To the Editor:

In the November 20, 2005, issue of Journal of Clinical Oncology, Ho et al¹ speculated that gefitinib-induced hepatitis with subsequent necrosis and fibrosis is inevitable when gefitinib is reinitiated with once-daily dosing, even with concurrent corticosteroids, necessitating treatment withdrawal despite marked clinical benefit.

To date, gefitinib-induced hepatitis has received little attention. However, in four phase I trials,^2-5 hepatotoxicity was a dose-limiting toxicity of gefitinib, as was diarrhea. Furthermore, in the Iressa (AstraZeneca, London, United Kingdom) Dose Evaluation in Advanced Lung Cancer (IDEAL 1) trial,⁶ 2% of patients receiving gefitinib alone at a dosage of 250 mg per day had grade 3 or 4 elevations of hepatic enzymes necessitating cessation of treatment. Therefore, countermeasures should be developed so that gefitinib-responders can continue treatment despite hepatotoxicity.

From September 2002 through December 2005, we administered oral gefitinib, 250 mg once daily, to 87 previously treated patients with advanced non–small-cell lung cancer (NSCLC) at our institution. One of these patients, a 61-year-old nonsmoking Japanese woman with lung adenocarcinoma, showed grade 2 elevation of hepatic enzymes after receiving gefitinib for 8 weeks. She had no previous history of liver disease or metastasis to the liver. The WHO performance status (PS) remained 0. During this period, a near complete response (CR) of the primary tumor and multiple lung metastases was confirmed. However, when gefitinib was discontinued owing to hepatotoxicity, the adenocarcinoma progressed and liver function improved in the following 4 weeks. As did Ho et al, we reinitiated gefitinib while carefully monitoring liver function; however, liver function deteriorated again to grade 2 within 1 week and to grade 3 over the following 5 weeks. Therefore, gefitinib was withdrawn for a second time despite the adenocarcinoma having regressed within 2 weeks of reinitiation. In the 6 weeks following withdrawal, liver function normalized but the lung cancer progressed. At this time, mutation analysis of lung cancer specimens obtained before first-line chemotherapy showed the presence of exon 19 deletion (at codons 746 through 750) of the epidermal growth factor receptor (EGFR) gene.⁷ Because the patient was a nonsmoking Japanese woman with a PS of 0 and mutation-positive adenocarcinoma that had not responded to previous platinum-based chemotherapy, we believed continuing the molecular targeted therapy was reasonable.^6-8 Then, unlike Ho et al, we reinitiated oral gefitinib, 250 mg once every 5 days, after receiving full informed consent for an unproven treatment schedule. Within 2 weeks, a near CR of the cancer was achieved with, at most, grade 1 elevation of hepatic enzymes. So far, a near CR with stable grade 1 hepatotoxicity has been maintained for 8 weeks with gefitinib administered according to this promising new treatment schedule.

Our hypothesis for the expected reduction of hepatotoxicity when gefitinib is administered once every 5 days is as follows. (1) Development of hepatitis depended on the dosage of gefitinib in the IDEAL 1 trial,⁶ where grade 1 to 4 elevations of hepatic enzymes were more common in patients receiving a dose of 500 mg per day (24%) than in patients receiving 250 mg per day (13%). (2) Furthermore, such hepatic toxicity was more common in Japanese patients (22%) than in non-Japanese patients (4%).⁶ Considering that, on average, Japanese patients weigh less than American patients, the reduced dosage of gefitinib might be preferable in Japanese patients, especially in those responding to gefitinib, although hepatotoxicity might also be due to ethnic differences in the hepatic metabolism of gefitinib. (3) In two phase I trials the maximum plasma concentration (C_max) and the area under the plasma concentration time curve (AUC) were dependent on the number of consecutive days gefitinib was administered,^2,3 such that the C_max and the AUC following administration of gefitinib at a dose of 225 mg per day for 14 consecutive days were 2.3-fold and 3.0-fold, respectively, those following a single 225 mg dose.³

Similarly, our hypothesis for the expected response to the reduced dose of gefitinib administered according to such a schedule is as follows. (1) When gefitinib was administered intermittently (for 14 consecutive days, every 28 days) to patients with NSCLC in two phase I trials, partial response (PR) was observed in 22% to 25% of patients, and stable disease (SD) was observed in 9% to 19% of patients,^2,3 without tumor regrowth being observed during the off treatment period. These results suggest that the efficacy of gefitinib would be maintained when treatment must be interrupted owing to any toxicity developing from continuous administration. (2) The mean terminal half-life (T_1/2) was not dependent on the number of consecutive days gefitinib was administered in two phase I trials,^2,3 where, obviously distinct from above-mentioned pharmacokinetics of C_max and AUC, the T_1/2 of 40 ± 8 hours after administration of gefitinib for 14 days at 225 mg per day was almost equal to the T_1/2 of 30 ± 5 hours after administration of a single 225 mg dose.³ (3) Antitumor activity was not dependent on the dosage of gefitinib in previous clinical trials.^2-6,8 In fact, PR rates achieved with doses of 250 mg per day and 500 mg per day were 18% and 19%, respectively, in the IDEAL 1 trial⁶ and 12% and 9%, respectively, in the IDEAL 2 trial.⁸ Furthermore, even when the dosage of gefitinib was less than 250 mg per day, PR or SD was achieved in all four phase I trials.^2-5 In particular, SD was maintained in one patient with NSCLC who received 50 mg of gefitinib per day for 14 consecutive days every 28 days. (4) In addition to a previous response to a lower dose of gefitinib, predictors of a favorable response to gefitinib identified in the IDEAL 1 and 2 trials were Asian ethnicity, female sex, nonsmoker status, good PS, and adenocarcinoma.^6,8 Moreover, a high EGFR gene copy number, identified by fluorescence in situ hybridization, and somatic mutations of the EGFR gene are possible surrogate biologic markers for response to gefitinib.⁷ Therefore, patients who have responded to daily gefitinib but have severe hepatotoxicity should be considered as candidates for treatment with our regimen if they also have several factors predicting a favorable response.

We have since treated a second Japanese nonsmoking woman with lung adenocarcinoma and a PS of 1 who showed grade 3 elevation of hepatic enzymes after daily administration of gefitinib. She has maintained SD and grade 0 hepatotoxicity with our new schedule for gefitinib. Regrettably, we have no evidence that our gefitinib schedule confers a survival benefit. Therefore, we believe clinical trials are warranted to evaluate this intermittent schedule following severe hepatotoxicity from daily administration of gefitinib, through the comparison of overall survival rates stratified by the presence of either EGFR gene mutations or high EGFR gene copy number. We recommend our intermittent schedule for patients who have responded to daily gefitinib but have discontinued treatment because of severe hepatotoxicity, especially if they have several factors predicting a favorable response.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.