Selumetinib plus docetaxel for KRAS-mutant

Selumetinib plus docetaxel for KRAS-mutant advancednon-small-cell lung cancer: a randomised, multicentre, placebo-controlled, phase 2 studyPasi A Jänne, Alice T Shaw, José Rodrigues Pereira, Gaëlle Jeannin, Johan Vansteenkiste, Carlos Barrios, Fabio Andre Franke, Lynda Grinsted, Victoria Zazulina, Paul Smith, Ian Smith, Lucio CrinòSummaryBackground No targeted therapies are available for KRAS-mutant non-small-cell lung cancer (NSCLC). Selumetinib is an inhibitor of MEK1/MEK2, downstream of KRAS, with preclinical evidence of synergistic activity with docetaxel in KRAS-mutant cancers. We did a prospective, randomised, phase 2 trial to assess selumetinib plus docetaxel in previously treated patients with advanced KRAS-mutant NSCLC.Methods Eligible patients were older than 18 years of age; had histologically or cytologically confi rmed stage IIIB–IV KRAS-mutant NSCLC; had failed fi rst-line therapy for advanced NSCLC; had WHO performance status of 0–1; had not received previous therapy with either a MEK inhibitor or docetaxel; and had adequate bone marrow, renal, and liver function. Patients were randomly assigned (in a 1:1 ratio) to either oral selumetinib (75 mg twice daily in a 21 day cycle) or placebo; all patients received intravenous docetaxel (75 mg/m² on day 1 of a 21 day cycle). Randomisation was done with an interactive voice response system and investigators, patients, data analysts, and the trial sponsor were masked to treatment assignment. The primary endpoint was overall survival, analysed for all patients with confi rmed KRAS mutations. This study is registered with , number NCT00890825.Findings Between April 20, 2009, and June 30, 2010, we randomly assigned 44 patients to receive selumetinib and docetaxel (selumetinib group) and 43 to receive placebo and docetaxel (placebo group). Of these, one patient in the selumetinib group and three in the placebo group were excluded from effi cacy analyses because their tumours were not confi rmed to be KRAS-mutation positive. Median overall survival was 9∙4 months (6∙8–13∙6) in the selumetinib group and 5∙2 months (95% CI 3∙8–non-calculable) in the placebo group (hazard ratio [HR] for death 0∙80, 80% CI 0∙56–1∙14; one-sided p=0∙21). Median progression-free survival was 5∙3 months (4∙6–6∙4) in the selumetinib group and 2∙1 months (95% CI 1∙4–3∙7) in the placebo group (HR for progression 0∙58, 80% CI 0∙42–0∙79; one-sided p=0∙014). 16 (37%) patients in the selumetinib group and none in the placebo group had an objective response (p<0∙0001). Adverse events of grade 3 or higher occurred in 36 (82%) patients in the selumetinib group and 28 (67%) patients in the placebo group. The most common grade 3–4 adverse events were neutropenia (29 [67%] of 43 patients in the selumetinib group vs 23 [55%] of 42 patients in the placebo group), febrile neutropenia (eight [18%] of 44 patients in the selumetinib group vs none in the placebo group), dyspnoea (one [2%] of 44 patients in the selumetinib group vs fi ve [12%] of 42 in the placebo group), and asthenia (four [9%] of 44 patients in the selumetinib group vs none in the placebo group).Interpretation Selumetinib plus docetaxel has promising effi cacy, albeit with a higher number of adverse events than with docetaxel alone, in previously treated advanced KRAS-mutant NSCLC. These fi ndings warrant further clinical investigation of selumetinib plus docetaxel in KRAS-mutant NSCLC.Funding AstraZeneca.IntroductionLung cancer is one of the leading causes of cancer mortality worldwide,1 with 5 year survival of less than 15% across all stages of disease in Europe.2 I n unselected patients with advanced non-small-cell lung cancer (NSCLC), there have been few improvements in outcomes for patients after chemotherapy.3 By contrast, kinase inhibitors have been particularly effective for subsets of patients with NSCLC defined by molecular oncogenic alterations such as EGFR muta t ions (gefi tinib and erlotinib)4,5 and ALK rearrangements (crizotinib).6KRAS is the most frequently mutated oncogene in NSCLC, with mutations detected in about 20% of tumours in white people. KRAS mutations also predict for absence of benefit from EGFR tyrosine kinase inhibitors,7 are a negative prognostic factor for survival,8 and may be associated with a reduced benefi t from adjuvant chemotherapy.9Although KRAS mutations were identified in lung cancer nearly 20 years ago,10 no approved therapies exist for KRAS-mutant NSCLC and few trials have specifi cally addressed this population of patients. Up to now, direct inhibition of KRAS has proven clinically challenging. Farnesyl transferase inhibitors, designed to specifi cally inhibit RAS proteins, showed little clinical effi cacy.11 Eff orts to inhibit mutant KRAS in NSCLC have thereforeLancet Oncol 2013; 14: 38–47Published OnlineNovember 28, 2012 /10.1016/S1470-2045(12)70489-8See Comment page 3 Lowe Center for ThoracicOncology and the Belfer Institute for Applied CancerScience, Dana-Farber CancerInstitute, Boston, MA, USA (P A Jänne MD);Massachusetts General Hospital Cancer Center,Boston, MA, USA(A T Shaw MD);InstitutoBrasileiro de CancerologiaTorácica, Sao Paulo, Brazil (J Rodrigues Pereira MD);Service de Pneumologie, Hôpital Gabriel Montpied, Clermont-Ferrand, France (G Jeannin MD); Respiratory Oncology Unit and Trial Unit, Department ofPneumology, UniversityHospital Gasthuisberg,Leuven, Belgium (Prof J Vansteenkiste MD);PUCRS School of Medicine,Centro de Pesquisa em Oncologia, Porto Alegre, Brazil(C Barrios MD);CACON,Hospital de Caridade de Ijuí,Ijuí, Brazil (F A Franke MD); AstraZeneca UK, Macclesfi eld,UK (L Grinsted MSc, V Zazulina MD, P Smith PhD,I Smith MD);and Medical Oncology Division, Hospital S Maria della Misericordia, Perugia, Italy (Prof L Crinò MD)Correspondence to: Dr Pasi A Jänne, Lowe Center for Thoracic Oncology, Dana-FarberCancer Institute, Boston,MA 02215, USApjanne@targeted eff ector proteins downstream in the RAS-RAF-MEK-ERK (MAPK) signalling pathway.The MAPK pathway converges at the MEK1/MEK2 (also known as MAP2K1 and MAP2K2) kinases, for which the only known substrates are the ERK1/ERK2 (also known as MAPK3 and MAPK1) kinases. MEK inhib-ition would therefore be expected to block ERK signal-ling irrespective of the upstream stimulus. Selumetinib (AZD6244, ARRY-142886; AstraZeneca, Alderley Park, Cheshire, UK) is an orally available, potent, selective, non-ATP competitive inhibitor of MEK1/MEK2 kinases.12 Preclinical data from KRAS-mutant NSCLC tumour xenografts showed that selumetinib signifi cantly sup-pressed tumour growth.13 Early phase clinical studies of selumetinib monotherapy showed target inhibition14 and tumour responses.15 However, in a phase 2 study, selumetinib monotherapy showed little clinical activity in an unselected pretreated population with NSCLC.16 Results of additional preclinical in-vivo studies have shown that the combination of selumetinib and docetaxel leads to greater tumour-growth inhibition or regression, and apoptosis,17,18 and this eff ect is synergistic.13 A phase 1 study of selumetinib plus docetaxel showed a man-ageable tolerability profi le in advanced solid tumours.19 The aim of this phase 2 trial was to assess the effi cacy and safety of selumetinib in combination with docetaxel as second-line treatment for patients with KRAS-mutant advanced NSCLC.MethodsStudy design and patientsWe did this double-blind, placebo-controlled, randomised phase 2 study in 67 sites across 12 countries. We included patients aged 18 years or older with histologically or cytologically confirmed, locally advanced or metastatic, KRAS-mutant NSCLC (stage IIIB–IV), after failure of fi rst-line chemotherapy. KRAS-mutation status was confi rmed before randomisation by an AstraZeneca-appointed cen-tral laboratory with Amplification Refractory Mutation System analysis,20 or by an AstraZeneca-agreed local la-boratory with agreed methods, or an accredited com-mercial laboratory with proprietary-accredited methods. No subsequent cross validation of mutational testing was done. Additional inclusion criteria were WHO perform-ance status 0–1, and adequate bone marrow, renal, and liver function. Patients were ineligible if they had received more than one chemotherapy, any previous treatment with a MEK inhibitor or docetaxel-containing regimen, or had mixed small cell and NSCLC histology. Patients with brain metastases or spinal-cord compression were eligible if they had received therapy and had been stable without steroids and anticonvulsants for at least 1 month.All patients provided written informed consent and the study was done in accordance with the Declaration of Helsinki. The protocol was approved by the institutional review board at each study site and complied with local country regulations.Randomisation and maskingEligible patients were randomly assigned, in a1:1 ratio (block size of four), to receive docetaxel incombination with either selumetinib (selumetinibgroup) or placebo (placebo group). Treatment-groupassignment was done by means of an interactive voiceresponse system at a central location. The voice responsesystem allocated randomisation numbers and medicationpack codes. Patients and study team were masked to thetreatment assigned; selu m etinib and placebo capsuleswere identical in appear a nce and presented in the samepackaging to ensure blinding of the medication.ProceduresPatients received intravenous docetaxel 75 mg/m² onday 1 of every 21 day cycle and were expected to receiveup to six cycles. Investigators were allowed to reduce thenumber of cycles if severe toxic eff ects developed, or toadminister more than six cycles when such treatmentdid not contravene local practice. I nvestigators wereadvised to consider administration of granulocyteFigure 1: Trial profi le30 patients in the selumetinib group completed the study, 13 were still receiving treatment at the time of data cutoff , and one discontinued treatment voluntarily. 29 patients in the placebo group completed the study and 14 were still receiving treatment at the time of data cutoff .colony-stimulating factor after occurrence of either febrile neutropenia or grade 4 neutropenia (according to the National Cancer I nstitute Common Terminology Criteria for Adverse Events [CTCAE], version 321), lasting for more than 7 days. Granulocyte colony-stimulating factor was administered in line with local practice and ASCO guidelines.22Patients received either selumetinib hydrogen sulphate capsules 75 mg twice a day or matched placebo until disease progression or unacceptable toxic eff ects occurred. The dose reduction algorithm in the study allowed for three steps: 75 mg once a day (fi rst dose reduction), 50 mg twice a day (dose adjustment), and 50 mg once a day (fi nal dose reduction). Once reduced or adjusted, the dose was not allowed to return to the previous level. Treatment with selumetinib or placebo was withheld if an adverse event of grade 3 or higher or an intolerable adverse event, irrespective of grade, was recorded despite optimum supportive care for all adverse events deemed at least partly related to treatment.After disease progression, patients could receive any subsequent therapy at the discretion of the treating physician. No crossover was allowed.The primary endpoint was overall survival, defi ned as the time from randomisation until the date of death from any cause. Secondary endpoints were progression-free survival (time from randomisation to docu-mentation of disease progression or death), proportion of patients with an objective response (ie, a complete or partial response), duration of response, change in tumour size, 6 month progression-free survival, and safety. Assessments of prevalence, severity, and change over time of NSCLC-specifi c symptoms were exploratory endpoints.Tumour response was based on investigator assess-ment of target and non-target lesions using CT or MRI at baseline, week 6, week 12, and then every 12 weeks,relative to date of randomisation. Tumour measure-ments according to Response Evaluation Criteria in Solid Tumors (RECI ST version 1.0) were used to estab l ish progression-free survival, the proportion of patients with an objective response, duration of response, and change in tumour size at 12 weeks. Confi rmation of response was done at the next scheduled RECIST assessment (no less than 4 weeks) after the date the criteria for response were fi rst met.We assessed safety throughout the study, from in-formed consent until 30 days after study treatment was discontinued. We graded adverse events according to the National Cancer Institute CTCAE (version 3).21Patient-reported outcomes were assessed with a Lung Cancer Specifi c Symptom Questionnaire (LSSQ), which includes the Lung Cancer Subscale (LCS). The LSSQ consists of questions GP1, GP2, GP4, and GF1 from the Functional Assessment of Cancer Therapy-Lung plus the Additional Concerns section, excluding questions B5, Q3, and L5.23 Patients completed questionnaires on day 1 and every 3 weeks until discontinuation from study treatment. Statistical analysisThe primary objective of the trial was to assess effi cacy in the selumetinib plus docetaxel group. The protocol was amended on Nov 1, 2010 (after enrolment was complete and before analysis), to change the primary endpoint from progression-free survival to overall survival; the sample size was not changed. This change was undertaken to allow decisions to be made based on data for overall survival without breaking the study blind on the earlier endpoint of progression-free survival, therefore maintaining trial integrity, and was agreed by local regulatory authorities and ethics committees.We planned overall survival analysis for after about 58 events, assuming a hazard ratio (HR) of 0∙57, giving roughly 80% power to show a statistically signifi cant diff erence for overall survival, with a one-sided 10% sig-nifi cance level. About 80 patients were to be randomised. We analysed efficacy data according to randomised treatment, excluding patients who were randomised withFigure 2: Kaplan-Meier estimates of overall and progression-free survival(A) Overall survival. (B) Progression-free survival. Symbols show censored observations.an unconfi rmed KRAS-mutation test. We analysed endpoints of progression-free survival and overall sur-vival with a Cox proportional hazards model, allowing for the eff ect of treatment and including terms for WHO performance status, sex, histology, and smoking status. All patients who received at least one dose of selumetinib or placebo were included in the safety population. Analysis of the LCS score was done post hoc once the study results were unmasked. Improvement rate is the proportion of patients with improvement; improvement was defined as an increase of at least 3 points from baseline in LCS score (0–28); worsening was defi ned as at least a 3 point decrease. No change was defi ned as a score within 2 points. We analysed time-to-deterioration with a Cox proportional hazards model, allowing for the effect of treatment and including terms for sex, WHO performance status, histology, and smoking status. Improvement rate was analysed with a logistic regression model, allowing for the eff ect of treatment and including terms for sex, WHO performance status, and smoking status. We summarised total LSSQ scores by treatment. Absolute values and changes from baseline of total LSSQ scores were summarised by visit (excluding screening). Compliance was also summarised by visit and treatment. We did data analysis with SAS (version 8.1). 95% CI values for median time-to-event data were calculated with the method proposed by Brookmeyer and Crowley24 with S-Plus software (version 8.1.1). This study is registered with , number NCT00890825. Role of the funding sourceThe trial was funded by the study sponsor and designed by the principal investigator (PAJ) and the sponsor. The sponsor was responsible for data collection, analysed the data, and had a role in data interpretation. The authors vouch for the completeness and accuracy of the data and the data analyses. This report was written by the lead author, with editorial assistance funded by the sponsor, and was reviewed and approved for publication by all coauthors and the sponsor. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit the publication. ResultsWe recruited patients between April 20, 2009, and June 30, 2010. 422 patients were screened; 299 (71%) were excluded because their KRAS-mutation status was negative or unknown. 87 (21%) patients were random i sed: 44 to the selumetinib group and 43 to the placebo group (fi gure 1). One patient was randomised to the placebo group and received doxetaxel but died before receiving placebo, thus 86 patients were included in the safety analysis. Four patients—one in the selumetinib group and three in the placebo group—were excluded from effi cacy analyses because their tumours were not confi rmed to be KRAS-mutation positive. Baseline characteristics were generally well balanced between the two groups (tables 1, 2, 3). We recorded slight imbalances between groups in the number of patients with locally advanced disease, previous response to therapy, previous radio t herapy, and previous treatment with vinca alkaloids (tables 1, 2).Median treatment duration was 117 days (range 7–435) for selumetinib and 68 days (7–463) for placebo. Median number of docetaxel cycles was fi ve (range one to nine) in the selumetinib group and four (one to 12) in the placebo group. Relative dose intensity of docetaxel was 86% (SD 18∙2) in the selumetinib group and 91% (17∙9) in the placebo group. 16 patients (36%) had a dose reduction and 24 patients(55%) had a dose interruption of selumetinib (table 4). Four patients (10%) had a dose interruption of placebo; there were no dose reductions of placebo (table 4).At the time of data cutoff(on May 1, 2011), 56 of 83 patients available for effi cacy analyses had died—29 (67%) patients in the selumetinib group and 27 (68%) patients in the placebo group. The survival status was known for 82 of 83 patients; median follow-up was 7∙2 months (95% CI 3∙5–10∙7). Median overall survival was 287 days (9∙4 months; 95% CI 6∙8–13∙6) in the selumetinib group and 157 days (5∙2 months; 3∙8–not calculable [NC]) in the placebo group; hazards were non-proportional (HR for death with selumetinib, 0∙80, 80% CI 0∙56–1∙14; one-sided p=0∙21; fi gure 2A). Overall survival was much the same when analysed with all patients, irrespective of con fi rmation of KRAS-mutation status (HR for death with selumetinib 0∙81, 80% CI 0∙58–1∙15; one-sided p=0∙22).35 (81%) patients in the selumetinib group and36 (90%) in the placebo group had progressed or died at the time of data cutoff .Median progression-free survival was 162 days (5∙3 months, 95% CI 4∙6–6∙4) inthe selu m etinib group and 63 days (2∙1 months, 1∙4–3∙7) in the placebo group (HR for progression with selumetinib 0∙58, 80% C I 0∙42–0∙79; one-sided p=0∙014, fi gure 2B). Most patients had progression of existing lesions (data not shown). 6-month progression-free survival was 37·1% (95%CI 22·3–51·9)in the selumetinib group and 15·8% (95%CI 4·2–27·4)in the placebo group (HR 0·54, 80% CI0·37–0·78; one-sided p=0·016).No patient had a complete response. 16 patients in the selumetinib group had a partial response (11 confi rmed, fi ve unconfirmed); thus, 16 (37%) of patients had an objective response (table 5). By contrast, no patient in the placebo group had an objective response (two-sided p<0∙0001, table 5). One of the two non-evaluable patients in the placebo group was deemed non-assessable because of unmeasurable non-target lesions. However, according to RECIST 1.1 criteria, this patientwould have had a partial response. Median duration ofresponse in the selu m etinib group was 182 days(6·0 months, 95% C I 3·8–non-calculable). Meandecrease in tumour size is shown in table 6 and awaterfall plot of change in tumour size at week 12 isshown in fi gure 3. Post-study use of anticancer therapywas much the same in the selumetinib (44% [19 of 43])and placebo (55% [22 of 40]) groups. A breakdown ofpost-discontinuation treatments can be found in theappendix. No crossover was allowed.The proportion of serious adverse events was higherin the selumetinib group than in the placebo group(table 7). The most frequent serious adverse eventswere febrile neutropenia (selumetinib group 14% [sixof 44]; placebo group 0), pneumonia (selumetinib 9%[four of 44]; placebo 0), neutropenia (selumetinib 7%See Online for appendixFigure 3: Change in tumour size at 12 weeks(A) Selumetinib plus docetaxel group. (B) Placebo plus docetaxel group. Week 12 tumour size was defi ned as the assessment within 7 days (before or after) of the scheduled assessment, or imputed with linear regression for patients with follow-up data showing a decrease but no assessment in window or, for those with follow-up data showing an increase, we imputed a 20% increase when progression (including death) occurred before the visit.[three of 44]; placebo 7% [three of 42]), and respiratory failure (selumetinib 7% [three of 44]; placebo group 5% [two of 42]).The number of adverse events leading to death was low in both groups (selumetinib group: respiratory failure n=3, pneumonia n=1; placebo group: respiratory failure n=1, pneumonitis n=1, lower respiratory tract infection n=1, cardiac arrest n=1). None were deemed by the investigator to be related to study treatment.Grade 3–4 neutropenia, febrile neutropenia, and asthenia were more common in the selumetinib group than in the placebo group; by contrast, grade 3–4 dys-pnoea was more common in the placebo group than the selumetinib group (table 8). The most frequent adverse events of any grade in the selumetinib group were neutropenia, diarrhoea, nausea, vomiting, peripheral oedema, rashes, and stomatitis, all of which occurred more frequently in the selumetinib group than in the placebo group (table 8). Dyspnoea, exertional dyspnoea, musculoskeletal chest pain, arthralgia, fatigue, and myalgia all occurred more frequently in the placebo group than in the selumetinib group (table 8).Febrile neutropenia or neutropenic infection and pyr-exia were more common in the selumetinib group than in the placebo group. Febrile neutropenia or neu t ropenic infections, together with respiratory infections, respiratory disorders, and other isolated adverse events contributed to the greater incidence of admission to hospital of patients in the selumetinib plus docetaxel group (table 7). Selumetinib seemed to increase the severity of neutropenia associated with docetaxel—18 (42%) patients had grade 4 reduction in neutrophils from baseline grade 0 in the selumetinib group, compared with ten (24%) of those in the placebo group; this outcome led to increased use of granulocyte-colony stimulating factor in the selumetinib plus docetaxel group (32% [14 of 44] vs 17% [seven of 42]). Discontinuation of docetaxel because of adverse events was much the same between groups (table 7).Overall compliance for completion of the LSSQ was 86% (37 of 43) for the selumetinib group and 85% (34 of 40) for the placebo group. I n post-hoc exploratory analyses, the proportion of patients who had an improvement in LCS score was greater for the selumetinib group (44%) than the placebo group (24%; odds ratio 2∙50; 80% CI 1∙34–4∙77; one-sided p=0∙029). Median time-to-deterioration of LCS score was 186 days (6∙1 months, 95% C I 2∙8–non-calculable) in the selumetinib group and 49 days (1∙6 months, 1∙4–3∙5) in the placebo group (HR for deterioration 0∙33, 80% CI 0∙22–0∙49; one-sided p=0∙0002; fi gure 4). DiscussionIn this phase 2 trial, we found that although the addition of the MEK1/MEK2 inhibitor selumetinib to docetaxel as second-line treatment did not signifi cantly improve overall survival for patients with KRAS-mutant NSCLC, the combination resulted in clinically meaningful and statistically signifi cant improvements in progression-free survival, responses, and patient-reported outcomes.To our knowledge, this is the first prospective study to assess patients with KRAS-mutant NSCLC (panel). Retrospective studies suggest that patients with NSCLC with KRAS-mutant tumours obtain less clinical benefi t from chemotherapy compared with the general NSCLC population.9,25 Our prospective study seems to support this notion, because patients who received docetaxel alone had fewer responses and reduced progression-free and overall survival compared with previous phase 3 trials of second-line docetaxel in a general NSCLC population.26,27 The mechanisms for the particularly poor effi cacy of docetaxel in KRAS-mutant NSCLC are unknown and warrant further investigation. Notably, overall survival and the proportion of patients with an objective response in the control group of our study were much the same as those in a retrospective subgroup analysis of patients with NSCLC with KRAS-mutant tumours receiving docetaxel (n=27).28The improved outcomes with selumetinib plus docetaxel were associated with an increase in adverse events and serious adverse events relative to placebo and docetaxel. The most common safety findings were as expected on the basis of monotherapy tolerability profi les of selumetinib and docetaxel.16,26,27 We noted a higher occurrence of adverse events including diarrhoea, vomiting, stomatitis, and dry skin with selumetinib plus docetaxel. Furthermore, the combination was also asso-ciated with greater neutropenic effects than would be expected with docetaxel alone. The increased toxic eff ects of the combination are thus a potential issue, but we anticipate this problem will be reduced with appropriateguidance and symptomatic management, such as the use of growth factor support to decrease the incidence of neutropenia and, subsequently, the occurrence of admission to hospital and infections. Although we recorded more admissions to hospital and dose alterations (interruption and reduction) with the combination of selumetinib and docetaxel than with placebo and docetaxel, these events did not result in a greater number of discontinuations because of adverse events or deaths.n addition to increased p roportions of responses and improved progression-free survival, we also noted a clinically meaningful improvement in disease-related symptoms. In post-hoc analyses, more patients had im-provements in lung cancer symptoms with selumetinib plus docetaxel than with placebo and docetaxel, and occurrence of dyspnoea (all grades) was lower in the selumetinib group. These benefi ts might be attributable to the cytoreductive eff ects of the treatment.The combination of the two drugs in our study seems to have a synergistic eff ect compared with monotherapy activity.16,27 This fi nding is consistent with preclinical data for selumetinib,13 but contrasts with most previous studies in NSCLC, in which addition of a targeted agent to chemotherapy has not resulted in improved effi cacy. For example, studies of vandetanib plus chemotherapy as second-line treatment in unselected patients showed little or no benefi t.29 Additional studies are therefore needed to understand why the combination studied here displays synergistic activity and whether this fi nding will also be recorded with other agents targeting MAPK signalling or other chemotherapies.Additional factors warrant further investigation to define a subpopulation of KRAS-mutant NSCLC in which the combination of selumetinib and docetaxel leads to improved effi cacy. Diff erences in effi cacy might arise depending on the type of KRAS mutation present.I n our study most mutations were found in codon 12 (table 3) but the sample size is too small to determine whether any of the specific codon 12 mutations or codon 13 mutations affect the effi cacy of selumetinib plus docetaxel. The therapeutic eff ect of specifi c KRAS mutations should be assessed in future clinical trials. Other factors, such as concurrent tumour suppressor loss, including P53 and LKB1,17 and dosing schedule, could also modulate the activity of the combination, thus, additional dosing and scheduling might need to be assessed. I n-vivo mechanistic drug sequencing studies have shown that administration of selumetinib after docetaxel, rather than before, induced more apoptosis.18 This fi nding could have important clinical implications for any dosing schedule of this combination.Potential limitations of our study include the small sample size and absence of independent confi rmation of progression-free survival and tumour response. The study design was felt to be appropriate for a phase 2 proof-of-concept trial and was suffi ciently powered toFigure 4: Kaplan-Meier estimate of time-to-deterioration of the lung cancer subscale Symbols show censored observations.。