Phase II preradiation R115777 (tipifarnib) in newly diagnosed GBM with residual enhancing disease

Abstract

Glioblastoma multiforme (GBM) is a lethal primary malignant brain tumor in adults. R115777 (tipifarnib) is an oral agent with antiproliferative effects, being a potent and selective inhibitor of farnesyltransferase. This multicenter, open-label phase II study was designed to evaluate the efficacy and safety of R115777 given after surgery and prior to radiation in patients with newly diagnosed and residual enhancing GBM. Following surgery, an MRI confirmed the presence of residual enhancing tumor. Patients on enzyme-inducing antiseizure drugs (EIASDs) received 600 mg twice per day, and those not on EIASDs received 300 mg twice per day. One to three monthly cycles of R115777 were administered, and radiation was initiated with progression or after three cycles. A cycle consisted of 3 weeks of continuous R115777 followed by a 1-week rest. MRI was done monthly. The primary end point was overall survival; secondary end points were tumor response rate and toxicity. A total of 28 confirmed GBM patients entered the study; 15 patients (54%) were on EIASDs. The overall median time of survival was 7.7 months. There were no tumor responses. Eight patients (29%) had stable disease as the best response. The study was stopped early due to progression of the disease in 12 patients (48%). A total of 24 patients (85%) were off study before the planned treatment schedule for radiation therapy. R115777 administered prior to radiation therapy in patients with newly diagnosed GBM and residual enhancing disease did not result in any measurable responses or improvement in survival. R115777 administered prior to radiation therapy is not recommended for patients with newly diagnosed GBM.

Glioblastoma multiforme (GBM) is the most common primary malignant brain tumor in adults and also the most lethal.¹ It is relatively treatment resistant, with an estimated median survival of 10–12 months. Surgical resection is usually the initial treatment, and there are data indicating that a gross total resection leads to a better result than a lesser resection.² Radiation has a definite role in treatment and, when given after surgery, can prolong survival.³ Randomized trials have demonstrated a dose–response curve up to 60 Gy. Higher doses have not been proven to be of benefit. Chemotherapy, until recently, was shown to be of only limited benefit, as documented in a meta-analysis.⁴ The best results to date have been reported in a randomized study of patients receiving daily temozolomide with radiation to 60 Gy followed by temozolomide, compared to radiation alone.⁵

Genetic mutations in tumor cells may lead to the development of resistance to cancer therapies.⁶ One strategy for reversing intrinsic resistance to treatment is to use therapies that target specific genetic mutations, such as ras. The activation of ras via mutation is common in human tumors, with an overall frequency of approximately 30%.⁷ ras genes encode a 21-kDa protein that serves as an intermediate signal transduction pathways critical for cellular processes such as growth, differentiation, and apoptosis.⁸ RAS protein function is dependent upon localization to the inner aspect of the plasma membrane. A critical step in RAS protein localization to the membrane is posttranslational modification via the addition of a 15-carbon farnesyl moiety by the enzyme farnesyltransferase,⁹ which is required for ras-mediated transformation of cells in culture.¹⁰ By inhibiting RAS posttranslational processing, farnesyltransferase inhibitors (FTIs) block oncogenic RAS activity. FTIs were initially shown to block the growth of ras-transformed mouse cells in soft agar and reverse the transformed morphology of v-H-ras-transformed fibro-blasts, but not to inhibit the growth of src-transformed, raf-transformed, or nontransformed fibroblasts.¹¹ More recently, FTIs have been shown to inhibit the growth of rodent tumors and human tumor xenografts irrespective of ras status.¹² Preclinical models of malignant glioma support a role of RAS signaling in malignant gliomas. Although RAS oncogenic activation alone does not cause malignant transformation of astrocytes, RAS activation coupled with activation of other signaling molecules does appear to contribute to astrocytic immortalization and transformation.

R115777 (tipifarnib; Zarnestra, Johnson & Johnson, New Brunswick, NJ, USA) is an orally bioavailable methyl-quinolone, which has been shown in vitro and in vivo to be a potent and selective inhibitor of farnesyltransferase. In tumor cells in culture, R115777 produced antiproliferative effects both in H-ras and K-ras models, and also inhibited proliferation of a cell line that overexpresses N-ras.¹³

In mouse xenograft studies, oral administration of R115777 twice a day for 15–32 days at doses of 6.25–100 mg/kg demonstrated significant antitumor effects in H-ras–transformed tumors and tumors bearing mutated K-ras. R115777 also produced antiangiogenic activity in the rat aortic ring model, an in vitro model of spontaneous microvessel outgrowth, and in the in vivo Matrigel mouse model of vascular endothelial growth factor–induced angiogenesis.¹⁴ RAS inhibition through the use of FTIs has been studied in U-87 human malignant glioma cells in both subcutaneous and intracranial nude mouse models.¹⁵ Objective regression of tumor growth was observed in both models, with significant prolongation of survival in the intracranial model without apparent toxicity. R115777 has also demonstrated radiosensitization in resistant human glioma cell lines (SF763 and U87).¹³

In humans, after oral intake, R115777 was rapidly absorbed, with peak plasma concentrations generally reached within 2–4 h postdosing (capsule). Plasma concentrations declined biphasically with time. The half-life associated with the first phase of drug elimination was about 2–3 h; the half-life associated with the second and terminal phases was highly variable across patients, with a median value of 16 h. R115777 did not accumulate upon multiple dosing. Steady-state levels were reached within 2–3 days after drug intake and were maintained throughout twice-daily dosing (up to 8 weeks) in vitro. R115777 inhibits the metabolism of specific CYP3A4, CYP2D6, and CYP2C8-10 substrates, indicating the possibility of clinically relevant interactions with coadministered drugs that are metabolized mainly by these human liver cytochrome P-450 forms.¹⁶

Phase I and II studies have been completed in a number of tumor types, including breast, pancreas, and small-cell lung cancer.¹⁷ An intermittent dosing schedule was found to be the most tolerable. A North American Brain Tumor Consortium study of R115777 in patients with recurrent gliomas (33 patients with GBM, 5 with anaplastic astrocytoma, and 3 with anaplastic oligodendroglioma, who were not receiving enzyme-inducing antiseizure drugs (EIASDs) and had failed initial therapy) placed on 300 mg R115777 twice a day for 21 days on and 7 days off found modest activity against recurrent GBM.¹⁸ In a parallel study in patients on EIASDs, 23 patients received doses ranging from 300 mg to 700 mg twice a day, for 21 days every four weeks, and four patients had lack of progression of their disease at six months.¹⁹ It is therefore reasonable to investigate the activity of this agent in patients with newly diagnosed glioblastomas. In view of the possible drug interactions through P-450–inducing anticonvulsants, two cohorts of patients were investigated.

Based on the above information, a phase II trial of R115777 was given following surgery but prior to any radiation in patients with newly diagnosed GBM with residual contrast-enhancing disease was instituted. It should be noted that at the time this study was initiated, combined therapy with temozolomide and radiation was not the standard of care.

Materials And Methods

Eligibility

Eligible patients were 18 years or older and had supratentorial histologically confirmed grade IV astrocytoma (GBM) with no prior radiation, chemotherapy, or immunologic or biologic therapy. All patients who entered the study were required to have measurable contrast-enhancing tumor demonstrated by postoperative MRI; a KPS score ≥60; normal hematologic, hepatic, and renal function; and a Mini-Mental Status Exam score ≥15. Written informed consent was obtained from all patients before they entered the study.

Treatment Plan

Patients started on R115777 as a single agent after recovery from surgery. Patients were divided into two strata based on their use of EIASDs. Group A consisted of those taking phenytoin, carbamazepine, phenobarbital, primidone, or oxcarbazine. This group was given R115777 orally at 600 mg twice a day. Patients not on EIASDs were give R115777 300 mg twice a day. Treatment was administered for 3 weeks followed by a 1-week rest. After the first cycle, MRI was performed to evaluate tumor status. If the contrast-enhancing tumor decreased in size or remained stable (less than a 25% increase in volume of tumor and no new lesions), a second cycle of R115777 was administered. If the same criteria were met after the second cycle, then a third cycle was given.

A 2-week rest was given after the third cycle. External radiation was then administered—a total dose of 60 Gy. The first 46 Gy was delivered in 2-Gy fractions to the tumor plus the edema plus a 2-cm margin. The final 14 Gy was given to the contrast-enhancing tumor with a 2-cm margin. All treatment was delivered with 6 MV or greater beams. Following a 2-week rest postradiation, R115777 was reinstituted at the preradiation dose for all patients not exhibiting progression, unacceptable toxicity, or refusal to continue. For patients not responding to R115777 at any time prior to the completion of three cycles, external radiation was immediately administered. No postradiation drug was given to this group.

Efficacy Criteria

The primary efficacy end point was overall survival. The survival time was defined from the time of initial histologic diagnosis to death, or censored if the patient was alive at the time of last contact. We expected an estimated median survival time of 13.2 months, if the treatment was effective. Standardized response criteria were used with MRI scans employing volumetric analysis and neurologic examinations.⁹ A complete response (CR) was complete disappearance of all tumor on MRI, off all glucocorticoids, and a stable or improving neurologic status for at least 4 weeks. A partial response (PR) was 50% or more reduction in tumor size on volumetric MRI, on a stable or decreasing dose of glucocorticoids, and a stable or improving neurologic status for at least 4 weeks. Progressive disease (PD) was neurologic abnormalities not explained by causes unrelated to tumor progression (e.g., anticonvulsant or corticosteroid toxicity, electrolyte abnormalities, or hyperglycemia) or a greater than 25% increase in the volume of the tumor by MRI scan. If a patient had less than a 25% increase in tumor volume but worsening of neurologic status or was on a stable or increasing dose of steroids, or if new lesions appeared on serial MRI, further study treatment was discontinued and radiotherapy was instituted. Stable disease (SD) was defined as a patient whose clinical status and MRI/CT volumetrics did not meet the criteria for PR or PD.

Study Design and Statistical Methods

The study was a nonrandomized open-label, multicenter phase II clinical trial to evaluate the efficacy and safety of R115777 given prior to radiation therapy (RT) for treating newly diagnosed GBM patients. The study was designed to estimate a hazard rate with a total of 42 death events expected. It was powered for detecting a 30% reduction in hazard rate compared to the New Approaches to Brain Tumor Therapy (NABTT) historical database among the same patient population. The study was also planned to estimate the tumor response rate, progression-free survival, and the toxicities associated with the treatment.

All the data were collected and processed through the NABTT central office. Continuous variables were depicted by using descriptive statistics. Categorical variables were summarized by frequency tables, and estimated probabilities with 95% confidence intervals (CIs) were calculated using exact binomial distribution. Survival probability, median time of survival, and progression-free survival with 95% CIs were estimated using the Kaplan-Meier method. All the estimations and data summarization were done with SAS software, version 9.0 (SAS Institute, Inc., Cary, NC, USA).

Results

From August 28, 2003, until April 2004, 28 patients entered the study (17 male, 11 female; Table 1). Four patients had biopsy only, and 24 had a surgical resection. All histologies were classified as GBM. Fifteen patients were taking EIASDs, and 13 were not. Of those on EIASDs, 14 were on phenytoin and 1 on carbamazepine. The median age was 58 years (range, 34–82 years), and the median KPS score was 90 (range, 60–100). One patient completed three cycles of R115777, external radiation, and two maintenance cycles of drug. One patient completed three cycles and radiation. Three patients completed 2 cycles, 2 patients 1.5 cycles, 15 completed 1 cycle (54%), and 6 patients completed <1 cycle (21%). All patients were off the study due to disease progression except for one patient who was off study due to toxicity and another for other reasons. Twenty of the 28 patients (71%) had only one cycle or less of the R11577 treatment before disease progression.

Table 1.

Patient characteristics (n = 28)

Characteristic	No. of Patients	Percent
Sex
Male	17	60.7
Female	11	39.2
Age
Median	58
Range	34–82
KPS score
Median	90
Range	60–100
Anticonvulsants	21
P450	13	46.42
Non-P450	15	53.57

Toxicity

An overview of the toxicities is shown in Table 2. One grade 5 toxicity in the lower dose cohort was reported: a pulmonary embolism. There were a total of 17 grade 4 toxicities, of which 10 were hematologic, and all were reversible. There were 48 grade 3 toxicities, with multiple reported toxicities in the same patient. Three patients had rash, three decreased platelets, and two a low absolute neutrophil count (Table 3).

Table 2.

Overview of adverse events

Measure	Dose (mg twice daily)
	300	600	Total
Number of patients	13	15	28
Adverse events
Grade 5	1	0	1
Grade 4	1	16	17
Grade 3	21	27	48
Grade 2	40	74	114
Grade 1	128	174	302
Missing	0	1	1
Total	191	292	483
Relationship of reported side effect to R115777
Definite	0	50	50
Probable	31	33	64
Possible	46	72	118
Unlikely	38	40	78
None	75	88	163
Missing	1	9	10
Total number of dose-limiting toxicities	0	0	0

Table 3.

Number of patients with grade 3 and 4 events with possible, probable, or definite relationship to R115777, among the 28 patients in this study

Toxicity	Dose (mg twice daily)
	300	600	Total (%)
Absolute neutrophil count	1	1	7
Depressed level of consciousness		1	4
Diarrhea	1		4
Dyspnea	1		4
Fatigue		1	4
Febrile neutropenia		1	4
Hallucinations	1		4
Hemoglobin		1	4
Hyperglycemia	1		4
Hypocalcemia		1	4
Incontinence		1	4
Neuropathy—motor	1		4
Neuropathy—sensory		1	4
Neurology—other	1		4
Platelets	1	2	11
Rash/desquamation	1	2	11
Seizure	1		4
Speech impairment	1		4
Transfusion: packed RBCs		1	4
WBCs	1	1	7

Response

No patient was reported as having responded to therapy (50% reduction in tumor volume, on a stable dose of steroids, and stable neurologic status for 4 weeks). Eight patients had SD as the best response. Eighteen patients had PD, and two were not evaluable (Table 4). To date, 25 of 28 patients have died. The estimated median time of overall survival was 234.5 days (95% CI, 124, 448), and the estimated median time of progression-free survival was 42 days (95% CI, 38, 61) (Fig. 1). The study was stopped early due to the relatively rapid progression of disease in 24 patients (85%), who were off the study before the planned treatment schedule for RT.

Table 4.

Patient response to therapy

Best Response	Frequency	Percent
Progressive disease	18	64.29
Stable disease	8	28.57
Not evaluable	2	7.14

Fig. 1.

Overall survival and progression-free survival among 28 glioblastoma multiforme patients with contrast-enhancing tumor demonstrated by postoperative MRI who were treated with R115777 (tipifarnib).

Discussion

A number of agents have been tested in patients with residual disease following biopsy or partial resection of GBM and prior to radiation. Paclitaxel was studied in 15 assessable patients treated at the maximum tolerated dose of 140 mg/m² in the EIASD-negative group and 200 mg/m² in the EIASD-positive group; 1 of 15 patients showed a radiographic response, and the median survival was 355 days.²⁰ Gemcitabine was studied in 21 newly diagnosed GBM patients with prior to radiation; the median progression-free survival was 11 weeks, and the progression-free survival at 4 months was 24%.²¹ In the 18 of 21 patients who received radiation, the median progression-free survival was 8 months, and progression-free survival at 12 months was 17%. The median overall survival was 11 months; there was no survival advantage to preradiation gemcitabine.²¹

Nimustine hydrochloride plus cisplatin followed by radiation was tested in 22 patients with residual measurable postoperative disease, resulting in one patient with CR from chemotherapy, 36% with PR, 14% SD, and 45% PD; the median time to progression was 5.9 months, and median overall survival of 14.9 months.²² Postoperative temozolomide was studied in 36 patients for a maximum of four cycles prior to radiation, resulting in 11% with CR, 31% PR, and 28% SD; the median progression-free survival was 3.9 months, and median overall survival 13.2 months.²³ A combination of carboplatin and cyclophosphamide was tested prior to radiation in 17 patients with measurable postoperative GBM, resulting in one PR and two SD; the median survival was 7.6 months, and survival at 1 year was 33%.²⁴ Twenty-four patients were given irinotecan following surgery and prior to radiation, resulting in three minor responses and seven SD among the 22 evaluable patients; the time to progression was 9 weeks, and 6-month progression-free survival was 26%.²⁵ The NABTT Consortium studied 9-aminocamptothecin in 22 newly diagnosed patients with GBM and residual disease in a phase I study; there were no CRs or PRs.²⁶ A combination of irinotecan and carmustine was given to 37 newly diagnosed glioblastoma patients, resulting in 3 patients (11%) with CR and 11 patients (39%) with SD.²⁷

In the present study, an interim analysis was not formally planned. During the trial, all the serious adverse events were reported to the NABTT central office within 24 h, and the central office was notified by the principal investigator if any patient went off study. A fast off-study rate raised concerns, and a decision for an unplanned interim look was made by the principal investigator (RL), with a consensus from the NABTT central office. The interim analysis was performed on the 28 enrolled patients on September 9, 2004. It focused mainly on the secondary end points: progression-free survival, tumor response rate, and patient on-trial status. The results revealed that all 28 patients were off the study at the time of the interim analysis. The median time of progression-free survival was 42 days; there were no CRs or PRs. Eight patients had SD as the best response. The principal investigator, NABTT central office (Data Safety Monitoring Committee), and Cancer Therapy Evaluation Program had a conference call on May 4, 2004, to review these results, and the study was closed to further accrual. On May 6, 2004, the study was formally terminated early based on the fast and high off-study rate, which caused poor adherence of the treatment plan. The survival data were last updated on December 28, 2005. Twenty-five of the 28 patients had died. The estimated median time of overall survival was 7.7 months (95% CI, 4.1, 14.7), which is in the range reported by other studies of patients with measurable postoperative disease who were treated with systemic agents prior to radiation.

To our knowledge, this is the first study reporting on a molecular-targeting compound delivered as a single agent in newly diagnosed GBMs with documented residual disease treated prior to radiation. Although there is preclinical evidence that FTIs are active in GBM, these results suggest that R115777 has little to no activity in this clinical setting and patient population. Other factors contributing to the poor patient outcome are the measurable postoperative residual disease required to measure response, and the delay in administering RT. The NABTT CNS Consortium has not used this clinical trial design since temozolomide administered with radiation was shown to result in prolonged survival. Although this study conclusively demonstrates that R115777 has no significant activity as a single agent in previously untreated GBM, it remains possible that FTIs may be more effective in combination with other systemic agents or when given concomitantly with radiation.