Chemotherapy, radiotherapy, chemoradiotherapy and combination therapy in localised and locally advanced pancreatic cancer
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
A systematic review of the available evidence, both published and unpublished, on the management of localised and locally advanced pancreatic cancer is planned. The interventions to be evaluated are chemotherapy and radiotherapy, in the adjuvant (for localised pancreatic cancer), neoadjuvant (for localised and locally advanced pancreatic cancer) and palliative settings (for locally advanced pancreatic cancer). We will study chemotherapy administered either as single agent or as combination therapy, including with novel agents. We will assess radiotherapy, given either alone or combined with chemotherapy as chemoradiation, chemoradiation followed on by chemotherapy or chemotherapy with follow‐on chemoradiation.
Background
Description of the condition
Pancreatic cancer is the 10th most common cancer in the UK and USA in terms of incidence (Jemal 2010; ONS 2010), but is among the fourth or fifth leading causes of cancer death (Jemal 2010; ONS 2010b). The only treatment with potential for cure is resection (surgical removal), but even in specialised centres just 10% to 15% of diagnosed patients have resectable disease (Stathis 2010). Of the remainder, 30% to 35% have locally advanced disease and 50% have metastatic disease (Stathis 2010).
Description of the intervention
In patients with resectable disease, adjuvant chemotherapy has improved overall survival (Neoptolemos 2010) or disease‐free survival (Oettle 2007), and more than doubled the five‐year survival rates from 10% to nearly 25% (Van Laethem 2012).
The majority of patients will require palliative treatment (chemotherapy or chemoradiotherapy with or without follow‐on chemotherapy), with a median survival of 6 to 11 months in patients with locally advanced pancreatic cancer and two to six months in those with metastatic disease (Stathis 2010). The role of chemotherapy in advanced pancreatic cancer, and in particular for gemcitabine‐based combinations over single‐agent gemcitabine (overall survival advantage), has been established both in randomised controlled trials and subsequent aggregate data meta‐analyses of these (Sultana 2007).
Radiation therapy (in conjunction with chemotherapy) in both the adjuvant and advanced disease scenarios remains controversial (Johung 2012; Neoptolemos 2011; Sultana 2007b).
How the intervention might work
Pancreatectomy with standard lymphadenectomy is advocated for resectable disease, but patients tend to succumb to recurrence either loco‐regionally or in the liver (Abrams 2001; Hishinuma 2006; Koshy 2005; Sperti 1997). Adjuvant treatment following curative resection acts by targeting micro‐metastatic disease (Chua 2005), thereby improving outcomes. The attractiveness of neoadjuvant therapy lies in the fact that nearly 20% to 30% of resected patients fail to receive adjuvant therapy on the grounds of delayed recovery from major surgery, co‐morbidities, patient choice and early recurrence (Yeo 1995). Palliative chemotherapy provides an overall survival benefit and improves symptom control/quality of life compared to best supportive care (Burris 1997; Sultana 2007; Yip 2006). Chemoradiotherapy/combination therapy has been administered in locally advanced pancreatic cancer patients with the intention of down‐staging the disease.
Why it is important to do this review
In an individual patient data (IPD) meta‐analysis, adjuvant chemotherapy resulted in a 25% reduction in the risk of death (hazard ratio (HR) 0.75, 95% confidence interval (CI) 0.64 to 0.90) compared to no chemotherapy. In contrast, there was no significant difference between chemoradiation versus no chemoradiation (HR 1.09, 95% CI 0.89 to 1.32) (Stocken 2005).
A subsequent aggregate data meta‐analysis of adjuvant treatment concluded that chemotherapy improved median survival by three months (95% CI 0.3 to 5.7 months; P = 0.03), but did not impact on five‐year survival rates, possibly due to the low numbers at risk at this time point (Boeck 2007). It included two further randomised controlled trials (RCTs) on chemotherapy versus best supportive care (Kosuge 2006; Oettle 2007), compared to the previously published individual patient data meta‐analyses. A criticism of this study was the methodology of the meta‐analysis, utilising median survival and rates at different time points. These have been shown to not be the ideal surrogate measures for meta‐analyses of survival data (Michiels 2005).
Aggregate data meta‐analyses of randomised controlled trials of chemotherapy (Heinemann 2008; Hu 2011; Sultana 2007; Xie 2010; Yip 2006) in advanced pancreatic cancer patients have suggested a survival advantage for gemcitabine combination chemotherapy. The role of chemoradiotherapy, with or without follow‐on chemotherapy, is controversial in the management of advanced pancreatic cancer. The only level 1a evidence in this area found a significant survival benefit for chemoradiotherapy over radiotherapy, but not for chemoradiotherapy with follow‐on chemotherapy compared to chemotherapy alone (Sultana 2007b). However, the heterogeneity in the trials included, combined with the fact that both staging and radiotherapy techniques had progressed since the inception of the majority of the trials, limited the conclusions that could be drawn.
The role of neoadjuvant/preoperative therapy in pancreatic cancer is less clear cut. Most studies have been retrospective or prospective cohort/phase I/II trials. A comprehensive systematic review by Gillen et al of 111 prospective (n = 78) (including phase I/II studies) and retrospective (n = 33) studies found that the median survival for resectable patients who received neoadjuvant chemotherapy and went on to have a resection was 23.3 months (95% CI 12 to 54 months) (Gillen 2010). Another meta‐analysis evaluated gemcitabine given alone or in combination with radiotherapy prior to surgery in localised and locally advanced pancreatic cancer (Andriulli 2011). In the 707 patients analysed, they found little support for neoadjuvant therapy in resectable patients, and only marginal benefit in locally advanced disease (39% were resectable on restaging, of whom 68% could undergo resection). A limitation of this study was the inability to assess survival data. Another meta‐analysis of both prospective and retrospective cohort studies found a higher risk of perioperative death in patients receiving preoperative chemoradiotherapy, though there was no difference in overall complications (Laurence 2011). Meta‐analysis of prospective phase II studies recommended preoperative treatment only in the context of locally advanced disease (Assifi 2011).
Updated aggregate data meta‐analyses incorporating recent and ongoing RCTs, and using the most appropriate methods for survival analyses, are warranted.
Objectives
A systematic review of the available evidence, both published and unpublished, on the management of localised and locally advanced pancreatic cancer is planned. The interventions to be evaluated are chemotherapy and radiotherapy, in the adjuvant (for localised pancreatic cancer), neoadjuvant (for localised and locally advanced pancreatic cancer) and palliative settings (for locally advanced pancreatic cancer). We will study chemotherapy administered either as single agent or as combination therapy, including with novel agents. We will assess radiotherapy, given either alone or combined with chemotherapy as chemoradiation, chemoradiation followed on by chemotherapy or chemotherapy with follow‐on chemoradiation.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials.
Types of participants
Patients with localised or locally advanced pancreatic cancer.
Types of interventions
We will assess chemotherapy (systemic/regional), radiotherapy, chemoradiotherapy, chemoradiotherapy with follow‐on chemotherapy, chemotherapy with follow‐on chemoradiotherapy, neoadjuvant therapy and novel agents.
Types of outcome measures
Primary outcomes
Overall survival.
Secondary outcomes
Progression‐free survival or time to progression, response rate, adverse events and quality of life.
Search methods for identification of studies
Electronic searches
We will identify trials by searching the following electronic databases:
-
Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library) (Appendix 1);
-
MEDLINE (Appendix 2);
-
EMBASE (Appendix 3);
-
ISI Web of Science;
-
CancerLit;
-
Current Contents.
We will search the databases (details in appendices) using the terms: explode 'pancreas neoplasms' or 'pancreatic neoplasms' or 'pancreas tumour' or 'pancreatic tumour' or 'pancreas cancer' or 'pancreatic cancer' or 'adenocarcinoma pancreas' and explode 'chemotherapy' or 'treatment' and 'randomised controlled trials' (publication type). We will search all databases from their inception.
Searching other resources
We will scan registries of the National Cancer Institute Physician Data Query, the UK Co‐ordinating Committee on Cancer Research and the National Clinical Trials Registry for ongoing studies. We will screen the bibliographies of the selected trials and review articles and we will contact experts in the field.
We will explore proceedings/reports of relevant meetings and conferences (American Society of Clinical Oncology, American Association of Cancer Research, the European Cancer Conference, European Society of Medical Oncology, American Gastroenterological Association, European Pancreatic Club, American Association of Pancreatology, British Society of Gastroenterology and the United European Gastroenterology Week) (last five years).
Data collection and analysis
Selection of studies
Two independent review authors (AS, PG) will scan the abstract of every trial identified by the search to determine eligibility using an evaluation form. Authors will not be blind to source. We will then select full articles for further assessment if the abstract suggests the study includes patients with localised/locally advanced pancreatic carcinoma and deals with the interventions outlined before. If these criteria are unclear from the abstract, we will retrieve the full article for clarification. We will exclude papers not meeting the inclusion criteria. We will resolve any disagreements by discussion. We will keep a record of all included and excluded trials, with explanation for the latter.
The inclusion criteria will be RCTs dealing with patients with localised/locally advanced pancreatic ductal adenocarcinoma. Non‐randomised or retrospective studies, trials dealing with surgical resection of tumours and cancers other than pancreas cancers will be excluded.
Data extraction and management
Two independent review authors (AS, RJ) will extract raw data from the publications and resolve any disagreements by discussion. We will extract the following characteristics from each RCT: primary author; number of patients randomised; patient inclusion and exclusion criteria; methodological quality; sample size estimation; intention‐to‐treat analyses; intervention regimens; mean/median age; sex; stage and site of disease; surgery performed; resection margin status; lymph node involvement; CA19‐9 level; previous treatment and its efficacy; median follow‐up; response to treatment; overall survival; progression‐free survival; adverse events and quality of life. We will enter the characteristics and outcomes of the included trials and details of the excluded trials into RevMan 2011.
Assessment of risk of bias in included studies
We will use the Cochrane 'Risk of bias' assessment tool as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Measures of treatment effect
We will summarise dichotomous data (response rate/adverse events) using risk ratios and 95% confidence intervals pooled using the Mantel‐Haenszel method for combining trials (Deeks 2001). We will summarise time‐to‐event data for overall survival and progression‐free survival or time to progression in terms of log hazard ratios and variances. We will summarise quality of life assessments using the standardised mean difference.
Primary outcome
Overall survival will be defined as time from randomisation to death. We will also include alternative definitions such as time from initiation of therapy/surgery to death and note this as a potential source of heterogeneity.
Secondary outcomes
-
Progression‐free survival or time to progression will be defined as time from randomisation until disease progression. We will also note alternative definitions such as time from initiation of treatment and the criteria used to define 'progression' (i.e. imaging guidelines such as Response Evaluation Criteria In Solid Tumors (RECIST)/World Health Organization (WHO) or clinical criteria) and consider these as potential sources of heterogeneity.
-
Response rate will be defined as overall response and, where available, we will note details of partial or complete response, stable disease or progression. We will record the criteria used to define response rate such as RECIST or WHO criteria and these could be potential sources of heterogeneity.
-
Grade 3/4 adverse events will be defined as side effects occurring during the study, from the date of randomisation until either the end of the study or death. We will note the criteria used to define adverse events, such as different versions (one or two) of the common toxicity criteria or alternatively the WHO criteria.
-
Quality of life will be as defined by validated questionnaire, with note being made of the type of questionnaire used (Hospital Anxiety Depression Score, European Organization for Research and Treatment of Cancer quality of life questionnaire).
Unit of analysis issues
We do not anticipate unit of analysis issues, as we will only be including randomised controlled trials.
Dealing with missing data
We will contact authors of the primary trial reports when necessary to clarify data and to provide missing information. To minimise the potential for within‐study selective reporting bias, we will include trials wherein the primary outcome measure (overall survival) is not reported if all other inclusion criteria are met. We will contact the authors of these trials and ask them to supply data for the primary outcome measure, if collected.
Assessment of heterogeneity
We will assess heterogeneity by visual inspection of the forest plot, using the Chi² test and by interpretation of the I² statistic. We will carry out sensitivity analyses using the common chemotherapy agent arm across studies as a means to assess between‐trial variability.
We will record data to assess sources of methodological and clinical heterogeneity. We anticipate three sources of heterogeneity and will assess these as follows:
-
We will describe the quality of selected studies for the different components assessed in each trial.
-
We will assess clinical heterogeneity by qualitatively examining the variability across trials in terms of participant and intervention characteristics.
-
We will examine the characteristics of the sample in relation to suspected effect modifiers, namely the stage of the disease, duration of disease, preceding surgery, previous treatment, recurrent disease and type of chemotherapy (agent used, dosage and route of administration).
Assessment of reporting biases
We will use contour‐enhanced funnel plots to assess publication bias (Sterne 2011). We will also give consideration to within‐study selective reporting bias (Williamson 2005).
Data synthesis
We plan to conduct aggregate data meta‐analyses.
Subgroup analysis and investigation of heterogeneity
A priori subgroup analysis is not planned. Once the results of each study have been summarised using an effect measure, we will compute an average value of the effect across studies using a fixed‐effect model. In addition, we will also conduct a random‐effects meta‐analysis as a sensitivity analysis.
Sensitivity analysis
Sensitivity analysis based on the degree of risk of bias is planned. We will assess the influence of each study by systematically removing each one from the analysis and evaluating the effect on the result.
