Promising results from a Phase II trial of induction chemotherapy and sequential radiotherapy for advanced nasopharyngeal carcinoma (NPC) at The University of Texas M. D. Anderson Cancer Center (Houston, TX) and two retrospective reviews of the authors' historical experience with NPC demonstrated that distant failure was directly correlated with advanced lymph node status. Furthermore, local control was excellent for patients with T1–3 disease that was managed with radiation alone or with a sequential approach involving chemotherapy. Neoadjuvant chemotherapy (primarily with cisplatin + 5-fluorouracil) was associated with a significantly decreased risk of distant metastasis and with improved survival. Based on these findings, the authors evaluated a novel induction regimen involving docetaxel and carboplatin for patients with previously untreated T1–2N2–3M0 NPC.
Docetaxel (80 mg/m2 on Day 1) and carboplatin (to an area under the time-concentration curve of 6 on Day 1) were administered every 21 days for 3 cycles, after which radiotherapy was administered. NPC was restaged with magnetic resonance imaging and nasopharyngoscopy 3 weeks after the completion of chemotherapy and 6 weeks after the completion of radiotherapy.
Over 5 years, 18 patients were enrolled in the study. Grade 4 neutropenia and Grade 2 fatigue were observed in 51% and 28% of chemotherapy courses, respectively. After chemotherapy, 2 patients had complete responses, 14 had partial responses, 1 had a minor response, and 1 had progressive disease. The latter two patients and one patient who had a partial response underwent off-study chemoradiotherapy. After radiotherapy or chemoradiotherapy, 12 patients had complete responses and 6 had partial responses. Seven patients had recurrent disease; two had local recurrences, and five had distant metastases.
Nasopharyngeal carcinoma (NPC), although infrequent in North America, is a common malignancy in other areas of the world, such as southern China. NPC with World Health Organization (WHO) type 2 or 3 histology is distinct from other squamous malignancies of the head and neck. Patients with this type of NPC are more likely to present with extensive lymph node involvement and have a higher incidence of eventual distant metastases; in addition, this type of NPC tends to be more chemosensitive and radiosensitive. Historically, early-stage disease has been treated successfully with radiotherapy; however, in cases of advanced disease, although radiotherapy results in high initial response rates, disease recurrence is common.1 In a retrospective review of 378 patients (75% of whom were white) who received radiotherapy alone at The University of Texas M. D. Anderson Cancer Center (Houston, TX) between 1954 and 1992, the actuarial survival rates at 5, 10, and 20 years were 48%, 34%, and 18%, respectively. Overall, 27% of patients developed distant metastases, and 49% died of NPC.2 Qin et al.3 documented a nearly identical 5-year survival rate in Chinese patients who received radiotherapy alone. The combined data from two retrospective reviews at The University of Texas M. D. Anderson Cancer Center suggested that advanced lymph node status was directly correlated with risk of distant metastasis and that radiotherapy alone was quite effective in achieving locoregional control of T1–3 primary tumors.2, 4 Furthermore, neoadjuvant chemotherapy (most commonly with cisplatin + 5-fluorouracil) was found to be associated with a reduced risk of metastasis and with improved survival in a matched-cohort study conducted by Geara et al.2
Because of the biology and natural history of NPC, several investigators have studied the addition of chemotherapy to radiotherapy in an effort to reduce distant metastasis rates, improve local control, and increase survival. A recent metaanalysis of 1528 patients from 6 randomized trials of chemotherapy and radiotherapy versus radiotherapy alone confirmed that chemotherapy provided benefit for patients with locally advanced NPC.5–10 The addition of chemotherapy improved disease-free survival rates by 37%, 40%, and 34% at 2, 3, and 4 years, respectively, with all improvements being statistically significant. Furthermore, chemotherapy improved overall survival rates by 20%, 19%, and 21% at 2, 3, and 4 years, respectively; however, only the improvement observed at 4 years was statistically significant.11 In addition to the randomized studies included in the metaanalysis, several other nonrandomized studies have suggested that there is a benefit associated with the addition of neoadjuvant chemotherapy to radiotherapy.4, 12–16 Nonetheless, it has been relatively difficult to demonstrate that neoadjuvant chemotherapy is associated with a definite benefit, especially in areas in which NPC is endemic.17, 18 In North America, primarily because of the results of the Intergroup 0099 Phase III trial, concurrent chemoradiotherapy with cisplatin and adjuvant chemotherapy with cisplatin + 5-fluorouracil have become standards of care for patients with advanced NPC.6
In the current study, we evaluated a risk-based approach that was based on experience at The University of Texas M. D. Anderson Cancer Center suggesting that patients with advanced N status and early T status would benefit from neoadjuvant chemotherapy. Specifically, we sought to reduce the risk of distant recurrence while eliminating the need to add concurrent chemotherapy to radiotherapy; this strategy was based on a predicted high rate of locoregional control with radiotherapy alone. We used a combination of docetaxel and carboplatin, based on the activity of docetaxel as a single agent against squamous cell carcinoma of the head and neck and on the activity and tolerability associated with this combination in other solid tumors. We hypothesized that this regimen would improve on the efficacy of cisplatin + 5-fluorouracil while reducing toxicity, logistic complexity, and the amount of inconvenience experienced by the patient.
MATERIALS AND METHODS
Patient Eligibility and Evaluation
Patients who had histologically proven NPC with known WHO classification and T1–2N2–3M0 disease were eligible for the current study. Patients were required to have measurable disease, including at least 1 bidimensionally measurable lesion, no previous chemotherapy, a life expectancy of at least 3 months, Zubrod performance status ≤ 2, absolute granulocyte count ≥ 1500 per μL, platelet count ≥ 100,000 per μL, a total serum bilirubin level within normal limits, a serum glutamic-oxaloacetic transaminase level ≤ 1.5 times the upper limit of normal (ULN), and an alkaline phosphatase level ≤ 5 times the ULN. Patients were excluded for having any previous malignancy, except for adequately treated basal or squamous cell skin malignancy, cervical malignancy in situ, and any other malignancy from which the patient had been free of disease for more than 5 years. Pregnant and lactating women also were excluded. Patients who previously received radiotherapy to the head and neck, except for those who had nonmelanomatous skin malignancies outside the radiotherapy treatment volume, were not eligible. Patients were informed of the investigational nature of the study and provided written informed consent in accordance with institutional guidelines.
Pretreatment evaluation included a detailed medical history, physical examination, and assessment of performance status. A fiberoptic examination of the upper aerodigestive tract, including the nasopharynx, oropharynx, and larynx, was completed ≤ 4 weeks before treatment. A chest X-ray and a computed tomography (CT) or magnetic resonance imaging (MRI) scan of head and neck, including the cervical and supraclavicular lymph node–bearing regions, were obtained within 3 weeks of entry into the study. A sonogram or CT scan of the abdomen and a bone scan were obtained if the patient's alkaline phosphatase levels were elevated. For all patients, laboratory values, including a complete blood count (CBC) with differential, a platelet count, and albumin, calcium, electrolyte, blood urea nitrogen (BUN), creatinine, alkaline phosphatase, total bilirubin, lactate dehydrogenase (LDH), and serum glutamic-pyruvic transaminase (SGPT) levels, were measured before enrollment; pregnancy testing was performed when appropriate. Patients were examined by a multidisciplinary team before study entry. Neck dissection was considered for patients with persistent adenopathy following irradiation. An oncologic dentist evaluated every patient before treatment. Teeth were extracted when it was deemed necessary to do so, and a prophylactic fluoride treatment program was initiated.
Chemotherapy Treatment Plan and Dose Modifications
Docetaxel (80 mg/m2) was administered via a 1-hour intravenous infusion and was immediately followed by carboplatin (to an area under the time-concentration curve of 6) administered intravenously over 30 minutes on Day 1 of each 21-day cycle. (Carboplatin doses were determined using the Calvert formula and the Cockcroft–Gault formula for creatinine clearance.19, 20) These were the starting doses for each patient (dose level 0). In total, three courses of induction chemotherapy were planned.
Patients received 8 mg dexamethasone orally twice daily for 3 consecutive days, beginning 1 day before the start of each treatment cycle, to minimize the likelihood of an anaphylactic reaction or fluid retention. Intravenous administration of docetaxel was not accepted in lieu of oral administration. Chemotherapy was administered based on the patient's actual body weight–derived surface area, with no increase in the dosage of either agent for any weight gain that occurred during the study. In contrast, dose reduction was recommended when weight loss occurred. Antiemetics were chosen at the discretion of the attending physician.
Patients received weekly CBCs and platelet counts. Measurement of electrolyte, magnesium, BUN, creatinine, alkaline phosphatase, LDH, SGPT, and total bilirubin levels was performed immediately before each cycle of induction chemotherapy. Patients were evaluated by a medical oncologist before each course of chemotherapy; evaluation included a directed medical history and physical examination.
Chemotherapy was held if the absolute neutrophil count (ANC) was < 1200 per mL or if the platelet count was < 100,000 per mL on Day 22 of a given cycle (i.e., Day 1 of the following cycle); CBCs and platelet counts were obtained weekly. Treatment was administered when the ANC was > 1200 per mL and the platelet count was > 100,000 per mL. If adequate recovery of these counts was not evident after 2 weeks of treatment delay, the patient was withdrawn from the study and referred for immediate radiotherapy.
Doses were adjusted according to the system that exhibited the greatest degree of toxicity. Toxic effects were graded using the National Cancer Institute (NCI) Common Toxicity Criteria. Doses that were reduced due to toxicity were not reescalated to their initial levels. In addition, treatment delays to allow recovery from nonhematologic toxicities did not last for more than 3 weeks. In the event that recovery from a toxic reaction required more than 3 weeks, chemotherapy was discontinued and radiotherapy was initiated. A maximum of two 25% reductions in docetaxel dose (i.e., from 80 to 60 mg/m2 and from 60 to 45 mg/m2) were allowed for each patient.
A 25% dose reduction for both chemotherapeutic agents (i.e., dose level −1) was required when any of the following occurred: Grade 4 neutropenia for > 5 days, febrile neutropenia, neutropenia with infection, or Grade 4 thrombocytopenia. Use of granulocyte–colony-stimulating factor in lieu of dose reduction was permitted in cycles following prolonged Grade 4 neutropenia or febrile neutropenia if either was the only dose-limiting toxicity. When bilirubin levels were elevated, treatment was held until they returned to normal. When transaminase levels were 2.5–5 times the ULN, the docetaxel dose was reduced by 25%. Doses were not reduced when transaminase levels were < 2.5 times the ULN, whereas docetaxel administration was held when transaminase levels were > 5 times the ULN. For Grade 3 or 4 neurotoxicity, patients were withdrawn from the study; for Grade 2 neurotoxicity, treatment was held until recovery to Grade 0–1 neurotoxicity occurred, with a 25% reduction in docetaxel dose in subsequent courses. Carboplatin doses in each course were calculated based on serum creatinine levels immediately before the start of the course and on calculated clearance. In the event of renal dysfunction, this calculation procedure adjusted carboplatin doses appropriately. No docetaxel dose adjustments due to renal dysfunction were necessary. There also were no dose reductions resulting from hypersensitivity reactions to docetaxel. If the initial reaction was a Grade 4 reaction according to the NCI Common Toxicity Criteria, the patient received no further treatment and was withdrawn from the protocol therapy. If a second severe reaction (Grade 3) occurred despite additional premedication, the patient was withdrawn from the study. In addition, no dose reductions resulted from edema. Patients who experienced new-onset or symptomatic edema or other signs of increasing fluid retention were treated with oral diuretics. Other toxic reactions that were less severe than Grade 2 were managed symptomatically, with no dose adjustments. If other toxicities (except for alopecia and anemia) were more severe than Grade 3, administration of the drug responsible for the toxicity was withheld until resolution of the reaction to Grade ≤ 1 or to baseline, after which chemotherapy was reinitiated with a dose reduction of 25%.
Radiotherapy Treatment Plan
Megavoltage equipment (either linear accelerators or 60Co machines) was used to supply appropriate photon energies (1–18 megavolts) and a wide range of electron energies (6–20 megaelectron-volts). Wedges and/or compensating filters were immobilized with Aquaplast casts (WFR/Aquaplast, Wyckoff, NJ). The radiation beam was shaped using customized Cerrobend blocking. Treatment verification with portal imaging for each new field was repeated at least once weekly. The primary tumor and upper neck were treated with lateral opposed fields that provided a wide margin. Additional fields and changes in field orientation were applied as required for disease in the orbital, ethmoid, and paravertebral areas. After the subclinical dose was administered, macroscopic disease was encompassed with a margin of at least 1 cm. An anterior field was used to treat the lower neck and supraclavicular fossa.
Radiotherapy was initiated 3 weeks after the completion of induction chemotherapy. Radiation was delivered via 1 fraction per day on 5 days per week for 7 weeks. The daily fraction size was 200 centigrays (cGy). Sites of macroscopic disease received a tumor dose of 6600–7000 cGy, whereas sites of subclinical disease received 5000 cGy. The following normal tissue tolerance levels were adhered to: spinal cord, 45 Gy; optic pathway, 54 Gy; and brain stem, 54 Gy. CT dosimetry was performed for all patients. Doses delivered to macroscopic disease sites had specified target volumes (the tumor plus a 1 cm margin). Dose inhomogeneity was limited to a maximum of 5%. If treatment was delivered via electrons, the doses administered to subclinical and macroscopic disease sites were 100% and 90%, respectively. The doses administered to the lower neck and supraclavicular fossa were specified, provided that there was no macroscopic disease in these sites. The attending radiation oncologist examined all patients at least once weekly during treatment, and more often if necessary, to assess and manage acute reactions to radiotherapy.
Evaluation of Response
Three weeks after the completion of induction chemotherapy and 6 weeks after the completion of radiotherapy, responses at the primary and lymph node sites were evaluated with a CT or MRI scan of the head and neck and with nasopharyngoscopy. A complete response (CR) was defined as the complete disappearance of all objective evidence (radiographic and physical) of disease. A partial response (PR) was defined as a decrease of ≥ 50% in the sum of the products of the diameters of all measured lesions, with no increase in the size of any lesion and no new lesions. Because exact measurement of nasopharyngeal lesions is not always possible, any visually significant decrease in lesion size was considered to indicate a PR. A minor response (MR), other than in the nasopharynx, was defined as any response that was less extensive than a PR. Progressive disease (PD) was defined as any unequivocal increase (≥ 25%) in the sum of the products of the diameters of all measurable lesions or the appearance of an unequivocal new lesion.
Survival duration was calculated from the first day of treatment until the date of death or last contact with the patient. The duration of disease-free survival was calculated from the date on which CR was documented until the date on which PD was diagnosed.
Statistical Methods and Study Design
Using historical data from Dimery et al.,16 who reported that cisplatin + 5-fluorouracil yielded a CR rate of 20%, we set a target CR rate of 40% for treatment with docetaxel + carboplatin. Based on the Simon optimal two-stage design, 17 patients were to be accrued in the first stage of the trial. If CRs after neoadjuvant chemotherapy were observed in 3 or fewer patients, the trial was to be closed, with the CR rate estimated to be ≤ 20%. If 4 or more patients experienced a CR, then additional patients would be accrued until the cohort included 37 patients. The regimen was to be deemed effective if 10 or more CRs were observed in the total study cohort. The power and α level for the sample size calculation and stopping rule were 80% and 0.05, respectively.
Eighteen patients were enrolled in the study between January 1998 and November 2002. Their baseline characteristics are shown in Table 1. All patients were evaluable for response and had T1–2N2–3M0 disease. A breakdown of the specific disease stages is shown in Table 2. Staging was performed according to the American Joint Committee on Cancer staging system.
Table 1. Patient Characteristics
WHO: World Health Organization.
Total no. of patients
No. evaluable for response
No. evaluable for toxicity
Median age in yrs (range)
WHO type 1
WHO type 2
WHO type 3
Performance status (%)
Table 2. Disease Stage Distribution (Based on the 1992 American Joint Committee on Cancer Staging System) among Patients with Nasopharyngeal Carcinoma
Seventeen patients completed all three courses of induction chemotherapy. The remaining patient was withdrawn from the study after two courses due to PD. Two patients underwent dose reduction due to neutropenic fever. The remaining patients received all treatment courses at dose level 0. A total of 53 courses were completed: 49 at dose level 0 and 4 at dose level −1. Three patients received concurrent chemoradiotherapy instead of the radiotherapy planned for use in the study; these three patients received cisplatin + 5-fluorouracil, carboplatin, or cisplatin as part of concurrent chemotherapy.
Treatment Response, Recurrence, and Survival
All 18 patients were evaluable for response to induction chemotherapy. After completing induction chemotherapy, 2 patients experienced CR, 14 experienced PR, 1 experienced MR, and 1 had PD. The overall response rate to chemotherapy was 89%. The patient who experienced PD, the patient who experienced MR, and one patient who experienced PR received subsequent chemoradiotherapy. After chemoradiotherapy, one of these three patients experienced CR, while the other two experienced PR. Of the 15 patients who received planned radiotherapy after induction chemotherapy, 11 experienced CR, whereas 4 had PR at final restaging (Table 3).
One patient, who had a PR after all treatment, was lost to follow-up 7 months after the completion of the trial, at which time she exhibited no evidence of recurrent disease. The other three patients who had a PR after radiotherapy exhibited no evidence of disease during long-term follow-up. Two patients had local recurrences. One of these two patients had a PR after treatment with docetaxel + carboplatin and a CR after all treatment; local recurrence developed 16 months after diagnosis, and the patient was free of disease 4 years after the administration of additional chemotherapy. The other patient had an MR after induction chemotherapy and received initial chemoradiotherapy; this patient currently is 6 months from completing additional radiotherapy with concurrent chemotherapy for suspected residual disease and is free of PD. Five patients have experienced distant metastases; two of these patients have died. To date, all recurrences have been observed in the first 2 years after diagnosis, and both deaths occurred in the third year after diagnosis. The median follow-up duration among survivors is 2 years. The estimated 3-year progression-free and overall survival rates were 54.3% (95% confidence interval, 33.8–87.2%) and 74.1% (95% confidence interval, 48.4–100%), respectively (Figs. 1, 2).
The only common hematologic toxicity was neutropenia. Grade 4 neutropenia was observed in 27 of 53 treatment courses (51%), febrile neutropenia was observed in 4 of 53 courses (8%), and Grade 3 neutropenia was observed in 11 of 53 courses (21%). There were no episodes of Grade 2, 3, or 4 anemia and no episodes of Grade 3 or 4 thrombocytopenia. In addition, there were no Grade 4 nonhematologic toxicities and no significant nonhematologic adverse events. Grade 3 nonhematologic toxicities were uncommon. The only common Grade 2 nonhematologic toxicities were fatigue, alopecia, and nausea (Table 4).
No Grade 3–4 nonhematologic toxicity was observed.
Out of a total of 18 patients.
Out of a total of 53 treatment courses.
Grade 4 neutropenia
Grade 2 nausea
Grade 2 emesis
Grade 2 edema
Grade 2 sensory neuropathy
Grade 2 ototoxicity
Grade 2 fatigue
Grade 2 stomatitis
Grade 2 alopecia
In the current Phase II study, we used a risk-based approach to estimate the complete and overall response rates and the survival rate among patients receiving docetaxel + carboplatin for NPC that was locally advanced based on lymph node status only. Although radiotherapy alone often is successful in controlling local and regional disease in patients with this type of NPC, the risk of distant metastasis is high; however, neoadjuvant therapy has been found to be associated with a decreased risk of distant recurrence.
We chose to investigate the docetaxel + carboplatin regimen in the hope that it would prove superior to the cisplatin + 5-fluorouracil regimen in terms of CR rate. The docetaxel + carboplatin regimen was quite active, resulting in an overall response rate of 89%. Nonetheless, based on CR rates and institutional historical controls,16 our experience indicates that this regimen is not likely to be superior to the cisplatin + 5-fluorouracil regimen.
One possible reason for the decreased CR rate in the current study compared with the CR rate reported by Dimery et al.16 is the evolution of imaging techniques (from CT to MRI) for the nasopharynx. Eleven patients (61%) in the current trial were evaluated with serial MRI, whereas all patients in the study conducted by Dimery et al. underwent CT. One would expect MRI, the more sensitive technique, to reduce the number of CRs by identifying residual disease that was not detectable by CT. Nonetheless, both patients who experienced CR after chemotherapy in the current study underwent MRI. The overall response rates for the two regimens were similar (89% vs. 93%), and the patient populations were comparable with respect to WHO classification, age, gender, and race. Furthermore, the current, specifically targeted population had NPC with advanced N status and early T status, whereas patients in the study conducted by Dimery et al. had T1–4 disease, with no lymph node involvement evident in 11% of the study population.
Although the docetaxel + carboplatin regimen produced a lower CR rate compared with 5-fluorouracil + cisplatin in historical controls, response rates were similar to those observed in other chemotherapy trials involving such patient populations. Specifically, in 15 trials involving patients with locally advanced NPC, the overall rates of response to combination chemotherapy ranged from 62% to 98%, while CR rates ranged from 5% to 66%.1 These results are comparable to the 89% overall response rate and the 11% CR rate observed in the current study. In addition, 17 patients with locally advanced NPC who were treated with docetaxel, cisplatin, 5-fluorouracil, and leucovorin (TPFL) had survival durations similar to those observed in patients in the current study.21 Of the 17 patients who were treated with TPFL, 10 were alive and without evidence of disease, 2 had local recurrences, and 5 had distant metastases at the conclusion of the study. These results are nearly identical to the results of the current study, in which 11 patients are alive and without evidence of disease, 2 patients had local recurrences, and 5 patients had distant metastases.
The taxoids paclitaxel and docetaxel both produce an overall response rate of approximately 40% in patients with recurrent squamous malignancies of the head and neck.22, 23 Paclitaxel and docetaxel are the most active single agents identified for treating this type of disease, which typically is much less chemosensitive than NPC is in the recurrent setting. To our knowledge, at the beginning of the current trial, there were no published data on the activity of taxoids against NPC. Since then, several small, nonrandomized studies have investigated the use of paclitaxel, usually in combination with carboplatin, to treat advanced, recurrent, or metastatic NPC.24–29 This combination has been shown to have significant activity against NPC, with response rates of 22% for paclitaxel alone and 57–75% for paclitaxel in combination with other agents as first-line therapy for recurrent or metastatic NPC.24, 26, 29 In a recent Phase II trial, docetaxel and cisplatin exhibited little activity in patients with recurrent or metastatic NPC; however, only nine patients were enrolled in this study, and some had received previous cisplatin therapy.30 It is clear that most studies on the use of taxoids with platinum compounds to treat NPC have yielded promising results.
Also noteworthy is that the docetaxel + carboplatin regimen represents a less toxic, more convenient outpatient regimen compared with cisplatin + 5-fluorouracil; the former regimen does not have logistic complications such as 5 days of continuous intravenous infusion and central access, the requirement of frequent hospitalization for administration, and associated mucosal, renal, neurologic, and emetic toxicities. The current study confirmed that the docetaxel + carboplatin regimen was well tolerated and conveniently administered.
In the current study, we used a risk-based approach that was founded on the results of studies demonstrating that patients who have NPC with advanced N status and early T status require systemic chemotherapy to reduce the risk of distant metastases but do not require concurrent chemoradiotherapy, due to the excellent locoregional control achieved with radiotherapy alone.2, 4, 16 The current regimen yielded a CR rate that was less than the rate yielded by 5-fluorouracil and cisplatin in historical controls, although the response and survival rates observed in the current study were similar to those observed in several other trials involving patients with locally advanced NPC.1, 21 In addition, the docetaxel + carboplatin regimen was well tolerated. We believe that the current risk-based approach merits further study; however, given that 5 years were required to enroll 18 patients in the current study, a multiinstitutional effort will be necessary for further trials involving patients with this type of NPC in geographic areas in which the disease is not endemic.