Pathology in disgrace?

In a recent Editorial in the journal, Bosman1 addresses the uncertainty of histopathological grading of dysplasia as a tool for treatment decision and raises the question of alternative approaches. Drawing the attention to molecular markers and quantitative pathology, he argues forcefully that straightforward molecular markers are unlikely to give the unambiguous information that the clinicians are seeking. He then refers to two recently published studies2, 3 and calls into doubt the diagnostic usefulness of quantitative diagnostic pathology. But each of these positions is open to challenge. For example, the inherent drawback of the use of single gene probes4 can be overcome in principle by the use of composite molecular markers. Thus, gross estimation of nuclear DNA content (DNA ploidy) has been shown to have considerable prognostic impact on the clinical outcome in premalignant lesions of the head and neck5. It is significant that the point-value genetics underlying chromosomal imbalances are now beginning to be revealed6-9 and the resulting DNA aneuploidy – reflecting genetic instability – may itself induce further genetic changes. Therefore, gross aberrations in the DNA content of cells should be viewed as more than proxy variables that manifest themselves secondary to the carcinogenic process. Analysis of the DNA content of cells in, for example, cancerous or precancerous lesions, is not more laborious, time-consuming or costly than the many molecular and immunohistochemical analyses commonly performed in pathology laboratories, nor does it require very sophisticated equipment. Well-established protocols for such analysis exist and have been published in consensus reports10, 11. A direct comparison of traditional histological assessment of premalignancies and the gross estimation of DNA content as a prognostic marker in oral premalignancies strongly indicated the lack of prognostic information in traditional histological assessment12, 13. By contrast, measuring the DNA content in the same lesions predicted with a considerable degree of certainty which premalignancies would subsequently develop carcinomas. In a multivariate analysis, DNA aneuploidy was a highly significant prognostic factor (p<0.001), with adjusted odds ratio for developing a subsequent carcinoma of 27.7 (95% confidence interval 18.4–36.8) in the DNA aneuploid premalignant lesions, compared with DNA diploid premalignancies5. With respect to other measures in quantitative pathology, there exist a number of promising methods for assessing tissue architecture in a strictly quantitative and highly reproducible manner14, and which also provide prognostic information of considerable impact15. These results apply not only to early-stage carcinomas of the tongue, but also to other solid tumours, such as carcinomas of the prostate14. Tissue architecture may be considered as a composite structural marker and its assessment is an obvious possibility to achieve objectivity and to improve the lack of reproducibility in histological grading which has been demonstrated for premalignant lesions1, 12, 13, 16. Again, such quantitative analysis of tissue architecture can be performed on routinely stained tissue sections, using off-the-shelf computer equipment14, 15. Diagnostic pathology is by no means left in disgrace. New diagnostic methods with a considerably improved prognostic impact already exist and deserve implementation within the scope of routine diagnostics.