The object of this retrospective study was to describe a series of patients with petrous bone cholesteatomas, paying particular attention to classification, diagnosis, surgical strategy, results, complications and recurrences. Furthermore, the study was designed to evaluate the impact of imaging techniques on an early diagnosis. Topographically, the petrous bone cholesteatomas of the present series were grouped using Sanna's classification and different surgical approaches were used. High resolution CT and/or MRI were used to follow-up the patients. The case notes of 52 patients with petrous bone cholesteatomas who were referred to our hospital for surgery between 1987 and 2003 were reviewed postoperatively. There were 45 primary cases and 7 recurrences. The facial nerve had been infiltrated and compressed by the cholesteatoma in 18 patients. Fourteen were managed with cable grafts using sural nerve or great auricular nerves. About 26 patients with preoperative grade I confirmed their normal facial function in 23 cases. In the other ten patients, the preoperative facial paralysis was due to compression by the cholesteatoma and its removal provided partial recovery of facial function in four patients. Our study compared two observation periods (1987-1996 and 1997-2003) when the diffusion and the availability of imaging techniques in our national health system had considerably increased. Two important factors emerged: firstly, the number of less extensive surgical approaches was higher in the more recent observation period, proving that cholesteatomas smaller in size had been diagnosed. Secondly, preoperative facial paralysis was less frequent in the same period-falling to 25% of cases of total facial paralysis from the 45.8% of the earlier period-practically half as much. The partial paralyses instead increased slightly, demonstrating that otologists have become more sensitive to and pay more attention to this symptom.