- Email: [email protected]
A prospective, blinded analysis of thymidylate synthase and p53 expression as prognostic markers in the adjuvant treatment of colorectal cancer
original article
Annals of Oncology 17: 1810–1817, 2006 doi:10.1093/annonc/mdl301 Published online 13 September 2006
A prospective, blinded analysis of thymidylate synthase and p53 expression as prognostic markers in the adjuvant treatment of colorectal cancer S. Popat1*, Z. Chen2, D. Zhao3, H. Pan2, N. Hearle1, I. Chandler1, Y. Shao3, W. Aherne4 & R. S. Houlston1
Received 9 May 2006; revised 9 July 2006; accepted 10 July 2006
Background: Despite previous studies, uncertainty has persisted about the role of thymidylate synthase (TS) and p53 status as markers of prognosis in colorectal cancer (CRC).
original article
Patients and methods: A total of 967 patients accrued to a large adjuvant trial in CRC were included in a prospectively planned molecular substudy, and of them, 59% had rectal cancer and about 90% received adjuvant chemotherapy (either systemically or randomly allocated to intraportal 5-fluorouracil infusion or both). TS and p53 status were determined, blinded to any clinical data, by immunohistochemistry using a validated polyclonal antibody or the DO-7 clone, respectively, and their relationships with overall survival were examined. Results: High TS expression was observed in 58% and overexpression of p53 in 60% of tumours. TS expression correlated with tumour stage, and p53 overexpression, with rectal cancers. There was no evidence that either marker was significantly associated with survival by either univariate (TS hazard ratio (HR) = 0.94, 95% CI 0.76–1.18 and P = 0.6 and p53 HR = 0.98, 95% CI 0.78–1.23 and P = 0.9) or multivariate analyses (TS HR = 0.99, 95% CI 0.79–1.25 and P = 0.9 and p53 HR = 0.98, 95% CI 0.78–1.23 and P = 0.8). Conclusions: Neither TS nor p53 expression has significant prognostic value in the adjuvant setting of CRC. Key words: colorectal cancer, p53, prognosis, thymidylate synthase
introduction Colorectal cancer (CRC) is one of the commonest cancers, affecting around 700 000 individuals each year worldwide [1] and in many Western countries such as United States it is the commonest cause of cancer death in nonsmoking men [2]. The natural history of CRCs are not necessarily similar, and tumour molecular profile is likely to play an important role in determining the prognosis for individual patients [3, 4]. Adjuvant chemotherapy is the standard of care for stage III CRC [5–8], but its role in stage II CRC is limited to those at greater risk of relapse as defined by histopathological criteria [9, 10]. Despite curative surgery, however, many patients eventually relapse even following adjuvant therapy [11], while many of those who currently receive adjuvant therapy might never have relapsed. The identification of robust molecular markers to supplement standard clinical and pathological staging systems is therefore of particular relevance for patients with CRC, especially in the light of recently introduced drugs against CRC [8, 12]. *Correspondence to: Dr S. Popat, Section of Cancer Genetics, Institute of Cancer Research, Sutton, SM2 5NG, UK. Tel: +44-(0)-208-642-6011; Fax: +44-(0)-208-643-0373. E-mail: [email protected]
ª 2006 European Society for Medical Oncology
Two molecular markers that have been the subject of much investigation are expression level of the enzyme thymidylate synthase (TS) and mutation status of TP53 [13–15]. TS is a ratelimiting enzyme involved in DNA synthesis [16], and competitive inhibition of its activity appears to be the major mechanism for the antitumour effect of fluoropyrimidines [17], the cornerstone of therapy for CRC. In vitro data indicate that TS expression level is a determinant of fluoropyrimidine sensitivity [16, 18], suggesting that TS expression might determine tumour sensitivity in vivo [13, 16, 18]. The tumour suppressor TP53, is critically involved in the control of cell cycle and apoptosis [19–22] and is commonly mutated in CRC [23]. Mutation commonly results in expression of protein with abnormal conformation, characterised by an increased nuclear half-life, which is readily detected as p53 overexpression by immunohistochemistry [20, 24]. Several in vitro studies have reported a relationship between TP53 mutation status and sensitivity to a number of cytotoxic agents, including fluoropyrimidines [25–27]. Several studies have previously investigated the prognostic effects of TS and p53 expression in CRC. Most of these studies were relatively small and mainly involved retrospectively
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
1 Section of Cancer Genetics, Institute of Cancer Research, Sutton, UK; 2Clinical Trial Service Unit, Richard Doll Building, Oxford, UK; 3Department of Abdominal Surgery, Tumour Hospital, Chinese Academy of Medical Sciences, Beijing, People’s Republic of China; 4Cancer Research UK Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, UK
original article
Annals of Oncology
methods patients and samples Details about the study design, objectives and patient population of the original clinical trial have previously been reported [30]. Briefly, the trial was set up in 1994 to assess the effects of a 7-day continuous intraportal vein infusion of 5-fluorouracil (5-FU), started immediately after tumour resection, on long-term survival in patients with stage I–III CRC [30, 31]. Two hundred hospitals throughout China participated in the trial, and >10 000 patients had been entered in a randomised trial from 1994 to 1998 with the current cycle of follow-up on survival continuing to the middle of June 2003 (median 6 years, in survivors). Of the collaborating hospitals, 39 agreed to participate in a substudy of molecular markers that involved collection of a pair of paraffin-embedded, formalin-fixed tissue blocks (one tumour tissue and one normal tissue) for each randomly assigned patient. All required approvals for the main trial and the molecular substudy were obtained from relevant organisations and ethical committees in China, and informed consent was obtained from individual patients before the investigation. Overall, a total of 967 sets of samples were collected prospectively during the course of the trial. Following official approval by the Chinese Administration for Human Genetic Materials samples, all pathological specimens were sent to the UK Clinical Trial Service Unit (CTSU) in Oxford, where they were anonymised before forwarding to the Section of Cancer Genetics for laboratory analysis.
immunohistochemistry methods Three-micron sections from a representative part of the primary tumour were cut onto silane-coated glass slides and assessed for TS and p53 expression by the avidin–biotin complex immunohistochemical technique (Vectastain Elite ABC kit; Vector Laboratories Inc., Burlingame, CA) [32]. Negative and positive control slides were included in each staining run. The
Volume 17 | No. 12 | December 2006
negative control consisted of replacing primary antibody with phosphatebuffered saline [pH 7.6/0.1% (by vol) Tween solution] whereas positive controls were sections of tumour known to stain heavily for TS or p53. Tumour sections were deparaffinised in Histoclear (National Diagnostics, Hull, UK) and hydrated in decreasing concentrations of ethanol. Endogenous peroxidase activity was quenched with 5% hydrogen peroxide in methanol for 20 min. For p53 (but not TS) immunohistochemistry, antigen retrieval was performed using a microwave oven-based method. Specifically, sections were incubated in boiling 10 mmol/l citric acid buffer (pH 6.0) for 10 min and then cooled in running water. Nonspecific background staining was blocked with 20% goat serum for 20 min. Sections were then incubated with appropriate primary antibodies at a 1:100 dilution, in a humidified chamber at room temperature for 60 min, using either a validated rabbit polyclonal antibody to recombinant human TS [33, 34] or an anti-p53 mouse monoclonal antibody (clone DO-7; Dako Corp., Glostrup, Denmark). After rinsing, a biotinylated anti-rabbit or anti-mouse secondary antibody was applied for 30 min followed by further washing, and avidin–biotin–peroxidase complexes (Vectastain Elite ABC kit; Vector Laboratories Inc.). Immunostaining was developed by applying freshly prepared 0.05% 3,39-diaminobenzidine tetrahydrochloride (Vectastain Elite ABC kit; Vector Laboratories Inc.). Sections were counterstained in Mayer’s Haematoxylin (Sigma Chemical Co., St Louis, MO), dehydrated in a series of ethanols, cleared in Histoclear (National Diagnostics) and mounted with glass coverslips using DePeX (BDH, Poole, UK).
immunohistochemistry evaluation All slides were randomly allocated to independent assessment by two observers blinded to clinical information. TS expression was categorised semi-quantatively into four groups (0–3) based on chromagen intensity, with the highest tumour staining detected being used as the reference for grading. Grades 0 and 1, representing none and minimal staining, respectively, were defined as the ‘low’ expression group, whereas grades 2 and 3 were defined as the ‘high’ expression group. p53 immunoreactivity was dichotomised into positive or negative based on staining of malignant nuclei, with a threshold of 10%. Thresholds to assign marker status were those currently established and used by the vast majority of researchers in this field [13, 14], thereby allowing a comparative assessment of our dataset. Level of scoring agreement between the two observers was recorded and was in excess of 85% (see Results section). In cases of disagreement, marker status was discussed and determined by consensus. All immunohistochemistry and results scoring was performed blinded to clinical outcome and recorded on a separate database before merging with the main database kept in Oxford CTSU for the final data analysis.
statistical analysis Differences in demographic data and between different categories of markers were tested for significance using the v2 test for categorical variables and the Student’s t-test or Mann–Whitney nonparametric test for continuous variables. The level of agreement between assessors was assessed using the j statistic. A Cox regression model was used with individual marker as the exposure variables and OS (from time of surgery to time of death or end of current follow-up) as the outcome. The analyses were adjusted simultaneously for sex, age, tumour size, grade (World Health Organization), stage and sites as well as use of post-operative adjuvant therapies. The cumulative probability of survival was calculated by the Kaplan–Meier method and groups were compared by the stratified log-rank test. Information on progression-free survival was not available for all randomly assigned patients, nor were most histopathological features, e.g. T-stage, and number of lymph nodes resected at time of surgery. For all patients randomized to the clinical study, information was available neither for progression-free survival nor for most histopathological features, e.g. T-stage, and number of lymph nodes resected at time of surgery. All P values were two-sided, with
doi:10.1093/annonc/mdl301 | 1811
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
collected samples and clinical data with unblinded analytic methods, and the results have been inconsistent [13, 14]. In addition, meta-analyses published on the prognostic utility of both markers [13, 14] were not based on individual patient data, and results may have been biased by a variety of factors [28]. To help address such deficiencies, we present results from a large prospectively designed molecular study of about 1000 earlystage CRC patients, accrued into a multi-centre randomised trial of adjuvant therapy. The prespecified hypotheses tested were that TS expression level and p53 expression status are markers of overall survival (OS) in potentially curatively resected CRC, and the results are reported as per the Reporting Recommendations for Tumour Marker Prognastic Studies (REMARK) guidelines [29]. These guidelines were recently published as a standard for reporting tumour marker prognostic studies [29], as a major recommendation of the National Cancer Institute–European Organisation for Research and Treatment of Cancer. Using this standard, relevant information on study design, preplanned hypotheses, patient and specimen characteristics, assay methods and statistical analysis methods are specified, to enable transparent and complete reporting so that the relevant information be available to other researchers in the field to assess the usefulness of the data and understand the context of the conclusions. These guidelines complement the prospective, blinded, large patient population design of this study and together address many of the inadequacies of previously reported studies in this field.
original article
Annals of Oncology
patient demographics Among the 967 patients with samples available for investigation of TS and p53 expression, the overall characteristics were similar to those in the main trial (Table 1). The mean age was 55 years, with 45% aged below 55 years and 53% were male. There were more rectal (60%) than colon cancers (40%), and most (57%) had stage II disease. Neoadjuvant radiotherapy was hardly given, but about 80% of patients received adjuvant systemic chemotherapy (chiefly 5-FU based) on top of randomly allocated 7-day intraportal 5-FU infusion. So, overall in this
TS and p53 expression Of the 967 samples tested, 953 and 951 were assessable for TS and p53 expression, respectively, with the remainder nonevaluable due to either insufficient tumour present or nonrepresentative specimens. For TS analysis, a further 174 (18%) samples were nonevaluable due to repeatedly poor staining, resulting in 779 samples with available TS data (Table 2). Of these, 58% (449) expressed high TS levels (grade 3, 22% and grade 2, 36%), with the remainder
Table 1. Comparison of baseline characteristics between those with tumour samples collected and those without in the original randomised trial Characteristic Sex Male Female Age at diagnosis (years) <45 45–54 55–64 >65 Tumour site Colon Rectum Others Tumour size (cm) <4.0 4.0–5.9 6.0–7.9 >8.0 Stage I II III Tumour differentiation Well/moderate Poor Neoadjuvant radiotherapy Given Not given Nontrial adjuvant chemotherapy Given Not given Randomised treatment Intraportal 5-FU Control
Tumours assessed (%)
Tumours not assessed (%)
509 (53) 458 (47)
5358 (54) 4601 (46)
235 201 303 228
2555 2156 3110 2138
(24) (21) (31) (24)
378 (39) 573 (59) 16 (2)
3545 (36) 6293 (63) 121 (1)
207 394 203 163
1967 4588 2091 1313
(21) (41) (21) (17)
(20) (46) (21) (13)
76 (8) 555 (57) 336 (35)
883 (9) 5445 (55) 3631 (36)
696 (87) 103 (13)
NA (–) NA (–)
8 (1) 959 (99)
141 (1) 9818 (99)
766 (79) 201 (21)
7989 (80) 1970 (20)
471 (49) 496 (51)
4979 (50) 4980 (50)
5-FU, 5-fluorouracil and NA, not ascertained.
1812 | Popat et al.
(26) (22) (31) (21)
Table 2. Patient characteristics according to marker expression status Characteristic
Sex Male Female Age at diagnosis (years) <45 45–54 55–64 >65 Tumour site Colon Rectum Others Tumour size (cm) <4.0 4.0–5.9 6.0–7.9 >8.0 Stage I II III Tumour differentiation Well/moderate Poor Neoadjuvant radiotherapy Given Not given Nontrial adjuvant chemotherapy Given Not given Randomised treatment Portal 5-FU Control
TS status No. with high expression (%)
p53 status No. positive (%)
246 (60) 203 (55)
243 (61) 210 (59)
113 96 121 119
108 93 153 99
(58) (59) (51) (65)
(57) (59) (67) (55)
182 (61) 260 (56) 7 (47)
151 (52)* 294 (65) 8 (53)
96 185 98 70
93 191 101 68
(60) (58) (59) (52)
(59) (63) (63) (50)
39 (74)* 259 (59) 151 (53)
35 (66) 248 (58) 170 (61)
399 (59) 48 (49)
395 (60) 56 (58)
2 (29) 447 (58)
5 (71) 448 (60)
365 (58) 84 (57)
377 (61) 76 (55)
213 (56) 236 (59)
219 (60) 234 (60)
TS, thymidylate synthase and 5-FU, 5-fluorouracil. *P < 0.05.
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
results
patient population, 10% received intraportal 5-FU alone, 39% received both intraportal 5-FU and adjuvant (nontrial) 5-FU, and 10% received neither intraportal 5-FU nor adjuvant chemotherapy. The 5-year survival rate for patients investigated was 57.8% [standard error (SE) 1.7%] and was similar to that observed for all patients in the main trial (59.5%, SE 0.5%). Efficacy data from the main trial are yet to be reported.
values <0.05 regarded as conventionally statistically significant. Assuming a control survival rate of 60% and 50% of patients with high TS expression or p53 overexpression [13, 14], then analysis of tissue samples from 750 patients will have 80% power to detect an absolute difference of 10% in OS associated with the expression of either of these markers.
original article
Annals of Oncology
having low level of expression (grade 0, 3% and grade 1, 39%; Figure 1). For p53 status, a further 196 (20%) samples were nonevaluable due to either repeatedly poor staining or tissue destruction at antigen retrieval, resulting in available data from 755 (78% of the total) patients for the present analysis (Table 2). Of these samples, just over half demonstrated p53 overexpression (453, 60%; Figure 1). Level of agreement between observers for both TS and p53 expression was high (j > 0.94). Of the 754 patients with both TS and p53 results available (78% of the total), a significant relationship between TS and p53 status was observed (P = 0.0006), with tumours expressing high levels of TS more likely to overexpress p53.
discussion The present study is one of the largest investigations of the prognostic value of molecular markers in CRC reported to date.
Figure 1. Immunohistochemical analysis of tumours for thymidylate synthase (TS) and p53. Tumour showing (A) high-level TS expression, (B) low-level TS expression, (C) p53 overexpression and (D) no p53 overexpression.
Volume 17 | No. 12 | December 2006
doi:10.1093/annonc/mdl301 | 1813
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
marker status and clinical correlates Significant associations were observed between marker status and several clinical variables (Table 2). TS status was associated with stage (P = 0.02), with tumours less likely to express high TS levels if advanced at resection. These associations were not observed for p53 status, where marker status was instead associated with site (rectal tumours more frequently being p53 positive, P < 0.001). There was no interaction between markers and tumour grade. At the end of the current cycle of follow-up, 400 (41%) patients had died, mostly from disease recurrence, with median follow-up of 51 months (0–100). OS was significantly associated with CRC stage (HR = 1.68, 95% CI 1.37–2.06 and P < 0.0001), and marginally significantly with grade (HR = 1.35, 95% CI 0.99–1.84 and P = 0.06), after correction for co-variates, with 5-year survival rates of 61.5% (SE 2.2%) and 46.4% (SE 2.8%) for stage II and III CRC, and 61.7% (SE 1.9%) and 45.5% (SE 5.1%) for well/moderately and poorly differentiated disease, respectively. This survival rate was
consistent with that previously reported from China [35, 36]. No statistically significant difference in OS was observed between tumours with or without high TS expression, with a 5-year survival rate of 59.0% (SE 2.9%) in patients with high TS levels compared with 56.8% (SE 3.2%) in those with low TS levels (HR = 0.94, 95% CI 0.76–1.18 and P = 0.6; Figure 2). Similarly, no statistically significant difference in OS was observed between the two different p53 groups, with a 5-year survival rate of 58.0% (SE 2.9%) in patients with p53-positive tumours compared with 57.8% (SE 3.4%) in p53-negative tumours (HR = 0.98, 95% CI 0.78–1.23 and P = 0.9; Figure 3). The effect of TS and p53 expression in patients treated by surgery alone was not specifically assessed due to small number of patients involved. When all standard prognostic clinical variables were included as co-variables in a Cox proportional hazards model, there was again no evidence that these two markers were significantly associated with OS (HR = 0.99, 95% CI 0.79–1.25 and P = 0.9 for TS and HR = 0.98, 95% CI 0.78–1.23 and P = 0.8 for p53). When restricting analysis to a subset of patients with stage III rectal cancers (n = 167), however, there was an indication that high TS levels were associated with a poor OS (HR = 1.55, 95% CI 1.01– 2.38 and P = 0.04), whereas for p53 there was also a borderline significant relationship with OS (HR = 1.58, 95% CI 0.99–2.50 and P = 0.05). No significant relationship was found with OS in the stage I–II rectal cohorts for either marker (TS HR = 0.87, 95% CI 0.60–1.26 and P = 0.5 and p53 HR = 0.82, 95% CI 0.55–1.22 and P = 0.3). These subsets were, however, analysed in a post hoc manner, and should therefore be interpreted with caution.
original article
Figure 3. Overall survival in all patients stratified by p53 status.
Despite the use of prospective study design and blinded laboratory assessment of TS and p53 status to minimise the potential biases seen in previous such studies [29], the present study provides little evidence that TS or p53 overexpression is of any significant prognostic value in the adjuvant setting of CRC.
1814 | Popat et al.
Although it involved a Chinese patient population that has not been extensively studied previously, these results were generally consistent with many, though not all, previous studies conducted in Western populations [37–41]. While there is some suggestion that TS and p53 might be associated with poor OS in a subset of patients with stage III rectal cancer, this result may be due chiefly or wholly to the play of chance, given the large number of subgroups examined, the small number of patients involved in that particular subset and the marginally statistical significance of the results. It is highly plausible that TS expression or p53 status may influence CRC prognosis, given that TS is a target for 5-FU [42] and also a regulator of p53 expression [43, 44]. There is also evidence that p53 itself could potentially modulate 5-FU resistance [17]. Several studies have previously reported on the relationships between TS expression, p53 status and survival in CRC but results have been inconsistent, with some showing a positive relationship [45, 46], while others show no such relationship [33, 40, 47–49]. In studies of patients with advanced CRC, high TS expression has been shown in some to be associated with poorer OS [13], whereas many other studies in the adjuvant setting have generally failed to document a relationship between levels of expression and OS, particularly in the setting of adjuvant 5-FU-based regimens [38–41]. It is unclear what accounts for the apparent discrepancy between observations in the advanced and adjuvant settings. While it is possible that the level of TS expression in primary resected tumour may not reflect its expression level in metastases [47, 50] or associated lymph nodes [37], it is equally possible that much of the difference is methodologically related. By contrast to the present study, most previous studies, especially those in advanced CRC, were very small typically involving only about 100 patients and tended to use retrospectively retrieved tissue samples and clinical data, as well as unblinded laboratory assays for molecular markers. Consequently, the results could be subject to large biases. By using a large prospectively designed study with central unblinded and established molecular assays, it is possible to minimise much of the problems commonly found in previous studies so as to provide an unbiased and reliable, though not necessarily conclusive, assessment of prognostic significance of TS and p53. There is suggestion from a pooled analysis of 584 patients that TS level may predict poor OS only in patients treated by surgery alone without adjuvant 5-FU [13]. If this were true, then it may suggest that adjuvant 5-FU could potentially abrogate a poorer prognosis ascribed by high TS expression, possibly a consequence of 5-FU scheduling. This hypothesis, however, cannot be tested directly in the present study, as nearly all patients received some form of 5-FU-based chemotherapy following surgery. In vitro data indicate that induction of high TS expression could potentially lead to 5-FU resistance in some types of tumour cells but is dependent on whether 5-FUmediated cytotoxicity in cells is mediated primarily by TS inhibition or by nucleotide misincorporation [51, 52]. The latter non-TS-directed effect may have been the more predominant mechanism of 5-FU action when administered as a bolus regimen [53]. Unfortunately, no detailed information was available in the present study regarding the regimen of adjuvant 5-FU administered, so that the question as to whether any
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
Figure 2. Overall survival in all patients stratified by thymidylate synthase status.
Annals of Oncology
original article
Annals of Oncology
Volume 17 | No. 12 | December 2006
In summary, we have utilised standard molecular methods to evaluate TS and p53 as prognostic markers in the largest patient cohort from a single clinical trial reported to date. Our findings provide little evidence to indicate that in the adjuvant setting, TS or p53 expression is predictive of OS, and these markers should not be used in routine clinical decision making. We cannot, however, exclude a role for TS in patients treated in the adjuvant setting by surgery alone and further studies, particularly in those with good-prognosis stage II CRC, are warranted. This will help better define the prognostic utility of this marker without the potential covariate of adjuvant chemotherapy.
acknowledgements This work was supported by grants from the Cancer Research UK and the Association for International Cancer Research. SP was in receipt of a Clinician Scientist Fellowship from the UK Department of Health and IC a Clinical Training Fellowship from St George’s Hospital Medical School. The UK Medical Research Council, the British Heart Foundation and Cancer Research UK provided core funding to the CTSU.
references 1. Ferlay J, Bray F, Pisani P et al. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC CancerBase No. 5. Lyon, France: IARC Press 2001. 2. American Cancer Society. Cancer facts and figures 2004. Atlanta: American Cancer Society; 2004. 3. Kahlenberg MS, Sullivan JM, Witmer DD et al. Molecular prognostics in colorectal cancer. Surg Oncol 2003; 12: 173–186. 4. Johnston PG. Of what value genomics in colorectal cancer? Opportunities and challenges. J Clin Oncol 2004; 22: 1538–1539. 5. Kerr DJ, Gray R, McConkey C et al. Adjuvant chemotherapy with 5-fluorouracil, L-folinic acid and levamisole for patients with colorectal cancer: non-randomised comparison of weekly versus four-weekly schedules—less pain, same gain. QUASAR Colorectal Cancer Study Group. Ann Oncol 2000; 11: 947–955. 6. Efficacy of adjuvant fluorouracil and folinic acid in colon cancer. International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) investigators. Lancet 1995; 345: 939–944. 7. NIH Consensus Conference. Adjuvant therapy for patients with colon and rectal cancer. JAMA 1990; 264: 1444–1450. 8. Andre T, Boni C, Mounedji-Boudiaf L et al. Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. N Engl J Med 2004; 350: 2343–2351. 9. Figueredo A, Charette ML, Maroun J et al. Adjuvant therapy for stage II colon cancer: a systematic review from the Cancer Care Ontario Program in evidencebased care’s gastrointestinal cancer disease site group. J Clin Oncol 2004; 22: 3395–3407. 10. Benson AB III, Schrag D, Somerfield MR et al. American Society of Clinical Oncology recommendations on adjuvant chemotherapy for stage II colon cancer. J Clin Oncol 2004; 22: 3408–3419. 11. Moertel CG, Fleming TR, Macdonald JS et al. Fluorouracil plus levamisole as effective adjuvant therapy after resection of stage III colon carcinoma: a final report. Ann Intern Med 1995; 122: 321–326. 12. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: 2335–2342. 13. Popat S, Matakidou A, Houlston RS. Thymidylate synthase expression and prognosis in colorectal cancer: a systematic review and meta-analysis. J Clin Oncol 2004; 22: 529–536.
doi:10.1093/annonc/mdl301 | 1815
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
differences in this aspect of patient management may have biased outcome cannot be investigated directly. Most previous molecular analyses have been conducted in Western populations, and there is little direct information about the molecular features of CRC in populations such as China, where disease incidence is still much lower than that in Western Europe and North America. The present study in China showed that, although there is a difference between Chinese and Western patients in the main types of CRC observed (more rectal rather than colon cancer in China), the proportion of patients with high levels of TS expression was generally similar to that in other reported series in the Western populations [13], as was the proportion of tumours overexpressing p53 [14]. Although the frequency of p53 overexpression in this dataset was marginally higher than that assumed for power calculations, this difference did not impact on the power of our study. Previous reports have indicated an increased incidence of the 3R/3R 28-base-pair variable number of tandem repeat (VNTR) polymorphism in Chinese patients [54], which could result in a greater proportion of high TS expression. Although TS genotype was not specifically assessed, the lack of overrepresentation by tumours expressing high TS levels may be accounted for by either an unknown prevalence of the functional G/C single-nucleotide polymorphism within the third VNTR tandem repeat [55, 56] or a complex relationship between genotype and expression when measured by immunohistochemistry [57]. By comparison with TS expression, the prognostic value of p53 status has been less well established. Although a number of studies have reported a poorer OS in patients with TP53 mutations or p53 overexpression, there has been considerable variation in findings [14, 15], due perhaps in part to the same methodological problems as in studies of TS (nonblinded investigators, small samples sizes and retrospective analyses) and in part to the method of ascertaining p53 status. A number of studies have investigated TP53 directly by mutational analysis, while others have adopted indirect methods on the basis of either immunohistochemistry or allelic imbalance at chromosome 17p [14, 15, 58, 59], and each of them is also likely to be subject to large experimental variation. In particularly, exons screened for mutations have varied between studies, as have microsatellite markers genotyped. Although immunohistochemistry-based studies have used a variety of antibodies, the most commonly used one, as in the present study, is DO-7 [15]. Despite this and the large number of patients involved, the present study did not provide any supportive evidence that p53 overexpression is of prognostic significance in the adjuvant setting. In immunohistochemistry-based analyses, however, although p53 overexpression is used as a surrogate marker of TP53 mutation, it does not fully correlate with mutation status [60]. Mutation does not necessarily correlate with p53 expression [61] or nuclear aggregation. Discordance between TP53 sequence variation and expression status by immunohistochemistry has been observed in up to 30% of tumours [60] with some tumours completely lacking protein expression nonfunctional secondary to mutation [61], and other nontruncating functional sequence variants undetectable by immunohistochemistry [62]. This could therefore potentially obscure any real relationship between longterm survival prognosis and TP53 mutation.
original article
1816 | Popat et al.
38. Nanni O, Volpi A, Frassineti GL et al. Role of biological markers in the clinical outcome of colon cancer. Br J Cancer 2002; 87: 868–875. 39. Edler D, Glimelius B, Hallstrom M et al. Thymidylate synthase expression in colorectal cancer: a prognostic and predictive marker of benefit from adjuvant fluorouracil-based chemotherapy. J Clin Oncol 2002; 20: 1721–1728. 40. Allegra CJ, Parr AL, Wold LE et al. Investigation of the prognostic and predictive value of thymidylate synthase, p53, and Ki-67 in patients with locally advanced colon cancer. J Clin Oncol 2002; 20: 1735–1743. 41. Tomiak A, Vincent M, Earle CC et al. Thymidylate synthase expression in stage II and III colon cancer: a retrospective review. Am J Clin Oncol 2001; 24: 597–602. 42. Danenberg PV. Thymidylate synthetase—a target enzyme in cancer chemotherapy. Biochim Biophys Acta 1977; 473: 73–92. 43. Ju J, Pedersen-Lane J, Maley F et al. Regulation of p53 expression by thymidylate synthase. Proc Natl Acad Sci USA 1999; 96: 3769–3774. 44. Chu E, Copur SM, Ju J et al. Thymidylate synthase protein and p53 mRNA form an in vivo ribonucleoprotein complex. Mol Cell Biol 1999; 19: 1582–1594. 45. Allegra CJ, Paik S, Colangelo LH et al. Prognostic value of thymidylate synthase, Ki-67, and p53 in patients with Dukes’ B and C colon cancer: a National Cancer Institute-National Surgical Adjuvant Breast and Bowel Project collaborative study. J Clin Oncol 2003; 21: 241–250. 46. Lenz HJ, Danenberg KD, Leichman CG et al. p53 and thymidylate synthase expression in untreated stage II colon cancer: associations with recurrence, survival, and site. Clin Cancer Res 1998; 4: 1227–1234. 47. Westra JL, Hollema H, Schaapveld M et al. Predictive value of thymidylate synthase and dihydropyrimidine dehydrogenase protein expression on survival in adjuvantly treated stage III colon cancer patients. Ann Oncol 2005; 16: 1646–1653. 48. Berglund A, Edler D, Molin D et al. Thymidylate synthase and p53 expression in primary tumor do not predict chemotherapy outcome in metastatic colorectal carcinoma. Anticancer Res 2002; 22: 3653–3659. 49. Slebos RJ, Baas IO, Clement M et al. Clinical and pathological associations with p53 tumour-suppressor gene mutations and expression of p21WAF1/Cip1 in colorectal carcinoma. Br J Cancer 1996; 74: 165–171. 50. Aschele C, Debernardis D, Tunesi G et al. Thymidylate synthase protein expression in primary colorectal cancer compared with the corresponding distant metastases and relationship with the clinical response to 5-fluorouracil. Clin Cancer Res 2000; 6: 4797–4802. 51. Parker WB, Cheng YC. Metabolism and mechanism of action of 5-fluorouracil. Pharmacol Ther 1990; 48: 381–395. 52. Geoffroy FJ, Allegra CJ, Sinha B et al. Enhanced cytotoxicity with interleukin-1 alpha and 5-fluorouracil in HCT116 colon cancer cells. Oncol Res 1994; 6: 581–591. 53. Aschele C, Sobrero A, Faderan MA et al. Novel mechanism(s) of resistance to 5-fluorouracil in human colon cancer (HCT-8) sublines following exposure to two different clinically relevant dose schedules. Cancer Res 1992; 52: 1855–1864. 54. Marsh S, Collie-Duguid ES, Li T et al. Ethnic variation in the thymidylate synthase enhancer region polymorphism among Caucasian and Asian populations. Genomics 1999; 58: 310–312. 55. Mandola MV, Stoehlmacher J, Muller-Weeks S et al. A novel single nucleotide polymorphism within the 59 tandem repeat polymorphism of the thymidylate synthase gene abolishes USF-1 binding and alters transcriptional activity. Cancer Res 2003; 63: 2898–2904. 56. Kawakami K, Watanabe G. Identification and functional analysis of single nucleotide polymorphism in the tandem repeat sequence of thymidylate synthase gene. Cancer Res 2003; 63: 6004–6007. 57. Popat S, Wort R, Houlston RS. Relationship between thymidylate synthase (TS) genotype and TS expression: a tissue microarray analysis of colorectal cancers. Int J Surg Pathol 2004; 13: 127–133. 58. Russo A, Bazan V, Iacopetta B et al. The TP53 colorectal cancer international collaborative study on the prognostic and predictive significance of p53 mutation: influence of tumor site, type of mutation, and adjuvant treatment. J Clin Oncol 2005; 23: 7518–7528.
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
14. Petersen S, Thames HD, Nieder C et al. The results of colorectal cancer treatment by p53 status: treatment-specific overview. Dis Colon Rectum 2001; 44: 322–333. 15. Munro AJ, Lain S, Lane DP. P53 abnormalities and outcomes in colorectal cancer: a systematic review. Br J Cancer 2005; 92: 434–444. 16. Santi DV, McHenry CS, Sommer H. Mechanism of interaction of thymidylate synthetase with 5-fluorodeoxyuridylate. Biochemistry 1974; 13: 471–481. 17. Longley DB, Harkin DP, Johnston PG. 5-Fluorouracil: mechanisms of action and clinical strategies. Nat Rev Cancer 2003; 3: 330–338. 18. Berger SH, Jenh CH, Johnson LF et al. Thymidylate synthase overproduction and gene amplification in fluorodeoxyuridine-resistant human cells. Mol Pharmacol 1985; 28: 461–467. 19. Lane DP, Benchimol S. p53: oncogene or anti-oncogene? Genes Dev 1990; 4: 1–8. 20. Rodrigues NR, Rowan A, Smith ME et al. p53 mutations in colorectal cancer. Proc Natl Acad Sci USA 1990; 87: 7555–7559. 21. Bartek J, Bartkova J, Vojtesek B et al. Aberrant expression of the p53 oncoprotein is a common feature of a wide spectrum of human malignancies. Oncogene 1991; 6: 1699–1703. 22. Levine AJ, Momand J, Finlay CA. The p53 tumour suppressor gene. Nature 1991; 351: 453–456. 23. Vogelstein B, Fearon ER, Hamilton SR et al. Genetic alterations during colorectaltumor development. N Engl J Med 1988; 319: 525–532. 24. Finlay CA, Hinds PW, Tan TH et al. Activating mutations for transformation by p53 produce a gene product that forms an hsc70-p53 complex with an altered halflife. Mol Cell Biol 1988; 8: 531–539. 25. Bunz F, Hwang PM, Torrance C et al. Disruption of p53 in human cancer cells alters the responses to therapeutic agents. J Clin Invest 1999; 104: 263– 269. 26. Longley DB, Boyer J, Allen WL et al. The role of thymidylate synthase induction in modulating p53-regulated gene expression in response to 5-fluorouracil and antifolates. Cancer Res 2002; 62: 2644–2649. 27. Longley DB, Allen WL, McDermott U et al. The roles of thymidylate synthase and p53 in regulating Fas-mediated apoptosis in response to antimetabolites. Clin Cancer Res 2004; 10: 3562–3571. 28. Piedbois P, Buyse M. Meta-analyses based on abstracted data: a step in the right direction, but only a first step. J Clin Oncol 2004; 22: 3852–3859. 29. McShane LM, Altman DG, Sauerbrei W et al. Reporting recommendations for tumor marker prognostic studies (REMARK). J Natl Cancer Inst 2005; 97: 1180– 1184. 30. Chen ZM, Shao YF. Rationale, design and organization of the UK-China collaborative colorectal cancer trial of 5-FU intra-portal infusion. Chin J Oncol 1994; 6: 473–476. 31. Chen ZM, Shao YF. The UK-China collaborative colorectal cancer trial of 5-FU intra-portal infusion: progress and baseline characteristics of the first 3000 patients randomised. Chin J Oncol 1995; 6: 467–469. 32. Hsu SM, Raine L, Fanger H. The use of antiavidin antibody and avidin-biotinperoxidase complex in immunoperoxidase technics. Am J Clin Pathol 1981; 75: 816–821. 33. Findlay MP, Cunningham D, Morgan G et al. Lack of correlation between thymidylate synthase levels in primary colorectal tumours and subsequent response to chemotherapy. Br J Cancer 1997; 75: 903–909. 34. Van der Wilt CL, Smid K, Aherne GW et al. Evaluation of immunohistochemical staining and activity of thymidylate synthase in cell lines. Adv Exp Med Biol 1993; 338: 605–608. 35. Wang JP, Yang ZL, Dong WG et al. Multi-variate regression analysis of clinicopathological characteristics and prognosis of colorectal cancer. Chin J Oncol 2003; 25: 59–61. 36. Zhao DB, Gao JD, Shan Y et al. Characteristics of metastasis and recurrence following curative resection for colonic carcinoma. Chin J Gastrointest Surg 2006; 9: 10–13. 37. Ohrling K, Edler D, Hallstrom M et al. Detection of thymidylate synthase expression in lymph node metastases of colorectal cancer can improve the prognostic information. J Clin Oncol 2005; 23: 5628–5634.
Annals of Oncology
Annals of Oncology
59. Iacopetta B, Russo A, Bazan V et al. Functional categories of TP53 mutation in colorectal cancer: results of an international collaborative study. Ann Oncol 2006; 17: 842–847. 60. Soong R, Robbins PD, Dix BR et al. Concordance between p53 protein overexpression and gene mutation in a large series of common human carcinomas. Hum Pathol 1996; 27: 1050–1055.
original article 61. Liu Y, Bodmer WF. From the cover: analysis of P53 mutations and their expression in 56 colorectal cancer cell lines. Proc Natl Acad Sci USA 2006; 103: 976–981. 62. Resnick MA, Inga A. Functional mutants of the sequence-specific transcription factor p53 and implications for master genes of diversity. Proc Natl Acad Sci USA 2003; 100: 9934–9939.
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
Volume 17 | No. 12 | December 2006
doi:10.1093/annonc/mdl301 | 1817
Annals of Oncology 17: 1810–1817, 2006 doi:10.1093/annonc/mdl301 Published online 13 September 2006
A prospective, blinded analysis of thymidylate synthase and p53 expression as prognostic markers in the adjuvant treatment of colorectal cancer S. Popat1*, Z. Chen2, D. Zhao3, H. Pan2, N. Hearle1, I. Chandler1, Y. Shao3, W. Aherne4 & R. S. Houlston1
Received 9 May 2006; revised 9 July 2006; accepted 10 July 2006
Background: Despite previous studies, uncertainty has persisted about the role of thymidylate synthase (TS) and p53 status as markers of prognosis in colorectal cancer (CRC).
original article
Patients and methods: A total of 967 patients accrued to a large adjuvant trial in CRC were included in a prospectively planned molecular substudy, and of them, 59% had rectal cancer and about 90% received adjuvant chemotherapy (either systemically or randomly allocated to intraportal 5-fluorouracil infusion or both). TS and p53 status were determined, blinded to any clinical data, by immunohistochemistry using a validated polyclonal antibody or the DO-7 clone, respectively, and their relationships with overall survival were examined. Results: High TS expression was observed in 58% and overexpression of p53 in 60% of tumours. TS expression correlated with tumour stage, and p53 overexpression, with rectal cancers. There was no evidence that either marker was significantly associated with survival by either univariate (TS hazard ratio (HR) = 0.94, 95% CI 0.76–1.18 and P = 0.6 and p53 HR = 0.98, 95% CI 0.78–1.23 and P = 0.9) or multivariate analyses (TS HR = 0.99, 95% CI 0.79–1.25 and P = 0.9 and p53 HR = 0.98, 95% CI 0.78–1.23 and P = 0.8). Conclusions: Neither TS nor p53 expression has significant prognostic value in the adjuvant setting of CRC. Key words: colorectal cancer, p53, prognosis, thymidylate synthase
introduction Colorectal cancer (CRC) is one of the commonest cancers, affecting around 700 000 individuals each year worldwide [1] and in many Western countries such as United States it is the commonest cause of cancer death in nonsmoking men [2]. The natural history of CRCs are not necessarily similar, and tumour molecular profile is likely to play an important role in determining the prognosis for individual patients [3, 4]. Adjuvant chemotherapy is the standard of care for stage III CRC [5–8], but its role in stage II CRC is limited to those at greater risk of relapse as defined by histopathological criteria [9, 10]. Despite curative surgery, however, many patients eventually relapse even following adjuvant therapy [11], while many of those who currently receive adjuvant therapy might never have relapsed. The identification of robust molecular markers to supplement standard clinical and pathological staging systems is therefore of particular relevance for patients with CRC, especially in the light of recently introduced drugs against CRC [8, 12]. *Correspondence to: Dr S. Popat, Section of Cancer Genetics, Institute of Cancer Research, Sutton, SM2 5NG, UK. Tel: +44-(0)-208-642-6011; Fax: +44-(0)-208-643-0373. E-mail: [email protected]
ª 2006 European Society for Medical Oncology
Two molecular markers that have been the subject of much investigation are expression level of the enzyme thymidylate synthase (TS) and mutation status of TP53 [13–15]. TS is a ratelimiting enzyme involved in DNA synthesis [16], and competitive inhibition of its activity appears to be the major mechanism for the antitumour effect of fluoropyrimidines [17], the cornerstone of therapy for CRC. In vitro data indicate that TS expression level is a determinant of fluoropyrimidine sensitivity [16, 18], suggesting that TS expression might determine tumour sensitivity in vivo [13, 16, 18]. The tumour suppressor TP53, is critically involved in the control of cell cycle and apoptosis [19–22] and is commonly mutated in CRC [23]. Mutation commonly results in expression of protein with abnormal conformation, characterised by an increased nuclear half-life, which is readily detected as p53 overexpression by immunohistochemistry [20, 24]. Several in vitro studies have reported a relationship between TP53 mutation status and sensitivity to a number of cytotoxic agents, including fluoropyrimidines [25–27]. Several studies have previously investigated the prognostic effects of TS and p53 expression in CRC. Most of these studies were relatively small and mainly involved retrospectively
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
1 Section of Cancer Genetics, Institute of Cancer Research, Sutton, UK; 2Clinical Trial Service Unit, Richard Doll Building, Oxford, UK; 3Department of Abdominal Surgery, Tumour Hospital, Chinese Academy of Medical Sciences, Beijing, People’s Republic of China; 4Cancer Research UK Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, UK
original article
Annals of Oncology
methods patients and samples Details about the study design, objectives and patient population of the original clinical trial have previously been reported [30]. Briefly, the trial was set up in 1994 to assess the effects of a 7-day continuous intraportal vein infusion of 5-fluorouracil (5-FU), started immediately after tumour resection, on long-term survival in patients with stage I–III CRC [30, 31]. Two hundred hospitals throughout China participated in the trial, and >10 000 patients had been entered in a randomised trial from 1994 to 1998 with the current cycle of follow-up on survival continuing to the middle of June 2003 (median 6 years, in survivors). Of the collaborating hospitals, 39 agreed to participate in a substudy of molecular markers that involved collection of a pair of paraffin-embedded, formalin-fixed tissue blocks (one tumour tissue and one normal tissue) for each randomly assigned patient. All required approvals for the main trial and the molecular substudy were obtained from relevant organisations and ethical committees in China, and informed consent was obtained from individual patients before the investigation. Overall, a total of 967 sets of samples were collected prospectively during the course of the trial. Following official approval by the Chinese Administration for Human Genetic Materials samples, all pathological specimens were sent to the UK Clinical Trial Service Unit (CTSU) in Oxford, where they were anonymised before forwarding to the Section of Cancer Genetics for laboratory analysis.
immunohistochemistry methods Three-micron sections from a representative part of the primary tumour were cut onto silane-coated glass slides and assessed for TS and p53 expression by the avidin–biotin complex immunohistochemical technique (Vectastain Elite ABC kit; Vector Laboratories Inc., Burlingame, CA) [32]. Negative and positive control slides were included in each staining run. The
Volume 17 | No. 12 | December 2006
negative control consisted of replacing primary antibody with phosphatebuffered saline [pH 7.6/0.1% (by vol) Tween solution] whereas positive controls were sections of tumour known to stain heavily for TS or p53. Tumour sections were deparaffinised in Histoclear (National Diagnostics, Hull, UK) and hydrated in decreasing concentrations of ethanol. Endogenous peroxidase activity was quenched with 5% hydrogen peroxide in methanol for 20 min. For p53 (but not TS) immunohistochemistry, antigen retrieval was performed using a microwave oven-based method. Specifically, sections were incubated in boiling 10 mmol/l citric acid buffer (pH 6.0) for 10 min and then cooled in running water. Nonspecific background staining was blocked with 20% goat serum for 20 min. Sections were then incubated with appropriate primary antibodies at a 1:100 dilution, in a humidified chamber at room temperature for 60 min, using either a validated rabbit polyclonal antibody to recombinant human TS [33, 34] or an anti-p53 mouse monoclonal antibody (clone DO-7; Dako Corp., Glostrup, Denmark). After rinsing, a biotinylated anti-rabbit or anti-mouse secondary antibody was applied for 30 min followed by further washing, and avidin–biotin–peroxidase complexes (Vectastain Elite ABC kit; Vector Laboratories Inc.). Immunostaining was developed by applying freshly prepared 0.05% 3,39-diaminobenzidine tetrahydrochloride (Vectastain Elite ABC kit; Vector Laboratories Inc.). Sections were counterstained in Mayer’s Haematoxylin (Sigma Chemical Co., St Louis, MO), dehydrated in a series of ethanols, cleared in Histoclear (National Diagnostics) and mounted with glass coverslips using DePeX (BDH, Poole, UK).
immunohistochemistry evaluation All slides were randomly allocated to independent assessment by two observers blinded to clinical information. TS expression was categorised semi-quantatively into four groups (0–3) based on chromagen intensity, with the highest tumour staining detected being used as the reference for grading. Grades 0 and 1, representing none and minimal staining, respectively, were defined as the ‘low’ expression group, whereas grades 2 and 3 were defined as the ‘high’ expression group. p53 immunoreactivity was dichotomised into positive or negative based on staining of malignant nuclei, with a threshold of 10%. Thresholds to assign marker status were those currently established and used by the vast majority of researchers in this field [13, 14], thereby allowing a comparative assessment of our dataset. Level of scoring agreement between the two observers was recorded and was in excess of 85% (see Results section). In cases of disagreement, marker status was discussed and determined by consensus. All immunohistochemistry and results scoring was performed blinded to clinical outcome and recorded on a separate database before merging with the main database kept in Oxford CTSU for the final data analysis.
statistical analysis Differences in demographic data and between different categories of markers were tested for significance using the v2 test for categorical variables and the Student’s t-test or Mann–Whitney nonparametric test for continuous variables. The level of agreement between assessors was assessed using the j statistic. A Cox regression model was used with individual marker as the exposure variables and OS (from time of surgery to time of death or end of current follow-up) as the outcome. The analyses were adjusted simultaneously for sex, age, tumour size, grade (World Health Organization), stage and sites as well as use of post-operative adjuvant therapies. The cumulative probability of survival was calculated by the Kaplan–Meier method and groups were compared by the stratified log-rank test. Information on progression-free survival was not available for all randomly assigned patients, nor were most histopathological features, e.g. T-stage, and number of lymph nodes resected at time of surgery. For all patients randomized to the clinical study, information was available neither for progression-free survival nor for most histopathological features, e.g. T-stage, and number of lymph nodes resected at time of surgery. All P values were two-sided, with
doi:10.1093/annonc/mdl301 | 1811
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
collected samples and clinical data with unblinded analytic methods, and the results have been inconsistent [13, 14]. In addition, meta-analyses published on the prognostic utility of both markers [13, 14] were not based on individual patient data, and results may have been biased by a variety of factors [28]. To help address such deficiencies, we present results from a large prospectively designed molecular study of about 1000 earlystage CRC patients, accrued into a multi-centre randomised trial of adjuvant therapy. The prespecified hypotheses tested were that TS expression level and p53 expression status are markers of overall survival (OS) in potentially curatively resected CRC, and the results are reported as per the Reporting Recommendations for Tumour Marker Prognastic Studies (REMARK) guidelines [29]. These guidelines were recently published as a standard for reporting tumour marker prognostic studies [29], as a major recommendation of the National Cancer Institute–European Organisation for Research and Treatment of Cancer. Using this standard, relevant information on study design, preplanned hypotheses, patient and specimen characteristics, assay methods and statistical analysis methods are specified, to enable transparent and complete reporting so that the relevant information be available to other researchers in the field to assess the usefulness of the data and understand the context of the conclusions. These guidelines complement the prospective, blinded, large patient population design of this study and together address many of the inadequacies of previously reported studies in this field.
original article
Annals of Oncology
patient demographics Among the 967 patients with samples available for investigation of TS and p53 expression, the overall characteristics were similar to those in the main trial (Table 1). The mean age was 55 years, with 45% aged below 55 years and 53% were male. There were more rectal (60%) than colon cancers (40%), and most (57%) had stage II disease. Neoadjuvant radiotherapy was hardly given, but about 80% of patients received adjuvant systemic chemotherapy (chiefly 5-FU based) on top of randomly allocated 7-day intraportal 5-FU infusion. So, overall in this
TS and p53 expression Of the 967 samples tested, 953 and 951 were assessable for TS and p53 expression, respectively, with the remainder nonevaluable due to either insufficient tumour present or nonrepresentative specimens. For TS analysis, a further 174 (18%) samples were nonevaluable due to repeatedly poor staining, resulting in 779 samples with available TS data (Table 2). Of these, 58% (449) expressed high TS levels (grade 3, 22% and grade 2, 36%), with the remainder
Table 1. Comparison of baseline characteristics between those with tumour samples collected and those without in the original randomised trial Characteristic Sex Male Female Age at diagnosis (years) <45 45–54 55–64 >65 Tumour site Colon Rectum Others Tumour size (cm) <4.0 4.0–5.9 6.0–7.9 >8.0 Stage I II III Tumour differentiation Well/moderate Poor Neoadjuvant radiotherapy Given Not given Nontrial adjuvant chemotherapy Given Not given Randomised treatment Intraportal 5-FU Control
Tumours assessed (%)
Tumours not assessed (%)
509 (53) 458 (47)
5358 (54) 4601 (46)
235 201 303 228
2555 2156 3110 2138
(24) (21) (31) (24)
378 (39) 573 (59) 16 (2)
3545 (36) 6293 (63) 121 (1)
207 394 203 163
1967 4588 2091 1313
(21) (41) (21) (17)
(20) (46) (21) (13)
76 (8) 555 (57) 336 (35)
883 (9) 5445 (55) 3631 (36)
696 (87) 103 (13)
NA (–) NA (–)
8 (1) 959 (99)
141 (1) 9818 (99)
766 (79) 201 (21)
7989 (80) 1970 (20)
471 (49) 496 (51)
4979 (50) 4980 (50)
5-FU, 5-fluorouracil and NA, not ascertained.
1812 | Popat et al.
(26) (22) (31) (21)
Table 2. Patient characteristics according to marker expression status Characteristic
Sex Male Female Age at diagnosis (years) <45 45–54 55–64 >65 Tumour site Colon Rectum Others Tumour size (cm) <4.0 4.0–5.9 6.0–7.9 >8.0 Stage I II III Tumour differentiation Well/moderate Poor Neoadjuvant radiotherapy Given Not given Nontrial adjuvant chemotherapy Given Not given Randomised treatment Portal 5-FU Control
TS status No. with high expression (%)
p53 status No. positive (%)
246 (60) 203 (55)
243 (61) 210 (59)
113 96 121 119
108 93 153 99
(58) (59) (51) (65)
(57) (59) (67) (55)
182 (61) 260 (56) 7 (47)
151 (52)* 294 (65) 8 (53)
96 185 98 70
93 191 101 68
(60) (58) (59) (52)
(59) (63) (63) (50)
39 (74)* 259 (59) 151 (53)
35 (66) 248 (58) 170 (61)
399 (59) 48 (49)
395 (60) 56 (58)
2 (29) 447 (58)
5 (71) 448 (60)
365 (58) 84 (57)
377 (61) 76 (55)
213 (56) 236 (59)
219 (60) 234 (60)
TS, thymidylate synthase and 5-FU, 5-fluorouracil. *P < 0.05.
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
results
patient population, 10% received intraportal 5-FU alone, 39% received both intraportal 5-FU and adjuvant (nontrial) 5-FU, and 10% received neither intraportal 5-FU nor adjuvant chemotherapy. The 5-year survival rate for patients investigated was 57.8% [standard error (SE) 1.7%] and was similar to that observed for all patients in the main trial (59.5%, SE 0.5%). Efficacy data from the main trial are yet to be reported.
values <0.05 regarded as conventionally statistically significant. Assuming a control survival rate of 60% and 50% of patients with high TS expression or p53 overexpression [13, 14], then analysis of tissue samples from 750 patients will have 80% power to detect an absolute difference of 10% in OS associated with the expression of either of these markers.
original article
Annals of Oncology
having low level of expression (grade 0, 3% and grade 1, 39%; Figure 1). For p53 status, a further 196 (20%) samples were nonevaluable due to either repeatedly poor staining or tissue destruction at antigen retrieval, resulting in available data from 755 (78% of the total) patients for the present analysis (Table 2). Of these samples, just over half demonstrated p53 overexpression (453, 60%; Figure 1). Level of agreement between observers for both TS and p53 expression was high (j > 0.94). Of the 754 patients with both TS and p53 results available (78% of the total), a significant relationship between TS and p53 status was observed (P = 0.0006), with tumours expressing high levels of TS more likely to overexpress p53.
discussion The present study is one of the largest investigations of the prognostic value of molecular markers in CRC reported to date.
Figure 1. Immunohistochemical analysis of tumours for thymidylate synthase (TS) and p53. Tumour showing (A) high-level TS expression, (B) low-level TS expression, (C) p53 overexpression and (D) no p53 overexpression.
Volume 17 | No. 12 | December 2006
doi:10.1093/annonc/mdl301 | 1813
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
marker status and clinical correlates Significant associations were observed between marker status and several clinical variables (Table 2). TS status was associated with stage (P = 0.02), with tumours less likely to express high TS levels if advanced at resection. These associations were not observed for p53 status, where marker status was instead associated with site (rectal tumours more frequently being p53 positive, P < 0.001). There was no interaction between markers and tumour grade. At the end of the current cycle of follow-up, 400 (41%) patients had died, mostly from disease recurrence, with median follow-up of 51 months (0–100). OS was significantly associated with CRC stage (HR = 1.68, 95% CI 1.37–2.06 and P < 0.0001), and marginally significantly with grade (HR = 1.35, 95% CI 0.99–1.84 and P = 0.06), after correction for co-variates, with 5-year survival rates of 61.5% (SE 2.2%) and 46.4% (SE 2.8%) for stage II and III CRC, and 61.7% (SE 1.9%) and 45.5% (SE 5.1%) for well/moderately and poorly differentiated disease, respectively. This survival rate was
consistent with that previously reported from China [35, 36]. No statistically significant difference in OS was observed between tumours with or without high TS expression, with a 5-year survival rate of 59.0% (SE 2.9%) in patients with high TS levels compared with 56.8% (SE 3.2%) in those with low TS levels (HR = 0.94, 95% CI 0.76–1.18 and P = 0.6; Figure 2). Similarly, no statistically significant difference in OS was observed between the two different p53 groups, with a 5-year survival rate of 58.0% (SE 2.9%) in patients with p53-positive tumours compared with 57.8% (SE 3.4%) in p53-negative tumours (HR = 0.98, 95% CI 0.78–1.23 and P = 0.9; Figure 3). The effect of TS and p53 expression in patients treated by surgery alone was not specifically assessed due to small number of patients involved. When all standard prognostic clinical variables were included as co-variables in a Cox proportional hazards model, there was again no evidence that these two markers were significantly associated with OS (HR = 0.99, 95% CI 0.79–1.25 and P = 0.9 for TS and HR = 0.98, 95% CI 0.78–1.23 and P = 0.8 for p53). When restricting analysis to a subset of patients with stage III rectal cancers (n = 167), however, there was an indication that high TS levels were associated with a poor OS (HR = 1.55, 95% CI 1.01– 2.38 and P = 0.04), whereas for p53 there was also a borderline significant relationship with OS (HR = 1.58, 95% CI 0.99–2.50 and P = 0.05). No significant relationship was found with OS in the stage I–II rectal cohorts for either marker (TS HR = 0.87, 95% CI 0.60–1.26 and P = 0.5 and p53 HR = 0.82, 95% CI 0.55–1.22 and P = 0.3). These subsets were, however, analysed in a post hoc manner, and should therefore be interpreted with caution.
original article
Figure 3. Overall survival in all patients stratified by p53 status.
Despite the use of prospective study design and blinded laboratory assessment of TS and p53 status to minimise the potential biases seen in previous such studies [29], the present study provides little evidence that TS or p53 overexpression is of any significant prognostic value in the adjuvant setting of CRC.
1814 | Popat et al.
Although it involved a Chinese patient population that has not been extensively studied previously, these results were generally consistent with many, though not all, previous studies conducted in Western populations [37–41]. While there is some suggestion that TS and p53 might be associated with poor OS in a subset of patients with stage III rectal cancer, this result may be due chiefly or wholly to the play of chance, given the large number of subgroups examined, the small number of patients involved in that particular subset and the marginally statistical significance of the results. It is highly plausible that TS expression or p53 status may influence CRC prognosis, given that TS is a target for 5-FU [42] and also a regulator of p53 expression [43, 44]. There is also evidence that p53 itself could potentially modulate 5-FU resistance [17]. Several studies have previously reported on the relationships between TS expression, p53 status and survival in CRC but results have been inconsistent, with some showing a positive relationship [45, 46], while others show no such relationship [33, 40, 47–49]. In studies of patients with advanced CRC, high TS expression has been shown in some to be associated with poorer OS [13], whereas many other studies in the adjuvant setting have generally failed to document a relationship between levels of expression and OS, particularly in the setting of adjuvant 5-FU-based regimens [38–41]. It is unclear what accounts for the apparent discrepancy between observations in the advanced and adjuvant settings. While it is possible that the level of TS expression in primary resected tumour may not reflect its expression level in metastases [47, 50] or associated lymph nodes [37], it is equally possible that much of the difference is methodologically related. By contrast to the present study, most previous studies, especially those in advanced CRC, were very small typically involving only about 100 patients and tended to use retrospectively retrieved tissue samples and clinical data, as well as unblinded laboratory assays for molecular markers. Consequently, the results could be subject to large biases. By using a large prospectively designed study with central unblinded and established molecular assays, it is possible to minimise much of the problems commonly found in previous studies so as to provide an unbiased and reliable, though not necessarily conclusive, assessment of prognostic significance of TS and p53. There is suggestion from a pooled analysis of 584 patients that TS level may predict poor OS only in patients treated by surgery alone without adjuvant 5-FU [13]. If this were true, then it may suggest that adjuvant 5-FU could potentially abrogate a poorer prognosis ascribed by high TS expression, possibly a consequence of 5-FU scheduling. This hypothesis, however, cannot be tested directly in the present study, as nearly all patients received some form of 5-FU-based chemotherapy following surgery. In vitro data indicate that induction of high TS expression could potentially lead to 5-FU resistance in some types of tumour cells but is dependent on whether 5-FUmediated cytotoxicity in cells is mediated primarily by TS inhibition or by nucleotide misincorporation [51, 52]. The latter non-TS-directed effect may have been the more predominant mechanism of 5-FU action when administered as a bolus regimen [53]. Unfortunately, no detailed information was available in the present study regarding the regimen of adjuvant 5-FU administered, so that the question as to whether any
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
Figure 2. Overall survival in all patients stratified by thymidylate synthase status.
Annals of Oncology
original article
Annals of Oncology
Volume 17 | No. 12 | December 2006
In summary, we have utilised standard molecular methods to evaluate TS and p53 as prognostic markers in the largest patient cohort from a single clinical trial reported to date. Our findings provide little evidence to indicate that in the adjuvant setting, TS or p53 expression is predictive of OS, and these markers should not be used in routine clinical decision making. We cannot, however, exclude a role for TS in patients treated in the adjuvant setting by surgery alone and further studies, particularly in those with good-prognosis stage II CRC, are warranted. This will help better define the prognostic utility of this marker without the potential covariate of adjuvant chemotherapy.
acknowledgements This work was supported by grants from the Cancer Research UK and the Association for International Cancer Research. SP was in receipt of a Clinician Scientist Fellowship from the UK Department of Health and IC a Clinical Training Fellowship from St George’s Hospital Medical School. The UK Medical Research Council, the British Heart Foundation and Cancer Research UK provided core funding to the CTSU.
references 1. Ferlay J, Bray F, Pisani P et al. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC CancerBase No. 5. Lyon, France: IARC Press 2001. 2. American Cancer Society. Cancer facts and figures 2004. Atlanta: American Cancer Society; 2004. 3. Kahlenberg MS, Sullivan JM, Witmer DD et al. Molecular prognostics in colorectal cancer. Surg Oncol 2003; 12: 173–186. 4. Johnston PG. Of what value genomics in colorectal cancer? Opportunities and challenges. J Clin Oncol 2004; 22: 1538–1539. 5. Kerr DJ, Gray R, McConkey C et al. Adjuvant chemotherapy with 5-fluorouracil, L-folinic acid and levamisole for patients with colorectal cancer: non-randomised comparison of weekly versus four-weekly schedules—less pain, same gain. QUASAR Colorectal Cancer Study Group. Ann Oncol 2000; 11: 947–955. 6. Efficacy of adjuvant fluorouracil and folinic acid in colon cancer. International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) investigators. Lancet 1995; 345: 939–944. 7. NIH Consensus Conference. Adjuvant therapy for patients with colon and rectal cancer. JAMA 1990; 264: 1444–1450. 8. Andre T, Boni C, Mounedji-Boudiaf L et al. Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. N Engl J Med 2004; 350: 2343–2351. 9. Figueredo A, Charette ML, Maroun J et al. Adjuvant therapy for stage II colon cancer: a systematic review from the Cancer Care Ontario Program in evidencebased care’s gastrointestinal cancer disease site group. J Clin Oncol 2004; 22: 3395–3407. 10. Benson AB III, Schrag D, Somerfield MR et al. American Society of Clinical Oncology recommendations on adjuvant chemotherapy for stage II colon cancer. J Clin Oncol 2004; 22: 3408–3419. 11. Moertel CG, Fleming TR, Macdonald JS et al. Fluorouracil plus levamisole as effective adjuvant therapy after resection of stage III colon carcinoma: a final report. Ann Intern Med 1995; 122: 321–326. 12. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: 2335–2342. 13. Popat S, Matakidou A, Houlston RS. Thymidylate synthase expression and prognosis in colorectal cancer: a systematic review and meta-analysis. J Clin Oncol 2004; 22: 529–536.
doi:10.1093/annonc/mdl301 | 1815
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
differences in this aspect of patient management may have biased outcome cannot be investigated directly. Most previous molecular analyses have been conducted in Western populations, and there is little direct information about the molecular features of CRC in populations such as China, where disease incidence is still much lower than that in Western Europe and North America. The present study in China showed that, although there is a difference between Chinese and Western patients in the main types of CRC observed (more rectal rather than colon cancer in China), the proportion of patients with high levels of TS expression was generally similar to that in other reported series in the Western populations [13], as was the proportion of tumours overexpressing p53 [14]. Although the frequency of p53 overexpression in this dataset was marginally higher than that assumed for power calculations, this difference did not impact on the power of our study. Previous reports have indicated an increased incidence of the 3R/3R 28-base-pair variable number of tandem repeat (VNTR) polymorphism in Chinese patients [54], which could result in a greater proportion of high TS expression. Although TS genotype was not specifically assessed, the lack of overrepresentation by tumours expressing high TS levels may be accounted for by either an unknown prevalence of the functional G/C single-nucleotide polymorphism within the third VNTR tandem repeat [55, 56] or a complex relationship between genotype and expression when measured by immunohistochemistry [57]. By comparison with TS expression, the prognostic value of p53 status has been less well established. Although a number of studies have reported a poorer OS in patients with TP53 mutations or p53 overexpression, there has been considerable variation in findings [14, 15], due perhaps in part to the same methodological problems as in studies of TS (nonblinded investigators, small samples sizes and retrospective analyses) and in part to the method of ascertaining p53 status. A number of studies have investigated TP53 directly by mutational analysis, while others have adopted indirect methods on the basis of either immunohistochemistry or allelic imbalance at chromosome 17p [14, 15, 58, 59], and each of them is also likely to be subject to large experimental variation. In particularly, exons screened for mutations have varied between studies, as have microsatellite markers genotyped. Although immunohistochemistry-based studies have used a variety of antibodies, the most commonly used one, as in the present study, is DO-7 [15]. Despite this and the large number of patients involved, the present study did not provide any supportive evidence that p53 overexpression is of prognostic significance in the adjuvant setting. In immunohistochemistry-based analyses, however, although p53 overexpression is used as a surrogate marker of TP53 mutation, it does not fully correlate with mutation status [60]. Mutation does not necessarily correlate with p53 expression [61] or nuclear aggregation. Discordance between TP53 sequence variation and expression status by immunohistochemistry has been observed in up to 30% of tumours [60] with some tumours completely lacking protein expression nonfunctional secondary to mutation [61], and other nontruncating functional sequence variants undetectable by immunohistochemistry [62]. This could therefore potentially obscure any real relationship between longterm survival prognosis and TP53 mutation.
original article
1816 | Popat et al.
38. Nanni O, Volpi A, Frassineti GL et al. Role of biological markers in the clinical outcome of colon cancer. Br J Cancer 2002; 87: 868–875. 39. Edler D, Glimelius B, Hallstrom M et al. Thymidylate synthase expression in colorectal cancer: a prognostic and predictive marker of benefit from adjuvant fluorouracil-based chemotherapy. J Clin Oncol 2002; 20: 1721–1728. 40. Allegra CJ, Parr AL, Wold LE et al. Investigation of the prognostic and predictive value of thymidylate synthase, p53, and Ki-67 in patients with locally advanced colon cancer. J Clin Oncol 2002; 20: 1735–1743. 41. Tomiak A, Vincent M, Earle CC et al. Thymidylate synthase expression in stage II and III colon cancer: a retrospective review. Am J Clin Oncol 2001; 24: 597–602. 42. Danenberg PV. Thymidylate synthetase—a target enzyme in cancer chemotherapy. Biochim Biophys Acta 1977; 473: 73–92. 43. Ju J, Pedersen-Lane J, Maley F et al. Regulation of p53 expression by thymidylate synthase. Proc Natl Acad Sci USA 1999; 96: 3769–3774. 44. Chu E, Copur SM, Ju J et al. Thymidylate synthase protein and p53 mRNA form an in vivo ribonucleoprotein complex. Mol Cell Biol 1999; 19: 1582–1594. 45. Allegra CJ, Paik S, Colangelo LH et al. Prognostic value of thymidylate synthase, Ki-67, and p53 in patients with Dukes’ B and C colon cancer: a National Cancer Institute-National Surgical Adjuvant Breast and Bowel Project collaborative study. J Clin Oncol 2003; 21: 241–250. 46. Lenz HJ, Danenberg KD, Leichman CG et al. p53 and thymidylate synthase expression in untreated stage II colon cancer: associations with recurrence, survival, and site. Clin Cancer Res 1998; 4: 1227–1234. 47. Westra JL, Hollema H, Schaapveld M et al. Predictive value of thymidylate synthase and dihydropyrimidine dehydrogenase protein expression on survival in adjuvantly treated stage III colon cancer patients. Ann Oncol 2005; 16: 1646–1653. 48. Berglund A, Edler D, Molin D et al. Thymidylate synthase and p53 expression in primary tumor do not predict chemotherapy outcome in metastatic colorectal carcinoma. Anticancer Res 2002; 22: 3653–3659. 49. Slebos RJ, Baas IO, Clement M et al. Clinical and pathological associations with p53 tumour-suppressor gene mutations and expression of p21WAF1/Cip1 in colorectal carcinoma. Br J Cancer 1996; 74: 165–171. 50. Aschele C, Debernardis D, Tunesi G et al. Thymidylate synthase protein expression in primary colorectal cancer compared with the corresponding distant metastases and relationship with the clinical response to 5-fluorouracil. Clin Cancer Res 2000; 6: 4797–4802. 51. Parker WB, Cheng YC. Metabolism and mechanism of action of 5-fluorouracil. Pharmacol Ther 1990; 48: 381–395. 52. Geoffroy FJ, Allegra CJ, Sinha B et al. Enhanced cytotoxicity with interleukin-1 alpha and 5-fluorouracil in HCT116 colon cancer cells. Oncol Res 1994; 6: 581–591. 53. Aschele C, Sobrero A, Faderan MA et al. Novel mechanism(s) of resistance to 5-fluorouracil in human colon cancer (HCT-8) sublines following exposure to two different clinically relevant dose schedules. Cancer Res 1992; 52: 1855–1864. 54. Marsh S, Collie-Duguid ES, Li T et al. Ethnic variation in the thymidylate synthase enhancer region polymorphism among Caucasian and Asian populations. Genomics 1999; 58: 310–312. 55. Mandola MV, Stoehlmacher J, Muller-Weeks S et al. A novel single nucleotide polymorphism within the 59 tandem repeat polymorphism of the thymidylate synthase gene abolishes USF-1 binding and alters transcriptional activity. Cancer Res 2003; 63: 2898–2904. 56. Kawakami K, Watanabe G. Identification and functional analysis of single nucleotide polymorphism in the tandem repeat sequence of thymidylate synthase gene. Cancer Res 2003; 63: 6004–6007. 57. Popat S, Wort R, Houlston RS. Relationship between thymidylate synthase (TS) genotype and TS expression: a tissue microarray analysis of colorectal cancers. Int J Surg Pathol 2004; 13: 127–133. 58. Russo A, Bazan V, Iacopetta B et al. The TP53 colorectal cancer international collaborative study on the prognostic and predictive significance of p53 mutation: influence of tumor site, type of mutation, and adjuvant treatment. J Clin Oncol 2005; 23: 7518–7528.
Volume 17 | No. 12 | December 2006
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
14. Petersen S, Thames HD, Nieder C et al. The results of colorectal cancer treatment by p53 status: treatment-specific overview. Dis Colon Rectum 2001; 44: 322–333. 15. Munro AJ, Lain S, Lane DP. P53 abnormalities and outcomes in colorectal cancer: a systematic review. Br J Cancer 2005; 92: 434–444. 16. Santi DV, McHenry CS, Sommer H. Mechanism of interaction of thymidylate synthetase with 5-fluorodeoxyuridylate. Biochemistry 1974; 13: 471–481. 17. Longley DB, Harkin DP, Johnston PG. 5-Fluorouracil: mechanisms of action and clinical strategies. Nat Rev Cancer 2003; 3: 330–338. 18. Berger SH, Jenh CH, Johnson LF et al. Thymidylate synthase overproduction and gene amplification in fluorodeoxyuridine-resistant human cells. Mol Pharmacol 1985; 28: 461–467. 19. Lane DP, Benchimol S. p53: oncogene or anti-oncogene? Genes Dev 1990; 4: 1–8. 20. Rodrigues NR, Rowan A, Smith ME et al. p53 mutations in colorectal cancer. Proc Natl Acad Sci USA 1990; 87: 7555–7559. 21. Bartek J, Bartkova J, Vojtesek B et al. Aberrant expression of the p53 oncoprotein is a common feature of a wide spectrum of human malignancies. Oncogene 1991; 6: 1699–1703. 22. Levine AJ, Momand J, Finlay CA. The p53 tumour suppressor gene. Nature 1991; 351: 453–456. 23. Vogelstein B, Fearon ER, Hamilton SR et al. Genetic alterations during colorectaltumor development. N Engl J Med 1988; 319: 525–532. 24. Finlay CA, Hinds PW, Tan TH et al. Activating mutations for transformation by p53 produce a gene product that forms an hsc70-p53 complex with an altered halflife. Mol Cell Biol 1988; 8: 531–539. 25. Bunz F, Hwang PM, Torrance C et al. Disruption of p53 in human cancer cells alters the responses to therapeutic agents. J Clin Invest 1999; 104: 263– 269. 26. Longley DB, Boyer J, Allen WL et al. The role of thymidylate synthase induction in modulating p53-regulated gene expression in response to 5-fluorouracil and antifolates. Cancer Res 2002; 62: 2644–2649. 27. Longley DB, Allen WL, McDermott U et al. The roles of thymidylate synthase and p53 in regulating Fas-mediated apoptosis in response to antimetabolites. Clin Cancer Res 2004; 10: 3562–3571. 28. Piedbois P, Buyse M. Meta-analyses based on abstracted data: a step in the right direction, but only a first step. J Clin Oncol 2004; 22: 3852–3859. 29. McShane LM, Altman DG, Sauerbrei W et al. Reporting recommendations for tumor marker prognostic studies (REMARK). J Natl Cancer Inst 2005; 97: 1180– 1184. 30. Chen ZM, Shao YF. Rationale, design and organization of the UK-China collaborative colorectal cancer trial of 5-FU intra-portal infusion. Chin J Oncol 1994; 6: 473–476. 31. Chen ZM, Shao YF. The UK-China collaborative colorectal cancer trial of 5-FU intra-portal infusion: progress and baseline characteristics of the first 3000 patients randomised. Chin J Oncol 1995; 6: 467–469. 32. Hsu SM, Raine L, Fanger H. The use of antiavidin antibody and avidin-biotinperoxidase complex in immunoperoxidase technics. Am J Clin Pathol 1981; 75: 816–821. 33. Findlay MP, Cunningham D, Morgan G et al. Lack of correlation between thymidylate synthase levels in primary colorectal tumours and subsequent response to chemotherapy. Br J Cancer 1997; 75: 903–909. 34. Van der Wilt CL, Smid K, Aherne GW et al. Evaluation of immunohistochemical staining and activity of thymidylate synthase in cell lines. Adv Exp Med Biol 1993; 338: 605–608. 35. Wang JP, Yang ZL, Dong WG et al. Multi-variate regression analysis of clinicopathological characteristics and prognosis of colorectal cancer. Chin J Oncol 2003; 25: 59–61. 36. Zhao DB, Gao JD, Shan Y et al. Characteristics of metastasis and recurrence following curative resection for colonic carcinoma. Chin J Gastrointest Surg 2006; 9: 10–13. 37. Ohrling K, Edler D, Hallstrom M et al. Detection of thymidylate synthase expression in lymph node metastases of colorectal cancer can improve the prognostic information. J Clin Oncol 2005; 23: 5628–5634.
Annals of Oncology
Annals of Oncology
59. Iacopetta B, Russo A, Bazan V et al. Functional categories of TP53 mutation in colorectal cancer: results of an international collaborative study. Ann Oncol 2006; 17: 842–847. 60. Soong R, Robbins PD, Dix BR et al. Concordance between p53 protein overexpression and gene mutation in a large series of common human carcinomas. Hum Pathol 1996; 27: 1050–1055.
original article 61. Liu Y, Bodmer WF. From the cover: analysis of P53 mutations and their expression in 56 colorectal cancer cell lines. Proc Natl Acad Sci USA 2006; 103: 976–981. 62. Resnick MA, Inga A. Functional mutants of the sequence-specific transcription factor p53 and implications for master genes of diversity. Proc Natl Acad Sci USA 2003; 100: 9934–9939.
Downloaded from http://annonc.oxfordjournals.org/ at New York Medical College on September 3, 2015
Volume 17 | No. 12 | December 2006
doi:10.1093/annonc/mdl301 | 1817