Taxanes for advanced non‐small cell lung cancer
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To assess the effects of taxanes as part of a combined or single‐agent therapy versus other agents or best supportive care as first‐ or second‐line treatment for advanced non‐small cell lung cancer (NSCLC).
A secondary objective is to assess different modes or schemes of administration of taxanes in patients with this disease.
Background
Description of the condition
Lung cancer is a major public health concern, with high incidence and mortality. At around two million new cases each year, it is the second most common cancer in men as well as women and by far the leading cause of cancer death in both (Ferlay 2015).
Approximately 85% to 90% of all lung cancers are diagnosed as non‐small cell lung cancer (NSCLC) (Jemal 2011), which consists mainly of adenocarcinoma (40%), squamous cell carcinoma (25% to 30%), and large‐cell (undifferentiated) carcinoma (10% to 15%) (Meza 2015; Riaz 2012). At the time of diagnosis, nearly 70% of all NSCLCs have spread either locally or to distant regions of the body (Meza 2015), which is referred to as advanced NSCLC (stage IIIB‐IV) (Sobin 2009).
The five‐year survival rate for stage IIIB NSCLC is about 5%, while metastatic, or stage IV NSCLC, has a 5‐year survival rate of about 1% (Siegel 2017). While surgery to remove the cancer is the best option in early‐stage NSCLC, systemic treatment is needed in advanced disease. There are currently several options available for people with advanced stage lung cancer whose goal is to prolong survival and improve quality of life. These are based on chemotherapy, radiotherapy, and most recently immunotherapy (Carbognin 2015).
For patients with stage IIIB NSCLC, concurrent chemoradiotherapy is the treatment of choice. If this treatment is not possible, sequential approaches of induction chemotherapy followed by definitive radiotherapy represent a valid and effective alternative (Postmus 2017). Platinum based therapy is recommended, and taxanes have been tested in this setting.
For patients with stage IV NSCLC, the treatment strategy should take into account factors like histology, molecular pathology, age, performance status (PS), co‐morbidities and patient preferences (Novello 2016). When genetic alterations are present (epidermal growth factor receptor (EGFR) mutations and anaplastic lymphoma kinase (ALK) rearrangements, which account for only about 10% and 5% of these patients, respectively), directed systemic therapy with tyrosine kinase inhibitors is recommended (Attatrian 2017). For most stage IV NSCLC patients with EGFR‐ and ALK‐negative disease, chemotherapy with platinum doublets (that is, a platinum‐based regimen with the addition of a third‐generation cytotoxic, which could consist of a taxane) is the recommended first‐line treatment option if PS is 0 to 2. In people with programmed death ligand 1 (PD‐L1) expression on at least 50% of cells, immunotherapy is the treatment of choice (Reck 2016).
Patients with a poor PS (3,4) should be offered best supportive care (BSC) in the absence of documented EGFR mutations or ALK rearrangements. There is concern about the benefits of using cytotoxic treatment rather than BSC alone in this setting.
Second‐line therapy could be offered to patients with PS 0 to 2 and clinical or radiological progress after first‐line chemotherapy. In this context, combination chemotherapy regimens do not seem to provide a substantial benefit over single‐agent treatments. Taxanes are also an option for these patients. Immune check‐point‐inhibitor antibodies are a new class of therapy, useful in subjects with PD‐L1 expression in tumour cells (Ellis 2017).
Description of the intervention
Taxanes have been used either as part of a platinum‐based regimen (chemotherapy with platinum doublets) or alone in stage IIIB and stage IV NSCLC patients.
Taxanes are antimicrotubule cytotoxic agents, first isolated in 1971 from the bark of the Pacific yew tree (Taxus brevifolia), and they have been extensively used in several tumour sites such as the breast and lung (NSCLC), among others.
Paclitaxel and docetaxel are the most widely available taxanes; pharmacologically, their volume of distribution is large with a rapid uptake by tissues (except the central nervous system) and therefore they have long half‐lives of elimination. Their metabolism is mainly hepatic, and there are therefore drug interactions with p450 inducers and inhibitors. Given that the time of treatment above the threshold concentration is the most important factor determining cytotoxicity, several studies have investigated the best administration schedules. Concerns about the safety of taxanes have always been a topic of debate around their use. Some adverse reactions are similar among all taxanes and include haematological (neutropenia, thrombocytopenia) and gastrointestinal (nausea and vomiting, diarrhoea) effects, as well as hair loss and fatigue. Other reactions are agent‐specific, such as hypersensitivity and neurotoxicity with paclitaxel, and fluid retention, nail toxicity and neurotoxicity with docetaxel.
In order to improve efficacy and reduce toxicity, new taxanes have been developed. The nanoparticle albumin‐bound form of paclitaxel (nab‐paclitaxel) is the most studied. It has improved anti‐tumour activity and has a better global toxicity profile (Gradishar 2006; Hirsh 2014)
How the intervention might work
Taxanes are considered third‐generation chemotherapy agents and have been used in the treatment of advanced NSCLC. Recent guidelines have recommended taxanes in specific clinical scenarios, while in other situations their use remains unclear (ESMO 2015; ESMO 2016).
Taxanes act by improving microtubule assembly and stabilisation, preventing them from breaking down after the cell division, and blocking further division of cancer cells (Rowinsky 1990;Rowinsky 1992).
Why it is important to do this review
Taxanes are a key component of modern chemotherapy regimens for treating patients with advanced NSCLC, usually in combination with a platinum agent, which is the cornerstone therapy for this condition. The specific situations where taxanes have been used as well as the modalities of their administration have been very diverse and, therefore, the possible alternatives are also very numerous. For this reason, there are still some uncertainties about the effectiveness of taxanes compared with other drugs (as a component of a platinum doublet) or used as a single agent, or compared with best supportive care for the most advanced disease. Therefore, a Cochrane Review will provide the necessary evidence to clarify the role of taxanes in the therapeutic management of NSCLC patients.
Objectives
To assess the effects of taxanes as part of a combined or single‐agent therapy versus other agents or best supportive care as first‐ or second‐line treatment for advanced non‐small cell lung cancer (NSCLC).
A secondary objective is to assess different modes or schemes of administration of taxanes in patients with this disease.
Methods
Criteria for considering studies for this review
Types of studies
Published and unpublished randomised controlled trials (RCTs).
Types of participants
People with pathologically confirmed NSCLC, unresectable or metastatic disease (stage IIIB or IV), wild‐type or unknown status for EGFR, and wild‐type or unknown status for ALK.
Types of interventions
We will consider any RCT assessing a systemic treatment containing a taxane used either as first‐ or second‐line therapy for people with advanced NSCLC. Taxanes could be used as a monotherapy or combined with other agents.
We will consider the following comparisons according to the stage and the modality of administration.
-
Stage IIIB.
-
Chemoradiotherapy using a platinum‐based chemotherapy including a taxane versus a platinum‐based chemotherapy without a taxane.
-
-
For stage IV.
-
Platinum‐based chemotherapy plus a taxane versus platinum‐based chemotherapy plus another agent (P + T versus P + O) as first‐line treatment.
-
Platinum‐based chemotherapy regimens containing different taxanes (P + T1 versus P + T2), as first‐line treatment.
-
Platinum‐based chemotherapy regimens containing different schedules of the same taxane (P + T1a versus P + T1b), as first‐line treatment.
-
Chemotherapy regimens where a taxane is combined with a different agent (T + O1 versus T + O2), as first‐line treatment.
-
Combined therapy versus single agent chemotherapy (O + T versus O; O1 + O2 versus T), as first‐line treatment.
-
Single agent chemotherapy with a taxane versus other agent (T versus O), as second‐line treatment.
-
Taxanes versus best supportive care (T versus BSC), as second‐line treatment.
-
Note: we will also include studies where treatment consists of a targeted therapy in EGFR‐NSCLC.
Note: we will exclude any comparisons with immunotherapy.
Types of outcome measures
Primary outcomes
-
Overall survival (OS), defined as the time from randomisation until death from any cause. We will consider censoring living patients at their date of the last follow‐up for the OS outcome.
-
One‐year survival rate (1ySR), defined by the percentage of patients alive at one year from randomisation for first‐line comparisons.
-
Health‐related quality of life (HRQoL), measured by any validated tool.
-
Drug toxicities (according to National Cancer Institute Common Toxicity Criteria) (NCI Common Toxicity Criteria).
Secondary outcomes
-
Progression free survival (PFS) defined as the time from randomisation to disease progression or death from any cause. We will consider censoring living patients at their date of the last follow‐up for the PFS outcome.
-
Objective response rate (ORR), classified according to the Response Evaluation Criteria in Solid Tumors (RECIST) (Eisenhauer 2009).
Search methods for identification of studies
We will perform the search for trials in accordance with the Cochrane Lung Cancer Review Group recommendations with no limits regarding study publication date or language.
Electronic searches
We will search the following databases from inception until date of search.
-
Cochrane Lung Cancer Group Trial Register.
-
The Cochrane Central Register of Controlled Trials (CENTRAL) (latest issue).
-
MEDLINE, accessed via PubMed.
-
Embase.
-
LILACS.
-
Opengray.
The search strategies to be used for CENTRAL, MEDLINE and Embase are presented in Appendix 1.
We will search all databases using both controlled vocabulary (namely MeSH in MEDLINE and EMTREE in Embase) and a wide range of free‐text terms. We will perform the MEDLINE search using the Cochrane highly sensitive search strategy and precision‐maximising version (2008 version) as described in the Cochrane Handbook for Systematic Reviews of Interventions (Chapter 6.4.11.1 and detailed in box 6.4.b)) (Higgins 2011b)
Searching other resources
We will search for relevant clinical practice guidelines on NSCLC as well as other reviews on this condition and will then handsearch the list of included RCTs. We will consider guidelines published from 2015 onwards.
We will handsearch the proceedings of the American Society of Clinical Oncology (ASCO) and the European Society of Medical Oncology (ESMO) meetings. We will consider proceedings published from 1995 onwards.
We will also search Clinical Trials.gov and the European register for drug trials (EudraCT) from inception onwards.
We will also review Pharmacovigilance databases and reported adverse events in Open Vigil (OpenVigil 2.1‐MedDRA (data 2004Q1‐2017Q2), openvigil.sourceforge.net).
Data collection and analysis
The processes of study selection, data extraction, risk of bias assessment, and data analysis will follow standard Cochrane methods (Higgins 2011b). For the first three processes we will use Covidence software (www.covidence.org).
Selection of studies
We will independently examine all titles and abstracts of studies found in the search and select those potentially eligible using the aforementioned inclusion criteria for a more detailed assessment. We will make a final decision after obtaining a full‐text copy of these studies. We will resolve any discrepancies by consensus after discussion with a third review author (GU). We will perform title and abstract screening in duplicate using Covidence software (CL, MV, EM).
We will record reasons for exclusion and provide a study selection flow chart.
Data extraction and management
Two review authors (CL, MV) will independently extract and record data on a standardised data extraction form. We will resolve disagreements with the participation of a third review author (MB). We will include the following information from individual studies on the data extraction forms.
-
Publication details (source).
-
Eligibility criteria and reasons for exclusion.
-
Characteristics of participants (e.g. age, sex, performance status, histology, EGFR mutation status, stage, previous treatments).
-
Details of intervention (e.g. drug, dose, regimen, scheme, duration).
-
Outcome measures (primary and secondary), including the definition and time point.
-
Study design (methods and other relevant details) (e.g. setting, method of allocation, allocation concealment, blinding, follow‐up, type of analyses, e.g. intention‐to‐treat, modified intention‐to‐treat).
-
Risk of bias assessment.
-
Results (number of participants allocated in each group, and for each outcome of interest, missing participants, summary data for each group, estimate of effect with confidence interval and P value and subgroup analyses).
-
Funding.
For studies with redundant publications, we will extract data from the manuscript with the most updated version of the study.
We will perform data extraction and management in duplicate using Covidence software (CL, MV, MB).
Assessment of risk of bias in included studies
Two review authors (JPB, GU) will independently assess potential sources of bias in the included studies using the Cochrane Collaboration's 'Risk of bias' assessment tool (Higgins 2011a). We will assess risk of bias in included studies in duplicate using Covidence software (JPB, GU), resolving any differences in judgement through discussion with a third review author (EM).
We will assess the following domains: randomisation (sequence generation), allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other bias.
We will rate each domain as being at 'high', 'low', or 'unclear' risk of bias for each included study, using criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). We will report these judgements for each domain and study in the 'Risk of bias' table.
We will complete a 'Risk of bias' table for each included study and will summarise risks of bias across studies.
Measures of treatment effect
For each outcome, we will calculate summary estimates of treatment effect (with 95% confidence interval (CI)) for each comparison.
For dichotomous outcomes (1ySR, objective response rate and toxicity), we will use risk ratios (RRs).
For continuous outcomes (health‐related quality of life), we will present mean differences (MDs) of measures using the same scale, and standardised mean differences (SMDs) for measures using different scales.
For time‐to‐event endpoints (OS, PFS), we will report hazard ratios (HRs). Whenever possible, we will extract the HRs and associated variances directly from the publications. In case these data are unavailable, we will extract data according to the method described by Tierney 2007.
Unit of analysis issues
The primary unit of analysis will be the participant. We do not anticipate including trials with non‐standard designs, such as cross‐over trials and cluster‐randomised trials.
Dealing with missing data
In case the published data are incomplete or are provided in a format that is not useful for the meta‐analysis, we will contact the study authors in order to collect this information. If despite several attempts we cannot obtain the required data (e.g. standard deviations or correlation coefficients), we will attempt to estimate these from other available statistics such as P values according to the methods described in theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). We will extract data presented only in graphs (Kaplan‐Meier curves) using the GetData Graph Digitizer or similar software. Otherwise, when data are not amenable to meta‐analysis, we will present the results of these studies in a descriptive manner, discussing them in the context of the full set of results that are available.
We will carefully review the population of analysis of each primary study and use the intention‐to‐treat population in our analysis.
Assessment of heterogeneity
We will explore statistical heterogeneity with the I² statistic, considering 50% the threshold to define substantial heterogeneity and 80% the threshold to define extreme heterogeneity. In the latter case, we will not pool data in meta‐analysis.
If we identify sufficient trials, we will examine possible sources of heterogeneity by performing subgroup analyses.
Assessment of reporting biases
Where possible, we plan to explore selective reporting bias by checking consistency between the trial protocol (or the information provided in the trial registry) and the published article.
We will use funnel plots for testing publication bias, whenever the number of studies for the outcome and comparison is large enough (Egger 1997; Macaskill 2001). We will not conduct a funnel plot test for asymmetry when the number of studies in a meta‐analysis is less than 10.
Data synthesis
We will use Review Manager 5 (RevMan 5) software for data synthesis (RevMan 2014). We will combine treatment effect estimates of individual trials using fixed‐effect and random‐effects models. We will base the choice of the model on clinical and methodological diversity across trials, distribution of effect sizes and number and size of trials. We will estimate absolute differences in median OS and PFS by using the HR and the average control group median OS and PFS ((median/HR) − median), providing 95% confidence intervals.
Subgroup analysis and investigation of heterogeneity
Whenever possible, we will investigate the influence of the following variables on the results.
-
Sex.
-
Age (elderly versus non‐elderly patients).
-
Histologic type (squamous vs non‐squamous).
-
Performance status (0 to 1 versus 2 for stages IIIB and IV, and 0 to 2 versus 3 to 4 for stage IIIB).
Sensitivity analysis
If we identify sufficient trials, we will restrict the analysis to trials at low risk of bias. We will also perform a sensitivity analysis comparing academic and commercial trials.
Summary of findings table
We will generate 'Summary of findings' tables using the following outcomes: overall survival (OS), one‐year survival rate (1ySR), health‐related quality of life (HRQoL), drug toxicities (according to National Cancer Institute Common Toxicity Criteria) (NCI Common Toxicity Criteria), progression‐free survival (PFS) and objective response rate (ORR). We will use the five GRADE criteria for appraising the quality of the body of evidence (study limitations, consistency of effect, imprecision, indirectness and publication bias).
When data aggregation is not possible, we will present the results of individual studies in tables or graphics in a narrative manner and will discuss them.
