Key points

Introduction

Head and neck cancer is the seventh most common cancer worldwide, with an estimated 686,328 new cases diagnosed in 2012. Most patients present with locally advanced disease (stage III or IVA-B). Nonsurgical management using radiation-based therapy is a definitive curative-intent treatment strategy in patients with unresectable disease and is often favored as an organ-sparing strategy. The benefit of chemotherapy delivered concurrently with radiation is well established in locally advanced head and neck cancer (LAHNC). The addition of chemotherapy concomitant with radiation improves survival and locoregional control compared with radiation alone. Multiple meta-analyses have shown the superiority of chemoradiation compared with radiation alone in LAHNC, with an absolute survival benefit at 5 years of 5% to 9% depending on anatomic location. Concurrent chemoradiotherapy (CRT) consisting of bolus cisplatin 100 mg/m ² is an established regimen in LAHNC.

Sequential chemoradiotherapy (ST) integrates induction chemotherapy followed by CRT with the goal of reducing distant metastases and maximizing local regional control and organ preservation. The optimal induction chemotherapy regimen was defined by 2 landmark randomized phase III trials, TAX 323 and TAX 324, which showed superiority of docetaxel, cisplatin, and 5-fluorouracil (TPF) compared with cisplatin/5-fluorouracil (PF) with respect to improved survival, progression-free survival (PFS), and locoregional and distant failure. Subsequently the GORTEC study in larynx cancer also confirmed the superiority of TPF compared with PF. Based on these results, TPF is the induction chemotherapy regimen for comparison in randomized trials evaluating ST versus CRT.

Several randomized clinical trials comparing head-to-head ST and CRT have had methodological limitations, including poor accrual, leading to equivocal results. Most recently, one randomized trial showed a survival advantage with ST compared with CRT, fueling further controversy and debate. This article discusses the latest evidence and state-of-the-art approaches to ST and CRT, highlighting ongoing controversies and future directions.

Introduction

Head and neck cancer is the seventh most common cancer worldwide, with an estimated 686,328 new cases diagnosed in 2012. Most patients present with locally advanced disease (stage III or IVA-B). Nonsurgical management using radiation-based therapy is a definitive curative-intent treatment strategy in patients with unresectable disease and is often favored as an organ-sparing strategy. The benefit of chemotherapy delivered concurrently with radiation is well established in locally advanced head and neck cancer (LAHNC). The addition of chemotherapy concomitant with radiation improves survival and locoregional control compared with radiation alone. Multiple meta-analyses have shown the superiority of chemoradiation compared with radiation alone in LAHNC, with an absolute survival benefit at 5 years of 5% to 9% depending on anatomic location. Concurrent chemoradiotherapy (CRT) consisting of bolus cisplatin 100 mg/m ² is an established regimen in LAHNC.

Sequential chemoradiotherapy (ST) integrates induction chemotherapy followed by CRT with the goal of reducing distant metastases and maximizing local regional control and organ preservation. The optimal induction chemotherapy regimen was defined by 2 landmark randomized phase III trials, TAX 323 and TAX 324, which showed superiority of docetaxel, cisplatin, and 5-fluorouracil (TPF) compared with cisplatin/5-fluorouracil (PF) with respect to improved survival, progression-free survival (PFS), and locoregional and distant failure. Subsequently the GORTEC study in larynx cancer also confirmed the superiority of TPF compared with PF. Based on these results, TPF is the induction chemotherapy regimen for comparison in randomized trials evaluating ST versus CRT.

Several randomized clinical trials comparing head-to-head ST and CRT have had methodological limitations, including poor accrual, leading to equivocal results. Most recently, one randomized trial showed a survival advantage with ST compared with CRT, fueling further controversy and debate. This article discusses the latest evidence and state-of-the-art approaches to ST and CRT, highlighting ongoing controversies and future directions.

Summary of recent randomized clinical trials comparing sequential chemoradiotherapy versus concurrent chemoradiotherapy

The head-to-head comparison of ST and CRT in LAHNC has recently been reported in 4 randomized phase III clinical trials ( Table 1 ). Notably, of the 4 studies, 3 were methodologically underpowered to detect a survival benefit. Two studies, PARADIGM and DeCIDE, lacked statistical power for their primary end point, overall survival (OS), because of early closure and a higher than expected survival in the control arm (3-year OS ∼78% compared with 55% at the time of study design) likely secondary to a large proportion of human papillomavirus (HPV)–related oropharynx cases. Only 1 trial was methodologically powered to detect a survival difference and this study showed a survival benefit favoring the ST arm compared with CRT.

PARADIGM is a randomized phase III trial of ST versus CRT in patients with LAHNC. In the ST arm, patients who achieved complete clinical response at the primary site following induction chemotherapy received weekly carboplatin (area under the curve, 1.5) and daily radiation fractionation, whereas the remaining patients received weekly docetaxel (20 mg/m ² ) with accelerated radiation therapy (RT). Patients in the CRT-alone arm received 2 cycles of bolus cisplatin with accelerated concomitant boost over 6 weeks (as in RTOG 0129). No differences in PFS (hazard ratio [HR], 1.07; confidence interval [CI], 0.59–1.92; P = .82) or OS (HR, 1.09; CI, 0.59–2.03; P = .77) were detected, although statistical power was limited because only 145 of the 300 planned patients enrolled. In an unplanned subset analysis, there were non–statistically significant trends noted favoring ST in nonoropharynx sites and CRT for the oropharynx sites. ST did not significantly reduce distant recurrence rates; however, distant recurrences were rare. Patients who did not achieve complete clinical response to induction chemotherapy seemed to have much poorer survival, suggesting that TPF may be used to identify biologically bad cancers and subsequent novel salvage regimens in this patient population are needed. No significant differences in toxicity were noted between the arms and no treatment-related deaths occurred in the study.

DeCIDE is a multicenter phase III study comparing CRT consisting of docetaxel, 5-fluorouracil (5-FU), hydroxyurea, and hyperfractionated radiotherapy (DFHX-RT) versus 2 cycles of TPF induction followed by DFHX-RT. Eligibility required N2 and N3 disease, and prior organ-sparing surgery was allowed. This study enrolled 280 patients (initially planned for 400 patients). No significant difference was identified between the 2 arms for the primary end point, OS (HR, 0.91; CI, 0.59–1.41; P = .68), or for recurrence-free survival or distant failure-free survival. There was no difference in survival between the 2 treatment arms in oropharynx patients or nonoropharynx patients, or among patients with N2a or N2b disease. The cumulative incidence of distant recurrence without prior locoregional recurrence favored the ST arm ( P = .043). Toxicities were higher in the ST arm, with increased severe neutropenia during CRT (14% vs 4%, P = .023) and more deaths during treatment (13 deaths [13.5%] in the ST arm versus 4 deaths [3.8%] in the CRT arm).

The Spanish Head and Neck Cancer Cooperative Group conducted a 3-arm phase III trial of 439 patients with previously untreated unresectable LAHNC comparing TPF followed by CRT versus PF followed by CRT versus CRT alone. This study was designed before the establishment of TPF as superior to PF. The primary end point was PFS and time to treatment failure (TTF). This study was not powered to detect survival differences. The initial abstract presentation reported a significant reduction in TTF favoring ST compared with CRT; however, subsequent published intent-to-treat analysis found no difference in PFS, TTF, or OS. Notably, CRT consisted of conventional bolus cisplatin every 3 weeks for all treatment arms. Deaths caused by study treatment toxicity were reported as 7 in the TPF ST arm, 4 in the PF ST arm, and 2 in the CRT arm, raising concerns for excessive toxicity with concurrent bolus cisplatin-radiation following induction chemotherapy. Forty-five patients (29%) in the TPF ST arm and 33 patients (21%) in the PF ST arm did not received CRT for reasons other than disease progression.

The GSTTC Italian Study Group is the only randomized phase III trial showing an improvement in OS with ST compared with CRT. This multicenter study of 421 patients with LAHNC was designed to evaluate (1) OS benefit of ST versus CRT, and (2) grade 3 to 4 in-field mucosal toxicity of CRT versus cetuximab-radiation. The 2 × 2 factorial design consisted of 4 arms as follows: (1) A1, CRT (PF × 2 concomitant to standard radiation fractionation); (2) A2, cetuximab concomitant to radiation; (3) B1, 3 cycles of TPF followed by CRT; (4) B2, 3 cycles of TPF followed by cetuximab-radiation. ST significantly improved PFS (HR, 0.73; 95% CI, 0.57–0.94; P = .015), and OS (HR, 0.72; 95% CI, 0.55–0.96; P = .025) compared with CRT. Superiority of ST compared with CRT was independent of type of CRT; however, in an unplanned subgroup analysis, greater OS benefit was observed with ST using concurrent cetuximab-radiation (HR, 0.57; 95% CI, 0.34–0.93), rather than cisplatin/5-FU–radiation (HR, 0.8; 95% CI, 0.56–1.15). There was no difference in grade 3 to 4 nonhematologic toxicities, including stomatitis, between the ST and CRT arms. Compliance with CRT was not affected by induction TPF. These results, presented in abstract form, are noteworthy and further the argument that ST may be superior to CRT. Controversy still remains because definitive conventional bolus cisplatin-radiation was not a comparator arm in this trial. These results also raise the question of what is the optimal CRT following TPF induction. Further evaluation is required to determine whether TPF followed by cetuximab-radiation (bioradiation) is a preferred ST regimen for therapeutic efficacy and tolerance. Subset analysis to identify which patients are most likely to benefit from ST is of great interest.

Larynx preservation randomized trials: sequential chemoradiotherapy versus concurrent chemoradiotherapy

ST and CRT are both validated as organ preservation strategies in locally advanced larynx cancer. Induction chemotherapy was first established as part of an organ preservation strategy following the landmark Department of Veterans’ Affairs Laryngeal Cancer Study Group larynx trial in 1991, which randomized patients to either 3 cycles of PF induction chemotherapy followed by definitive RT or primary surgery followed by postoperative RT. Patients without at least a partial response to induction chemotherapy, and those with evidence of disease progression during or after induction, were managed with surgery and postoperative RT. The larynx preservation rate was 64% in the induction chemotherapy group. The 2-year survival rate was equivalent in both treatment arms (68%).

Concurrent chemoradiotherapy was subsequently evaluated as an alternative to induction chemotherapy followed by RT alone in the RTOG 91-11 study. RTOG 91-11 randomized 547 patients with T2, T3, or low-volume T4 supraglottic or glottic larynx cancer to CRT with bolus cisplatin, PF induction chemotherapy followed by RT alone in responders, or RT alone. In all 3 groups, definitive RT consisted of 70 Gy in 35 fractions. The initial report in 2003 after a median follow-up of 3.8 years concluded that CRT was superior to PF followed by RT, or RT alone for larynx preservation rate (88% vs75% vs70% respectively) at 2 years. CRT and induction chemotherapy were equivalent for the primary end point of laryngectomy-free survival (LFS). OS rates were similar among all 3 groups. Following RTOG 91-11, CRT was adopted in North America, whereas induction chemotherapy continued to be favored in Europe.

Notably, the updated long-term results of RTOG 91-11, with a median of 10.8 years’ follow-up, support a more favorable role of induction compared with CRT. The 10-year LFS and OS rates for CRT and induction chemotherapy were similar, and both were superior to RT alone. However, after about 4.5 years, the Kaplan-Meier curves for LFS and OS show clear separation favoring induction chemotherapy compared with CRT (10-year OS 39% vs 28%). This trial was not designed to assess whether this difference was statistically significant but it represents a potentially clinically meaningful difference. Although CRT was associated with a superior larynx preservation rate compared with induction, this was coupled with a statistically significant increase in non-larynx cancer–related deaths in the CRT arm compared with the induction chemotherapy arm (69.8% vs 52.8% at 10 years; P = .03). This finding highlights that larynx preservation is an imperfect end point and ultimately success cannot be claimed if many patients succumb to late treatment toxicity that compromises the function of a preserved larynx or ultimately survival.

For larynx cancer, it remains unclear whether an ST approach using modern induction chemotherapy regimens of TPF would result in greater survival benefit or reduction in distant metastases compared with CRT alone. The superiority of TPF compared with PF induction in larynx preservation has been shown, but the optimal CRT regimen following TPF induction chemotherapy is less defined. TREMPLIN is a phase II trial in locally advanced larynx/hypopharynx squamous cell carcinoma that randomized 153 patients to TPF (docetaxel and cisplatin 75 mg/m ² each on day 1 and fluorouracil 750 mg/m ² on days 1–5) followed by conventional bolus cisplatin concurrent with radiation or TPF followed by cetuximab-radiation. The primary end point of larynx preservation at 3 months was similar, as were larynx function preservation and OS. Compliance was higher with cetuximab-RT, but there were fewer local failures in the cisplatin-RT arm. Despite this, after salvage surgery (which only occurred in the cetuximab-RT arm) the ultimate local failure rates were comparable between both arms.