BACKGROUND: One criticism of current video systems for endoscopic surgery is that two-dimensional (2D) images lack depth perception and may impair surgical dissection. To objectively measure the efficacy of 3D endoscopy, we designed a training model with specific tasks to show potential differences between 2D and 3D endoscopy. Its clinical value was then evaluated during endoscopic sinus and skull base surgical cases. METHODS: Fifteen subjects were grouped according to endoscopic experience: novices and nonnovices. A training model was constructed to include five tasks: incision manipulation; ring transfer; nerve hook; distance estimation, visual only; and distance estimation, visual and tactile. Each participant was assessed with both a standard 2D endoscope and a 3D endoscope. The clinical value of a 3D endoscope (Visionsense, Ltd., Petach Tikva, Israel) was then examined in four endoscopic sinus cases and four skull base cases. RESULTS: Of the subjects, six (40%) were novices. Overall, the errors committed during any one task were not significantly different between systems. Novices trended toward more success during the nerve hook task using the 3D system. With size cueing versus visualization alone, distance estimation was significantly more accurate. Novices tended to prefer the 3D system and experienced surgeons disliked the initial learning curve. Advantages were particularly noticed during skull base surgery; subjectively improved depth perception was beneficial during vascular dissection. CONCLUSION: Three-dimensional endoscopy may improve depth perception and performance for novices. The 3D endoscope is a safe and feasible tool for endoscopic sinus and skull base surgery; it is promising for improving microneurosurgical dissection precision transnasally.