To Study the Role of Personalized Risk Factors for Poor Response to Angiostatic Therapy Based on Regression Models of the Disease

Purpose : to study the role of personalized risk factors for poor response to angiostatic therapy based on a regression model of the disease. Patients and methods . The study had a retrospective cohort design, including 84 patients (92 eyes) with a newly diagnosed MNV and evidence of disease activity. In 18 (19.5 %) people, the process was bilateral. Personalized retinal morphometrics and intraocular pressure (IOP) in eyes with MNV treated with fixed intravitreal injections (IVI) of Eylea in correlation with treatment response were studied. Treatment response was ranked into five types. Treatment results were assessed clinically and morphometrically by optical coherence tomography (OCT). Based on the results of correlations, logistic regression models of the disease were built. They studied the role of various factors in the treatment of the disease and the strength of their association with an adverse outcome. Results . 2 models of the disease have been developed and presented in the form of logistic regression equations with formulas for calculating the prognosis of the effectiveness of treatment with angiostatics. The strength of the risk factors association for an adverse response was analyzed, such as the height of pigment epithelial detachment, the presence of intraretinal edema, personalized intraocular pressure indicators of the affected eye, the patient’s disease belonging to the hereditary form of AMD, the presence of bad habits (smoking). It is shown that an increase in the height of pigment epithelial detachment (marker PED_H = 170 microns) by 1 micron, with other unchanged predictors, increases the ratio of the chance of a negative outcome to a positive one by 1%. At this marker value, the sensitivity of the poor response to angiostatic treatment was 0.493 and the specificity was 0.737. When intraretinal edema occurs, the ratio of the chance of a negative outcome to a positive outcome increases by 68 %. In conditions of increased intraocular pressure, the risk of a poor response increases by 4 times. The prognosis for a smoker is 30% worse compared to a nonsmoking patient. Hereditary AMD increases the risk of developing a poor response to treatment by 25 %. Conclusion. A logistic regression model of the disease is presented, which correctly predicts 94 % of non-response, which makes the prognosis reliable, reliable and accurate. Modifiable (smoking, PED_H > 170 microns, presence of IRF, increased IOP) and non-modifiable (hereditary norm of the disease) risk factors that are important in practical ophthalmology and increase the likelihood of developing an unfavorable outcome of angiostatic treatment were analyzed.