provide branch-specific information of glycan species. Correct and cost-efficient preprocessing of chromatographic data is the major prerequisite for subsequent analyses ranging from inference of structural isomers to biomarker discovery and prediction of humoral immune response from characterized changes in glycosylation. The complexity of glycomic chromatograms poses a number of challenges for developing automated data annotation and quantitation algorithms, which frequently necessitated manual or semi-manual approaches to preprocessing, most notably to peak detection and integration. Such procedures are meticulous and time-consuming, and may be a source of confounding due to their dependence on human labelers. Although liquid chromatography is a mature field and a number of methods have been developed for automatic peak detection outside the area of glycomics analysis, we found that hardly any of them are suitable for automatic integration of UPLC glycomic profiles without substantial modifications. In this chapter, we illustrate practical challenges of automatic peak detection of UPLC glycomics chromatograms. We outline a robust, semi-supervised method ACE (Automatic Chromatogram Extraction) for automated alignment and detection of glycan peaks in chromatograms, developed by Pharmatics Limited (UK) in collaboration with Genos Limited (Croatia). Application of the tool requires minimal human interference, which results in a significant reduction in the time and cost of IgG glycomics signal integration using Waters Acquity UPLC instrument (Milford, MA, USA) in several human cohorts with blind technical replicas.

Glycosylation of plasma IgG in colorectal cancer prognosis.

In this study we demonstrate the potential value of Immunoglobulin G (IgG) glycosylation as a novel prognostic biomarker of colorectal cancer (CRC). We analysed plasma IgG glycans in 1229 CRC patients and correlated with survival outcomes. We assessed the predictive value of clinical algorithms and compared this to algorithms that also included glycan predictors. Decreased galactosylation, decreased sialylation (of fucosylated IgG glycan structures) and increased bisecting GlcNAc in IgG glycan structures were strongly associated with all-cause (q < 0.01) and CRC mortality (q = 0.04 for galactosylation and sialylation). Clinical algorithms showed good prediction of all-cause and CRC mortality (Harrell's C: 0.73, 0.77; AUC: 0.75, 0.79, IDI: 0.02, 0.04 respectively). The inclusion of IgG glycan data did not lead to any statistically significant improvements overall, but it improved the prediction over clinical models for stage 4 patients with the shortest follow-up time until death, with the median gain in the test AUC of 0.08. These glycan differences are consistent with significantly increased IgG pro-inflammatory activity being associated with poorer CRC prognosis, especially in late stage CRC. In the absence of validated biomarkers to improve upon prognostic information from existing clinicopathological factors, the potential of these novel IgG glycan biomarkers merits further investigation.