In 1983, the "experts" of a the Inflammatory Bowel Disease-Dysplasia Morphology Study Group met under the leadership of Riddell and Yardley in order to reach an agreement in regard to nomenclature of so-called precancerous lesions in inflammatory bowel disease. The major goals of the group were the same that had motivated the previous symposium: to develop standardized definitions of terms pertaining to inflammatory bowel disease and dysplasia; to establish criteria for diagnosing dysplasia and for differentiating it from other epithelial changes occurring in inflammatory bowel disease; to develop a system for grading and classifying dysplastic changes in inflammatory bowel disease; and to assess the accuracy and precision of the grading system for dysplasia. [24] The definition of dysplasia made by the members of the group read as follows: "Dysplasia is defined as an unequivocal neoplastic alteration of the colonic epithelium. It should be stressed that such dysplastic epithelium not only may be a marker or precursor of carcinoma, but may itself be malignant and associated with direct invasion into the underlying tissue. This definition is analogous to the definition of dysplasia in adenomas of the colon in the absence of IBD [inflammatory bowel disease], in neoplastic lesions of the rest of the gastrointestinal track, and in other epithelia." [24]

The definition of dysplasia as "unequivocal neoplastic alteration of colonic epithelium" is imprecise without further specification. Is it confined to mucosal epithelium and glandular elements or has it passed beyond it? What are the morphologic criteria needed to recognize this "neoplastic alteration"?

The group recommended three major categories, namely, "negative for dysplasia," "indefinite for dysplasia," and "positive for dysplasia." The latter category, in contrast to the previous system of Morson that graded dysplasia as mild, moderate, and severe, was graded just as low grade and high grade. Parenthetically, Morson was also a member of the group. As the authors mentioned, the term "indefinite for dysplasia" was used when the pathologists were unable to differentiate reactive changes from true neoplastic changes and was usually applied to equivocal cases of so-called low grade dysplasia. This category was further subdivided into "probably positive," "probably negative," and "unknown." [24]

The classification and definition of dysplasia in IBD made by the "the experts" of the Inflammatory Bowel Disease-Dysplasia Morphology Study Group was a struggle in vain. Some new terms were created for pathologists to define so-called precursor states to differentiate them from carcinoma and the inflammatory process. However these terms made the subject more complex. In short, the classification was confusing, not repeatable, and not understandable by clinicians. All categories were shown in subsequent studies to be highly subjective, and the precision and accuracy of this grading system for dysplasia was soon called into question, especially in regard to "indefinite for dysplasia" or low grade dysplasia. Studies that assessed the intraobserver and interobserver variability revealed that there was no meaningful statistical agreement by using these definitions among pathologists. For example, in Jensen's work that assessed the observer variability of dysplasia using kappa statistics, he and his associates pointed out that although the results of their study was better than previous ones, there still were marked differences in the intraobserver and interobserver reliability, while interobserver variation was much higher than intraobserver variation. [25] It is clear that disagreement among diagnoses of such lesions makes epidemiologic studies and studies of adenoma or cancer of the colon less reliable and produces problems for further research. No doubt, a precise diagnosis is a cornerstone for evaluation of risk factors, frequency of cancer, management of a patient, and evaluation of any program for prevention of colonic cancer. This classification not only did not work for pathologists to define a lesion exactly, namely, carcinoma at very early stage of development, but also did not help in patient management. The creation of all these terms just left the clinicians confronted with many probabilities far away from a correct diagnosis. However, in spite of complexity and subjectivity, the classification has been accepted by most Western pathologists and has been used for more than 10 years.

Pascal, in 1994, continued to consider adenoma as a precursor of colorectal carcinoma and pointed out that low grade dysplasia is really adenoma or adenoma is just low grade dysplasia. He believed other adenomas that show high grade dysplasia warrant mention. [26] This is what he said about definition of an adenoma: "High grade dysplasia in an adenoma is recognized by the same criteria as those just described for its recognition in IBD. What, then, is low grade dysplasia in noncolitis patients? It is the adenoma itself. Analogous to the transformation of nonneoplastic mucosa to low grade dysplasia in IBD, and also analogous to original concept of the transformation of the normal cervical epithelium to mild dysplasia, adenomatous epithelium can be equated with low grade dysplasia whenever it occur in the bowel. There is no reason to attempt to change the nomenclature of adenomas, but in examining adenomas for additional cancer risk factor only one grade of dysplasia, namely high grade, warrants mention." [26]

In 1997, the Japanese Society for Cancer of the Colon and Rectum published the first English edition (from the fifth Japanese edition) of the book General Rules for Clinical and Pathological Studies on Cancer of the Colon, Rectum, and Anus. [27] The goals of the general rules, as described by the members of the society, were that the criteria were simple, universally applicable, useful, adequate for statistical studies, and internationally acceptable. In this guideline, the authors of the book did not use the terms dysplasia, precancer, precursor, intraepithelial, or intramucosal neoplasia, which Western pathologists had coined in their previous works. First, they separated criteria for histologic classification in surgical specimens from those in colorectal incisional biopsies. The histologic changes in colorectal biopsies were classified in five groups: the first included normal mucosa without atypia and the fifth was for definite carcinoma. Group 3 included "neoplastic lesion with low-grade atypia" that encompassed adenoma with mild or moderate atypia and lesions difficult to diagnose as neoplastic or non-neoplastic ones. Group 4 included "neoplastic lesions with high grade atypia," which encompassed adenoma with severe atypia, borderline lesions between benign and malignant ones, and adenocarcinoma of highly well-differentiated type. [27]

The nomenclature system proposed by members of Japanese Society for Cancer of the Colon and Rectum was totally different from that suggested by the Inflammatory Bowel Disease-Dysplasia Morphology Study Group in 1983. Japanese pathologists had different concepts about definition of the terms adenoma, carcinoma, precancer, precursor, dysplasia, atypia, and neoplasia, and they interpreted the biopsy specimens differently. They never used terminologies such as precancer or dysplasia. Nevertheless, in comparison with the classifications in Western countries, the frequency of early carcinoma in Japan (the carcinoma that has proved by histopathologic examination not to have passed through the submucosal layer) is by far more common than advanced carcinoma (the carcinoma that has proved to have passed muscularis propria). [28–30] This means not only that different criteria and terminologies existed in Japan compared with the West, but also that Japanese and Western pathologists had different views and interpretations of the same morphologic changes. This was highlighted beautifully in a study undertaken by Schlemper and his associates in 1998. [31]

This study included eight pathologists—four from Japan and four from Western countries (USA and Europe) specializing in gastrointestinal neoplasias—who were shown microscopic slides of 20 Japanese patients. The diagnoses in both groups were compared by statistical analysis and the result was very impressive. All of the diagnoses made by all Western pathologists as a whole differed greatly from that of their Japanese counterparts. There was disagreement between the Western and Japanese viewpoints of benign versus malignant for half of the slides. A colectomy specimen of a large polyp was diagnosed as adenoma by half of the Western pathologists and as definite carcinoma by the other half, depending on whether submucosal invasion was interpreted to be absent or present, but this specimen was diagnosed by all Japanese pathologists as definite carcinoma owing to nuclear features such as vesicular nuclei and the disorganized glandular structure. Cytopathologic features such as enlarged, hyperchromatic, or vesicular nuclei with prominent nucleoli were mentioned frequently by both Western and Japanese pathologists, although they were interpreted differently. These cytopathologic features were findings of so-called low grade as well as high grade dysplasia according to the Western pathologists, whereas for Japanese pathologists, they all were morphologic clues for the diagnosis of carcinoma. This is one conclusion from their observations: "This study of colorectal neoplasms shows that Japanese pathologists often diagnose "carcinoma" when Western pathologist use the term "adenoma" to indicate the presence of still benign neoplastic epithelial proliferation with no metastatic potential." [31] The authors stated that the reason for discrepancy in diagnosis of adenoma and carcinoma between Western countries and Japan is due not only to the use of different nomenclature and criteria (for example, invasion as an obligatory criterion for carcinoma in the West but not in Japan; the term dysplasia is used in the West but not in Japan), but also because of differing views and interpretations of histopathologic slides: "We found that there are some nuclear and structural features that Japanese pathologists judge to be clues for the diagnosis of carcinoma but that most western pathologists either do not take into consideration at all (such as rounded nuclei and variable shape of glands) or do not attach similar importance to with regard to severity of dysplasia (such as marked hyperchromatism of nuclei and enlarged prominent nucleoli). The different interpretation of such nuclear and structural features could account for lesions being diagnosed as low grade dysplasia in the west and as carcinoma in Japan." [31]

Because of these large differences among pathologists in definition and interpretation of adenomatous changes, two years later in 2000, (and also in another meeting in 2002 for its revision) 31 well-known and influential pathologists in the field of gastrointestinal neoplasia from 12 different countries were invited to meet in Vienna to review different specimens from colorectal, as well as esophageal and gastric biopsies, to discuss their results, and to develop a new consensus terminology. [32] The extent of diagnostic agreement between those with Western and Japanese viewpoints was assessed using kappa statistics. The results of this meeting evolved into another classification of dysplasia/neoplasia which is celebrated as The Vienna Classification of Gastrointestinal Epithelial Neoplasia (Table 1). In this classification, what are called "epithelial neoplastic lesions" are grouped into five categories. As the authors pointed out, this classification was applicable throughout specimens from the entire gastrointestinal tract and for the diagnosis of both biopsy and resected specimens.

A glance at the table reveals that this classification, just as with previous ones, is a mixture of those from the Japanese Society for Cancer of the Colon and Rectum Classification (5 categories) and from the IBD Study Group Classification (dysplasia instead of atypia). Categorization into five groups was a desirable goal of the Japanese group, including the term dysplasia, which suited the Western view. As shown in Table 1, to reach a better communication between different pathologists from the West to the East, they equated the term dysplasia with neoplasia in categories 1 and 2 but considered it to be the same as adenoma in category 3 and 4. The term dysplasia, however, is not defined in the entire article and the exact place of dysplasia in this classification is not clear; the term is equated with both neoplasia and adenoma. It is interesting that in another classification of neoplasias of the gastrointestinal tract published by WHO in 2001, the term dysplasia was replaced by "intraepithelial neoplasia" and these neoplasias were graded as low and high, the latter encompassing carcinoma in situ. [33] Parenthetically, the Vienna classification underwent revision several times.

However, in spite of several classifications proposed by "the experts" to reach the same understandable language among pathologists themselves and between pathologists and gastroenterologists working in the field of adenomatous polyps and longstanding ulcerative colitis of the colon, it seems that not all pathologists follow this classification and not all clinicians understand the terminology. In an article written by Rex, Ulbright, and Cummings (a gastroenterologist with two pathologists), in 2005, they point out that clinicians still interpret the pathology report of adenomatous polyps differently and there is no general consensus in terminology. They report serious errors in decision-making for management of a patient with colonic polyps that resulted from the misunderstanding of the meaning and implications of a pathology report. They emphasize the importance of correct communication between clinicians and pathologist in managing patients with colorectal polyps: "Confusion may arise for endoscopists because of varying terminology used in the pathology report. For example, the term "atypia" in the description of an adenoma is poorly defined and difficult to interpret. Pathologist often uses this term when they are uncertain whether a lesion is reactive or dysplastic, usually in the setting of inflammation or ulceration. However, in our experience, clinicians often interpret the term "atypia" to mean that higher grade of dysplasia is present." [34]

Based on the article written by Rex et al, West and Mitsuhashi stressed why pathologists may not follow the recommendation proposed by previous classifications: "There are several reasons why pathologists may not follow these recommendations. For example, some may want to convey to their clinical colleagues that a patient has a biologically more advance lesion than high-grade dysplasia sensu stricto when carcinoma in situ or intramucosal carcinoma is present, in particular when a polyp is biopsied or incompletely excised and there may be residual and possibly more advance neoplasm. Moreover, it is difficult to accept the logic of calling a lesion high grade dysplasia in a polypectomy, when the same lesion had been correctly called adenocarcinoma is a biopsy. Using different diagnoses for the same lesion depending on the completeness of excision is a practice that does not make much sense." [35] The authors pointed out how much confusion still exists between pathologists because of different recommendations from different classifications. "It might seem that pathologists are being faced with using different systems of classification depending on whether they are dealing with complete or incomplete excisions, polypectomies or surgical resections, and gastroenterologists or surgeons!" [35]

In a study in 2006, Haboubi states that recommendations for treatment and follow up of dysplasia based on initial classification of it proposed by IBD-Dysplasia Study Group in 1983 [24] not only created more problems with controversy but presented fewer solutions. [36] Among many of his criticisms, he also pointed out that the problem with the term dysplasia is in defining it morphologically with precision and accuracy, and that, for example, such a term as "indefinite for dysplasia" is not understandable. This is what he said about the matter of dysplasia in ulcerative colitis: Once the biopsy is taken the problem is the significant inter and intra-observer variation. This variation is to such an extent that it was recommended to make such a diagnosis of dysplasia there is a need for the biopsy to be seen by two different people, one of them should have a special interest in gastrointestinal pathology. The main problem still lies between differentiations of low grade dysplasia from indeterminate dysplasia rather than high grade dysplasia in which there is a much better concordance. This however may minimize the errors but will not totally eliminate having negative colectomies for patients with particular low grade dysplasia. The other problem is the differentiation between understanding what the term "indefinite for dysplasia" means." [36]

Clearly, it is not possible nor logical to request a second opinion from an expert in gastrointestinal pathology with every biopsy diagnosed as dysplasia. Even if this were possible, it shows how unreliable the diagnosis of dysplasia is. Some studies have shown that nearly 83% of gastroenterologists are not confident with a diagnosis of dysplasia in patients with ulcerative colitis and send the specimen to another pathologist for second opinion. [37] Nevertheless difficulty in diagnosis of dysplasia in colonic polyps or in ulcerative colitis is not confined to general pathologists. [25, 31] In another study by Rex et al., the accuracy of pathologic reading of colorectal polyps by "general pathologists" in community practice was assessed, and the findings showed that there is a great inconsistency among general pathologists in diagnosing high grade dysplasia in polyps. They also noted that that problem is not limited to "non-expert" pathologists, but exists even among the experts. "We found high error rates in identification of high grade dysplasia . . . However, previous studies have indicated a high degree of interobserver variation for interpretation of dysplasia and tubular versus villous histologic features even among experts." [38]

To solve these disagreements in diagnosing dysplasia, the authors suggest that colonoscopists and pathologists should work closely with and educate each other to increase the quality of interpretation. This collaboration should be done as part of a continuous quality improvement program that could be monitored at the level of individual hospitals and practices that manage the patient with adenomatous polyp of the colon.

However, for us, all the confusion and problems result from use of the term "dysplasia" per se. This is not the fault of colonoscopists or pathologists—both expert and non-expert—who are unable to communicate and understand each others' scientific terminology. A language composed of meaningless and never-defined terms, such as dysplasia with all its grades, precancer, precursor, premalignancy, adenoma with atypicality, malignant adenoma, adenoma malignum, invasive adenoma, malignant polyp, adenoma with adenocarcinoma grade I-benign, colorectal intraepithelial neoplasia, and so on, is useless.

Although, as Morson for the first time pointed out, the cytopathologic changes of so-called adenoma (or "dysplasia") in ulcerative colitis are identical to those seen in sporadic adenomatous polyp (see the selected quotations at the beginning of this article), some authors attempted to separate adenoma in ulcerative colitis from sporadic adenoma. In 1981, Blackstone et al., categorized adenoma in ulcerative colitis as either flat or associated with a raised lesion or mass that they coined "dysplasia-associated lesion or mass." [39] Since then, some studies were undertaken to forge histologic criteria for diagnosing DALM and distinguishing it from sporadic adenoma. [40,41] That distinction was very important from a clinical point of view because, since it was believed that the natural history and biology of DALM and flat dysplasia were different from sporadic adenoma, subsequent management would be different.

In our opinion, categorizing dysplasia to "flat" and "elevated" lesions and separating these lesions from sporadic adenomatous polyps only complicated the problem. For us, all these lesions are in the same category. In actuality, all are carcinoma in situ that arises in different settings. So-called flat adenoma ("dysplasia") sporadically occurred in patients without ulcerative colitis, and cytopathologic features of so-called flat adenoma are identical to adenomatous polyp on histologic grounds. During the past two decades, there has been a growing interest in molecular genetics or antigenic expressions of these lesions utilizing PCR and immunohistochemistry. [42–44] All these works, however, fail to show a distinctive genetic profile for each of these lesions. For example, in a recent article by Odze et al., the loss of heterozygosity of APC and p16 genes in so-called polypoid dysplasia associated with ulcerative colitis (or DALM) was evaluated, and they compared the results with those in "flat dysplasia" in ulcerative colitis and those in "sporadic adenomas" (or adenomatous polyp), not related to ulcerative colitis. They concluded that genetic alterations in adenoma-like DALMs associated with ulcerative colitis are similar to non-colitic sporadic adenomas. [45]

In addition to the difficulty in definition and histopathologic diagnosis of dysplasia among pathologists, there is even debate among gastroenterologists and organiziations such as the American College of Gastroenterology (ACG) regarding the treatment of adenoma or dysplasia in ulcerative colitis and follow-up of the patients. Opinions differ among gastroenterologists as to how to treat and manage the patient with a "dysplasia report" mostly because of inconsistencies on the definition of dysplasia. Studies on management and follow-up of the patients with surveillance programs used by gastroenterologists to treat the patients showed widespread variation in practice, as well as confusion regarding the definition and the implications of dysplasia. The first in 1995, by Bernstein et al. reported a survey of 89 U.S. gastroenterologists and GI fellows regarding treatment of dysplasia in ulcerative colitis; only 19% of respondents correctly identified the definition of dysplasia, and the mean number of biopsy specimens taken during a surveillance examination ranged from 7.8 to 9.1. [46] In another study from the United Kingdom done by Eaden et al. 341 gastroenterologists participated in their survey and the authors again found great inconsistencies in the management of the patients with a pathologic report of dysplasia in ulcerative colitis. [47]

More recently, in 2007, Rodriguez and her coworker conducted another survey by mailing a questionnaire to 1,000 gastroenterologists in the U.S and evaluated current management of those patients with ulcerative colitis to which the pathologist added a diagnosis of "dysplasia" in first or subsequent biopsies during the surveillance program. Although the authors claimed that in contrast to previous studies most gastroenterologists in the United States follow guidelines suggested by American College of Gastroenterology (ACG), different approaches to dysplasia are still present. "A major limitation of a surveillance program is failure to take appropriate action when dysplasia is found. The majority of the physicians in this study who recommended continued surveillance rather than colectomy indicated that they would follow patients with LGD [low grade dysplasia] more closely. However, this is still of concern, because many respondents were not taking large numbers of biopsy specimens, and, even with an "adequate" biopsy protocol, dysplasia may be missed on subsequent examinations." [37]

Although the above study is the largest and best survey on surveillance in the United States, the response rate from all gastroenterologists was very low at only 33%. The authors acknowledged this low rate of response and pointed out that this may have produced a bias in their study, if those who did not respond to the questionnaire did not know or follow those recommendations made by ACG for treatment of dysplasia in ulcerative colitis. "A low response rate may have skewed our results toward those physicians who know the recommendations of practice guidelines, because those unfamiliar with them might tend to not respond. This could make surveillance practices in the Unite States appear to be more in accordance with guidelines than they truly are in reality. Another important study limitation is the potential for recall bias. Physicians may be well aware of surveillance recommendation and thus report the recommended strategy when answering questionnaires, yet their actual practice might differ significantly." [37]