Abstract
Purpose:
Replication of diabetic maculopathy based on a model initiated with normal macular capillaries from adaptive optics scanning laser ophthalmoscopy imaging. Using the model with actual capillary anatomy, to test if relatively simple physiologically based assumptions can predict the pathology seen in diabetic patients. Giving a deeper understanding of the actual pathophysiology of ischemic and edematous diabetic maculopathy. To apply the model to vascular networks determined in individual patients allowing prediction of natural history and response to therapeusis with anti-VEGF agents.
Methods:
With a multiscale Compucell3D model using input vascular structure of multiple subjects' macular capillaries from AOSLO, models were made which included a)flow and pressure calculations for each point on the vascular map from the arterioles, through the capillaries, to the venules, b)an underlying cellular field with a basal VEGF secretion which increases with hypoxia and declines with cellular VEGF uptake, c)oxygen distributions determined by diffusion from the vessels and consumption by the cellular field,d)a probabilitic model of capillary closure based on VEGF levels modelling 'stickiness' of leukocytes and resultant retinal capillary closure,e)recurrent iterations of the resultant oxygen distribution and consequent VEGF distribution subsequent to each capillary closure.
Results:
The application of the Compucell3D model to the vascular anatomy of a number of subjects yielded results highly similar to those seen in actual patients. A pattern is seen in which the loss of a given capillary increases the susceptibility of surrounding capillaries to closure. It does this through the intermediary of local VEGF concentrations. Areas of ischemia and edema increase in size over time. Certain patterns of vascular anatomy seem to increase vulnerability to capillary closure and development of macular edema. There has been inadequate time to allow comparison of the model to the actual clinically observed patterns of macular capillary closure seen in individual patients. This would be the test of whether this modelling approach can provide an important aspect of individualized medicine in diabetic retinopathy.
Conclusions:
The Compucell3D modelling of the development of diabetic maculopathy yielded results highly comparable to those seen in patients but further verification of clinical utility will require longitudinal studies.