Signal pattern enumeration of Urovysion Fluorescence in Situ Hybridization test is tedious and requires great experience. Our aim was to eliminate human interaction by automating the process, using an adoptable, automated image acquisition, and analysis system. Methods: For extensive analytical analysis control, cell populations were used, while preliminary clinical study was performed on 21 patients with clinical suspicion for bladder cancer. All investigations were carried out using an automated user-trainable workstation (Metafer4-Metacyte). Results: The system identified nuclei with a specificity and sensitivity of 92.7 and 96.6%, respectively, while signal detection accuracy was 81.1% on average. Both analytical and diagnostic accuracy of automated analysis was comparable to manual approach (94.8 and 71% vs. 97.9 and 76%, respectively), but classification accuracy increased with degree of polysomy, thus diagnostic sensitivity in low grade, low stage cases was poor. Conclusion: It is possible to automate signal enumeration of Urovysion using a user-trainable system, and achieve efficiency comparable to manual analysis. Previously introduced automated immunophenotypic targeting should further increase diagnostic sensitivity, while resulting in a comprehensively automated method. However, the problem of reduced detection accuracy in cases featured with low polysomy is likely to remain a great challenge of automated signal enumeration.
- Adoptable workstation
- Automated-spot counting
ASJC Scopus subject areas
- Medical Laboratory Technology