Clinical applications of optical coherence tomography in corneal surgery: Intraoperative optical coherence tomography imaging and exploration of predictors for graft detachment in Descemet membrane endothelial keratoplasty

The aim of this research was to study the utility of intraoperative OCT and investigate predictive factors of graft detachment in DMEK surgery. The first part of this dissertation focuses on the applications of intraoperative OCT in ophthalmic and in particular corneal surgery.
Chapter 2 summarizes the current research and practical applications of intraoperative OCT. The review shows that surgeons are increasingly using intra-operative OCT during surgery and that this has a positive impact on surgical decision making. The benefits of OCT for the surgeon have been extensively studied, but the added value for patients is still unclear.
Chapter 3 reports on our initial clinical experiences with intraoperative OCT during DMEK surgery and how this impacted the surgical protocol. The result of the study showed promising evidence that intraoperative OCT may lead to lower endothelial cell loss and incidence of postoperative complications in Descemet membrane endothelial keratoplasty surgery. This led to the conceptualization of an intraoperative OCT optimized surgical protocol was conceptualized consisting of OCT guidance for determining orientation, unfolding of the graft, and confirming adherence, while refraining from prolonged pressuring the eye. In chapter 4 the results of the ADVISE-trial are presented: a non-inferiority randomized control trial in which we compare the intraoperative OCT-optimized surgical DMEK protocol with the conventional practice; with prolonged overpressure of the eye and without support of intraoperative OCT. The study revealed the intraoperative OCT-optimized surgical DMEK protocol was non-inferior and overpressure can be obviated. In addition, the intraoperative OCT-optimized surgical DMEK protocol led to an improved surgical decision-making in 40% of surgeries and shorter surgical duration.
Chapter 5 describes a case where intra-operative OCT proved crucial in the diagnosis and treatment of an infant with severe corneal edema caused by a rare corneal disease.
In Chapter 6, we describe the development of an image analysis algorithm that can automatically determine the orientation of the graft in the eye using intra-operative OCT. The algorithm can automatically segment the graft and determine the top and bottom of the graft, eliminating the current time-consuming manual review of OCT images by surgeons.
In the second part of the dissertation data-driven methods and advanced analytical methods are used to decode the causes of graft detachment following endothelial keratoplasty, a frequent and burdensome complication.

In Chapter 7, machine learning methods are used to determine the relative contribution of patient and donor characteristics and surgical factors on graft detachment using data from the Dutch Organ Transplant Registry and the surgical protocols of corneal surgeons nationwide. In chapter 8 surgical videos were analyzed to assess the contribution of surgical handling and tissue manipulation on graft detachment.
In chapter 9 we present a predictive model using optical coherence tomography scans taken one day after surgery to identify patients who will develop a graft detachment in the following days. The model was able to accurately identify which patients will develop a detachment and how large the area will be aiding clinical decision-making and tailored patient-care.