Biopsy of Colorectal Polyps Is Not Adequate for Grading of Neoplasia

• Introduction
• Patients and Methods
• Ethical Considerations
• Statistical Methods
• Results
• Discussion
• References

Background and Study Aim: Valid tissue sampling of colorectal adenomas is crucial for their management in terms of treatment and follow-up. The aim of this study was to assess the validity of a cold biopsy sample as representative for the whole polypectomy specimen, with regard to histopathological features.
Patients and Methods: As part of the Norwegian Colorectal Cancer Prevention trial, 442 participants (60 % men) who fulfilled the criterion of colonoscopic recovery of adenoma that had been biopsied at flexible sigmoidoscopy, had their adenomas subsequently removed by polypectomy (snare resection) at colonoscopy. Logistic regression analysis was used to determine which variables contributed to the histopathological discrepancy between cold biopsy and polypectomy specimens.
Results: Among the 532 colorectal adenomas biopsied at flexible sigmoidoscopy and removed by colonoscopy, the assessment of intraepithelial neoplasia (dysplasia) status was changed in 51 adenomas (10 %), and 38 (7 %) of them had been underestimated at biopsy compared with polypectomy. Likewise, the assessment of villousness was changed in 45 adenomas (9 %), being upgraded in 26 (6 %) at polypectomy compared with biopsy. In a multivariate model, the diameter of neoplasia at polypectomy was positively associated with increased risk of the underestimation of intraepithelial neoplasia and/or villousness influencing a diagnosis of advanced colorectal neoplasia, when cold biopsy and polypectomy specimens were compared (P _trend = 0.01). Among 56 cases of advanced neoplasia, 35 (63 %) showed only low-grade intraepithelial neoplasia on biopsy.
Conclusions: Biopsy-based diagnosis underestimated histopathological diagnosis in about 10 % of colorectal adenomas detected by flexible sigmoidoscopy screening, but advanced neoplasia was underestimated in more than 60 %. Efforts must be made to obtain polypectomy specimens to secure precise diagnosis.

Introduction

The main importance of colorectal adenomas is their relationship to colorectal cancer. Information about the histopathological features of polyps is crucial for their management, as non-neoplastic, low-risk, and advanced neoplasia will require completely different strategies for treatment and follow-up. Since advanced adenomas are defined by size, grade of intraepithelial neoplasia (dysplasia), and villousness, valid tissue sampling is a prerequisite for recommendations on therapy and surveillance.

In the presence of multiple, proximally located colonic polyps, re-intubation at the same colonoscopy session may be needed for each polyp, to retrieve polypectomized lesions too large to be extracted through the biopsy channel. This may be exhausting for the patient and resource-demanding. It is simpler to just take a biopsy prior to or after a snare polypectomy without retrieving the whole polyp. A prerequisite for this must be that the biopsy specimen is representative of the histology of the whole polyp. Very few studies have addressed this question [1], in spite of some reports on a proximal shift of site distribution of colorectal neoplasia, and an increasing temptation to reduce the number of colonoscopy re-intubations by limiting tissue sampling to volumes retrievable through the biopsy channel [2].

The present study was part of the NORwegian Colorectal CAncer Prevention (NORCCAP) trial, recruiting individuals from an average-risk population of men and women aged 50 - 64 years [3] [4]. The aim of the study was to elucidate the validity of a cold biopsy sample as representative for the whole polypectomy specimen with regard to histopathological features.

Patients and Methods

A detailed description of the NORCCAP study design has been published elsewhere [3] [4]. Briefly, this is a prospective, controlled study where 20 780 men and women (1 : 1), aged 50 - 64 years, have been drawn randomly from the population registry of the city of Oslo and the county of Telemark. Invitees were offered a once-only screening examination by means of flexible sigmoidoscopy or a combination of flexible sigmoidoscopy and a three-day fecal occult blood test (FOBT).

We used disposable cold biopsy forceps and polypectomy snares (Radialjaw biopsy forceps, and Sensation snares; Boston Scientific Corporation, Watertown, Massachusetts, USA). Polyp location was estimated using endoscopic landmarks in addition to distance from the anal verge measured at withdrawal of the endoscope. Polyp size was estimated in situ, by using closed (2-mm) or open (7-mm) biopsy forceps held against the lesion at flexible sigmoidoscopy, and measured after polypectomy. The number of biopsies taken from each polyp was left to the endoscopist to decide. Biopsy specimens were immersed in formalin fixative. More than one biopsy fragment examined by the pathologist was used to express multiplicity of biopsies taken. Any polyp ≥ 10 mm or any biopsy-verified adenoma < 10 mm qualified for a full colonoscopy.

Polyps were removed at colonoscopy using a diathermy snare. Biopsy samples of polyps taken at flexible sigmoidoscopy and polyps removed at colonoscopy were examined by the same or a different pathologist with a special interest in gastrointestinal pathology. A handout including reproduction of text and figures concerning intraepithelial neoplasia, from the World Health Organization (WHO) guidelines [5], was given to the participating pathologists at consensus meetings, to reduce interobserver variation. The histological classification followed the WHO guidelines with some modifications [3]. An adenoma in the gastrointestinal tract is equivalent to intraepithelial neoplasia. Mild/moderate and severe intraepithelial neoplasia were defined as low-grade and high-grade intraepithelial neoplasia, respectively. The only clinical information available to the examining pathologist was that the specimen in question represented a biopsy/biopsies of a polyp or a polypectomy specimen from the colon or rectum. The pathologist examining the polypectomy specimen did not check the histopathological diagnosis of the biopsy specimen when examining the polypectomy specimen, nor was he/she aware of whether the polyp taken at colonoscopy was a recovered or new lesion or had been missed at flexible sigmoidoscopy. In those cases where the cold biopsy specimen showed cancer, participants were recommended to undergo full colonoscopy to exclude synchronous neoplasia before operation. Cases of cancer diagnosed at biopsy proved not to be suited for polypectomy and were therefore excluded from the present analysis. Thus, malignant lesions in the present study represent lesions with a benign diagnosis at biopsy proving to be cancer after examination of the polypectomy specimen.

To minimize the risk of erroneous classification of polyps missed at flexible sigmoidoscopy being registered as recovered at colonoscopy, cases with more than one adenoma in any one bowel segment or less than 10 cm apart in neighbouring segments were excluded. Discrepancy in intraepithelial neoplasia was defined as either an up- or downgrading of intraepithelial neoplasia in the polypectomy specimen when compared with the initial biopsy. The same applied for discrepancy in villousness. An advanced neoplasia was defined as an adenoma measuring ≥ 10 mm in diameter and/or with villous components and/or showing high-grade intraepithelial neoplasia or adenocarcinoma.

Ethical Considerations

The Regional Ethics Committee and the National Institute of Data Inspection approved the NORCCAP study protocol. Written informed consent was obtained from all participants before entering the trial.

Statistical Methods

Logistic regression analysis was used to determine which variables contributed to the histopathological discrepancy between cold biopsy and polypectomy specimens. Inter- and intraobserver variation was based on the ability to reproduce three diagnostic morphological categories (low-grade and high-grade intraepithelial neoplasia, and cancer) and the presence of villous components (villous/tubulovillous versus tubular adenoma). The variables were: multiplicity of biopsies taken (multiple or single), same or different pathologist and maximum diameter of each neoplastic polyp, in addition to villousness and intraepithelial neoplasia. The association between the independent variables and any discrepancy (dichotomized into yes/no) in the evaluation of histopathological features was expressed as odds ratios (ORs) with 95 % confidence interval (CI). A univariate analysis was carried out initially for each variable, and subsequently all variables were included in a multivariate logistic regression analysis. The statistical software SPSS 11.0 (SPSS, Chicago, Ilinois, USA) was applied for these analyses.

Results

Among 12 960 screenees, 2154 (17 %) participants had neoplasms verified by biopsy at flexible sigmoidoscopy. A total of 442 (21 %) fulfilled the criterion defining colonoscopic recovery of adenoma that had been biopsied at flexible sigmoidoscopy and removed by snare resection at colonoscopy (265 men [60 %] and 177 women [40 %]). The numbers of participants having one, two, three, or four neoplasias recovered for polypectomy were 361 (82 %), 73 (16 %), 7 (1.8 %), and 1 (0.2 %), respectively, totalling 532 colonic neoplasias. The size distribution of neoplasias was 193 measuring 2 - 5 mm (36 %), 196 measuring 6 - 9 mm (37 %), and 143 measuring ≥ 10 mm (27 %).

The discrepancy in intraepithelial neoplasia and villousness observed by examination of the cold biopsy and polypectomy specimens is given in Tables [1] and [2]. Importantly, the assessment of intraepithelial neoplasia status was changed in 51 (10 %) adenomas after polypectomy. Of these, 35 (7 %) were diagnosed as low-grade intraepithelial neoplasia based on biopsy, but assessed as high-grade intraepithelial neoplasia or malignant after polypectomy. In addition, 3 out of 13 (23 %) assessed as high-grade intraepithelial neoplasia at biopsy turned out to be carcinomas, thus 38 (7 %) out of 532 neoplastic polyps would have been underestimated if the diagnosis had been based on biopsy alone. Conversely, for 34 adenomas with high-grade intraepithelial neoplasia at flexible sigmoidoscopy biopsy, in 13 (38 %) the status assessment was changed to low-grade intraepithelial neoplasia after polypectomy.

In a total of 45 (9 %) out of 528 adenomas, the assessment of villousness status was changed after examination of the polypectomy specimen. After polypectomy, 26 (6 %) adenomas judged as tubular at flexible sigmoidoscopy were assessed to be tubulovillous/villous adenoma, and would have been erroneously classified as tubular adenomas by biopsy alone. In 19 (29 %) adenomas the tubulovillous/villous classification at flexible sigmoidoscopy was changed to tubular adenoma.

For neoplasias measuring 2 - 5 mm, 6 - 9 mm and ≥ 10 mm in diameter, the underestimations of intraepithelial neoplasia were 4 %, 10 % and 24 %, respectively (Table [3]). There was no underestimation of intraepithelial neoplasia or villousness for the smallest neoplasias measuring 2 - 5 mm when more than one biopsy was taken.

Multiple biopsies were taken in 30 % of the smallest, 38 % of the intermediate and 55 % of the largest group of lesions (Table [3]). In the two largest categories, underestimation of intraepithelial neoplasia and villousness ranged from 7 % to 20 % for single biopsy and 15 % to 27 % for multiple biopsies, none of the comparisons of single versus multiple biopsies reaching significance level.

In a multivariate analysis, statistical significance was not observed for underestimation of high-grade intraepithelial neoplasia and/or villousness for multiple biopsies when adjusting for whether the same or a different pathologist had examined the two histology preparations and diameter of neoplasia at polypectomy (Table [4]). The risk of underestimation increased significantly when different pathologists examined the two histology preparations. Underestimation of intraepithelial neoplasia and/or villousness influencing a diagnosis of advanced neoplasia was positively associated with polyp size (P _trend = 0.01).

Discussion

The most crucial point in tissue sampling during colonoscopy is not to miss advanced neoplasia. In the present study, however, this occurred even for lesions measuring 2 - 5 mm for single biopsy specimens, but not when multiple biopsies were taken (Table [3]). This indicates that polypectomy and retrieval of the entire polypectomy specimen should be aimed for whenever possible. Smaller polypectomized lesions may be retrieved through the biopsy channel, but very often re-intubation may be required at the same colonoscopy session in the presence of multiple polyps.

It was disturbing to find that 35 of 56 lesions (63 %) with high-grade intraepithelial neoplasia or carcinoma (one case) had been classified only as low-grade intraepithelial neoplasia at biopsy. For villousness the miss rate was 36 % (26 out of 72). This suggests that examination of the entire polypectomized specimen is the only acceptable method on which further treatment and surveillance should be based, unless biopsy indicates a need for surgery (e. g. carcinoma or high-grade intraepithelial neoplasia in a lesion unsuitable for endoscopic resection). A biopsy-based treatment strategy after polypectomy may, however, result in unnecessary surgery if the biopsy sample shows carcinoma. Retrieval of the entire polypectomy specimen might have shown free resection margins and a different recommendation might have been posed to the patient.

Since biopsy and resection specimens were obtained at separate endoscopy sessions, it was crucial to minimize the risk of erroneously identifying a neighbouring polyp as the initially biopsied polyp to be recovered and resected. This risk may be considered negligible in participants with a single neoplasia (82 %). In the remaining group of participants with two (16 %) or more (2 %) neoplasias, the criteria of location in different segments and distance apart were chosen to further the correct recovery classification of lesions without excluding the entire group of participants with multiple lesions. The risk of missing colorectal lesions at one session, to be detected at the next, has been estimated to be in the range of 0 % - 6 % for polyps ≥ 1 cm and 15 % - 20 % for polyps < 1 cm [6] [7]. The present study has shown that the discrepancy between histopathological diagnoses based on biopsy and resection specimens increases with increasing size of the lesion, as shown previously in a small study by Pugliese et al. [1]. This indicates, as expected, that the morphological heterogeneity of lesions increases with size. This was supported by the lack of morphological discrepancy for the smallest category of lesions, measuring 2 - 5 mm in diameter. It may be argued that these smallest lesions may suffer considerably from diathermy artefacts and render only a very small tissue sample for histological evaluation, thus further limiting the probability of detecting any possible discrepancy.

The histopathological features of advanced neoplasias (carcinoma, high-grade intraepithelial neoplasia, or villousness) are particularly subject to interobserver variations between pathologists. In addition to selecting highly qualified pathologists for the task, efforts had been made to further standardize histopathological evaluations within the NORCCAP trial. Nevertheless, in our multivariate analysis, interpathologist variation still proved its role as an independent variable, together with size of lesions, in explaining the discrepancy between the histopathological features in biopsy and polypectomy specimens. These findings have been observed by others, pointing out interobserver variation with regard to intraepithelial neoplasia as a particular challenge [8] [9] [10] [11]. This emphasizes the importance of quality assurance programs in this (and other) diagnostic procedures when subjective judgement may play a pivotal role.

Obtaining multiple biopsies was unexpectedly associated with increased probability of discrepancy between the assessments of intraepithelial neoplasia/villousness status based on biopsy compared with examination of the polypectomized lesion. This is most likely due to the association between multiplicity of biopsies and diameter of the neoplasia. The larger the volume of the neoplasia, the smaller the fraction of the obtained biopsies relative to the neoplasia volume. Accordingly, the probability of obtaining a biopsy with a histopathologic diagnosis different from the polypectomized lesion, is not related to the multiplicity of biopsies as such, but rather to the diameter of the lesion. This is also supported by the observation that the association between biopsy multiplicity and discrepancy lost its significance when adjusted for the diameter of the polypectomized lesion.

It is well known that the topographical distribution of different histological features is not uniform throughout an adenoma [12] [13] [14]. Small foci of high-grade intraepithelial neoplasia and villousness may conceivably be removed by the cold biopsy, leaving the bulk of the adenoma with less severe histological features to be seen in the polypectomized lesion. This may explain the overestimation of severity observed in 38 % of our cases with high-grade intraepithelial neoplasia at biopsy. Further, it may also contribute to the disagreement observed between the two pathologist assessments of the same neoplastic lesion. It must be kept in mind that our interpathologist variation contains this component of interspecimen variation in addition to the inherent interobserver variation usually reported in the literature [8] [9]. Our degree of disagreement is comparable to that reported in these studies, in which the same specimens have been evaluated by different pathologists, suggesting that the pure interpathologist variation in our study is somewhat smaller than reported in the literature.

In conclusion, biopsy-based diagnosis underestimated the histopathological diagnosis in about 10 % of colorectal adenomas detected by flexible sigmoidoscopy screening, but advanced neoplasia was underestimated in more than 60 %. Efforts should be made to obtain polypectomy specimens to secure precise diagnosis.

References

• 1 Pugliese V, Gatteschi B, Aste H. et al . Value of multiple forceps biopsies in assessing the malignant potential of colonic polyps.  Tumori. 1981;  67 57-62
  PubMedSearch in Google Scholar
• 2 Thiis-Evensen E, Hoff G S, Sauar J. et al . Population-based surveillance by colonoscopy: effect on the incidence of colorectal cancer.  Scand J Gastroenterol. 1999;  34 414-20
  CrossrefPubMedSearch in Google Scholar
• 3 Bretthauer M, Gondal G, Larsen I K. et al . Design, organization and management of a controlled population screening study for detection of colorectal neoplasia.  Scand J Gastroenterol. 2002;  37 568-573
  CrossrefPubMedSearch in Google Scholar
• 4 Gondal G, Grotmol T, Hofstad B. et al . The Norwegian colorectal cancer prevention (NORCCAP) screening study. Baseline finding and implementations for clinical work-up in age groups 50 - 64 years.  Scand J Gastroenterol. 2003;  38 635-642
  CrossrefPubMedSearch in Google Scholar
• 5 Jass J, Sobin L H. Histologic typing of intestinal tumours. Berlin; Springer Verlag 1989
  Search in Google Scholar
• 6 Rex D K, Culter C S, Lemmel G T, Rahmani E Y, Clark D W, Helper D J, Lehman G A, Mark D G. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies.  Gastroenterology. 1997;  112 24-28
  CrossrefPubMedSearch in Google Scholar
• 7 Hixson L J, Fennerty M B, Sampliner R E, Garewal H S. Prospective blinded trail of the colonoscopic miss-rate of large colorectal polyps.  Gastrointest Endosc. 1991;  37 125-127
  CrossrefPubMedSearch in Google Scholar
• 8 Rex D K, Alikhan M, Cummings O. et al . Accuracy of pathologic interpretation of colorectal polyps by general pathologists in community practice.  Gastrointest Endosc. 1999;  50 468-474
  CrossrefPubMedSearch in Google Scholar
• 9 Costantini M, Sciallero S, Giannini A. et al . Interobserver agreement in the histologic diagnosis of colorectal polyps. The experience of the multicenter adenoma colorectal study (SMAC).  J Clin Epidemiol. 2003;  56 209-214
  CrossrefPubMedSearch in Google Scholar
• 10 Brown L JR, Smeeton N C, Dixon M F. Assessment of dysplasia in colorectal adenomas: an observer variation and morphometric study.  J Clin Pathol. 1985;  38 174-179
  PubMedSearch in Google Scholar
• 11 Jensen P, Krogsgaard M R, Christiansen J. et al . Observer variability in the assessment of type and dysplasia of colorectal adenomas, analyzed using kappa statistics.  Dis Colon Rectum. 1995;  38 195-198
  CrossrefPubMedSearch in Google Scholar
• 12 Fung C HK, Goldman H. The incidence and significance of villous change in adenomatous polyps.  Am J Clin Pathol. 1970;  53 21-25
  PubMedSearch in Google Scholar
• 13 Lane N, Fenoglio C M. Observation on the adenoma as precursor to ordinary large bowel carcinoma.  Gastrointest Radiol. 1976;  1 111-119
  CrossrefPubMedSearch in Google Scholar
• 14 Silverberg S G. Focally malignant adenomatous polyps of the colon and rectum.  Surg Gynecol Obstet. 1970;  131 103-114
  PubMedSearch in Google Scholar

T. Grotmol, M. D., Ph. D.

The Cancer Registry of Norway

Institute of Population-Based Cancer Research · Montebello · N-0310 Oslo · Norway

Fax: +47-22-451370

Email: tom.grotmol@kreftregisteret.no

References

• 1 Pugliese V, Gatteschi B, Aste H. et al . Value of multiple forceps biopsies in assessing the malignant potential of colonic polyps.  Tumori. 1981;  67 57-62
  PubMedSearch in Google Scholar
• 2 Thiis-Evensen E, Hoff G S, Sauar J. et al . Population-based surveillance by colonoscopy: effect on the incidence of colorectal cancer.  Scand J Gastroenterol. 1999;  34 414-20
  CrossrefPubMedSearch in Google Scholar
• 3 Bretthauer M, Gondal G, Larsen I K. et al . Design, organization and management of a controlled population screening study for detection of colorectal neoplasia.  Scand J Gastroenterol. 2002;  37 568-573
  CrossrefPubMedSearch in Google Scholar
• 4 Gondal G, Grotmol T, Hofstad B. et al . The Norwegian colorectal cancer prevention (NORCCAP) screening study. Baseline finding and implementations for clinical work-up in age groups 50 - 64 years.  Scand J Gastroenterol. 2003;  38 635-642
  CrossrefPubMedSearch in Google Scholar
• 5 Jass J, Sobin L H. Histologic typing of intestinal tumours. Berlin; Springer Verlag 1989
  Search in Google Scholar
• 6 Rex D K, Culter C S, Lemmel G T, Rahmani E Y, Clark D W, Helper D J, Lehman G A, Mark D G. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies.  Gastroenterology. 1997;  112 24-28
  CrossrefPubMedSearch in Google Scholar
• 7 Hixson L J, Fennerty M B, Sampliner R E, Garewal H S. Prospective blinded trail of the colonoscopic miss-rate of large colorectal polyps.  Gastrointest Endosc. 1991;  37 125-127
  CrossrefPubMedSearch in Google Scholar
• 8 Rex D K, Alikhan M, Cummings O. et al . Accuracy of pathologic interpretation of colorectal polyps by general pathologists in community practice.  Gastrointest Endosc. 1999;  50 468-474
  CrossrefPubMedSearch in Google Scholar
• 9 Costantini M, Sciallero S, Giannini A. et al . Interobserver agreement in the histologic diagnosis of colorectal polyps. The experience of the multicenter adenoma colorectal study (SMAC).  J Clin Epidemiol. 2003;  56 209-214
  CrossrefPubMedSearch in Google Scholar
• 10 Brown L JR, Smeeton N C, Dixon M F. Assessment of dysplasia in colorectal adenomas: an observer variation and morphometric study.  J Clin Pathol. 1985;  38 174-179
  PubMedSearch in Google Scholar
• 11 Jensen P, Krogsgaard M R, Christiansen J. et al . Observer variability in the assessment of type and dysplasia of colorectal adenomas, analyzed using kappa statistics.  Dis Colon Rectum. 1995;  38 195-198
  CrossrefPubMedSearch in Google Scholar
• 12 Fung C HK, Goldman H. The incidence and significance of villous change in adenomatous polyps.  Am J Clin Pathol. 1970;  53 21-25
  PubMedSearch in Google Scholar
• 13 Lane N, Fenoglio C M. Observation on the adenoma as precursor to ordinary large bowel carcinoma.  Gastrointest Radiol. 1976;  1 111-119
  CrossrefPubMedSearch in Google Scholar
• 14 Silverberg S G. Focally malignant adenomatous polyps of the colon and rectum.  Surg Gynecol Obstet. 1970;  131 103-114
  PubMedSearch in Google Scholar

T. Grotmol, M. D., Ph. D.

The Cancer Registry of Norway

Institute of Population-Based Cancer Research · Montebello · N-0310 Oslo · Norway

Fax: +47-22-451370

Email: tom.grotmol@kreftregisteret.no