Journal of Surgical Oncology
Metachronous Colorectal Cancer: A Competing Risks Analysis with Consideration for a Stratified Approach to Surveillance Colonoscopy NICHOLAS J. BATTERSBY, MRCS, BMedSc,1 ALEX COUPLAND, MBBS,1 GEORGE BOULIOTIS, PhD(MedStat), MSc(Math),2 NAZZIA MIRZA, FRCS,1 AND J. GRAHAM WILLIAMS, MChFRCS1* 1 Colorectal Surgery, The Royal Wolverhampton Hospital, Wolverhampton, United Kingdom MRC-Midland Hub for Trials Methodology Research, University of Birmingham, Birmingham, United Kingdom
2
Background: The incidence of metachronous cancer will become an important clinical consideration as the life expectancy of the population increases and as rates of curative resection improve. Objective: To assess the pattern of metachronous cancer development following curative resection of colorectal cancer in an unselected patient population offered postoperative colonoscopic surveillance. Method: Prospective clinical follow‐up after curative colorectal cancer resection and surveillance colonoscopy with or without polypectomy in accordance with the national guidelines. Actuarial analysis and competing risk analysis were performed to account for death and recurrence and to stratify for age, gender, stage, and tumor site. Results: Five hundred thirty‐eight patients with median follow‐up 4 years 2 month (0–16) years. Fifteen patients (3%) developed metachronous cancer, at a median time interval of 90 months from primary resection. Thirteen metachronous cancer patients (87%, 13/15) underwent one to five surveillance colonoscopies: nine patients were asymptomatic at time of diagnosis of metachronous cancer. Competing risks analysis suggests that the adjusted cumulative incidence in males aged 55 is 4% at 10 years compared with 1% in females aged 85 years old. Conclusions: A patient aged under 65 at the time of the primary curative resection carries a 2% 5‐year risk of metachronous cancer, implying that 3 year surveillance colonoscopy is justified. Whereas patients aged over 75 carry less than a 2% 10‐year risk, implying that it is seldom warranted to repeat the colonoscopy more frequently than every 5 years. A stratified approach to the frequency of surveillance colonoscopy requires further consideration.
J. Surg. Oncol. . ß 2013 Wiley Periodicals, Inc.
KEY WORDS: metachronous; colorectal cancer; surveillance colonoscopy; competing risks analysis
INTRODUCTION Colorectal cancer is the third most common cancer diagnosed in the UK [1]. Patients who have had one colorectal cancer are at risk of developing additional, metachronous, colorectal cancers [2,3]. Metachronous colorectalcancers are tumors that develop at a site remote from the primary tumor, which are histologically separate and occur 12 months or more after surgery, in a completely investigated colon [4].This definition draws the distinction between synchronous tumors, present at the time of the primary resection, and from recurrence of the primary cancer. A number of case‐series with long‐term follow‐up have reported metachronous cancer rates of 0.7–9% [5–9]. In 1993, the national polyp study published evidence that colonoscopy and polypectomy enabled early detection of cancer and reduced the subsequent incidence of cancer; several studies have since supported these findings [10–12]. The primary aim of this study was to report the time dependent risk of metachronous cancer in an unselected population of patients who had previously undergone a curative colorectal cancer resection, in a unit, which applies current surveillance colonoscopy guidelines and takes a proactive approach to colonoscopy and polypectomy [13–15]. This is the first study reporting on metachronous cancer rate to use competing risks analysis in order to account for cancer recurrence and mortality.
METHODOLOGY Details of 572 patients who underwent colorectal cancer resection under the care of a single consultant colorectal surgeon (JGW) between
ß 2013 Wiley Periodicals, Inc.
May 1995 and February 2012 were prospectively recorded in a database that included patient demographics, AJCC tumor stage and site, operation performed, cancer recurrence, metachronous cancer, and mortality. Endoscopic data was also recorded. Patients with synchronous colorectal cancer, who underwent panproctocolectomy, and those who did not undergo resection of the primary cancer or who underwent a palliative operation were excluded. Data collection and analysis were performed according to the Hospital Trust ethics and clinical governance policy. Each patient underwent pre‐operative colonoscopy to identify synchronous lesions (or within 3 months of surgery where an obstructing lesion prevented complete investigation, or after emergency presentation). All colonoscopies were performed in a Joint Advisory Group (JAG) UK colonoscopy training center following bowel preparation using sodium picrosulphate or Fleet Phosphosoda1
Conflict of interest: none. This data was presented in part as a podium presentation at ASCRS 2012, San Antonio, Texas. *Correspondence to: J. Graham Williams, Colorectal Surgery, The Royal Wolverhampton Hospital, Wolverhampton WV10 0QP, United Kingdom. Fax: þ01902 695691. E‐mail: [email protected] Received 28 May 2013; Accepted 28 October 2013 DOI 10.1002/jso.23504 Published online in Wiley Online Library (wileyonlinelibrary.com).
2
Battersby et al.
(Lynchburg, Virginia 24502). The majority of examinations were performed by the senior author, or by a supervised (JGW) trainee. Follow‐up colonoscopy protocol was in accordance with current American Gastroenterological Association and the American Cancer Society guidelines [14,15]. The majority of patients were reviewed in the out‐patient department every 6 months for 3 years and annually up to 5 years. Where out‐patient follow‐up did not occur, further information on the patient’s status was obtained from hospital records and the patient’s General Practitioner.
Among various candidates, a modified Fine and Gray [18] model for metachronous events that adjust for cancer recurrence and death as competing events was used. Finally, model predictions were inputted into plausible clinical scenarios for estimating probabilistic differences among various groups of patients. Statistical significance was taken as P < 0.05 for all tests. Best‐fit model was selected based on the Akaike and Bayesian Information Criteria (AIC and BIC). Survival analysis was implemented with the Stata software [16,17].
Statistical Analysis
RESULTS
In the first part of the analysis, non‐parametric tests for assessing univariate differences related to patients’ characteristics (Chi‐squared or Fisher’s exact test). Among various covariates we generated the polypectomy rate as the total number of polyps divided by the total number of procedures. We also categorized age into four groups (less than 55, 55–64, 65–74, 75 plus). Finally, we grouped the location of cancer into three groups (rectum, right colon, and other) in order to limit strong imbalances among the groups. In the second part, we analyzed patient survival from the date of diagnosis of the index cancer, focusing on the occurrence of a metachronous cancer. Survival analysis included an additional step that of competing risks model [16,17]. Although semiparametric ordinary Cox regression could be extended to accommodate competing risks, we evaluated factors for modeled time varying covariates; as such a flexible parametric model was used in order to provide the most precise data fit.
Demographics Of the 572 patients undergoing resection of colorectal cancer, 34 patients were excluded from analysis (panproctocolectomy—3, synchronous cancer—3, colonic stent—1, and unresectable disease— 27). The remaining 538 patients had a median age of 70 years 11 months at the time of surgery; 295 (56%) were male. The AJCC stage, site of the index cancer and the operations performed are summarized in Table I.
Metachronous Cancers During follow‐up, 15 patients developed a metachronous cancer, giving a crude rate of 2.7% (15/538). The risk of metachronous cancer increased with time; the cumulative incidence was 1.8%, 4%, and 8% at 5, 10, and 12 years, respectively (Fig. 1). The median time interval
TABLE I. Demographic Data Metachronous patients, (n ¼ 15) Variables Age (IQ range)
Gender Colorectal tumor site (n ¼ 538) Cecum and ascending colon Transverse Descending Sigmoid Rectosigmoid Rectum AJCC stage (n ¼ 538) I II III Unknown Operation (n ¼ 538) AP resection Low anterior resection High anterior resection Left hemicolectomy Right hemicolectomy Hartmann’s procedure Transverse or extended right hemicolectomy Subtotal colectomy Pelvic exenteration Sigmoid resection Transanal resection of cancer Proctocolectomy
Primary cancer, (n ¼ 523)a 70 years 11 months (61 years 10 months– 77 years 11 months) 228 Female (44%) 295 Male (56%)
67 years 2 months (59 years 9 months– 75 years 10 months) 7 Female (47%) 8 Male (53%)
Metachronous 75 years 10 months (65years 11months– 81years 9months)
143 (27%) 26 (5%) 83 (16%) 31 (6%) 78 (15%) 162 (31%)
5 (33%) 0 0 4 (27%) 1 (6.7%) 5 (33%)
3 (20%) 3 (20%) 1 (6.7%) 2 (13%) 0 6 (40%)
95 (18%) 213 (41%) 207 (40%) 8 (1%)
4 (27%) 7 (47%) 4 (27%)
7 (47%) 5 (33%) 3 (20%)
41 (8%) 111 (21%) 77 (15%) 78 (15%) 144 (28%) 26 (5%) 26 (5%)
1 (6.7%) 4 (27%) 2 (13%) 4 (27%) 4 (27%) 0 0
1 (6.7%)
9 (2%) 3 (0.5%) 2 (0.5%) 6 (1%) —
0 0 0 0 —
2 (13%)
No significant difference between the groups was identified. a Patients who underwent a curative resection and did not develop a metachronous cancer.
Journal of Surgical Oncology
Index
1 (6.7%) 3 (20%) 1 (6.7%) 3 (20%)
4 (27%)
Metachronous Colorectal Cancer
3
Of the remaining six patients who developed a metachronous cancer: two presented with a change in bowel habit; two as emergencies with large bowel obstruction from a cecal and sigmoid tumor; one patient declined colonoscopic follow‐up and later presented with profuse bleeding and one patient had a palpable rectal polyp found in the outpatient clinic (Fig. 2).
Colonoscopic Follow‐up Follow‐up colonoscopy was performed on 361 patients. Of the 175 patients who did not undergo colonoscopy; 70 patients did not survive 1 year, 73 patients declined or were not deemed fit enough for colonoscopy, and 32 patients were lost to colonoscopic follow‐up. In total 613 colonoscopies were performed (Median 2 per patient, range 1–5). Polyps were identified and removed during 163 of these procedures (polypectomy rate 26.6%). Thirteen of the 15 patients (87%) who developed a metachronous cancer underwent at least one colonoscopy during follow‐up (Fig. 2). The polypectomy rate in this group (30%) was similar to the remaining population. In the nine asymptomatic metachronous cancer patients the median time from the last colonoscopy to diagnosis was 30 months. Four patients did not undergo colonoscopy for over 50 months (Fig. 2). Fig. 1. The cumulative incidence for metachronous cancer development from the time of the initial colorectal cancer resection. The number of patients at risk is shown below the x‐axis. The actuarial analysis is censored for recurrence and death.
between index and metachronous cancer diagnosis was 7 years and 6 months (range 2–14years). Nine of the 15 patients who developed a metachronous cancer were asymptomatic at the time of diagnosis by surveillance colonoscopy (Fig. 2). The metachronous cancers were staged as AJCC 1 in 47% of cases compared with 25% of index cancers.
Survival Survival was examined for all patients: median follow‐up was 4 years and 2 months and maximum follow‐up was 16 years and 4 months. Two hundred seventeen (40%)of the patients died; three had developed metachronous cancer. One hundred fifty two patients (28%) developed recurrent disease; one local recurrence was at the site of a metachronous cancer. A competing‐risks regression was carefully selected for modeling the time to metachronous tumors as a main event of interest after controlling for competing events (death and recurrence) and also adjusting for the effects of age, gender, stage, site, and number of polyps at initial
Fig. 2. A timeline between the primary cancer resection and metachronous cancer diagnosis for the 15 effected patients. Each line represents a patient with metachronous cancer. The light blue line represents asymptomatic patients (patients 2–5, 9, 10, 13–15) and the red line patients with symptoms (patients 1, 6–8, 11, 12). The red þ reflects the timing of colonoscopies, the black x represents a colonoscopy where one or more adenomatous polypectomies were performed. Both patients who did not receive colonoscopies had declined follow‐up, patient 6 presented with large bowel obstruction, patient 7 with rectal bleeding. There are relatively few colonoscopies performed beyond 60 months, as current guidelines recommend [13]. Journal of Surgical Oncology
4
Battersby et al.
Fig. 3. Adjusted cumulative incidence for metachronous cancer in males and females (four scenarios). Regression results for adjusted cumulative incidence of metachronous cancer accounting for death and recurrence by applying the flexible model for competing risks. Scenarios are based on age and adjusting for tumor stage and tumor site: S1 aged 55; S2 aged 65; S3 aged 75; S4 aged 85. Only age was a significant time dependent risk factor for metachronous cancer development.
colonoscopy. Despite the satisfactory fit and the flexibility of the model, only age had a significant time varying impact. As a continuous variable, age was significantly associated with a risk of developing metachronous cancer (P ¼ 0.018), HR 1.20 (CI 1.05–1.35). Thus, for a male patient aged 85, 75, 65, and 55 there was a 0.7%, 1.5%, 2.3%, and 3.1% 5‐year risk of developing metachronous cancer, respectively. This is a competing risk analysis and hence recurrence and the increased likelihood of all cause mortality in older patients is accounted for. For a transparent interpretation of the regression results, we calculate the cumulative incidence of metachronous cancer for four categories of males and females with the following fixed characteristics: AJCC1 stage, Right sided colon cancer, gender, and setting age at 55, 65, 75, and 85 year of age. Figure 3 provides the results for both genders.
DISCUSSION The number of patients at risk of developing a metachronous cancer will rise as a result of increasing longevity [19], coupled with a significant increase in colorectal cancer survival as a consequence of earlier detection of colorectal cancer by national screening initiatives, increased public awareness, improved staging, a higher proportion of curative resections, and myriad new treatment options [10,12,20,21]. Since the 1970s the 10‐year survival rate from colorectal cancer has doubled from 22% to 45% [7].Therefore surveillance and long‐term follow‐up for these patients needs careful consideration [2,3]. Following curative colorectal cancer resection, surveillance colonoscopy is intended to achieve both early detection of metachronous cancers and detection and removal of pre‐malignant lesions, thus lowering the risk of subsequent cancer formation [10–12]. In this series we describe the metachronous colorectal cancer rate detected during clinical and endoscopic follow‐up for a cohort of over Journal of Surgical Oncology
500 patients. The primary aim of the study was to report the time dependent metachronous cancer risk in an unselected population of colorectal cancer patients initially treated by curative resection of the primary cancer. In order to account for cancer recurrence and mortality, competing risks analysis was used. We believe this is the first study to use this technique when assessing the risk of developing metachronous colorectal cancer. A number of published series have suggested that metachronous cancers are more likely to occur in the first 3 years [22,23].These series used 6 months as the cut off between synchronous and metachronous cancer; using 1 year as the cut‐off we avoided an early peak incidence from missed synchronous cancers. The number of patients at risk from metachronous cancer falls over time as a consequence of death and recurrent disease. To circumvent this, we used competing risks analysis to account for cancer recurrence and deaths. Our data suggests that the cumulative incidence of developing metachronous cancer increases with time (Fig. 1). At 5 years there is a 1.8% risk which increases slightly to a 4% risk at 10 years. Despite the satisfactory fit and the flexibility of the competing risks model, only age had a significant time varying impact on metachronous cancer risk, highlighting the importance of prolonged colonoscopic follow‐up in the younger patients. In this series the crude metachronous cancer rate was 2.7%, which despite the aggressive approach to polypectomy, is similar to previously published series 0.8–7.2% [5–9,23].We have no way of knowing the metachronous cancer rate had surveillance colonoscopy not been undertaken, although it would appear that despite regular colonoscopic surveillance there is a persistent risk of developing a second cancer. There are two possible explanations for this. Firstly, small, flat polyps may have been missed at follow‐up colonoscopy. Secondly, although we regard the tumor adenoma sequence as occurring over a 10‐year period [24,25],these metachronous cancers appear to develop over a
Metachronous Colorectal Cancer narrower timeframe, which could reflect more aggressive tumorigenesis [26,27]. Considering the first explanation; other authors have accounted for tumors arising within the first 3 years as missed tumors because of inadequate endoscopy [28,29].It is possible that a few of our patients had a missed synchronous cancer. We think this unlikely because the colon had been assessed immediately prior to or soon after resection of the index cancer. Similarly, the polyp detection rate is within current guidelines for high quality colonoscopy [23,30]. Colonoscopies in this series were performed by, or in the presence of the senior author who has an audited cecal completion rate of 95% (JGW) and polyp detection rate of over 25%. In addition, the endoscopy unit is 1 of 16 UK JAG accredited endoscopy training centres. Thus we believe that using an experienced endoscopist, who is an accredited trainer, minimizes the likelihood that synchronous lesions or significant polyps were missed. Considering the second explanation; some patients may have had an undiagnosed hereditary non‐polyposis colorectal cancer syndrome (HNPCC), a microsatellite instability (MSI) or other de novo genetic mutations leading to an accelerated adenoma to carcinoma sequence [31,32].Mulder et al. [22] recently reported the metachronous cancer incidence from the Rotterdam Cancer Registry. An intriguing finding was that patients diagnosed with a metachronous lesion within 3 years of resection were significantly more likely to have a poorly differentiated cancer than patients diagnosed after 3 years. The study was limited by uncertainty over the timing and frequency of endoscopic follow‐up, but this finding lends support to the argument that more aggressive tumorigenesis plays a role in development of a metachronous cancer. The higher adjusted cumulative incidence in younger patients may also reflect more aggressive tumorigenesis, however, these patients were not routinely genetically screened. No colonoscopic follow‐up occurred in 175 of the 538 patients eligible for the reasons outlined above. This limitation may alter the likelihood and timing of metachronous cancer diagnosis. However, our aim was to report the risk of metachronous cancer development in an unselected population who underwent curative primary colorectal resection. The non‐surveyed patients still present with metachronous cancer and to ensure optimal generalizability these patients have been included in the analysis. Finally, once follow‐up reaches 10 years, the effect of a small number of randomly occurring metachronous cancers may distort the results. This is a well‐documented concern, and with a relatively small risk set therefore we have understated the values beyond 10 years [33]. Current BSG and ACPGBI guidelines recommend that following investigation for a synchronous lesion, surveillance endoscopy should be 5 yearly until benefit is outweighed by comorbidity [13]. The Australian National Health & Medical Research Council (NHMRC) and the US Multi‐Society Task Force on Colorectal Cancer suggest 3–5 yearly follow‐up colonoscopy [23,34]. One third of the patients were diagnosed with metachronous cancer more than 50 months after a surveillance colonoscopy. An important consideration is that the recommended 5‐year delay between colonoscopic follow‐up may be too long for adequate and early detection of polyps and cancers. It is possible that earlier endoscopy would have diagnosed these cancers at an earlier stage or perhaps as adenomas. Several authors have suggested that the minimum frequency of colonoscopy should be 3 years [23,29,35]. Interpreting the competing risks analysis, a patient aged 55 has a 3% 5‐year and 4% 10‐year metachronous cancer risk, where as a patient aged 65 has a 2% 5‐year and 3% 10‐year risk. We also believe that decreasing the interval between surveillance colonoscopies deserves further consideration in this group. Equally the overall risk in the older subgroup is less that 2% at 10 years for those aged over 75 and less than 1.5% for those over 85 years. For this group an assessment of patient fitness and a discussion with the patient about the oncological risk that they will tolerate should determine whether surveillance continues. Journal of Surgical Oncology
5
CONCLUSION In this series, we report the metachronous colorectal cancer rate detected during clinical and endoscopic follow‐up in a district general hospital. Using competing risks analysis we identify that it is the youngest patients who carry the greatest time dependent risk of developing metachronous colorectal cancer. As colorectal cancer survival and overall life expectancy are increasing, we believe follow‐up should continue while patients remain willing and fit enough for intervention. A patient aged under 65 at the time of the primary curative resection carries a 2% 5‐year risk of metachronous cancer, implying that 3‐year surveillance colonoscopy is justified. Whereas those aged over 75 carry less than a 2% 10‐year risk, implying that it is seldom warranted to repeat the colonoscopy more frequently than every 5 years. A stratified approach to the frequency of surveillance colonoscopy requires further consideration.
AUTHOR CONTRIBUTIONS NJB substantially contributed to conception, design, data acquisition, analysis and interpretation of data; drafting the article and approved the final published version. AC and NM substantially contributed to acquisition of data; revising draft critically for important intellectual content and approved the final published version. GB substantially contributed to the interpretation and presentation of data; revising draft critically for important intellectual content; and approved the final published version. JGW substantially contributed to conception, design, data acquisition and interpretation of data; drafting the article and revising it critically for important intellectual content; and approved the final published version.
REFERENCES 1. Office for National Statistics: Survival rates in England, patients diagnosed 2001–2006 followed up to 2007. 2. Cali RL, Pitsch RM, Thorson AG, Watson P, Tapia P, Blatchford GJ, Christensen MA: Cumulative incidence of metachronous colorectal cancer. Dis Colon Rectum 1993;36:388–393. 3. Warren S, Gates O: Multiple primary malignant tumors: A survey of the literature and statistical study. Am J Cancer 1932;16:1358– 1414. 4. Kaibara N, Koga S, Jinnai D: Synchronous and metachronous malignancies of the colon and rectum in Japan with special reference to a coexisting early cancer. Cancer 1984;54:1870–1874. 5. Heald RJ: Synchronous and metachronous carcinoma of the colon and rectum. Ann R Coll Surg Engl 1990;72:172–174. 6. Finan PJ, Ritchie JK, Hawley PR: Synchronous and “early” metachronous carcinomas of the colon and rectum. BJS 1987;74: 945–947. 7. Bülow S, Svendsen LB, Mellemgaard A: Metachronous colorectal carcinoma. BJS 1990;77:502–505. 8. Rennert G, Robinson E, Rennert HS, Neugut AI: Clinical characteristics of metachronous colorectal tumors. Int J Cancer 1995;60:743–747. 9. Park IJ, Yu CS, Kim HC, Jung YH, Han KR, Kim JC: Metachronous colorectal cancer. Colorectal Dis 20068:323–327. 10. Winawer SJ, Zauber AG, Ho MN, O’Brien MJ, Gottlieb LS, Sternberg SS, Waye JD, Schapiro M, Bond JH, Panish JF, Ackroyd F, Shike M, Kurtz RC, Hornsby‐Lewis L, Gerdes H, Stewart ET, and the National Polyp Study Workgroup: Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. NEJM 1993;329:1977–1981. 11. Hoff G, Grotmol T, Skovlund E, Bretthauer M: Risk of colorectal cancer seven years after flexible sigmoidoscopy screening: Randomised controlled trial. BMJ 2009;338:b1846. 12. Atkin WS, Edwards R, Kralj‐Hans I, Wooldrage K, Hart AR, Northover JM, Parkin DM, Wardle J, Duffy SW, Cuzick J; UK Flexible Sigmoidoscopy Trial Investigators: Once‐only flexible
6
13.
14. 15.
16. 17. 18. 19.
20.
21.
22.
Battersby et al. sigmoidoscopy screening in prevention of colorectal cancer: A multicentre randomised controlled trial. Lancet 2010;375:1624– 1633. Cairns SR, Scholefield JH, Steele RJ, Dunlop MG, Thomas HJ, Evans GD, Eaden JA, Rutter MD, Atkin WP, Saunders BP, Lucassen A, Jenkins P, Fairclough PD, Woodhouse CR; British Society of Gastroenterology; Association of Coloproctology for Great Britain and Ireland: Guidelines for colorectal cancer screening and surveillance in moderate and high risk groups (update from 2002). Gut 2010;59:666–689. Müller AD, Sonnenberg A: Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case‐control study of 32,702 veterans. Ann Intern Med 1995;123:904–910. Winawer SJ, Fletcher RH, Miller L, Godlee F, Stolar MH, Mulrow CD, Woolf SH, Glick SN, Ganiats TG, Bond JH, Rosen L, Zapka JG, Olsen SJ, Giardiello FM, Sisk JE, Van Antwerp R, Brown‐ Davis C, Marciniak DA, Mayer RJ: Colorectal cancer screening: Clinical guidelines and rationale. Gastroenterology 1997;112:594– 642. Royston P, Lambert P: Further development of flexible parametric models for survival analysis. Stata J 2009;9:265–290. Royston P, Lambert P: Flexible parametric survival analysis using Stata: Beyond the Cox model, Texas: Stata Press; 2011.p. 347. Fine J, Gray R: A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:495–509. Mills J: What are the chances of surviving to age 100? Demographic analysis [Internet]. UK; 2012. Available from: http://www.ons.gov. uk/ons/rel/lifetables/historic‐and‐projected‐mortality‐data‐from‐the‐ uk‐life‐tables/2010‐based/rpt‐surviving‐to‐100.html Sebag‐Montefiore D, Stephens RJ, Steele R, Monson J, Grieve R, Khanna S, Quirke P, Couture J, de Metz C, Myint AS, Bessell E, Griffiths G, Thompson LC, Parmar M: Preoperative radiotherapy versus selective postoperative chemoradiotherapy in patients with rectal cancer (MRC CR07 and NCIC‐CTG C016): A multicentre, randomised trial. Lancet 2009;373:811–820. Martling AL, Holm T, Rutqvist LE, Moran BJ, Heald RJ, Cedemark B: Effect of a surgical training programme on outcome of rectal cancer in the County of Stockholm. Stockholm Colorectal Cancer Study Group, Basingstoke Bowel Cancer Research Project. Lancet 2000356:93–96. Mulder SA, Kranse R, Damhuis RA, Ouwendijk RJT, Kuipers EJ, Van Leerdam ME: The incidence and risk factors of metachronous colorectal cancer: An indication for follow‐up. Dis Colon Rectum 2012;55:522–531.
Journal of Surgical Oncology
23. Hollington P, Tiong L, Young G: Timing and detection of metachronous colorectal cancer. ANZ J Surg 2011;81:272–274. 24. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL: Genetic alterations during colorectal‐tumor development. NEJM 1988;319: 525–532. 25. Morson B: The polyp‐cancer sequence in the large bowel. Proc R Soc Med 1974;67:451–457. 26. Lynch HT, Smyrk T, Jass JR: Hereditary nonpolyposis colorectal cancer and colonic adenomas: Aggressive adenomas? Semin Surg Oncol 1995;11:406–410. 27. Ahlquist DA: Aggressive polyps in hereditary nonpolyposis colorectal cancer: Targets for screening. Gastroenterology 1995; 108:1590–1592. 28. Bensen S, Mott LA, Dain B, Rothstein R, Baron J: The colonoscopic miss rate and true one‐year recurrence of colorectal neoplastic polyps. Polyp Prevention Study Group. Am J Gastroenterol 1999;94:194–199. 29. Green RJ, Metlay JP, Propert K, Catalano PJ, Macdonald JS, Mayer RJ, Haller DG: Surveillance for second primary colorectal cancer after adjuvant chemotherapy: An analysis of Intergroup 0089. Ann Intern Med: 2002;136:261–269. 30. Viel JF, Studer JM, Ottignon Y, Hirsch JP; Franche‐Comté Polyp Surveillance Study Group: Predictors of colorectal polyp recurrence after the first polypectomy in private practice settings: A cohort study. PLoS ONE 2012;7:e50990. 31. Fajobi O, Yiu CY, Sen‐Gupta SB, Boulos PB: Metachronous colorectal cancers. BJS 1998;85:897–901. 32. Fante R, Roncucci L, Di Gregorio C, Tamassia MG, Losi L, Benatti P, et al.: Frequency and clinical features of multiple tumors of the large bowel in the general population and in patients with hereditary colorectal carcinoma. Cancer 1996;77:2013–2021. 33. Singer JD, Willet JB: Applied longitudinal data analysis: Modeling change and event occurrence. In: Singer JD, Willett JB, editors. The cumulative hazard function, 1st edition. USA: Oxford University Press; 2003.Chapter 13.4. ISBN0195152964. 34. Rex DK, Kahi CJ, Levin B, Smith RA, Bond JH, Brooks D, et al.: Guidelines for colonoscopy surveillance after cancer resection: A consensus update by the American Cancer Society and the US Multi‐Society Task Force on Colorectal Cancer. Gastroenterology 2006;130:1865–1871. 35. Bouvier A‐M, Latournerie M, Jooste V, Lepage C, Cottet V, Faivre J: The lifelong risk of metachronous colorectal cancer justifies long‐term colonoscopic follow‐up. Eur J Cancer 2008;44:522–527.
Metachronous Colorectal Cancer: A Competing Risks Analysis with Consideration for a Stratified Approach to Surveillance Colonoscopy NICHOLAS J. BATTERSBY, MRCS, BMedSc,1 ALEX COUPLAND, MBBS,1 GEORGE BOULIOTIS, PhD(MedStat), MSc(Math),2 NAZZIA MIRZA, FRCS,1 AND J. GRAHAM WILLIAMS, MChFRCS1* 1 Colorectal Surgery, The Royal Wolverhampton Hospital, Wolverhampton, United Kingdom MRC-Midland Hub for Trials Methodology Research, University of Birmingham, Birmingham, United Kingdom
2
Background: The incidence of metachronous cancer will become an important clinical consideration as the life expectancy of the population increases and as rates of curative resection improve. Objective: To assess the pattern of metachronous cancer development following curative resection of colorectal cancer in an unselected patient population offered postoperative colonoscopic surveillance. Method: Prospective clinical follow‐up after curative colorectal cancer resection and surveillance colonoscopy with or without polypectomy in accordance with the national guidelines. Actuarial analysis and competing risk analysis were performed to account for death and recurrence and to stratify for age, gender, stage, and tumor site. Results: Five hundred thirty‐eight patients with median follow‐up 4 years 2 month (0–16) years. Fifteen patients (3%) developed metachronous cancer, at a median time interval of 90 months from primary resection. Thirteen metachronous cancer patients (87%, 13/15) underwent one to five surveillance colonoscopies: nine patients were asymptomatic at time of diagnosis of metachronous cancer. Competing risks analysis suggests that the adjusted cumulative incidence in males aged 55 is 4% at 10 years compared with 1% in females aged 85 years old. Conclusions: A patient aged under 65 at the time of the primary curative resection carries a 2% 5‐year risk of metachronous cancer, implying that 3 year surveillance colonoscopy is justified. Whereas patients aged over 75 carry less than a 2% 10‐year risk, implying that it is seldom warranted to repeat the colonoscopy more frequently than every 5 years. A stratified approach to the frequency of surveillance colonoscopy requires further consideration.
J. Surg. Oncol. . ß 2013 Wiley Periodicals, Inc.
KEY WORDS: metachronous; colorectal cancer; surveillance colonoscopy; competing risks analysis
INTRODUCTION Colorectal cancer is the third most common cancer diagnosed in the UK [1]. Patients who have had one colorectal cancer are at risk of developing additional, metachronous, colorectal cancers [2,3]. Metachronous colorectalcancers are tumors that develop at a site remote from the primary tumor, which are histologically separate and occur 12 months or more after surgery, in a completely investigated colon [4].This definition draws the distinction between synchronous tumors, present at the time of the primary resection, and from recurrence of the primary cancer. A number of case‐series with long‐term follow‐up have reported metachronous cancer rates of 0.7–9% [5–9]. In 1993, the national polyp study published evidence that colonoscopy and polypectomy enabled early detection of cancer and reduced the subsequent incidence of cancer; several studies have since supported these findings [10–12]. The primary aim of this study was to report the time dependent risk of metachronous cancer in an unselected population of patients who had previously undergone a curative colorectal cancer resection, in a unit, which applies current surveillance colonoscopy guidelines and takes a proactive approach to colonoscopy and polypectomy [13–15]. This is the first study reporting on metachronous cancer rate to use competing risks analysis in order to account for cancer recurrence and mortality.
METHODOLOGY Details of 572 patients who underwent colorectal cancer resection under the care of a single consultant colorectal surgeon (JGW) between
ß 2013 Wiley Periodicals, Inc.
May 1995 and February 2012 were prospectively recorded in a database that included patient demographics, AJCC tumor stage and site, operation performed, cancer recurrence, metachronous cancer, and mortality. Endoscopic data was also recorded. Patients with synchronous colorectal cancer, who underwent panproctocolectomy, and those who did not undergo resection of the primary cancer or who underwent a palliative operation were excluded. Data collection and analysis were performed according to the Hospital Trust ethics and clinical governance policy. Each patient underwent pre‐operative colonoscopy to identify synchronous lesions (or within 3 months of surgery where an obstructing lesion prevented complete investigation, or after emergency presentation). All colonoscopies were performed in a Joint Advisory Group (JAG) UK colonoscopy training center following bowel preparation using sodium picrosulphate or Fleet Phosphosoda1
Conflict of interest: none. This data was presented in part as a podium presentation at ASCRS 2012, San Antonio, Texas. *Correspondence to: J. Graham Williams, Colorectal Surgery, The Royal Wolverhampton Hospital, Wolverhampton WV10 0QP, United Kingdom. Fax: þ01902 695691. E‐mail: [email protected] Received 28 May 2013; Accepted 28 October 2013 DOI 10.1002/jso.23504 Published online in Wiley Online Library (wileyonlinelibrary.com).
2
Battersby et al.
(Lynchburg, Virginia 24502). The majority of examinations were performed by the senior author, or by a supervised (JGW) trainee. Follow‐up colonoscopy protocol was in accordance with current American Gastroenterological Association and the American Cancer Society guidelines [14,15]. The majority of patients were reviewed in the out‐patient department every 6 months for 3 years and annually up to 5 years. Where out‐patient follow‐up did not occur, further information on the patient’s status was obtained from hospital records and the patient’s General Practitioner.
Among various candidates, a modified Fine and Gray [18] model for metachronous events that adjust for cancer recurrence and death as competing events was used. Finally, model predictions were inputted into plausible clinical scenarios for estimating probabilistic differences among various groups of patients. Statistical significance was taken as P < 0.05 for all tests. Best‐fit model was selected based on the Akaike and Bayesian Information Criteria (AIC and BIC). Survival analysis was implemented with the Stata software [16,17].
Statistical Analysis
RESULTS
In the first part of the analysis, non‐parametric tests for assessing univariate differences related to patients’ characteristics (Chi‐squared or Fisher’s exact test). Among various covariates we generated the polypectomy rate as the total number of polyps divided by the total number of procedures. We also categorized age into four groups (less than 55, 55–64, 65–74, 75 plus). Finally, we grouped the location of cancer into three groups (rectum, right colon, and other) in order to limit strong imbalances among the groups. In the second part, we analyzed patient survival from the date of diagnosis of the index cancer, focusing on the occurrence of a metachronous cancer. Survival analysis included an additional step that of competing risks model [16,17]. Although semiparametric ordinary Cox regression could be extended to accommodate competing risks, we evaluated factors for modeled time varying covariates; as such a flexible parametric model was used in order to provide the most precise data fit.
Demographics Of the 572 patients undergoing resection of colorectal cancer, 34 patients were excluded from analysis (panproctocolectomy—3, synchronous cancer—3, colonic stent—1, and unresectable disease— 27). The remaining 538 patients had a median age of 70 years 11 months at the time of surgery; 295 (56%) were male. The AJCC stage, site of the index cancer and the operations performed are summarized in Table I.
Metachronous Cancers During follow‐up, 15 patients developed a metachronous cancer, giving a crude rate of 2.7% (15/538). The risk of metachronous cancer increased with time; the cumulative incidence was 1.8%, 4%, and 8% at 5, 10, and 12 years, respectively (Fig. 1). The median time interval
TABLE I. Demographic Data Metachronous patients, (n ¼ 15) Variables Age (IQ range)
Gender Colorectal tumor site (n ¼ 538) Cecum and ascending colon Transverse Descending Sigmoid Rectosigmoid Rectum AJCC stage (n ¼ 538) I II III Unknown Operation (n ¼ 538) AP resection Low anterior resection High anterior resection Left hemicolectomy Right hemicolectomy Hartmann’s procedure Transverse or extended right hemicolectomy Subtotal colectomy Pelvic exenteration Sigmoid resection Transanal resection of cancer Proctocolectomy
Primary cancer, (n ¼ 523)a 70 years 11 months (61 years 10 months– 77 years 11 months) 228 Female (44%) 295 Male (56%)
67 years 2 months (59 years 9 months– 75 years 10 months) 7 Female (47%) 8 Male (53%)
Metachronous 75 years 10 months (65years 11months– 81years 9months)
143 (27%) 26 (5%) 83 (16%) 31 (6%) 78 (15%) 162 (31%)
5 (33%) 0 0 4 (27%) 1 (6.7%) 5 (33%)
3 (20%) 3 (20%) 1 (6.7%) 2 (13%) 0 6 (40%)
95 (18%) 213 (41%) 207 (40%) 8 (1%)
4 (27%) 7 (47%) 4 (27%)
7 (47%) 5 (33%) 3 (20%)
41 (8%) 111 (21%) 77 (15%) 78 (15%) 144 (28%) 26 (5%) 26 (5%)
1 (6.7%) 4 (27%) 2 (13%) 4 (27%) 4 (27%) 0 0
1 (6.7%)
9 (2%) 3 (0.5%) 2 (0.5%) 6 (1%) —
0 0 0 0 —
2 (13%)
No significant difference between the groups was identified. a Patients who underwent a curative resection and did not develop a metachronous cancer.
Journal of Surgical Oncology
Index
1 (6.7%) 3 (20%) 1 (6.7%) 3 (20%)
4 (27%)
Metachronous Colorectal Cancer
3
Of the remaining six patients who developed a metachronous cancer: two presented with a change in bowel habit; two as emergencies with large bowel obstruction from a cecal and sigmoid tumor; one patient declined colonoscopic follow‐up and later presented with profuse bleeding and one patient had a palpable rectal polyp found in the outpatient clinic (Fig. 2).
Colonoscopic Follow‐up Follow‐up colonoscopy was performed on 361 patients. Of the 175 patients who did not undergo colonoscopy; 70 patients did not survive 1 year, 73 patients declined or were not deemed fit enough for colonoscopy, and 32 patients were lost to colonoscopic follow‐up. In total 613 colonoscopies were performed (Median 2 per patient, range 1–5). Polyps were identified and removed during 163 of these procedures (polypectomy rate 26.6%). Thirteen of the 15 patients (87%) who developed a metachronous cancer underwent at least one colonoscopy during follow‐up (Fig. 2). The polypectomy rate in this group (30%) was similar to the remaining population. In the nine asymptomatic metachronous cancer patients the median time from the last colonoscopy to diagnosis was 30 months. Four patients did not undergo colonoscopy for over 50 months (Fig. 2). Fig. 1. The cumulative incidence for metachronous cancer development from the time of the initial colorectal cancer resection. The number of patients at risk is shown below the x‐axis. The actuarial analysis is censored for recurrence and death.
between index and metachronous cancer diagnosis was 7 years and 6 months (range 2–14years). Nine of the 15 patients who developed a metachronous cancer were asymptomatic at the time of diagnosis by surveillance colonoscopy (Fig. 2). The metachronous cancers were staged as AJCC 1 in 47% of cases compared with 25% of index cancers.
Survival Survival was examined for all patients: median follow‐up was 4 years and 2 months and maximum follow‐up was 16 years and 4 months. Two hundred seventeen (40%)of the patients died; three had developed metachronous cancer. One hundred fifty two patients (28%) developed recurrent disease; one local recurrence was at the site of a metachronous cancer. A competing‐risks regression was carefully selected for modeling the time to metachronous tumors as a main event of interest after controlling for competing events (death and recurrence) and also adjusting for the effects of age, gender, stage, site, and number of polyps at initial
Fig. 2. A timeline between the primary cancer resection and metachronous cancer diagnosis for the 15 effected patients. Each line represents a patient with metachronous cancer. The light blue line represents asymptomatic patients (patients 2–5, 9, 10, 13–15) and the red line patients with symptoms (patients 1, 6–8, 11, 12). The red þ reflects the timing of colonoscopies, the black x represents a colonoscopy where one or more adenomatous polypectomies were performed. Both patients who did not receive colonoscopies had declined follow‐up, patient 6 presented with large bowel obstruction, patient 7 with rectal bleeding. There are relatively few colonoscopies performed beyond 60 months, as current guidelines recommend [13]. Journal of Surgical Oncology
4
Battersby et al.
Fig. 3. Adjusted cumulative incidence for metachronous cancer in males and females (four scenarios). Regression results for adjusted cumulative incidence of metachronous cancer accounting for death and recurrence by applying the flexible model for competing risks. Scenarios are based on age and adjusting for tumor stage and tumor site: S1 aged 55; S2 aged 65; S3 aged 75; S4 aged 85. Only age was a significant time dependent risk factor for metachronous cancer development.
colonoscopy. Despite the satisfactory fit and the flexibility of the model, only age had a significant time varying impact. As a continuous variable, age was significantly associated with a risk of developing metachronous cancer (P ¼ 0.018), HR 1.20 (CI 1.05–1.35). Thus, for a male patient aged 85, 75, 65, and 55 there was a 0.7%, 1.5%, 2.3%, and 3.1% 5‐year risk of developing metachronous cancer, respectively. This is a competing risk analysis and hence recurrence and the increased likelihood of all cause mortality in older patients is accounted for. For a transparent interpretation of the regression results, we calculate the cumulative incidence of metachronous cancer for four categories of males and females with the following fixed characteristics: AJCC1 stage, Right sided colon cancer, gender, and setting age at 55, 65, 75, and 85 year of age. Figure 3 provides the results for both genders.
DISCUSSION The number of patients at risk of developing a metachronous cancer will rise as a result of increasing longevity [19], coupled with a significant increase in colorectal cancer survival as a consequence of earlier detection of colorectal cancer by national screening initiatives, increased public awareness, improved staging, a higher proportion of curative resections, and myriad new treatment options [10,12,20,21]. Since the 1970s the 10‐year survival rate from colorectal cancer has doubled from 22% to 45% [7].Therefore surveillance and long‐term follow‐up for these patients needs careful consideration [2,3]. Following curative colorectal cancer resection, surveillance colonoscopy is intended to achieve both early detection of metachronous cancers and detection and removal of pre‐malignant lesions, thus lowering the risk of subsequent cancer formation [10–12]. In this series we describe the metachronous colorectal cancer rate detected during clinical and endoscopic follow‐up for a cohort of over Journal of Surgical Oncology
500 patients. The primary aim of the study was to report the time dependent metachronous cancer risk in an unselected population of colorectal cancer patients initially treated by curative resection of the primary cancer. In order to account for cancer recurrence and mortality, competing risks analysis was used. We believe this is the first study to use this technique when assessing the risk of developing metachronous colorectal cancer. A number of published series have suggested that metachronous cancers are more likely to occur in the first 3 years [22,23].These series used 6 months as the cut off between synchronous and metachronous cancer; using 1 year as the cut‐off we avoided an early peak incidence from missed synchronous cancers. The number of patients at risk from metachronous cancer falls over time as a consequence of death and recurrent disease. To circumvent this, we used competing risks analysis to account for cancer recurrence and deaths. Our data suggests that the cumulative incidence of developing metachronous cancer increases with time (Fig. 1). At 5 years there is a 1.8% risk which increases slightly to a 4% risk at 10 years. Despite the satisfactory fit and the flexibility of the competing risks model, only age had a significant time varying impact on metachronous cancer risk, highlighting the importance of prolonged colonoscopic follow‐up in the younger patients. In this series the crude metachronous cancer rate was 2.7%, which despite the aggressive approach to polypectomy, is similar to previously published series 0.8–7.2% [5–9,23].We have no way of knowing the metachronous cancer rate had surveillance colonoscopy not been undertaken, although it would appear that despite regular colonoscopic surveillance there is a persistent risk of developing a second cancer. There are two possible explanations for this. Firstly, small, flat polyps may have been missed at follow‐up colonoscopy. Secondly, although we regard the tumor adenoma sequence as occurring over a 10‐year period [24,25],these metachronous cancers appear to develop over a
Metachronous Colorectal Cancer narrower timeframe, which could reflect more aggressive tumorigenesis [26,27]. Considering the first explanation; other authors have accounted for tumors arising within the first 3 years as missed tumors because of inadequate endoscopy [28,29].It is possible that a few of our patients had a missed synchronous cancer. We think this unlikely because the colon had been assessed immediately prior to or soon after resection of the index cancer. Similarly, the polyp detection rate is within current guidelines for high quality colonoscopy [23,30]. Colonoscopies in this series were performed by, or in the presence of the senior author who has an audited cecal completion rate of 95% (JGW) and polyp detection rate of over 25%. In addition, the endoscopy unit is 1 of 16 UK JAG accredited endoscopy training centres. Thus we believe that using an experienced endoscopist, who is an accredited trainer, minimizes the likelihood that synchronous lesions or significant polyps were missed. Considering the second explanation; some patients may have had an undiagnosed hereditary non‐polyposis colorectal cancer syndrome (HNPCC), a microsatellite instability (MSI) or other de novo genetic mutations leading to an accelerated adenoma to carcinoma sequence [31,32].Mulder et al. [22] recently reported the metachronous cancer incidence from the Rotterdam Cancer Registry. An intriguing finding was that patients diagnosed with a metachronous lesion within 3 years of resection were significantly more likely to have a poorly differentiated cancer than patients diagnosed after 3 years. The study was limited by uncertainty over the timing and frequency of endoscopic follow‐up, but this finding lends support to the argument that more aggressive tumorigenesis plays a role in development of a metachronous cancer. The higher adjusted cumulative incidence in younger patients may also reflect more aggressive tumorigenesis, however, these patients were not routinely genetically screened. No colonoscopic follow‐up occurred in 175 of the 538 patients eligible for the reasons outlined above. This limitation may alter the likelihood and timing of metachronous cancer diagnosis. However, our aim was to report the risk of metachronous cancer development in an unselected population who underwent curative primary colorectal resection. The non‐surveyed patients still present with metachronous cancer and to ensure optimal generalizability these patients have been included in the analysis. Finally, once follow‐up reaches 10 years, the effect of a small number of randomly occurring metachronous cancers may distort the results. This is a well‐documented concern, and with a relatively small risk set therefore we have understated the values beyond 10 years [33]. Current BSG and ACPGBI guidelines recommend that following investigation for a synchronous lesion, surveillance endoscopy should be 5 yearly until benefit is outweighed by comorbidity [13]. The Australian National Health & Medical Research Council (NHMRC) and the US Multi‐Society Task Force on Colorectal Cancer suggest 3–5 yearly follow‐up colonoscopy [23,34]. One third of the patients were diagnosed with metachronous cancer more than 50 months after a surveillance colonoscopy. An important consideration is that the recommended 5‐year delay between colonoscopic follow‐up may be too long for adequate and early detection of polyps and cancers. It is possible that earlier endoscopy would have diagnosed these cancers at an earlier stage or perhaps as adenomas. Several authors have suggested that the minimum frequency of colonoscopy should be 3 years [23,29,35]. Interpreting the competing risks analysis, a patient aged 55 has a 3% 5‐year and 4% 10‐year metachronous cancer risk, where as a patient aged 65 has a 2% 5‐year and 3% 10‐year risk. We also believe that decreasing the interval between surveillance colonoscopies deserves further consideration in this group. Equally the overall risk in the older subgroup is less that 2% at 10 years for those aged over 75 and less than 1.5% for those over 85 years. For this group an assessment of patient fitness and a discussion with the patient about the oncological risk that they will tolerate should determine whether surveillance continues. Journal of Surgical Oncology
5
CONCLUSION In this series, we report the metachronous colorectal cancer rate detected during clinical and endoscopic follow‐up in a district general hospital. Using competing risks analysis we identify that it is the youngest patients who carry the greatest time dependent risk of developing metachronous colorectal cancer. As colorectal cancer survival and overall life expectancy are increasing, we believe follow‐up should continue while patients remain willing and fit enough for intervention. A patient aged under 65 at the time of the primary curative resection carries a 2% 5‐year risk of metachronous cancer, implying that 3‐year surveillance colonoscopy is justified. Whereas those aged over 75 carry less than a 2% 10‐year risk, implying that it is seldom warranted to repeat the colonoscopy more frequently than every 5 years. A stratified approach to the frequency of surveillance colonoscopy requires further consideration.
AUTHOR CONTRIBUTIONS NJB substantially contributed to conception, design, data acquisition, analysis and interpretation of data; drafting the article and approved the final published version. AC and NM substantially contributed to acquisition of data; revising draft critically for important intellectual content and approved the final published version. GB substantially contributed to the interpretation and presentation of data; revising draft critically for important intellectual content; and approved the final published version. JGW substantially contributed to conception, design, data acquisition and interpretation of data; drafting the article and revising it critically for important intellectual content; and approved the final published version.
REFERENCES 1. Office for National Statistics: Survival rates in England, patients diagnosed 2001–2006 followed up to 2007. 2. Cali RL, Pitsch RM, Thorson AG, Watson P, Tapia P, Blatchford GJ, Christensen MA: Cumulative incidence of metachronous colorectal cancer. Dis Colon Rectum 1993;36:388–393. 3. Warren S, Gates O: Multiple primary malignant tumors: A survey of the literature and statistical study. Am J Cancer 1932;16:1358– 1414. 4. Kaibara N, Koga S, Jinnai D: Synchronous and metachronous malignancies of the colon and rectum in Japan with special reference to a coexisting early cancer. Cancer 1984;54:1870–1874. 5. Heald RJ: Synchronous and metachronous carcinoma of the colon and rectum. Ann R Coll Surg Engl 1990;72:172–174. 6. Finan PJ, Ritchie JK, Hawley PR: Synchronous and “early” metachronous carcinomas of the colon and rectum. BJS 1987;74: 945–947. 7. Bülow S, Svendsen LB, Mellemgaard A: Metachronous colorectal carcinoma. BJS 1990;77:502–505. 8. Rennert G, Robinson E, Rennert HS, Neugut AI: Clinical characteristics of metachronous colorectal tumors. Int J Cancer 1995;60:743–747. 9. Park IJ, Yu CS, Kim HC, Jung YH, Han KR, Kim JC: Metachronous colorectal cancer. Colorectal Dis 20068:323–327. 10. Winawer SJ, Zauber AG, Ho MN, O’Brien MJ, Gottlieb LS, Sternberg SS, Waye JD, Schapiro M, Bond JH, Panish JF, Ackroyd F, Shike M, Kurtz RC, Hornsby‐Lewis L, Gerdes H, Stewart ET, and the National Polyp Study Workgroup: Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. NEJM 1993;329:1977–1981. 11. Hoff G, Grotmol T, Skovlund E, Bretthauer M: Risk of colorectal cancer seven years after flexible sigmoidoscopy screening: Randomised controlled trial. BMJ 2009;338:b1846. 12. Atkin WS, Edwards R, Kralj‐Hans I, Wooldrage K, Hart AR, Northover JM, Parkin DM, Wardle J, Duffy SW, Cuzick J; UK Flexible Sigmoidoscopy Trial Investigators: Once‐only flexible
6
13.
14. 15.
16. 17. 18. 19.
20.
21.
22.
Battersby et al. sigmoidoscopy screening in prevention of colorectal cancer: A multicentre randomised controlled trial. Lancet 2010;375:1624– 1633. Cairns SR, Scholefield JH, Steele RJ, Dunlop MG, Thomas HJ, Evans GD, Eaden JA, Rutter MD, Atkin WP, Saunders BP, Lucassen A, Jenkins P, Fairclough PD, Woodhouse CR; British Society of Gastroenterology; Association of Coloproctology for Great Britain and Ireland: Guidelines for colorectal cancer screening and surveillance in moderate and high risk groups (update from 2002). Gut 2010;59:666–689. Müller AD, Sonnenberg A: Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case‐control study of 32,702 veterans. Ann Intern Med 1995;123:904–910. Winawer SJ, Fletcher RH, Miller L, Godlee F, Stolar MH, Mulrow CD, Woolf SH, Glick SN, Ganiats TG, Bond JH, Rosen L, Zapka JG, Olsen SJ, Giardiello FM, Sisk JE, Van Antwerp R, Brown‐ Davis C, Marciniak DA, Mayer RJ: Colorectal cancer screening: Clinical guidelines and rationale. Gastroenterology 1997;112:594– 642. Royston P, Lambert P: Further development of flexible parametric models for survival analysis. Stata J 2009;9:265–290. Royston P, Lambert P: Flexible parametric survival analysis using Stata: Beyond the Cox model, Texas: Stata Press; 2011.p. 347. Fine J, Gray R: A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:495–509. Mills J: What are the chances of surviving to age 100? Demographic analysis [Internet]. UK; 2012. Available from: http://www.ons.gov. uk/ons/rel/lifetables/historic‐and‐projected‐mortality‐data‐from‐the‐ uk‐life‐tables/2010‐based/rpt‐surviving‐to‐100.html Sebag‐Montefiore D, Stephens RJ, Steele R, Monson J, Grieve R, Khanna S, Quirke P, Couture J, de Metz C, Myint AS, Bessell E, Griffiths G, Thompson LC, Parmar M: Preoperative radiotherapy versus selective postoperative chemoradiotherapy in patients with rectal cancer (MRC CR07 and NCIC‐CTG C016): A multicentre, randomised trial. Lancet 2009;373:811–820. Martling AL, Holm T, Rutqvist LE, Moran BJ, Heald RJ, Cedemark B: Effect of a surgical training programme on outcome of rectal cancer in the County of Stockholm. Stockholm Colorectal Cancer Study Group, Basingstoke Bowel Cancer Research Project. Lancet 2000356:93–96. Mulder SA, Kranse R, Damhuis RA, Ouwendijk RJT, Kuipers EJ, Van Leerdam ME: The incidence and risk factors of metachronous colorectal cancer: An indication for follow‐up. Dis Colon Rectum 2012;55:522–531.
Journal of Surgical Oncology
23. Hollington P, Tiong L, Young G: Timing and detection of metachronous colorectal cancer. ANZ J Surg 2011;81:272–274. 24. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL: Genetic alterations during colorectal‐tumor development. NEJM 1988;319: 525–532. 25. Morson B: The polyp‐cancer sequence in the large bowel. Proc R Soc Med 1974;67:451–457. 26. Lynch HT, Smyrk T, Jass JR: Hereditary nonpolyposis colorectal cancer and colonic adenomas: Aggressive adenomas? Semin Surg Oncol 1995;11:406–410. 27. Ahlquist DA: Aggressive polyps in hereditary nonpolyposis colorectal cancer: Targets for screening. Gastroenterology 1995; 108:1590–1592. 28. Bensen S, Mott LA, Dain B, Rothstein R, Baron J: The colonoscopic miss rate and true one‐year recurrence of colorectal neoplastic polyps. Polyp Prevention Study Group. Am J Gastroenterol 1999;94:194–199. 29. Green RJ, Metlay JP, Propert K, Catalano PJ, Macdonald JS, Mayer RJ, Haller DG: Surveillance for second primary colorectal cancer after adjuvant chemotherapy: An analysis of Intergroup 0089. Ann Intern Med: 2002;136:261–269. 30. Viel JF, Studer JM, Ottignon Y, Hirsch JP; Franche‐Comté Polyp Surveillance Study Group: Predictors of colorectal polyp recurrence after the first polypectomy in private practice settings: A cohort study. PLoS ONE 2012;7:e50990. 31. Fajobi O, Yiu CY, Sen‐Gupta SB, Boulos PB: Metachronous colorectal cancers. BJS 1998;85:897–901. 32. Fante R, Roncucci L, Di Gregorio C, Tamassia MG, Losi L, Benatti P, et al.: Frequency and clinical features of multiple tumors of the large bowel in the general population and in patients with hereditary colorectal carcinoma. Cancer 1996;77:2013–2021. 33. Singer JD, Willet JB: Applied longitudinal data analysis: Modeling change and event occurrence. In: Singer JD, Willett JB, editors. The cumulative hazard function, 1st edition. USA: Oxford University Press; 2003.Chapter 13.4. ISBN0195152964. 34. Rex DK, Kahi CJ, Levin B, Smith RA, Bond JH, Brooks D, et al.: Guidelines for colonoscopy surveillance after cancer resection: A consensus update by the American Cancer Society and the US Multi‐Society Task Force on Colorectal Cancer. Gastroenterology 2006;130:1865–1871. 35. Bouvier A‐M, Latournerie M, Jooste V, Lepage C, Cottet V, Faivre J: The lifelong risk of metachronous colorectal cancer justifies long‐term colonoscopic follow‐up. Eur J Cancer 2008;44:522–527.
