INFLUENCE OF OUTLIERS ON THE ASSOCIATION BETWEEN LABORATORY DATA AND HISTOPATHOLOGICAL FINDINGS IN LIVER-BIOPSY

Discriminant analysis techniques were used to predict the histopatholo gical findings in liver biopsy specimens in asymptomatic patients with slightly to moderately raised routine liver tests. Moderate to severe fibrosis and/or inflammation were treated as indication for biopsy. T wo methods were used to classify patients. One was the dichotomous dis crimination between ''biopsy necessary'' or ''biopsy not necessary-gro ups of patients. The other involved combining two discriminant functio ns trained separately for recognition of fibrosis or inflammation, and then combined to predict the biopsy necessity. Detection of outliers by standard techniques, directly available in the SPSS-X package, was performed before starting discrimination procedures. Both 'sharp'' ass ignment rules and continuous scoring rules were applied to the classif ication problem. The correct classification rate reached over 85% for the algorithms tested. In the majority of cases the classification was found to be ''non-doubtful''. Elimination of outliers (especially by standardized residuals) improved the global correct classification rat e, but only slightly improved assignment to the ''biopsy necessary'' g roup. Routine and complementary laboratory findings were found to be t he most discriminating; answers to questionnaire and ultrasound examin ation were less important. Selection of the most diagnostic features b ased on ''clean'' data without outliers enabled us to find interesting medical associations, which were previously masked by extremely asymp tomatic values outlying from the main body of the ''biopsy necessary'' group.