ABSTRACT: Squamous cell carcinomas of the head and neck are highly responsive to induction chemotherapy. However, randomized trials have failed to demonstrate a survival advantage with the addition of induction chemotherapy to locoregional therapy consisting of surgery and/or radiation therapy. Currently, concomitant radiation and chemotherapy has emerged as a standard and has optimized locoregional control in head and neck cancer. In this setting, the addition of induction chemotherapy may further improve outcome by enhancing both locoregional and distant control. As interest in induction regimens is renewed, we elected to conduct a systematic review of trials of induction chemotherapy for locoregionally advanced head and neck cancer. The most studied combination— cisplatin(Drug information on cisplatin) plus fluorouracil(Drug information on fluorouracil) (5-FU)—achieves objective response rates of about 80%. In a meta-analysis, induction with platinum/ 5-FU resulted in a small survival advantage over locoregional therapy alone. The introduction of a taxane into induction chemotherapy regimens has produced promising results. Induction chemotherapy should be the subject of further clinical research in head and neck cancer. Randomized clinical trials in which the control arm is concurrent chemoradiotherapy and the experimental arm is induction chemotherapy followed by concurrent chemoradiotherapy are planned. Platinum/taxane combinations are the preferred regimens for further study in the induction setting and a suitable platform with which to investigate the addition of novel targeted agents.
Head and neck cancer affects about 39,000 persons and results in 11,000 deaths annually in the United States.[1] Patients with head and neck cancer commonly have locoregionally advanced disease at presentation, which is associated with a poor long-term outcome: Traditional treatment with surgery and/ or radiation produces a 5-year survival rate of 40% or less.[2] Historically, disease recurrence has been predominantly locoregional, whereas distant failure occurs in 20% to 30% of patients. The use of chemotherapy has been introduced with a binary objective- to achieve organ preservation and improve survival. Indeed, in recent years, chemotherapy has been successfully employed in a variety of clinical settings and has established an integral role in the curative management of locoregionally advanced head and neck squamous cell carcinoma. The timing of chemotherapy in this setting can be either (1) prior to locoregional therapy (ie, induction, neoadjuvant, or preoperative), (2) concurrent with definitive radiation therapy, or (3) after locoregional therapy with or without concomitant radiation therapy (ie, adjuvant or postoperative). Currently, the data strongly support the use of chemotherapy concurrently with radiation therapy either as primary therapy or in the postoperative setting. Concurrent chemoradiotherapy has become the standard nonsurgical treatment for locoregionally advanced head and neck cancer. However, the optimal chemotherapeutics for use concurrently with radiation to treat head and neck cancer have not yet been determined. A favored regimen, supported by results of cooperative group randomized trials in the United States, is cisplatin at 100 mg/m2 every 3 weeks during conventional fractionation radiation. In unresectable disease and nasopharyngeal cancer, a survival advantage was demonstrated with the use of the above regimen over radiotherapy alone,[3,4] whereas in laryngeal cancer, concurrent chemoradiotherapy with cisplatin resulted in a higher rate of organ preservation.[5] Finally, the combination of cisplatin and radiation therapy was superior to radiation therapy alone after a potentially curative surgical resection.[6,7] In addition to single-agent cisplatin, a number of combination chemotherapy regimens, predominantly platinum/fluorouracil (5-FU), have been studied concurrently with radiation and produced superior results over radiation alone in randomized trials.[8]
Induction Chemotherapy: Rationale and Controversy
Squamous cell carcinomas of the head and neck are highly responsive to induction chemotherapy. A complete response (CR) to induction chemotherapy, and especially a pathologic CR, is predictive of improved survival.[9-13] It is reasonable to hypothesize that induction chemotherapy can improve patient outcome by downstaging the primary tumor and eradicating locoregional and distant micrometastases. Nevertheless, with a few notable exceptions, multiple randomized trials conducted in the 1980s and 1990s that used standard locoregional therapy- surgery and/or radiation-as a control failed to demonstrate that the addition of induction chemotherapy to locoregional treatment results in survival benefit.[14,15] In several of these studies, the addition of induction chemotherapy decreased the rate of distant metastasis, which usually did not translate into a survival benefit. Moreover, a deleterious effect on survival from induction chemotherapy was suggested in some studies.[16,17] Domenge et al reported the only positive trial, a study that enrolled 318 patients with locally advanced oropharyngeal cancer.[18] Disease-free survival was improved in the induction chemotherapy arm, but the difference did not reach statistical significance (P = .11); however, overall survival was superior in the induction chemotherapy arm (P < .05).[18] A study by Paccagnella et al, comparing induction chemotherapy with cisplatin/ 5-FU followed by surgery and/ or radiation, showed a survival benefit for induction therapy but only in patients with unresectable tumors.[19] In a meta-analysis by Pignon et al, when all induction chemotherapy studies were examined together, there was no survival benefit with induction chemotherapy.[ 20] However, when the subset of trials with cisplatin (or carboplatin(Drug information on carboplatin)) plus 5-FU was examined, a small (3%) but statistically significant improvement in survival was observed with the addition of induction chemotherapy (hazard ratio [HR] = 0.88, 95% confidence interval [CI] = 0.79-0.97). Nevertheless, the superior treatment was concomitant delivery of chemotherapy and radiation, which resulted in a meaningful survival benefit of 8% at 5 years (HR = 0.81, 95% CI = 0.76-0.88). A number of studies have directly compared induction chemotherapy followed by single-modality radiotherapy to concurrent chemoradiotherapy.[5,21] A study by Taylor et al showed that concurrent therapy on an alternate-week schedule produced better disease control than induction chemotherapy followed by radiation.[21] Furthermore, Forastiere et al reported that locoregional control and laryngeal preservation rates were superior with concurrent vs sequential chemotherapy and radiation.[ 5] Therefore, concurrent chemoradiotherapy has emerged as the preferred treatment strategy for locally advanced head and neck cancer. However, it is unknown whether induction chemotherapy would be of benefit when the primary therapy is concurrent chemoradiotherapy. When the patterns of failure in randomized trials of concurrent chemotherapy and radiation vs radiation alone were analyzed, it was evident that the survival benefit was most likely a result of improved locoregional control.[8] More recently, phase II studies have reported that high rates of local control approaching 90% can be achieved with intensive chemoradiotherapy regimens.[22,23] A reversal of the historical pattern of relapse (ie, local more common than distant) was observed, with a predominance of distant failure.[24] In that context, eradication of distant micrometastasis that is potentially achievable with induction chemotherapy may become critical. In conclusion, the potential role of induction chemotherapy in improving the survival of patients treated with concurrent chemoradiotherapy needs to be revisited. We conducted a review of the activity and toxicity profile of various induction chemotherapy regimens. The review of these clinical trials may assist in the identification of the most efficacious regimens for use in future trials of induction chemotherapy.
Methods
We performed a computerized search of Cancerlit, Medline, and American Society of Clinical Oncology website databases in order to identify phase II and III clinical studies of induction chemotherapy for locoregionally advanced head and neck cancer. Trials in the English literature that included previously untreated patients with locoregionally advanced squamous cell cancer of the head and neck were selected and included in this review. We recorded the objective response rates, pathologic response rates (when reported), and toxicities of the regimens. Survival data were primarily reviewed in randomized trials. It should also be noted that most studies included heterogeneous patient populations usually without eligibility restrictions by site or resectability. Moreover, there were significant differences in the methods used to assess and report objective response rates between trials. Although survival end points may be relevant to the efficacy of induction chemotherapy regimens,[ 25] subsequent locoregional therapy is a major confounding factor. Therefore, we elected to use objective response rates as a surrogate of chemotherapy efficacy, acknowledging the limitations of this assumption.
Differences in Response Among Head and Neck Sites
Some reports have suggested differences in chemotherapy responsiveness among squamous cell carcinomas that arise from different head and neck sites. Nasopharyngeal cancer may be different biologically and appears to be the most chemotherapy-responsive head and neck tumor.[26-33] Separate clinical trials of induction chemotherapy for nasopharyngeal cancer are warranted. The oral cavity was shown to be a less responsive site in one study,[34] but contradicting results have also been reported.[35] A recent trial reported an objective response rate of 82% in the primary tumor among patients with resectable oral cavity squamous cell carcinoma.[36] Limited experience has been reported for other less common sites of head and neck cancer[37] such as the paranasal sinuses, but it is likely that squamous cell carcinomas arising from these sites and other more common sites are equally chemotherapy responsive. In this review, the studies examined mainly enrolled patients with tumors affecting four major primary sites: oral cavity, oropharynx, hypopharynx, and larynx.
Optimal Number of Treatment Cycles
Although no randomized trial data have suggested an optimal number of chemotherapy cycles in this setting, it is widely accepted that two to four cycles are optimal as induction therapy.Studies with cisplatin and 5-FU have shown that the response rate is higher with three vs two cycles,[38,39] whereas a plateau may be reached after three cycles.[14] Di Blasio et al observed no further increase in the rate of objective response with the extension of induction chemotherapy to five cycles; however, there the CR rate improved (see Table 1).[14] 
Shin et al reported that the rate of complete response to carboplatin/ ifosfamide(Drug information on ifosfamide)/paclitaxel at the primary sites increased from 23% after two cycles to 60% after four cycles.[40] Moreover, the Department of Veterans Affairs (VA) laryngeal cancer study and the European Organization for Research and Treatment of Cancer (EORTC) trial in resectable laryngeal and hypopharyngeal cancers, respectively,[41,42] as well as other randomized trials[36] have employed three cycles of induction chemotherapy with cisplatin and 5-FU. Finally, a recent phase III, randomized EORTC trial in patients with unresectable head and neck cancer incorporated four cycles of induction chemotherapy in both arms-cisplatin/5-FU/docetaxel [Taxotere] or cisplatin/5-FU (Figure 1).[43]
Activity of Combination Regimens
Cisplatin and 5-FU
Cisplatin and 5-FU have been the most commonly used agents in the induction chemotherapy of head and neck cancer.[44] When these agents are combined, standard doses are cisplatin at 100 mg/m2 on day 1 and 5-FU at 1,000 mg/m2/d as a continuous infusion for 5 days, repeated every 3 weeks. This dosing schedule was proven to be superior to cisplatin plus bolus 5-FU in a small randomized study performed at Wayne State University.[ 45] In this study of 42 patients with locally advanced head and neck cancer, both the overall objective response rate (72% vs 20%) and the CR rate (22% vs 10%) were significantly higher for the continuous infusion vs bolus 5-FU treatment arm. Toxicities were also less pronounced in the continuous-infusion arm except for mucositis, which was usually mild in severity and tolerable. Multiple phase II and III trials using PF (cisplatin [Platinol]/5-FU) as induction therapy for locally advanced head and neck cancer have been conducted (Table 2).[5,9,14-17,19,34,36,38,41,43, 46-57] Older studies[54,56,57] utilized clinical/endoscopic evaluation of response only, but more recent studies[ 5,16,48] routinely employed computed tomography (CT) in addition to endoscopy. The OR and CR rates achieved in 24 trials with cisplatin/ 5-FU ranged from 54% to 94% (mean: 80%) and 7% to 66% (mean: 32%), respectively. Pathologic CR rates, which were reported in a minority of studies, ranged between 22% and 33%.[36,46,51,53,54] Toxicities observed in four large trials of induction chemotherapy with the standard cisplatin/5-FU regimen are listed in Table 3.[15,34,38,52-57] The toxic death rate during induction cisplatin/ 5-FU is low, ie, 3% or less.[9,36]
Carboplatin and 5-FU
Carboplatin was introduced as a safer but potentially equally efficacious agent to cisplatin for the treatment of head and neck cancer. However, the combination of carboplatin and 5-FU (CF) was shown to be inferior to PF in the recurrent or metastatic disease setting.[ 61] In the induction setting, a number of phase II trials as well as a phase III trial of CF have been conducted (Table 4).[48,58-60]
CF was compared with PF in a randomized trial conducted in Spain.[48] De Andres et al randomized 96 patients with stage IV squamous cell carcinoma of the head and neck to receive either CF or standard PF. The study was terminated at interim analysis, when superiority of the control arm was demonstrated. The overall response rate (ORR) was higher in the PF arm (92% vs 76%). Moreover, the 5-year disease-specific survival rate (49% vs 25%, P = .03) and 5-year disease-free survival rate (47% vs 24%, P = .02) were significantly better in the PF arm. Differences in the toxicity profile were also noted. Hematologic toxicity predominated in the CF arm, whereas mucositis and nausea/vomiting were more frequently seen with PF. In conclusion, CF should be considered an inferior regimen. However, carboplatin may be substituted for cisplatin in cases of intolerable cisplatin-associated toxicity.
Cisplatin/5-FU Plus Leucovorin
The modulation of 5-FU was a topic of intense research in the 1980s and 1990s. Many phase II clinical studies investigated the efficacy of standard PF when leucovorin was added to the regimen (PFL); leucovorin was administered in various formulations in these studies (Table 5).[62-66] The mean ORR in these five phase II clinical trials of PFL was 78%, with a mean clinical CR rate of 37%. A major downside of PFL was that it results in high rates of severe mucositis (14% to 41%)-higher than what is expected with standard PF-which suggests 5-FU modulation by leucovorin in the normal mucosa.[62-66] Clark et al reported that 27% of patients developed grade 3/4 stomatitis in the first PFL cycle, with a significant percentage of patients requiring hospital admission for severe stomatitis with dehydration,[62] whereas Vokes et al observed that 14% of patients developed grade 3 mucositis, which in some cases was of very early onset, occurring during the first 5 days of the 5-FU infusion.[ 63] Other authors reported that 36% to 41% of patients developed grade 3 or 4 mucositis.[64-66]
Thus, although no randomized trial has compared PFL to PF in the setting of induction chemotherapy for head and neck cancer, the added benefit provided by leucovorin seems to be negligible. Moreover, the toxicities added to the regimen may be considerable. Finally, two phase II trials conducted by groups in Chicago studied the addition of interferon alfa-2b(Drug information on interferon alfa-2b) (Intron-A) to PFL.[67,68] The combination was highly active, resulting in clinical CRs in 51% to 66% of patients. Despite its high activity, PFL/ interferon alfa-2b had an unfavorable toxicity profile, with high rates of severe mucositis (52%-54%) and myelosuppression.[ 67,68]
Cisplatin/Bleomycin-Based Regimens
Cisplatin/bleomycin-based combinations have been extensively studied for the treatment of head and neck cancer. As induction chemotherapy, cisplatin plus bleomycin(Drug information on bleomycin) (PB) demonstrated ORRs of 48% to 70% and CR rates of 0% to 17%.[69,70] The Head and Neck Contracts Program (HNCP) trial randomized patients with stage III/IV resectable head and neck cancer to one of three arms: (1) one cycle of induction chemotherapy with PB followed by standard therapy (surgery and postoperative radiotherapy); (2) induction chemotherapy with PB and standard therapy followed by maintenance chemotherapy, which consisted of six cycles of monthly cisplatin; or (3) standard therapy alone.[71] The ORR observed with one cycle of PB in the HNCP study was low-37%, with a 3% CR rate-and there was no survival advantage with the use of chemotherapy. Toxicities associated with PB included nausea/vomiting (grade 3/4 vomiting in 25%), bleomycin-related rash (grade 3 in 2%), mucositis (grade 3 in 1%), neutropenia, and anemia. Other investigators evaluated the addition of methotrexate(Drug information on methotrexate) to PB with or without leucovorin; we identified four trials in which these regimens yielded ORRs between 58% and 77% (mean: 68%) and CR rates of 9% to 26% (mean: 20%).[12,72-74] In two other studies of PB-containing combinations, the incorporation of vincristine resulted in an ORR of 79% and a CR rate of 29%,[38] whereas the addition of vindesine(Drug information on vindesine) plus mitomycin(Drug information on mitomycin) produced an ORR of 50%, a 10% CR rate in the primary site, 27% ORR, and 9% response rate in the lymph nodes.[75] Weaver et al reported lower rates of CR for two cycles of PB plus vincristine vs three cycles of standard PF (29% vs 54%, P = .04) in a nonrandomized comparison.[38]
Finally, Schuller and colleagues randomized 158 patients with locoregionally advanced head and neck cancer to receive surgery plus radiation with or without induction chemotherapy consisting of three courses of PB plus vincristine and methotrexate.[76] These investigators reported an ORR of 70% and a CR rate of 19%. However, there was no statistically significant difference in survival between the two treatment arms.[76] In conclusion, PB and multiagent PB-based regimens do not optimize antitumor activity in the induction setting. Currently, the use of bleomycin is exceedingly rare for the treatment of patients with head and neck cancer.
Platinum-Plus-Taxane Combination Regimens
The combination of a platinum (carboplatin or cisplatin) and a taxane (paclitaxel or docetaxel(Drug information on docetaxel)) has shown marked antitumor activity in head and neck cancer. A potential advantage of the platinum-plus-taxane combination is that the incidence of stomatitis seen with it is less than what is observed when 5-FU is part of the induction regimen, which may decrease complications of subsequent radiotherapy.
• Platinum and Paclitaxel(Drug information on paclitaxel)—Carboplatin and paclitaxel is well tolerated and active as induction chemotherapy in patients with head and neck cancer. In seven clinical trials (Table 6),[77- 83] the ORR ranged from 66% to 89% (mean: 82%) and CR rates from 13% to 48% (mean: 30%). Although every-3-week administration of carboplatin and paclitaxel is common, weekly regimens with very promising antitumor activity have also been studied by groups at the University of Chicago[78,79] and Brown University.[ 77] Potential severe toxicities with carboplatin and paclitaxel include febrile neutropenia (~5%), thrombocytopenia, neuropathy, hypersensitive reactions, and myalgias/arthralgias. The Minnie Pearl Cancer Research Network conducted a phase II trial that evaluated the addition of continuousinfusion 5-FU to carboplatin/paclitaxel as induction chemotherapy for locally advanced head and neck cancer in a multicenter, community-based setting.[ 84] Induction chemotherapy consisted of paclitaxel at 200 mg/m2 and carboplatin at an area under the concentration- time curve [AUC] of 6, both administered on days 1 and repeated on day 22, and 5-FU at 225 mg/m2/d as a 24-hour continuous intravenous infusion on days 1 to 43. This was fol- lowed 1 to 3 weeks later by radiation to 68.4 Gy plus concurrent paclitaxel at 50 mg/m2 and carboplatin at an AUC of 2, given weekly for six doses. A total of 123 patients were enrolled in the study, the majority of whom had oropharyngeal (41%), laryngeal (18%), or nasopharyngeal (15%) primaries. In 119 evaluable patients, the ORR was 70% (CRs noted in 9%).[84] With a median follow-up of 24 months, the 3-year progression-free and overall survival rates were 53% and 51%, respectively. Toxicities with the induction regimen included mucositis (grade 3/4: 12%) and neutropenia (grade 4: 8%), and 24% of patients required hospitalization for treatmentrelated complications. Subsequent chemoradiotherapy was feasible, but radiation dose omissions and delays were necessary: 58% of patients received the planned radiation dose, and 61% of patients received all six planned doses of weekly carboplatin and paclitaxel with the radiation. During concurrent chemoradiotherapy, 11% of patients experienced grade 3/ 4 radiation dermatitis, and 75% of patients developed grade 3/4 mucositis, whereas 23% required enteral or parenteral alimentation for more than 6 months. Myelosuppression was infrequent; 4% of patients developed grade 3/4 neutropenia. Overall, this study showed that aggressive induction chemotherapy followed by concurrent chemoradiotherapy can be delivered in the community setting. Efficacy results were promising. Surprisingly, there is a paucity of induction trials with cisplatin and paclitaxel. This combination was studied in a phase I/II trial in head and neck cancer by Hitt et al.[85] Twenty-eight patients with locally advanced, unresectable squamous cell head and neck cancer received cisplatin at 75 mg/m2 plus paclitaxel at 175 to 300 mg/m2 for the first cycle with intrapatient dose escalation, repeated every 21 days for three cycles. Patients receiving paclitaxel doses of 200 mg/m2 and above received granulocyte colony-stimulating factor (G-CSF [Neupogen]) support. In 27 evaluable patients, the ORR was 78% and the CR rate was 48%. Despite increased dose intensity, the regimen was relatively well tolerated without dose-limiting hematologic toxicities. Four patients developed grade 3 myalgias.[55]
• Cisplatin/5-FU/Paclitaxel—Hitt et al investigated the triple combination of cisplatin/5-FU/paclitaxel as induction chemotherapy in head and neck cancer.[86] In this phase II trial conducted in Spain, 70 patients with locally advanced squamous cell head and neck cancer, 60% of whom had unresectable disease, were treated with pac-litaxel, 175 mg/m2 as a 3-hour infusion on day 1; cisplatin, 100 mg/m2 on day 2; and 5-FU, 500 to 750 mg/m2/d as a 24-hour continuous infusion on days 2 to 6, repeated every 3 weeks for a total of three cycles, followed by surgery and/or radiation therapy. In 43 patients, weekly paclitaxel was given concurrently with radiation. The 5-FU dose was reduced from 750 to 500 mg/m2/d due to the excessive toxicity, mainly mucositis and neutropenic sepsis, observed in the first 14 patients enrolled. A total of 70 patients were enrolled, 90% of whom received all three cycles of induction. Major toxicities of cisplatin/5-FU/paclitaxel in patients treated at the lower 5-FU dose included grade 3/4 neutropenia in 29% of patients (with one case of neutropenic fever), grade 3/4 mucositis in 5%, and grade 3 peripheral neuropathy in 5%. The ORR in 69 evaluable patients was 87%, including 58% CRs. The 5-year time to progression and overall survival rates of 56% and 44%, respectively, were encouraging. Following this favorable experience, a phase III trial was launched that compared standard PF with PF (using a 5-FU dose of 500 mg/m2/d) plus paclitaxel at 175 mg/m2 for the induction chemotherapy of head and neck cancer.[87] Preliminary results favored the triple combination, and updated results of this study are pending. Finally, a dose-dense cisplatin/5-FU/ paclitaxel regimen, with cycles repeated every 2 weeks, was studied by the same group in patients with recurrent or unresectable head and neck cancer.[ 88] The activity was very promising (ORR of 83% in recurrent patients and 91% in unresectable disease patients), and toxicities were acceptable: grade 4 neutropenia occurred in one patient (3%) but led to sepsis and death; grade 3/4 mucositis in five patients (15%); and grade 3 neuropathy in one patient (3%).
• Platinum/Paclitaxel Plus Ifosfamide—The combination of a platinum (carboplatin or cisplatin), paclitaxel, and ifosfamide showed impressive antitumor activity (ORR: 58%-59%, CR rate: 17%) in patients with recurrent/metastatic head and neck cancer in two consecutive phase II trials at M. D. Anderson Cancer Center.[89,90] Given these promising results, Shin et al conducted a trial of carboplatin/ paclitaxel/ifosfamide as induction therapy in patients with locally advanced head and neck cancer.[40] The induction regimen consisted of paclitaxel at 175 mg/m2 on day 1, ifosfamide at 1,000 mg/m2 as a 2-hour infusion on days 1 to 3 with mesna(Drug information on mesna), and carboplatin at an AUC of 6 on day 1, repeated every 3 to 4 weeks. Patients achieving an objective response after the first two cycles of induction chemotherapy received two additional cycles (ie, a total of four). Of 52 evaluable patients, 31% had a CR and 50%, a partial response (ORR: 81%). Five patients (9%) developed neutropenic fever. Grade 3/4 thrombocytopenia and anemia occurred in three (6%) and four (7%) patients, respectively. Grade 3/4 fatigue developed in four patients (7%), arthralgia/ myalgia in two (4%), peripheral neuropathy in two (4%), and one patient died of an anaphylactic reaction. Khuri et al used paclitaxel/ifosfamide/ cisplatin as sole therapy for selected patients with intermediate-stage (T2-T4, N0/N1) supraglottic or glottic laryngeal cancer that was deemed potentially resectable by conservation laryngeal surgery.[91] The study regimen consisted of paclitaxel at 175 mg/ m2 on day 1, ifosfamide at 1,000 mg/ m2 as a 2-hour infusion on days 1 to 3 with mesna, and cisplatin at 60 mg/m2 on day 1. After three cycles of paclitaxel/ ifosfamide/cisplatin, patients achieving a partial response proceeded to surgery, whereas patients achieving a pathologic CR received an additional three cycles of chemotherapy and no local treatment. Updated results with a median follow- up of 36 months were recently presented.[92] In a total of 29 patients who were treated on study, 11 (38%) achieved a CR, 9 of whom have remained disease-free without locoregional therapy (median followup: 36 months), and 18 (62%) had a partial response. Overall, the results with platinum/paclitaxel/ifosfamide are favorable but represent single-institutional experience. The contribution of ifosfamide in the activity of these regimens is difficult to discern.
• Cisplatin/Docetaxel-Based Combinations—Docetaxel has shown high levels of activity as a single agent and in combination regimens in head and neck cancer. Cisplatin and docetaxel have been combined in phase II trials in recurrent or metastatic head and neck cancer with overall response rates between 33% and 54%.[93-95] This is at least comparable to the activity of cisplatin/ 5-FU and cisplatin/paclitaxel in the recurrent/metastatic disease setting. The combination of cisplatin and docetaxel is well tolerated even in the palliative care setting. The US Food and Drug Administration approved cisplatin/ docetaxel for first-line therapy of metastatic non-small-cell lung cancer. The commonly used doses are 75 mg/m2 for each drug on day 1. This regimen is associated with acceptable toxicities, including about a 5% incidence of febrile neutropenia.[96] There has been limited experience with cisplatin/docetaxel as induction chemotherapy for locoregionally advanced head and neck cancer. The EORTC conducted a phase II trial of docetaxel, 100 mg/m2, followed by cisplatin, 75 mg/m2, both administered on day 1 and repeated every 3 weeks, without prophylactic colony-stimulating factor support.[94] In a subgroup of 18 patients with chemotherapy- naive locally advanced or recurrent head and neck cancer without distant metastasis, the regimen showed substantial activity with an ORR of 89%. Another phase II trial of cisplatin/docetaxel in patients with locally advanced unresectable squamous cell head and neck cancer reported a lower response rate of 55% (30% CR rate) in 37 evaluable patients.[ 97] Finally, preliminary results have been reported from a randomized phase II trial of docetaxel, 85 mg/m2 on day 1, plus cisplatin, 40 mg/m2 on days 1 to 2 vs standard PF in patients with stage III/IV squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx.[98] Both regimens were repeated every 3 weeks to a maximum of four cycles. Objective response rates were comparable between the two treatment arms (ORR/ CR rate: 69%/15% for docetaxel/cisplatin vs 71%/12% for PF). However,the toxicity profile of the two regimens was different. Docetaxel/cisplatin resulted in a higher rate of grade 3/4 neutropenia but less frequent grade 3/4 mucositis and nausea/vomiting.[98]
• Cisplatin/5-FU/Docetaxel—Docetaxel-With the goal of improving the efficacy of PF, many groups have studied the addition of docetaxel to the combination (TPF), and in some cases, the addition of leucovorin (TPFL), in a number of phase I and II trials (Table 7).[99-104] Locoregional therapy varied considerably among these studies. Efficacy results have been very promising. When examined collectively, these trials showed that TPF or TPFL produced 2-year survival rates of 42% to 82% and ORRs that ranged from 71% to 100%.[105] The predominant severe toxicities encountered with TPF are hematologic, with grade 3/4 neutropenia seen in 83% to 95% of patients, resulting in neutropenic fever in 16% to 19%, and grade 3/4 stomatitis in 17% to 30%.[99,101] Treatment-related deaths occurred rarely (0%-4%).[99,101] As expected, toxicities were more pronounced when each drug in the regimen was administered in full doses. On the basis of encouraging efficacy results, two phase III randomized trials comparing TPF with standard PF have been conducted, one of which, conducted by the EORTC, recently re- ported results.[43] This study, presented by Vermorken et al, randomized 358 patients with unresectable head and neck cancer to receive one of two induction regimens: (1) cisplatin at 75 mg/m2 on day 1, 5-FU at 750 mg/ m2 continuous infusion * 5 days, and docetaxel at 75 mg/m2 on day 1, with prophylactic ciprofloxacin(Drug information on ciprofloxacin), or (2) standard PF (cisplatin at 100 mg/m2 on day 1, 5-FU at 1,000 mg/m2 continuous infusion * 5 days). Induction chemotherapy was followed by single-modality radiation therapy (various fractionation schemes allowed) in both arms. TPF proved superior to PF in terms of ORR (68% vs 54%, P = .007), progression- free survival (HR = 0.72; 95% CI = 0.56-0.91), and overall survival (HR = 0.73; 95% CI = 0.57- 0.94).[43] The triplet regimen was generally better tolerated, as it resulted in lower rates of severe nausea/ vomiting and mucositis, apparently due to the use of lower doses of cisplatin and 5-FU in this regimen compared with standard PF. A second randomized trial of TPF (TAX 324) is using somewhat different dosages for cisplatin/5-FU: cisplatin at 100 mg/m2 and 5-FU at 1,000 mg/m2 as a 4-day infusion. Patients in both arms subsequently receive radiation plus weekly carboplatin. Planned randomized trials of chemoradiotherapy with or without induction chemotherapy from the Southwest Oncology Group (SWOG) and the Dana-Farber Cancer Institute will be employing variants of TPF as induction chemotherapy in the experimental arm.
Induction Chemotherapy for Organ Preservation
A major goal of nonsurgical therapy is to achieve organ preservation in patients with resectable head and neck tumors. The avoidance of laryngectomy is the best example of an organpreservation goal. Three randomized trials that included about 600 patients with locally advanced but resectable laryngeal or hypopharyngeal squamous cell carcinoma have compared induction chemotherapy followed by radiation therapy with surgery that included laryngectomy followed by radiation therapy.[41,42,106] A meta-analysis showed a nonsignificant trend toward worse local control (12% vs 25%) and overall survival (39% vs 45% at 5 years) with the nonsurgical approach.[18,107] There was significant heterogeneity among the three trials; moreover, one of them had a small sample size (< 70),[106] which limited the interpretation of meta-analysis results. Induction chemotherapy may decrease the rate of mutilating surgery for other sites, such as the oral cavity, as shown in a randomized trial by Licitra et al.[36] In laryngeal cancer, a subsequent trial to the VA laryngeal study three-arm trial (which contained no surgical arm) showed that optimal laryngeal preservation is achieved with concurrent chemoradiotherapy.[5] Patients with resectable stage III/IV laryngeal cancer were randomly assigned to one of three arms: (1) the laryngealpreservation strategy used in the VA laryngeal study, consisting of induction chemotherapy with three cycles of cisplatin and 5-FU, followed by radiotherapy for responders, (2) radiation plus concurrent cisplatin at 100 mg/m2 administered every 3 weeks, or (3) radiation alone. Locoregional control (78% vs 61%, P = .003) and laryngealpreservation rates were superior with concurrent chemoradiotherapy vs induction chemotherapy (84% vs 72%, P = .005). An advantage in overall survival could not be demonstrated with the use of chemotherapy, despite a decrease in the rate of distant metastasis in the two chemotherapy arms. Based on the results of this study, concurrent chemoradiotherapy emerged as the preferred strategy to achieve organ preservation in laryngeal cancer. This may hold true for other primary sites as well.
Planned Randomized Trials of Induction Chemotherapy
The feasibility of platinum-based induction chemotherapy followed by chemotherapy given concurrently with radiotherapy has been demonstrated in multiple clinical trials (see Table 8), [78,79,81,82,84,87,108] and many groups are launching phase III trials of induction chemotherapy. The general schema of these trials is shown in Figure 2. The control arm consists of concurrent chemoradiotherapy, but regimens and eligibility criteria vary considerably. For example, SWOG is planning a trial in resectable oropharyngeal cancer with cisplatin plus standard radiation as the control arm and cisplatin/ docetaxel/5-FU followed by cisplatin plus standard radiation as the experimental arm. Based on their observations with the T-FHX regimen (paclitaxel, 5-FU, hydroxyurea, and hyperfractionated radiation),[78,79] Vokes et al at the University of Chicago will enroll patients with advanced N-stage disease and will use TPF as the induction regimen; both arms will receive concurrent chemoradiotherapy with docetaxel, 5-FU, hydroxyurea, and hyperfractionated radiation. Finally, Posner and colleagues at the Dana-Farber Cancer Institute designed a phase III trial in which the control arm will be cisplatin at 100 mg/m2 for two cycles plus accelerated boost radiotherapy-based on the phase II Radiation Therapy Oncology Group experience-and the experimental arm will be cisplatin/5-FU/ docetaxel for three cycles, based on the Dana-Farber experience,[109] followed by radiation plus chemotherapy. The concomitant chemoradiotherapy regimen in the experimental arm is determined according to the response to induction chemotherapy: Patients who achieve a CR will receive conventional radiation plus carboplatin, but patients with less than a CR will receive a more aggressive regimen with accelerated concomitant boost radiotherapy plus docetaxel.
Conclusions
Induction chemotherapy remains experimental in head and neck cancer. It is likely that certain patient groups derive benefit from induction chemotherapy, but this has yet to be proven in randomized trials. Concurrent chemoradiotherapy has evolved as a standard primary or postoperative treatment in locoregionally advanced head and neck cancer. With the evolution of standard therapy for locoregionally advanced squamous cell head and neck cancer, the role of induction chemotherapy needs to be revisited. It is postulated that induction chemotherapy may optimize disease control by eradication of locoregional and distant micrometastasis. Distant control is particularly important in the setting of improved locoregional control rates seen with concurrent chemoradiotherapy. Our review identified several active combination regimens in head and neck cancer. Cisplatin/5-FU, which has long been a standard for induction chemotherapy, is being surpassed by platinum/ taxane-based combinations. Carboplatin/paclitaxel appears to be better tolerated than PF and achieves similar response rates, although no randomized study has yet compared the two regimens. On the other hand, carboplatin/ 5-FU has proven inferior to cisplatin/ 5-FU. Docetaxel is another taxane with promising activity in head and neck cancer. The triple combination of taxane/cisplatin/5-FU was shown to be superior to cisplatin/5-FU in two randomized trials; results of a third study are pending. For many authors, docetaxel/cisplatin/5-FU is an emerging standard for induction chemotherapy. With newer, more effective systemic therapies, the rationale for revisiting the role of induction therapy is strengthened. Currently, a number of randomized trials are being designed to define the role of induction chemotherapy in head and neck cancer when the primary treatment is concurrent chemoradiotherapy. These trials are of major importance for the evolution of head and neck cancer therapeutics. Finally, platinum/taxane combinations may be good platform regimens with which to investigate the addition of novel therapeutics in head and neck cancer, including molecular-targeted therapies such as inhibitors of epidermal growth factor receptor. Such clinical studies are expected to attract the interest of investigators in the coming years.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
