Urothelial papillary lesions. Development of a Bayesian Belief Network for diagnosis and grading

The diagnosis and grading of urothelial papillary lesions are affected by uncertainties which arise from the fact that the knowledge of histopathology is expressed in descriptive linguistic terms, words and concepts. A Bayesian Belief Network (BBN) was used to reduce the problem of uncertainty in diagnostic clue assessment, while still considering the dependencies between elements in the reasoning sequence. A shallow network was designed and developed with an open-tree topology, consisting of a root node containing four diagnostic alternatives (papilloma, papillary carcinoma grade 1, papillary carcinoma grade 2 and papillary carcinoma grade 3) and eight first-level descendant nodes for the diagnostic features. Six of these nodes were based on cell features and two on the architecture. The results obtained with prototypes of relative likelihood ratios showed that belief in the diagnostic alternatives is very high and that the network can identify papilloma and papillary carcinoma, including their grade, with certainty. In conclusion, a BBN applied to the diagnosis and grading of urothelial papillary lesions is a descriptive classifier which is readily implemented and allows the use of linguistic, fuzzy variables and the accumulation of evidence presented by diagnostic clues