Design and evaluation of a new bioelectrical impedance sensor for micro-surgery: application to retinal vein cannulation.

PURPOSE: Nowadays, millions of people suffer from retinal vein occlusion, a blind-making eye disease. No curative treatment currently exists for this vascular disorder. However, a promising treatment consists in injecting a thrombolytic drug directly inside the affected retinal vessel. Successfully puncturing miniature vessels with diameters between 50 and 400 [Formula: see text] remains a real challenge, amongst others due to human hand tremor, poor visualisation and depth perception. As a consequence, there is a significant risk of double-puncturing the targeted vessel. Sub-surfacic injection of thrombolytic agent could potentially lead to severe retinal damage. METHODS: A new bio-impedance sensor has been developed to visually display the instant of vessel puncture. The physical working principle of the sensor has been analysed, and a representative electrical model has been derived. Based on this model, the main design parameters were derived to maximise the sensor sensitivity. A detailed characterisation and experimental validation of this concept were conducted. RESULTS: Stable, repeatable and robust impedance measurements were obtained. In an experimental campaign, 35 puncture attempts on ex vivo pig eyes vessels were conducted. A confusion matrix shows a detection accuracy of 80% if there is a puncture, a double puncture or no puncture. The 20% of inaccuracy most probably comes from the limitations of the employed eye model and the experimental conditions. CONCLUSIONS: The developed bio-impedance sensor has shown great promise to help in avoiding double punctures when cannulating retinal veins. Compared to other puncture detection methods, the proposed sensor is simple and therefore potentially more affordable. Future research will include validation in an in vivo situation involving vitreoretinal surgeons.

Robot-assisted retinal vein cannulation in an in vivo porcine retinal vein occlusion model.

PURPOSE: To evaluate the feasibility of robot-assisted retinal vein cannulation for retinal vein occlusion. METHODS: Prospective experimental study performed in in vivo porcine eyes. A standard three port pars plana vitrectomy was followed by laser-induced branch retinal vein occlusion. Consequently, a retinal vein cannulation with the help of a surgical robot and a microneedle was performed. Complete success was defined as a stable intravenous position of the needle tip confirmed by blood washout for at least 3 min. Secondary outcomes were the occurrence of intra-operative complications and technical failures. RESULTS: Cannulation was successful in 15 of 18 eyes with a complete success rate (duration of infusion of more than 3 min) of 73% after exclusion of two eyes from analysis due to failure in establishing a blood clot. There were no technical failures regarding the robotic device. The intravessel injections of ocriplasmin in two of two eyes led to a clot dissolution. In a subset of five eyes, a second cannulation attempt at the border of the optic disc resulted in a stable intravessel position and infusion during 362 (±138) seconds. CONCLUSION: Robot-assisted retinal vein cannulation with prolonged infusion time is technically feasible. Human experiments are required to analyse the clinical benefit of this new therapy.