Second‐line chemotherapy for locally advanced or metastatic small cell lung cancer
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The objective of this systematic review is to explore the benefits of second‐line chemotherapy in patients with resistant (a relapse within three months of first treatment) or relapsed (a relapse after three months of first treatment) SCLC.
Background
Chemotherapy is an established first‐line treatment for small cell lung cancer (SCLC), based on outcomes of trials undertaken 30 years ago. SCLC accounts for up to one in five patients that present with a lung neoplasm. The prognosis of this disease is poor in part due to the aggressive nature of this cancer type. Patients are divided into limited and extensive stage disease, with median survival of 8‐10 months in those with extensive disease (Hanna 2006; Mavroudis 2001; Schiller 2001). Current therapy in the United Kingdom normally consists of a doublet of a platinum‐based compound, with etoposide as recommended by the National Institute for Health and Clinical Excellence (NICE 2005). Survival benefits have been demonstrated compared with placebo and have been subjected to a prior Cochrane review (Pelayo 2009). First‐line treatment rarely results in a cure, because only a minority of patients (10%) exhibit a complete response to treatment. Two‐thirds of patients have a partial response, and one‐fifth have no response. As a result of this, the choice of second‐line therapy requires exploration.
Objectives
The objective of this systematic review is to explore the benefits of second‐line chemotherapy in patients with resistant (a relapse within three months of first treatment) or relapsed (a relapse after three months of first treatment) SCLC.
Methods
Criteria for considering studies for this review
Types of studies
We will consider published or unpublished randomised controlled or quasi‐randomised controlled trials regardless of blinding or language of publication. We will exclude studies that do not include full data on the primary treatment and/or data on patient follow up. We will exclude duplicate publications of the same data. We will include cross‐over trials where patients have been randomised into the first period.
Types of participants
We will include all patients with a histological diagnosis of SCLC (limited and extensive stage) who have completed one previous course of combination (at least two agents) chemotherapy.
Types of interventions
Any second‐line chemotherapy (any single agent or combination of different drug treatments, at any dosage or number of cycles). Please note that, in order to avoid overlap with the updated review 'Chemotherapy versus best supportive care for extensive SCLC' (Pelayo 2009), we will exclude comparisons of second‐line chemotherapy versus best supportive care or placebo in patients with extensive disease (relapsed or progressive).
Types of outcome measures
Main outcome measure
1. Survival
Additional outcome measures
1. Quality of life scores
2. Toxic events
3. Disease remission time
4. Response rate
Search methods for identification of studies
We will search the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, latest issue); MEDLINE (1950 to date); EMBASE (1950 to date); CINAHL; ISI Web of Science and LILACS; all via OVID where available, using the search strategy detailed below.
1 exp Carcinoma, Small Cell/
2 lung.ti,ab.
3 pulmonary.ti,ab.
4 bronchogenic.ti,ab.
5 or/1‐4
6 small cell lung.ti,ab.
7 oat cell.ti,ab.
8 round cell.ti,ab.
9 6 or 7 or 8
10 cancer.ti,ab.
11 carcinoma$.ti,ab.
12 malignan$.ti,ab.
13 tumor$.mp. or tumour$.ti,ab. [mp=title, original title, abstract, name of substance word, subject heading word]
14 (tumor$ or tumour$).ti,ab.
15 neoplasm$.ti,ab.
16 exp Pneumonectomy/
17 Pneumonectom$.ti,ab.
18 or/10‐17
19 exp Antineoplastic Combined Chemotherapy Protocols/
20 exp Antineoplastic Agents/
21 (second‐line or second line).ti,ab.
22 (((rescue adj3 chemotherapy) or failed) adj3 chemotherapy).ti,ab.
23 ((relaps$ or recurrent or progressive or resistant) adj3 small cell lung cancer).ti,ab.
24 19 or 20 or 21 or 22 or 23
25 5 and 9 and 18 and 24
26 (nonsmall cell lung or non‐small cell lung).ti,ab.
27 6 not 26
28 25 and 27
We will also search databases for ongoing trials, for example http://clinicaltrials.gov and http://controlled‐trials.com, using a similar search strategy. We will apply a trials filter to the search strategy following the style of box 6.4.c ("Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity‐maximizing version (2008 revision); Ovid format") of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008).
We will not impose a language restriction.
Searching other resources
We will search for additional studies from:
1. the reference lists of retrieved trial materials and review articles;
2. conference abstracts and proceedings.
We will contact experts and those with an interest in this field.
Data collection and analysis
Selection of studies
Two authors (AD and YEO) will independently screen the title and abstract of publications identified by the searches, being "overly inclusive" if there is doubt as prescribed by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). We will not be blinded to the names of the authors, institutions or journal of publication. Where there is disagreement between AD and YEO regarding the suitability for inclusion or exclusion, on the basis of methodological issues, we will consult a third party (SP) for additional guidance. We will include reasons for any notable exclusions. Where necessary, we will request any additional clarification from trialists and add these trials to a list "awaiting assessment" in the interim.
Data Extraction and Management
We will obtain data from the trials shortlisted using a standard extraction form. Using this form, we will record data concerning participant characteristics, intervention details and outcome measures. For outcome measures, we will focus principally on survival and, as stated above in the Types of outcome measures section above, we will also consider Quality of Life (QoL) scores, toxic events, disease remission time and response rate.
Assessment of risk of bias in included studies
Two authors (AD and YEO) will independently assess the risk of bias of the shortlisted studies, using the Cochrane recommended tool (Higgins 2008).
We will use the following framework for the assessment:
Sequence generation adequate?
YES: Adequately generated allocation sequence
UNCLEAR: Not reported in enough detail to make an assessment, and cannot be verified by contacting study authors
NO: Not adequate or not used
Allocation concealment?
YES: Allocation adequately concealed
UNCLEAR: Not reported in enough detail to make an assessment, and cannot be verified by contacting study authors
NO: Not adequately concealed
Blinding of participants, personnel and outcome assessors adequate?
YES: Knowledge of allocation of intervention sufficiently prevented
UNCLEAR: Not reported in enough detail to make an assessment, and cannot be verified by contacting study authors
NO: Knowledge of allocation of intervention not sufficiently prevented
Incomplete outcome data adequately addressed?
YES: Incomplete outcome data were adequately addressed (attritions/exclusions (and reasons) reported; numbers of participants of each intervention comparable to total randomised participants. We will note here whether review authors saw fit to re‐include any exclusions.)
UNCLEAR: Not reported in enough detail to make an assessment, and cannot be verified by contacting study authors
NO: Not adequately addressed
Where there are disagreements on the assessments, we will consult SP for additional guidance.
Measures of treatment effect
These are according to categories described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). We will use the following effect measures in the case of each different data type (associated outcome measure stated).
Types of outcome measures
Main outcome measure
1. Mortality
Additional outcome measures
1. QoL scores
2. Toxic events
3. Disease remission time
4. Response rate
Dichotomous data (mortality)
We intend to calculate odds ratios (OR). In the event that there are incomplete follow‐up data in either treatment or control group, we will consider hazard ratios.
Continuous data (QoL scores, toxic events)
We will analyse continuous outcomes where the mean and standard deviation are included in the trial reports or they can be calculated from available data. For the meta‐analysis of continuous outcomes, we will calculate the difference in means (or “mean difference”) between groups wherever possible. Where this is not possible, we will attempt to analyse the data using standardised mean differences (SMD). We will undertake analysis of continuous data using parametric tests for more normally distributed data and non‐parametric tests for skewed data. If count data do not appear normal, we will conduct a Poisson regression analysis. We will treat disease remission time as survival data where possible; however, if trial reporting precludes this, we will treat disease remission time as continuous data (Parmar 1998).
Unit of analysis issues
In the event that cross‐over trials are identified, we will use the methods recommended by Elbourne 2002 to incorporate these into our meta‐analysis.
Dealing with missing data
Where limitations of missing data in trial reports are identified, we will contact trial authors to request the necessary information. Where information remains unavailable, we will not statistically aggregate the trial, and will report outcomes narratively alongside the meta‐analyses where relevant. We will agree upon a minimum of events upon in deriving summary statistics for trials that do not have the number of events reported (Parmar 1998).
Assessment of heterogeneity
We will consider an I2 value of more than 50% existence of statistical heterogeneity. We will calculate the I2 value using The Cochrane Collaboration's Review Manager 5 ( RevMan 2008 ) software.
If results for endpoints of interest are sufficiently homogenous, we will undertake meta‐analysis. The quality of the studies will be the main source of homogeneity.
Data synthesis
We will combine data using RevMan 5 ( RevMan 2008 ) software in accordance with guidance provided by the Cochrane Handbook and the editorial board of the Cochrane Lung Cancer Review Group
Subgroup analysis and investigation of heterogeneity
We plan to consider whether observed treatment effects are consistent across a priori defined patient subgroups, such as chemotherapy received and disease stage.
The subgroups will be as follows.
Chemotherapy received (platinum‐based therapy, topoisomerase inhibitors, others)
Stage (limited or extensive)
If there is substantial heterogeneity and differences of effect, we will carry out subgroup analyses across the categories identified above to examine whether there is any evidence that particular subgroups of patients benefit more or less from chemotherapy.
Sensitivity analyses
We will re‐calculate analyses where heterogeneity is identified to exclude any trials that are clear outliers; we will present the results of these analyses alongside the original analysis where heterogeneity was identified for the purpose of comparison.
