Optimal path planning and re-planning are essential requirements for efficient robotic and laparoscopy surgical education as they guide instruments through the operating field. Ensuring collision free paths and training under uncertainty, with limited dependence on an expert surgeon's assistance, are key aspects to be considered. In this paper, a hybrid path planning method is introduced to assist in surgical training that combines offline and online approaches for trajectory generation. The proposed offline-online path planner integrates features of offline approaches that guarantee safe navigation with real-time re-planning capability of online planners. It reduces memory storage requirements of the trajectories for online path planning, by more effectively saving them as segments, catering to the portability of the surgical systems. The efficacy of the proposed approach is demonstrated in a simulated training environment for hand-eye coordination tasks on the Computer Assisted Surgical Trainer (CAST) platform.