GASTROINTESTINAL SURGERY Ann R Coll Surg Engl 2012; 94: 574–578 doi 10.1308/003588412X13373405387771 Management of colorectal polyp cancers S Naqvi, S Burroughs, HS Chave, G Branagan Salisbury NHS Foundation Trust, UK ABSTRAcT INTRODUCTION Management of malignant colorectal polyps is controversial. The options are resection or surveillance. Resec- tion margin status is accepted as an independent predictor of adverse outcome. However, the rate of adverse outcome in polyps with a resection margin of <1mm has not been investigated. METHODS A retrospective search of the pathology database was undertaken. All polyp cancers were included. A single histopathologist reviewed all of the included polyp cancers. Polyps were divided into three groups: clear resection margin, involved resection margin and unknown resection margin. Polyps were also analysed for tumour grade, morphology, Haggitt/ Kikuchi level and lymphovascular invasion. Adverse outcome was defined as residual tumour at the polypectomy site and/or lymph node metastases in the surgical group and local or distant recurrence in the surveillance group. RESULTS Sixty-five polyps (34 male patients, mean age: 73 years, range: 50–94 years) were included. Forty-six had clear polyp resection margins; none had any adverse outcomes. Sixteen patients had involved polyp resection margins and twelve of these underwent surgery: seven had residual tumour and two of these patients had lymph node metastases. Four underwent surveillance, of whom two developed local recurrence. Three patients had resection margins on which the histopathologist was unable to comment. All patients with a clear resection margin had no adverse outcome regardless of other predictive factors. CONCLUSIONS Polyp cancers with clear resection margins, even those with <1mm clearance, can be treated safely with sur- veillance in our experience. Polyp cancers with unknown or involved resection margins should be treated surgically. KEYwORdS Colorectal neoplasm – Colonic polyp – Endoscopy Accepted 21 August 2012 cORRESpONdENcE TO Shehryer Naqvi, Salisbury District Hospital, Odstock Road, Salisbury, Wiltshire SP2 8BJ, UK T: +44 (0)7752 406 654; E: [email protected] Polyp cancers in the colorectum are defined as adenomas 2mm.6 We instituted a retrospective study of all patients with within which an invasive carcinoma has developed and in- a proven polyp cancer over a ten-year period at Salisbury Dis- vaded by direct continuity through the muscularis mucosa trict Hospital to investigate whether a clear resection margin into the submucosa.1 With the advent of the National Health of any distance is associated with an adverse outcome. Service (NHS) Bowel Cancer Screening Programme in 2006, the number of polyp cancers identified at colonoscopy in Methods the UK has increased. The incidence of malignant colorec- tal polyps in the screening programme was 1.88% between Cases of malignant colorectal polyps between March 2000 2006 and 2009.2 There is a debate over how to treat patients and September 2010 were identified retrospectively using after endoscopic polypectomy for polyp cancer. The options the histology database. Our inclusion criteria were any mac- are either a formal surgical resection or surveillance.3,4 roscopic polypoid adenomas with a focus of carcinoma in- Since the cancer has invaded into the submucosa, it has vading into the submucosa. Any cases in which dysplastic the potential to spread via lymphatics and blood vessels. The cells did not invade through the muscularis mucosae (high- incidence of lymph node metastases in malignant polyps is grade dysplasia) were excluded, as were polypoid cancers around 6.7%.3–15 There is also the risk that having resected (ie a lesion with the macroscopic appearance of a polyp but the polyp endoscopically, it may not have been adequately constituting entirely malignant tissue when examined his- removed, leaving a risk of local recurrence. tologically). The notes for these cases were reviewed and It is widely accepted that resection margin status is a reli- data collected on polyp histology, outcomes, and the length able prognostic factor in predicting adverse outcome in re- and nature of follow-up. sected malignant polyps.1,4–7,16 However, most authors state Endoscopy reports for each patient were reviewed and that a clear resection margin is anywhere from 1mm5 to the morphology of polyps was noted as either pedunculated 574 Ann R Coll Surg Engl 2012; 94: 574–578 2167 Naqvi.indd 574 16/10/2012 11:54:14 NAqvI BURROUgHS CHAvE BRANAgAN MANAGEMENT Of cOLOREcTAL pOLYp cANcERS Table 2 Haggitt/Kikuchi levels versus adverse outcomes Haggit/ Haggit/ Haggit/ Haggit Not Kikuchi Kikuchi Kikuchi 4 noted 1 2 3 Clear 11 8 16 9 2 margin Adverse 0 0 0 0 0 outcome Involved 0 0 3 6 7 margin figure 1 Flow diagram of polyp cancers with clear resection Adverse 0 0 1 3 4 margins and rates of adverse outcome outcome residual residual residual tumour tumour tumour 1 recur- 1 recur- rence rence Unknown 2 0 0 0 1 margin Adverse 1 recur- 0 0 0 0 outcome rence Totals 13 8 (12%) 19 15 10 (20%) (29%) (23%) (15%) figure 2 Flow diagram of polyp cancers with involved margins and rates of adverse outcome Statistical significance between the clear resection mar- gin and involved resection margin groups was calculated Table 1 comparison of clear versus involved resection using Fisher’s exact test. margins Resection No adverse Adverse P-value Results margin outcomes outcomes Between March 2000 and September 2010, 68 patients were Clear 46 0 <0.0001 identified who had malignant colorectal polyps based on Involved 7 9 their original histology reports. After re-examination, three patients were excluded for not meeting the inclusion cri- teria (1 polypoid cancer and 2 high grade dysplasia). The or sessile. All specimens were re-examined by a single his- remaining 65 polyps (33 male patients) with a mean age of topathologist at Salisbury District Hospital. Data were col- 73 years (range: 50–93 years) fit the inclusion criteria. lected on resection margin, tumour grade, vascular inva- Forty-six polyps (71%) were identified with cancer-free sion and level of invasion based on studies by Haggitt et al14 resection margins. In this group, 21 (45%) had a resection and Kikuchi et al.15 margin of 0.1–1mm. Sixteen polyps (25%) had cancer in- Patients were divided into four groups: i) resection mar- volving the resection margin. Three polyps (5%) had resec- gin clear by >1mm, ii) equivocal resection margin (0.1– tion margins on which the histopathologist was unable to 1mm), iii) margin involved with tumour, and iv) unknown comment. resection margin status (cases in which the histopathologist Thirty patients (46%) underwent surgery with the re- was unable to comment on the resection margin). Adverse mainder undergoing surveillance with a mean follow-up outcomes were divided into those for patients who had sur- duration of 2.7 years (range: 0.5–5 years). Endoscopic sur- gery and those for patients who were treated conservatively veillance was variable. However, all patients received at with surveillance. Adverse outcomes in the surgical group least a yearly colonoscopy until year two and then a three- were defined as presence of residual cancer at the site of yearly colonoscopy. Forty-one patients (63%) had staging polypectomy and/or lymph node metastasis. Those in the computed tomography and only one revealed any spread surveillance group were defined as local or distant recur- (involved margin group). rence. Follow-up data reviewed included length of follow- In the group that received surgery, three patients (10%) up, patient status and method of surveillance. had lymph node metastases and residual cancer was found Ann R Coll Surg Engl 2012; 94: 574–578 575 2167 Naqvi.indd 575 15/10/2012 09:37:59 NAqvI BURROUgHS CHAvE BRANAgAN MANAGEMENT Of cOLOREcTAL pOLYp cANcERS Table 3 polyp morphology versus adverse outcomes clear Adverse outcome Involved Adverse outcome Unknown Adverse outcome Pedunculated 27 0 3 3 1 0 Sessile 6 0 6 4 1 0 Not noted 13 0 7 3 1 1 Table 4 Vascular invasion versus adverse outcomes Vascular invasion Adverse outcome Clear margin 6 0 Involved margin 3 2 residual tumour Unknown margin 1 0 Table 5 poor tumour grade versus adverse outcomes poor grade Adverse outcome Clear margin 1 0 Involved margin 2 2 Unknown margin 0 – figure 3 graph comparing polyp morphology with resection margin status monary fibrosis, and one is deceased but did not die of their disease. The difference in adverse outcome between the clear in seven patients (23%). In the group that received surveil- margin groups overall and the involved margin group was lance, four patients (11%) had a local or distant recurrence. statistically significant (Table 1). Twenty-five patients (38%) had a resection margin of Three patients (5%) had margins that the histopatholo- >1mm, seven of whom underwent surgery. None of these gist was unable to assess. None of these patients underwent patients had residual tumour at the site of polypectomy surgery. Two had significant co-morbidities precluding sur- on post-operative histological examination and none had gery. One patient who had a re-excision of her original pol- involved lymph nodes. The remaining 18 patients under- yp owing to incomplete initial resection was found to have went surveillance for a mean follow-up period of 2.8 years a local recurrence at one year and died of her disease. The (range: 0.5–5 years) (Fig 1). Three have since died of causes second patient completed five years of follow-up with no re- unrelated to their malignancy. None of these patients had currence. The third completed four years of follow-up with local or distant recurrence on follow-up. One patient was no recurrence. noted to have recurrent polyps at repeat colonoscopy but no Results were stratified against Haggitt and Kikuchi14,15 evidence of malignancy. levels (Table 2). In the clear resection margin group, there Twenty-one patients (32%) had a resection margin was an even spread between all Haggitt and Kikuchi lev- of 0.1–1mm, eleven of whom underwent surgery. None of els. However, there were no adverse outcomes regardless these patients had residual tumour at the site of polypec- of Haggit/Kikuchi level. In the involved margin group, all tomy on post-operative histological examination and none recorded Haggitt/Kikuchi levels were either Haggitt 3 or 4 had any involved lymph nodes. The remaining 10 patients and associated with an even spread of adverse outcomes. underwent surveillance with a mean follow-up duration Polyp morphology was divided into sessile, pedunculated of 3.2 years (range: 0.5–5 years) (Fig 1). Two were lost to or unknown as documented in endoscopy reports and sub- follow-up. The remainder have had no evidence of local or sequent histological examination (Table 3). Overall, there distant recurrence. were 31 pedunculated polyps (47%), 13 sessile polyps (20%) Sixteen patients (25%) had resection margins involved and 21 (32%) whose morphology was not noted. Most pol- at polypectomy (Fig 2). Twelve patients underwent surgery. yps (52%) with a clear resection margin were pedunculated Seven patients (44%) had residual tumour at the polypec- and most polyps in the involved margin group were either tomy site, three of whom had lymph node metastases. The sessile or not noted (81%) (Fig 3). There were no adverse remaining patients were free of residual cancer and lymph outcomes in the clear resection margin group regardless of node metastases. Of the four patients who did not have sur- polyp morphology. gery, two had adverse outcomes: one was deemed unfit for Vascular invasion (Table 4) was noted in ten cases surgery and died of her disease, one declined treatment and (15%). Six of these patients were in the clear resection mar- died of her disease, one was unfit for surgery but had no gin group and none had any adverse outcomes at either evidence of recurrence at two years when he died of pul- surgery or long-term follow-up. Three patients were in the 576 Ann R Coll Surg Engl 2012; 94: 574–578 2167 Naqvi.indd 576 15/10/2012 09:37:59 NAqvI BURROUgHS CHAvE BRANAgAN MANAGEMENT Of cOLOREcTAL pOLYp cANcERS involved margins group; two of these had residual tumour data suggested that a resection margin of >0mm in subse- in their resected specimens at surgery. quently surgically resected specimens was significantly as- We found only three patients (5%) with a poor grade tu- sociated with no residual cancer. Tumour involvement at mour on histology (Table 5). One patient was in the clear the excision margin was associated with residual tumour. margin group and had recurrent polyps but no malignancy. As with other studies looking at malignant colorectal The remaining two were in the involved margin group and polyps,4 the major limitation in our study was its retrospec- both had residual tumour in their resected specimens at tive nature. This is especially reflected in the group of pa- surgery. tients in the unknown margin group. We acknowledge that no current multidisciplinary team setting would accept a re- port with no assessment of resection margin. These patients discussion were included as resection margin status has been poorly Since the introduction of the NHS Bowel Cancer Screening noted historically. Programme in 2006, an increasing number of polyp cancers All of the polyp cancers identified in our cohort were have been identified as incidental findings at endoscopic re-examined by a single histopathologist. The original his- polypectomy, which matches our own data. This highlights topathology reports for each specimen were also reviewed. the need for a paradigm to decide which patients should be These data revealed eight reports between 2000 and 2005 treated conservatively and which require surgical manage- with no mention of resection margin status. In three of these ment.3 reports the pathologist was unable to comment on the re- A number of histological and endoscopic parameters section margin due to the alignment of the specimen. If the have been investigated to determine which resected polyps morphology of the polyps was compared with the resection are at highest risk of lymph node metastasis and/or local margin status of the original reports, it was found that five recurrence. These include presence of carcinoma at the (62%) of the eight patients with an unknown resection mar- resection margin,1,4–7,16 morphology of the polyp,14,15 grade gin had sessile polyps whereas only six (13%) had sessile of carcinoma4,7,11,13,14 and presence of lymphovascular inva- polyps in the clear resection margin group. The suggestion sion.4,6,17 is that because sessile polyps are removed piecemeal at en- We investigated primarily the association between clear doscopy, it is often difficult for the pathologist to identify the resection margin and adverse outcome. Resection margin resection margin clearly. The fact that only three polyps had status is widely accepted as an independent risk factor for resection margins on which the histopathologist could not adverse outcome.1,4–7,16 Our own data revealed 16 patients comment highlights that the situation is improving with ex- with involved margins, of whom 5 (31%) had an adverse perience. outcome. None of these five had another positive predictor In our study, polyp morphology itself did not affect the of adverse outcome. More strikingly, in the clear resection rate of adverse outcome. Haggit et al postulated that all ses- margin group, there were no adverse outcomes despite 15 sile polyps were high risk and were assigned level 4 in their specimens (23%) with 1 or more adverse risk factors (poor classification.14 However, Kikuchi et al noted that 32 of 105 tumour grade, vascular invasion, Haggitt/Kikuchi 4). sessile polyps were level Sm1 in their classification system Similarly, Netzer et al found that of 24 patients with in- and had the same risk of adverse outcome as all other Sm1 volved margins, 9 (37.5%) had an adverse outcome and in polyps.15 Our data support this and other studies4 in that 5 of these, resection margin status was the only risk fac- there was a higher proportion of sessile polyps in the high- tor.6 Seitz et al postulated that the use of diathermy to resect risk involved margin group but if the resection margin was colorectal polyps induces post-diathermy necrosis in any re- clear, they should be treated as any other polyp with a clear sidual tumour remnant.16 This would explain why there was resection margin. no residual tumour in five patients with a positive resection Ten specimens with vascular invasion were found in margin. our dataset. Six (60%) of these were in the clear resection A resection margin of 0.1–1mm is considered an in- margin group and had no adverse outcomes. Three (30%) volved margin and most studies in the literature would ad- were in the involved margin group and two (66%) of these vise surgery. Cooper et al found that 21.4% of cases with had an adverse outcome. This suggests vascular invasion is cancer at or near the resection margin (0.1–1mm) had an a significant risk factor for adverse outcome in polyp can- adverse outcome but it was not clear how many of these had cers although our numbers were too small to draw any firm a measurable cancer–margin distance.5 Netzer et al noted, conclusions. Evidence in the literature suggests that vascu- however, that all patients with a clearly cancer-free resec- lar invasion does not correlate well with outcomes in polyp tion margin, even those with a cancer–margin distance of cancers and is therefore of poor prognostic value.4,18 <2mm, had no adverse outcome.6 Nevertheless, their study Poorly differentiated tumours are rare in polyp can- still advised a resection margin of ≥2mm as ‘safe’. Our own cers. We found only three specimens (4%). In our literature data found 20 patients with a resection margin of 0.1–1mm. search, the mean incidence was 3.1%.4,7,11,13,14 Some stud- None of these patients had an adverse outcome in either ies have postulated that it is a significant risk factor on ac- surgical or surveillance groups. This included nine cases count of its aggressive nature 4,5,7 but it has not been found (45%) with a resection margin of 0.1–0.5mm. to be significant in around half the studies.6,15 In our data- A study by the Northern Region Colorectal Cancer Audit set, poorly differentiated tumour grade was not associated Group on 386 polyp cancers supports this outcome.17 Their with adverse outcome in the clear resection margin group. Ann R Coll Surg Engl 2012; 94: 574–578 577 2167 Naqvi.indd 577 15/10/2012 09:38:00 NAqvI BURROUgHS CHAvE BRANAgAN MANAGEMENT Of cOLOREcTAL pOLYp cANcERS 5. Cooper HS, Deppisch LM, gourley WK et al. Endoscopically removed malignant However, two poorly differentiated tumours were found in colorectal polyps: clinicopathologic correlations. Gastroenterology 1995; 108: the involved margin group and both resulted in adverse 1,657–1,665. outcomes. Further research is needed on a large cohort of 6. Netzer P, Forster C, Biral R et al. Risk factor assessment of endoscopically polyps with poorly differentiated tumours before firm con- removed malignant colorectal polyps. Gut 1998; 43: 669–674. 7. Morson BC, Whiteway JE, Jones EA et al. Histopathology and prognosis of clusions can be reached. malignant colorectal polyps treated by endoscopic polypectomy. Gut 1984; 25: 437–444. conclusions 8. Nivatvongs S, Rojanasakul A, Reiman HM et al. The risk of lymph node metastasis in colorectal polyps with invasive adenocarcinoma. Dis Colon In our study resection margin status is an independent risk Rectum 1991; 34: 323–328. 9. Chantereau MJ, Faivre J, Boutron MC et al. Epidemiology, management, and factor for adverse outcome. A polyp cancer with an involved prognosis of malignant large bowel polyps within a defined population. Gut resection margin should be considered for surgical resec- 1992; 33: 259–263. tion. A clear resection margin (of any distance, even those 10. Richards WO, Webb WA, Morris SJ et al. Patient management after endoscopic ≤1mm) can be considered low risk and therefore managed removal of the cancerous colon adenoma. Ann Surg 1987; 205: 665–672. 11. Hermanek P, gall FP. Early (microinvasive) colorectal carcinoma. Pathology, non-surgically. diagnosis, surgical treatment. Int J Colorectal Dis 1986; 1: 79–84. The incidence of polyp cancers is set to rise with in- 12. Coverlizza S, Risio M, Ferrari A et al. Colorectal adenomas containing invasive creased numbers being identified in the bowel cancer carcinoma. Pathologic assessment of lymph node metastatic potential. Cancer screening programme and, consequently, a clear treatment 1989; 64: 1,937–1,947. algorithm needs to be devised to treat patients correctly and 13. Kyzer S, Bégin LR, gordon PH, Mitmaker B. The care of patients with colorectal polyps that contain invasive adenocarcinoma. Endoscopic safely. Polyp specimens removed piecemeal are much more polypectomy or colectomy? Cancer 1992; 70: 2,044–2,050. difficult to interpret by the histopathologist and have the po- 14. Haggitt RC, glotzbach RE, Soffer EE, Wruble LD. Prognostic factors in tential to lead to a resection margin status that is difficult or colorectal carcinomas arising in adenomas: implications for lesions removed by impossible to assess. The aim of polypectomy should there- endoscopic polypectomy. Gastroenterology 1985; 89: 328–336. 15. Kikuchi R, Takano M, Takagi K et al. Management of early invasive colorectal fore be to remove the specimen intact. cancer. Risk of recurrence and clinical guidelines. Dis Colon Rectum 1995; 38: 1,286–1,295. References 16. Seitz U, Bohnacker S, Seewald S et al. Is endoscopic polypectomy an adequate 1. Mitchell PJ, Haboubi NY. The malignant adenoma: when to operate and when therapy for malignant colorectal adenomas? Presentation of 114 patients and to watch. Surg Endosc 2008; 22: 1,563–1,569. review of the literature. Dis Colon Rectum 2004; 47: 1,789–1,796. 2. Lee T, Nickerson C, McNally R, Rutter R. Colorectal polyp cancers in the NHS 17. gill M, Rutter M, Holtham S. Management and short term outcome of Bowel Cancer Screening Programme. Presented at: National Cancer Research malignant colorectal polyps in the North of England. Colorectal Dis 2012 Jun Institute Cancer Conference; November 2010; Liverpool. 18. [Epub ahead of print.] 3. Colacchio TA, Forde KA, Scantlebury vP. Endoscopic polypectomy: inadequate 18. geraghty JM, Williams CB, Talbot IC. Malignant colorectal polyps: venous treatment for invasive colorectal carcinoma. Ann Surg 1981; 194: 704–707. invasion and successful treatment by endoscopic polypectomy. Gut 1991; 32: 4. Hassan C, Zullo A, Risio M et al. Histologic risk factors and clinical outcome 774–778. in colorectal malignant polyp: a pooled-data analysis. Dis Colon Rectum 2005; 48: 1,588–1,596. 578 Ann R Coll Surg Engl 2012; 94: 574–578 2167 Naqvi.indd 578 15/10/2012 09:38:00