Predictive models of choroidal neovascularization and geographic atrophy incidence applied to clinical trial design

Hyperpigmentary abnormalities in age-related macular degeneration: association with progression and impact on visual sensitivity

BACKGROUND/AIMS

To investigate the additional prognostic value of quantifying the extent of colour fundus photography (CFP)-defined hyperpigmentary abnormalities (HPAs) compared with their presence alone for predicting progression to late-stage age-related macular degeneration (AMD) and to understand their association with visual sensitivity in individuals with intermediate AMD.

METHODS

140 participants with bilateral large drusen underwent multimodal imaging and microperimetry at baseline and then every 6 months for up to 3 years. Baseline CFPs were graded for the presence of HPAs and their extent was quantified. Optical coherence tomography (OCT) scans were used to quantify drusen volume. Predictive models for progression to late AMD (including OCT signs of atrophy) were developed using either HPA presence or extent. The association between HPA extent with mean visual sensitivity (both overall and sector based) was also evaluated. All models were adjusted for the confounders of baseline age and drusen volume.

RESULTS

The predictive performance for late AMD development was not significantly different for HPA presence or extent (p=0.92). Increasing HPA extent in each sector, but not its overall extent in an eye, was associated with reduced sector-based visual sensitivity (p<0.001 and p=0.671, respectively).

CONCLUSION

In a cohort with bilateral large drusen, quantifying HPA extent did not improve the prediction of late AMD development compared with presence alone. HPA extent was associated with more local, rather than generalised, reductions in visual sensitivity. These findings suggest that quantification of HPA extent adds little to the prediction of AMD progression, but that it provides an imaging biomarker of visual dysfunction.

Reticular Pseudodrusen on the Risk of Progression in Intermediate Age-Related Macular Degeneration

PURPOSE

To examine the association between reticular pseudodrusen (RPD) and progression to late age-related macular degeneration (AMD) in individuals with intermediate AMD.

DESIGN

Prospective cohort study.

METHODS

Two hundred eighty eyes from 140 participants with bilateral large drusen underwent multimodal imaging (MMI), including optical coherence tomography (OCT), near-infrared reflectance (NIR), fundus autofluorescence, and color fundus photography (CFP), at 6-monthly intervals up over a 36-month follow-up period. The presence of RPD per eye was determined based on either a combined MMI criterion, or each individual imaging modality, and their extent measured on combined OCT and NIR imaging. The association between the presence of RPD on different imaging modalities, and their extent, with the development of late AMD (including OCT-defined atrophy) was evaluated.

RESULTS

The presence of RPD on MMI, or any of its individual modalities, at baseline was not significantly associated with an increased rate of developing late AMD, with or without adjusting for risk factors for AMD progression (age, drusen volume on OCT, and pigmentary abnormalities on CFP; all P ≥ 0.205). The extent of RPD present was also not significantly associated with an increased rate of developing late AMD, with or without adjustment for risk factors for AMD progression (both P ≥ 0.522).

CONCLUSIONS

In this cohort with bilateral large drusen, the presence of RPD was not significantly associated with an increased risk of developing late AMD. Additional longitudinal studies in all stages of AMD are needed to understand the implications of RPD on vision loss in this condition.

Hyporeflective Cores within Drusen: Association with Progression of Age-Related Macular Degeneration and Impact on Visual Sensitivity

PURPOSE

To examine the association between hyporeflective cores within drusen (HCD) and disease progression in age-related macular degeneration (AMD) and with visual function.

DESIGN

Longitudinal observational study.

PARTICIPANTS

280 eyes from 140 participants with bilateral large drusen, without late AMD.

METHODS

Multimodal imaging (MMI) and microperimetry were performed at baseline, and then every 6 months for up to 3 years. Baseline optical coherence tomography (OCT) scans were graded for the presence of HCD and used to calculate drusen volume. The total area of the drusenoid lesions containing hyporeflective cores (HCD extent) on color fundus photographs (CFPs) was calculated. CFPs were also graded for the presence of pigmentary abnormalities. The association between HCD extent with progression to late AMD (including OCT signs of atrophy) and visual sensitivity measured by microperimetry at baseline and its rate of change over time was evaluated, with and without adjustment for confounders of drusen volume, pigmentary abnormalities, and age.

MAIN OUTCOME MEASURES

Time to develop late AMD and visual sensitivity.

RESULTS

Twenty (7%) eyes from 12 (9%) individuals were found to have HCD at baseline, which was associated with a non-significantly increased rate of progression to late AMD (unadjusted P = 0.050). HCD extent was significantly associated with an increased rate of progression to late AMD (unadjusted P = 0.034) and lower visual sensitivity at baseline (unadjusted P < 0.001). However, these associations were no longer significant (P ≥ 0.264 for both) after adjusting for known risk factors for AMD progression. HCD extent was also not associated with a faster rate of visual sensitivity decline prior to the development of late AMD, with or without adjustment (P ≥ 0.674 for both). Increasing age and larger drusen volume was associated with HCD extent (P ≤ 0.041).

CONCLUSIONS

In a cohort with bilateral large drusen, HCD presence and extent was not independently associated with an increased rate of progression to late AMD over 3 years, nor with lower visual sensitivity or an increased rate of visual sensitivity decline prior to the development of late AMD, after adjusting for known risk factors of disease progression.

Evaluating the Validity of the Age-Related Eye Disease Study Grading Scale for Age-Related Macular Degeneration: AREDS2 Report 10

IMPORTANCE

To test potential treatments for age-related macular degeneration (AMD), clinical trials need standardized outcome measures that are valid for predicting AMD progression in different study populations.

OBJECTIVE

To evaluate the validity of the Age-Related Eye Disease Study (AREDS) detailed and simple AMD severity scales by comparing rates of development of late AMD (neovascular AMD and/or central geographic atrophy) between AREDS and AREDS2 participants.

DESIGN, SETTING, AND PARTICIPANTS

Both AREDS (1992-2001) and AREDS2 (2006-2012) enrolled patients from academic and community-based retinal practices across the United States. In AREDS (n = 4519), participants with varying severity of AMD-from no AMD to late AMD in 1 eye-were enrolled. In AREDS2 (n = 4203), participants with bilateral large drusen or large drusen in the study eye and late AMD in the fellow eye were enrolled.

MAIN OUTCOMES AND MEASURES

Five-year incidence of late AMD, assessed by annual masked centralized fundus photograph grading.

RESULTS

In AREDS, the mean (SD) age of the patients was 69.3 (5.7) years, and 2519 (55.7%) were female. In AREDS2, the mean (SD) age of the patients was 73.1 (7.7) years, and 2388 (56.8%) were female. The 5-year rates of late AMD did not differ between AREDS2 and AREDS participants within nearly all baseline AMD detailed severity scale levels: levels 1 to 3: 2.4% vs 0.5% (difference, 1.9%; 95% CI, -0.2% to 4.0%; P < .001); level 4: 6.5% vs 4.9% (difference, 1.6%; 95% CI, -1.7% to 4.8%; P = .34); level 5: 8.0% vs 5.6% (difference, 2.4%; 95% CI, -1.2% to 5.9%; P = .22); level 6: 12.8% vs 13.7% (difference, -0.9%; 95% CI, -4.8% to 3.1%; P = .66); level 7: 26.2% vs 27.8% (difference, -1.5%; 95% CI, -6.6% to 3.5%; P = .54); and level 8: 46.4% vs 44.7% (difference, 1.7%; 95% CI, -7.5% to 10.9%; P = .72). Within simple scale levels, AREDS2 and AREDS 5-year rates did not differ significantly except for level 1 (9.4% vs 3.1%, P = .02; level 2: 12.8% vs 11.8%, P = .65; level 3: 26.3% vs 25.9%, P = .90; and level 4: 45.6% vs 47.3%, P = .57).

CONCLUSIONS AND RELEVANCE

The AREDS detailed and simple AMD severity scales were useful measures for assessing the risk of developing late AMD in the AREDS2 population; these data suggest that they should be useful tools for clinical trials of AMD treatments.

Weibull regression with Bayesian variable selection to identify prognostic tumour markers of breast cancer survival

As data-rich medical datasets are becoming routinely collected, there is a growing demand for regression methodology that facilitates variable selection over a large number of predictors. Bayesian variable selection algorithms offer an attractive solution, whereby a sparsity inducing prior allows inclusion of sets of predictors simultaneously, leading to adjusted effect estimates and inference of which covariates are most important. We present a new implementation of Bayesian variable selection, based on a Reversible Jump MCMC algorithm, for survival analysis under the Weibull regression model. A realistic simulation study is presented comparing against an alternative LASSO-based variable selection strategy in datasets of up to 20,000 covariates. Across half the scenarios, our new method achieved identical sensitivity and specificity to the LASSO strategy, and a marginal improvement otherwise. Runtimes were comparable for both approaches, taking approximately a day for 20,000 covariates. Subsequently, we present a real data application in which 119 protein-based markers are explored for association with breast cancer survival in a case cohort of 2287 patients with oestrogen receptor-positive disease. Evidence was found for three independent prognostic tumour markers of survival, one of which is novel. Our new approach demonstrated the best specificity.

Understanding the undelaying mechanism of HA-subtyping in the level of physic-chemical characteristics of protein

The evolution of the influenza A virus to increase its host range is a major concern worldwide. Molecular mechanisms of increasing host range are largely unknown. Influenza surface proteins play determining roles in reorganization of host-sialic acid receptors and host range. In an attempt to uncover the physic-chemical attributes which govern HA subtyping, we performed a large scale functional analysis of over 7000 sequences of 16 different HA subtypes. Large number (896) of physic-chemical protein characteristics were calculated for each HA sequence. Then, 10 different attribute weighting algorithms were used to find the key characteristics distinguishing HA subtypes. Furthermore, to discover machine leaning models which can predict HA subtypes, various Decision Tree, Support Vector Machine, Naïve Bayes, and Neural Network models were trained on calculated protein characteristics dataset as well as 10 trimmed datasets generated by attribute weighting algorithms. The prediction accuracies of the machine learning methods were evaluated by 10-fold cross validation. The results highlighted the frequency of Gln (selected by 80% of attribute weighting algorithms), percentage/frequency of Tyr, percentage of Cys, and frequencies of Try and Glu (selected by 70% of attribute weighting algorithms) as the key features that are associated with HA subtyping. Random Forest tree induction algorithm and RBF kernel function of SVM (scaled by grid search) showed high accuracy of 98% in clustering and predicting HA subtypes based on protein attributes. Decision tree models were successful in monitoring the short mutation/reassortment paths by which influenza virus can gain the key protein structure of another HA subtype and increase its host range in a short period of time with less energy consumption. Extracting and mining a large number of amino acid attributes of HA subtypes of influenza A virus through supervised algorithms represent a new avenue for understanding and predicting possible future structure of influenza pandemics.

The benefits of using genetic information to design prevention trials

Clinical trials for preventative therapies are complex and costly endeavors focused on individuals likely to develop disease in a short time frame, randomizing them to treatment groups, and following them over time. In such trials, statistical power is governed by the rate of disease events in each group and cost is determined by randomization, treatment, and follow-up. Strategies that increase the rate of disease events by enrolling individuals with high risk of disease can significantly reduce study size, duration, and cost. Comprehensive study of common, complex diseases has resulted in a growing list of robustly associated genetic markers. Here, we evaluate the utility--in terms of trial size, duration, and cost--of enriching prevention trial samples by combining clinical information with genetic risk scores to identify individuals at greater risk of disease. We also describe a framework for utilizing genetic risk scores in these trials and evaluating the associated cost and time savings. With type 1 diabetes (T1D), type 2 diabetes (T2D), myocardial infarction (MI), and advanced age-related macular degeneration (AMD) as examples, we illustrate the potential and limitations of using genetic data for prevention trial design. We illustrate settings where incorporating genetic information could reduce trial cost or duration considerably, as well as settings where potential savings are negligible. Results are strongly dependent on the genetic architecture of the disease, but we also show that these benefits should increase as the list of robustly associated markers for each disease grows and as large samples of genotyped individuals become available.

Inclusion of genotype with fundus phenotype improves accuracy of predicting choroidal neovascularization and geographic atrophy

PURPOSE

The accuracy of predicting conversion from early-stage age-related macular degeneration (AMD) to the advanced stages of choroidal neovascularization (CNV) or geographic atrophy (GA) was evaluated to determine whether inclusion of clinically relevant genetic markers improved accuracy beyond prediction using phenotypic risk factors alone.

DESIGN

Cohort study.

PARTICIPANTS

White, non-Hispanic subjects participating in the Age-Related Eye Disease Study (AREDS) sponsored by the National Eye Institute consented to provide a genetic specimen. Of 2415 DNA specimens available, 940 were from disease-free subjects and 1475 were from subjects with early or intermediate AMD.

METHODS

DNA specimens from study subjects were genotyped for 14 single nucleotide polymorphisms (SNPs) in genes shown previously to associate with CNV: ARMS2, CFH, C3, C2, FB, CFHR4, CFHR5, and F13B. Clinical demographics and established disease associations, including age, sex, smoking status, body mass index (BMI), AREDS treatment category, and educational level, were evaluated. Four multivariate logistic models (phenotype; genotype; phenotype + genotype; and phenotype + genotype + demographic + environmental factors) were tested using 2 end points (CNV, GA). Models were fitted using Cox proportional hazards regression to use time-to-disease onset data.

MAIN OUTCOME MEASURES

Brier score (measure of accuracy) was used to identify the model with the lowest prediction error in the training set. The most accurate model was subjected to independent statistical validation, and final model performance was described using area under the receiver operator curve (AUC) or C-statistic.

RESULTS

The CNV prediction models that combined genotype with phenotype with or without age and smoking revealed superior performance (C-statistic = 0.96) compared with the phenotype model based on the simplified severity scale and the presence of CNV in the nonstudy eye (C-statistic = 0.89; P<0.01). For GA, the model that combined genotype with phenotype demonstrated the highest performance (AUC = 0.94). Smoking status and ARMS2 genotype had less of an impact on the prediction of GA compared with CNV.

CONCLUSIONS

Inclusion of genotype assessment improves CNV prediction beyond that achievable with phenotype alone and may improve patient management. Separate assessments should be used to predict progression to CNV and GA because genetic markers and smoking status do not equally predict both end points.

FINANCIAL DISCLOSURE(S)

Proprietary or commercial disclosure may be found after the references.