- Email: [email protected]
False-positive magnetic resonance imaging of small internal auditory canal tumors: A clinical, radiologic, and pathologic correlation study
False-positive magnetic resonance imaging of small internal auditory canal tumors: A clinical, radiologic, and pathologic correlation study MAJ MOISES A. ARRIAGA, MD, LTC DAVID CARRIER, MD, and COL GLENN D. HOUSTON, ODS,
Lackland AFB, Texas Magnetic resonance imaging with gadolinium facilitates the early diagnosis of internal auditory canal tumors at a small enough stage to permit increasing application of hearing preservation surgical techniques. Surgeons report successful removal of tumors as small as 3 mm, which are diagnosed with enhanced magnetic resonance imaging. A retrospective study was performed to determine the risk of false-positive "tumor" diagnosis with enhanced magnetic resonance imaging. We reviewed the imaging records, office notes, and surgical records of 112 consecutive "tumors" involving the internal auditory canal treated by the Wilford Hall USAFMedical Center Neurotology Service between July 1991 and July 1994. Two categories of false-positive magnetic resonance imaging were identified: [I] surgically confirmed absence of internal auditory canal neoplasm and (2) spontaneous resolution of the internal auditory canal lesions on subsequent, enhanced magnetic resonance images. Overall, eight false-positive scans were identified. Three were surgically confirmed as false-positive, and five resolved on subsequent imaging studies. All cases were smaller than 6 mm and involved the distal internal auditory canal (fundus]. The surgically confirmed cases were approached through a middle fossa technique with successful hearing preservation. The overall rate of surgical false-positive results was 3.5% [3 cases in 86 surgeries]. However, the overall falsepositive rate for intracanalicular "tumors" was 32% (8 cases in 25 intracanalicular lesions]. Although hearing preservation is more likely in small lesions, the surgeon must consider the possibility that an internal auditory canal lesion smaller than 6 mm may actually represent a nonneoplastic process. Enhancing lesions limited to the internal auditory canal fundus may be treated by reimaging the patient in 6 months after the first image rather than by prompt surgical exploration. [OTOLARYNGOLHEADNECKSURG1995;113:61-70,]
Ill M a g n e t i c resonance imaging (MRI) is currently the most sensitive imaging technique for identifying internal auditory canal (IAC) and cerebellopontine angle neoplasms. Early identification of vestibular schwannomas enhances preservation of hearing and facial function after microsurgical removal. Gadolinium-enhanced, Tl-weighted MRI permits accu-
From the Departments of Otolaryngology-Head and Neck Surgery (Dr. Arriaga), Radiology (Dr. Carrier), and Pathology (Dr. Houston), Wilford Hall Medical Center. The opinions expressed in this article are those of the author and do not necessarily represent those of the Air Force or the Department of Defense or other departments of the United States ,government. Received for publication Sept. 18, 1994; accepted Feb. 10, 1995. Reprint requests: Moises Arriaga, MD, Director Otology/Neurotology, Department of Otolaryngology 2200 Berquist Dr., Suite 1, Wilford Hall Medical Center, PSSO, Lackland AFB, TX 78236. 23/1/64150
rate diagnosis of tumors as small as 2 mm. 1 With specific imaging sequences, most nonneoplastic entities may be excluded, such as cholesterol granuloma, epidermoid, adjacent bone marrow in the porous acousticus, and lipomas. Nonetheless, there remains a small subset of patients with small enhancing lesions in the IAC identified falsely as neoplasms. Recent experience with such cases prompted a review of our patients in an effort to identify imaging features of these enhancing lesions that would allow a more accurate preoperative diagnosis. METHODS AND MATERIAL
A retrospective study was performed to determine the risk of false-positive "tumor" diagnosis with enhanced MRI. We reviewed the imaging records, office notes, and surgical records of 112 consecutive IAC "tumor" patients treated by the Wilford Hall USAF Medical Center Neurotology Service be61
62
ARRIAGA et al.
Otolaryngology Head and Neck Surgery
July t995
Fig. I. Case t. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the right IAC (arrow). C, D and E, Postgadolinium, T~-weighted images at t0, 16, and 33 months of follow-up show essentially no c h a n g e (arrows). The original interpretation of E questioned the possibility of slight increase in size from D; however, no such c h a n g e was evident in reviewing the entire sequence.
tween July 1991 and July 1994. The lesions were entirely intracanalicular in 25 patients. A total of 86 surgical resections of tumors involving the IAC were performed during this time, with 15 of those surgeries for completely intracanalicular lesions. The absence of neoplasm in false-positve cases was confirmed either pathologically or by resolution (partial or complete) on follow-up MRI scans.
All patients were imaged on a 1.5-T GE Signa scanner (General Electric, Milwaukee, Wis.) with pregadolinium and postgadolinium, T~-weighted (TR 600, TE 11) images through the IACs. A 3-mm slice thickness with zero gap, FOV 16, 256 x 192 matrix was used, with two excitations on the pregadolinium sequence and four excitations on the postgadolinium sequence. Additionally, fast-spin echo
Otolaryngology Head and Neck Surgery
Volume t13
ARRIAGA et al.
Number t
63
Fig. 1 (cont.] F, Photomicrograph reveals dense hyalinized fibrous connective tissue and calcified debris. (Hematoxylin and eosin stain; original magnification x 100.} G, Photomicrograph of F at a higher power demonstrating dense, hyalinized fibrous connective tissue with cellular and calcified debris. (Hematoxylin and eosin stain; original magnification x 200.]
Table 1. False-positive MRI of the IAC: Clinical, radiologic, and pathologic features Case No.
Age [yr]/sex
Image findings
1
38/F
1/90 Punctate, 2-3 mm, fundus 10/90 No change 6/91 No change 9/92 ? Minimal growth 12/92 Punctate, 2-3 mm, fundus 3/93 No change 8/94 Ill-defined, partially resolved 5/92 Punctate, 2-3 mm, fundus 6/93 No change 8/94 Resolved 2/94 Ill-defined, 4 mm 7/94 Resolved 2/92 Punctate, 2-3 mm, fundus 7/92 ? increase 12/92 ? Decrease 5/93 ? Increase 8/91 Punctate, 3 mm, fundus 10/92 Ill-defined 8/94 Partially resolved fundus 10/90 IDdefined, 3 mm, fundus 4/91 and 9/92 No change 5/93 Resolved 5/94 Globular 6 x 4 mm
2
69/M
3
57/M
4
32/F
5
44/M
6
70/M
7
54/M
8
41/M
T2-weighted (TR 3600, effective T E 100), 3-mm slice thickness, zero gap, FOV 24, 512 × 256, 2 NEX images were obtained through the same area. After fixation with formalin, the surgical specimens were embedded in paraffin, sectioned at multiple levels, and stained with hematoxylin and eosin in preparation of microscopic evaluation. RESULTS
Eight false-positive "tumors" were identified among the 112 lesions treated during the period of
Clinical
Surgery/pathology
Vertigo/hearing Middle fossa (1/93) Hyaline Hearing Contralateral hearing Hearing Vertigo/hearing
Middle fossa (6/93) Neuronal degeneration
Hearing Hearing/tinnitus Hearing
Middle fossa (5/94)
the study. The false-positive lesions were entirely intracanalicular, in the fundus of the IAC, and smaller than 6 mm in maximum dimension. Three of the eight false-positive tumors were confirmed as nonneoplastic by histopathology among the 86 surgical resections for IAC tumors performed during this time period. The other five lesions either resolved completely or evolved into a less tumor-like appearance on subsequent MRI. The remaining 21 patients who did not undergo operations include 4 patients referred elsewhere for care because of mili-
64
ARRIAGA et al.
Otolaryngology H e a d a n d Neck Surgery July 1995
Fig. 2. Case 5. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the left IAC (arrow). C, Postgadolinium, T~-weighted image at 6 months of follow-up shows lesion slightly larger (arrow). D, Postgadolinium, T~-weighted image at 10 months of follow-up shows lesion slightly smaller (arrow), probably because of partial volume averaging. E and F, Postgadolinium, T~-weighted images at 15 months of follow-up show lesion divided between two slices and apparently larger that al I 0-month study (arrows). However, when compared with 6-month study, it is probably unchanged. Varying size due to slight changes of the lesion relative to slice position.
tary eligibility issues, 1 patient treated with stereotactic radiation therapy, 9 elderly patients with stable lesions and severe medical problems prohibiting general anesthesia, 3 patients in whom the
diagnosis is uncertain and repeat imaging is pending, 2 neurofibromatosis-2 patients with stable tumor in an only-hearing ear, and 2 patients with stable residual tumor operated on more than 10 years ago.
Otolaryngology Head and Neck Surgery Volume 113 Number I
ARRIAGA et al:
65
Fig. 2 (cont.). G, Fast-spin echo T2-weighted image also at 15 months of follow-up confirms space-occupying lesion displacing cerebrospinal fluid (arrow). H, Photomicrograph reveals section of peripheral nerve exhibiting mild vacuolar c h a n g e and osseous debris. (Hematoxylin and eosin stain; original magnification x 40.) I, Higher magnification of H exhibiting peripheral nerve fibers with mild vacuolar c h a n g e and foci of calcified debris. (Hematoxylin and eosin stain; original magnification x 200.) J, Photomicrograph of H and I at a higher magnification demonstrating vacuolar c h a n g e in nerve fibers. (Hematoxylin and eosin stain; original magnification × 400.)
The clinical features, radiologic findings, and histopathologic interpretations are summarized in Table 1 for the eight false-positive lesions. The series consisted of two women and six men, with a mean age of 50.6 years. All the patients except one had serial MRI scans at least 4 months apart to identify changing radiographic features. Follow-up times on the patient s not operated On ranged from 6 to 36 months. The false-positive scans demonstrated two patterns of enhancement in the distal IAC (fundus). In one pattern the enhancement was sharply circumscribed either punctate (cases 1, 2, 3, 5, and 6) or globular (case 8). The other pattern was ill-defined, which occurred in two lesions (cases 4 and 7). Except for case 3, all patients exhibited asymmetric sensorineural hearing loss in the ear with the abnormal imaging. Patient 3 had hearing loss and
tinnitus in the contralateral ear. All eight patients admitted unsteadiness to varying degrees; however, patients 1 and 5 described initial episodes of severe vertigo with subsequent chronic unsteadiness (both were surgically confirmed to be false-positive). During follow-up, no patient demonstrated significant progression of clinical symptoms including hearing loss or tinnitus. Interestingly, patient 7 had severe, disabling tinnitus on the side of his enhancing MRI lesion. Although his lesion completely resolved during the 3-year follow-up period, his tinnitus symptoms were unimproved. The timing and rationale for microsurgical removal of the lesions in the three surgically confirmed false-positive cases involved clinical symptoms, radiologic findings, and patient preference. Patients 1 and 5 had strong family histories of malignant brain tumors. Despite reassurance of the benign nature of
66
ARRIAGA et al.
Otolaryngology Head and Neck Surgery July 1995
Fig. 3. Case 8. A and B, Pregadolinium a n d postgadolinium, T~-weighted axial images through the IACs at initial visit show a globular focus of e n h a n c e m e n t in the distal right IAC (arrow). Findings were consistent with vestibular s c h w a n n o m a on original interpretation. C, Intraoperative view of right middle cranial fossa dissection of IAC in this patient. G, Geniculate ganglion; P, o p e n e d g e of dura at porous acousticus. Arrow identifies junction of the labyrinthine segment of the facial nerve with the IAC portion of the nerve.
the lesions, they were both insistent on microsurgical removal when serial imaging demonstrated no regression and "possible slight growth" of their lesions. In retrospect, and comparing each patient's entire sequence of images, the lesions were stable. Figure 1A-E demonstrates the images during a 2-year period before surgery in patient 1, and Fig. 2A-F demonstrates the images during an 18month period preceding surgery in patient 5. Both imaging features and personal factors prompted surgery for patient 8 (Fig. 3A and B). He was a professional musician for whom any additional hearing loss would have had dire career consequences. Moreover, the globular appearance of his MR! enhancement was highly suggestive of schwannoma. Surgical removal in all three false-positive cases was performed through the middle fossa approach. Hearing was retained within 15 dB of preoperative
thresholds for pure tones and speech reception, and speech discrimination was within 10% of preoperative levels. All three patients have retained their preoperative level of function and independence with good vestibular compensation. Intraoperatively, patients 1 and 5 had abnormal, nodular, distal superior vestibular nerves. The histopathologic findings in patients 1 and 5 are most consistent with sequellae of a previous inflammatory or degenerative lesion. Patient 5 demonstrated vacuolar, neural change with fibrosis in the superior vestibular nerve (Fig. 2G, H, and I). Patient 1 had dense hyalinized fibrous tissue removed from the distal superior vestibular nerve (Fig. 1F-I). These findings are consistent with a reactive inflammatory mechanism, although ischemic or hemorrhagic injury could produce similar findings. Both patients had initial severe vertigo followed by chronic un-
Otolaryngology H e a d a n d N e c k Surgery
Volume 1t3
Number t
ARRIAGA et al.
67
Fig. 4. Case 7. A a n d B, Pregadolinium a n d postgadolinium, T~-weighted axial images through the IACs at initial visit show an ill-defined focus of e n h a n c e m e n t in the fundus of the left IAC (arrow). O and D, Postgadolinium, T~-weighed images at 6 a n d 23 months of follow-up are u n c h a n g e d (arrows). E, Postgadolinium, T~-weighted i m a g e at 35 months of follow-up demonstrates resolution of enh a n c e m e n t (arrow).
steadiness and hearing loss before surgery, symptoms which clinically support a cochleovestibular neuritis. Case 8 merits special mention. The radiographic features were distinct from those of the other seven
cases because the focus of enhancement was globular and larger than in the rest of the cases. Intraoperatively, just as with the other cases, there was a distinctly abnormal appearance to the distal superior vestibular nerve. There was an adhesive, hem-
68
Otolaryngology Head and Neck Surgery July 1995
ARRIAGA et al.
Fig. 5. Case 6. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the right IAC (arrow). C, Postgadolinium, T~-weighted image at 14 months of follow-up shows conversion of lesion to an ill-defined, barely perceptible focus of enhancement (arrow).
orrhagic rind around the distal vestibular nerve; however, this area of abnormality was smaller than the 6 × 4 mm preoperative size estimate from the MRI. The distal halves of the superior and inferior vestibular nerves were sent as specimens; however, only a small fragment of nerve was actually processed for histopathology. The histopathology of this specimen was interpreted as a normal nerve. There are three possible explanations. (1) Cochlear schwannoma may have been left in situ. This is unlikely because the entire IAC was fully exposed (Fig. 3C) and the cochlear nerve appeared normal intraoperatively. (2) The pathologic portion of the specimen may have been lost in processing. Very small specimens are often adherent to instruments and containers or even inadvertent suctioning during passage. There is no way to disprove this possibility; however, the specimen did not have the usual gross appearance and texture of a schwannoma. (3) The MRI was simply overly sensitive and
identified arachnoid adhesions as tumor. We favor the third explanation. The lesions that were not operated on (cases 2, 3, 4, 6, and 7) either resolved completely or evolved to an ill-defined, partially resolved appearance on follow-up scans. Cases 4 and 7 began as ill,defined lesions and resolved totally on follow-up images (Fig. 4A-E). In comparison, cases 2 and 6 began as sharply circumscribed, punctate lesions and became ill-defined and partially resolved during the time of follow-up (Figs. 5A-C). Case 3 began as a punctate lesion and nearly completely resolved during the follow-up period. DISCUSSION
Gadolinium-enhanced MRI is highly sensitive for intracanalicular IAC pathology and has become the gold standard for preoperative evaluation of the IAC. ~This technique has nearly eliminated the need for the previous gold standards of computed tomog-
Otolaryngology Head and Neck Surgery Volume 113
Number I
raphy scans after the instillation of subarachnoid air or radiopaque contrast. Because of its very high sensitivity, a negative gadolinium-enhanced MRI essentially eliminates the possibility of an IAC neoplasm. 2 Because of its lower specificity, the converse is not true. Enhancement within the IAC may be caused by a variety of processes other than neoplasm. Our review of the literature identified 12 reported cases of surgically proven nonneoplastic enhancing lesions in the IAC. These include Ramsay-Hunt syndrome, 3vestibulocochlear neuritis, arteriovenous malformation, 4 and arachnoiditis) False-positive diagnosis of an intracanalicular neoplasm is a very undesirable situation given the potential morbidity, patient discomfort, and cost of surgery. Haid and Wolf 6 recently reported that their rate of surgically confirmed false-positive IAC tumor explorations was 2% (8 false-positive findings in 400 surgical explorations). In their cases the operative finding was of arachnoidal adhesions suggesting an arachnoiditis in each case without a pathologic diagnosis. In our three surgically confirmed falsepositive cases, two demonstrated neural inflammatory sequellae, and one had intraoperative evidence of arachnoiditis. Our false-positive rate of tumor surgery for lesions involving the IAC was 3.5% (3 false-positive explorations in 86 surgeries) and 20% for intracanalicular lesions (3 false-positive explorations in 15 surgeries). The most critical finding for clinical decisions is the overall false-positive rate for intracanalicular lesions. In this series 32% of intracanalicular lesions were incorrectly diagnosed as neoplasms (8 falsepositive findings in 25 intracanalicular lesions). If false-positive results are possible, what are the clinical features that can alert the clinician? Von Glass et al. s argued that the absence of vestibular findings should dissuade the clinician from diagnosing a schwannoma. However, two of our surgically confirmed cases had prominent vestibular symptoms before surgery. In one of our cases, the imaging studies were performed for contralateral symptoms (case 3); clinical observation with serial imaging was the prudent course of action. In our experience, small, enhancing, punctate, or ill-defined lesion in the fundus should raise the suspicion of a lesion or process other than a neoplasm. All five of the lesions that we did not operate resolved completely or became ill-defined on follow-up scans. Han et al.'s series 4 of three surgically confirmed false-positive lesions also identified that intense, focal enhancement limited to the lateral
ARRIAGA e t al.
69
IAC was suspicious for a nonneoplastic lesion. Occasionally, cases may be quite dramatic with enhancement of the entire IAC, positive air computed tomography scan, and appropriate audiologic findings and may still represent only arachnoiditis. The globular pattern of enhancement in our case 8 was consistent with schwannoma and similar to the cases reported by Von Glass et al. 5 and Haid and Wolf. 6 Serial imaging offers insight into the biologic behavior of the enhancing lesion rather than the static information from a single MRI. Case 8 was the only lesion without serial imaging; perhaps a second image 6 months later would have indicated resolution. Even serial imaging has its pitfalls. When lesions smaller than the 3-mm slice thickness are examined, the lesion may appear to change size from one scan to the next because of slight changes in the position of the lesion relative to the slice. This may result in division of the lesion between two images or partial volume averaging, either of which may alter the apparent size of the lesion (Fig. 2E-F). When there is an apparent slight change in size compared with that shown in the last study, the clinician and radiologist must compare the current study with all available old scans. For example, case 5 demonstrated a "slight increase in size" on the last scan done before surgery. However, when compared with the original scan in the series no real change was found. The rationale for surgery on a small intracanalicular schwannoma is total tumor removal with hearing and facial nerve preservation. Excellent data exist to support improved hearing preservation 7 and improved facial function preservation 8 for smaller rather than larger lesions. However, it has not been shown that there is a difference in functional outcome if schwannomas are removed at a 2- to 3-ram size rather than a 6- to 8-mm size. In larger tumors we have demonstrated that a "clinical break point" of significantly worse facial nerve outcomes occurs once lesions reach 2.5 cm. 9 Regardless of approach, excellent outcomes have been reported for tumors smaller than 1 cm. Furthermore, because average reported growth rates are less than 2 mm per year, 1° there is little risk to reimaging a small intracanalicular lesion in 6 months and subsequent reimaging annually if there is no significant growth. 1 Surgical approach selection for small IAC lesions is also important. Our lesions involved the fundus of the I A C - a region ideally suited for the middle fossa approach. In one report, the authors used translabyrinthine exposure only to find that there was no
Otolaryngology Head and Neck Surgery July 1995 -
70
ARRIAGA et al.
neoplasm. 1I Middle fossa exposure offers the possibility of hearing preservation with distal IAC exposure. CONCLUSIONS
1. Small enhancing intracanalicular lesions of the IAC fundus may represent nonneoplastic pathology. 2. Most small enhancing intracanalicular lesions of the IAC fundus should be treated by reimaging the patient 6 months after the first image rather than prompt surgical exploration. 3. Comparison studies should be compared with the entire imaging series, not just the previous study, because apparent growth on subsequent MRIs may reflect positioning of the tumor in relation to the slices rather than actual tumor growth. 4. The middle fossa approach offers the possibility of hearing preservation for small, distal IAC lesions, especially if the ultimate diagnosis is of a nonneoplastic pathologic process. REFERENCES
1. Strasnick B, Glasscock ME, Haynes D, McMenomey SO, Minor LB. The natural history of untreated acoustic neuromas: Laryngoscope 1994;104:1115-9.
2. Sidman JD, Carasco VN, Whaley RA, Pillsbury HC. The new gold standard for diagnosing cerebellopontine angle tumors. Arch Otolaryngol Head Neck Surg 1989;115:1244-7. 3. Anderson RE, Laskoff JM. Ramsay Hunt syndrome mimicking intracanalicular acoustic neuroma on contrast-enhanced MR. AJNR Am J Neuroradiol 1990;11:409. 4. Han MH, Jabour BA, Andrews JC, et al. Nonneoplastic enhancing lesions mimicking intracanalicular acoustic neuromas on gadolinium-enhanced MR images. Radiology 1991; 179:795-6. 5. Von Glass W, Haid CT, Cidlinsky K, et al. False-positive MR imaging in the diagnosis of acoustic neurinomas. OTOLARYNGOLHEADNECKSURG1991;104:863-7. 6. Haid CT, Wolf SR. False positive MRI findings of the eighth cranial nerve simulating a tumour lesion [Abstract]. Presented at the Fourth International Symposium and Workshops on Inner Ear Medicine and Surgery, Snowmass-Aspen, Colo., July 17, 1994. 7. Jackler RK, Pitts LH. Selection of surgical approach to acoustic neuroma. Otolaryngol Clin North Am 1992;25:36187. 8. Tos M, Thomsen J, Marsen A. Results of translabyrinthine removal of 300 acoustic neuromas related to tumour size. Acta Otolaryngol (Stockh) 1988;(suppl 452):38-51. 9. Arriaga M, Long S, Nelson R. Clinical correlation of acoustic neuroma volume. Am J Otol 1993;14:465-8. 10. Nedzelski JM, Schessel DA, Pfleiderer A. Conservative management of acoustic neuromas. Otolaryngol Clin North Am 1992;25:691-705. 11. Barbara M, Saliola S, Filipo R. False-positive MRI in a patient with otoneurological pathology. J Laryngol Oto11993; 107:465-7.
Lackland AFB, Texas Magnetic resonance imaging with gadolinium facilitates the early diagnosis of internal auditory canal tumors at a small enough stage to permit increasing application of hearing preservation surgical techniques. Surgeons report successful removal of tumors as small as 3 mm, which are diagnosed with enhanced magnetic resonance imaging. A retrospective study was performed to determine the risk of false-positive "tumor" diagnosis with enhanced magnetic resonance imaging. We reviewed the imaging records, office notes, and surgical records of 112 consecutive "tumors" involving the internal auditory canal treated by the Wilford Hall USAFMedical Center Neurotology Service between July 1991 and July 1994. Two categories of false-positive magnetic resonance imaging were identified: [I] surgically confirmed absence of internal auditory canal neoplasm and (2) spontaneous resolution of the internal auditory canal lesions on subsequent, enhanced magnetic resonance images. Overall, eight false-positive scans were identified. Three were surgically confirmed as false-positive, and five resolved on subsequent imaging studies. All cases were smaller than 6 mm and involved the distal internal auditory canal (fundus]. The surgically confirmed cases were approached through a middle fossa technique with successful hearing preservation. The overall rate of surgical false-positive results was 3.5% [3 cases in 86 surgeries]. However, the overall falsepositive rate for intracanalicular "tumors" was 32% (8 cases in 25 intracanalicular lesions]. Although hearing preservation is more likely in small lesions, the surgeon must consider the possibility that an internal auditory canal lesion smaller than 6 mm may actually represent a nonneoplastic process. Enhancing lesions limited to the internal auditory canal fundus may be treated by reimaging the patient in 6 months after the first image rather than by prompt surgical exploration. [OTOLARYNGOLHEADNECKSURG1995;113:61-70,]
Ill M a g n e t i c resonance imaging (MRI) is currently the most sensitive imaging technique for identifying internal auditory canal (IAC) and cerebellopontine angle neoplasms. Early identification of vestibular schwannomas enhances preservation of hearing and facial function after microsurgical removal. Gadolinium-enhanced, Tl-weighted MRI permits accu-
From the Departments of Otolaryngology-Head and Neck Surgery (Dr. Arriaga), Radiology (Dr. Carrier), and Pathology (Dr. Houston), Wilford Hall Medical Center. The opinions expressed in this article are those of the author and do not necessarily represent those of the Air Force or the Department of Defense or other departments of the United States ,government. Received for publication Sept. 18, 1994; accepted Feb. 10, 1995. Reprint requests: Moises Arriaga, MD, Director Otology/Neurotology, Department of Otolaryngology 2200 Berquist Dr., Suite 1, Wilford Hall Medical Center, PSSO, Lackland AFB, TX 78236. 23/1/64150
rate diagnosis of tumors as small as 2 mm. 1 With specific imaging sequences, most nonneoplastic entities may be excluded, such as cholesterol granuloma, epidermoid, adjacent bone marrow in the porous acousticus, and lipomas. Nonetheless, there remains a small subset of patients with small enhancing lesions in the IAC identified falsely as neoplasms. Recent experience with such cases prompted a review of our patients in an effort to identify imaging features of these enhancing lesions that would allow a more accurate preoperative diagnosis. METHODS AND MATERIAL
A retrospective study was performed to determine the risk of false-positive "tumor" diagnosis with enhanced MRI. We reviewed the imaging records, office notes, and surgical records of 112 consecutive IAC "tumor" patients treated by the Wilford Hall USAF Medical Center Neurotology Service be61
62
ARRIAGA et al.
Otolaryngology Head and Neck Surgery
July t995
Fig. I. Case t. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the right IAC (arrow). C, D and E, Postgadolinium, T~-weighted images at t0, 16, and 33 months of follow-up show essentially no c h a n g e (arrows). The original interpretation of E questioned the possibility of slight increase in size from D; however, no such c h a n g e was evident in reviewing the entire sequence.
tween July 1991 and July 1994. The lesions were entirely intracanalicular in 25 patients. A total of 86 surgical resections of tumors involving the IAC were performed during this time, with 15 of those surgeries for completely intracanalicular lesions. The absence of neoplasm in false-positve cases was confirmed either pathologically or by resolution (partial or complete) on follow-up MRI scans.
All patients were imaged on a 1.5-T GE Signa scanner (General Electric, Milwaukee, Wis.) with pregadolinium and postgadolinium, T~-weighted (TR 600, TE 11) images through the IACs. A 3-mm slice thickness with zero gap, FOV 16, 256 x 192 matrix was used, with two excitations on the pregadolinium sequence and four excitations on the postgadolinium sequence. Additionally, fast-spin echo
Otolaryngology Head and Neck Surgery
Volume t13
ARRIAGA et al.
Number t
63
Fig. 1 (cont.] F, Photomicrograph reveals dense hyalinized fibrous connective tissue and calcified debris. (Hematoxylin and eosin stain; original magnification x 100.} G, Photomicrograph of F at a higher power demonstrating dense, hyalinized fibrous connective tissue with cellular and calcified debris. (Hematoxylin and eosin stain; original magnification x 200.]
Table 1. False-positive MRI of the IAC: Clinical, radiologic, and pathologic features Case No.
Age [yr]/sex
Image findings
1
38/F
1/90 Punctate, 2-3 mm, fundus 10/90 No change 6/91 No change 9/92 ? Minimal growth 12/92 Punctate, 2-3 mm, fundus 3/93 No change 8/94 Ill-defined, partially resolved 5/92 Punctate, 2-3 mm, fundus 6/93 No change 8/94 Resolved 2/94 Ill-defined, 4 mm 7/94 Resolved 2/92 Punctate, 2-3 mm, fundus 7/92 ? increase 12/92 ? Decrease 5/93 ? Increase 8/91 Punctate, 3 mm, fundus 10/92 Ill-defined 8/94 Partially resolved fundus 10/90 IDdefined, 3 mm, fundus 4/91 and 9/92 No change 5/93 Resolved 5/94 Globular 6 x 4 mm
2
69/M
3
57/M
4
32/F
5
44/M
6
70/M
7
54/M
8
41/M
T2-weighted (TR 3600, effective T E 100), 3-mm slice thickness, zero gap, FOV 24, 512 × 256, 2 NEX images were obtained through the same area. After fixation with formalin, the surgical specimens were embedded in paraffin, sectioned at multiple levels, and stained with hematoxylin and eosin in preparation of microscopic evaluation. RESULTS
Eight false-positive "tumors" were identified among the 112 lesions treated during the period of
Clinical
Surgery/pathology
Vertigo/hearing Middle fossa (1/93) Hyaline Hearing Contralateral hearing Hearing Vertigo/hearing
Middle fossa (6/93) Neuronal degeneration
Hearing Hearing/tinnitus Hearing
Middle fossa (5/94)
the study. The false-positive lesions were entirely intracanalicular, in the fundus of the IAC, and smaller than 6 mm in maximum dimension. Three of the eight false-positive tumors were confirmed as nonneoplastic by histopathology among the 86 surgical resections for IAC tumors performed during this time period. The other five lesions either resolved completely or evolved into a less tumor-like appearance on subsequent MRI. The remaining 21 patients who did not undergo operations include 4 patients referred elsewhere for care because of mili-
64
ARRIAGA et al.
Otolaryngology H e a d a n d Neck Surgery July 1995
Fig. 2. Case 5. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the left IAC (arrow). C, Postgadolinium, T~-weighted image at 6 months of follow-up shows lesion slightly larger (arrow). D, Postgadolinium, T~-weighted image at 10 months of follow-up shows lesion slightly smaller (arrow), probably because of partial volume averaging. E and F, Postgadolinium, T~-weighted images at 15 months of follow-up show lesion divided between two slices and apparently larger that al I 0-month study (arrows). However, when compared with 6-month study, it is probably unchanged. Varying size due to slight changes of the lesion relative to slice position.
tary eligibility issues, 1 patient treated with stereotactic radiation therapy, 9 elderly patients with stable lesions and severe medical problems prohibiting general anesthesia, 3 patients in whom the
diagnosis is uncertain and repeat imaging is pending, 2 neurofibromatosis-2 patients with stable tumor in an only-hearing ear, and 2 patients with stable residual tumor operated on more than 10 years ago.
Otolaryngology Head and Neck Surgery Volume 113 Number I
ARRIAGA et al:
65
Fig. 2 (cont.). G, Fast-spin echo T2-weighted image also at 15 months of follow-up confirms space-occupying lesion displacing cerebrospinal fluid (arrow). H, Photomicrograph reveals section of peripheral nerve exhibiting mild vacuolar c h a n g e and osseous debris. (Hematoxylin and eosin stain; original magnification x 40.) I, Higher magnification of H exhibiting peripheral nerve fibers with mild vacuolar c h a n g e and foci of calcified debris. (Hematoxylin and eosin stain; original magnification x 200.) J, Photomicrograph of H and I at a higher magnification demonstrating vacuolar c h a n g e in nerve fibers. (Hematoxylin and eosin stain; original magnification × 400.)
The clinical features, radiologic findings, and histopathologic interpretations are summarized in Table 1 for the eight false-positive lesions. The series consisted of two women and six men, with a mean age of 50.6 years. All the patients except one had serial MRI scans at least 4 months apart to identify changing radiographic features. Follow-up times on the patient s not operated On ranged from 6 to 36 months. The false-positive scans demonstrated two patterns of enhancement in the distal IAC (fundus). In one pattern the enhancement was sharply circumscribed either punctate (cases 1, 2, 3, 5, and 6) or globular (case 8). The other pattern was ill-defined, which occurred in two lesions (cases 4 and 7). Except for case 3, all patients exhibited asymmetric sensorineural hearing loss in the ear with the abnormal imaging. Patient 3 had hearing loss and
tinnitus in the contralateral ear. All eight patients admitted unsteadiness to varying degrees; however, patients 1 and 5 described initial episodes of severe vertigo with subsequent chronic unsteadiness (both were surgically confirmed to be false-positive). During follow-up, no patient demonstrated significant progression of clinical symptoms including hearing loss or tinnitus. Interestingly, patient 7 had severe, disabling tinnitus on the side of his enhancing MRI lesion. Although his lesion completely resolved during the 3-year follow-up period, his tinnitus symptoms were unimproved. The timing and rationale for microsurgical removal of the lesions in the three surgically confirmed false-positive cases involved clinical symptoms, radiologic findings, and patient preference. Patients 1 and 5 had strong family histories of malignant brain tumors. Despite reassurance of the benign nature of
66
ARRIAGA et al.
Otolaryngology Head and Neck Surgery July 1995
Fig. 3. Case 8. A and B, Pregadolinium a n d postgadolinium, T~-weighted axial images through the IACs at initial visit show a globular focus of e n h a n c e m e n t in the distal right IAC (arrow). Findings were consistent with vestibular s c h w a n n o m a on original interpretation. C, Intraoperative view of right middle cranial fossa dissection of IAC in this patient. G, Geniculate ganglion; P, o p e n e d g e of dura at porous acousticus. Arrow identifies junction of the labyrinthine segment of the facial nerve with the IAC portion of the nerve.
the lesions, they were both insistent on microsurgical removal when serial imaging demonstrated no regression and "possible slight growth" of their lesions. In retrospect, and comparing each patient's entire sequence of images, the lesions were stable. Figure 1A-E demonstrates the images during a 2-year period before surgery in patient 1, and Fig. 2A-F demonstrates the images during an 18month period preceding surgery in patient 5. Both imaging features and personal factors prompted surgery for patient 8 (Fig. 3A and B). He was a professional musician for whom any additional hearing loss would have had dire career consequences. Moreover, the globular appearance of his MR! enhancement was highly suggestive of schwannoma. Surgical removal in all three false-positive cases was performed through the middle fossa approach. Hearing was retained within 15 dB of preoperative
thresholds for pure tones and speech reception, and speech discrimination was within 10% of preoperative levels. All three patients have retained their preoperative level of function and independence with good vestibular compensation. Intraoperatively, patients 1 and 5 had abnormal, nodular, distal superior vestibular nerves. The histopathologic findings in patients 1 and 5 are most consistent with sequellae of a previous inflammatory or degenerative lesion. Patient 5 demonstrated vacuolar, neural change with fibrosis in the superior vestibular nerve (Fig. 2G, H, and I). Patient 1 had dense hyalinized fibrous tissue removed from the distal superior vestibular nerve (Fig. 1F-I). These findings are consistent with a reactive inflammatory mechanism, although ischemic or hemorrhagic injury could produce similar findings. Both patients had initial severe vertigo followed by chronic un-
Otolaryngology H e a d a n d N e c k Surgery
Volume 1t3
Number t
ARRIAGA et al.
67
Fig. 4. Case 7. A a n d B, Pregadolinium a n d postgadolinium, T~-weighted axial images through the IACs at initial visit show an ill-defined focus of e n h a n c e m e n t in the fundus of the left IAC (arrow). O and D, Postgadolinium, T~-weighed images at 6 a n d 23 months of follow-up are u n c h a n g e d (arrows). E, Postgadolinium, T~-weighted i m a g e at 35 months of follow-up demonstrates resolution of enh a n c e m e n t (arrow).
steadiness and hearing loss before surgery, symptoms which clinically support a cochleovestibular neuritis. Case 8 merits special mention. The radiographic features were distinct from those of the other seven
cases because the focus of enhancement was globular and larger than in the rest of the cases. Intraoperatively, just as with the other cases, there was a distinctly abnormal appearance to the distal superior vestibular nerve. There was an adhesive, hem-
68
Otolaryngology Head and Neck Surgery July 1995
ARRIAGA et al.
Fig. 5. Case 6. A and B, Pregadolinium and postgadolinium, T~-weighted axial images through the IACs at initial visit show a punctate focus of enhancement in the fundus of the right IAC (arrow). C, Postgadolinium, T~-weighted image at 14 months of follow-up shows conversion of lesion to an ill-defined, barely perceptible focus of enhancement (arrow).
orrhagic rind around the distal vestibular nerve; however, this area of abnormality was smaller than the 6 × 4 mm preoperative size estimate from the MRI. The distal halves of the superior and inferior vestibular nerves were sent as specimens; however, only a small fragment of nerve was actually processed for histopathology. The histopathology of this specimen was interpreted as a normal nerve. There are three possible explanations. (1) Cochlear schwannoma may have been left in situ. This is unlikely because the entire IAC was fully exposed (Fig. 3C) and the cochlear nerve appeared normal intraoperatively. (2) The pathologic portion of the specimen may have been lost in processing. Very small specimens are often adherent to instruments and containers or even inadvertent suctioning during passage. There is no way to disprove this possibility; however, the specimen did not have the usual gross appearance and texture of a schwannoma. (3) The MRI was simply overly sensitive and
identified arachnoid adhesions as tumor. We favor the third explanation. The lesions that were not operated on (cases 2, 3, 4, 6, and 7) either resolved completely or evolved to an ill-defined, partially resolved appearance on follow-up scans. Cases 4 and 7 began as ill,defined lesions and resolved totally on follow-up images (Fig. 4A-E). In comparison, cases 2 and 6 began as sharply circumscribed, punctate lesions and became ill-defined and partially resolved during the time of follow-up (Figs. 5A-C). Case 3 began as a punctate lesion and nearly completely resolved during the follow-up period. DISCUSSION
Gadolinium-enhanced MRI is highly sensitive for intracanalicular IAC pathology and has become the gold standard for preoperative evaluation of the IAC. ~This technique has nearly eliminated the need for the previous gold standards of computed tomog-
Otolaryngology Head and Neck Surgery Volume 113
Number I
raphy scans after the instillation of subarachnoid air or radiopaque contrast. Because of its very high sensitivity, a negative gadolinium-enhanced MRI essentially eliminates the possibility of an IAC neoplasm. 2 Because of its lower specificity, the converse is not true. Enhancement within the IAC may be caused by a variety of processes other than neoplasm. Our review of the literature identified 12 reported cases of surgically proven nonneoplastic enhancing lesions in the IAC. These include Ramsay-Hunt syndrome, 3vestibulocochlear neuritis, arteriovenous malformation, 4 and arachnoiditis) False-positive diagnosis of an intracanalicular neoplasm is a very undesirable situation given the potential morbidity, patient discomfort, and cost of surgery. Haid and Wolf 6 recently reported that their rate of surgically confirmed false-positive IAC tumor explorations was 2% (8 false-positive findings in 400 surgical explorations). In their cases the operative finding was of arachnoidal adhesions suggesting an arachnoiditis in each case without a pathologic diagnosis. In our three surgically confirmed falsepositive cases, two demonstrated neural inflammatory sequellae, and one had intraoperative evidence of arachnoiditis. Our false-positive rate of tumor surgery for lesions involving the IAC was 3.5% (3 false-positive explorations in 86 surgeries) and 20% for intracanalicular lesions (3 false-positive explorations in 15 surgeries). The most critical finding for clinical decisions is the overall false-positive rate for intracanalicular lesions. In this series 32% of intracanalicular lesions were incorrectly diagnosed as neoplasms (8 falsepositive findings in 25 intracanalicular lesions). If false-positive results are possible, what are the clinical features that can alert the clinician? Von Glass et al. s argued that the absence of vestibular findings should dissuade the clinician from diagnosing a schwannoma. However, two of our surgically confirmed cases had prominent vestibular symptoms before surgery. In one of our cases, the imaging studies were performed for contralateral symptoms (case 3); clinical observation with serial imaging was the prudent course of action. In our experience, small, enhancing, punctate, or ill-defined lesion in the fundus should raise the suspicion of a lesion or process other than a neoplasm. All five of the lesions that we did not operate resolved completely or became ill-defined on follow-up scans. Han et al.'s series 4 of three surgically confirmed false-positive lesions also identified that intense, focal enhancement limited to the lateral
ARRIAGA e t al.
69
IAC was suspicious for a nonneoplastic lesion. Occasionally, cases may be quite dramatic with enhancement of the entire IAC, positive air computed tomography scan, and appropriate audiologic findings and may still represent only arachnoiditis. The globular pattern of enhancement in our case 8 was consistent with schwannoma and similar to the cases reported by Von Glass et al. 5 and Haid and Wolf. 6 Serial imaging offers insight into the biologic behavior of the enhancing lesion rather than the static information from a single MRI. Case 8 was the only lesion without serial imaging; perhaps a second image 6 months later would have indicated resolution. Even serial imaging has its pitfalls. When lesions smaller than the 3-mm slice thickness are examined, the lesion may appear to change size from one scan to the next because of slight changes in the position of the lesion relative to the slice. This may result in division of the lesion between two images or partial volume averaging, either of which may alter the apparent size of the lesion (Fig. 2E-F). When there is an apparent slight change in size compared with that shown in the last study, the clinician and radiologist must compare the current study with all available old scans. For example, case 5 demonstrated a "slight increase in size" on the last scan done before surgery. However, when compared with the original scan in the series no real change was found. The rationale for surgery on a small intracanalicular schwannoma is total tumor removal with hearing and facial nerve preservation. Excellent data exist to support improved hearing preservation 7 and improved facial function preservation 8 for smaller rather than larger lesions. However, it has not been shown that there is a difference in functional outcome if schwannomas are removed at a 2- to 3-ram size rather than a 6- to 8-mm size. In larger tumors we have demonstrated that a "clinical break point" of significantly worse facial nerve outcomes occurs once lesions reach 2.5 cm. 9 Regardless of approach, excellent outcomes have been reported for tumors smaller than 1 cm. Furthermore, because average reported growth rates are less than 2 mm per year, 1° there is little risk to reimaging a small intracanalicular lesion in 6 months and subsequent reimaging annually if there is no significant growth. 1 Surgical approach selection for small IAC lesions is also important. Our lesions involved the fundus of the I A C - a region ideally suited for the middle fossa approach. In one report, the authors used translabyrinthine exposure only to find that there was no
Otolaryngology Head and Neck Surgery July 1995 -
70
ARRIAGA et al.
neoplasm. 1I Middle fossa exposure offers the possibility of hearing preservation with distal IAC exposure. CONCLUSIONS
1. Small enhancing intracanalicular lesions of the IAC fundus may represent nonneoplastic pathology. 2. Most small enhancing intracanalicular lesions of the IAC fundus should be treated by reimaging the patient 6 months after the first image rather than prompt surgical exploration. 3. Comparison studies should be compared with the entire imaging series, not just the previous study, because apparent growth on subsequent MRIs may reflect positioning of the tumor in relation to the slices rather than actual tumor growth. 4. The middle fossa approach offers the possibility of hearing preservation for small, distal IAC lesions, especially if the ultimate diagnosis is of a nonneoplastic pathologic process. REFERENCES
1. Strasnick B, Glasscock ME, Haynes D, McMenomey SO, Minor LB. The natural history of untreated acoustic neuromas: Laryngoscope 1994;104:1115-9.
2. Sidman JD, Carasco VN, Whaley RA, Pillsbury HC. The new gold standard for diagnosing cerebellopontine angle tumors. Arch Otolaryngol Head Neck Surg 1989;115:1244-7. 3. Anderson RE, Laskoff JM. Ramsay Hunt syndrome mimicking intracanalicular acoustic neuroma on contrast-enhanced MR. AJNR Am J Neuroradiol 1990;11:409. 4. Han MH, Jabour BA, Andrews JC, et al. Nonneoplastic enhancing lesions mimicking intracanalicular acoustic neuromas on gadolinium-enhanced MR images. Radiology 1991; 179:795-6. 5. Von Glass W, Haid CT, Cidlinsky K, et al. False-positive MR imaging in the diagnosis of acoustic neurinomas. OTOLARYNGOLHEADNECKSURG1991;104:863-7. 6. Haid CT, Wolf SR. False positive MRI findings of the eighth cranial nerve simulating a tumour lesion [Abstract]. Presented at the Fourth International Symposium and Workshops on Inner Ear Medicine and Surgery, Snowmass-Aspen, Colo., July 17, 1994. 7. Jackler RK, Pitts LH. Selection of surgical approach to acoustic neuroma. Otolaryngol Clin North Am 1992;25:36187. 8. Tos M, Thomsen J, Marsen A. Results of translabyrinthine removal of 300 acoustic neuromas related to tumour size. Acta Otolaryngol (Stockh) 1988;(suppl 452):38-51. 9. Arriaga M, Long S, Nelson R. Clinical correlation of acoustic neuroma volume. Am J Otol 1993;14:465-8. 10. Nedzelski JM, Schessel DA, Pfleiderer A. Conservative management of acoustic neuromas. Otolaryngol Clin North Am 1992;25:691-705. 11. Barbara M, Saliola S, Filipo R. False-positive MRI in a patient with otoneurological pathology. J Laryngol Oto11993; 107:465-7.