Grading of oral squamous cell carcinomas – Intra and interrater agreeability: Simpler is better?
Steigen, Sonja E; Søland, Tine Merete; Nginamau, Elisabeth Sivy; Laurvik, Helene; Costea, Daniela-Elena; Johannessen, Anne Christine; Jebsen, Peter Wilhelm; Bjerkli, Inger-Heidi; Uhlin-Hansen, Lars; Hadler-Olsen, Elin Synnøve
Peer reviewed, Journal article
Published version
Åpne
Permanent lenke
https://hdl.handle.net/1956/23235Utgivelsesdato
2020Metadata
Vis full innførselSamlinger
Originalversjon
https://doi.org/10.1111/jop.12990Sammendrag
Background: Numerous studies have been presented on histological grading of oral squamous cell carcinomas (OSCC) for predicting survival, but uncertainty of their usefulness rises due to discordances of results. A scoring system should be robust and well validated, and intra‐ and interrater agreement can be used as a tool to visualize the strength of reproducibility. Methods: Here, we present an intra‐ and inter‐observer study on evaluation of OSCC using some of the most common histopathological parameters. The observers were from different Norwegian university hospitals, and calibration to ensure accuracy was first performed. Percentage of the agreement was calculated for the score made by the individual observer at different times, as well as between pairs of observers. Results: The evaluation made by the same observer at two different time points (intrarater) correlated better than observations made by different participants (interrater). In an attempt to increase the rate of agreement, many of the parameters were either dichotomized into simply low‐ and high grade, or to a three‐tier system when more than three options in the original design. This increased the concurrence with 15.4% for the intrarater and with 23% for the interrater comparisons. Conclusion: High agreement for histopathological parameters can be difficult to obtain on hematoxylin and eosin staining in scoring systems with many options. A simpler system might be more advantageous to achieve higher degree of reproducibility.