Is it possible to make the unequivocal diagnosis of "dysplasia", especially "low-grade dysplasia" in Barrett's mucosa?
K. Geboes (Leuven)
Multistep genetic changes are associated with the occurrence and development of human malignant tumors . This multistep process is associated with and most probably depends particularly upon defects in genes controlling cell proliferation and cell death. During carcinogenesis a normal cell progresses subsequently to premalignant, in situ cancer (high grade dysplasia), invasive cancer, metastasis and eventually recurrence. Genomic instability leads to the generation of single and later multiple aneuploid populations of cells. One of these, by further genetic changes eventually acquires the capacity for invasion. Molecular cytogenetic techniques help tracing the clonal changes, evolution and expansion. The genetic changes not only involve genes and their products controlling cell proliferation and cell death. Other cell processes can also be altered. Using in situ hybridization an aberrant expression of cytokeratin mRNAs has been found in squamous cell carcinoma of the esophagus. A decrease and/or loss of expression of genes coding for the esophageal differentiation-related cytokeratins (cytokeratin 4 mRNA) is noted while an expression of genes, not present in the normal esophagus appears (cytokeratin 8, 18, 19) . These findings confirm earlier data obtained using immunohistochemistry . Cytokeratins are molecules that belong to the complex intracellular network of filaments, building an intracellular structure known as cytoskeleton. Cytokeratins, which belong to the intermediate filament subgroup of cytoskeletal filaments are important for the cell shape and epithelial cell cohesion. As they undergo also changes during tumorigenesis it is clear that cell shape and structure and cellular cohesion will also be altered. These alterations are part of the morphological changes known as dysplasia.
Dysplasia is defined as an "unequivocal, non invasive (confined within the basement membrane), neoplastic transformation of the epithelium, that can be recognized by microscopy (because of cytological and architectural changes). Dysplasia is thus the morphological expression of one of the steps in carcinogenesis. Dysplasia can be subdivided into low- and high-grade (or variants of this classification) depending on the severity of architectural and cytological alterations. Low-grade dysplasia, the less severe form, is thus the earliest morphologic expression of neoplasia detectable by routine histology which occurs as one of the early steps in carcinogenesis.
Cell proliferation in Barrett's esophagus
In the gastrointestinal tract normal epithelial cell proliferation is confined to a particular area of the mucosa, usually located in the basal part of the crypts. This is also the area, where, in case of epithelial cell damage and loss, reactive changes will occur during the healing process. In addition this is also the area where abnormal proliferation will begin during carcinogenesis, although in neoplasia the proliferative compartment usually expands towards the surface. In Barrett's esophagus (BE) the epithelial cell proliferation is also limited to the basal part of the crypts. The length of the proliferative compartment and the labelling index indicating the cells in cycle are however variable . This is to be expected as Barrett's mucosa is a metaplastic condition. Subsequently, the appearance of nuclear atypism in this area is not uncommon and has been noted on several occasions [5, 6].
The question as to whether "low-grade dysplasia" or "dysplasia" can be detected unequivocally or reliably, based on changes occurring deep in the epithelium refers therefore precisely to the area where normal cell proliferation occurs and where, in Barrett's mucosa nuclear abnormalities may be more pronounced and a tendency to stratification exists. These alterations in Barrett's can be explained as reactive changes or belong to the metaplastic nature of the condition. The question will therefore be answered in different steps.
Low-grade dysplasia, deep in the epithelium
Can (low-grade) dysplasia exist in the epithelium, deep in the mucosa?
The answer to this question must be "yes". Arguments in favor of a positive answer come from both theoretical considerations and practical findings. The epithelium deep in the mucosa is the area where normal proliferation occurs and where abnormal proliferation (the beginning of neoplasia) will start, given the very nature of the cells in this compartment and the nature of carcinogenesis. In practice it is clear that dysplasia can be found in the deeper part of the epithelium. Low-grade dysplasia of the esophageal squamous epithelium is exactly localized in this area . Low-grade dysplasia limited to the deeper part of the epithelium can also be observed in columnar mucosa in the stomach and duodenum and even in BE dysplastic lesions, limited to the pits can be found. In our experience such lesions were however, when found in BE, usually associated with more important lesions involving the whole epithelium or with lesions of high-grade dysplasia or frank carcinoma. The question therefore is not whether such a lesion exists but rather whether it is possible to diagnose such a lesion reliably. Two possible conditions can be considered : basal changes associated with a denuded surface and basal changes covered with a normal surface epithelium.
Is it possible to make a reliable diagnosis of (low-grade) dysplasia limited
to the deeper epithelium when the surface is denuded?
When the surface is denuded subsequently to epithelial cell damage and hence there is an erosion or ulceration, it is impossible to make a reliable diagnosis of low-grade dysplasia because the changes induced by inflammation and as part of the healing process, make a reliable diagnosis impossible. Reactive changes will include an increase of nuclear size, the appearance of a larger nucleolus, tendency to stratification and increased mitoses because there is a need for cell proliferation. Hence more immature (less well differentiated) cells will be present in the proliferative compartment which is the deeper part of the epithelium. Therefore a reliable diagnosis will not be possible.
Is it possible to make a reliable diagnosis of (low-grade) dysplasia limited
to the deeper epithelium when the surface is normal?
Interobserver studies have shown that the agreement for the diagnosis of low-grade dysplasia in BE is less good than for high-grade dysplasia even when the surface epithelium is involved . Therefore the reliability of this diagnosis in cases of a normal surface epithelium is most probably low. The real reason in fact why it is difficult to diagnose reliably (low-grade) dysplasia deep in the epithelium is the lack of sensitivity of the morphological criteria available, especially when compared with benign reactive changes. In both conditions nuclei can be hyperchromatic, the N/C ratio is increased, pseudostratification and mild cellular and nuclear pleomorphism can occur in the deep epithelium [5, 6]. An if it is impossible to make a reliable diagnosis it is better to be cautious because an authentic diagnosis of dysplasia implies a considerable psychological burden for the patient and considerable costs for follow-up.
Can additional techniques help to resolve the diagnostic problem?
Several techniques have been developed allowing a more in depth study of the alterations that occur at the cellular and tissular level during carcinogenesis. The nuclear presence of abnormal products such as mutant p53 can be identified using immunohistochemistry and appropriate antibodies. Flow cytometry can identify aneuploid cell populations and comparative genomic hybridization (with fluorescent in situ hybridization) can identify chromosomal gains and losses. Application of these techniques can certainly help to support a diagnosis of dysplasia. Some of these procedures are however not available everywhere and are still rather expensive. Others such an immunohistochemistry are not consistently positive . Furthermore they try in fact to identify other phenomena which do not necessarily appear during the process of carcinogenesis at the same moment as dysplasia.
Although (low-grade) dysplasia can exist in the deeper part of the epithelium, even when the surface is mature, an unequivocal diagnosis is impossible because the morphologic criteria available to do so, lack sensitivity. This lack is due to the fact that the alterations occurring during reactive changes are highly comparable and occur in the same area, making a reliable distinction impossible. The use of additional techniques may however, in practice, help to solve some diagnostic problems.
3. Geboes K, Haustermans K, Lerut T, Van der Schueren E. Cytokeratin 14, 18 and 19 expression in normal epithelium and in squamous cancer of the esophagus. In Nabeya K, Hanaoki T, Nogami H, eds. Recent advances in diseases of the esophagus. Berlin: Springer Verlag, 1993:483-487.
9. Muzeau F, Fléjou JF, Potet F, Belghiti J, Thomas G, Hamelin R. Profil des mutations du gène p53 et expression anormale de la protéine p53 dans les deux formes de cancer de l'sophage. Gastroenterol Clin Biol 1996;20:430-437.