Colorectal cancer (CRC) is a major public health issue in

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2010;8: In Vivo Diagnosis and Classification of Colorectal Neoplasia by Chromoendoscopy-Guided Confocal Laser Endomicroscopy SILVIA SANDULEANU,* ANN DRIESSEN, ENCARNA GOMEZ GARCIA, WIM HAMEETEMAN,* ADRIAAN DE BRUÏNE, and AD MASCLEE* *Department of Gastroenterology and Hepatology, Department of Pathology, and Department of Clinical Genetics, University Hospital Maastricht, Maastricht, The Netherlands See Editorial on page 318. BACKGROUND & AIMS: Colorectal cancer surveillance guidelines rely on clinicohistologic features of adenomas. Unfortunately, in common practice, recording of these features lacks precision and uniformity, which might hamper appropriate follow-up decisions. Confocal laser endomicroscopy (CLE) is a novel technology that allows real-time in vivo microscopy of the mucosa and provides accurate histopathology. The aims of this study were (1) to define and validate differential features of adenomatous and nonadenomatous colorectal polyps by chromoendoscopy-guided CLE (C-CLE) and (2) to assess predictive value of this technique for diagnosis of colorectal neoplasia. METHODS: Patients at risk for colorectal cancer were prospectively investigated by using CLE. During extubation, fluorescein 10% was used in conjunction with acriflavine hydrochloride 0.05% to characterize global tissue architecture as well as cytonuclear features of colorectal epithelium. Ex vivo histology was used as gold standard. Reproducibility tests were performed. RESULTS: In total, 116 colorectal polyps from 72 patients were examined. Ex vivo histology showed 68 adenomas, 6 invasive carcinomas, 30 hyperplastic polyps, and 12 inflammatory polyps. C-CLE of adenomas revealed lack of epithelial surface maturation, crypt budding, altered vascular pattern, and loss of cell polarity. In contrast, C-CLE of nonadenomatous polyps revealed epithelial surface maturation, and minor abnormalities of crypt architecture and of vascular pattern, and maintained cell polarity. Adenoma dysplasia score reliably discriminated high-grade dysplasia from low-grade dysplasia (accuracy, 96.7%). Interobserver agreement was high (K coefficients: pathologist, 0.92; endomicroscopist, 0.88). In vivo histology predicted ex vivo data with sensitivity of 97.3%, specificity of 92.8%, and accuracy of 95.7%. CON- CLUSIONS: C-CLE accurately discriminates adenomatous from nonadenomatous colorectal polyps and enables evaluation of degree of dysplasia during ongoing endoscopy. This technology might offer considerable potential to ultimately fine-tune surveillance programs, particularly in highrisk groups. Keywords: Colorectal Adenomas; Hyperplastic Polyps; Confocal Endomicroscopy. View this article s video abstract at Colorectal cancer (CRC) is a major public health issue in developed countries. 1 Early detection, accurate characterization, and resection of the precursor lesions, adenomatous polyps, are essential for the prevention of this malignancy. 2,3 However, a large proportion of patients with adenomas also have nonadenomatous lesions such as hyperplastic and inflammatory polyps. 4 Conventional endoscopy has limited ability to discriminate adenomatous from nonadenomatous colorectal polyps. Therefore, all identified lesions are routinely removed. This approach results in substantial increase of endoscopic workload, additional pathology costs, and possibly a higher rate of complications. Currently, decisions regarding endoscopic treatment of adenomas and surveillance after polypectomy rely on clinicohistologic features, namely size, multiplicity, grade of dysplasia, and villosity. 5,6 Unfortunately, in common practice, recording of these features frequently lacks precision and uniformity. As such, superficially taken biopsies, sampling error, and loss of polyps for histopathology are common events, making it difficult to draw firm conclusions regarding follow-up. Moreover, there is well-known interobserver variation among pathologists in assessment of villosity and grading of dysplasia. 7,8 All these factors together might lead to underdiagnosis or overdiagnosis of colorectal neoplasia and, hence, suboptimal treatment and surveillance practices. Several studies addressing the outcome of CRC screening indicated that colonoscopy with polypectomy fails to prevent progression to CRC in 10% 25% of cases. 9,10 A large population-based study has shown that colonoscopy might prevent death from left-sided but not from right-sided CRC. 11 Therefore, gastroenterologists and pathologists should unify their efforts to secure precise diagnosis and classification of colorectal neoplasia. Recently, new enhanced imaging techniques evolved that might potentially alleviate these practical issues. It has been shown that chromoendoscopy with methylene blue or indigo carmine and targeted biopsies significantly improves the detection of intraepithelial neoplasia compared with conventional colonoscopy in individuals at increased risk for CRC. 12,13 Abbreviations used in this paper: ADS, adenoma dysplasia score; C-CLE, chromoendoscopy-guided confocal laser endomicroscopy; CI, confidence interval; CRC, colorectal cancer; HGD, high-grade dysplasia; HP, hyperplastic polyp; IC, invasive carcinoma; IP, inflammatory polyp; LGD, low-grade dysplasia; SA, serrated adenoma by the AGA Institute /10/$36.00 doi: /j.cgh

2 372 SANDULEANU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 4 Confocal laser colonoscopy combines the classic white-light endoscopy with real-time microscopy. 14,15 Patients receive a contrast agent, fluorescein, and an endoscopically directed laser excites the targeted tissue. Confocal images are generated simultaneously with the endoscopic images. This technique landscapes large surface and subsurface areas, allowing generation of real-time, in vivo histologic images. Kiesslich et al 16 showed that chromoendoscopy-guided confocal laser endomicroscopy (C-CLE) enables accurate detection and classification of colorectal neoplasia. For this purpose, 2 histopathologic elements are mainly considered, crypt morphology and vascular pattern, which permit differentiation between normal mucosa, regenerative changes, and neoplasia. Nevertheless, additional aspects need to be clarified before embedding this technique in clinical practice. Differentiation of low-grade dysplasia (LGD) and high-grade dysplasia (HGD) might be of relevance in this regard. Acriflavine hydrochloride is a contrast agent that specifically labels the nuclei and hence pinpoints cytonuclear alterations during confocal endomicroscopy. 17,18 We performed a prospective study to systematically evaluate differential features of adenomatous and nonadenomatous colorectal polyps by C-CLE. Fluorescein was used in conjunction with acriflavine to characterize global tissue architecture and cytonuclear features of colorectal epithelium. Special attention was paid to define objective criteria for grading of dysplasia in adenomatous lesions during endomicroscopy examination. Subjects and Methods Subjects Between December 2007 and February 2009, consecutive subjects attending our outpatient clinic for counseling and surveillance of individuals at risk for CRC were recruited in this prospective study. The following subgroups were included: (1) patients with history of advanced adenomas or CRC, (2) patients with Lynch syndrome, and (3) patients with familial CRC. The latter subgroup consisted of subjects with a dominant CRC history, but without known mutation. 19 Exclusion criteria were age 18 years, suboptimal bowel preparation, known polyposis syndromes, pregnancy, lactation, previous documented allergy to intravenously administered sodium fluorescein, and coagulopathy that precludes biopsy sampling or polypectomy. Written informed consent was obtained from all participants, and the study was approved by the local Medical Ethics Committee. Study Procedure Bowel preparation consisted of 3-day low residue diet and 4 L hypertonic polyethylene glycol solution (Klean-Prep). Only patients with adequate bowel preparation (visualization of more than 90% of colonic mucosa) were included. Cecal intubation was documented. All CLE procedures were performed by a single experienced endoscopist (S.S.), who had completed the CLE training program in Mainz and performed 25 complete procedures before commencing the study. The endoscopist was blinded to histologic data and remained blinded during data collection until data analysis. Figure 1 outlines the study methodology. Endoscopies were performed by using the Pentax EC3870CIFK (Pentax Medical Co, Montvale, NJ) colonoscope and ISC-1000 CLE system. In short, the principle of this technique relies on use of a single optical fiber to deliver a pinpointed argon laser light into a Figure 1. Study methodology. specimen. Light emerging from this focused point after administration of an exogenous fluorophore (eg, fluorescein intravenously or acriflavine hydrochloride topically) is recaptured into the same optical system. Confocal images were collected at a scan rate of 0.8 frames/s ( pixels) or 1.6 frames/s ( pixels). The optical plane thickness was 7 m, with a lateral resolution of 0.7 m. The range of z-axis was m below the surface layer, and the field of view was m. After ileocecal intubation, butylscopolamine (20 mg bolus) was administered intravenously to reduce colonic peristalsis. In case of suspect lesions, selective chromoscopy with 5 10 ml indigo carmine 0.4% (SERB S.A.S.; Lab Pharm, Paris, France) was applied via a spraying catheter. Sequential CLE images were generated after intravenous administration of 5 ml fluorescein 10% (SA Alcon-Couvreur NV, Puurs, Belgium). Subsequently, indigo carmine was washed out by using tap water applied through the cleansing channel of the endoscope. Then 5 10 ml acriflavine hydrochloride 0.05% (Sigma-Aldrich Chemie BV, Zwijndrecht, The Netherlands) was topically applied and was allowed to stain the mucosa for about 1 minute. Experimental data suggest that this fluorophore, displaying antiseptic properties, might carry the risk for mutagenicity. 18 Nevertheless, these observations were not substantiated in human studies. After topical application of acriflavine, washing out was performed again. At this point, targeted biopsies were taken from normal-appearing mucosa and from focal lesions. Finally, lesions were endoscopically removed by cold biopsy, snare polypectomy, or by endoscopic mucosal resection.

3 April 2010 DIAGNOSIS OF COLORECTAL NEOPLASIA BY C CLE 373 Endoscopic and Histopathologic Classification of Colorectal Lesions After chromoendoscopy, the morphology of lesions was characterized according to Paris classification. 20,21 Size of lesions was classified as diminutive ( 5 mm), small (6 9 mm), or large ( 10 mm). All specimens were examined by an experienced gastrointestinal pathologist (A.D.), who was unaware of patient information, endoscopic data, and CLE findings. Tissue specimens were sectioned at 4- m intervals, immediately fixed in 10% buffered formalin solution, and subsequently stained with hematoxlyin-eosin. Intraepithelial neoplasia was classified according to the modified Vienna classification. 22 Advanced adenomas were defined as adenomas 10 mm, or containing HGD, or any villous component. 6 Confocal Endomicroscopy: Image Acquisition and Characterization of Lesions Figure 2 depicts the sequence of events for the acquisition of confocal images during endoscopy. Confocal images were collected in a 4-quadrant clockwise fashion from normal colonic mucosa (within 10 cm away from the lesion), as well as from focal lesions. For each colorectal lesion, normal (control) data were available. In case of diminutive, small, and large polyps, images were obtained from at least 6, 8, and 12 separate endomicroscopic fields, respectively. Biopsies were taken 1 mm below the mucosal erythematous landmark induced by the suction port stabilization, as described elsewhere. 16 In case of adenomatous polyps, grading of dysplasia was based on worst area approach. 23 During the procedure, the endoscopist classified all confocal images obtained. General architecture of normal mucosa and of focal lesions was initially characterized after intravenous administration of sodium fluorescein 10% by using the Mainz classification. 16 Cytonuclear features were analyzed in detail after topical administration of ml acriflavine hydrochloride 0.05%. The complete digital video recording of confocal images was also obtained, allowing further review when necessary. Only good quality images, defined as clear pattern recognition, were considered. Study Design and Statistical Calculations This study consisted of 2 consecutive parts: (1) definition and standardization of data, in which diagnostic criteria were described, on an explorative basis; and (2) validation study, in which consecutive patients were prospectively investigated during 1 year. A total of 60 adenomas were found in 72 patients. In our experience, prevalence of adenomas with HGD in patients with Lynch syndrome is about 20% (unpublished data). On the basis of this assumption, we expected that 12 (20%) adenomas would harbor HGD and 48 LGD. The power to detect a difference in adenoma dysplasia score (ADS) of a minimum 3 points, with a standard deviation of 2 points (on the basis of data from the explorative study) and with a significance level ( ) of 5% by using a t test, is As previously shown, 24 it is expected that the power of Mann Whitney U test in this case is Interobserver agreement was tested by using the proportion of agreement (Pa) and weighted Cohen K coefficient (wk). The strength of agreement was considered as slight, ; fair, ; moderate, ; substantial, Figure 2. Algorithm of confocal imaging acquisition during the study ; and high (excellent), (SPSS 16.0; SPSS Inc, Chicago, IL). 25 Nonparametric tests (Mann Whitney U) were used to assess differences between continuous variables. Two-sided P values.05 were assumed to indicate statistical significance. Results Classification of Colorectal Neoplasia by Confocal Laser Endomicroscopy Definition and standardization of data. A total of 2860 CLE images from 20 subjects with histologically proven normal mucosa (subgroup I, n 5), nonadenomatous (subgroup II, n 7), or adenomatous (subgroup III, n 8, of which 4 with LGD and 4 with HGD) lesions were examined. Herewith, we propose a classification of adenomatous and nonadenomatous colorectal lesions by using CLE, as summarized in Table 1. In addition, in an attempt to provide objective criteria to differentiate LGD from HGD by CLE, the ADS was calculated by adding up scores corresponding to the following parameters: (1) epithelial surface maturation, (2) crypt morphology, (3) vascular pattern, and (4) cytonuclear atypia. Each parameter was scored from 0 2, according to severity of changes (Table 2). Reproducibility estimates for confocal laser endomicroscopy image interpretation. This initial set of CLE images was also examined by a second endomicroscopist (W.H.) and a pathologist (A.D.), who were blinded to the patient s records and endoscopy data. The interobserver agreement for classification into the 3 subgroups between endoscopist and pathologist was Pa, 0.94 (95% confidence interval [CI], ) and wk, 0.92 (95% CI, ), whereas the interobserver agreement between endoscopists was Pa, 0.91 (95% CI, ) and wk, 0.88 (95% CI, ), respectively. Validation Study For this purpose, a second series of consecutive subjects were prospectively investigated. Characteristics of study population. Seventy-eight patients were invited to participate in the study. Six patients were excluded as a result of suboptimal bowel preparation (n 4), incomplete colonoscopy related to severe diverticular disease of the sigmoid (n 1), and lack of histopathology (n 1). A total of 72 patients (40 men and 32 women; median age, 56 years [range, years] were included. Twenty-four (33.3%) were patients with history of advanced adenomas and/or CRC;

4 374 SANDULEANU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 4 Table 1. Systematic Classification of Colorectal Lesions by Using C-CLE General architecture Cytonuclear features Normal mucosa Nonadenomatous polyps Adenomatous polyps Regular (uniform) architecture of surface and glandular epithelium Regular honey-comb appearance of vascular pattern Slightly disturbed architecture: enlarged, branch-like, elongated crypts (HP, IP) Increased number of cells in the crypts (mucosal folding, stellar aspect) (HP) Mild alterations of vascular pattern: faint aspect (HP) or slightly dilated, irregular vessels (IP) Inflammatory infiltrate of lamina propria, decreased crypt/stroma ratio (IP) Disturbed architecture: mild irregularity of the crypts (LGD, SA), eventual villous transformation, simple to complex crowding (HGD), causing increased crypt/stroma ratio to completely altered morphology, crypt destruction (IC) Mild to moderate alterations of vascular pattern: dilated vessels, irregular aspect (LGD, HGD); neoangiogenesis, with capillary leakage (IC) Epithelial cells are uniformly lined up along the basement membrane Normal cell polarity of surface and glandular epithelium, normal aspect of mucin-producing goblet cells (epithelial cell maturation) Epithelial cells are morphologically normal; preserved cell polarity Depletion of goblet cells Incomplete to lack of epithelial surface maturation (LGD, HGD, IC) Slight cytonuclear atypia: basally localized, pencillate nuclei, loss of cell polarity with pseudostratification (LGD, SA) to severe cytonuclear atypia: more apically localized, enlarged, roundish nuclei, depletion of goblet cells (HGD) Islands of malignant cells (IC) 20 (27.8%) patients had Lynch syndrome, whereas 28 (38.9%) cases were patients with familial (non-lynch) CRC. The median duration of ileocecal intubation and total procedure time were 15 minutes (range, 8 35 minutes) and 62 minutes (range, minutes), respectively. The procedure was safe and generally well-tolerated, without any serious adverse events. All patients experienced transient mild yellow discoloration of urine caused by excretion of sodium fluorescein. One patient had a mild allergic reaction, possibly as a result of fluorescein. Classification of colorectal lesions by selective chromoscopic colonoscopy. In total, 116 colorectal polyps from 72 patients were examined and classified (Table 3). Table 2. ADS to Discriminate HGD From LGD During Endomicroscopy Examination LGD HGD Epithelial surface maturation : normal 1: incomplete maturation 2: lack of epithelial surface maturation Crypt architecture : normal 1: enlarged, slightly crowded crypts 2: crowding, distorted crypts Vascular pattern : normal 1: slightly increased vascular pattern, preserved hexagonal pattern 2: increased, distorted vessels Cytonuclear atypia 1 2 0: basal, regular nuclei 1: pseudostratification of regular, pencillate nuclei 2: pseudostratification of irregular, large, round, more apically localized nuclei Ademona dysplasia score Characterization of lesions by confocal laser endomicroscopy, with focus on general architecture and cytonuclear features. In total, 928 endomicroscopy fields were scanned. Seventy-eight percent of all CLE images were considered of good quality. Interpretation of the rest of images was precluded as a result of movement artifacts (8%), insufficient resolution (6%), or a combination of these factors (8%). Classification of colorectal lesions was performed in real time during ongoing confocal endomicroscopy, as described in Table 1. Special attention was paid to general architecture, namely crypt architecture and vascular pattern, as well as to cytonuclear features, in particular alterations of cell polarity, size, and morphology of nuclei. Figure 3 illustrates examples of normal mucosa, adenomatous and nonadenomatous colorectal polyps by C-CLE, and corresponding targeted histopathologic specimens. Histopathologic classification of focal lesions. Histopathology showed 68 adenomatous lesions, of which 44 were adenomas with LGD, 16 adenomas with HGD, and 8 were serrated adenomas. There were also 6 invasive carcinomas, 4 well-differentiated and 2 poorly differentiated. In addition, 42 nonadenomatous (30 hyperplastic and 12 inflammatory) polyps were analyzed. Correspondence between histopathology and confocal laser endomicroscopy images. Intraepithelial neoplasia and/or cancer could be predicted in vivo by using the proposed classification of colorectal lesions (Table 1), with sensitivity of 97.3% (72/74), specificity of 92.8% (39/42), and accuracy of 95.7% (111/116). Two serrated adenomas were misclassified as hyperplastic polyps by C-CLE. Both lesions were diminutive. Conversely, 1 hyperplastic and 2 inflammatory (pseudo)polyps were misclassified as adenomas with LGD by C-CLE. Two of these lesions were diminutive, and 1 was small. Characterization of advanced adenomas by confocal laser endomicroscopy. In this study, 32 advanced colorectal lesions were examined. Of them, 12 were large tubu-

5 April 2010 DIAGNOSIS OF COLORECTAL NEOPLASIA BY C CLE 375 Table 3. Characteristics of the Study Population Location Morphology (Paris classification) Patients No. of polyps Size a (mm) Right colon Left colon Rectum 0-Ip 0-Is 0-IIa 0-IIb/c History of advanced adenoma and/or 46 6 (2 18) CRC (n 24) Family history of CRC (n 28) 42 8 (3 32) Lynch syndrome (n 20) 28 8 (4 24) a Values are median (ranges). lar adenomas ( 10 mm), 16 were adenomas with HGD with/ without a villous component, and 4 were villous adenomas with LGD. CLE showed villous features in 6 lesions that were diagnosed as tubular adenomas by ex vivo histopathology. Also, HGD was found by CLE examination in a large flat adenoma (II-a), whereas ex vivo histology showed LGD only. This indicates potential sampling error at biopsy. Differentiation between high-grade and low-grade dysplasia by confocal laser endomicroscopy (adenoma dysplasia score). Endomicroscopy images of adenomatous polyps with histologically proven LGD (n 44) and HGD (n 16) were examined. ADS was calculated by summing scores for epithelial surface maturation, crypt morphology, vascular pattern, and cytonuclear atypia. Median ADS was significantly higher in adenomas with HGD versus those with LGD (7 [range, 4 8] vs 2 [range, 1 5]; P.0001) (Figure 4.). ADS 5 discriminated HGD from LGD with 93.7% sensitivity, 97.7% specificity, and accuracy of 96.7%. Clinical examples are depicted in Figure 5. Discussion This study showed that C-CLE allows accurate differentiation of adenomatous and nonadenomatous colorectal polyps during ongoing endoscopy. C-CLE predicted histopathology with sensitivity of 97.3%, specificity of 92.8%, and accuracy of 95.7%. In line with histopathology principles, special attention was paid to describe general architectural as well as cytonuclear features of colorectal mucosa during CLE examination. We propose a systematic, reproducible classification of colorectal lesions by means of CLE. The following features were characteristic for adenomatous polyps: simple to complex crowding of glands (crypt budding), altered vascular pattern, lack of epithelial surface maturation, depletion of goblet cells, loss of cell polarity, and pseudostratification of irregular nuclei. In addition, cancerous tissue showed crypt destruction and neoangiogenesis with interstitial leakage of fluorescein. In contrast, nonadenomatous polyps were characterized by only minor abnormalities of crypt architecture (star-shaped, branch-like), minor alterations of vascular pattern, epithelial surface maturation, preserved appearance of goblet cells, and maintained cell polarity. Our results are in agreement with literature data indicating that CLE, either endoscopy-based or probed-based techniques, enables accurate in vivo characterization of colorectal neoplasia. 16,26 28 This study also showed that C-CLE is a reproducible tool for classification of colorectal neoplasia when performed in a different center from the one where the basic concept was developed. The Mainz classification provides basic guidelines for pattern recognition to distinguish among normal mucosa, regenerative changes, and neoplasia. However, a systematic approach with focus on architectural changes and cytonuclear atypia is needed when presence and grading of dysplasia are to be evaluated. We defined the ADS, which reliably discriminates HGD from LGD during CLE. In the present study, 32 advanced neoplasias were examined by C-CLE. In 7 (21.8%) of them, histologic diagnosis was upgraded by C-CLE from nonadvanced to advanced adenomas. This was due to detection of either a villous component or of HGD. It is conceivable that systematic scanning of large mucosal areas during C-CLE minimizes the risk of sampling error and thereby might improve diagnostic accuracy. Of note, a considerable proportion of adenomas in this study was small. Although most of these lesions were correctly classified during CLE, 5 (4.3%) small lesions were misclassified. Because small polyps are more difficult to target, small-scale correlations remain difficult for both routine histopathology and CLE examination. Colonoscopic screening and surveillance have steadily increased worldwide. Surprisingly, the protection against CRC remains far from perfect, among other factors, also as a result of inherent limitations of endoscopic techniques. Surveillance policies are usually based on clinicopathologic features of adenomas. Recording of these features in routine practice frequently lacks precision and uniformity, making it difficult to draw firm conclusions regarding follow-up intervals. Biopsy specimens do not represent histology of the whole polyp, leading to underestimation of advanced colorectal neoplasia in up to 60% of cases. 29 Quality improvement of colonoscopic examination in terms of detection and classification of lesions is therefore mandatory. Our data indicate that C-CLE offers considerable potential to refine the diagnosis of colorectal lesions and, hence, to delineate subgroups of patients at high risk for CRC. We believe the strength of this study is that patients were consecutively included via our multidisciplinary outpatient clinic, providing a survey of current practice. During endomicroscopy, systematic examination was performed in an attempt to minimize the risk of sampling error. In addition, optical biopsies of normal-appearing mucosa were taken as control data. The study results reinforce the need to adopt reliable features and terminology for classification of colorectal neoplasia. However, such classification might remain difficult, because most histopathologic features are not pa-

6 376 SANDULEANU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 4 thognomonic and might not clearly distinguish among clinically relevant entities. 30 It is also plausible that progression of adenomas to CRC is merely a continuous process in which different stages coexist rather than a sequential process. Figure 4. ADS was calculated by adding up scores corresponding to the following parameters: epithelial surface maturation, crypt morphology, vascular pattern, and cytonuclear atypia. Median ADS was significantly higher in adenomas with HGD vs those with LGD (P.0001). A few comments have to be made concerning methodologic aspects. Additional use of acriflavine enabled individual cell characterization and highlighted mucosal cytonuclear features. Although some nuclear features, namely size, intracellular location, morphology, and pseudostratification, could be more precisely characterized by combined use of fluorescein and acriflavine, other important aspects (eg, nuclear hyperchromasia, aspect of nucleoli) could not be addressed. This information is particularly important for differentiation of HGD from LGD and for differentiation of hyperplastic polyps from serrated adenomas. The last are putative precursors of CRC via the newly described serrated pathway. 31 In this regard, new contrast agents should be Figure 3. (A) Normal colonic mucosa. (A-1) Endomicroscopy (acriflavine) shows regular crypt architecture. Crypt detail showing normal nuclei (white dots) within the epithelial cells (arrow). (A-2) The corresponding histologic specimens of normal colonic mucosa (hematoxylineosin staining, 30 original magnification; crypt detail, 50 original magnification). (B) Hyperplastic polyp. The luminal openings of crypts show stellar appearance (arrows) by CLE (acriflavine) (B-1) and by conventional histology (B-2, hematoxylin-eosin, 20 original magnification). (C) Serrated adenoma. (C-1) CLE (acriflavine) shows epithelial surface maturation (arrow), regular crypts, and pseudostratification of nuclei (arrowhead). (C-2) Corresponding histologic specimens ( 20 original magnification). (D) Adenomatous polyp. Slight irregularity of the crypts in a tubular adenoma by endomicroscopy (fluorescein) (D-1) and by conventional histology (D-2, 15 original magnification). (E) Invasive carcinoma. (E-1) Endomicroscopy of a poorly differentiated adenocarcinoma shows cryptal destruction, distorted vessels, leakage of fluorescein into the surrounding tissue (neoangiogenesis, arrow), and islands of malignant (dark) cells (arrowhead). (E-2) Corresponding histologic specimens ( 20 original magnification).

7 April 2010 DIAGNOSIS OF COLORECTAL NEOPLASIA BY C CLE 377 Figure 5. Grading of dysplasia in adenomatous polyps. (A) Adenoma with LGD. Endoscopic appearance of flat adenoma (Paris IIa) in a patient with familial CRC (A-1). CLE (acriflavine) shows enlarged, slightly irregular crypts and pseudostratification of nuclei within the crypt (A-2, arrow), and the surface epithelium (A-3, arrowhead). Conventional histology of targeted biopsies shows similar findings (A-4, 15 original magnification). (B) Adenoma with HGD. Endoscopic appearance of adenoma (Paris IIa IIc) in a patient with familial CRC (B-1). Endomicroscopy (acriflavine) shows complex crowding of crypts (B-2, arrow), reaching the surface epithelium, that clearly lacks normal differentiation (B-3, arrowhead). Similar findings by conventional histology (B-4, 12 original magnification). explored to accurately detect and characterize dysplastic tissue. A promising development was reported by Hsiung et al. 32 These authors applied a fluorescein-bound heptapeptide topically to detect colonic dysplasia during endomicroscopy examination. In summary, we have shown that chromoendoscopy in conjunction with confocal endomicroscopy allows in vivo differentiation of adenomatous and nonadenomatous colorectal polyps during ongoing colonoscopy. We propose a systematic classification of colorectal lesions by CLE, yielding high interobserver agreement. The ADS represents a step toward more objective grading of dysplasia. This technology might offer considerable potential to ultimately fine-tune surveillance programs, particularly in high-risk groups. References 1. International Agency for Research on Cancer. Globocan 2002: cancer incidence, prevalence, and mortality worldwide. Available at: Accessed February 26, Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal tumor development. N Engl J Med 1988;319: Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy: the National Polyp Study Workgroup. N Engl J Med 1993;329: Ansher AF, Lewis JH, Fleischer DE, et al. Hyperplastic colonic polyps as a marker for adenomatous colonic polyps. Am J Gastroenterol 1989;84: Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale update based on new evidence. Gastroenterology 2003;124: Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. Gastroenterology 2006;130: Chandler I, Houlston RS. Interobserver agreement in grading of colorectal cancers findings from a nationwide web-based survey of histopathologists. Histopathology 2008;52: Terry MB, Neugut AI, Bostick RM, et al. Reliability in the classification of advanced colorectal adenomas. Cancer Epidemiol Biomarkers Prev 2002;11: Loeve F, van Ballegooijen M, Boer R, et al. Colorectal cancer risk in adenoma patients: a nation-wide study. Int J Cancer 2004; 111: Bressler B, Paszat LF, Vinden C, et al. Colonoscopic miss rates for right-sided colon cancer: a population based analysis. Gastroenterology 2004;127: Baxter NN, Goldwaser MA, Paszat LF, et al. Association of colonoscopy and death from colorectal cancer. Ann Intern Med 2009;150: Kiesslich R, Neurath MF. Chromoendoscopy and other novel imaging techniques. Gastroenterol Clin North Am 2006;35: Bernstein CN. The color of dysplasia in ulcerative colitis. Gastroenterology 2003;124: Sakashita M, Inoue H, Kashida H, et al. Virtual histology of colorectal lesions using laser-scanning confocal microscopy. Endoscopy 2003;35: Polglase AL, McLaren WJ, Skinner SA, et al. A fluorescence endomicroscope for in vivo microscopy of the upper- and the lower GI tract. Gastrointest Endosc 2005;62: Kiesslich R, Burg J, Vieth M, et al. Confocal laser endomicroscopy for diagnosing intraepithelial neoplasias and colorectal cancer in vivo. Gastroenterology 2004;127: Kiesslich R, Goetz M, Angus EM, et al. Identification of epithelial gaps in human small and large intestine by confocal endomicroscopy. Gastroenterology 2007;133: Wainwright M. Dyes in the development of drugs and pharmaceutical. Dyes Pigments 2008;76: Dove-Edwin J, de Jong AE, Adams J, et al. Prospective results of surveillance colonoscopy in dominant familial colorectal cancer

8 378 SANDULEANU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 4 with and without Lynch syndrome. Gastroenterology 2006;130: Participants in the Paris Workshop. The Paris endoscopic classification of superficial neoplastic lesions: oesophagus, stomach, and colon. Gastrointest Endosc 2003;58:S3 S Kudo S. Early colorectal cancer. Tokyo, Japan: Igaku-Shoin, Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of gastrointestinal neoplasia. Gut 2000;47: Hamilton S. Tumours of the digestive systen. Lyon: IARC Press, Zimmerman DW, Zumbo BD. The relative power of the Wilcoxon- Mann Whitney test and Student t test under simple bounded transformations. J Gen Psychology 1990;117: Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 2009;1: Kiesslich R, Goetz M, Rafoud K, et al. Staging of colorectal neoplasia with confocal laser endomicroscopy using two contrast agents simultaneously. Gastrointest Endoscopy 2008; 67:AB Wallace MB, Fockens P. Probe-based confocal laser endomicroscopy. Gastroenterology 2009;136: Sanduleanu S, Driessen A, Hameeteman W, et al. Inflammatory cloacogenic polyp: diagnostic features by confocal endomicroscopy. Gastrointest Endosc 2009;69: Gondal G, Grotmol T, Hofstad B, et al. Biopsy of colorectal polyps is not adequate for grading of neoplasia. Endoscopy 2005;37: Montgomery EA. Biopsy interpretation of gastrointestinal tract mucosa. Philadelphia: Lippincott Williams & Wilkins, 2006: East JE, Saunders BP, Jass JR. Sporadic and syndromatic hyperplastic polyps and serrated adenomas of the colon: classification, molecular genetics, natural history, and clinical management. Gastroenterol Clin N Am 2008;37: Hsiung P-L, Hardy J, Friedland S, et al. Detection of colonic dysplasia in vivo using a targeted heptapeptide and confocal microendoscopy. Nat Med 2009;14: Reprint requests Address requests for reprints to: S. Sanduleanu, MD, PhD, Department of Internal Medicine, Division of Gastroenterology and Hepatology, Maastricht University Medical Center, Postbox 5800, 6202 AZ Maastricht, The Netherlands. s.sanduleanu@mumc.nl; fax: (31) Acknowledgments The authors thank M. Frusch-Brouwer and Ton Mestrom for outstanding technical assistance. Conflicts of interest These authors disclose the following: This study was investigatorinitiated. Dr S. Sanduleanu and Prof Dr A. Masclee were the recipients of an educational grant from Pentax B.V., The Netherlands. The remaining authors disclose no conflicts.