Redefining the Role of Induction Chemotherapy in Head and Neck Cancer

Abstract

The role of induction chemotherapy in locoregionally advanced squamous cell head and neck cancer continues to be debated. The dramatic responsiveness of this tumor to induction chemotherapy initially promised an improvement in locoregional control and survival. However, multiple phase III clinical trials and metaanalyses exploring a number of different drug combinations and treatment schedules have failed to demonstrate any consistent impact on these end points. A marginal survival benefit was achieved using fluorouracil and platinum induction chemotherapy in the large MetaAnalysis of Chemotherapy on Head and Neck Cancer, but this benefit was dwarfed by the clear cut and consistent improvement identified from concurrent chemoradiotherapy schedules. As such, enthusiasm for induction therapy waned. A frequent finding from the induction studies was a decrease in the likelihood of distant metastases. The lack of impact on overall survival from induction chemotherapy, despite this improvement in distant recurrence, was felt to reflect the limited importance of distant metastases in the natural history of the disease. With increasingly aggressive concurrent chemoradiotherapy schedules, however, locoregional control has improved, and the impact of distant metastatic disease on overall survival has become more important. This is the rationale behind the current sequential treatment schedules, which explore induction chemotherapy to address the risk of distant metastases, followed by concurrent chemoradiotherapy to optimize locoregional control. The recent recognition that a taxane, when added to the well-studied fluorouracil and cisplatin combination, can improve the success of induction chemotherapy has spawned a new generation of phase III studies exploring the impact of three-drug induction therapy on overall survival. In this issue, Worden et al from the University of Michigan explore another potential role for induction chemotherapy. The premise of their approach is that induction chemotherapy can serve as a predictive tool, allowing appropriate selection of the definitive head and neck cancer management strategy. Those patients responding to induction chemotherapy can be approached with an organ preservation strategy. For those patients in whom induction chemotherapy is unsuccessful, early surgical resection would be more appropriate. The prognostic rationale here is entirely reasonable. It has frequently been reported that patients who respond to chemotherapy do better than those patients who do not respond, although this observation likely reflects the multiple other tumorand patient-related factors that determine responsiveness to chemotherapy, including disease extent and performance status. In head and neck cancer, this has been further extended by the recognition that those patients who respond to chemotherapy are also the patients most likely to respond to radiotherapy. This predictive potential of chemotherapy served as one of the rationales behind the study design of the initial larynx preservation trial from the Department of Veterans’ Affairs Laryngeal Cancer Study Group. In that study, patients not responding to two courses of induction chemotherapy with fluorouracil and cisplatin underwent early laryngectomy and postoperative radiation. However, responders received additional chemotherapy followed by definitive radiation alone. This approach allowed larynx preservation in 64% of patients, without compromising survival when compared with standard surgery and postoperative radiation. This treatment algorithm has been further refined by the University of Michigan group. In their trials, a single cycle of induction chemotherapy is used to define optimal definitive management. Patients responding to this single course of chemotherapy are treated with definitive radiation with concurrent single-agent cisplatin. Patients who do not respond to the single induction course undergo early surgical resection. The use of only a single induction chemotherapy course seems to limit its role to that of a chemoselection tool, rather than exploiting any potential impact of the chemotherapy on locoregional control, survival, or distant metastases. In their 2006 report in this journal, the Michigan group felt that this approach was successful for advanced laryngeal cancer. Contrary to the expectation of an inferior survival in the chemotherapy nonresponders, the nonresponders who underwent an early laryngectomy seemed to have an equivalent survival to the chemotherapy responders treated with definitive nonoperative chemoradiotherapy. The current report reviews this institution’s experience in patients with oropharynx cancer. In these patients, however, early surgical resection was successful in only four of the 11 induction chemotherapy nonresponders, an overall treatment success that was distinctly inferior to that of the chemotherapy responders treated nonoperatively. Although chemoselection seemed to be a successful strategy in patients with larynx cancer, this was not the case for those with oropharynx tumors, leading the authors to suggest a need to develop alternative treatment approaches for nonresponders. Why did this same group of investigators find such disparate results between their patients with larynx and oropharynx cancer? Are these reports a valid test of this treatment algorithm? Clearly, it is difficult to draw firm conclusions from this kind of phase II experience. A true evaluation of this concept of chemoselection would require a randomized trial comparing definitive chemoradiotherapy JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 26 NUMBER 19 JULY 1 2008