1

Introduction

Tonsillar carcinomas are the most common cancers of the oropharynx; this subsite comprises approximately 70% to 80% of the cancers of this region . Early stage tonsil cancers are generally limited to the tonsillar complex, which involves the anterior and posterior tonsillar pillar and the superior constrictor muscle. Although size appears to be the determinant of T stage as per the American Joint Committee on Cancer , local spread has been shown to be a significant prognostic indicator particularly in predicting response to treatment . Clinically apparent cervical metastases in tonsillar cancers range from 66% to 76% at presentation . The presence of occult neck disease in the clinically N0 neck is estimated at 24% to 35% . Consequently, the guidelines by the American Head and Neck Society suggest neck dissection or irradiation for oropharyngeal cancers . Several reports suggest that human papillomavirus (HPV) has largely redefined this disease with recent estimates reporting a rising incidence of HPV-positive tonsil cancers . The implications of this shift remain to be determined; however, evidence suggests that HPV-positive pathology may confer a positive effect on prognosis .

Early stage (I and II) tonsil cancer can be managed using single-modality treatment—either surgery or radiation alone—with equivalent survival outcome as demonstrated in multiple retrospective reviews . Many have favored radiation as a treatment given its ability to address retropharyngeal and parapharyngeal nodes , areas not addressed in conventional neck dissections. Historically, a combination of chemotherapy and radiation therapy was preferred for advanced-stage cancers largely because of the functional implications involved in initial surgical approaches . In addition, when these patients undergo surgery, postoperative treatment intensification is often required, which further influences primary treatment selection. In the past few years, reports have renewed interest in transoral resection of tonsillar cancers, both early and late stage ; this renewed interest chiefly lies in the improvement in preoperative imaging and ability to resect these cancers with minimal functional sequelae.

For more than a decade at the Cleveland Clinic, we have treated patients with early T stage tonsil cancers via a primary surgical approach, radiation alone, or combined adjuvant treatment contingent upon the extent of nodal disease. Until the recent widespread use of intensity-modulated radiation therapy in the management of these tumors, surgery was used as an alternative to radiation in attempts to avoid the xerostomia associated with conventional 3-field technique and to pathologically stage the neck. Because of a variety of reasons, however, many surgical patients ultimately receive adjuvant treatment, which potentially offsets the benefit of selecting surgery up front, unless radiation dosages are decreased in the postoperative setting. Our first objective was to identify whether offering patients surgery, as intended single modality at initial presentation, would achieve equivalent survival rates and locoregional control (LRC) rates to patients pursuing radiation treatments as intended single modality. Secondly, we sought to identify which patients staged with early T stage tonsil cancer could avoid the need for postoperative adjuvant treatment. Examining these objectives will help to determine the implications of low T stage tonsil cancer on initial management decisions.

2

Materials and methods

Approval for this retrospective analysis was obtained from the Cleveland Clinic Institutional Review Board. We identified patients from the Cleveland Clinic Head and Neck tumor registry who were treated for T1 and T2 tonsillar cancers. This registry has prospectively collected data on all patients treated for head and neck malignancies at our institution since January 1, 1997. Patients included in this study are those registered into the database up until September 1, 2008, allowing sufficient time for patient follow-up. We retrospectively reviewed the charts of patients treated for T1 and T2 tonsillar cancers with histologically proven squamous cell carcinoma. We excluded patients with N2b to N3 disease, as these patients were not felt to be adequate surgical candidates and would typically be treated with concurrent chemoradiation therapy, reserving surgery for salvage. Other exclusion criteria were the presence of distant metastatic disease; previously treated head and neck cancers; and concomitant cancers, either second primaries or other coexisting malignancies. The treatment selection of either the surgery or radiation arm was based on the recommendation of a multidisciplinary tumor board, taking into consideration surgeon preference, patient comorbidities, patient preference, and clinical status of neck disease.

All radical tonsillectomies were performed at the Cleveland Clinic with curative intent as defined by negative intraoperative margins. Radical tonsillectomies involved en bloc resection of the tonsil, anterior and posterior tonsillar pillars, and underlying superior constrictor muscle as a deep margin. Resection took place under loupe magnification with electrocautery. Depending on the extent and location of the tumor, adjacent soft palate or tongue base was taken for margin control. A minority of patients (3) required a mandibulotomy for adequate exposure to the oropharynx. Reconstructive efforts were based on the extent of the defect and included healing by secondary intent, closure with acellular dermis, locoregional flaps, and free flaps. All patients (17/17) underwent a staging neck dissection, which involved an ipsilateral selective neck dissection involving at least levels 2 to 4. Indications for adjuvant radiation included the presence of extracapsular nodal spread and close margins at the primary site. Radiation was also used for local and regional recurrences for purposes of salvage. Patients with extracapsular nodal spread and locoregional recurrences are typically given concurrent chemotherapy unless it is medically contraindicated. Over the course of this study, there was also some variation in the total dose, which ranged from 58 to 72 Gy, reflecting the differences in indications from postoperative adjuvant treatment to salvage. In our surgical group, the only patients receiving concurrent chemotherapy were those with locoregional recurrence.

For the radiation group, all radiation treatments were performed at the Cleveland Clinic. There was some heterogeneity with respect to dose administered ranging from 68.8 to 74.4 Gy. Again, the technique varied from conventional 3-field radiation to intensity-modulated radiation therapy when patients were treated with radiation for intended single-modality treatment—a majority of patients receiving external beam radiation. For the conventional approach, opposed lateral fields were used weighted 3:2 toward ipsilateral side; the fields were brought off the spinal cord at 42.0 Gy and then coned down at 54.0 and 64.0 Gy. Subsequent neck dissections were performed typically 12 weeks after completion of treatment at the recommendation of a multidisciplinary tumor board after assessing tumor response by examination; computed tomographic imaging; and, more recently, positron emission tomography/computed tomography.

The patient data collected included age at original diagnosis, smoking and drinking history, clinical stage, pathologic stage, HPV status, histologic features, primary treatment modality, adjuvant treatments, indications for adjuvant treatments, date of treatment completion, surgical procedure(s) performed, type of reconstruction, use of nasogastric (NG) feeding tube or gastrostomy (G) tube, date of recurrence either at the neck or primary location, and date of last follow-up or death. All patients were staged according to the American Joint Committee on Cancer criteria . Because many of these tumors were diagnosed and treated before routine HPV testing, which commenced in 2006 for all oropharyngeal cancers at our institution, HPV typing is not present for all patients and, thus, was not incorporated in the analyses. Human papillomavirus status was tested by in situ hybridization .

2

Materials and methods

Approval for this retrospective analysis was obtained from the Cleveland Clinic Institutional Review Board. We identified patients from the Cleveland Clinic Head and Neck tumor registry who were treated for T1 and T2 tonsillar cancers. This registry has prospectively collected data on all patients treated for head and neck malignancies at our institution since January 1, 1997. Patients included in this study are those registered into the database up until September 1, 2008, allowing sufficient time for patient follow-up. We retrospectively reviewed the charts of patients treated for T1 and T2 tonsillar cancers with histologically proven squamous cell carcinoma. We excluded patients with N2b to N3 disease, as these patients were not felt to be adequate surgical candidates and would typically be treated with concurrent chemoradiation therapy, reserving surgery for salvage. Other exclusion criteria were the presence of distant metastatic disease; previously treated head and neck cancers; and concomitant cancers, either second primaries or other coexisting malignancies. The treatment selection of either the surgery or radiation arm was based on the recommendation of a multidisciplinary tumor board, taking into consideration surgeon preference, patient comorbidities, patient preference, and clinical status of neck disease.

All radical tonsillectomies were performed at the Cleveland Clinic with curative intent as defined by negative intraoperative margins. Radical tonsillectomies involved en bloc resection of the tonsil, anterior and posterior tonsillar pillars, and underlying superior constrictor muscle as a deep margin. Resection took place under loupe magnification with electrocautery. Depending on the extent and location of the tumor, adjacent soft palate or tongue base was taken for margin control. A minority of patients (3) required a mandibulotomy for adequate exposure to the oropharynx. Reconstructive efforts were based on the extent of the defect and included healing by secondary intent, closure with acellular dermis, locoregional flaps, and free flaps. All patients (17/17) underwent a staging neck dissection, which involved an ipsilateral selective neck dissection involving at least levels 2 to 4. Indications for adjuvant radiation included the presence of extracapsular nodal spread and close margins at the primary site. Radiation was also used for local and regional recurrences for purposes of salvage. Patients with extracapsular nodal spread and locoregional recurrences are typically given concurrent chemotherapy unless it is medically contraindicated. Over the course of this study, there was also some variation in the total dose, which ranged from 58 to 72 Gy, reflecting the differences in indications from postoperative adjuvant treatment to salvage. In our surgical group, the only patients receiving concurrent chemotherapy were those with locoregional recurrence.

For the radiation group, all radiation treatments were performed at the Cleveland Clinic. There was some heterogeneity with respect to dose administered ranging from 68.8 to 74.4 Gy. Again, the technique varied from conventional 3-field radiation to intensity-modulated radiation therapy when patients were treated with radiation for intended single-modality treatment—a majority of patients receiving external beam radiation. For the conventional approach, opposed lateral fields were used weighted 3:2 toward ipsilateral side; the fields were brought off the spinal cord at 42.0 Gy and then coned down at 54.0 and 64.0 Gy. Subsequent neck dissections were performed typically 12 weeks after completion of treatment at the recommendation of a multidisciplinary tumor board after assessing tumor response by examination; computed tomographic imaging; and, more recently, positron emission tomography/computed tomography.

The patient data collected included age at original diagnosis, smoking and drinking history, clinical stage, pathologic stage, HPV status, histologic features, primary treatment modality, adjuvant treatments, indications for adjuvant treatments, date of treatment completion, surgical procedure(s) performed, type of reconstruction, use of nasogastric (NG) feeding tube or gastrostomy (G) tube, date of recurrence either at the neck or primary location, and date of last follow-up or death. All patients were staged according to the American Joint Committee on Cancer criteria . Because many of these tumors were diagnosed and treated before routine HPV testing, which commenced in 2006 for all oropharyngeal cancers at our institution, HPV typing is not present for all patients and, thus, was not incorporated in the analyses. Human papillomavirus status was tested by in situ hybridization .