Letter to the Editor: “Observer variability for Lung-RADS categorisation of lung cancer screening CTs: impact on patient management”
by Daniel Herrera, John Campaña, and Nelson Bedoya (danielherrera50@hotmail.com)
Observer variability for Lung-RADS categorisation of lung cancer screening CTs: impact on patient managementDear Editor,
We had read carefully the paper from doctors Sarah J. van Riel, Colin Jacobs, Ernst Th. Scholten, Rianne Wittenberg, Mathilde M. Winkler Wille, Bartjan de Hoop, Ralf Sprengers, Onno M. Mets, Bram Geurts and Mathias Prokop, Observer variability for Lung-RADS categorisation of lung cancer screening CTs: impact on patient management [1]. We want to thank for the big effort to make this research which was aimed only to the benefit of the patient. Highlighting the big impact of the radiologist’s role in the lung cancer diagnosis based on images, suggests us how these positive and reliable results can be used as a guidance, not only in lung screening, but also in the identification of other types of cancers.
For this reason, we want to give our opinion to increase the accuracy and remove any doubt that the results obtained in this great study are infallible and correct. In the article, the substantial result of the interobserver agreement is mentioned by means of the use of the Kappa coefficient. This is quite significant and allows to determine the adequate result of the use of LUNG-RADS and the determination of lung cancer. However, it would be interesting to see the behavior of other statistical methods, such as the Kendall coefficient or the interclass interval, since the results could vary in comparison with those obtained through the Kappa test, taking into account that their results should not be considered as a sign of universal good agreement, because these depend both on the prevalence in each class, and of the number of categories [2-4].
It would also be interesting to compare the results of the correlations of this study with a Gold standard, in order to obtain new valuable information. Another point important to consider is that, in phase III studies, at least 10 observers are recommended, and preferably from quite a few institutions, since this will give a better opportunity to evaluate the frequency, nature and magnitude of the differences between them [5].
Our intention with these opinions is to minimize the doubts and objections that may arise in the variability of the use of the LUNG-RADS, and to have the opportunity to know the results when applying the different alternatives mentioned, in order to refine this important study.