INTRODUCTION

Two randomized phase III trials^1,2 have demonstrated a survival benefit for concurrent chemoradiotherapy (CT + RT) over sequential therapy in patients with locally advanced or inoperable non–small-cell lung cancer (NSCLC) with a favorable prognosis. Furuse et al¹ reported that concurrent CT + RT was superior to sequential therapy with a significant improvement in 5-year survival (16% v 9%). Similarly, Radiation Therapy Oncology Group (RTOG) trial 94-10 supports the role of concurrent chemotherapy (CT) and once-daily radiation, with 4-year survival of 21% v 12% for sequential therapy (P = .03).² Such intensive therapy, however, is associated with increased acute esophagitis.³ In a quality-adjusted survival (QAS) analysis of several RTOG lung cancer studies, Movsas et al⁴ found that the QAS of induction CT followed by RT was nearly equivalent to that of concurrent CT + RT. Reduction in esophageal or upper GI and lung toxicities led to the greatest improvement in QAS. This analysis provided the rationale to test the ability of amifostine (AM) to reduce the high rate of esophagitis encountered in concurrent CT + RT regimens.

AM (Ethyol; WR-2721; Medimmune Inc, Gaithersburg, MD) is dephosphorylated at the tissue site to a free thiol, WR-1065, which acts as a potent scavenger of oxygen free radicals induced by ionizing radiation.^5,6 Multiple small randomized trials have produced mixed results regarding the ability of AM to reduce the rates of esophagitis and/or pneumonitis in locally advanced NSCLC.^7-12 The primary purpose of this randomized study was to test the hypothesis that the cytoprotectant AM could reduce the severity of esophagitis during concurrent CT + RT within the setting of a large cooperative group trial. The secondary objective of this trial was to evaluate differences in patient-reported swallowing symptoms and quality of life between the two arms.

PATIENTS AND METHODS

RESULTS

Between September 1998 and March 2002, 243 patients were enrolled (average accrual, 5.7 patients per month) onto the study; 120 patients were randomly assigned to the AM arm and 122 patients were randomly assigned to the no-AM arm (one patient was ineligible). The baseline demographics were well balanced with no significant differences between treatment arms regarding patient age, sex, performance status, histology, stage, or mean esophageal length in the radiation field (Table 1). At a median follow-up of 30.8 months (range, 4 to 50.2 months), no differences were observed in median survival (AM, 17.3 months; no AM, 17.9 months; P = .87; Fig 2) or median time to progression (AM, 9.2 months; no AM, 9.2 months; P = .77).

The majority of patients received RT per protocol or with minor variations (80% on the AM arm v 85% on the no-AM arm; P = not significant). No significant difference was observed in the number of RT treatment breaks more than 5 days (15% for AM v 10% for no AM; P = .24). Ninety-five percent of patients on the AM arm and 98% on the no-AM arm received CT according to protocol or with a minor or acceptable variation; 29% of patients received AM per protocol (34% males v 23% females); 43% had a minor or acceptable deviation (45% males v 36% females). The remaining patients had either a major deviation (10% deemed acceptable; 12% unacceptable) or the use of AM was terminated (6%). There was greater compliance with AM among males (Wilcoxon test P = .03). The main reasons that AM was terminated prematurely were toxicities (27%) and patient refusal (24%).

Toxicities observed during induction CT and within 90 days from the start of CT + RT are listed in Tables 2 and 3, respectively. The AM arm featured significantly higher rates of acute nausea (P = .03), vomiting (P = .007), cardiovascular toxicity, mostly transient hypotension (P = .0001), and acute infection or febrile neutropenia (P = .03). Nevertheless, patients experienced significantly less weight loss with AM (3.2% v 4.5%; P = .045). No significant differences in late toxicities were noted between arms (Table 4).

Regarding the primary end point of esophagitis, two of the three measurements did not show significant differences in favor of AM. There was no difference in the maximum CTC esophagitis grade between the two arms (Table 3). Thirty percent of patients in the AM arm developed esophagitis ≥ grade 3 versus 34% on the no-AM arm (P = .9). Similarly, there was no difference between the arms based on the PDL. However, based on the patient's daily swallowing diaries (Fig 3), the swallowing symptom AUC during CT + RT was significantly lower with AM (z test P = .025; Table 5), indicating that patients in the AM arm perceived less difficulty swallowing. Among patients older than 65 years and women, this reduction in swallowing symptoms with AM was more prominent (P = .003 and P = .006, respectively). The patient swallowing AUC was significantly lower with AM given at or close to protocol specifications than without AM (P = .039). This decrease was not observed among patients who received AM with a major deviation compared with those who did not receive AM (P = .85; Table 5).

The pretreatment EORTC QLQ-C30 and LC-13 forms were completed in 108 of 120 patients (90%) in the AM arm and 113 of 122 patients (93%) on the no-AM arm. Of these patients, 66% and 60% completed the QOL forms at the 6-week follow-up, respectively. Overall, the global, functional, and symptom QOL scales were not significantly different between the two arms. However, the symptom of pain (on the EORTC QLQ-C30 form) showed more clinically meaningful (≥ 10 points) improvement (24 of 71 [34%] AM v 14 of 68 [21%] no AM) and less deterioration (17 of 71[24%] AM versus 34 of 68[51%] no AM) at 6 weeks follow-up (v pretreatment) in favor of the AM arm (P = .003). A more in-depth discussion of correlates of QOL end points will be the subject of a separate article.

DISCUSSION

The primary objective of this randomized trial was to test the ability of AM to reduce the rate of esophagitis in patients receiving concurrent hyperfractionated RT + CT for locally advanced or inoperable NSCLC. Overall, AM did not reduce esophagitis ≥ grade 3 (per the NCI-CTC criteria) in this setting. Similar to prior studies,^7-12 RTOG 98-01 showed no evidence of tumor protection from AM: median survival and time to progression results were similar for both arms.

Four randomized phase II or III studies of AM in lung cancer have demonstrated the ability of this agent to reduce esophagitis or pneumonitis (Table 6). All of these trials used clinician-rated assessments of esophagitis. Antonadou et al⁷ randomly assigned 146 lung cancer patients receiving thoracic radiation to receive or not receive AM (340 mg/m² IV before daily RT). They found significant reductions in esophagitis ≥ grade 2 and pneumonitis (P < .001). In a subsequent randomized phase II trial by the same group,⁸ 73 stage III NSCLC patients received concurrent CT (either paclitaxel 60 mg/m² or carboplatin AUC 2) once weekly during standard RT with or without AM (300 mg/m² IV daily). They reported a significant reduction in esophagitis ≥ grade 3 and pneumonitis.⁸ Komaki et al¹⁰ reported the results of a randomized trial of AM in 62 patients who received concurrent CT (cisplatin and oral etoposide) and HRT (1.2 Gy bid to 69.6 Gy). In this study, AM was administered 500 mg IV (before the morning RT fraction) for the first 2 days of each week. Esophagitis ≥ grade 3 was significantly lower in the AM arm (16% v 35%; P = .02), as was the rate of severe pneumonitis (0% v 16%; P = .02). In a randomized phase II study testing subcutaneous AM (500 mg daily before RT), Koukourakis et al⁹ reported a reduction in esophageal toxicity among 60 patients receiving thoracic RT (WHO grade 0 to 1 v 3 to 4; P = .08; grade 0 to 1 v 2 to 4; P = .02).

Two other randomized phase III trials in stage III NSCLC did not demonstrate a significant benefit of AM. Leong et al¹¹ studied 60 patients treated with two cycles of paclitaxel/carboplatin followed by radiation (64 Gy) with concurrent weekly paclitaxel (60 mg/m²). In this double-blinded study, patients were randomly assigned to receive 740 mg/m² IV AM versus placebo before each dose of CT. Although Leong et al¹¹ noted a trend toward less grade 3 esophagitis in the AM arm (5% compared with 15% in the control arm), this finding was not statistically significant.¹¹ Moreover, the preliminary results of another randomized phase III trial in stage III NSCLC (N = 100) of concurrent RT (64.8 Gy) with weekly paclitaxel (50 mg/m²) and carboplatin (AUC 2) followed by two cycles of gemcitabine and cisplatin with or without AM (200 mg/m² IV daily before RT and 500 mg IV weekly before CT) has so far demonstrated no benefit of AM in reducing esophagitis.¹² Because a subsequent analysis with longer follow-up in more than 100 patients also demonstrated lack of esophageal protection with AM, this randomized trial was terminated early (N. Senzer, personal communication, May, 2004).

Of all of the studies testing the role of AM in lung cancer, RTOG 98-01 was the largest and was performed in the setting of a multi-institutional cooperative group protocol. It is not uncommon for a strategy initially reported as promising in small, single-institution trials subsequently to prove negative in the context of a larger collaborative group effort. For example, intense third-generation CT regimens were believed to be superior to standard (cyclophosphamide, doxorubicin, vincristine, and prednisone) CT in advanced non-Hodgkin's lymphoma based on several small trials,^22-24 until this was disproven in a large randomized, cooperative group trial.²⁵

Nevertheless, it is important to analyze potential differences between RTOG 98-01 and the prior positive studies. One important issue in RTOG 98-01 relates to compliance with the AM delivery: 72% of patients received AM per protocol or with a minor or acceptable variation. The difficulty with the AM compliance seemed to be due to its underlying toxicity (mostly nausea or vomiting, transient hypotension, and infection or febrile neutropenia) and/or patient refusal. Similarly, Leong et al¹¹ reported significant increases in hypotension, vomiting, and chills in the AM arm. It has been suggested that subcutaneous administration of AM may reduce its associated toxicities. In a randomized phase II study, AM (500 mg) was administered subcutaneously before daily RT.⁹ Overall, there was little hypotension or nausea; yet 15% had AM interrupted because of fever, rash, or asthenia. Among thoracic RT patients (N = 60), there was a significant reduction in esophageal toxicity.⁹ Interestingly, in RTOG 98-01, patient-reported swallowing symptoms (based on the AUC analysis) were significantly lower only in those who received AM per protocol or with a minor deviation (Table 5). Less weight loss also corresponded with AM use (3.2% v 4.5%; P = .045), although this finding may have been due partly to the increased administration of intravenous fluids in the AM arm.

Unlike most of the prior positive trials with AM, RTOG 98-01 mandated HRT. The AM dose and scheduling used in RTOG 98-01 was a modification of the approach by Buntzel et al.²⁶ In that study, head and neck cancer patients treated with RT (daily to 60 Gy) and concurrent carboplatin (70 mg/m² days 1 to 5, and 21 to 25) were randomly assigned to receive AM (500 mg before each carboplatin dose) versus placebo. They reported a significant decrease in dysphagia (P = .005), hematologic toxicity (P = .002), and mucositis (P < .001). In RTOG 98-01, AM was administered four times per week; theoretically, 40% of all RT fractions should have been protected by AM, whereas in the study by Buntzel et al,²⁶ only 33% of all RT fractions were preceded by AM. It is possible that more RT fractions need to be directly protected by AM (and/or a higher dose of AM used) to have an impact on CT + RT esophagitis. However, Komaki et al¹⁰ demonstrated a significant reduction in esophagitis and pneumonitis using AM (500 mg IV) only twice weekly in the context of HRT and concurrent CT for NSCLC.

The optimal method for administering AM in the context of HRT protocols remains unknown. In RTOG 98-01, the initial plan was to deliver AM before the first daily fraction of radiation. However, this led to a long treatment day, which adversely affected accrual. After the first 15 patients, the RTOG Lung Steering Committee decided to modify the study, administering the AM before the afternoon RT treatment. In a recent preclinical study in which rodents received 4 Gy bid over 5 days, AM was administered subcutaneously either at 50 mg/kg (equivalent to 325 mg/m² in humans) or 100 mg/kg (equivalent to 650 mg/m² in humans) delivered either before the morning radiation fraction, the afternoon RT fraction, or both.²⁷ At the 100 mg/kg dose, a single dose in the morning seemed to provide better radioprotection than a single dose in the afternoon. However, this was not the case for the 50 mg/kg dosage (closer to the dosing in RTOG 98-01). More recent preclinical data provide support for AM administration before each RT dose.²⁸ Although the timing issue deserves additional study, RTOG 98-01, in using AM four afternoons per week, should have been able to achieve similar results to those of Komaki et al,²⁹ in which the AM was administered only two mornings per week. It also remains unclear whether future studies should continue using a flat dose of 500 mg or base dosage on body-surface area. In women, paradoxically, there seemed to be some benefit with AM despite lower compliance than in males; this may be based on the relative lower body-surface area of females.

Subsequent to the implementation of RTOG 98-01, the role of HRT for NSCLC has been called into question. RTOG 94-10² has since demonstrated that optimal 4-year survival was achieved with concurrent CT and once-daily RT (21%) rather than twice-daily (17%) radiation. However, the long-term results of another randomized trial suggest that HRT and CT may still be worthwhile in locally advanced NSCLC, with a promising 4-year survival result of 24%.³⁰ More recently, Belani et al³¹ reported a median survival of 20 months with induction CT followed by RT delivered tid. Currently, the standard of care for patients with good performance status with locally advanced NSCLC is concurrent CT and once-daily radiation.

A key strength of RTOG 98-01 is that it used a variety of end points to study esophagitis, including the NCI-CTC toxicity criteria, a PDL, and a patient daily swallowing diary. Although AM yielded no impact on the clinically based NCI-CTC toxicity end point or the similar PDL criteria, significant differences were noted in the patient swallowing diaries. On the basis of these diaries (Table 5), the swallowing symptom AUC during CT + RT was significantly lower with AM (P = .025). Although a placebo effect cannot be excluded, the fact that the swallowing AUC values started at the same level (at the beginning of CT + RT) and only diverged toward the end of CT + RT (when one would expect esophagitis to become clinically manifest) suggests this is not a placebo effect, which typically occurs early but dissipates over time (Fig 3).

This study suggests that our current clinical toxicity end points, primarily reflecting the physician's and not the patient's perspective, may not be sufficiently sensitive. Indeed, RTOG 98-01 is unique in that it is the only study of AM in lung cancer that included a prospective, validated, QOL instrument.^13,14 Although overall QOL was not significantly different between the two arms, the symptom of pain showed significantly more clinically meaningful improvement and less deterioration at 6 weeks of follow-up (v pretreatment) in the AM arm (P = .003). It is possible that patients have difficulty separating swallowing symptoms from the overriding symptom of pain; this hypothesis will be explored in a separate article focusing on QOL.

In conclusion, RTOG 98-01 does not support the hypothesis that AM reduces esophagitis in the context of CT and twice-daily RT for patients with locally advanced NSCLC. However, patient-derived self-assessments suggest a possible advantage to AM that should be explored further. Preliminary results seem promising with subcutaneous AM,⁹ a simpler method of administration that produces less nausea, vomiting, and hypotension. Other strategies to minimize esophagitis should also be investigated. Emerging data using three-dimensional conformal radiation with CT suggest promising results with significantly less esophagitis.³² Additional improvements may be obtained by minimizing esophageal exposure with more sophisticated technology, such as intensity-modulated RT.³³ A novel biologic strategy under development involves the use of manganese superoxide dismutase.³⁴ Ultimately, the goal of treatment must be to improve not only the quantity, but also the quality of survival of patients with lung cancer.

Authors' Disclosures of Potential Conflicts of Interest

The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Consultant/Advisory Role: Corey Langer, Bristol-Myers Squibb, MedImmune; Maria Werner-Wasik, MedImmune; Ritsuko Komaki, MedImmune; Mitchell Machtay, MedImmune; Todd Wasserman, MedImmune. Stock Ownership: Maria Werner-Wasik, MedImmune; Rita Axelrod, Bristol-Myers Squibb. Honoraria: Maria Werner-Wasik, MedImmune; Ritsuko Komaki, MedImmune Oncology; Mitchell Machtay, MedImmune; Benjamin Movsas, MedImmune; Todd Wasserman, MedImmune. For a detailed description of these categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and Disclosures of Potential Conflicts of Interest found in Information for Contributors in the front of each issue.