Pit pattern analysis using acetic-acid magnification chromoendoscopy in predicting histopathology of small colorectal polyps. The diagnostic yield and intra-/inter-observer reproducibility

Introduction
Acetic-acid magnification chromoendoscopy gains significance in gastrointestinal endoscopy. Cellular mechanisms allowing acetic acid to be used in chromoendoscopy have been recognized [1]. Studies concerning its usefulness in detection of intestinal metaplasia, dysplasia and early cancer in Barrett’s esophagus [2-6] as well as early gastric carcinoma [6] have been published. Acetic acid was also proposed as a support in magnification chromoendoscopy to judge small colorectal polyps [7, 8].
Pit pattern analysis can be implemented in differential diagnosis between non-neoplastic and neoplastic colorectal lesions [9, 10]. The classification of pit patterns described by Kudo [10] distinguishes 6 types of pit patterns: I (typical of normal colonic mucosa), II (attributable to hyperplastic polyps in most cases) and types III (L and S), IV and V (usually neoplastic lesions).
Studies dealing with the use of the acetic acid in colonoscopy are scarce [7, 8]. Intra-observer and inter-observer reproducibility in classifying small colorectal polyps to particular pit patterns by acetic-acid magnification chromoendoscopy have not been established.

Aim
The aim of the study was to analyse the use of acetic-acid magnification chromoendoscopy in prospective analysis of pit patterns of small colorectal polyps. The particular aims were to assess (1) the diagnostic yield of acetic-acid chromoendoscopy and Kudo’s classification of pit patterns in predicting neoplastic histopathology of small colorectal polyps and (2) intra- and inter-observer reproducibility in classifying small colorectal polyps to non-neoplastic and neoplastic pit patterns.

Material and methods
This prospective study was carried out during one month (February 2006) in the endoscopy unit that is a part of the tertiary referral centre for gastrointestinal diseases. The unit works on the basis of an open access system. Every endoscopist participating in this study performed about 200 procedures with magnifying colonoscopy before the study. A single gastroenterologist (M.K.) with a 15-year experience in colonoscopy and 2-year experience in magnification chromoendoscopy performed all colo-noscopies. Possible contraindications were evaluated in every patient. The performance of the examination did not infringe the safety of patients in any aspect.
The study was performed on patients referred for diagnostic endoscopy, in whom at least one polyp of less than 1 cm was found. The patients referred for therapeutic or urgent endoscopy and those with no colorectal polyps were not included in the study. Patients with a history of inflammatory bowel disease, hereditary nonpolyposis colorectal cancer, familial adenomatous polyposis, with a history of previous colonoscopy or referred for surveillance endoscopy were not included to the study.
At the first stage of the study, colonoscopy was performed. Three to four litres of a polyethylene glycol solution was used for bowel preparation. Standard sedoanalgesia (with propofol or a combination of pentanyl, midazolam and short-acting barbiturates) was used. Sedation with monitoring (ECG, pulsoximetry) was done by an anesthesiologist. Magnification colonoscope CF Q160Z (Olympus Corp., Tokyo, Japan) with magnification controller Olympus MAJ-570 (Olympus Corp., Tokyo, Japan) were used. Videoendoscopy processor Olympus EXERA CV 160 (Olympus Corp., Tokyo, Japan) and monitor Olympus OEV 203 (Olympus Corp., Tokyo, Japan) were also used. At this stage, the patients in whom colonoscopy revealed lesions attributable to the inflammatory bowel disease, polyposis syndrome and/or colorectal cancer were excluded.
At the second stage, acetic acid magnification chromoendoscopy was performed in patients with at least 1 colorectal polyp with a diameter of no more than 1 cm. The diameter and localization of the lesion were recorded. An open biopsy forceps technique was used to establish the size of the polyp. A solution (1.5%) of acetic acid of approximately 10 ml (5-15 ml), using a syringe with a cushion of air, was spread over the area of the polyp through the working channel of the colonoscope. The polyp was later inspected under the magnification of 50-150× (Figures 1, 2) and its pit pattern according to Kudo’s classification was defined as non-neoplastic (I and II) or neoplastic (III L, III S, IV, and V) [10]. Pit patterns were judged in real time. Frozen images of each polyp were recorded for further evaluations. The images were stored as bit maps of approx. 1 MB each, resolution: 96 dpi vertically and 96 dpi horizontally, 720 × 576 pixels, 24-bit color). The films (compression Dvmax codec, approx. 40-50 MB, sequence duration 10-15 s, 25 pictures/s, 24-bit color, 720 × 576 pixels). The original images were transmitted using the standard protocol: digital imaging and communication in medicine (DICOM) to the server that hosted the database. Recording EndoRAAD interface (a module of NetRAAD), software provided by the University Health Care (UHC), Lublin, Poland, was used.
Each polyp was removed either by snare polypectomy or a cold biopsy technique. Specimens were fixed in 4% buffered formalin, processed routinely and stained with hematoxylin and eosin. The histopathologist was not aware of the pit pattern classification of polyps.
At the third stage, the recordings which were digitally stored on the hard disc were anonymized. They were evaluated and attributed to the type of pit patterns by the same endoscopist 4 weeks later. The same recordings were also evaluated by two independent endoscopists (H.B., G.R.). Every endoscopist was blinded for pit pattern judgement proposed by the other. Recordings were not changed and/or graphically modified in any way.
Four pit pattern judgments were established for each polyp: one - judged during the first examination in real time, second - four weeks later by the same endoscopist on the basis of anonymized recordings for intra-observer reproducibility and the next two judgements - established on the basis of recordings - by two independent endoscopists. Inter-observer reproducibility was assessed using video recordings by
3 endoscopists comparing judgements of pit patterns
(3 pairs of comparisons). For simplicity, pit patterns were classified into 2 groups - non-neoplastic (I, II) vs. neoplastic (III L, III S, IV, V). The results of the histopathologic examination were categorized to neoplastic polyps (adenomas, adenocarcinomas) and non-neoplastic polyps (hyperplastic polyps, polyps with normal mucosa).
Assessment of the diagnostic yield of the pit pattern in predicting neoplastic histopathology consisted in the relation between pit patterns conformed to neoplastic and non-neoplastic lesion (evaluated as a diagnostic test) and the result of the histopathologic examination (considered as the gold standard). Sensitivity, specificity, positive (PPV) and negative (NPV) predictive value, and diagnostic accuracy (correct classification rate) of diagnoses of neoplastic pit patterns in predicting neoplastic histopathology of colorectal polyps were calculated.
For assessing intra-observer reproducibility, the percentage of agreement and k coefficient (and 95% of confidence intervals) for pit patterns (non-neoplastic vs. neoplastic) evaluated twice by the same endoscopist (during the colonoscopy and later using the recording) was calculated. Similarly, for assessing inter-observer reproducibility, the percentage of agreement and
k coefficients (and 95% of confidence intervals) for 3 pairs of comparisons for pit patterns (recordings evaluated by three endoscopists) were calculated.
Before investigation, the informed consent for acetic-acid magnification chromocolonoscopy was obtained from each patient. During collection, analysis and interpretation of data, the patients were provided with full anonymity. The examination was performed in conformity with guidelines of the Helsinki Declaration, after written permission was obtained from every examined person.

Results
Demographic data and characteristics of patients are given in Table I. All patients were referred for colonoscopy due to symptoms compatible with clinical scenarios described by the European Panel on the Appropriateness of Gastrointestinal Endoscopy (EPAGE, www.epage.ch). More precisely, the patients fulfilled criteria described as scenarios 8-12 and 18. Clinical scenario number 8 is iron-deficiency anemia (malabsorption syndrome excluded); clinical scenario number 9 is hematochezia (without inflammatory bowel disease - IBD; clinical scenario number 10 is uncomplicated lower abdominal pain of at least 2 months’ duration, without known IBD, anemia or FOBT-positive stools; clinical scenario number 11 is change in bowel habits (predominantly constipation) of at least 2 months’ duration without pain, known IBD, anemia or FOBT-positive stools; clinical scenario number 12 is uncomplicated diarrhoea after exclusion of infectious or malabsorption aetiology and without known IBD, no anemia or bleeding and not referred for colorectal cancer screening; clinical scenario number 18 comprises miscellaneous indications like a lesion at recent barium enema or sigmoidoscopy, FOBT-positive stools, fulminant colitis, endometriosis and unexplained weight loss [11-14]. None of the examined patients showed contraindications for colonoscopy, there were no inconclusive colonoscopies and no complications, either.
Sixteen polyps were localized in rectum, 24 in the sigmoid colon, 3 in the ascending colon, 4 in the transverse colon and 3 in the descending colon. Pit pattern judgement performed in real time (during chromocolonoscopy) revealed 5 polyps with pit pattern I, 25 polyps presented pattern II (Figure 1), 18 polyps - pattern III L (Figure 2) and 2 polyps - pattern IV. Pit patterns III S and V were not encountered. At histopathology, 19 polyps were found to be neoplastic (18 adenomas, 1 adenocarcinoma) and 31 polyps were non-neoplastic (25 hyperplastic and 6 with normal mucosa). Sensitivity, specificity, PPV, NPV and diagnostic accuracy of pit patterns in prediction of neoplastic histopathology of small colorectal polyps were: 63.2% (95% CI 38-84), 74.2% (95% CI 55-88), 60% (95% CI 36-81), 76.7% (95% CI 58-90) and 70% (95% CI 55-82), respectively. Intra-observer reproducibility in classifying polyps to non-neoplastic and neoplastic based on pit patterns was very good [correct classification rate 94% (95% CI 83-99), k value 0.88 (95% CI 0.7-1.0)]. The results of the assessment of inter-observer reproducibility are collected in Table II.

Discussion
The implementation of acetic acid in gastrointestinal endoscopy was promising because the agent is cheap, safe and easily accessible [2, 4, 5, 7, 8]. Besides, the use of acetic acid is rapid and reversible, hence potentially repeatable [2, 5, 8]. In this study we confirm that the technique is easy to use and not associated with side effects. Acetic-acid chromoendoscopy can be used for differentiation of neoplastic and non-neoplastic polyps, thus superfluous polypectomy can be avoided [7].
Togashi et al. [7] reported that the acetic acid is very efficient in highlighting pit patterns, which allows predicting with high accuracy histopathology of small colorectal polyps. In the presented study, the diagnostic yield of the pit pattern analysis in predicting neoplastic histopathology of small colorectal polyps using acetic-acid magnification chromoendoscopy was unsati-sfactory. Our results are clearly inferior to the results of Togashi et al. [7] and are not as good as in indigo-carmine enhanced magnification endoscopy [15, 16]. The source of such different results could be technical aspects as magnifying chromoendoscopy with acetic acid is an operator-dependent method [3, 5].
Fortun et al. [4] pointed out that the pit pattern analysis by acetic-acid magnification endoscopy is characterized in the diagnosis of intestinal metaplasia, dysplasia or early cancer in Barrett’s esophagus by
a moderate to substantial level of inter- and intra-observer disagreement, however, in other studies there was a high variability of these parameters [17, 18]. Huang et al. [9] demonstrated that inter- and intra-observer agreements in the judgement of the colonic pit pattern in indigo carmine chromocolonoscopy were remarkable, with mean k values exceeding 0.7 and 0.8, respectively. However, the authors admit that the endoscopic prediction of histopathological diagnoses based on pit pattern analyses is complex and ‘somewhat artificial’. In the study by Kiesslich et al. [19] inter-observer variation in the judgement of pit patterns during indigo carmine magnification chromocolonoscopy did not exceed 50%. In the present study, intra-observer reproducibility in classifying polyps to non-neoplastic and neoplastic pit patterns was high; however, inter-observer repro-ducibility was unsatisfactory.
In the present study, simplifications in inter-pretation of the pit pattern, grouping pit patterns in non-neoplastic and neoplastic categories were made. Huang et al. [9] suggested that such approach could lead to diminishing the potential of Kudo’s classification. This approach is in the authors’ opinion acceptable in the study with a limited number of cases. Another limitation in our study is potential biases resulting from interpretation of recordings, which can be somewhat different from the real time evaluation of colorectal lesions [9]. We tried to overcome such limitations by interpreting each polyp not only using still pictures but also videofilms. It should also be remembered that different pit patterns may be present within one polyp (e.g. in serrated adenoma) [9].
In conclusion, acetic-acid magnification endoscopy is of a limited value in predicting neoplastic histo-pathology of small colorectal polyps. Intra-observer reproducibility in classifying small colorectal polyps to particular pit patterns is acceptable, although its inter-observer reproducibility is unsatisfactory.

Reference
1. Lambert R, Rey JF, Sankaranarayanan R. Magnification and chromoscopy with the acetic acid test. Endoscopy 2003; 35: 437-45.
2. Canto MI. Acetic-acid chromoendoscopy for Barrett’s esophagus: the “pros”. Gastrointest Endosc 2006; 64: 13-16.
3. Conio M. Esophageal chromoendoscopy in Barrett’s esophagus: “cons”. Gastrointest Endosc 2006; 64: 9-12.
4. Fortun PJ, Anagnostopoulos GK, Kaye P, et al. Acetic acid-enhanced magnification endoscopy in the diagnosis of specialized intestinal metaplasia, dysplasia and early cancer in Barrett’s oesophagus. Aliment Pharmacol Ther 2006; 23: 735-42.
5. Rey JF, Inoue H, Guelrud M. Magnification endoscopy with acetic acid for Barrett’s esophagus. Endoscopy 2005; 37: 583-6.
6. Hoffman A, Kiesslich R, Bender A, et al. Acetic acid-guided biopsies after magnifying endoscopy compared with random biopsies in the detection of Barrett’s esophagus: a prospective randomized trial with crossover design. Gastrointest Endosc 2006; 64: 1-8.
7. Togashi K, Hewett DG, Whitaker DA, et al. The use of acetic acid in magnification chromocolonoscopy for pit pattern analysis of small polyps. Endoscopy 2006; 38: 613-6.
8. Kawamura YJ, Togashi K, Sasaki J, Konishi F. Acetic acid spray in colonoscopy: an alternative to chromoendoscopy. Gut 2005; 54: 313.
9. Huang Q, Fukami N, Kashida H, et al. Interobserver and intra-observer consistency in the endoscopic assessment of colonic pit patterns. Gastrointest Endosc 2004; 60: 520-6.
10. Kudo S. Endoscopic mucosal resection of flat and depressed types of early colorectal cancer. Endoscopy 1993; 25: 455-61.
11. De Bosset V, Gonvers JJ, Burnand B, et al. Appropriateness of colonoscopy: iron-deficiency anemia. Endoscopy 1999; 8:
627-30.
12. Gonvers JJ, de Bosset V, Froehlich F, et al. Appropriateness of colonoscopy: hematochezia. Endoscopy 1999; 8: 631-6.
13. De Bosset V, Gonvers JJ, Vader JP, et al. Appropriateness of colonoscopy: lower abdominal pain or constipation. Endoscopy 1999; 8: 637-40.
14. Gonvers JJ, Bochud M, Burnand B, et al. Appropriateness of Colonoscopy: diarrhea. Endoscopy 1999; 8: 641-6.
15. Su MY, Ho YP, Chen PC, et al. Magnifying endoscopy with indigo carmine contrast for differential diagnosis of neoplastic and nonneoplastic colonic polyps. Dig Dis Sci 2004; 49: 1123-7.
16. Fu KI, Sano Y, Kato S, et al. Chromoendoscopy using indigo carmine dye spraying with magnifying observation in the most reliable method for differential diagnosis between non-neoplastic and neoplastic colorectal lesions: a prospective study. Endoscopy 2004; 36: 1089-93.
17. Mayinger B, Oezturk Y, Stolte M, et al. Evaluation of sensitivity and inter- and intra-observer variability in the detection of intestinal metaplasia and dysplasia in Barrett’s esophagus with enhanced magnification endoscopy. Scand J Gastro-enterol 2006; 41: 349-56.
18. Meining A, Rösch T, Kiesslich R, et al. Inter- and intra-observer variability of magnification chromoendoscopy for detecting specialized intestinal metaplasia at the gastroesophageal junction. Endoscopy 2004; 36: 160-4.
19. Kiesslich R, Thanka Nadar BJ, et al. Interobserver variation of the pit pattern classification: an internet-based survey [abstract] [German] [cited 2007 Jun 7]. Available from: URL: http://www.thieme.de/abstracts/endoheute/abstracts2002/
daten/62.html.
20. Kiesslich R, Neurath MF, Galle PR. Chromoendoscopy and magnifying endoscopy in patient with gastroesophageal reflux disease. Useful or negligible? Dig Dis 2004; 22: 142-7.