摘要Background Endoscopic transsphenoidal approach is a minimally invasive surgical technique for the removal of sellar and parasellar lesions, which has been progressively accepted by neurosurgeons. However, frustration is often expressed by neurosurgeons when first attempting endoscopic endonasal pituitary surgery. To overcome the learning curve from microscope to endoscope in a smooth way, a new human nasal model has been developed. The present study assessed this new model of the human paranasal sinuses for endonasal surgery training, particularly for endonasal pituitary surgery training.Methods The procedure for endonasal trenssphenoidal endoscopy was performed using this model. Three approaches were used to observe the endonasal structures and sphenoidal sinus: paraseptal; middle turbinectomy; and middle meatal. Attempts were made to identify anatomical landmarks in the nasal cavity and sphenoidal sinus. Model landmarks were compared with those in a cadaver and a real patient.Results This model precisely reproduced nasal bone structure. Compared with cadavers and living bodies, intranasal structures displayed very good color and texture, providing a close facsimile of the operative environment, and good morphology, with similar hardness and tactile feel on resection. All intranasal anatomical landmarks were easily identified,including choanae, inferior, middle and superior turbinates, and even the natural ostium of the sphenoidal sinus.Conclusion This human nasal model is very useful for training neurosurgeons in endoscopic endonasal transsphenoidal pituitary surgery, but typical anatomical landmarks in the posterior wall of the sphenoidal sinus in this model should be improved.