A Novel Robotic Laser Therapy For Presbyopia: Artificial Intelligence-Based Prediction And Optimization Of Clinical Outcomes

To evaluate a Virtual Eye Simulation Analyzer (VESA) within a virtual clinical trial (VCT) to plan surgical treatment based upon anterior segment ocular coherence tomography (AS-OCT) imaging using a treatment algorithm and predict central optical power results following laser scleral microporation treatment. 

Small Business Innovation Research (SBIR) grant application experimental study

A VCT was created using artificial intelligence (AI) and deep machine learning in a Finite Element Model (FEM). We recreated in silico human eyes with age- and patient-specific parameters, optical and biomechanical properties. VESA ran simulations of LSM treatments based upon preoperative AS-OCT imaging using algorithms and compared results to actual procedure results from published human clinical data of 26 eyes of 13 presbyopic patients enrolled in an IRB Registered pilot study. 

VESA successfully demonstrated age-related, biomechanical changes contributing to the loss of dynamic range of focus (DRoF). Pretreatment planning using scleral thickness measurements and mathematical algorithms correctly directed treatments at 85% depth. Post-operatively, In silico predictions of central optical power (COP) were compared to their recorded ETDRS distance corrected near visual acuities (DCNVA) at 40cm.  Mean predicted improvement in reading ADD and DCNVA was not significantly different from the actual mean reading ADD and DCNVA following LSM (P>0.05).

VESA successfully simulated biomechanical changes of DRoF with progressive age, and correctly predicted results of human clinical studies using LSM. The multifactorial nature of presbyopia and limited image-based technologies measuring biomechanical processes of accommodation is problematic.  VESA can be used to optimize future treatment outcomes through predictive diagnostics.​