Reproducibility, repeatability, and level of difficulty of two methods for tumor budding evaluation in oral squamous cell carcinoma

BackgroundThis study aimed to analyze the reproducibility, repeatability, and level of difficulty of two methods for tumor budding evaluation in oral squamous cell carcinoma (OSCC): staining by hematoxylin and eosin (HE) and immunostaining for multicytokeratin. MethodsThe evaluation of tumor budding was performed by three examiners in 103 samples of OSCC, using the two methods. A Likert-type scale was used to measure the difficulty in the assessment. The interexaminer agreement (reproducibility) was estimated using Fleiss's kappa and the intra-examiner agreement (repeatability) was estimated using Cohen's kappa. The agreement between the two methods was evaluated using Cohen's Kappa. The Friedman test was used to compare the three examiners' perceived levels of difficulty of assessment. The Wilcoxon test was used to compare the level of difficulty of the evaluation between the two methods. ResultsReproducibility by the immunostaining method for multicytokeratin was substantial, being higher than the only fair agreement by the HE. Repeatability by the HE ranged from moderate to substantial among examiners, regardless of the examiner's experience. Repeatability by the immunostaining method for multicytokeratin did not vary among examiners, showing almost perfect agreement. The agreement between the two methods ranged from fair to moderate among examiners, being lower in the less experienced examiner. All the examiners presented greater difficulty in the evaluation by the HE. ConclusionIn view of the unsatisfactory agreement between the two methods of tumor budding evaluation in OSCC, it is recommended that this evaluation should be performed by the immunostaining method for multicytokeratin, considering its higher reproducibility, greater replicability, and lower difficulty compared to the HE.