Argiris and colleagues report a systematic review evaluating the activity and potential role of induction chemotherapy in patients with previously untreated, locoregionally advanced squamous cell head and neck cancer. They consider both phase II and III published trials. The data reviewed in their paper, and their thoughtful synthesis and interpretation of these data, highlight certain themes:
- The majority of previously untreated patients will have a major response to platinum-based induction chemotherapy; a significant minority will have a clinical complete response.
- While induction chemotherapy is feasible when administered prior to locoregional treatment, it can produce significant toxicity. There is a small but real treatment-related mortality rate.
- Cisplatin with infusional fluorouracil(Drug information on fluorouracil) (5-FU) has been the most extensively investigated induction chemotherapy regimen. Historically, no other regimen has proven clearly superior with regard to antitumor effect, and this drug combination is the one to which new regimens are compared.
- The substitution of carboplatin(Drug information on carboplatin) for cisplatin(Drug information on cisplatin) in combination regimens may ameliorate treatment toxicity but does not improve-and may adversely affect-disease control outcomes.
- Recent encouraging experience with taxane-based induction regimens has raised interest in the induction approach.
The authors have focused on objective response rate as their outcome of interest, and appropriately acknowledge its limitations. Their point regarding the changes in how response is assessed over time, and the importance of considering this issue in interpreting results, deserves emphasis. Response rates, while valuable, are nonetheless imperfect predictors of outcomes of greater importance to both the patient and the treating team, such as survival, organ preservation/function, and quality of life. Indeed, the randomized data evaluating the efficacy of induction chemotherapy on survival are illustrative in this regard. Despite the high response rate, when evaluated in randomized studies and meta-analyses, the addition of induction chemotherapy to locoregional treatment may affect the pattern of failure but does not lead to a clear nor consistent improvement in overall survival as an end point (compared to simply proceeding with locoregional treatment alone), a point emphasized by the authors.[ 4] A review of practice guidelines further supports this interpretation of these data by most experts. For example, there is little mention of induction chemotherapy in the treatment guidelines for locoregionally advanced head and neck cancer published by the National Comprehensive Cancer Network, with the exception of those for advanced hypopharyngeal cancer treated with larynx preservation intent. Furthermore, the results of the Intergroup Larynx Preservation Study, in which concurrent cisplatin and radiation led to the highest larynx preservation rate in patients with advanced larynx cancer (among treatment arms that also include induction cisplatin/ 5-FU and radiation alone), have made concurrent-not induction-chemotherapy/ radiation the preferred approach in this setting. Methodologic Limitations
Why have high response rates not clearly translated into survival benefit? The authors comment on possible biologic and methodologic reasons. We feel the latter deserve further emphasis. Using other clinical settings where adjunctive chemotherapy has demonstrated survival benefit for patients with solid tumors, as well as the concurrent chemotherapy/radiation data in advanced head and neck cancer (absolute increase at 5 years: 8%) as guides, one would anticipate that the magnitude of survival benefit associated with the use of induction chemotherapy in locoregionally advanced head and neck cancer in absolute terms will be modest, if real.[4,7] Trials must be designed very carefully to reliably detect such a difference, and both head and neck cancer and sequential combined-modality treatment approaches pose challenges in this regard. For example, medical comorbidity and second primary tumors are frequent issues in patients with squamous cell head and neck cancer, representing competing causes of mortality.[8,9] "Stage III and IV disease," a common eligibility criterion in combined-modality studies, can include a broad spectrum of T and N combinations as well as primary sites with considerable associated variation in prognostic expectations, risk and pattern of failure, and likelihood of benefit from added chemotherapy. The toxicity and response associated with induction may affect compliance with and performance of the planned locoregional treatment. If conservative sample size estimates are not applied, such factors can lead to underpowered studies, efficacy esti-mates with broader confidence intervals than desired, and interventions in which the potency has decreased, collectively decreasing a study's ability to detect the anticipated difference in outcome. Importance of Phase III Data
The authors considered both phase II and phase III studies in their review. Since there can be a disconnect between promising response rates in phase II trials and subsequent validation in phase III trials, the latter studies are more relevant with respect to how different therapies compare in efficacy, and allow for more definitive conclusions. Some of the data interpretations made by the authors are best considered in this framework. For example, the authors note that historically, other cisplatin-based combinations had less antitumor efficacy than cisplatin/5-FU, but the available randomized studies that address this question are very limited. In a randomized study in the recurrent disease setting, cisplatin/5-FU offered no advantage in response or survival compared to the combination of cisplatin, methotrexate(Drug information on methotrexate), bleomycin(Drug information on bleomycin), and vincristine. There is great interest in the taxanes as part of a strategy to improve outcomes in head and neck cancer patients, and certainly as a drug class they have significant activity. However, phase III data in the recurrent disease setting demonstrated that methotrexate yielded comparable response and survival outcomes when compared to two different schedules of paclitaxel(Drug information on paclitaxel). Similarly, paclitaxel and cisplatin offered no efficacy advantage compared to 5-FU and cisplatin, although the former had a more favorable side-effect profile.[11,12] Key Study
We agree that the recently reported randomized data from Vermorken et al comparing docetaxel(Drug information on docetaxel) (Taxotere)/cisplatin/ 5-FU vs cisplatin/5-FU, both followed by radiation alone, in patients with locoregionally advanced disease are particularly important. These investigators demonstrated that the triplet arm had statistically significant improvements in response rate, progression- free survival, and overall survival, without significantly increased toxicity. These data are consistent with an efficacy advantage with the triplet, making further phase III investigation of induction strategies in this setting particularly attractive. Further followup on this study and those of similar design is eagerly awaited. It should be emphasized that in the study reported by Vermorken and colleagues, locoregional treatment entailed radiation alone, not concurrent chemotherapy/radiation therapy. Whether induction chemotherapy provides incremental benefit when followed by concomitant therapy compared with concomitant therapy alone is not definitively known but is the subject of important and potentially treatment-paradigm-shifting studies that are planned or in progress, as highlighted by the authors. The results are eagerly awaited. Conclusions
In summary, the available results of studies incorporating a taxane with cisplatin and 5-FU as induction therapy are very encouraging.[3,13] The apparent increased efficacy of this type of regimen over cisplatin/5-FU, combined with an enhanced appreciation among clinical investigators regarding how to address the methodologic challenges of evaluating induction therapy, support current investigational efforts to clarify the role of induction chemotherapy in this setting. Consistent and careful acute and late toxicity reporting, as well as prospective quality-of-life data, will be of great importance in assessing the relative risks and benefits in such studies. Well done randomized studies will be fundamental to resolving this management question, and we are hopeful that this "second generation" of induction studies will show consistent outcome benefit with the incorporation of chemotherapy prior to locoregional treatment. Until more data are available, however, concurrent chemotherapy/ radiation alone, without induction therapy, remains an appropriate standard treatment option for such patients and the standard to which new approaches are compared. Available data suggest that when concurrent chemotherapy is utilized with radiation, it should be platinumbased.[ 4,7,14]