Mieke Bus

45 3 cardiology and ophthalmology. As it involves the use of a medical device instead of a phar- macological product, the optimal way to proceed is following the IDEAL evaluation criteria, described by the Balliol collaboration. (9, 10) This evaluation equals surgical innovation to the different phases of the pharmacological research. CLE has been described for the IDEAL stage 1 (Innovation phase). PDD, OCT and NBI are framed as IDEAL stage 2A or developmen- tal phase. To develop these technologies into a diagnostic test, results should be compared with the standard test at use. In the case of PDD and NBI, the standard test is WL-URS (includ- ing biopsies), where OCT uses histology as gold standard. It is recommended to combine the IDEAL methodology with the QUADAS for assessing accuracy of a diagnostic test. These clinical studies should provide a full diagnostic accuracy analysis according to the QUADAS recommendations, including sensitivity, specificity, positive predictive value and negative predictive value. However, upper urinary tract tumours are rare, resulting in a lim- ited study population and therefore require a multicentre approach when entering IDEAL phase 2b (exploration phase) or 3 (assessment phase). So far, only two studies on optical diagnostics in the upper urinary tract are registered at clinicaltrials.gov, using similar search terms as described in the method section. Both stud- ies comprise optical coherence tomography and only one study is recruiting patients at this moment. Conclusions NBI, SPIES and PDD aim at improved detection of upper urinary tract tumours. OCT and CLE aim at providing real-time, minimally invasive and objective prediction of his- topathological diagnosis. Optical diagnostics might overcome the limitations of the current diagnostic standard of upper urinary tract tumours to determine which tumours can be treated endoscopically. Although these optical techniques show promising results, more, better conducted prospective studies should be done before they can be implemented in the diagnostic work up of upper urinary tract tumours. Acknowledgements The authors like to thank Aristeo Lopez and Joseph Liao, Stanford University for generously providing the CLE figure used in this paper.