In the overall series, 99% were independent in their activities. The operative mortality rate was 0.5% (two of 410 patients). All patients with tumors up to 1.0 cm had a good result, as did 95% in the 1.0-1.9-cm group and 92% in the 2.0-2.9-cm group. Even patients with large tumors had an 80% chance of having a good outcome. The most common reasons for the fair results were impaired balance, gait, and coordination. Dysarthiia or diplopia occurred in a few patients. In six of the 40 patients with a fair result there was a significant preoperative deficit which improved but still limited the patient's activity. In about 5% of the entire series a significant headache problem lasted longer than expected; in two patients this prevented full recovery, placing them into the fair result category.
The poor result and deaths are recorded in Table 24.2. The poor
Extent of Tumor Removal and Recurrence
The goal of the operation is total removal of the tumor. However, this goal must be tempered by surgical judgment which considers the need to preserve and improve function, as well as the long-term results. My experience indicates that there is a place for subtotal and radical subtotal removal of acoustic neuromas because the recurrence rate has been low and in larger tumors the incidence of postoperative neurological problems has been low, with no poor resufts. This has been especially true in elderly patients.
Radical subtotal removal describes those procedures in which a small fragment of tumor is left, usually because it is densely adherent to the facial nerve or brainstem (Fig. 24.11). The term subtotal removal is used to describe those patients in whom an extensive removal of the tumor is done but a portion of the rim of the capsule is left attached to the brainstem and cranial nerves (Fig. 24.12). Table 24.3 records the relationship between tumor size and the extent of removal.
Table 24.4 records the reasons for
doing a radical subtotal or subtotal removal. The most common reasons
were adherence of the tumor to the facial nerve or brainstem, age
(70 years or older), and the patient's request. After carefully
considering all the treatment options some patients now request
a less than total removal to reduce the risk of damage to the facial
nerve and neurological disability.
After the decision to do radical subtotal or subtotal removal, the recurrence rate needs careful evaluation. An important observation is seen in Table 24.5. Over the 14 years of this series none of the 35 patients with radical subtotal removals have had a recurrence requiring treatment. These patients have been observed for 1-12 years (average, 4.4 years). In seven of the patients their tumors could not be seen on follow-up scans.
Following subtotal removal nine of 52 (17%) patients had recurrence requiring treatment (average follow-up, 4.2 years) (Table 24.5).
Recurrence can also occur following apparent total removal of the tumor. The known recurrences rate in this series was 0.9% (three of 326). The first patient was 69 years of age. She exhibited a large recurrence over 2 years and had another apparent total removal. Subsequent MRI scans over 2 years have shown no recurrence. The second patient had recurrence diagnosed 5 years after surgery. She was given fractionated radiation therapy and, subsequently, there was no change over 3 years. The last patient had a small tumor seen on her first MRI scan 9 years after surgery. Repeat MRI over 2 years has shown no change.
Wazen el al. (58) found that in nine of 13 patients with subtotal removal (age, 66-81 years) there was no growth in the residual tumors on followup ranging from 6 months to 15 years. KJemink et al. (17) reported 20 patients who had incomplete removal of the tumor to reduce the operative risks. Two groups were defined, a subtotal group (resection of less than 95% of the tumor) and a near-total group (resection of 95% or more of the tumor). The subtotal group included mostly elderly patients (mean age, 68.5 years) with large tumors. The near-total group consisted of young patients (mean age, 45.8 years). The mean length of follow-up was 5 years, and only one patient showed regrowth. Lownie and Drake (27) reported that nine of 11 patients followed for 10-22 years after radical intracapsular removal did not have recurrence. The two recurrences were at 2 and 3 years postoperatively.
The low incidence of recurrence and the ability to effectively treat recurrence when it does occur suggest that we should consider more often a radical subtotal or subtotal removal in patients with large tumors, particularly in the elderly.
There were four patients with this complication. One epidural and one cerebellar hemorrhage were diagnosed on CT scans when the patients were showing increasing neurological symptoms. Reoperation was followed by a good recovery. Another patient with an epidural hematoma had the sudden onset of coma. Intubation and opening of the incision in the intensive care unit were followed by a good recovery. The last patient developed a hemorrhage into the cerebellum and brainstem during removal of a large tumor. She did not recover.
There were three patients with this complication. One had a mild disability and he responded to steroids and intermittent mannitol therapy. The other two were not doing as well as expected 12-18 hours after surgery. CT scans documented the cause, and resection of the lateral cerebellum was followed by recovery with mild residual deficits.
CEREBROSPINAL FLUID LEAK
The incidence of cerebrospinal leak has been 8%. In this series 6.5% subsided with the use of a lumbar drain for 72 hours. In the other 1.5% a transmastoid obliteration was done and there has not been any recurrence.
Two patients required a ventreuloperitoneal shunt for hydrocephalus within a few weeks of the operation. In two patients a lumbar drain was used for 72 hours and there was no recurrence of the problem. One patient developed a pseudomeningocele and a loculated cyst in the cerebellopontine angle and cerebellum, causing brainstem compression, 4 months after his operation. The symptoms were relieved with a cyst-topeiitoneal shunt.
Four patients had a wound infection. Bacterial meningitis occurred in five, including two with wound infections. All responded to antibiotic treatment and, when necessary, debiidement of the incision.
A chronic granulomatous meningitis starting 1 week after operation was developed by one patient. All cultures and biopsy were negative. She did not respond to several medical programs of treatment. She died several months later.
Two patients had aseptic meningitis. They responded to steroid treatment.
Some patients note mild unsteadiness that clears over several days to several weeks. More severe impain-nent in walking and difficulties with coordination and dysarthria take longer to improve. If there is transient difficulty in focusing the eyes or diplopia it usually clears over days to a few weeks.
Transient vertigo may occur, especially after removal of small tumors when vestibular ftmction was present preoperatively. The loss of flmction in the vestibular nerve is gradually accommodated by central compensation (14).
The postoperative neurological problems causing fair and poor results were discussed earlier. Approximately 8% had persistent deficits in coordination or gait. Several patients had increased facial numbness, which was not a problem except in the few patients who also had facial paralysis. Recovery of the numbness was variable. One patient had an isolated 1 lth nerve palsy. Mild trouble swallowing was noted in a few patients but this improved. One patient (age, 76 years) required a tracheostomy and gastrostomy for severe preoperative dysphagia. Several months after operation she had recovered to the point where these could be removed and she went from a chronic care facility back to her own home. Three patients noted impaired sensation in the oliposite extremities.
Five patients had deep venous thrombus and two had pulmonary emboli. We now use altemating-compression thigh-high air boots beginning when the patient enters the operating room. There were no serious postoperative cardiac problems. Electrocardiogram (ECG) changes and any cardiopulmonary symptoms were immediately evaluated by a cardiologist. When there were mild changes on the ECG these tended to resolve in the early postoperative period. One patient had an intraoperative ECG change which led to cessation of the operation, and a subtotal removal was done, but the ECG changes resolved postoperatively.
Persistent headache remains a significant problem in at least 5% of the patients. It usually improves with time and responds to a program of physical therapy and, in a few patients, local blocks.
Facial Nerve Function
The facial nerve is so involved with tumor in some patients that it cannot be saved. A decision has to be made whether to leave a small piece of tumor capsule with the nerve (radical subtotal removal) or to divide the nerve and do a nerve graft. I have left tumor in patients who, in the preoperative discussion, requested this to reduce the risk of facial paralysis, knowing that they might need further surgery in the future. The results of long-term follow-up on these patients are encouraging.
In the early postoperative period the status of the facial nerve gives an indication of what the final result will be. If there is a complete paralysis (grade 6), it is unlikely the patient will recover beyond grade 3. A patient at grade 2 will usually make a good recovery, as will a high percentage of grade 3 patients. A delayed facial weakness is not uncommon, but rarely does it progress to a complete paralysis and there is a high probability of a good early recovery.
If the patient awakens from anesthesia with facial paralysis and the inability to adequately close the eyelids or develops delayed facial paralysis, the cornea must be protected. Management of the paralyzed face is a complex problem. Initially the eyelids are approximated with tape, artificial tears are used regularly during the day, and an ophthalmic ointment is used at night. The use of tarsorrhaphies and, more recently, gold weights is essential to the maintenance of a healthy cornea and prevention of visual loss and incapacitating pain. Which oculoplastic procedure is best suited for a particular patient depends upon the patient's age, skin laxity, and the presence or absence of corneal anesthesia. All of these procedures have the advantage of being reversible. Improved results may occur by using a modification of the classic hypoglossal-facial anastomosis with partial division of the hypoglossal nerve and anastomosis of half of the nerve to the lower branch of the facial nerve, combined with one of the eye procedures and a temporalis transposition flap.
Cochlear Nerve Function
In 1984 our group published the results of a series of 22 patients in whom an attempt was made to save hearing (42). Subsequent publications have updated this series (29-31, 43). In 1988 Gardner and Robertson (1 1) reviewed the reports on hearing preservation published in the English literature from 1954 to 1986. Several publications have subsequently discussed hearing preservation (3-5, 8, 10, 12, 18, 21-25, 29-31, 33, 37-39, 42, 43, 45, 46, 48-52, 54, 59). In spite of extensive clinical interest in this subject, our experience is similar to that of Cohen (3), who found that the results of hearing preservation have not improved over the past decade.
The question of what constitutes serviceable hearing has been discussed by several authors, with the most common criteria being a speech reception threshold of <50 db and a speech discrimination score (SDS) of 50% or more (11, 51, 59). On the other hand, Whittaker and Luetje (59) support the definition of a SDS of 70% or better for serviceable hearing, while we have used a SDS of 35% or better because for some patients this has been very useful (39, 42).
Our results show that for patients with at least a 35% SDS the chance of saving useful hearing near the preoperative level is 60% if the tumor is intracanalicular or extends less than 5 nim into the posterior fossa and 36% if it extends 0.6-1.5 cm. Tumors 2 cm or larger have a low probability of hearing preservation even if there is excellent preoperative hearing. Preservation of hearing has been reported in a few patients following removal of large tumors, including one of my patients (37).
In an attempt to help preserve hearing during removal of an acoustic neuroma, we have investigated the monitoring of auditory evoked potentials using a system developed by Dr. Robert Levine (21-25, 30, 31, 42). Electrocochleography (ECochG) monitors the status of the cochlea and the auditory nerve peripheral to the tumor, and brainstem auditory evoked potential (BAEP) monitors the neural activity central to the tumor. ECochG is a near-field potential that provides rapid feedback of the compound action potential of the auditory nerve (Nl) probably generated near the cochlea, and cochlear microphonics potentials which are generated by the hair cells of the inner ear. BAEP is a far-field potential which has a slower feedback. In practice, only wave V, which is generated within the brainstem, is monitored because the other potentials are much smaller and often undetectable.
The short-latency ECochG potentials are the most useful for monitoring during operation because they are generally not affected by anesthesia, they are almost always detectable, and they have inunediate feedback (22). On the other hand BAEPS, while useful, are undetectable in some patients even when there is useful hearing and it may take up to 1 minute or more to obtain satisfactory recordings because of the small amplitude of the potentials.
By monitoring both ECochG and BAEP, the entire portion of the auditory system at risk during an acoustic neuroma operation can be monitored. The presence or absence of Nl indicates the integrity of the auditory nerve peripheral to the tumor, wave V is an indication of auditory nerve activity central to the tumor, and the cochlear microphonics indicate the status of the cochlea, which is at risk from interruption of blood supply or from damage to other structures essential for cochlear function, such as the posterior semicircular canals.
When the status of Nl and wave V at the end of the operation are correlated with the hearing outcome, it is found that if Nl and wave V are lost there is no hearing. In our series, if wave V and NI were present, all but one patient had useful hearing. If Nl was present and wave V was lost or never detected, the results were not predictable (23). A fundamental limitation of the monitoring is related to how the individual nerve fibers react to the injury. They may stop conducting completely, there may be too few fibers left to generate a gross potential that can be detected, or they may conduct a modified or desynchronized impulse.
The hope is that monitoring will give an indication of early hearing compromise that is reversible and will allow the surgeon to alter the dissection (23, 42, 43). This has likely been the case in some patients in whom a change occurred which recovered when the dissection was stopped or altered. Monitoring has not made a definite difference in the outcome when there has been abrupt loss of function without warning that does not recover, presumably due to interruption of vascular supply, when gradual loss of function occurs, or when there is no change in any waveform during the operation. However, we believe that monitoring has helped us to better understand the problems in preserving hearing function. Slavit et al. (52) compared two matched series of patients with and without auditory monitoring and concluded that there is a benefit from intraoperative monitoring, but Compton et al. (5) found it to be of little help.
The long-term results of hearing preservation have been evaluated. In the report of Shelton et al. (50), 14 of 25 patients (56%) who underwent removal of an acoustic neuroma by the middle fossa approach suffered a significant loss of hearing in the operated ear over a mean follow-up of 8 years (range, 3-20 years). On the other hand Palva et al. (45) reported a significant loss in only two of 13 patients during the first 4 years following suboccipital removal. Rosenberg et al. (48) did not observe a significant decline in nine patients observed for 1.3-11 years. McKenna et al. (29), reporting our series of 18 patients with follow-up ranging from 3.4 to 10.4 years (mean, 5.4 years), found four patients (22%) with a significant decline in hearing. Changes did not correlate with tumor size, preoperative hearing, intraoperative changes in hearing, the interval between initial symptoms and surgery, sex, or age.
Concern about recurrence following removal of an acoustic neuroma with preservation of the cochlear nerve has been discussed in the literature. Thedinger et al. (55) emphasized that inadequate exposure of the lateral end of the internal auditory canal may be associated with leaving a remnant of tumor. Neely (33) reported that in patients in whom all of the tumor appeared to have been removed residual tumor was found in the cochlear nerve, and he concluded that, "histologic data suggest that complete tumor removal in attempts to preserve hearing may be beyond our surgical capabilities." However, Sami et al. (49) reported no recurrence in 16 patients who had removal of intracanalicular acoustic neuromas with anatomic preservation of the cochlear and facial nerves, who had been observed for 1-8 years. In the series reported in this paper an attempt to preserve hearing was made for 119 patients with tumors of 2.0 cm. Follow-up CT and MRI scans have shown no definite recurrence. A few patients have an area of gadolinium enhancement in the internal auditory canal on MRI scans. Whether this represents residual tumor or postoperative scar is unknown, but follow-up scans have remained unchanged.
Management of Tumor in the Only Hearing Ear
The largest experience with acoustic neuromas involving the only hearing ear is in patients with bilateral acoustic neuromas (43). However, this problem is occasionally encountered in patients with a unilateral tumor when the opposite ear is deaf because of previous infection or trauma. The following guidelines for treatment of these patients are suggested.
In this series there were two patients, including one with a 1.5-cm tumor who did not want to risk loss of hearing and had a decompression with subtotal removal. The hearing loss has stabilized. The other patient
has been reported previously (37). The patient had a large tumor where total removal was required to relieve brainstem compression, and she did enjoy retention of preoperative hearing.
Pensak et al. (46) reviewed this problem and reported two patients in whom unilateral acoustic neuromas, 1.0 cm and 2.0 cm, in the only hearing ear were treated with complete removal and retention of unstable hearing.
Management in Elderly Patients (70 Years and Older)
Patients 70 years of age and older were considered for surgery if they had a progressive or disabling neurological deficit or symptoms. In 22 of 29 patients there was a good result, six had a fair result, and only one had a poor result (Table 24.2). Total removal was done in five, all of whom had tumors of 2.5 cm or less. A radical subtotal removal was done in five, and 19 had a subtotal removal. No patient has had to stay in a nursing home and even the patients with fair results are able to care for themselves and do many activities. Only two, described previously, required a second operation for recurrence.
Management in Patients with Preoperative Hydrocephalus
In contrast to 25 years ago, we now rarely see a patient who has symptoms of high-pressure hydrocephalus. Some patients with enlarged ventricles on their scans have no symptoms of hydrocephalus and the ventricles become smaller after tumor removal. However, a few elderly patients with large tumors and large ventricles are seen who have symptoms suggesting normal-pressure hydrocephalus. These patients have been treated with subtotal removal of the tumor and placement of a ventriculoperitoneal shunt, usually in the same operation but occasionally at separate operations. They have generally had good long-term results.