Biomarker-based phase I dose-escalation, pharmacokinetic, and pharmacodynamic study of oral apricoxib in combination with erlotinib in advanced nonsmall cell lung cancer


  • Presented in part as an oral presentation at the 100th Annual Meeting of the American Association for Cancer Research, April 18-22, 2009, Denver, Colorado.



Apricoxib, a novel once-daily selective cyclooxygenase-2 inhibitor, was investigated in combination with erlotinib for recurrent stage IIIB/IV nonsmall cell lung cancer to determine the maximum tolerated dose, dose-limiting toxicity, and recommended phase II dose (RP2D) based on changes in urinary prostaglandin E2 metabolite (PGE-M).


Patients received escalating doses of apricoxib (100, 200, and 400 mg/day) in combination with erlotinib 150 mg/day until disease progression or unacceptable toxicity. Urinary PGE-M was used to assess biologic activity and inform the optimal biologic dose.


Twenty patients were treated (3 at 100 mg; 3 at 200 mg; 14 at 400 mg apricoxib) with a median of 4 cycles (range, 2-14 cycles); 8 patients (40%) received prior EGFR-directed therapies. No dose-limiting toxicity was observed. Study drug-related adverse events (AEs) included diarrhea, rash, dry skin, anemia, fatigue, and increased serum creatinine; 4 patients had grade ≥3 drug-related AEs (diarrhea, perforated duodenal ulcer, hypophosphatemia, and deep vein thrombosis). The RP2D was 400 mg/day based on safety, biologic activity based on decreases in urinary PGE-M, and pharmacokinetics. One patient had a partial response, and 11 had stable disease. Stable disease was observed in patients who had received prior EGFR inhibitor therapy but was greater in patients not previously treated with an EGFR inhibitor. Seventeen patients had elevated urinary PGE-M at baseline, and 14 (70%) had a decrease from baseline, which was associated with disease control.


Apricoxib plus erlotinib was well tolerated and yielded a 60% disease control rate. A phase II trial is currently investigating 400 mg/day apricoxib plus 150 mg/day erlotinib in patients selected based on change in urinary PGE-M. Cancer 2011. © 2010 American Cancer Society.

Studies have shown that the epidermal growth factor receptor (EGFR) and cyclooxygenase-2 (COX-2) play an important role in the biology of nonsmall cell lung cancer (NSCLC),1, 2 and EGFR inhibitors have demonstrated clinical activity, particularly in patients with EGFR mutations.3-6 Overexpression of COX-2 is associated with a poor prognosis in NSCLC,1, 7 and COX-2 inhibitors have demonstrated synergy when combined with EGFR inhibitors in preclinical models.8-10 Inflammatory mediators, including prostaglandin-E2 (PGE2), can promote epithelial-to-mesenchymal transition and increase resistance to EGFR tyrosine kinase inhibitors (TKIs) in lung cancer.8, 11 These studies provide a strong rationale for combining a COX-2 inhibitor with an EGFR TKI for the treatment of NSCLC. The current phase I study investigated the combination of apricoxib plus erlotinib in patients with recurrent or stage IIIB/IV NSCLC.

Apricoxib is a novel once-daily selective COX-2 inhibitor not yet approved for patient use. Preclinical studies have demonstrated that apricoxib has potent analgesic, anti-inflammatory, and antitumor effects.12, 13 In xenograft models of human NSCLC, apricoxib inhibited tumor growth more effectively than celecoxib.12 The preclinical safety profile of apricoxib is consistent with a selective COX-2 inhibitor with minimal gastrointestinal toxicity at exposures demonstrating anti-inflammatory and antitumor activity.13

Selection of patients based on tumor biology can identify a subset of patients who are more likely to benefit from EGFR and COX-2 targeted therapy. This concept was recently validated in 2 large studies that selected NSCLC patients to enrich the population for EGFR mutations. These studies demonstrated high response rates and significant clinical benefit from EGFR TKIs.3, 5 Similarly, the combination of celecoxib plus carboplatin/gemcitabine significantly improved survival in a subset of patients with moderate to high COX-2 expression.7 Another promising biomarker for COX-2 activity is the stable metabolite of PGE2 (PGE-M).14 Urinary PGE-M correlates with intratumoral PGE2 levels,15 and patients with NSCLC have elevated levels of urinary PGE-M.14, 16 More importantly, 2 studies have demonstrated that a decrease in urinary PGE-M in response to treatment with celecoxib plus chemotherapy correlated with either an objective tumor response or improved survival.15, 17 These studies suggest that PGE-M may predict response to COX-2 inhibition and may be a useful biomarker for patient selection.

The primary objective of the current study was to determine the recommended phase II dose of apricoxib in combination with erlotinib (150 mg) as daily oral therapy for advanced NSCLC. A secondary objective was to explore the value of selecting patients for a subsequent phase II study based on changes in urinary PGE-M during treatment with apricoxib.


Patient Selection

Adult (≥18 years of age) patients with histologically confirmed stage IIIB (pleural effusion) or stage IV NSCLC that was measurable or evaluable by Response Evaluation Criteria in Solid Tumors (RECIST) were eligible. Patients with central nervous system (CNS) metastases were eligible if their CNS disease was asymptomatic and stable after radiotherapy for ≥2 weeks and they did not require corticosteroids for ≥1 week. Patients must have failed at least 1 prior chemotherapy regimen or refused chemotherapy. Prior treatment with an EGFR TKI or ongoing treatment with erlotinib (150 mg/day) at study entry was allowed, but patients with disease progression at study entry while actively receiving an EGFR inhibitor were excluded. No patient entered the study on erlotinib. Of the 8 patients with prior erlotinib or gefitinib treatment, 2 had a partial response, 3 had stable disease (SD), and 2 had progressive disease as their best response to prior EGFR TKI therapy. In 1 patient, the previous response to erlotinib was unknown. Patients must have had an Eastern Cooperative Oncology Group performance status of 0, 1, or 2 and adequate organ function.

Patients were ineligible if they had received prior treatment with a COX-2 inhibitor for metastatic NSCLC, radiation therapy <2 weeks prior to study entry, or chemotherapy, noncytotoxic investigational agents, or high-dose corticosteroids within 3 weeks of initiating study treatment. Patients were excluded if they had a history of myocardial infarction, stroke, or cardiac intervention within 12 months; a history of gastrointestinal bleeding, ulceration, or perforation; or hypersensitivity to apricoxib, erlotinib, aspirin, or other nonsteroidal anti-inflammatory drugs (NSAIDs). Concurrent use of COX-2 inhibitors or other NSAIDs (within 2 days) or aspirin (within 7 days) of the first dose of study drug, antiplatelet drugs, or anticoagulants was not permitted. Low-dose aspirin was not allowed. Use of potent CYP3A4 inhibitors or inducers was prohibited during administration of erlotinib. Standard treatments for erlotinib-related rash and diarrhea were recommended. Use of proton pump inhibitors (PPIs) was not allowed initially; however, after the occurrence of a perforated duodenal ulcer in 1 patient in cohort 2, it was suggested that patients be placed on PPIs after the pharmacokinetics (PK) blood draws for apricoxib and erlotinib were completed. Repeat PK analysis of erlotinib following initiation of a PPI was not performed. Solubility of apricoxib is not affected by pH, and therefore absorption would not be expected to change in the presence of PPIs. The study was approved by the institutional ethics review boards, and all patients provided written informed consent.

Study Design

This was a phase I, multicenter dose-escalation study designed to determine the recommended phase II dose of apricoxib (Tragara Pharmaceuticals, San Diego, CA) in combination with a fixed dose of erlotinib (Tarceva; OSI Pharmaceuticals, Melville, NY). This was defined either by the optimal biologic dose (OBD) or the maximum tolerated dose (MTD). The OBD was defined as the dose at which the average urinary PGE-M level was reduced from baseline by >75% in the first 3 patients in a dosing cohort (cohorts 1 and 2) or the first 3 patients to exhibit a decrease in PGE-M (amendment 3) and complete 1 cycle of study treatment. The MTD was defined as the dose level below the dose that caused dose-limiting toxicity (DLT) in 1 of 3 or >1 of 6 patients in a dosing cohort. Nine additional patients were enrolled at the recommended phase II dose to gather further PK data.

DLT was defined as a treatment-emergent adverse event (TEAE) during cycle 1 that the investigator considered to be possibly, probably, or definitely related to study drug, including any of the following: grade 4 neutropenia lasting at least 5 days; grade 3 febrile neutropenia or grade 4 thrombocytopenia of any duration; or grade 3 or greater nonhematologic toxicities, including grade 3 or greater diarrhea, nausea, or vomiting despite adequate treatment. Overall survival was not assessed.


Cohorts of 3 patients received escalating doses of apricoxib daily in a fasting condition during cycle 1 and thereafter with food in combination with daily oral erlotinib (according to the manufacturers' recommendations, either 2 hours after or 1 hour before meals) at a fixed dose of 150 mg, beginning on day 8 of cycle 1. Each cycle was 28 days. The starting dose of apricoxib was 100 mg/day (administered as a 100-mg tablet), with escalation to 200 mg/day in cohort 2. Thereafter, the dose-escalation schema was determined based on an interim PK assessment (day 8, cycle 1) of dosing cohorts 1 and 2 to compare the human PK profile of the milled apricoxib agent used in this study with the unmilled agent used in earlier PK studies. Dose escalation to 400 mg/day in cohort 3 was based on the observation that the peak plasma concentration (Cmax) and area under the concentration time curve (AUC) in cohorts 1 and 2 were <2 times the Cmax and AUC using the unmilled drug substance. Three to six patients were treated at each dose level according to a 3 + 3 design. Patients were removed from the study if they had disease progression or intolerable toxicity, if they refused treatment, or at the investigator's discretion. Patients could continue to receive study treatment until disease progression or intolerable toxicity.


Patients were evaluated at baseline and every 2 cycles thereafter by computed tomography for tumor response and disease progression according to RECIST version 1.0.18 Objective responses were to be confirmed at least 4 weeks later, and in the case of SD, follow-up measurements must have met the SD criteria at least once after study entry at a minimum interval of 6 weeks (42 days). Safety was assessed using National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), Version 3.0. Urine samples were collected for assessment of PGE-M levels at baseline, end of cycle 1 (cohorts 1 and 2) or day 15 of cycle 1 (cohort 3+), day 1 of cycle 2, and end of study. Where possible, tumor and/or plasma was collected for determination of EGFR mutation status.


Pharmacokinetic analysis of apricoxib was performed using blood drawn on day 1 predose, days 1-3 postdose, days 8, 15, and 22 of cycle 1, and day 1 (predose) of each subsequent cycle. PK analysis of erlotinib was performed at steady state on day 1 of cycle 2. PK calculations were performed using noncompartmental methods (WinNonlin, version 5.0.1, Pharsight Corporation, Mountain View, CA). Full PK data were collected on all patients.

PGE-M Assay

Urinary PGE-M was quantified by liquid chromatography/tandem mass spectrometry as previously described.14 The upper limit of normal (ULN) for urinary PGE-M in men (10.4 ng/mg creatinine [Cr]) and women (6.0 ng/mg Cr) had been previously established using this assay,14 and these values were used to define elevated PGE-M among patients enrolled in the current study.

EGFR Mutation Determination

Genomic DNA was extracted from 0.8-1.5 mL of plasma using a QIAamp DNA Blood mini kit (Qiagen, Valencia, CA). DNA was eluted in 50 μL of sterile ddH2O and the concentration and purity measured by a NanoDrop ND-1000 spectrophotometer.

Scorpion-ARMS for EGFR mutations

Specimens were analyzed for mutations in EGFR using a DxS EGFR29 mutation test kit (DxS Ltd., Manchester, UK) following the manufacturer's protocol. Each EGFR test kit is supplied with probes specific for exon 19 deletions, T790M and L858R mutations. Each kit also contains a control probe specific to a non-polymorphic site used to assess total DNA content. Five microliters of template DNA was used in each reaction. All reactions were performed on an iQ5 Real-Time Polymerase Chain Reaction detection system (BioRad, Hercules, CA) using these conditions: 95°C for 10 minutes, 45 cycles of 95°C for 30 seconds, and 61°C for 60 seconds, with fluorescent readings (set for FAM and HEX) collected after each cycle. Data were analyzed using iQ5 Optical System software (version 2.0; (BioRad, Hercules, CA). The cycle threshold (CT) was defined as the fractional cycle number at which the fluorescence passes the fixed threshold, set above baseline, as determined by the software. For each probe set, a cutoff deltaCT, defined as the CT of the mutant minus the CT of the control, has been established by the manufacturer as the point at which a positive signal could be a result of background fluorescence of the primer on wild-type DNA.

Statistical Design

Patients who completed at least 1 cycle (both drugs for at least 21 of 28 days) and had at least 1 subsequent safety follow-up visit were included in the analysis for determining the OBD and MTD. Patients were deemed not evaluable and replaced if they failed to complete cycle 1 for reasons other than toxicity. All patients who completed 2 cycles of study treatment were considered evaluable for efficacy unless there was clear evidence of clinical progression prior to completing cycle 2. Descriptive statistics were used to summarize all safety, efficacy, and apricoxib and erlotinib blood concentration data using SAS statistical software (version 8.01).



A total of 27 patients signed informed consent and were screened. Of these, 7 patients never received any study drug (4 were ineligible, 1 had disease progression, 1 withdrew consent, 1 was lost to follow-up), and 20 patients were enrolled and treated between October 2007 and January 2009. Baseline patient and disease characteristics for the 20 treated patients are shown in Table 1. Median age was 67.5 years, 65% of patients were female, and 40% of patients were never smokers. The majority of patients had stage IV NSCLC, and histopathology was varied. The majority of patients had received at least 1 systemic chemotherapy regimen, and 8 (40%) patients had received prior EGFR TKI therapy. Three patients were enrolled into cohort 1 (100 mg apricoxib), 3 into cohort 2 (200 mg), and 14 into cohort 3 (400 mg). One patient in cohort 2 did not meet all eligibility criteria because of the time frame of prior treatment with an NSAID but was enrolled by waiver. A dose of 400 mg apricoxib was selected for cohort 3 based on the interim PK and urinary PGE-M analysis of cohorts 1 and 2.

Table 1. Demographics and Baseline Disease Characteristics
 Apricoxib DoseOverall (N = 20)
100 mg (n = 3)200 mg (n = 3)400 mg (n = 14)
  • ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor.

  • a

    Prior erlotinib in 7 patients; prior gefitinib in 1 patient.

Median age, years67.060.068.567.5
Sex, n (%)
 Female3 (100%)1 (33%)9 (64)13 (65)
 Male02 (67%)5 (36)7 (35)
Ethnic origin, n (%)
 White2 (67%)3 (100%)10 (71%)15 (75%)
 Asian1 (33%)04 (29%)5 (25%)
ECOG performance status, n (%)
 001 (33%)4 (29%)5 (25%)
 13 (100%)2 (67%)9 (64%)14 (70%)
 2001 (7%)1 (5%)
Disease stage, n (%)
 IV3 (100%)3 (100%)12 (86%)18 (90%)
 IIIB002 (14%)2 (10%)
Histopathology, n (%)
 Squamous cell carcinoma02 (67%)4 (29%)6 (30%)
 Adenocarcinoma1 (33%)04 (29%)5 (25%)
 Bronchiolo-alveolar adenocarcinoma1 (33%)02 (14%)3 (15%)
 Nonsmall cell carcinoma1 (33%)02 (14%)3 (15%)
 Adenosquamous carcinoma01 (33%)1 (7%)2 (10%)
 Papillary adenocarcinoma001 (7%)1 (5%)
Smoking history, n (%)
 Never smoked2 (67%)06 (43)8 (40)
 Smoker1 (33%)2 (67)1 (7)4 (20)
 Former smoker01 (33)7 (50)8 (40)
Prior chemotherapy, n (%)
 None001 (7%)1 (5%)
 1 prior regimen02 (67%)5 (36%)7 (35%)
 2 prior regimens1 (33%)1 (33%)3 (21%)5 (36%)
 ≥3 prior regimens2 (67%)05 (36%)7 (35%)
Prior EGFR-directed therapy
 Yesa1 (33%)07 (50%)8 (40%)
 No2 (67%)3 (100%)7 (50%)12 (60%)

A median of 4 cycles (range 2-14 cycles) of apricoxib and erlotinib were administered. Nineteen patients discontinued because of disease progression, and 1 patient discontinued of a study drug-related adverse event (grade 3, perforated duodenal ulcer).


The most frequently reported TEAEs were rash, diarrhea, dry skin, fatigue, anemia, anorexia, dyspnea, nausea, abdominal pain, anxiety, increased serum creatinine, decreased appetite, and dizziness (Table 2). The majority of TEAEs were mild to moderate in severity, and no grade 3 or 4 rash was reported. Grade 3 or greater TEAEs were reported in 8 patients (Table 3), and 4 patients developed grade 3 or greater study drug-related TEAEs (diarrhea, perforated duodenal ulcer, hypophosphatemia, and deep vein thrombosis). Following occurrence of an ulcer in 1 patient in cohort 2, proton pump inhibitors (PPIs) were suggested for all patients in cohort 3. Eleven patients in cohort 3 received PPIs. All grade 3 or greater TEAEs resolved with dose delays, no apricoxib dose modifications were required, and there was no DLT at any dose level.

Table 2. Summary of Treatment-Emergent Adverse Events Occurring in >10% of Patients
 Number of Patients (%)
Apricoxib DoseOverall (N = 20)
100 mg (n = 3)200 mg (n = 3)400 mg (n = 14)
Rash1 (33%)1 (33%)13 (93%)15 (75%)
Diarrhea1 (33%)2 (67%)10 (71%)13 (65%)
Dry skin3 (100%)07 (50%)10 (50%)
Fatigue2 (67%)06 (43%)8 (40%)
Anemia1 (33%)06 (43%)7 (35%)
Anorexia2 (67%)1 (33%)3 (21%)6 (30%)
Dyspnea1 (33%)05 (36%)6 (30%)
Nausea1 (33%)2 (67%)3 (21%)6 (30%)
Abdominal pain004 (29%)4 (20%)
Anxiety1 (33%)1 (33%)2 (14%)4 (20%)
Blood creatinine increased004 (29%)4 (20%)
Decreased appetite2 (67%)02 (14%)4 (20%)
Dizziness1 (33%)1 (33%)2 (14%)4 (20%)
Abdominal pain upper1 (33%)02 (14%)3 (15%)
Back pain003 (21%)3 (15%)
Constipation1 (33%)1 (33%)1 (7%)3 (15%)
Cough1 (33%)02 (14%)3 (15%)
Dehydration1 (33%)02 (14%)3 (15%)
Gastritis003 (21%)3 (15%)
Pruritus003 (21%)3 (15%)
Stomatitis1 (33%)02 (14%)3 (15%)
Weight decreased1 (33%)02 (14%)3 (15%)
Table 3. Summary of CTC Grade 3 or Greater Treatment-Emergent Adverse Events
 Number of Patients (%)
Apricoxib DoseOverall (N = 20)
100 mg (n = 3)200 mg (n = 3)400 mg (n = 14)
Fatigue002 (14%)2 (10%)
Diarrhea001 (7%)1 (5%)
Duodenal ulcer perforation01 (33%)01 (5%)
Hypoalbuminemia001 (7%)1 (5%)
Hypophosphatemia001 (7%)1 (5%)
Dyspnea001 (7%)1 (5%)
Pleural effusion001 (7%)1 (5%)
Pulmonary embolism001 (7%)1 (5%)
Anemia001 (7%)1 (5%)
Pneumonia001 (7%)1 (5%)
Syncope001 (7%)1 (5%)
Deep vein thrombosis001 (7%)1 (5%)

Four patients in cohort 3 had an increase in serum creatinine. Three of the 4 patients had a grade 1 increase, which returned to baseline without any change in drug dose. One patient had a grade 2 increase (from baseline of 1.36 to 2.0 mg/dL) that decreased to baseline when apricoxib was held. Apricoxib was restarted at the same dose level with no further increase in the patient's serum creatinine.

Three patients had a serious adverse event (SAE). One patient in cohort 2 had a study drug-related grade 3 perforated duodenal ulcer during cycle 3 and was hospitalized. In addition, 1 patient in cohort 3 had shortness of breath, pleural effusion, and pneumonia requiring hospitalization during cycle 4; these events were considered unrelated to the study drug and a result of disease progression. A second patient in cohort 3 had an SAE of disease progression during cycle 5 that was unrelated to the study drug. None of these SAEs occurred during cycle 1 and therefore were not DLTs. There were 3 deaths in the study or within 28 days of study discontinuation, all a result of disease progression. No serious cardiac AEs or prolongation of QTc were reported.

Antitumor Activity

One patient had a partial response (PR), and 11 patients had SD, for a disease control rate of 60% (Table 4). Median time to disease progression was 92 days (95% CI, 52-118 days). The PR occurred in a female patient in cohort 1; the duration of response was 302 days (from date of confirmed PR to date of progression), and the patient was progression free for 387 days (from date of first dose to date of progression). This patient, a 67-year-old white woman, never smoker, was EGFR wild type and had received prior chemotherapy but no prior EGFR TKI therapy. Her baseline PGE-M was high (113.5 ng/mg Cr) and decreased by 47% after 2 cycles. In total, she received 14 cycles of study drug.

Table 4. Best Tumor Response
 Apricoxib DoseOverall (N = 20)
100 mg (n = 3)200 mg (n = 3)400 mg (n = 14)
  • a

    Follow-up measurements must have met the SD criteria at least once after study entry at a minimum interval of 6 weeks.

Partial response1 (33%)001 (5%)
Stable diseasea2 (67%)3 (100%)6 (43%)11 (55%)
Progressive disease008 (57%)8 (40%)

Apricoxib plus erlotinib demonstrated activity regardless of previous treatment with an EGFR TKI (Table 5). Among 8 patients who had failed previous treatment with an EGFR TKI, 4 (50%) had SD. Among 12 patients who were not previously treated with an EGFR TKI, the disease control rate was 67% (1 PR and 7 SD for 3 cycles [n = 2] or 4+ cycles [n = 5]).

Table 5. Exploratory Analyses of Best Tumor Response and Time to Progression by Baseline and Change From Baseline Urinary Prostaglandin E2 Metabolite and Prior Anti-EGFR Therapy
 Patients, n (%)
Baseline PGE-MDecrease from Baseline PGE-MPrevious EGFR TKI Therapy
Elevated (n = 17)Normal (n = 3)Yes (n = 14)No (n = 6)Yes (n = 8)No (n = 12)
  • a

    1 Month = 28 days.

Partial response1 (6%)01 (7%)001 (8%)
Stable disease10 (59%)1 (33%)8 (64%)4 (67%)4 (50%)7 (58%)
PD6 (35%)2 (67%)5 (36%)2 (33%)4 (50%)4 (33%)
Median TTP, moa4.02.0NDND2.84.1
Range, cycles1.9-14.12.0-7.9NDND1.9-6.02.0-14.1

Urinary PGE-M Analysis

The majority of patients (17 of 20) had elevated PGE-M at baseline (Table 6). Overall, 14 patients (70%) had a decrease from baseline, and 11 patients (55%) had a >50% decrease from baseline urinary PGE-M, indicating that apricoxib has biologic activity against its target at the doses tested, although there was no clear dose-response relationship. Three patients had an increase from baseline PGE-M, and 3 patients had no change.

Table 6. Change From Baseline Urinary Prostaglandin E2 Metabolite During Cycle 1
Patient Number (Sex)Apricoxib Dose, mgBaseline PGE-M,ng/mg CraChange from Baseline, %Best Response
  • a

    Upper limit of normal for PGE-M is 10.4 ng&sold;mg Cr for men and 6.0 ng/mg Cr for women.

  • NC, no change; PR, partial response; SD, stable disease; PD, progressive disease.

002-001 (F)100113.5− 47PR
002-002 (F)10013.3− 59SD
002-003 (F)10016.0− 85SD
005-004 (M)20049.0− 84SD
005-006 (F)20011.2+ 45SD
005-007 (M)20024.0− 71SD
004-008 (F)4002.8− 92PD
002-011 (F)4006.1− 18SD
002-014 (F)40014.9− 74PD
002-015 (F)4004.1+ 255PD
004-017 (M)40018.6− 58PD
005-018 (M)40020.7− 80PD
002-019 (F)40010.8NCSD
004-020 (M)40015.2− 62SD
005-021 (F)40015.2− 26SD
002-022 (F)40024.2NCPD
001-023 (M)4004.7NCSD
002-024 (F)4009.0+ 8SD
001-026 (F)4008.7− 53PD
004-027 (M)40032.9− 86SD

Exploratory analyses were conducted to determine if disease control correlated with baseline PGE-M or change from baseline PGE-M (Table 5). Among 17 patients with elevated baseline PGE-M, 1 patient had a PR, and 10 patients had SD, for a disease control rate of 65%. In comparison, 1 of 3 patients (33%) with normal baseline PGE-M had SD. Median time to disease progression was also longer among patients with elevated baseline PGE-M (97 days [4 cycles] vs 55 days [2 cycles]). Similarly, a decrease from baseline PGE-M appears to be associated with a favorable response to apricoxib plus erlotinib. The patient with a PR had a high baseline level and a nearly 50% decrease from baseline, and most patients with SD had a decrease from baseline PGE-M. Among 3 patients with an increase in PGE-M during treatment, 1 had SD.

Apricoxib Pharmacokinetics

Apricoxib is a potent COX-2 inhibitor with an IC50 of 1.5 nM for inhibition of PGE2 production in cells. As shown in Figure 1, trough plasma levels of apricoxib remained above 53 ng/mL (the IC50) for 24 hours. The Cmax and AUC of apricoxib decreased slightly from 100 to 200 mg (likely due to high interpatient variability) but demonstrated a somewhat dose-proportional increase from 200 to 400 mg/day (Fig. 1). The 400-mg dose also provided more sustained exposure over 24 hours compared with lower doses, and some accumulation of drug was observed by day 8. The day 1 cycle 2 steady-state AUC of apricoxib in combination with erlotinib was similar to the single-agent AUC shown in Figure 1 (day 1 of cycle 1).

Figure 1.

Mean plasma concentrations of apricoxib by administered dose on day 1 of cycle 1 is shown. Trough plasma levels of apricoxib remained above the IC50 (dashed line) for 24 hours.

EGFR Mutation Analysis

Three patients (15%) had EGFR mutations detected, and none of these patients had received previous anti-EGFR TKI therapy. Patient 005-006, a 77-year-old woman, had an exon 21 mutation; patient 005-021, a 73-year-old woman, had an exon 19 mutation; and patient 001-026, a 42-year-old woman, had an exon 21 mutation. Patients 005-006 and 005-021 had SD for 8 and 3 cycles, respectively, and patient 001-026 progressed at the first tumor evaluation after cycle 2. No correlation between response and mutation status can be drawn because of the small number of mutations and the finding that the experimental kit used could not detect the T790M mutation because of issues with the probe sensitivity.

Determination of the Recommended Phase II Dose

No DLT occurred. Therefore, the recommended phase II dose was determined by the observed biologic activity of apricoxib based on changes from baseline urinary PGE-M levels and the observed PK profile. Although the protocol definition of the OBD (average decrease from baseline PGE-M >75%) was not achieved in any dosing cohort, there was a trend toward greater decreases in PGE-M in individual patients in cohort 3 compared with cohort 2. Further dose escalation of apricoxib beyond 400 mg was not undertaken because there was no clear increase in biologic activity compared with 200 mg and because of the drug-related SAE that occurred in cohort 2. Given that the 400-mg dose resulted in an increase in systemic exposure and maintained plasma levels above the threshold for COX-2 inhibition for a full 24 hours without resulting in any DLT, 400 mg was selected as the recommended phase II dose.


The combination of apricoxib and erlotinib was well tolerated. Overall, there was a slight increase in the incidence and severity of TEAEs with increasing doses of apricoxib, but no DLT was observed at the highest apricoxib dose tested. The most frequently observed study drug-related TEAEs (skin and gastrointestinal events) were consistent with the known safety profile of EGFR inhibitors, and the duodenal perforation and increased serum creatinine are consistent with the safety profile of COX-2 inhibitors. After 1 patient in cohort 2 developed a perforated ulcer, PPIs were suggested for all patients in cohort 3. This occurred before the erlotinib prescribing information was updated regarding coadministration of erlotinib and PPIs, and future studies will address this issue. An increase in gastrointestinal bleeding was reported previously with the combination of a COX-2 inhibitor and erlotinib;19 however, that study allowed low-dose aspirin and anticoagulant therapy. Based on the favorable safety profile, biologic activity, and PK profile of apricoxib, the recommended phase II dose of apricoxib was determined to be 400 mg/day in combination with 150 mg/day erlotinib.

A secondary objective of this study was to explore selection of patients for the subsequent phase II study based on the biomarker urinary PGE-M. In the absence of patient selection, COX-2 and EGFR inhibitors have demonstrated only modest antitumor activity in NSCLC. Several single-arm phase II studies investigating combinations of COX-2 inhibitors with standard chemotherapy in unselected NSCLC patients failed to demonstrate additive benefit.7, 15, 17, 20, 21 Similarly, several single-arm phase II studies that combined gefitinib with either celecoxib or rofecoxib demonstrated response rates similar to single-agent gefitinib.22-24 Therefore, patient selection is critical. In the case of EGFR inhibitors, selection of patients with EGFR mutations has recently been shown to yield high response rates and improved surival,3, 5 whereas EFGR TKIs yield response rates of only 10% to 15% in unselected populations. At initiation of the current study, there was insufficient data on PGE-M to support its use as a method for selection. Urinary PGE-M has been developed as a marker for evaluation of COX-2 inhibition and has only been tested in the setting of COX-2 inhibitors or other NSAIDS. Although a decrease in urinary PGE-M in response to treatment with celecoxib plus chemotherapy has been shown to correlate with objective tumor response and improved survival,15, 17 we sought further evidence to establish whether either baseline PGE-M or a decrease in response to treatment with apricoxib could be useful biomarkers for patient selection.

Exploratory analyses evaluating correlations between PGE-M and disease control suggested that baseline PGE-M may be a useful biomarker for patient selection, although no firm conclusions can be drawn from that analysis. Nevertheless, patients with elevated baseline PGE-M and those who subsequently had a decrease from baseline PGE-M during treatment appeared to have numerically better disease control. However, it must be noted that only 3 patients had normal baseline PGE-M at study entry. Patients with no prior EGFR therapy also appeared to benefit slightly more than those who had received prior EGFR therapy. Finally, no conclusions can be drawn regarding the subset of patients with an EFGR mutation. Although 3 patients (15%) were found to have an EGFR mutation using a plasma-based commercial assay, no correlation between response and EGFR mutational status could be drawn because of the small sample size.

In summary, this phase I study demonstrated that apricoxib could be safely administered in combination with erlotinib to patients with advanced NSCLC. This combination may overcome resistance and improve response to erlotinib.11, 25 A randomized, placebo-controlled, phase II study is currently ongoing to test the hypothesis that patients selected for treatment based on a ≥50% decrease from baseline PGE-M after a 5-day open-label run-in of apricoxib will benefit from treatment with daily apricoxib (400 mg) plus erlotinib (150 mg).


The authors thank the patients and their families for participating in this study; Jeff Riegel, PhD, for help in manuscript preparation; and Philip Mack, PhD, University of California, Davis, for analysis of EGFR mutations.


Supported by Tragara Pharmaceuticals, Inc., San Diego, California. Trial registry number: NCT00569114.