Surgical Approaches to the Oropharynx

With the sharp increase in the prevalence of HPV-associated oropharyngeal SCC, there has been a corresponding shift in the affected population to younger, nonsmoking patients. As discussed, the recent treatment paradigm largely consisted of intensive CRT protocols, just as for HPV− cancers, but with dramatically improved OS; this was a result of late recognition that a substantial portion of patients entered into trials had a new, more treatable form of SCCHN—caused by HPV. The increased risk for these younger patients of long-term sequelae, such as xerostomia, osteoradionecrosis, dysphagia, fibrosis, aspiration, and radiation-induced malignancy, in conjunction with newly available technology, led head and neck surgeons to explore minimally invasive surgical techniques in an effort to diminish the morbidity of surgical resection and reduce the need for high-dose CRT.

Minimally invasive surgical techniques have greatly expanded the scope of transoral approaches for oropharyngeal resections. Traditionally, transoral surgery was reserved for cancers of the tonsil, soft palate, and posterior pharyngeal wall less than 2 cm in greatest dimension. Larger neoplasms or tumors of the base of tongue were difficult to expose, especially at the deep margins. The development of endoscopes greatly improved transoral inspection of the oropharynx by magnification, tissue distention, and angled optics. In addition, advances in laser delivery systems and the wristed technology of the da Vinci robot have allowed increased freedom of motion in previously inaccessible (or intractable) subsites. These innovations have enabled surgeons to remove much larger neoplasms from the oropharynx with excellent oncologic control and functional outcomes.

Although most of the supporting evidence for transoral surgery is in early glottic cancer, there have been promising results with oropharyngeal cancer over the last decade. In 2003, Steiner and colleagues reported on a cohort of 48 patients with oropharyngeal cancers treated with transoral laser microsurgery (TLM), of which 94% were stage III/IV. Only 23 of the 48 patients required RT for either deep extension into the tongue base or positive resection margins. The 5-year local control for all T stages was 85% with no local recurrences in the T1 or T2 group and a 20% local recurrence rate for T3/T4 tumors. OS and DFS were 52% and 73%, respectively.

A recent, multicenter, prospective study of primary, TLM-resected, stage III/IV oropharyngeal cancers reported 204 patients. Two- and 5-year estimates demonstrated an OS of 89% and 78%, disease-specific survival of 91% and 84%, and DFS of 85% and 74%, respectively. The survival results are comparable to those expected using CRT as the primary treatment modality, with improved functional outcomes. A recent CRT study (n = 71) with 32% of the tumors staged T3 and T4 showed a 3-year OS of 83% and an LRC rate of 90%. In the TLM study, with 34% T3 and T4 tumors, the OS was 86% at 3 years and LRC of 93%. The matched CRT group in this report had a long-term G-tube rate of 35% compared with 3.4% in the TLM study.

Publications also indicate that TLM-based surgical treatment without adjuvant therapy can be used with excellent oncologic control for early T-stage oropharyngeal cancer. In 2009, Grant and colleagues reported 69 cases of oropharyngeal SCC staged T1–T4 (T1 = 25, T2 = 30, T3 = 12, T4 = 2) treated with TLM and a neck dissection, if indicated. None of the patients received adjuvant RT, either because it was not indicated or because the patient declined. Five-year OS was 86% with LRC of 90% for T1 and 94% for T2. None of the T3 or T4 patients had a local recurrence. In terms of functional outcomes, none of the patients required a long-term feeding tube and 98% reported normal or near normal swallowing, using the Salassa scale.

In addition to TLM, the feasibility of transoral robotic surgery (TORS) was investigated, first using a human cadaver, followed by a supraglottic laryngectomy in a canine model. Protocols were then established for positioning, use of 5-mm instruments, and hemostasis. The first documented oropharyngeal TORS series reported 3 patients with early T stage, base of tongue squamous cell carcinoma. En bloc resection with clear margins was achieved in all cases with no complications or measured change in swallowing function.

In a 2007 case series of 27 patients with tonsillar SCC who underwent primary resection with TORS, 24 patients were stage III/IVa. Neck dissections were performed on 26 of the 27 patients and 24 had at least one positive node. The cervical lymphadenectomy was performed 1 to 3 weeks after the TORS procedure in an effort to reduce lymphedema in the oropharynx after TORS. Tumor-free margins were achieved in all of the cases and there were no locoregional recurrences during the 2-year study period (with a minimum of 6-month follow-up). All of the patients who returned for evaluation (26 of 27) tolerated a regular diet without the need for a gastrostomy tube. One patient, with preexisting sleep apnea, required a tracheostomy due to its exacerbation by the surgical procedure.

Since 2007, several institutions have published series highlighting the efficacy of TORS for oropharyngeal SCC. These reports document excellent oncologic outcomes in series of 20 to 47 patients. Moore and coworkers reported that 12 of 45 patients (27%) needed no adjuvant RT, and at 1-year follow-up no local recurrences were detected. Weinstein and coworkers reported a series of 47 patients with stage III/IV oropharyngeal SCC. One patient had tumor at a margin of 2%. Disease-specific survival was 98% at 1 year and 90% at 2 years. After the TORS procedure, 18 of 47 patients (38%) needed no chemotherapy, and 5 patients (11%) did not require any adjuvant therapy. In a prospective, phase I single-arm study, 31 patients underwent TORS with selective neck dissection with a 100% disease-control rate and a 2-year gastrostomy dependency rate of 0%. Seven patients required no adjuvant therapy and 15 of the remaining 24 (63%) received RT alone (60 Gy to the primary site and dissected neck, 54 Gy to the contralateral neck and retropharyngeal nodes) based on the lack of nodal ECE. After recommended postoperative adjuvant therapy using contemporary practice standards, a substantial number of these patients were managed without adjuvant CRT, and sometimes without any RT whatsoever. Such results are contingent on appropriate case selection for transoral resection.

Functional outcome studies in patients with oropharyngeal cancer undergoing TORS resections have reported superior swallowing results when compared with published, nonsurgical protocols. In a direct comparison of TORS versus CRT with a matched group of patients with oropharyngeal cancer at a single institution, TORS was associated with better short-term eating ability (swallowing score 72 vs 43, P = .008), diet tolerance (43 vs 25, P = .01), and functional oral intake score (FOIS) (5.5 vs 3.3, P <.001) 2 weeks after completion of treatment, an index of acute treatment toxicity. At 12 months, the TORS group had returned to baseline swallowing status, whereas diet and FOIS remained impaired in the CRT group. In their report of 45 patients, Moore and coworkers found that only 5 of 22 patients who had gastrostomy tubes required their use beyond 20 days and all had their feeding tubes removed within 4 months. Dean and coworkers compared feeding tube dependence in patients undergoing TORS with those who had open surgery for post-CRT oropharyngeal SCC failures and found that although 43% of the open resection group were G-tube-dependent at 6 months, none of the TORS patients required feeding tube supplementation. Such comparisons are affected by case selection, and randomized trials will be required to ascertain whether patients experience differences in outcome when treated with either best current surgical or nonsurgical approaches. Such studies are currently underway.

Although much attention has been devoted to improved outcomes of oropharyngeal SCC in the HPV era, there remains a significant portion of patients at risk for local relapse and other causes of limited survival. This cohort is primarily represented by those with a significant smoking history (>10 pack-years) whose tumors are not caused by HPV. Surgical resection is one potential means of augmenting current CRT with a potential to improve local control and OS. Interestingly, a SEER (Surveillance, Epidemiology, and End Results) analysis of early stage tonsillar patients treated with a nononcologic tonsillectomy in addition to standard therapy suggested improved survival even when controlling for age and decade of treatment. Perhaps surgical reduction of tumor volume, even with a nononcologic resection, improved patient outcomes. This finding supports a hypothesis that surgical resection could potentially benefit high-risk cohorts. Although traditional surgical approaches have been associated with significant morbidity, TORS offers an opportunity to intensify local therapy, delivering an oncologic procedure in combination with well-established adjuncts, including CRT. Efforts to establish the proposition through prospective trials have been frustrated by poor accrual.

Radiation Therapy

Definitive RT schedules for SCCHN were studied extensively before the acceptance of combined modality regimens. Altered fractionation regimens relied on manipulating treatment time to exploit radiobiological principles such as reassortment of cells into the sensitive phase of the cell cycle (hours), repair of normal tissues (hours), reoxygenation (hours to days), and repopulation of tumor cells (weeks). In general, conventional dose and fractionation consist of 2.0 Gy per day to a total dose of 70.0 Gy. Pure hyperfractionation is delivery of the same total dose in the same overall treatment time but with more fractions (eg, 2 fractions per day). Total dose must be increased to account for the lower dose per fraction to achieve equivalent local control. When hyperfractionation is applied clinically, typically 2 fractions are delivered per day (with a 6-hour interfraction interval to allow for normal tissue repair). A common “hyperfractionation” schedule is 81.6 Gy delivered in 1.2-Gy fractions over about 7 weeks. This schedule allows for a higher total dose with a minimal increase in late toxicity because of the lower dose per fraction. Another approach, acceleration with concomitant boost, delivers conventional fractionation during the first portion of treatment, while the final 12 treatments are delivered twice per day. This approach effectively shortens the overall time to deliver a curative dose, a treatment strategy designed specifically to address tumor repopulation. Other approaches have been developed to accelerate the treatment course and address repopulation, including a 6 fraction per week strategy used in DAHANCA 7. Clinical trials reveal that hyperfractionation and accelerated fractionation regimens have improved LRC and, in some trials, also survival. In a meta-analysis of 15 phase III trials with more than 6000 patients, altered fraction RT was associated with a 3.4% absolute benefit in 5-year OS when compared with conventional fractionation. Although altered fractionation in combination with chemotherapy increases toxicity without a statistically significant improvement in clinical outcomes in comparison with standard fractionation, definitive RT-alone regimens (including altered fractionation) are again being proposed as a deintensification strategy.

In addition to the improved understanding of the radiobiology of SCCHN, there have been technological improvements in the delivery of RT. The most widely used technique for improving the therapeutic ratio of RT is intensity-modulated radiation therapy (IMRT). IMRT uses fundamentally different means of radiation planning and delivery when compared with conventional head and neck RT. Specifically, conventional RT combines 3 or 4 large fields in which the RT dose to most normal structures (eg, parotid glands, pharyngeal walls) is the same as the dose to the target volumes of the tumor and regional lymphatics. IMRT combines numerous small and irregularly shaped radiation beams of variable intensities to achieve a radiation dose distribution that is highly conformal to the target volumes. In addition, accelerated hyperfractionation, which was previously accomplished by delivered 2 fractions per day for the last 12 treatments, is now possible using a one fraction per day approach with IMRT. Here, the dose delivered is varied based on disease burden using a concept described as a simultaneous integrated boost (SIB). In the case of an SIB, the radiation dose per fraction varies throughout the disease risk areas from 1.6 to 2.2 Gy per fraction. Completing the definitive treatment course in 30 total fractions in comparison with the more common 35 fractions accelerates the treatment.

Although IMRT improves the dosimetry of RT, its effects on clinical outcomes are still being investigated. Small randomized trials of IMRT versus standard RT for nasopharynx cancer did not show a significant QoL benefit (despite showing significantly increased parotid gland flow with IMRT). In the UK PARSPORT trial, conventional RT was compared prospectively with that of IMRT in regards to parotid sparring and subsequent reduction of xerostomia in patients with pharyngeal carcinoma. The findings from this study at 12 and 24 months revealed that grade 2 or worse xerostomia was significantly lower in the IMRT group than in the conventional RT group. These findings were attributed to benefits in recovery of saliva secretion with IMRT compared with conventional treatment.

Systemic Therapy

The identification of HPV as a prognostic factor in oropharyngeal malignancies has promoted efforts to stratify patients into various risk categories; this has the potential to influence treatment. There may be patients that would benefit from deintensification to ameliorate toxicity (without a decline in survivorship), whereas others have a more guarded prognosis. In addition to its prognostic significance, HPV positivity predicts response to chemotherapy. As noted earlier, in ECOG 2399, patients with HPV+ cancers had higher response rates after induction chemotherapy (82% vs 55%, P = .01). OS at 2 years was significantly better in patients with HPV+ cancers compared with patients with HPV− cancers (95% vs 62%; P = .005). In ECOG 1308, complete clinical response at the primary site was used as a dynamic response biomarker; complete responders were treated with a 20% reduction in radiation dose (54.0 Gy vs 69.3 Gy). These findings support investigation of induction therapy as a means of selecting HPV+ patients that may benefit from dose-reduced IMRT.

The development of clinical risk classification for oropharyngeal SCC, integrating HPV and tobacco status, has also framed investigational priorities for high-risk disease. Deintensification strategies are being tested in patients with good risk HPV+ SCCHN (ECOG 1308; RTOG 1016; ECOG 3311); intensification approaches represent the major unmet need for HPV− disease, where survivorship has not improved despite increasingly aggressive application of CRT in a variety of means. In addition, most head and neck oncologists argue that intermediate-risk HPV+ disease demonstrates PFS unacceptable for deintensification (2-year PFS 60%–70%, based on RTGO 0129). The intermediate group identified in RTOG 0129 consists of those with HPV+ disease, greater than 10 pack-years of smoking along with N2b-N3 disease. In addition, those with HPV− disease along with a less than or equal to 10 pack-year history of smoking and T2–3 disease comprised this intermediate subset. Recent data also indicate intermediate PFS in patients with HPV+ disease and T3–4 tumors or N3 nodal disease, irrespective of smoking status. These subgroups are also viewed as suboptimal for deintensification.

The unexpected result of RTOG 0522, where the addition of cetuximab to cisplatin-RT did not improve PFS or OS in locally advanced HPV+ or HPV − SCCHN, prompted experiments in radiobiology to identify non-cross-resistant agents that may successfully increase response, and not simply toxicity. In preclinical models of SCCHN, cetuximab enhances the activity of docetaxel-RT but is less effective at enhancing platinum-RT. Perhaps cetuximab and cisplatin radiosensitize through a similar mechanism (specifically the repair of DNA double-strand breaks through nonhomologous end-joining), whereas docetaxel (a taxane microtubule inhibitor) radiosensitizes through a different mechanism. In the adjuvant high-risk setting, this did not inform design of the RTOG 0234 randomized phase II trial of cisplatin-cetuximab-RT or docetaxel-cetuximab-RT. The primary endpoint of that phase II trial, comparing DFS with the historical control of the RTOG 9501 trial’s cisplatin-RT arm, was significantly improved only for the cetuximab-docetaxel arm (hazard ratio [HR] 0.69, P = .012). As a result, it is important to ascertain whether the improved DFS is attributed to the taxane alone, or to its combination with cetuximab. In the high-risk adjuvant setting, a randomized phase II/III study of cisplatin-RT versus docetaxel-RT versus docetaxel-cetuximab-RT is being conducted (RTOG 1216).

Additional evidence supporting the use of cetuximab in SCCHN is derived from the recurrent/metastatic setting. In platinum-refractory recurrent/metastatic SCCHN, cetuximab monotherapy has a response rate of approximately 11%, providing evidence that cetuximab exerts its effects via pathways distinct from DNA damaging agents such as cisplatin or RT. In first-line therapy for recurrent/metastatic SCCHN, the addition of cetuximab to fluorouracil (5-FU)/platinum significantly improved OS in the EXTREME study. This study was the first randomized study to demonstrate a benefit in OS (10.1 vs 7.4 months, HR = 0.797, P = .036) with a molecular-targeted therapy added to the classical platinum plus 5-FU combination in the first-line treatment of recurrent or metastatic SCCHN.

Targeted Therapy and Immunotherapy

In addition to multidrug regimens for high-risk patients, there is heightened interest in immunotherapy. SCCHN is an immunosuppressive disease; patients demonstrate lower absolute lymphocyte counts than healthy subjects, impaired activity of natural killer cells (the major effector cell of innate immunity), poor antigen-presenting function by dendritic cells, and impaired function of tumor infiltrating T lymphocytes (the major effector cell of adaptive immunity) with a strong impact on clinical outcome. With progressive insight into the genetics of SCCHN, novel targets likely to be of interest in the definitive management of oropharyngeal SCC are increasingly recognized. The PI3-Kinase/Akt/mTOR signaling pathway is the most commonly mutated oncogenic pathway in SCCHN, and patients with HPV+ cancers are disproportionately affected.

In addition, a new therapeutic approach for p53-deficient cancers, a nearly universal phenotype in SCCHN on the basis of TP53 mutation (a typical result of tobacco abuse) or viral degradation of p53 protein, is under development. This synthetic-lethal stratagem pairs a DNA-damaging agent, such as cisplatin or RT, with a cell cycle checkpoint inhibitor (Wee1 or Chk1) to precipitate mitotic catastrophe. Epidermal growth factor receptor (EGFR)-independent activation of intracellular proliferative signaling also suggests novel targets for drug development, including accessory receptor tyrosine kinases (alternate HER family members, cMet, insulin growth factor receptor, fibroblast growth factor receptor, or vascular endothelial growth factor receptor), downstream EGFR signaling nodes (RAS, PI3-Kinase, STAT3, SRC), or cell cycle promoters (aurora kinase, CDK4/6). These potentially promising approaches warrant active investigation.