You are here
Exclusion of Gastrointestinal Cancer Patients With Prior Cancer From Clinical Trials: Is This Justified?
Clinical Colorectal Cancer, In Press, Corrected Proof, Available online 27 November 2015, Available online 27 November 2015
Strict eligibility criteria are necessary to maintain patient safety and scientific validity in clinical trials. However, this may lead to impaired generalizability of results. As survival in gastrointestinal (GI) cancer relates mainly to the GI malignancy, we hypothesized that previous cancers do not impact on survival and are not a rational exclusion criterion.
Materials and Methods
Patients treated with chemotherapy for a GI cancer in 2006 were identified from the electronic patient record at the Royal Marsden Hospital, London. Chart review was performed and patient age, gender, GI cancer stage, prior cancer stage, clinical trial availability/eligibility, and dates of cancer recurrence, death, and last follow-up were collated.
A total of 697 patients were identified. Fifty-four patients (8%) had a prior cancer; commonly breast (26%), prostate (17%), or colon (9%); most were stage I (42%) or II (37%). Two hundred ninety-seven (65%) patients had GI cancer recurrence, 7 (12%) patients had relapse of a prior cancer. Five hundred four (72%) patients have died, 170 (24%) are alive with no cancer, and 23 (3%) patients are alive with cancer. A total of 476 (94%) died of GI cancer, 2 (0.3%) of their prior cancer. Of all patients, 489 (70%) had an available trial, but 30% of patients with a prior cancer were ineligible for this reason. Overall and GI–cancer-specific survival were comparable for patients with/without a prior cancer.
Survival for patients with a GI cancer requiring chemotherapy relates to the GI cancer and rarely a prior cancer. These patients should not be excluded from clinical trial participation.
Keywords: Colorectal neoplasms, Esophageal neoplasms, Gastrointestinal neoplasms, Pancreatic neoplasms, Stomach neoplasms.
Randomized clinical trials are the backbone upon which novel, increasingly effective therapies for cancer are based, and patient participation in randomized trials may lead to enhanced survival in particular in the short term.1 and 2 However, the relevance of any clinical trial relies heavily on external validity, and trial generalizability may be significantly affected by factors such as exclusion criteria.3 Furthermore, although results from clinical trials provide the empirical evidence used to treat patients rationally, patient accrual often falls short of expectations. This is demonstrated by a study of 114 Medical Research Council sponsored trials that revealed that less than one-third achieved their target enrolment within the specified timeframe.4 Such low accrual rates may lead to insufficient statistical power and early trial termination. This represents a missed opportunity to answer a clinical and scientific question and a squandering of scant resources.
In theory, stringent eligibility criteria are necessary for clinical trials in order to maintain the safety of the patient and the scientific value of the protocol. Many of these criteria are common to all clinical trials, and along with “poor performance status,” one of the most frequent of these is the exclusion of patients with any previous invasive cancer.5 and 6 As the population ages and treatments of early stage tumors improve the proportion of patients with a prior history of cancer will inevitably increase – in 1971 there were 3 million cancer survivors in the United States; by 2007, this was 11.7 million.7 Excluding these patients may further decrease the pool of potential clinical trial participants and may limit the generalizability of trials that are performed. Although many suggestions have been made as to how to increase participation in clinical trials, the validity of this commonly used exclusion criteria has not been previously examined.
As survival for patients with advanced gastrointestinal (GI) cancers (esophagogastric, pancreatic, hepatocellular, cholangiocarcinoma, and colorectal cancer) is more likely to be determined by that cancer that any previous cancer, we hypothesized that previously treated, unrelated cancers are less likely to impact on survival than is suggested, and that, therefore, a history of a previously treated cancer may not be rational reason to exclude patients from clinical trials in GI malignancies.8, 9, 10, 11, and 12 In this study, we examine the survival outcomes of patients with and without a prior cancer treated with chemotherapy in the GI Unit of the Royal Marsden Hospital over a 1-year period, and how a history of a previous cancer in a patient affected clinical trial eligibility during that timeframe.
Materials and Methods
For this retrospective observational study, we collected data on patients with GI cancer (colorectal, esophagogastric, pancreatic, and hepatocellular cancer) who received treatment between January 1, 2006, and December 31, 2006, at the Royal Marsden Hospital, London. Patient details were extracted from the electronic patient record. Eligibility criteria included age of 18 years and over, diagnosis of colorectal, gastroesophageal, pancreatic, or hepatocellular cancer, and treatment with chemotherapy in the year 2006. The following information was collected following chart review: patient age, gender, GI cancer stage, prior cancer tissue of origin and stage, clinical trial availability and eligibility, and dates of cancer recurrence, death, cause of death, and last follow-up. The electronic records of the Gastrointestinal Clinical Trials Unit were reviewed to identify the opening and closing dates of pertinent clinical trials for the period under review, and the relevant protocols were reviewed with respect to eligibility criteria.
All non-trial patients in the Gastrointestinal Oncology Unit are followed according to departmental protocols. Patients receiving active treatment for metastatic disease are followed as per chemotherapy protocol while receiving chemotherapy and 3-monthly while on treatment breaks. Patients with resected neoadjuvantly or adjuvantly treated cancers are followed every 3 months for the first year, then 6-monthly for years 2 and 3 following resection, and then annually until year 5 when they are discharged. Patients with resected stage IV colorectal cancer are followed for 7 years. Survival outcomes are collected by hospital administrative staff from a national database. Patients participating in clinical trials were followed as per individual trial protocol.
Overall survival (OS) was calculated from date of diagnosis of GI cancer to date of death and GI–cancer-specific survival (GCSS) was calculated from date of diagnosis of GI cancer to date of death where cause of death was GI cancer (censored at date of death for other cause of death). Survival estimates were calculated according to the Kaplan-Meier method, and are presented with 95% confidence intervals (CIs). Differences in survival between groups were compared with the log-rank test. All analysis was performed in SPSS version 22. Multivariate Cox regression analysis was performed using the variables age, stage, and cancer subtype in order to assess the independent impact of these variables on GCSS and OS. The study was reviewed and approved by the Institutional Review Board (SE61) and did not require patient consent, nor did it have any influence on patient management. The year 2006 was chosen to allow for adequate follow up of at least 5 years for surviving patients and mature assessment of survival data.
A total of 697 patients were identified. Patient characteristics are presented in Table 1. The majority of patients were male (59%); the median age was 62 years for all patients. The most commonly treated GI tumor type was colorectal cancer (74%). Fifty-four (8%) patients had a previously treated cancer. Breast, prostate, and colon cancer together accounted for more than half of these previous cancers; almost 80% of previously treated cancers were Stage I or II. The median time from diagnosis of a previous cancer until current diagnosis was 7.9 years. Patients with a prior cancer were significantly older than patients without (median, 61 years vs. 67 years; P < .001), but were similar in all other baseline characteristics.
|No Prior Cancer||Prior Cancer||P Value|
|Diagnosis of primary|
|Stage of primary|
|Site of prior malignancy|
|Stage of prior malignancy|
Abbreviations: CRC = Colorectal cancer; EGC = esophagogastric cancer; HCC = hepatocellular carcinoma.
Of the 459 patients in the study with GI cancer that did not have metastatic cancer at presentation, 297 (65%) developed a recurrence of their GI cancer. Of the 54 patients with a previously treated cancer, 7 (12%) developed a recurrence of the previously treated cancer (however, 5 of these recurrences had occurred prior to their GI cancer diagnosis and been treated successfully with curative intent). Two patients had a recurrence of a previous cancer after their GI cancer diagnosis. These was 1 patient with a previous spinal cord glioma and 1 patient with a history of resected bladder cancer, and both of these patients died of their recurrent original non-GI cancer. To date, 504 (72%) patients have died, 170 (24%) are alive with no cancer, and 23 (3%) patients remain alive with cancer. Among the deceased patients, their primary GI cancer was the cause of death in 476 (94%) patients, followed by other reasons in 26 (5%) patients and their prior cancer in 2 (< 1%) patients.
Impact of Prior Cancer on Clinical Trial Recruitment
Of the 697 patients who were treated in the referent time period, 489 (70%) patients had an available clinical trial at the time of GI cancer treatment, and 113 (23% of patients with trial available; 16% of all patients) were enrolled into a clinical trial. Nine (2% of total) patients with a trial available were excluded because of a prior history of cancer. Of the 30 patients with prior cancer for whom a clinical trial was identified, 9 of these (30%) patients were excluded due to their previous malignancy. Thirteen patients were excluded from trial participation for other reasons; of the remainder who were trial eligible with an available trial (n = 8), 55% chose to participate in a clinical study. Both GCSS and OS were comparable between patients who participated and who did not participate in clinical trials. These were 35.4 months (95% CI, 25.4-45.5 months) versus 41.5 months (95% CI, 34.0-49.0 months) with a hazard ratio (HR) of 1.06 (95% CI, 0.82-1.37) and P = .662 for GCSS and 35.4 months (95% CI, 25.4-45.5 months) versus 40.0 months (95% CI, 32.8-47.2 months) with an HR of 1.00 (95% CI, 0.78-1.29) and P = .991 for median OS, respectively.
Impact of Prior Cancer on Survival
Median follow-up time was 83.2 months (95% CI, 77.9-88.5 months). Median GCSS was 41.4 months (95% CI, 36.7-46.1 months) for all patients. GCSS was 41.2 months (95% CI, 36.1-46.3 months) for patients with no history of a previous cancer versus 47.4 months (95% CI, 32.8-62.0 months) for prior cancer patients (P = .75; Figure 1). Median OS for all patients was 40.6 months (95% CI, 36.1-45.0 months). This was 40.6 months (95% CI, 36.1-45.1 months) for patients with no prior cancer patients versus 37.4 months (95% CI, 20.7-54.0 months) for patients who did have a prior cancer; these were not significantly different (P = .42; Figure 2). Survival was comparable for patients with curatively resected cancer (Stage I-III) and for patients with advanced disease (Stage IV), whether or not they had a prior cancer (P = .375 and .712 for interaction between stage and OS and disease-specific survival [DSS] respectively). Multivariate analysis demonstrated that, whereas for both DSS and OS age, stage and non-colorectal cancer histology were independently associated with survival, a history of a previous cancer was not (Table 2 and Table 3).
|Hazard Ratio||95% CI||P Value|
|Non CRC type||2.46||2.02-2.99||<.001|
Abbreviations: CI = Confidence interval; CRC = colorectal cancer.
|Hazard Ratio||95% CI||P Value|
|Non CRC type||2.40||1.98-2.92||<.001|
Abbreviations: CI = Confidence interval; CRC = colorectal cancer.
Stringent eligibility criteria in clinical trials may be justified to ensure patient safety and the scientific value of the protocol. However, this objective may impair generalizability of trial results as many patients are excluded due to common comorbidities; most frequently for kidney, infectious, cardiac, liver, or hematological-oncological disease, which may not impact on the research hypothesis in question.5 In our study, we provide evidence that, for patients with common GI malignancies, a prior history of a treated cancer does not result in inferior survival when compared with those patients without a previous malignancy.
Factors other than a history of a previous malignancy may influence the survival of cancer patients, and, in order to address the potential confounders of age, stage, and histological subtype within our study, we have performed multivariate analysis using these factors as variables. Unsurprisingly, age, the presence of metastatic disease, and non-colorectal cancer histology were independent predictors of survival. There are clear differences in the survival of GI cancers based on tissue of origin and stage. For example, median survival in clinical trials for patients with metastatic pancreatic cancer is less than 1 year, and for these patients, it is unlikely that a previous cancer poses a significant competing risk for death. Alternatively, for patients with less biologically aggressive metastatic cancers such as colorectal cancer, a previous cancer may represent a more frequent cause of death. However, regardless of cancer subtype, tumor stage remains the strongest predictor of survival in our analysis, and it may be possible to argue that for all patients with advanced or metastatic incurable cancer that a history of a prior malignancy is unlikely to be relevant with respect to survival.
In our study, age was also incrementally but independently associated with reduced survival. As patients in our study with a second cancer had a mean age 6 years older than those without a similar history, it is possible that this is the reason that the univariate Kaplan-Meier overall survival curve appears to diverge for patients with and without a previous cancer (we acknowledge that although this difference is non-significant, the analysis may be underpowered). However, despite the fact that increasing age predicts for worse survival, there is strong support for the enrolment of older patients in clinical trials in order to increase the representativeness of clinical trial populations.13 In contrast with increasing age, that a previous cancer does not appear to adversely affect survival provides further support for our argument against exclusion of these patients.
We note that in our study, the median time from a previous cancer diagnosis was almost 8 years, which raises the question of whether the lead time from previous cancer impacts on the chances of recurrence. The low absolute numbers of patients in our study with a previous cancer make it difficult to form any meaningful observations in this regard; however, we appreciate the point at which a patient may be considered “cured” of a cancer is variable – indolent hormone-sensitive breast cancer or lymphomas may recur after a period of decades, whereas late (> 5 years) recurrences of more aggressive epithelial cancers are rare. This clearly adds to the complexity of any proposal for including patients with prior cancers in clinical trials, but does not necessarily render the problem unsolvable. Due to the exponentially decreased risk of recurrence for most high grade cancers after 3 years, this would seem a very reasonable timeframe to suggest as a cancer-free baseline.
In this study population, the percentage of patients with a prior cancer was 7.7%, a proportion which likely reflects the GI cancer population studied (as second malignancies may be more common in other populations). It has been previously documented that patients with breast, prostate, and colon cancer are at significantly increased risk of second malignancies, and these patients represented the majority of patients in our study.14, 15, and 16 Rates of early diagnosis and cure of these cancers are high and may continue to increment with implementation of more comprehensive screening programs and increasing efficacy of treatment for early stage disease.17, 18, and 19 In the United States, the number of cancer survivors is increasing by 2% per annum, and, according to Surveillance, Epidemiology, and End Results Program data, as many as 18% of US cancers may be a second or subsequent malignancy.20 A recent study supports our argument – Gerber et al demonstrated that up to 80% of US United States Eastern Co-operative Oncology Group-sponsored or endorsed lung cancer trials excluded patients with a prior malignancy and that his affected up to 18% of lung cancer patients.21 The same group have also recently demonstrated that both lung cancer-specific survival and overall survival for lung cancer patients with a history of a prior cancer was not inferior to those without, thus supporting our hypothesis in another cancer group.22 Our study demonstrated that 2% of the total population were potentially denied participation in a clinical trial as a result of a previous cancer. Although this is a small proportion of the total compared with the previously mentioned study, it may be reflective of the patient make-up of our dataset, who were predominantly patients with colorectal cancer, which has a weaker link to environmental exposures such as tobacco. For example, 10% of patients in our study with gastroesophageal cancer had a history of a previous cancer, possibly reflecting a stronger link in this group to underlying alcohol and smoking etiologies. However, even if the figure of 2% is prevented from trial participation, we believe that this is unwarranted and unjust in principle. This is a significant proportion of patients to exclude from any clinical trial, especially given the lack of rationale for this evidenced in our study, and in particular from randomized trials where any effect (if one existed) would be nullified by randomization. Furthermore, when considering cancer survivors as a group, 30% of these patients were prevented from participating in trials as a result of their previous cancer. Furthermore, when a trial was available and this group was eligible, more than half of patients chose to participate. As advocacy groups for cancer survivors become increasingly involved in peer review of research strategies and funding, this inequity must surely be questioned.23
Limitations of our study include its retrospective nature and the inclusion of patients with differing cancer histologies. However, this is reflective of the patients treated at our cancer center, and we believe that it is generalizable to a wider population. We note that the survival of patient who did and did not participate in trials during this period was not different, which supports the external validity of these results. We chose one specific year (2006) in order to ascertain sufficient follow-up of patients to detect potential recurrences and also to provide a finite timescale in which to examine the trials that were on offer at that time. We acknowledge that our analysis could be underpowered despite containing almost 700 patients, as the proportion of patients with a prior cancer diagnosis is low and thus one might interpret any non-significant P values with caution. However, in performing multivariate analysis, we have attempted to demonstrate as robustly as possible within the dataset available that expected confounders of age, stage, and histological subtype are independently associated with survival and that a history of a previous cancer is not. Although expanding the time scale would provide us with greater power to examine each individual histology, doing this retrospectively would prove difficult due to a lack of previous electronic records, and prospective observational studies would require significant follow-up and may not provide additional information. Our study was conducted at a comprehensive cancer center, and this is evident in the high proportion of patients for whom a clinical trial was available (70%) and who participated in a trial (16% of total). Although this could potentially introduce a bias in survival times, this should apply equally to all patients as this is a single center study. The number of patients with a previous cancer is consistent with the literature, and does not appear to have been influenced by the nature of the institution; however, this may not be true of all cancers.24
In the UK, participation in clinical research is among the highest in developed countries; however, the aspiration is for all patients to be offered the opportunity to participate in clinical research should they so wish, and are eligible. Governments worldwide include reducing barriers to participation in trials as a stated goal.25 and 26 However, in order to improve on current recruitment rates to cancer clinical trials, it may be necessary to question the validity of commonly held beliefs regarding eligibility. Just as older patients should not be excluded by virtue of their age alone, we argue that neither should patients with a history of an adequately treated previous cancer.27 and 28 As significant health and socio-economic inequalities have previously been identified for cancer survivors, to remedy this one by providing equal access to clinical trials seems both sensible and fair.29, 30, and 31
Clinical Practice Points
- Exclusion of patients from clinical trials due to a previous cancer diagnosis is common; however, there is limited evidence to support this practice. As rates of cancer survival increase and a large group of patients are excluded from clinical trial participation it is necessary to challenge this paradigm.
- We demonstrate that, for patients with a new diagnosis of a gastrointestinal cancer and a history of a previous cancer, their survival relates to the current gastrointestinal cancer and not their previous malignancy. Furthermore, one-third of patients with a previous cancer were prevented from participating in a clinical trial due to their previous cancer.
- We conclude that patients with a previous cancer which had been previously treated should not be excluded from clinical trial participation. As survivorship issues become more prevalent, it is important that these patients have just access to clinical trials. Clinical trial generalizability is also improved by including patients reflective of the general population.
The authors have stated that they have no conflicts of interest.
Elizabeth Smyth, Clare Peckitt, Edward Armstrong, David Cunningham, and Ian Chau acknowledge funding support from the National Institute of Health Research, Royal Marsden/Institute of Cancer Research Biomedical Research Centre. Noelia Tarazona acknowledges an European Society of Medical Oncology Translational Research Fellowship. This work was conducted at the Royal Marsden Hospital, London, United Kingdom.
- 1 J.M. Unger, W.E. Barlow, D.P. Martin, et al. Comparison of survival outcomes among cancer patients treated in and out of clinical trials. J Natl Cancer Inst. 2014;106:dju002 Crossref
- 2 G.E. Vist, D. Bryant, L. Somerville, T. Birminghem, A.D. Oxman. Outcomes of patients who participate in randomized controlled trials compared to similar patients receiving similar interventions who do not participate. Cochrane Database Syst Rev. 2008;:MR000009
- 3 P.M. Rothwell. External validity of randomised controlled trials: “to whom do the results of this trial apply?”. Lancet. 2005;365:82-93 Crossref
- 4 M.K. Campbell, C. Snowdon, D. Francis, et al. Recruitment to randomised trials: strategies for trial enrollment and participation study. The STEPS study. Health Technol Assess. 2007;11:iii-ix 105
- 5 H.G. Van Spall, A. Toren, A. Kiss, R.A. Fowler. Eligibility criteria of randomized controlled trials published in high-impact general medical journals: a systematic sampling review. JAMA. 2007;297:1233-1240 Crossref
- 6 P.N. Lara Jr., R. Higdon, N. Lim, et al. Prospective evaluation of cancer clinical trial accrual patterns: identifying potential barriers to enrollment. J Clin Oncol. 2001;19:1728-1733
- 7 Cancer survivors–United States, 2007. MMWR Morb Mortal Wkly Rep. 2011;60:269-272
- 8 D. Cunningham, N. Starling, S. Rao, et al. Capecitabine and oxaliplatin for advanced esophagogastric cancer. N Engl J Med. 2008;358:36-46 Crossref
- 9 T. Conroy, F. Desseigne, M. Ychou, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med. 2011;364:1817-1825 Crossref
- 10 J.M. Llovet, S. Ricci, V. Mazzaferro, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med. 2008;359:378-390 Crossref
- 11 J. Valle, H. Wasan, D.H. Palmer, et al. Cisplatin plus gemcitabine versus gemcitabine for biliary tract cancer. N Engl J Med. 2010;362:1273-1281 Crossref
- 12 V. Heinemann, L.F. von Weikersthal, T. Decker, et al. FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer (FIRE-3): a randomised, open-label, phase 3 trial. Lancet Oncol. 2014;15:1065-1075 Crossref
- 13 A. Hurria, W. Dale, M. Mooney, et al. Designing therapeutic clinical trials for older and frail adults with cancer: U13 conference recommendations. J Clin Oncol. 2014;32:2587-2594 Crossref
- 14 L. Mellemkjær, S. Friis, J.H. Olsen, et al. Risk of second cancer among women with breast cancer. Int J Cancer. 2006;118:2285-2292
- 15 M. Van Hemelrijck, A. Feller, H. Garmo, et al. Incidence of second malignancies for prostate cancer. PloS One. 2014;9:e102596 Crossref
- 16 S. Noura, M. Ohue, Y. Seki, et al. Second primary cancer in patients with colorectal cancer after a curative resection. Dig Surg. 2009;26:400-405 Crossref
- 17 S.H. Lo, S. Halloran, J. Snowball, H. Seaman, J. Wardle, C. von Wagner. Colorectal cancer screening uptake over three biennial invitation rounds in the English bowel cancer screening programme. Gut. 2015;64:282-291 Crossref
- 18 M. Plescia, F. Wong, J. Pieters, D. Joseph. The National Breast and Cervical Cancer Early Detection Program in the era of health reform: a vision forward. Cancer. 2014;120(Suppl 16):2620-2624 Crossref
- 19 S. Loeb. Guideline of guidelines: prostate cancer screening. BJU Intl. 2014;114:323-325
- 20 K.C. Oeffinger, S.S. Baxi, D. Novetsky Friedman, C.S. Moskowitz. Solid tumor second primary neoplasms: who is at risk, what can we do?. Sem Oncol. 2013;40:676-689 Crossref
- 21 D.E. Gerber, A.L. Laccetti, L. Xuan, E.A. Halm, S.L. Pruitt. Impact of prior cancer on eligibility for lung cancer clinical trials. J Natl Cancer Inst. 2014;106:dju102
- 22 A.L. Laccetti, S.L. Pruitt, L. Xuan, E.A. Halm, D.E. Gerber. Effect of prior cancer on outcomes in advanced lung cancer: implications for clinical trial eligibility and accrual. J Natl Cancer Inst. 2015;107:djv102
- 23 I.M. Rich, Y. Andejeski, M.H. Alciati, et al. Perspective from the Department of Defense Breast Cancer Research Program. Breast Dis. 1998;10:33-45
- 24 E.C. Smyth, M. Hsu, K.S. Panageas, P.B. Chapman. Histology and outcomes of newly detected lung lesions in melanoma patients. Ann Oncol. 2012;23:577-582 Crossref
- 25 Federally Funded Research: The Catalyst for Progress Against Cancer. Vol 20142013. Available at: http://www.asco.org/advocacy/federally-funded-research-catalyst-progress-against-cancer. Accessed Dec 23, 2015.
- 26 Department of Health U. Increasing research and innovation in health and social care. 2013 [cited March 17, 2015]. Available at: https://www.gov.uk/government/policies/increasing-research-and-innovation-in-health-and-social-care/supporting-pages/health-research. Accessed Dec 23, 2015
- 27 S.F. Zafar, L.K. Heilbrun, P. Vishnu, et al. Participation and survival of geriatric patients in Phase I clinical trials: the Karmanos Cancer Institute (KCI) experience. J Geriatr Oncol. 2011;2:18-24 Crossref
- 28 J. Kazmierska. Do we protect or discriminate? Representation of senior adults in clinical trials. Rep Pract Oncol Radiother. 2012;18:6-10
- 29 J.H. Rowland, K.M. Bellizzi. Cancer survivorship issues: life after treatment and implications for an aging population. J Clin Oncol. 2014;32:2662-2668 Crossref
- 30 F. Mols, M.S. Thong, P. Vissers, T. Nijsten, L.V. van de Poll-Franse. Socio-economic implications of cancer survivorship: results from the PROFILES registry. Eur J Cancer. 2012;48:2037-2042 Crossref
- 31 P.B. Jacobsen, H.S. Jim. Consideration of quality of life in cancer survivorship research. Cancer Epidemiol Biomarkers Prev. 2011;20:2035-2041 Crossref
1 Department of Gastrointestinal Oncology, Royal Marsden Hospital, Sutton, UK
2 Department of Gastrointestinal Oncology, Royal Marsden Hospital, Sutton, UK
3 Royal Marsden Hospital, Sutton, UK
4 Royal Marsden Hospital, Sutton, UK
5 Heartlands Hospital, Bordesley Green East, Birmingham, UK
6 Department of Gastrointestinal Oncology, Royal Marsden Hospital, Sutton, UK
7 Department of Gastrointestinal Oncology, Royal Marsden Hospital, Sutton, UK
∗ Address for correspondence: Dr Ian Chau, MD, Royal Marsden Hospital, Downs Road, Sutton, SM2 5PT, United Kingdom.
© 2015 Elsevier Inc., All rights reserved.