Knowledge-guided histometry of the basal cell layer in prostatic intraepithelial neoplasia.

OBJECTIVE

To define a procedure for the fully automated evaluation of histopathologic sections of prostatic intraepithelial neoplasia.

STUDY DESIGN

The design is based on expert system-guided scene segmentation, the construction of a knowledge file and the defining of histometric measures for assessment of the basal cell layer.

RESULTS

A knowledge file specifying a total of 263 entities was constructed. Of 73 multimegapixel-sized images of cribriform glands, 71 were correctly and completely processed. A histometric measure allowing precise indexing of the degree of lesion progression, based on the deterioration of the basal cell layer, was derived.

CONCLUSION

The combination of image analytic and immunohistochemical procedures and fully automated processing provides a quantitative measure for evaluating prostatic intraepithelial neoplasia lesion progression and the effects of intervention.