`
`JOURNAL OF CLINICAL ONCOLOGY
`
`O R I G I N A L R E P O R T
`
`From the Montefiore Medical Center/
`Albert Einstein College of Medicine,
`Bronx; New York University School of
`Medicine, New York, NY; Institute for
`Drug Development, San Antonio, TX;
`Beth Israel Deaconess Medical Center
`and Deaconess Medical Center,
`Boston, MA; Dartmouth-Hitchcock
`Medical Center, Lebanon, NH; OSI
`Pharmaceuticals Inc, Boulder, CO; and
`Rush Cancer Institute, Chicago, IL.
`
`Submitted November 10, 2003; accepted
`April 28, 2004.
`
`Supported by OSI Pharmaceuticals, Inc,
`Melville, NY.
`
`Data was previously presented at the
`37th Annual Meeting of the American
`Society of Clinical Oncology, San Fran-
`cisco, CA, May 12-15, 2001; and at the
`10th World Congress on Lung Cancer,
`Vancouver, Canada, 2003.
`
`Authors’ disclosures of potential con-
`flicts of interest are found at the end of
`this article.
`
`Address reprint requests to Roma´n
`Pérez-Soler, MD, Department of Oncol-
`ogy, Hofheimer 100, Montefiore Medi-
`cal Center/Albert Einstein College of
`Medicine, 111 E 210th St, Bronx, NY
`10467; e-mail: rperezso@
`montefiore.org.
`
`© 2004 by American Society of Clinical
`Oncology
`
`0732-183X/04/2216-3238/$20.00
`
`DOI: 10.1200/JCO.2004.11.057
`
`Determinants of Tumor Response and Survival With
`Erlotinib in Patients With Non–Small-Cell Lung Cancer
`Roma´n Pérez-Soler, Abraham Chachoua, Lisa A. Hammond, Eric K. Rowinsky, Mark Huberman,
`Daniel Karp, James Rigas, Gary M. Clark, Pedro Santaba´rbara, and Philip Bonomi
`
`A
`
`B
`
`S
`
`T
`
`R
`
`A
`
`C
`
`T
`
`Purpose
`Erlotinib is a highly specific epidermal growth factor receptor (HER1/EGFR) tyrosine kinase inhibitor.
`This phase II study of erlotinib in patients with HER1/EGFR-expressing non–small-cell lung cancer
`previously treated with platinum-based chemotherapy evaluated tumor response, survival, and
`symptom improvement.
`
`Patients and Methods
`Fifty-seven patients received an oral, continuous daily dose of 150 mg of erlotinib. Assessments of
`objective response used WHO and Response Evaluation Criteria in Solid Tumors criteria. The European
`Organization for Research and Treatment of Cancer Quality of Life Questionnaire C30, supplemented
`with a lung cancer module, Quality of Life Questionnaire LC13, was used to measure health-related
`quality of life. Additional analyses were performed to identify predictors of response and survival.
`
`Results
`The objective response rate was 12.3% (95% CI, 5.1% to 23.7%). Responses were observed regardless
`of type or number of prior chemotherapy regimens. Median survival time was 8.4 months (95% CI, 4.8
`to 13.9 months), and the 1-year survival rate was 40% (95% CI, 28% to 54%). Erlotinib therapy was
`associated with tumor-related symptom improvement. The drug was well tolerated; drug-related
`cutaneous rash and diarrhea were observed in 75% and 56% of patients, respectively. One patient
`experienced toxicity consisting of severe grade 3 rash and diarrhea. Time since diagnosis and good
`performance status were significant predictors of survival in a multivariate Cox proportional hazards
`model, whereas HER1/EGFR staining intensity was not. Additionally, survival correlated with the
`occurrence and severity of rash.
`
`Conclusion
`Erlotinib was active and well tolerated in this patient population, and further clinical development is
`clearly warranted. Cutaneous rash seems to be a surrogate marker of clinical benefit, but this finding
`should be confirmed in ongoing and future studies.
`
`J Clin Oncol 22:3238-3247. © 2004 by American Society of Clinical Oncology
`
`INTRODUCTION
`
`Several human malignancies are associated
`with aberrant or overexpressed epidermal
`growth factor receptor (HER1/EGFR).1-3
`HER1/EGFR tyrosine kinase serves as a po-
`tential target for therapeutic intervention in
`human tumors including ovarian, head and
`neck, breast, bladder, lung, and other squa-
`mous cell carcinomas.4-7 Overexpression of
`HER1/EGFR has been directly related to
`chemoresistance and poor prognosis.8-10
`Several studies have shown that HER1/
`
`EGFR expression or overexpression is
`common in non–small-cell
`lung cancer
`(NSCLC) tumor samples.9-17
`Erlotinib (Tarceva; OSI-774; OSI Phar-
`maceuticals, Melville, NY) is a potent and
`selective inhibitor of HER1/EGFR tyrosine
`kinase. It is a direct and reversible enzyme
`inhibitor in vitro, with a median inhibitory
`concentration of 2 nmol/L (0.79 ng/mL).
`Erlotinib reduces HER1/EGFR autophos-
`phorylation in intact tumor cells with a me-
`dian inhibitory concentration of 20 nmol/L
`(7.9 ng/mL),
`inhibits epidermal growth
`
`3238
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`AVENTIS EXHIBIT 2081
`Mylan v. Aventis, IPR2016-00712
`
`
`
`Erlotinib Monotherapy in NSCLC
`
`factor-dependent cell proliferation at nanomolar concen-
`trations, and blocks cell-cycle progression at the G1 phase.18
`Oral administration of erlotinib to mice reduced the
`level of HER1/EGFR autophosphorylation in human tumor
`xenografts by over 70% for more than 12 hours. Daily
`administration markedly inhibited the growth of HN5 hu-
`man head and neck tumor and A431 squamous cell carci-
`noma xenografts in athymic mice, with near complete
`inhibition of tumor growth during a 20-day treatment reg-
`imen at the highest doses.19
`In two phase I erlotinib dose-escalation studies in pa-
`tients with advanced solid malignancies,20 14 patients
`achieved stable disease. The most common adverse events
`were diarrhea and rash, regardless of dose and schedule.
`Diarrhea was considered the dose-limiting adverse event. The
`maximum-tolerated dose and recommended phase II dose
`was 150 mg once daily on a continuous dosing schedule.
`Results from pharmacokinetic studies showed that er-
`lotinib is highly protein bound in humans (92% to 95%).
`The primary route of metabolism is oxidation by the he-
`patic cytochromes CYP3A4 and CYP3A5 and the pulmo-
`nary cytochrome CYP1A1. A potential
`for drug-drug
`interaction exists when erlotinib is coadministered with
`drugs that are highly protein bound or are CYP3A4 inhibi-
`tors or inducers.
`The high frequency of overexpression of HER1/EGFR
`in NSCLC provides a scientific rationale for evaluating the
`therapeutic effect of erlotinib in this tumor type. In addi-
`tion, there is a clear need for new therapeutics to treat
`patients with NSCLC, especially those patients with ad-
`vanced disease who have a poor prognosis after failure of
`platinum-based chemotherapy. The majority of these pa-
`tients will not benefit from additional chemotherapy, in-
`cluding taxane regimens. The current study was planned to
`estimate the objective tumor response rate of erlotinib ad-
`ministered as a single agent to patients with advanced (stage
`IIIB or IV) or recurrent metastatic HER1/EGFR-positive
`NSCLC who were previously treated with platinum-based
`combination chemotherapy. Secondary objectives were to
`estimate the stable disease rate, duration of response, time
`to disease progression, overall and 1-year survival, health-
`related quality-of-life (HRQOL) outcomes, and safety pro-
`file of erlotinib in this population.
`
`PATIENTS AND METHODS
`
`Eligibility Criteria
`The study population included male and female patients 18
`years of age or older. The main criteria for inclusion were docu-
`mented stage IIIB or IV advanced or recurrent metastatic NSCLC,
`disease progression or relapse after platinum-based therapy, mea-
`surable disease, and documentation of HER1/EGFR positivity.
`Additional criteria included Eastern Cooperative Oncology Group
`(ECOG) performance status of 0 to 2 and adequate bone marrow,
`hepatic, and renal function (total bilirubin and creatinine ⱕ 1.5 ⫻
`
`the upper limit of normal). Patients with brain metastases were
`eligible if they were clinically stable for at least 8 weeks.
`
`Procedures Performed at Screening
`The procedures performed at screening included a complete
`medical and surgical history, standard laboratory studies, ECG,
`pregnancy test with a negative result, tumor assessment, determi-
`nation of HER1/EGFR status by immunohistochemistry in a tu-
`mor specimen conducted by a central laboratory (IMPATH, Los
`Angeles, CA), with HER1/EGFR positivity defined as more than
`10% of cells staining positive, and administration of an HRQOL
`questionnaire. All patients gave written informed consent in ac-
`cordance with policies of local human subjects committees before
`screening and initiation of therapy.
`
`Procedures Performed During the Study
`Procedures that were to be completed at weeks 2, 4, 6, and 8,
`every 4 weeks throughout the study, and at the time of study
`discontinuation included an interval history and reassessment of
`performance status, brief physical and skin examination, weight,
`and vital signs, complete blood count with differential and platelet
`count, blood biochemistry, and urinalysis. Ophthalmologic eval-
`uations were to be repeated after 4 weeks of erlotinib therapy,
`within 2 weeks after any dose escalation, and at the time of study
`discontinuation only if a change from baseline had been detected
`with subsequent examinations.
`
`Treatment Plan
`Patients received erlotinib at an initial dose of 150 mg in a
`tablet formulation supplied by the sponsor (OSI Pharmaceuticals)
`that was self-administered orally once daily on a continuous basis.
`The dose was taken in the morning with up to 200 mL of water.
`The dose could be increased to 200 mg/d in patients who had
`received at least 4 weeks of continuous dosing at 150 mg/d and did
`not experience any drug-related adverse events during the previ-
`ous 4-week cycle. The dosage was to be decreased in 25- or 50-mg
`decrements if the patient experienced drug-related ocular toxici-
`ties of any National Cancer Institute Common Toxicity Criteria
`grade, had any drug-related adverse events subjectively considered
`intolerable, had any Common Toxicity Criteria ⱖ grade 3 drug-
`related adverse events not controlled with optimal supportive
`medication, or had undergone dose reduction for a drug-related
`adverse event that did not improve by at least one grade level to less
`than grade 3 within 2 weeks. Therapy could be continued after
`dose reduction to the minimum daily dose of 25 mg despite
`drug-related toxicity if the investigators and sponsor felt it was in
`the patient’s interest.
`Erlotinib treatment was planned for a minimum of 8
`weeks and was to continue for a maximum of 52 weeks unless
`disease progression or unmanageable toxicity occurred. Pa-
`tients with stable or responding disease for whom additional
`therapy beyond 52 weeks was deemed to be of potential benefit
`could continue erlotinib.
`
`Tumor Measurements
`Measurements of disease sites by clinical examination and
`radiographic imaging studies (x-ray, computed tomography scan,
`and magnetic resonance imaging) were collected at baseline before
`erlotinib therapy. The same methods were used every 8 weeks
`during the study to assess response. If a patient achieved a com-
`plete or partial response, tumor measurements were repeated 4
`weeks later to confirm the response.
`
`www.jco.org
`
`3239
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`
`
`Pérez-Soler et al
`
`HRQOL
`HRQOL was measured using the European Organization for
`Research and Treatment of Cancer Quality of Life Questionnaire
`C30, version 3.0.21 An additional lung cancer module, Quality of
`Life Questionnaire LC13, was used and is composed of 13 items; it
`is intended for use among a wide range of lung cancer patients
`varying in disease stage and treatment modality. Four additional
`questions were added to assess the effect of any rash, skin pain, and
`itching, and their impact on daily activities. These four questions
`were combined into a single rash score.
`The HRQOL questionnaire was administered at baseline,
`every 2 weeks during the first 2 months and monthly thereafter, at
`the end of study, and at the 1-month and 3-month follow-up
`visits. The questionnaire was to be completed during office visits
`before any other evaluations or assessment of adverse events.
`Changes in HRQOL scores during the study were compared with
`the baseline score.
`Criteria for Evaluation of Efficacy
`The primary efficacy variable was the overall response rate,
`which was defined as the percentage of patients with complete or
`partial responses. Responses were determined by the investigators
`according to WHO criteria.22
`Evaluation of the objective tumor response was performed
`every 8 weeks during treatment. Stable disease, duration of overall
`response, time to progression, and survival were assessed at
`2-week intervals for the first 2 months, every 4 weeks until study
`discontinuation, and at 1-month and 3-monthly intervals after
`treatment. Patients who met the criteria for complete or partial
`response had their response confirmed at least 4 weeks after the
`first determination of response. In addition, responses were also
`evaluated by the sponsor (OSI Pharmaceuticals) according to
`Response Evaluation Criteria in Solid Tumor (RECIST) criteria.23
`Statistical Analysis
`This was a single-arm, open-label, multicenter phase II study.
`A Gehan two-stage design was used to determine sample size.24 It
`was anticipated that a total of 37 patients would be enrolled to
`ensure that 33 patients would be fully assessable for response. With
`33 patients, the response rate of interest (10%) would be estimated
`with a maximum SE of 9%. A total of 57 patients were accrued to
`compensate for patients who had early progression and/or death
`before completing 8 weeks of study therapy. All 57 patients were
`included in all efficacy and safety analyses. The parameters of
`interest were estimated and presented with their 95% CIs using
`exact methods. All time-to-event variables, including duration of
`response and progression-free and overall survival, were analyzed
`using Kaplan-Meier product-limit survival estimates. Pretreat-
`ment characteristics were analyzed in univariate and multivariate
`logistic regression models for their ability to predict objective
`response and in univariate and multivariate Cox proportional
`hazards models for their ability to predict survival. Multivariate
`models were constructed using stepwise variable selection tech-
`niques. Changes in HRQOL from baseline, including impact of
`disease-related symptoms and rash, were evaluated using paired t
`tests. All analyses were performed using SAS/STAT User’s Guide
`version 8.2 for Windows (SAS Institute, Cary, NC).
`
`RESULTS
`
`A total of 84 patients were screened for the study; 57 were
`enrolled and treated with erlotinib. The most common
`
`Table 1. Summary of Baseline Patient and Disease Characteristics
`
`Demographic and Disease Characteristic
`
`No. of Patients
`(N ⫽ 57)
`
`Sex
`Female
`Male
`Race
`White
`Black
`Hispanic
`Asian
`Age
`31-39 years
`40-64 years
`65-69 years
`ⱖ 70 years
`ECOG performance status
`0
`1
`2
`Smoking classification
`Smoker
`Ex-smoker
`Never smoked
`NSCLC stage
`IV
`IIIB
`Histologic classification
`Adenocarcinoma
`Large cell
`Squamous cell
`NOS
`Pathologic grade
`Poorly differentiated
`Moderately differentiated
`Well differentiated
`Not assessable/not available
`HER1/EGFR expression
`Strong
`Weak to strong
`Weak
`HER1/EGFR stain
`10-19% cells
`20-39% cells
`40-59% cells
`60-79% cells
`80-100% cells
`
`34
`23
`
`52
`2
`2
`1
`
`3
`33
`7
`14
`
`6
`44
`7
`
`5
`42
`10
`
`48
`9
`
`35
`11
`9
`2
`
`24
`18
`8
`7
`
`32
`19
`6
`
`4
`17
`9
`6
`21
`
`%
`
`59.6
`40.4
`
`91.2
`3.5
`3.5
`1.8
`
`5.3
`57.9
`12.3
`24.6
`
`10.5
`77.2
`12.3
`
`8.8
`73.7
`17.5
`
`84.2
`15.8
`
`61.4
`19.3
`15.8
`3.5
`
`42.1
`31.6
`14.0
`12.3
`
`56.1
`33.3
`10.5
`
`7.0
`29.8
`15.8
`10.5
`36.8
`
`Abbreviations: ECOG, Eastern Cooperative Oncology Group; NSCLC,
`non–small-cell lung cancer; NOS, not otherwise specified; HER1/EGFR,
`epidermal growth factor receptor.
`
`reasons that screened patients were not enrolled onto the
`study were HER1/EGFR status (eight patients were nega-
`tive, one was too weak, and five had no tissue specimens
`available) and rapid deterioration (four patients). The 57
`patients were accrued from January 25, 2000, until February
`14, 2001. All patients were assessable for tumor response
`and toxicity. Table 1 shows the characteristics of the patient
`population. The median age was 62 years (range, 31 to 83
`
`3240
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`
`
`Erlotinib Monotherapy in NSCLC
`
`Table 2. Summary of Previous Therapy for NSCLC
`
`No. of
`Patients
`(N ⫽ 57)
`
`Characteristic
`
`Prior therapy for NSCLC
`Chemotherapy
`Surgery
`Radiation therapy, primary/metastatic
`Hormonal/immunologic therapy
`Time from initial diagnosis to erlotinib therapy
`⬍ 6 months
`6-12 months
`⬎ 12 months
`Median, months
`Range, months
`No. of prior chemotherapy regimens
`1
`2
`3 or more
`Median, No. of prior regimens
`Range, No. of prior regimens
`No prior platinum therapy
`Prior platinum-based treatment
`Prior docetaxel treatment
`Time from last regimen to erlotinib therapy
`⬍ 6 months
`6-12 months
`⬎ 12 months
`Median, months
`Range, months
`
`Abbreviation: NSCLC, non–small-cell lung cancer.
`
`57
`51
`42
`5
`
`5
`16
`36
`
`10
`24
`23
`
`2
`55
`15
`
`44
`8
`5
`
`%
`
`100.0
`89.4
`73.7
`8.7
`
`8.8
`28.1
`63.2
`
`17.7
`4.2-136.6
`
`17.5
`42.1
`40.4
`
`3.5
`96.5
`26.3
`
`77.2
`14.0
`8.8
`
`2
`1-8
`
`3
`0.5-26
`
`years), and the majority of patients (63%) had been diag-
`nosed with NSCLC more than 12 months before study
`enrollment. The median time from initial diagnosis to study
`entry was 17.7 months (range, 4 to 137 months). The ad-
`vanced stage of disease in these patients was characterized
`by multiple sites of distant metastases and the presence of
`lung cancer signs and symptoms at baseline. Fifty-four pa-
`tients (95%) reported symptoms at baseline, including fa-
`tigue (67%), dyspnea (61%), and cough (60%).
`
`Before study entry, the patients had received various
`therapies (Table 2). Eighty-two percent of patients had
`received two or more chemotherapy regimens. Fifteen pa-
`tients (26%) had received prior docetaxel. All but two pa-
`tients had been treated with at least one platinum-based
`combination chemotherapy regimen. One of the two pa-
`tients who had not received platinum therapy had been
`treated with gemcitabine-based combinations, and the
`other patient had been treated with two regimens of pacli-
`taxel. The majority of patients (44 of 57 patients, 77%) had
`documented disease progression during or within 6 months
`of their last chemotherapy regimen.
`
`Tumor Response and Survival
`Table 3 lists the antitumor response data. Two patients
`achieved a complete response, and five had a partial re-
`sponse, as determined by both WHO and RECIST criteria.
`The objective response rate (complete ⫹ partial response)
`was 12.3% (95% CI, 5.1% to 23.7%). On the basis of
`RECIST criteria, 22 patients (39%) had stable disease and
`28 patients (49%) had disease progression as their best
`tumor response. The number of prior chemotherapy regi-
`mens had no effect on response rates (12.8% for patients
`with two or more prior chemotherapy regimens v 12.3% for
`the whole group). One of the seven responders had not
`received prior platinum therapy; this 83-year-old patient,
`who had pre-existing neuropathy contraindicating plati-
`num therapy, had previously received two gemcitabine-
`based combinations, one of which included docetaxel. Of
`the 15 patients previously treated with docetaxel, four
`(27%) achieved complete or partial responses, and an
`additional five (33%) had stable disease. Of the 44 pa-
`tients with documented disease progression within 6
`months of their last chemotherapy treatment, three (7%)
`subsequently achieved a partial response with erlotinib,
`and 16 (36%) had stable disease.
`Thirty-five patients enrolled had adenocarcinoma,
`four of whom responded to therapy (11.4%; one complete
`and three partial responses). Of the remaining 22 patients,
`
`Table 3. Response to Erlotinib
`
`Investigator Best Response
`(WHO criteria)
`
`Response
`
`No. of Patients
`
`CR
`PR
`Stable disease
`Progressive disease
`Not assessable
`Overall response rate, CR ⫹ PR
`95% CI, %
`
`2
`5
`20
`28
`2
`7
`
`%
`
`3.5
`8.8
`35.1
`49.1
`3.5
`12.3
`
`5.1 to 23.7
`
`5.1 to 23.7
`
`Abbreviations: RECIST, Response Evaluation Criteria in Solid Tumor; CR, complete response; PR, partial response.
`
`www.jco.org
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`Sponsor Best Response
`(RECIST criteria)
`
`No. of Patients
`
`2
`5
`22
`28
`—
`7
`
`%
`
`3.5
`8.8
`38.6
`49.1
`—
`12.3
`
`3241
`
`
`
`Pérez-Soler et al
`
`(95% CI, 4.8 to 13.9 months), and the 1-year survival rate
`was 40% (95% CI, 28% to 54%; Fig 1).
`
`Determinants of Tumor Response and Survival
`Several pretreatment characteristics were analyzed in
`univariate and multivariate logistic regression models for
`their ability to predict the objective response rate. Table 4
`shows the results of these analyses. Time from last chemo-
`therapy was the only pretreatment characteristic that signif-
`icantly predicted the objective response rate in the
`multivariate analysis (P ⫽ .033), although time since initial
`diagnosis was marginally predictive (P ⫽ .086).
`The same pretreatment characteristics were analyzed in
`univariate and multivariate Cox proportional hazards
`models for their ability to predict survival (Table 5). Time
`since initial diagnosis and ECOG performance status were
`the only pretreatment characteristics that predicted survival
`in the multivariate model.
`Patients previously treated with docetaxel as second-
`line therapy had a higher objective response rate compared
`with the entire group (26.7% v 12.3% overall, respectively),
`a similar median survival time of 8.6 months (95% CI, 2.2 to
`20.8 months), and a 1-year survival rate of 47% (95% CI,
`21.3% to 73.4%).
`Exploratory analyses were conducted to investigate any
`potential relationship between rash and clinical outcomes.
`For these analyses, rash was defined as MedDRA (Medical
`Dictionary for Regulatory Activities, version 5.0; MedDRA
`MSSO, Reston, VA) codes that contain the terms rash,
`dermatitis, or acne. Rash was experienced by all seven pa-
`tients who had an objective response and by 21 (95%) of 22
`patients who had stable disease, but only 15 (54%) of 28
`patients (54%) who had progressive disease experienced
`rash (data not shown). Thus, rash was necessary but was not
`a sufficient condition for tumor response in this study.
`
`Fig 1. Overall survival of patients treated with erlotinib. Bullets represent
`patients still alive at time of analysis.
`
`11 had large-cell carcinoma, nine had squamous cell carci-
`noma, and two were not specified; three patients responded
`to therapy (13.6%), two with large-cell carcinoma (one
`complete and one partial response) and one with squamous
`cell carcinoma (partial response). Four of the seven re-
`sponding patients had adenocarcinoma (one complete re-
`sponse and three partial responses). Histologies were not
`further classified by subtype; specifically, BAC or BAC-like
`features were not characterized.
`The median duration of response was 19.7 weeks
`(range, 11.7 to 80.3 weeks). Progression-free survival was
`measured from the first erlotinib administration to the date
`of disease progression, start of subsequent anticancer treat-
`ment, death, or date of last contact, whichever occurred
`first. With five patients censored in the analysis, the median
`progression-free survival time was 9 weeks (95% CI, 8 to 15
`weeks). With nine patients still alive and censored in the
`analysis, the median overall survival time was 8.4 months
`
`Table 4. Univariate and Multivariate Logistic Regression Analyses to Predict Response (CR ⫹ PR)
`
`Factors
`
`Univariate
`P
`
`Multivariate
`P
`
`Factors in the final model
`Time from last chemotherapy ⬍ 6 v ⱖ 6 months
`Factors not in the final model
`Time since initial diagnosis, ⬍ 12 months v ⱖ 12 months
`HER1/EGFR staining intensity, weak, weak/strong, strong
`Stage of disease, IIIB v IV
`Sex, male v female
`No. of prior chemotherapy regimens, 1 v ⱖ 2
`Age, ⬎ 70 v ⱕ 70 years
`Histology, adenocarcinoma v other
`ECOG performance status, 0, 1, 2
`
`.021
`
`.031
`.28
`.22
`.50
`.81
`.50
`.80
`.34
`
`.033
`
`.086
`.19
`.28
`.39
`.50
`.65
`.79
`.90
`
`OR
`
`6.1
`
`95% CI
`
`1.2 to 32.0
`
`Abbreviations: CR, complete response; PR, partial response; OR, odds ratio; HER1/EGFR, epidermal growth factor receptor; ECOG, Eastern Cooperative
`Oncology Group.
`
`3242
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`
`
`Erlotinib Monotherapy in NSCLC
`
`Table 5. Univariate and Multivariate Analyses to Predict Survival
`
`Factors
`
`Univariate
`P
`
`Multivariate
`P
`
`Factors in the final model
`Time since initial diagnosis, ⬍ 12 months v ⱖ 12 months
`ECOG performance status, 0, 1, 2
`Factors not in the final model
`Stage of disease, IIIB v IV
`Time from last chemotherapy, ⬍ 6 v ⱖ 6 months
`HER1/EGFR staining intensity, weak, weak/strong, strong
`No. of prior chemotherapy regimens, 1 v ⱖ 2
`Age, ⬎ 70 v ⱕ 70 years
`Sex, male v female
`Histology, adenocarcinoma v other
`
`.0001
`.019
`
`.52
`.007
`.91
`.91
`.12
`.48
`.69
`
`.0007
`.04
`
`.14
`.27
`.42
`.48
`.58
`.59
`.68
`
`HR
`
`0.35
`1.92
`
`95% CI
`
`0.19 to 0.64
`1.02 to 3.61
`
`Abbreviations: HR, hazards ratio; ECOG, Eastern Cooperative Oncology Group; HER1/EGFR, epidermal growth factor receptor.
`
`In addition, patients who experienced rash had signif-
`icantly longer survival (Fig 2). The median survival of pa-
`tients without rash was 1.5 months compared with 8.5 and
`19.6 months for patients with a maximum of grade 1 rash
`and grade 2 or 3 rash (only one patient experienced grade 3
`rash), respectively. All pair-wise comparisons were statisti-
`cally significant. When rash was included in the multivari-
`ate analysis along with the factors listed in Table 5, rash was
`the most significant predictor of survival, with hazard ratios
`of 0.13 (95% CI, 0.06 to 0.30) and 0.05 (95% CI, 0.02
`to 0.15) for grade 1 and grade 2 or 3 rash, respectively
`(P ⬍ .0001 for each factor; Table 6).
`The median time to the first occurrence of rash, regard-
`less of severity, was 10 days (range, 2 to 44 days), whereas
`the median duration of erlotinib exposure was 9 weeks
`
`Fig 2. Survival of patients by grade of rash. Bullets represent patients (Pts)
`still alive at time of analysis.
`
`(range, 2 to 131 weeks). Therefore, it is unlikely that the rela-
`tionship between rash and increased survival can be explained
`simply by longer exposure to erlotinib. However, to investigate
`this possibility, we performed an additional multivariate anal-
`ysis in which rash, regardless of severity, was included as a
`time-dependent variable, and cumulative dose of erlotinib was
`included along with all of the factors in Table 5. Rash contin-
`ued to be a significant predictor of survival, with a hazard ratio
`of 0.24 (95% CI, 0.10 to 0.56; data not shown).
`Symptom Improvement and HRQOL
`The HRQOL questionnaires showed that patients in
`this study had multiple lung cancer–related symptoms and
`relatively good functional status, which was consistent with
`a European Organization for Research and Treatment of
`Cancer reference population with recurrent or metastatic
`NSCLC. The overall incidence of fatigue, dyspnea, and
`cough, which were the most frequent signs and symptoms
`captured by the HRQOL, decreased from 67%, 61%, and
`60% at baseline to 49%, 37%, and 39%, respectively, after
`initiation of erlotinib therapy (data not shown).
`Although the results must be treated cautiously be-
`cause the P values were not corrected for the large number
`of comparisons, some significant (P ⬍ .05) changes were
`seen. Pain was reduced at 2 weeks of therapy, and emotional
`functioning was increased during the first 4 weeks of ther-
`apy, but both subsequently returned to baseline levels. Di-
`arrhea and sore mouth were reported as increased from
`baseline while on therapy.
`Further analysis of the HRQOL scales suggested that
`responders sustained their quality of life longer than non-
`responders, but few comparisons reached statistical signif-
`icance. These relationships may serve as a historical
`comparison for future randomized studies.
`Adverse Events
`All 57 patients received at least one dose of erlotinib
`and were analyzed for safety. The majority of patients re-
`
`www.jco.org
`
`3243
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of Clinical Oncology. All rights reserved.
`
`
`
`Pérez-Soler et al
`
`Table 6. Univariate and Multivariate Analyses to Predict Survival Including Rash
`
`Factors
`
`Univariate
`P
`
`Multivariate
`P
`
`Factors in the final model
`Rash, grade 2/3, grade 2/3 v no rash
`Rash, grade 1, grade 1 v no rash
`Age, ⬎ 70 v ⱕ 70 years
`Time from last chemotherapy, ⬍ 6 v ⱖ 6 months
`Stage of disease, IIIB v IV
`Factors not in the final model
`ECOG performance status, 0, 1, 2
`Histology, adenocarcinoma v other
`HER1/EGFR staining intensity, weak, weak/strong, strong
`Sex, male v female
`No. of prior chemotherapy regimens, 1 v ⱖ 2
`Time since initial diagnosis, ⬍ 12 months v ⱖ 12 months
`
`⬍ .0001
`⬍ .0001
`.12
`.007
`.52
`
`.019
`.69
`.91
`.48
`.91
`.0001
`
`⬍ .0001
`⬍ .0001
`.017
`.038
`.057
`
`.40
`.48
`.50
`.52
`.74
`.81
`
`HR
`
`0.05
`0.13
`2.30
`0.44
`0.45
`
`95% CI
`
`0.02 to 0.15
`0.06 to 0.30
`1.16 to 4.58
`0.20 to 0.96
`0.20 to 1.02
`
`Abbreviations: HR, hazards ratio; ECOG, Eastern Cooperative Oncology Group; HER1/EGFR, epidermal growth factor receptor.
`
`ceived the 150 mg/d target dose. Five patients (9%) discon-
`tinued erlotinib as a result of an adverse event or withdrawal
`of consent, and two additional patients (4%) required dose
`reduction because of an adverse event after receiving mul-
`tiple cycles at 150 mg. The median duration of erlotinib
`exposure was 9 weeks (range, 2 to ⬎131 weeks).
`Table 7 lists the most common drug-related adverse
`events. Fifty-six patients (98%) had at least one drug-
`related adverse event, 38 patients (67%) had drug-related
`adverse events with a maximum severity of grade 1 or 2, and
`17 patients (30%) had at least one grade 3 drug-related adverse
`event. Less than 10% of patients showed signs of ocular toxicity
`attributable to erlotinib therapy; no incidence exceeded grade
`2. Dysphagia, pruritus, fatigue, dyspnea, decreased appetite,
`and anxiety were the only drug-related grade 3 events reported
`in two patients, and none was reported in more than two
`patients (4%). Interstitial pneumonia and grade 4 events were
`not reported.
`The incidence of drug-related rash and diarrhea during
`the study was of special interest. Rash, not otherwise speci-
`fied, was experienced by 67% of patients. When the defini-
`tion was expanded to include dermatitis and acne, the
`incidence increased to 75% (43 of 57 patients). Rash and/or
`diarrhea occurred as single events or concurrently in 89% of
`patients. Three patients (5%) experienced grade 3 derma-
`tologic events (one rash and two pruritus without associ-
`ated rash). Only one patient had grade 3 diarrhea (also with
`grade 3 rash). Grade 4 rash or diarrhea was not noted. The
`median time to the first occurrence of rash, regardless of
`severity, was 10 days (range, 2 to 44 days). The median time
`to the first occurrence of diarrhea, regardless of severity, was
`14 days (range, 1 to 420 days).
`Three patients (5%) discontinued therapy because of
`adverse events. Only one withdrawal was considered related
`to erlotinib (grade 2 nausea and vomiting). A total of 20
`
`patients (35%) died during treatment or within 30 days of
`the last dose of erlotinib. These deaths were a result of
`disease progression (85%) or respiratory events attributed
`to underlying disease (15%); none was a result of pneumo-
`
`Table 7. Incidence of Drug-Related Adverse Events Occurring in at
`Least 10% of Patients by CTC Grade
`
`Adverse Event
`
`Rash NOS
`Diarrhea
`Dry skin
`Pruritus
`Fatigue
`Nausea
`Appetite decreased
`Dyspnea
`Anxiety
`Vomiting
`Glossodynia
`Stomatitis
`Erythema
`Cough
`Depression
`Pain
`Arthralgia
`Constipation
`Dyspepsia
`Dysphagia
`Insomnia
`Weight decreased
`Dry mouth
`Dysgeusia
`Paresthesia
`
`No. of
`Patients
`
`38
`32
`20
`20
`16
`14
`13
`12
`12
`11
`10
`10
`10
`9
`9
`8
`8
`7
`7
`7
`7
`7
`6
`6
`6
`
`Maximum CTC Grade
`(No. of patients)
`
`%
`
`67
`56
`35
`35
`28
`25
`23
`21
`21
`19
`18
`18
`18
`16
`16
`14
`14
`12
`12
`12
`12
`12
`11
`11
`11
`
`1
`
`22
`23
`14
`14
`8
`10
`6
`7
`6
`9
`7
`7
`7
`6
`4
`3
`3
`6
`5
`5
`3
`7
`4
`3
`5
`
`2
`
`15
`8
`5
`4
`6
`4
`5
`3
`4
`2
`2
`2
`3
`3
`5
`4
`5
`1
`1
`0
`3
`0
`1
`3
`1
`
`3
`
`1
`1
`1
`2
`2
`0
`2
`2
`2
`0
`1
`1
`0
`0
`0
`1
`0
`0
`1
`2
`1
`0
`1
`0
`0
`
`4
`
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`0
`
`Abbreviations: CTC, Common Toxicity Criteria; NOS, not other-
`wise specified.
`
`3244
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`Downloaded from jco.ascopubs.org on June 17, 2016. For personal use only. No other uses without permission.
`Copyright © 2004 American Society of C