Endoscopy, Morphology, Morphometry and Molecular Markers: Predicting Cancer Risk in Colorectal Adenoma

Kjetil Søreide; Bjørn S Nedrebø; Andreas Reite; Kenneth Thorsen; Hartwig Kørner

Disclosures

Expert Rev Mol Diagn. 2009;9(2):125-137. 

In This Article

Morphometry & Quantitative Pathology Markers

Generally, morphometrics (from the Greek "morph," meaning shape or form, and "metron", meaning measurement) comprises methods of extracting measurements from shapes. It has been attempted to use morphometry, as a part of quantitative pathology, to classify tissue lesions in a metric (and preferably more objective; Figure 4) mode compared with the subjective assessment performed by pathologists.

Figure 4.

Morphology versus morphometry for adenoma assessment. Morphometric measurements performed by quantitative, digitalized assessment of histopathology slides compared with regular pathology assessment through the microscope.

The histopathologic grading of colorectal adenomatous dysplasia is subjective, and requires judgement by the pathologist.[40] In order to arrive at a more objectively defined risk assessment of a given lesion (Figure 4), several groups have performed image analysis to quantitatively assess colorectal glandular or epithelial characteristics.[59,60,61,62,63,64,65] Usually, cases of histologically normal mucosa and of low-, moderate- and high-grade dysplasia are examined using this technique, which has been improved by automated features and in more recent years use of knowledge-guided software. A large number of morphometric and densitometric features can be measured on each gland, such as gland shape, epithelial area, nuclear stratification, optical density and the length of axis. By discriminant analysis of the data, one or more morphometric features that provide the best discrimination between the various histologic groups may be revealed.

In the search for new methods of characterizing risk in adenomas, our research group has recently identified (by quantitative digitalized image analysis) a monotonous population of elongated cells, which discriminate high- from low-risk adenomas better than conventional histology type and grade.[66] However, with the inclusion of immunohistochemistry markers, in particular the antiapoptosis regulator survivin, the human telomerase reverse transcriptase (hTERT), as well as regulators of the cell cycle, the morphometric markers were still predictive,[67] but lost the predictive ability compared with other factors in a validation study.[68] Thus, while this may describe a recognized, epithelial cell type (round, elongated, sigar-shaped) within adenomas, the predictive ability for cancer is overshadowed by other factors.

Few studies have focused on the relationship between abnormal nuclear morphology and the molecular genetic alterations that are important in CRC development. One study demonstrated that nuclear morphology and biologic behaviour appeared to be influenced by accumulated alterations in cancer-associated genes.[63] Feulgen-stained isolated nuclei from 22 adenomas and 42 carcinomas that had been analyzed for ras gene mutations and allelic deletions on chromosomes 5q, 18q and 17p were characterized by computerized image analysis. Both nuclear area and the nuclear shape factor representing irregularity correlated with adenoma-carcinoma progression (r = 0.57 and r = 0.52 respectively; p < 0.0001), whereas standard nuclear texture, a parameter of chromatin homogeneity, was inversely correlated with progression (r = -0.80; p < 0.0001). The nuclear parameters were strongly interrelated (p < 0.0005). In multivariate analysis, the nuclear parameters were predominantly associated with adenoma-carcinoma progression (p < /= 0.0001) and were not influenced significantly by the individual molecular genetic alterations. Nuclear texture, however, was inversely correlated with fractional allelic loss, a global measure of genetic changes, in carcinomas (r = -0.39; p = 0.011). The findings indicate that nuclear morphology in colorectal neoplasia is strongly related to tumor progression. However, while these methods have been around for decades, the studies are heterogeneous in set-up, apply different end points (dysplasia evaluation, synchronous or metachronous cancer progression), and usually arrive at different conclusions and with different discriminators as important. Furthermore, it involves labor-intensive and, in part, costly methods that are not universally available in all pathology laboratories.

While morphometric measurements have been applied with a wide range of methods in colorectal neoplasia (from adenoma to carcinoma),[63,64,66,69,70,71,72] there is considerable heterogeneity in the studies performed, and a general applicability and consensus for clinical practice has thus far not been reached. Furthermore, morphometry methods are labour-intensive, time-consuming, and instruments for automated, digitalized measurements are not widespread or available in regular histopathology labs, thus making this technology unlikely to receive large input with current methodology and obtained results. Furthermore, one has trouble seeing how developments in this field could further improve the understanding of colorectal carcinogenesis. With current molecular technology (DNA microarrays, tissue arrays, PCR, comparative genomic hybridization, FISH etc.) the results from studies with clinical end points not only yields potential new biomarkers, they also contribute in the further understanding of underlying contributors to the molecular genesis of the disease. This is a major drawback with morphometry research as it can only provide a 'pattern-picture', yet very little direct or indirect information is derived on carcinogenesis per se, at least from the adenoma studies.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....