CONTROVERSIES

Ramon M. Esclamado, MD, Section Editor

MANAGEMENT OF TONSIL CANCER

Consultants: Carol R. Bradford, MD, Neal Futran, MD, DMD, Glenn Peters, MD

A 43-year old man is examined for evaluation of nous primaries are seen in the upper aerodiges- a persistent, right-sided sore for 4 months, tive tract and esophagus. The biopsy is positive and a painless, slowly enlarging right neck mass for invasive moderately well-differentiated squa- for 3 months. On further questioning, he has right mous cell cancer, therefore staged as T2 N2a M0. otalgia, but denies dysphagia, odynophagia, he- moptysis, hoarseness, or weight loss. He has a 40 1. What treatment options are available pack/year smoking history, but is otherwise in for this patient that can be offered outside of good health. a clinical trial? Physical examination revealsa3cm×3cm exophytic mass involving the right tonsillar fossa and posterior tonsillar pillar. There is no trismus, Dr. Bradford: This 43-year old male with a T2 the base and soft are not involved, N2a M0 moderately differentiated squamous cell and his teeth are in good repair. The remainder of carcinoma of the right tonsillar fossa could be the head and neck examination is remarkable treated either with primary surgery plus postop- only for a firm, freely mobile 4 cm×4cmright erative radiation therapy or radiation therapy level II lymph node. primarily followed by a modified neck dissection. A chest radiograph is normal, and a neck CT is A third alternative would be radiation therapy shown in Figure 1. Examination under anesthe- alone and close monitoring of the primary site sia confirmed the above findings, and no synchro- and neck.

Dr. Futran: Treatment options include (1) ex- Readers are invited to submit particularly difficult cases for consideration ternal beam radiation therapy alone to the pri- to Ramon M. Esclamado, MD, Editor, Controversies, Department of Oto- laryngology and Communicative Disorders/A71, The Cleveland Clinic mary lesion and both sides of the neck followed by Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. right modified neck dissection, (2) external beam From the Department of Otolaryngology, University of Michigan Medical radiation therapy as above with brachytherapy to Center, Ann Arbor, Michigan (Dr. Bradford); the Department of Otolaryn- gology, University of Washington, Seattle (Dr. Futran); Division of Otolar- the primary site followed by right modified radi- yngology, University of Alabama, Birmingham (Dr. Peters). cal neck dissection, (3) surgical extirpation of the Accepted for publication April 27, 1999. primary lesion with a modified radical neck dis-

Head & Neck 21: 657–662, 1999 section and likely flap reconstruction followed by external beam radiation therapy, or (4) for T2 le- CCC 1043-3074/99/0070657-06 © 1999 John Wiley & Sons, Inc. sions, I do not recommend chemotherapy con-

Management of Tonsil Cancer HEAD & NECK October 1999 657 most tonsillar fossa tumors are moderate- to well- differentiated lesions, the degree of differentia- tion would not significantly influence my decision making. (4) The presence of clinical trismus asso- ciated with the tumors suggests pterygoid muscle involvement and would influence me toward a primary surgical treatment plan.

Dr. Peters: Although there is considerable ef- fort being put forth trying to identify pretreat- ment prognostic indicators, at this point in time I still rely chiefly on the clinical stage in making treatment recommendations. Despite the obvious weaknesses and shortcomings that exist with the current staging system, there has not yet been FIGURE 1. Axial CT scan with contrast demonstrating level II identified a reliable biologic or molecular marker node with superficial primary tumor and no pterygoid muscle of to assist us in making these decisions. mandible involvement. Dr. Bradford: A lymphoepithelioma or poorly comitant with radiotherapy or as an adjunct to differentiated carcinoma of the tonsillar fossa radiation therapy. should be more radiosensitive than moderately or well-differentiated tumors. A thin or exophytic tu- Dr. Peters: Outside of a protocol setting, the mor probably also would be more radiosensitive. treatment options include radiation therapy or These factors would include my decision to treat surgery. I would be reluctant to use chemo- with primary radiotherapy, particularly the lym- therapy as an additional treatment modality un- phoepithelioma histology. less the patient was on a protocol of some sort. Radiation could be given as a single daily fraction, 3. Assuming that the patient has no treat- but I feel that this patient would likely receive an ment biases, what would be your treatment accelerated fractionation treatment program recommendation? (BID, concomitant boost, etc.). Surgery would in- clude resection of the primary along with neck Dr. Peters: I think that this is an ideal patient dissection and free flap reconstruction. The need for which I would recommend treatment on a pro- for postoperative radiation would be assessed tocol using combined modality chemotherapy and based on the path report. radiation therapy with surgery held in the sal- 2. Are there any tumor characteristics vage mode. If protocol treatment is not an option, that are helpful in determining a preferred then I would recommend accelerated fraction ra- treatment option? diation therapy, again with surgery held for sal- vage. The patient would be followed very closely Dr. Futran: (1) Location of the tumor in the during therapy, monitoring him for an objective tonsillar fossa: tumors located within the tonsil- response at the end of two cycles of chemotherapy lar fossa tend to have a higher rate of metastasis or 50 Gy of radiation. If an appropriately good than tumors arising from the anterior pillar, but response is noted, then I would feel fairly comfort- also tend to be more radiosensitive. Tumors aris- able in continuing nonsurgical therapy to its ing from the anterior tonsillar pillar tend to be completion. If a less than adequate response is more well differentiated and may extend to the seen in this patient then we would go on to sur- retromolar trigone, making surgery a more at- gery. At the conclusion of therapy, any remaining tractive primary option. Further, if the tumor ex- mass in the neck, demonstrated either by physi- tends onto the tongue base, primary radiation cal examination or CT, would need to be removed therapy has a poorer control rate and would with a neck dissection. I have been using selective influence me to use surgical extirpation as the neck dissections (leaving level V undissected) primary option. (2) Several studies suggest that more in this setting as I have found a large num- exophytic tonsil tumors are much more radiosen- ber of patients who have pathologically negative sitive than ulcerated or indurated tumors. (3) As necks. Another important aspect of treatment is

658 Management of Tonsil Cancer HEAD & NECK October 1999 nutrition and I would place a PEG prior to the mary tumor, the ultimate functional result would initiation of any therapy. be better. My reason for choosing this treatment ap- proach is that it provides the patient with the best 4. If surgery followed by postoperative chance for preserving nearly normal swallowing radiation was recommended, how would the and speech. As long as the patient is monitored primary tumor be approached, and what closely and his response to therapy is carefully type of neck dissection would you perform? established I do not feel that oncologic control is compromised. Dr. Bradford: The specifics of this treatment approach are as follows. I have a very low thresh- old for mandibulotomy to improve visualization Dr. Bradford: The advantage of surgery and for surgical extirpation. In a teaching environ- postoperative radiation therapy is that both the ment, the mandibulotomy combined with split primary site and neck receive combined modality provides excellent exposure with minimal mor- treatment. Historically, tonsillar fossa lesions bidity. One can fashion the mandibulotomy to have been treated with radiation therapy primar- preserve mental nerve continuity. A stairstep ily. However, several reports indicate that such mandibulotomy is planned medial to the mental an approach may decrease locoregional control.1,2 foramen. Prior to bony cuts, a 3D, high profile, 2.0 Given the improved results for reconstruction fol- mm plate is fashioned with three bicortical screws lowing surgical excision of tonsillar carcinomas, I on either side of the planned mandibulotomy. The favor an approach that involves primary surgery lip split incision is performed in the midline with followed by postoperative radiation therapy. An- a step at the vermilliocutaneous junction and a other factor in this treatment preference is the Z-plasty between this junction and the chin opportunity for an in-continuity neck dissection. crease. I prefer the chin to be split in the midline There is very little distance between the primary rather than around the chin subunit. If the pa- site and the level II lymph node on this CT scan. tient does not have tight dental junctions, one can On several occasions we have identified high ret- perform the mandibulotomy between teeth (usu- ropharyngeal nodes that were positive for squa- ally between the canine and lateral incisor). The mous carcinoma on histopathologic analysis. Re- lingual nerve can be preserved provided it is not section of these nodes followed by postoperative involved with tumor, although this does limit the radiation therapy seems to provide the best swing of the mandible. Radical resection of tumor chance for locoregional control of disease. Fur- can then be accomplished with 2-cm margins and thermore, histopathologic analysis of the lymph frozen section control. nodes provides an opportunity to enroll patients The neck dissection I would recommend for a in randomized prospective protocols comparing 4-cm jugulodigastric node is a modified versus a radiation therapy to chemotherapy plus radiation radical neck dissection. I would plan to sacrifice therapy for high-risk pathology (numerous posi- the sternocleidomastoid muscle. If cranial nerve tive nodes, extracapsular extension, close surgical XI is not involved with tumor it can be saved. margins). Direct tumor involvement would dictate sacrifice of the XIth cranial nerve. The internal or external Dr. Futran: For this particular patient and tu- jugular vessels can be saved for microvascular mor, external beam radiation therapy to a mean anastomosis only if free from tumor. If not, con- total dose of 6,500 centigray to the primary site tralateral vessel dissection (usually performed in and both necks, will provide an 80% chance of the setting of a limited selective neck dissection) local control and an excellent functional result. would be required if free tissue transfer is planned. The patient would also have a planned postradia- Optimum reconstruction for this anticipated 7 tion right modified radical neck dissection. Al- cm×7cmdefect of the tonsillar fossa and sur- though primary surgical extirpation and neck dis- rounding region could be achieved with either a section is an acceptable option followed by platysma myocutaneous flap or a radial forearm radiation therapy, the defect would likely require free fasciocutaneous flap. The platysma myocuta- a flap reconstruction and possible resection of the neous flap, however, is not advisable in the set- angle of the mandible. Although the functional ting of modified neck dissection when the sterno- result with proper reconstruction would be quite cleidomastoid muscle is not preserved.3 The good, if radiation therapy alone controls the pri- rationale for this caveat is lack of carotid artery

Management of Tonsil Cancer HEAD & NECK October 1999 659 protection due to coverage with only a thin cuta- primary nodal presentation, not specifically on neous flap. Therefore, a radial forearm free fas- the response to treatment. Several studies, par- ciocutaneous flap would be ideal for reconstruc- ticularly in the European literature, suggest that tion of this defect. Presumably the patient would even with complete clinical response of the neck lose approximately half of the . Func- to radiation therapy and chemotherapy protocols, tional reconstruction can be achieved by narrow- occult disease may still be present. Until there is ing the nasopharyngeal port with primary closure evidence that posttherapy necks that are clini- of the posterior aspect of the tonsillar/soft palate cally and radiographically negative have no occult defect (unpublished data, Chepeha DB and Tek- disease, a neck dissection is performed. Further, nos TN). The radial forearm free flap can then be the patient is aware at the time of primary treat- sutured into the anterior aspect of the defect. Uti- ment planning, that a neck dissection will be done lizing this technique has allowed functional res- regardless of the response. toration without velopharyngeal insufficiency in a series of patients with advanced tonsillar carci- Dr. Peters: If there was complete resolution of noma. the neck disease on physical examination and CT, Full course radiation therapy would be ad- I would not do a neck dissection. I would follow vised to commence 4 to 6 weeks postoperatively. the patient closely, repeating physical examina- Evidence of extracapsular extension on nodal pa- tions every 4 to 6 weeks for the first year and thology would make the patient eligible for ran- obtain another contrast CT scan in 6 months. domized protocols comparing radiation alone with chemotherapy plus radiation. Dr. Bradford: For nodal disease greater than 3 cm in greatest dimension, I favor a planned modi- Dr. Futran: First, I would assess whether the fied or radical neck dissection 6 weeks after ra- tumor is fixed to the mandible. If there is any diation therapy. I cannot comment on what type adherence to the mandible at the angle region, to of neck dissection to perform following chemora- get a proper margin, I would perform a composite diation since no prospective randomized trial has resection including the angle of the mandible via proven this treatment modality to be as effective lip split approach. If the tumor is not adherent to as surgery and postoperative radiation therapy the mandible, I would approach this patient by lip for advanced stage tonsillar or oropharyngeal split and midline mandibular split approach, re- squamous cell carcinoma. However, for advanced sect the primary tumor in continuity with a modi- laryngeal and hypopharyngeal carcinomas treated fied radical neck dissection, sparing only the spi- according to published protocols,4,5 planned neck nal accessory nerve. The optimal reconstruction dissections are advisable for residual palpable would be with a sensate radial forearm fasciocu- nodal disease following induction chemotherapy.6 taneous free flap if the mandible is spared. If the angle of the mandible is resected, I would use a 6. If definitive radiation or chemoradia- radial forearm osteofasciocutaneous sensate free tion resulted in a complete locoregional re- flap. sponse, then at 6 months a primary site re- currence occurred that was limited to the Dr. Peters: I would approach the primary tu- original primary tumor site, what salvage mor with a mandible sparing procedure such as a surgical procedure would you recommend? median mandibulotomy. The neck disease is fairly bulky and I feel it would be difficult to pre- Dr. Peters: If a recurrence developed at the site serve his spinal accessory nerve and the internal of the original primary and he was still without jugular vein. Therefore, I feel a classic radical evidence of recurrence in the neck, then I would neck dissection would be the operation of choice. resect the tonsil recurrence using a median man- dibulotomy and also perform an ipsilateral modi- 5. If definitive radiation or chemoradia- fied neck dissection. The reasons for the neck dis- tion were used first, and a complete locore- section include the need to sample the involved gional response was observed, would you nodes to rule out neck recurrence, “violation” of perform a planned neck dissection? the neck in resecting the recurrence in the oro- , and the need to provide recipient vessels Dr. Futran: Our philosophy at the University for the radial forearm flap that I would use for of Washington is to treat the neck based on the reconstruction.

660 Management of Tonsil Cancer HEAD & NECK October 1999 Dr. Bradford: If definitive radiation or chemo- on a treatment protocol, and if this were not an radiation resulted in a complete locoregional re- option, then accelerated fraction radiation therapy. sponse and 6 months later a primary site recur- He emphasizes the necessity for close monitoring rence occurred, I would perform the originally for an objective response after two cycles of che- planned surgical resection and reconstruction motherapy or 50 Gy of radiation with immediate (see #4 above). Clearly, in this instance free tissue salvage for inadequate responses. His rationale transfer is recommended due to the poor wound for this approach is that the best functional out- healing expected 6 months after radiation and/or come is achieved without compromising oncologic chemoradiation. control. Dr. Bradford favors surgical resection fol- lowed by radiation therapy and possibly adjuvant Dr. Futran: At this juncture, surgical extirpa- chemotherapy based on histopathologic param- tion is the only potential curative modality left. eters. She feels that combined modality treat- Therefore, the patient would require a resection ment to both the primary site and necks provides of the original tumor volume with appropriate superior locoregional control compared with de- margins by a lip splitting approach, which would finitive radiation therapy, provides objective his- include the angle of the mandible, tonsillar fossa, topathologic criteria for entry into adjuvant and necessary contiguous structures with frozen therapy protocols, and that sophisticated recon- section margin control and a modified neck dis- struction of the defects have reduced the surgical section if it hadn’t been done previously on the morbidity. Dr. Futran favors external beam ra- ipsilateral side. The defect would be reconstructed diation therapy with standard fractionation, with an osteofasciocutaneous sensate radial fore- along with a planned modified radical neck dis- arm free flap. The bone requirement is small and section. He feels that radiation therapy will pro- it is in a nontooth bearing area. The soft tissue vide a high likelihood of local control for this pa- attributes of the radial forearm flap make it an tient, and that the functional outcome will be ideal choice for this defect. Since this patient is a better than primary surgery. high risk for further recurrence, once the wounds Another alternative not discussed is treat- have sufficiently healed, consideration would be ment with concomitant chemoradiation. A re- made for the use of weekly methotrexate therapy cently published randomized trial comparing for 4–6 months. radiation versus concomitant chemoradiation7 demonstrated a significant improvement in local COMMENT control rates and disease-specific survival in a se- Definitive radiation therapy with surgical salvage ries of 100 patients with squamous cell cancer of or surgery and postoperative radiation therapy the head and neck. However, subsite analysis re- remain the mainstays of treatment for tonsillar vealed these differences to be significant for the squamous cell carcinoma. All of the consultants larynx and hypopharyngeal sites, with no differ- agree that chemoradiation treatment regimens ence in local control or disease survival for pa- should be performed in a protocol setting. There tients with orophyaryngeal cancer. are some tumor characteristics that may be help- If surgery were the given option, all the con- ful in selecting treatment, assuming the patient sultants would approach this tumor through a lip does not have any treatment biases. Factors that splitting incision and median or parmedian man- are felt to predict radiation responsiveness are dibulotomy and preserve the mandible unless in- tumors limited to the tonsillar fossa, superficial volved with tumor. This provides the best expo- or exophytic tumors, and poorly differentiated or sure, and preservation of mental nerve and lymphoepithelioma histology. Anterior pillar le- possibly lingual nerve function. None of the au- sions with retromolar trigone extension, signifi- thors would perform a selective neck dissection cant tongue base extension, or trismus are pref- sparing level V for a solitary 4-cm nodal metasta- erably treated surgically (Dr. Futran). Dr. Peters sis. The sensate radial forearm free flap is the bases his decision making primarily on clinical reconstructive option of choice. stage, and does not feel that a reliable biologic or The decision to perform a neck dissection in a molecular marker has been identified to help with neck that was originally clinically N+ that after treatment planning. radiation or chemoradiation is N0 is important, The authors vary considerably in their treat- since several studies have shown that there is a ment approach for this particular patient. Dr. Pe- 25%–30% incidence of residual disease in the ters favors chemotherapy and radiation therapy neck, even in N0, and failure rates for late sur-

Management of Tonsil Cancer HEAD & NECK October 1999 661 gical salvage of recurrent regional disease in this care center, particularly in patients from a rural situation are significant. Dr. Futran advocates setting. In addition, definitive prospective studies planned, post-therapy neck dissection in all pa- comparing oncologic outcomes, costs, functional tients with N+ disease at the time of primary outcomes, and quality of life are needed. treatment planning. Dr. Peters is willing to closely follow a patient if there is complete reso- lution of the neck disease both clinically and ra- REFERENCES diographically, while Dr. Bradford recommends a 1. O’Brien CJ, Castle GK, Steven GN, et al. Limitations of planned neck dissection for all patients with radiotherapy in the definitive treatment of squamous car- nodal disease originally 3 cm or greater. The role cinoma of the tonsillar fossa. Aus NZ J Surg 1992;62: 709–713. of a selective neck dissection in the posttreatment 2. Hicks WL, Kuriakose A, Loree TR, et al. Surgery versus neck that was originally N+ and is now NO re- radiation therapy as a single-modality treatment of tonsil- mains to be defined. lar fossa carcinoma: the Roswell Park Cancer Institute ex- perience. Laryngoscope 1998;108:1014–1019. In the event of a local recurrence 6 months 3. Esclamado RM, Burkey BB, Carroll WR, et al. The pla- after treatment, all of the authors are in agree- tysma myocutaneous flap: indications and caveats. Arch ment with a lip splitting approach, neck dissec- Otolaryngol Head Neck Surg 1994;120:32–35. 4. The Department of Veterans Affairs Laryngeal Cancer tion, and free flap reconstruction. Dr. Futran Study Group. Induction chemotherapy plus radiation com- would include segmental bony resection, while pared with surgery plus radiation in patients with ad- Drs. Bradford and Peters would spare the man- vanced laryngeal cancer. New Engl J Med 1991;324: 1685–1690. dible if uninvolved through a mandibulotomy ap- 5. Lefebvre J-L, Chevalier D, Luboinski B, et al. Larynx pres- proach. Drs. Bradford and Futran would resect at ervation in pyriform sinus cancer: preliminary results of a least the volume of the original tumor with mar- European Organization for Research and Treatment of Cancer Phase III trial. J Natl Cancer Inst 1996;88: gins. 890–899. The management of tonsillar carcinoma re- 6. Wolf GT, Fisher SG. Effectiveness of salvage neck dissec- mains challenging and without a uniform consen- tion for advanced regional metastases when induction che- motherapy and radiation are used for organ preservation. sus. Many other factors integrate into the deci- Laryngoscope 1992;102:934–939. sion making process, such as patient preferences 7. Adelstein DJ, Lavertu P, Saxton JP, et al. Mature results of and biases, surgeon and institutional philoso- a phase III randomized trial comparing concurrent chemo- radiotherapy with radiation therapy alone in Stage III and phies, geographical distances and social restric- IV squamous cell head and neck cancer. J Clin Oncol [Sub- tions involved with multiple visits to a tertiary mitted].

662 Management of Tonsil Cancer HEAD & NECK October 1999