High resolution T2-weighted magnetic resonance (MR) imaging has been proposed as a rapid, inexpensive means of investigating patients with sensorineural deafness, particularly to exclude vestibular schwannomas. Whether the accepted 'gold standard' of contrast-enhanced T1-weighted images can be omitted, however, remains controversial. Over a 22-month period the use of axial turbo-spin echo T2-weighted images (T2W) were prospectively compared with contrast-enhanced T1-weighted spin echo scans in the evaluation of 513 patients presenting with audiovestibular symptoms. A 2-D T2W turbo spin echo (TSE) sequence with 3 mm slices was used in 340 patients while a 3-D sequence with overlapping 1 mm slices was used in 173 patients. The T2-weighted image findings were documented and subsequently compared with contrast-enhanced images. With the 2-D sequence 24 patients (25 lesions) had internal auditory meatus (IAM)/cerebello-pontine angle (CPA) masses identified by contrast-enhanced T1-weighted images, all of which were seen on the T2-weighted TSE sequence; there was one false positive 'mass' on the T2-weighted scans and one false negative case of IAM dural enhancement on T1-weighted imaging; six were considered normal initially on the T2-weighted images although three were subtly abnormal in retrospect. With the 3-D sequence three acoustic neuromas were all identified correctly with no false positive and only one false negative result (labyrinthitis). The 2-D and 3-D images were judged technically inadequate for clinical assessment in 15 and nine per cent respectively. We conclude that mass lesions of the IAM/CPA can be reliably identified on T2W TSE imaging but labyrinthine lesions may be missed without contrast enhancement. This is of particular importance in planning the management of neurofibromatosis type 2. Non-neoplastic disorders of the inner ear are also likely to be missed.