Development and Validation of a Model to Predict Fast Progression in Glaucoma

Summary

Participation Requirements

Description

The purpose of this study is to obtain genetic data on a well-defined population of patients with glaucoma that had accurate measurements of rates of structural and functional change over time, develop a predictive model for fast progression in glaucoma, and then subsequently assess the efficacy of ...

The purpose of this study is to obtain genetic data on a well-defined population of patients with glaucoma that had accurate measurements of rates of structural and functional change over time, develop a predictive model for fast progression in glaucoma, and then subsequently assess the efficacy of the predictive model in a prospective cohort of patients. The investigators hypothesize that incorporating genetic and clinical data will lead to a powerful model to predict rates of change in the disease. Glaucoma is a neurodegenerative disease caused by progressive retinal ganglion cell (RGC) loss associated with characteristic structural changes in the optic nerve and retinal nerve fiber layer (RNFL). The neural insult can result in irreversible loss of vision and disability. Detection of progression plays a central role in the diagnosis and management of glaucoma. However, although most glaucoma patients will show some evidence of progression if followed long enough, the rate of deterioration can be highly variable among them. While most patients progress relatively slowly, others have aggressive disease with fast deterioration, which can eventually result in disability unless appropriate interventions take place. Many risk factors have been shown to be associated with glaucoma progression, such as older age, high intraocular pressure (IOP), thin corneas, presence of disc hemorrhage, among others. However, previous attempts to develop models that can predict future rates of disease progression have been limited. For example, a "risk calculator" for progressive glaucoma developed by De Moraes et al had an R2 of only 13% to explain rates of disease progression. Such limited predictive ability may be largely explained by different inherent susceptibility of individual eyes to risk factors such as IOP. In fact, it is well-known that certain eyes can develop rapid deterioration despite relatively low IOP levels and vice-versa. The development of accurate model to predict risk of fast progression in glaucoma would fulfill an extremely important unmet need by allowing tailoring treatment aggressiveness and monitoring according to risk. In addition, it would lead to a tool that could be used to identify patients that would most likely benefit from therapeutic interventions designed to protect the optic nerve, i.e, neuroprotective. Use of predictive statistical models could also lead to more effective clinical trials investigating new neuroprotective therapies, by targeting patient populations at most need. The study will involve two phases. Phase I will include patients selected from the Duke Glaucoma Registry (DGR), a database of electronic medical records developed by the Vision, Imaging and Performance (VIP) Laboratory of Dr. Medeiros. The database consists of adults at least 18 years of age with glaucoma or glaucoma suspect diagnoses who were evaluated at the Duke Eye Center or its satellite clinics between 2009 and 2018. The Duke University Institutional Review Board (IRB) approved the study involving database set up and analysis of progression under protocol PRO00103899 "Evaluation of progression of eye diseases from a large clinical population", with a waiver of informed consent due to the retrospective nature of this work. The present protocol will be prepared to involve consenting and collection of genetic material for developing a statistical model for predicting risk of fast progression in glaucoma. In addition, Phase II of this study will involve prospective recruitment of subjects for testing and validation of the predictive model. The database to be used in Phase I contains retrospective clinical information from baseline and follow-up visits, including patient diagnostic and procedure codes, medical history, photographs, and the results of all Spectralis Spectral-Domain Optical Coherence Tomography (SD OCT) and standard automated perimetry (SAP), acquired with the Humphrey Field Analyzer. SD OCT and SAP have been evaluated for quality and reliability. SD OCT images with quality less than 15 dB were excluded and SAP tests with false-positive results greater than 15% or fixation losses greater than 33% were also excluded. Phase I: Collection of genetic material Patients will be included if they had primary open-angle glaucoma (POAG) or suspicion of glaucoma based on International Classification of Diseases (ICD) codes at baseline. Selection of subjects will be based on analysis of retrospective de-identified data collected as part of PRO00103899 "Evaluation of progression of eye diseases from a large clinical population", PI Dr Felipe Medeiros. The database contains longitudinal data from subjects that have been evaluated at the Duke Eye Center and satellites. Classification of eyes as fast versus slow progressors will be based on historical OCT and SAP data available in the database. Preliminary analysis has identified 15,394 subjects in the database that have longitudinal SD OCT and SAP. Of these, 33% have been identified with at least fast progression by SD OCT and 15% by SAP, according to the definitions provided in the Table below. Tests performed after any diagnosis of retinal detachment, retinal or malignant choroidal tumors, non-glaucomatous disorders of the optical nerve and visual pathways, and venous or arterial retinal occlusion according to ICD codes are excluded. In addition, tests performed after treatment with pan-photocoagulation, according to Current Procedural Terminology (CPT) codes, are excluded from analysis of rates of progression. Using existing data on SD OCT RNFL thickness and SAP, eyes will be classified as slow, moderate, fast and catastrophic progressors The categorization of rates of progression will be useful for the purpose of patient recruitment for collection of genetic material, so that a target balance of subjects progressing at slow x fast rates can be achieved. However, the actual quantitative rate of progression will be available for analysis of the relationship with genetic markers as well. Of note, due to the nature of the structure and function relationship, these tests may identify different patients as fast progressors, but both are clinically relevant. Slow progressors will have slow progression on both tests. Collection of genetic material The Phase I of this proposal aims at recruiting 5,000 patients over a period of 2 years for obtaining blood samples which will be processed for extraction of DNA and genotyping. Patients will be selected from the DGR database and will be contacted by phone, mail or email using a standardized script and procedure. Patients will be invited for a single visit for blood collection. 10 ml of blood will be collected using standard procedures. The samples will be de-identified and receive a masked identification (ID) which will correspond to the existing masked ID in the DGR database. 2 ml will be sent to Genentech using frozen samples for genotyping using the Illumina Global Screening Array (GSA) kit. The remaining will be processed and stored at the Duke Eye Center. These samples may be maintained indefinitely or until they are exhausted. These samples will not be available for diagnostic or therapeutic purposes. The genetic test is for research purposes only. The research tests are not being used as diagnostic tests for any disease or illness. Subjects will receive no results from this study unless they indicate a risk of treatable serious illness, such as those described in the American College of Medical Genetics and Genomics report for release of incidental findings. Incidental Findings will be confirmed by an approved lab before providing to a subject or his/her physician. The investigators will contact the IRB when an incidental finding has been made that appears to be clinically actionable, and in particular, if it relates to a serious or life-threatening condition, to discuss the specific circumstances and how best to proceed. With the subject's permission, the blood sample may be shared anonymously with other investigators for research purposes. Such research will be strictly anonymous. The study will follow The Genetic Information Nondiscrimination Act (GINA) and PHI will be maintained in a confidential, secure location at Duke University Health System (DUHS), accessible only by authorized members of the study team, and will not be disclosed to third parties, except as described above, or required by law. Development of Predictive Statistical Model The development of the predictive model for fast progression will be performed in collaboration with Genentech. Genentech will be responsible for genotyping the samples. Data to be transferred to Genentech will include SD OCT images and visual field data, as well as general clinical data such as gender, race, age at time of testing (capped as >90 if older than 90), intraocular pressures over time, axial length, central corneal thickness, results of ophthalmic exam over time, ICD and CPT codes. All data will be De-identified before leaving Duke in accordance with the Safe Harbor Method. No PHI will be transmitted. Data will be transmitted to Genentech using Duke Box. Genentech will provide the PI with all results of genotyping in a suitable format. In addition, Genentech will provide the PI with the ability to obtain predictions using the developed predictive model. Phase II: Prospective Validation of Predictive Model For Phase II, new subjects will be prospectively recruited from Dr. Medeiros' clinic at Duke Eye Center for testing of the predictive model to identify fast progression. Baseline data on clinical variables and genotyping will be collected from these subjects and the relevant variables will be entered into the predictive model to obtain predictions. Patients will then be prospectively followed with cluster-based intensive structural and functional testing at Dr. Medeiros' lab. Testing will include SD OCT imaging of the optic nerve head, macula and retinal nerve fiber layer, as well as visual field testing with the Swedish Testing Algorithm (SITA) 24-2. Subjects will undergo 3 clusters of testing at baseline, 6 months and 1 year, with each cluster consisting of 5 repeated tests acquired within an interval of 2 to 3 weeks. The predictions from the model will then be compared to the actual prospective observations of progression to assess their efficacy. A sample size of 200 eyes of 100 subjects will be used for Phase II. Study Duration: Phase I - Blood Sample Collection: 2 years Phase II - Data exchange and development/validation of predictive model: 1 year

Tracking Information