Partial parotidectomies: morbidity and benign tumor recurrence rates in a series of 94 cases
abstract
Objectives Review the methods available for parotidectomy. Describe the technique of partial parotidectomy assisted by evoked electromyographic nerve location and the expected morbidity and benign and low-grade cancer tumor recurrence rates from this modified procedure.
Study Design From 1983 to 1999 the author performed or assisted in 94 parotidectomies (79 partial), all done by a single specialty group and all using evoked electromyographic nerve location. The cases were surveyed by reviewing all the hospital and office records on these cases and tabulating the type and extent of surgery, pathology, postoperative problems recorded, and long-term follow-up.
Methods Partial parotidectomy was elected in those cases of benign and low-grade malignant disease in which adequate tumor removal required a less than complete lobectomy or total parotidectomy. Heavy reliance was placed on proactive nerve location by an evoked electromyographic device with dissecting/stimulating hemostat. A retrospective review focusing on these cases was performed based on the patient charts and their continued documentation by the practice.
Results In 79 partial parotidectomies there were no documented facial nerve injuries and one incidence of recurrence of a benign mixed tumor and an acinic cell carcinoma, respectively.
Conclusion Partial parotidectomy has the advantages of reduced risk to the facial nerve, reduced operating time, possible outpatient surgery, and no apparent risk of increased recurrence of benign tumors.
INTRODUCTION
Parotidectomy can be classified in terms of the extent of resection as partial or total. Although the literature supports possible partial parotidectomy as an intraoperative option based on the surgical problem, otolaryngology residency programs have generally taught the facial nerve main trunk location and lateral lobe parotidectomy as an unwritten standard of care for removal of benign and low-grade malignant parotid masses.
The rational for the (superficial) lateral lobectomy parotidectomy is based on the recognition of the need for sparing the facial nerve. Blind lumpectomy or shelling out (finger dissection) of the tumor was found to carry an unacceptable risk of facial nerve paralysis. 1Anatomically, the main trunk of the facial nerve is the most reliable site for initial visualization because of its exit from the stylomastoid foramen. The lateral lobectomy proceeds logically from the removal of the tissue lateral to the facial nerve as the nerve is distally dissected and preserved. Experienced parotidectomy surgeons would often amputate the lateral lobe and suspend the dissection once the mass was effectively elevated, sparing further disturbance of facial nerve branches uninvolved in the mass. Deep lobe removal was often begun only after performance of a complete lateral lobectomy.
From an oncologic standpoint, the complete superficial lobectomy often makes no sense. The actual margin or cuff of normal tissue between the mass and the facial nerve, which is to be spared, is often minimal to nonexistent. Except for possible malignancy that might require sacrifice of the facial nerve, the final margin of resection is generally the shortest distance from the tumor mass to the facial nerve.
Recent reports seriously discuss partial parotidectomy as an option. Witt 2 has reported 76 parotidectomies of which he performed 59 partial parotidectomies. He excluded, automatically, all tumors larger than 3 cm with reduced mobility or involving the deep lobe. He always performed a main trunk facial nerve location with distal dissection until the tumor was adequately resected (2 cm margin, which is ultimately reduced to the shortest tumor-to-nerve distance) and reported no permanent nerve injuries and no recurrences over 10 years. His strategy was to reserve partial parotidectomy to those cases in which the frozen-section diagnosis was either benign or low-grade malignancy. Witt 3 has also reported, in a separate study, his conclusions that facial nerve (integrity) monitoring in parotidectomy is not the “standard of care,” although he routinely uses a disposable nerve stimulator at the end of the case for piece of mind.
Helmus 4 reported a series of 203 parotidectomies (146 “subtotal”). His surgical strategy was to anatomically locate a peripheral facial nerve branch distally and dissect it retrograde toward the tumor. He removed the tumor mass with a cuff of normal tissue using a Nd:YAG laser, precisely controlling bleeding, and direct visual nerve identification in establishing the margin. He used frozen section of the resected mass for “peace of mind.” There were no permanent facial nerve injuries and Helmus did not routinely use a nerve stimulator.
A recent three-part journal section, Clinical Challenges in Otolaryngology, provided a friendly debate of the adequacy of partial parotidectomy in tail of parotid lesions. The conclusion of these authors, was, indeed, that it was appropriate if the tumor was less than 3 to 4 cm in size, not in the deep lobe, and not malignant. 5
MATERIALS AND METHODS
The author's experience with parotidectomy has been an evolving one influenced by access beginning in 1983 to a dissecting/stimulating hemostat and later to an evoked electromyographic (EMG) nerve locator. The use of this equipment frees the surgeon from the need for main-trunk nerve location, or in many cases, from any nerve exposure at all. In a series of 94 cases performed with this equipment the author served as either surgeon or first assistant. Members of a single-specialty otolaryngology group performed all of these procedures and the lead surgeon always made the decision as to the method and extent of operation.
In this technique of surgery using evoked EMG nerve location, surface electrodes (sEMG) were applied during prepping of the patient either at the corner of the mouth or to the eyelid/forehead area constituting an anode, cathode, and ground. Electrode location was selected based on the branch of the facial nerve that the dissection was expected to first encounter. The skin was routinely sanded with 600-grit sandpaper before electrode application. Impedances were measured and were acceptable if under 3,000 Ohm. At the same time a stimulation ground electrode was likewise applied to the upper chest just above the clavicle on the side of the parotid lesion.
The surface electrodes were then attached to a Neurovision SE (RLN Systems, Inc., Jefferson City, Missouri) single-channel, automatic evoked EMG nerve locator. This device has a dedicated nerve stimulator output at 4 PPS slaved to a microprocessor-based evoked EMG input attached to the sEMG electrodes. The unit works by delivering a stimulation pulse as output and monitoring for a return compound motor action potential on the sEMG input channel. The microprocessor analyzes the timing, frequency content, and amplitude of the return signal and alerts the surgeon with an audio warning tone if the nerve is stimulated above a threshold. If stimulation is delivered, but no return signal occurs, the device provides a soft “click” to inform the surgeon.
Proactive nerve location is provided by the use of a dissecting/stimulating hemostat. This is a dedicated stimulator probe constructed by adding an electronic input to the finger-ring and by insulating the hemostat down to its tips. The dissecting/stimulating hemostat can then be used to perform blunt dissection with continuous nerve location. By running the stimulation at slightly higher intensities (1, 2, or 3 ma) the stimulating hemostat will give progressively wider margins of tissue depth to the nerve before alarming, often in the 0.5- to 1-cm range. This enables the nerve locator to become “virtual,” by alarming in the vicinity of the nerve rather than requiring actual visual identification or touching of the nerve.
RESULTS
We retrospectively examined each case record for recurrence rates for tumors, especially the benign mixed tumor, and nerve injury. Ninety-four cases were found (Table I), of which 39 were found to be pleomorphic adenoma on final pathology. Eighteen tumors were malignant and 66 were benign. Four surgeries were total parotidectomies, 11 were lateral lobectomy parotidectomies, and 79 were partial parotidectomies.
 |
Four of the benign mixed tumors had lateral lobectomy and 35 had partial parotidectomy. There was one case of recurrence of benign mixed tumor. In this case the initial excision was performed in December 1988 with inadequate removal of tumor recognized at the time of surgery because of splaying of the facial nerve by the tumor. This distortion of the main trunk of the facial nerve by the tumor was too near the stylomastoid foramen to allow for proximal identification of the nerve. Tumor was deliberately left on the nerve to prevent injury. Recurrence was clinically evident at 5 years and was resected completely in October 1993 with preservation of the facial nerve. An additional exploration of the wound was performed in December 1995 because of suspected mass that proved to be scar tissue with no recurrence of tumor. All three of these procedures were partial parotidectomies with no postoperative facial nerve injury, cosmetic defect, or Frey's syndrome evident. In addition, one case of low-grade acinic cell carcinoma of the parotid gland was recurrent and was treated by direct, simple re-excision. Frozen-section diagnosis of tumors was done as indicated, but no radical resection of the facial nerve was performed at the initial procedure solely on the basis of frozen section. All of the malignant lesions, except for the low-grade acinic cell carcinomas, were additionally treated by later, more extensive surgery, chemotherapy, or by radiation therapy. Five of the patients with malignant tumors have subsequently died from their disease. There were no incidences of iatrogenic facial nerve injury in the series of partial parotidectomies, although one of the lateral lobectomies had a 3-month facial nerve paralysis, which resolved spontaneously. Analysis of surgical times showed a steady drop in average skin-to-skin operating time throughout this series, from average times of approximately 90 to 100 minutes in the earliest cases, to approximately 50 minutes in the latest cases. Reduction in surgical times is attributed to increased surgeon reliance on the technique, reduced percentage of gland removal, and reduced need for identification of the main trunk of the facial nerve.
