Journal of Medical Imaging and Radiation Oncology 59 (2015) 474–479

MEDIC AL I MAG I N G —R A DI AT I O N O NCOLOGY—ORIGINAL ARTICLE

When is an acoustic neuroma not an acoustic neuroma? Pitfalls for radiosurgeons Daniel E Roos,1,2 Sandy G Patel,3 Andrew E Potter1 and Andrew C Zacest4,5 1 2 3 4 5

Department of Radiation Oncology, Royal Adelaide Hospital, Adelaide, South Australia, Australia School of Medicine, University of Adelaide, Adelaide, South Australia, Australia Department of Radiology, Royal Adelaide Hospital, Adelaide, South Australia, Australia Department of Neurosurgery, Royal Adelaide Hospital, Adelaide, South Australia, Australia School of Surgery, University of Adelaide, Adelaide, South Australia, Australia

D Roos BSc (Hons), MBBS, MD, FRANZCR; S Patel MBBS, FRANZCR; A Potter MBBS, BMedSc, FRANZCR; A Zacest MBBS, MS, FRACS, FFPMANZCA. Correspondence Prof. Daniel E Roos, Department of Radiation Oncology, Royal Adelaide Hospital, North Terrace, Adelaide, SA 5000, Australia. Email: [email protected] Conflict of interest: Nil. Submitted 16 November 2014; accepted 27 April 2015. doi:10.1111/1754-9485.12328

Abstract Introduction: Because acoustic neuroma (AN), also termed vestibular schwannoma, constitutes by far the commonest intracranial schwannoma and cerebello-pontine angle (CPA) tumour, there is a risk of overlooking rarer alternative diagnoses with similar clinical and/or radiological features. The purpose of this article is to highlight to radiosurgeons the potentially serious implications of this problem through illustrative case studies. Methods: Our linac stereotactic radiosurgery (SRS) technique has been previously described, with stereotactic headring fixation and treatment delivered via cones or micro-multileaf collimators using multiple arcs or static beams. Results: Between November 1993 and October 2014, we treated 132 patients referred with a clinical diagnosis of AN, the vast majority with 12 Gy marginal dose. Three of these (2.3%), evident either at the time of treatment (2) or subsequently (1), had features instead consistent with cochlear schwannoma, facial schwannoma and meningioma, respectively. Each warranted significant modification to standard AN outlining and fields. The meningioma progressed due to geographic miss. One other patient with recurrent facial schwannoma (not yet needing SRS) was also referred with an incorrect diagnosis of AN. Conclusion: When rare variants of common medical problems are not identified before referral, there is a risk that ‘blinkering’ can lead to misdiagnosis and suboptimal treatment. Radiosurgeons need to be particularly mindful of this issue with AN, which can mimic several other tumours occurring in the CPA region, albeit with different patterns of spread. Optimal imaging, high-quality radiology reporting and neuroradiology input at the time of SRS planning within the setting of a specialised multidisciplinary team are highly desirable. Key words: acoustic neuroma; cochlear schwannoma; facial schwannoma; meningioma; stereotactic radiosurgery.

Introduction Acoustic neuroma (AN) is a benign tumour of the eighth cranial nerve (VIII). Also termed vestibular schwannoma, or more accurately vestibulocochlear schwannoma, it arises from the nerve sheath of usually the vestibular division, rarely from the cochlear division (so-called ‘cochlear schwannoma’). Nonvestibulocochlear schwannomas account for less than

474

10% of intracranial schwannomas. Additionally, AN constitutes 7–8% of all primary intracranial tumours and 75–90% of cerebello-pontine angle (CPA) tumours. Although they often demonstrate the classical ‘ice-cream cone’ or pear shape, there are many variations ranging from barely perceptible intra-canalicular lesions without expansion of the internal auditory meatus (IAM), purely extra-canalicular within the CPA (20%), through to massive tumours compressing and distorting the

© 2015 The Royal Australian and New Zealand College of Radiologists

bs_bs_banner

Acoustic neuroma radiosurgery pitfalls

brainstem. They typically present with the triad of unilateral hearing loss, tinnitus and disequilibrium, but these symptoms and others including VII & V neuropathies do not necessarily correlate well with tumour size or shape.1,2 Management options comprise observation, microsurgery, stereotactic radiosurgery (SRS, 1–5 fractions), stereotactic radiotherapy (>5 fractions) or surgical debulking combined with SRS for the residuum. Importantly, the dominance of AN in tumours of the CPA and adjacent petrous temporal region may divert the attention of clinicians from alternative rarer diagnoses, with potentially serious therapeutic consequences. The differential is wide and includes benign tumours (meningioma, epidermoid cyst, lipoma, facial or trigeminal schwannoma, paraganglioma, primary cholesteatoma and dermoid), vascular lesions (haemangioma and dural arteriovenous fistula) and malignant tumours (metastasis including carcinomatous meningeal or retrograde perineural infiltration, lymphoma and primary melanoma). Symptoms from these lesions may overlap with those of AN. This issue is of particular relevance to radiation oncologists and neurosurgeons who perform radiosurgery. However, these clinicians may be misled by a putative diagnosis of AN from referring ENT specialists or other neurosurgeons who, in turn, may be influenced by an incorrect call from the reporting radiologist. General radiologists will very rarely encounter the other lesions, and even neuroradiology subspecialists will see them infrequently relative to AN. Of course, radiologists are also inevitably influenced by the clinical information (including provisional diagnosis) with which they are provided. In this fashion, a misdiagnosis can be transmitted down the referral chain due to inadvertent ‘blinkering’. Because biopsy is rarely performed in the setting of de novo SRS for benign lesions, such errors may not be detected before treatment. Even though the SRS doses for several of these entities are similar, their pattern of spread can be quite different, and non-recognition of this can lead to inadequate tumour outlining, geographic miss and hence treatment failure. The purpose of this article is to highlight the above issues by means of illustrative case studies.

Materials and methods Our linac radiosurgery unit was established in late 1993, initially utilising a Fischer-Leibinger system (Freiberg, Germany) with circular collimators, replaced in 2009 with a BrainLAB system (Munich, Germany) using either circular collimators or 5 mm (most recently 2.5 mm) micro-multileaf collimators. The planning process for AN has been described previously.3 It utilises stereotactic headring fixation, planning post-contrast CT with 1 mm/ 1 mm resolution T1W gadolinium enhanced 3T MRI fusion (since 2009), typically six non-coplanar arcs or

© 2015 The Royal Australian and New Zealand College of Radiologists

multiple static beams and a marginal dose of 12 Gy usually prescribed to the 80% isodose envelope encompassing the gross tumour volume without expansion. Generally, the maximum diameter considered suitable for SRS is 3 cm. Patients are initially assessed, and then followed-up in a weekly SRS clinic by one neurosurgeon and two radiation oncologists. They are referred by ENT surgeons or neurosurgeons with pre-existing imaging, usually from private radiology providers. The majority of cases are imaged with a high resolution volume T2 and a postcontrast volume T1 sequence. There is, however, considerable variation in sequence selection between institutions. Verbal consent was obtained from each of the patients included in this report, approval for which was granted by the Royal Adelaide Hospital’s Human Research Ethics Committee. Relevant data were prospectively recorded at each clinic attendance, supplemented where necessary by retrospective review of casenotes, imaging and radiology reports.

Results Between November 1993 and October 2014, we treated 132 patients referred with a clinical diagnosis of AN. Three of these (2.3%) were found to have atypical clinical and/or radiological features consistent with alternative diagnoses either during planning for SRS (2) or subsequently (1). They each received 12 Gy to the 80% isodose envelope. One other patient has not required treatment thus far (and so has not been included in the above statistic), but also turned out to have a different diagnosis. The relevant features of these four cases are summarised in Table 1. Further details regarding their management and outcomes are as follows: Case 1. We previously described this case that demonstrated lateral tumour extension to the geniculate ganglion at presentation. However, this extension had not been detected by the reporting radiologist on the pretreatment non-contrast MRI in 2006 and, unfortunately, remained occult on the SRS planning CT due to technical failure of contrast administration. This was prior to our routine use of contrast MRI fusion (Fig. 1). As a consequence, only the CPA and IAM components were outlined and treated, resulting in progression of the geniculate ganglion component.4 What became increasingly apparent on follow-up was that this lesion is almost certainly a CPA meningioma demonstrating extension laterally along both VII and VIII (hence into the geniculate ganglion/middle cranial fossa and vestibule/cochlea, respectively) and inferiorly involving the jugular foramen. Interestingly in retrospect, we did raise the possibility of meningioma in correspondence at initial referral because of the unusually broad dural base of the lesion around the IAM.

475

DE Roos et al.

Table 1. Patient and tumour characteristics

Age Sex Symptoms Dimensions (mm) Atypical features

Probable true diagnosis

Case 1

Case 2

Case 3

75 Female Deaf right ear, tinnitus, disequilibrium 28 × 26 × 23† Broad-based dural attachment. Infiltration into cochlea, vestibule, geniculate ganglion and jugular foramen Meningioma

55 Female Deaf right ear, disequilibrium, right facial twitching 19 × 10 × 8 Extension into cochlea

59 Male Deaf left ear, disequilibrium 20 × 13 × 10 Extension along VII to geniculate ganglion

Cochlear schwannoma

Facial schwannoma

Case 4 60 Female Deaf right ear 22 × 18 × 14 Involvement of mastoid, cochlea, vestibule and internal auditory meatus. Post-op changes in the mastoidectomy bed Debulked facial schwannoma (pathologically confirmed)

†Including lateral extension (not appreciated at the time of radiosurgery – see text).

On latest review at eight years, there had been both in-field and out-of-field progression with overall dimensions of 34 × 29 × 26 mm (Fig. 2). Consistent with this pattern of progression, the patient has developed ipsilateral V, VII and partial X neuropathies together with worsening disequilibrium. As a result of this landmark case in our experience, we have subsequently taken particular care to check for enhancement/extension lateral or inferior to the IAM and to consider alternative diagnoses to AN, as the remaining three cases illustrate. Case 2. We initially observed this patient from 2006 when she was referred with a right intra-canalicular AN, having already lost ipsilateral hearing. The initial non-

Fig. 1. Case 1: Non-contrast T1W MRI pre-radiosurgery showing a soft tissue mass at the geniculate ganglion (arrow), not reported (or appreciated) at the time of treatment.

476

contrast MRI did not detect cochlear involvement, but this was evident on follow-up contrast scans. This pattern of spread would be consistent with a cochlear schwannoma. Hence, the cochlea was included in the field when she proceeded to SRS in 2009 due to tumour growth (Fig. 3). On latest follow-up at five years, the tumour had shrunk 1–2 mm in all three dimensions, her balance had improved and the pretreatment ipsilateral facial twitching had resolved. Case 3. This patient’s lesion had been observed by his ENT surgeon since 2010. Contrast MRI at that time and non-contrast MRIs yearly thereafter were reported to

Fig. 2. Case 1: Contrast T1W MRI eight years after radiosurgery showing in-field and out-of-field progression involving the cerebello-pontine angle, cochlea, vestibule and geniculate ganglion. There was also extension into the jugular foramen inferiorly (not shown). This pattern of spread and the broad dural base strongly suggest meningioma.

© 2015 The Royal Australian and New Zealand College of Radiologists

Acoustic neuroma radiosurgery pitfalls

Fig. 3. Case 2: Contrast T1W MRI pre-radiosurgery showing extension into the cochlea (arrow), consistent with a cochlear schwannoma.

show a left intra-canalicular AN, gradually extending medially into the CPA. Following eventual referral for SRS, contrast planning CT and MRI in August 2014 clearly showed extension along the labyrinthine segment of VII to the geniculate ganglion, strongly suggesting that this lesion was a facial schwannoma. Unlike case 1 (meningioma), there was no involvement of the cochlea or vestibule. In retrospect, the component extending along VII was, in fact, visible on each of the prior private MRIs but had been missed successively by four different reporting radiologists. The target volume was designed to encompass all regions of enhancement (Fig. 4). At follow-up one month after SRS, he reported ipsilateral facial weakness developing three days post-treatment, gradually improving thereafter. This was most likely due to transient VII oedema from the radiation. Case 4. This patient was referred by a neurosurgeon in August 2013 with a diagnosis of residual right intracanalicular AN, operated upon by an ENT surgeon in 2005, with gradual progression subsequently. She also had an incidental 1.5 cm left posterior fossa meningioma (stable). However, on careful review of her history and imaging, it emerged that the surgery had entailed transmastoid debulking of a middle ear facial schwannoma (probably arising from the tympanic segment), extending into the external auditory meatus, mastoid antrum, cochlea, vestibule and fundus of the IAM medially. She had initially been referred to the neurosurgeon postoperatively in order to monitor the incidental meningioma. Along the way, the true nature of the right IAM lesion had become confused with AN.

© 2015 The Royal Australian and New Zealand College of Radiologists

Fig. 4. Case 3: Contrast radiosurgery planning T1W MRI showing extension along the labyrinthine segment and geniculate ganglion of VII, with overlaid 12 Gy, 6 Gy and 2 Gy isodose curves. Unlike case 1, there is no enhancement within the cochlea or vestibule. This is typical of a facial schwannoma.

This was reflected in interval MRI reports from several different radiologists who commented on the IAM component only, misinterpreting the more lateral enhancement as postoperative change and surgical packing (Fig. 5). In any event, with slow growth and intact VII function, continued observation was recommended at initial referral to the SRS clinic, and again at latest review one year later. However, all components of the residuum will need to be included if SRS becomes necessary in due course.

Discussion The dominance of AN in the CPA region can lead to potentially serious ‘blinkering’ to alternative diagnoses. As the above four case studies demonstrate, one cannot assume that reporting radiologists and referring ENT surgeons or neurosurgeons will always consider or detect other tumour types that may mimic AN clinically and radiologically. This is particularly the case if imaging contrast is not used, as can happen in the private system for logistics and cost reasons. It must be acknowledged that, in the absence of pathological confirmation for cases 1–3, the putative alternative diagnoses are based on their highly atypical pattern of spread demonstrated on MRI and indolent behaviour that excludes aggressive differentials. Misdiagnosis appears to be an uncommon event, comprising only 2.3% of our treated cases as far as we can tell. We

477

DE Roos et al.

Fig. 5. Case 4: Contrast T1W MRI at time of referral for a radiosurgery opinion showing postoperative changes in the right mastoidectomy bed (arrow right) and residual enhancement within the vestibule (arrow up), cochlea (arrow down) and internal auditory meatus. This was a pathologically confirmed facial schwannoma with direct extension into the above structures, probably arising from the tympanic segment, and debulked eight years previously.

think it unlikely that there have been others because of our rigorous long-term follow-up policy,3 and for more recently treated patients, our awareness of the issue due to the experience with case 1. Precise knowledge of the complex neuroanatomy of the CPA and petrous temporal bone explains the site of origin and extension of the various tumours in the region and can help distinguish between them. Emerging laterally from the pontomedullary junction, the paired VII and VIII are invested initially by glial nerve sheaths that soon give way to Schwann cell sheaths. Schwannomas tend to arise in this transition zone between central and peripheral myelin. In the case of VIII, the transition occurs further from the brainstem than for any other cranial nerve, typically within the IAM, explaining why AN usually originates intracanalicularly. However, although classically enlarging the IAM (the ‘cone’), further growth tends to be along the path of least resistance medially, giving rise to the CPA component (the ‘ice-cream’). They typically do not spread laterally into the labyrinth, but may do so into the vestibule, cochlea (Fig. 3) or both.5 Inferior extension into the jugular foramen should trigger consideration of an alternative diagnosis, most commonly meningioma (case 1). The course of VII in the facial canal within the petrous temporal bone is notably the longest intra-osseous path of any nerve in the body, explaining its vulnerability to local inflammatory processes, including iatrogenic (case

478

3). From the fundus (apex) of the IAM, it passes as the labyrinthine segment antero-laterally immediately posterior to the cochlea. At the geniculate ganglion, it gives off the greater superficial petrosal nerve (secreto-motor to the lacrimal land) and two other petrosal nerves into the floor of the middle cranial fossa. This accounts for the typical enhancing mass developing at this site if the geniculate ganglion is involved with a meningioma (Figs 1,2) or facial schwannoma (Fig. 4). Thereafter, VII deviates posteriorly into the middle ear below the lateral semicircular canal as the tympanic segment, originating from which, a facial schwannoma can infiltrate multiple adjacent structures (Fig. 5). Finally, it turns downwards as the mastoid segment to exit the skull through the stylomastoid foramen. A facial schwannoma originating in the CPA or IAM will not usually involve the vestibulocochlear apparatus (Fig. 4). In contrast, a CPA or IAM meningioma arising from either VII or VIII (and it may be impossible to tell which) can spread perineurally along both nerves and so may infiltrate all of these structures (Fig. 2). Because each of these lesions occupying similar ‘territory’ may compress and/or infiltrate the same base of skull structures including cranial nerves, they may therefore be indistinguishable clinically. For example, facial schwannomas may present with vestibulocochlear deficits without VII dysfunction until large (cases 3 and 4). Conversely, vestibulocochlear schwannomas can cause VII neuropathy even when small (case 2). Similarly, the radiological features overlap – both facial schwannoma and AN may undergo cystic degeneration, but solid lesions may be indistinguishable from meningiomas. Facial schwannomas of the CPA or IAM can appear identical to AN if there is no extension into the labyrinthine segment of VII, the so-called ‘labyrinthine tail’, typically associated with expansion of that part of the facial canal.6 It should be noted that failure to distinguish AN from other benign lesions treated with SRS may have little consequence with respect to radiation dose. Limited retrospective data suggest that non-vestibular schwannomas are appropriately treated with similar marginal doses to AN,1 and the range of doses reported for CPA meningiomas also typically overlaps with that of AN.7,8 Instead, the real issue is geographical miss from failure to recognise tumour extension atypical for AN (case 1), and accordingly, particular attention to the regions immediately lateral and inferior to the main tumour mass on MRI is critical. In summary, we present four case studies illustrating the importance of circumspection when radiosurgeons are referred patients with a putative diagnosis of AN. The overwhelming dominance of this tumour of the CPA/IAM region can lead to a ‘blinkered’ view with potentially serious clinical implications. The management of AN with SRS is a very complex and highly specialised area, best managed by a multidisciplinary team including

© 2015 The Royal Australian and New Zealand College of Radiologists

Acoustic neuroma radiosurgery pitfalls

experienced neurosurgeons, radiation oncologists and neuroradiologists. High-resolution MRI sequences with T1, T2 and post-contrast fat saturated T1 (spin echo or volume acquisition), high-quality radiology reporting and neuroradiology input at the time of SRS planning are highly desirable. Furthermore, close assessment of the temporal bone and posterior cisterns is recommended at each follow-up to detect unexpected behaviour early.

4.

5.

6.

References 1. Hasegawa T. Stereotactic radiosurgery for non-vestibular schwannomas. Neurosurg Clin N Am 2013; 24: 531–42. 2. Silk PS, Lane JI, Driscoll CL. Surgical approaches to vestibular schwannoma: what the radiologist needs to know. Radiographics 2009; 29: 1955–70. 3. Roos DE, Potter AE, Brophy BP. Stereotactic radiosurgery for acoustic neuromas: what happens

© 2015 The Royal Australian and New Zealand College of Radiologists

7.

8.

long term? Int J Radiat Oncol Biol Phys 2012; 82: 1352–5. Roos DE, Brophy BP, Taylor J. Lessons from a 17-year radiosurgery experience at the Royal Adelaide Hospital. Int J Radiat Oncol Biol Phys 2012; 82: 102–6. Salzman KL, Davidson HC, Harnsberger HR et al. Dumbbell schwannomas of the internal auditory canal. Am J Neuroradiol 2001; 22: 1368–76. Wiggins RH, Harnsberger HR, Salzman KL, Shelton C, Kertesz TR, Glastonbury CM. The many faces of facial nerve schwannoma. Am J Neuroradiol 2006; 27: 694–9. Park S-H, Kano H, Niranjan A, Flickinger JC, Lunsford LD. Stereotactic radiosurgery for cerebellopontine angle meningiomas. J Neurosurg 2014; 120: 708–15. Kim Y-H, Kim DG, Han JH et al. Radiosurgery for para-IAC meningiomas: the effect of radiation dose to the cochlea on hearing outcome. Int J Radiat Oncol Biol Phys 2012; 84: 675–80.

479

When is an acoustic neuroma not an acoustic neuroma? Pitfalls for radiosurgeons.

Because acoustic neuroma (AN), also termed vestibular schwannoma, constitutes by far the commonest intracranial schwannoma and cerebello-pontine angle...
661KB Sizes 69 Downloads 12 Views