Approximately 16.4 million people are affected by retinal vein occlusion (RVO) resulting from hypercoagulability, low blood flow or thrombosis in the central or the branched retinal veins. Most common current treatments for RVO aim to limit the damage. In recent years, an experimental procedure, retinal vein cannulation (RVC) has been studied in animal models as well as human eye models. RVC is a procedure for targeted delivery of a therapeutic agent into the occluded retinal vein for dissolving the thrombi. Although effective treatment has been demonstrated via RVC, performing this procedure manually still remains at the limits of human skills. RVC requires to precisely insert a thin cannula into a delicate thin retinal vein, and to maintain it inside the vein throughout the infusion. The needle-vein interaction forces are too small to sense even by an expert surgeon. In this work, we present an evaluation study of a handheld robotic assistant with a force-sensing microneedle for RVC. The system actively cancels hand tremor, detects venous puncture based on detected tool-tissue forces, and stabilizes the needle after venous puncture for reduced trauma and prolonged infusion. Experiments are performed cannulating the vasculature in fertilized chicken eggs. Results show 100% success in venous puncture detection and significantly reduced cannula position drift via the stabilization aid of the robotic system.