Impact of an increase in grading categories and double reporting on the reliability of breast cancer grade

Abstract

The present study illustrates the effect of using the consolidated total score (obtained on a 7-point scale by adding the tubular, nuclear and mitotic scores) as well as double reporting on the interobserver reliability of breast cancer grading by the Nottingham method. 50 consecutive breast cancer slides were graded independently by 5 pathologists. The interobserver reliability between the individual pathologists for final grade, total score and grade components was estimated by quadratic weighted kappa (k(w)). Similarly, the inter-observer reliability of double reported scores (obtained by averaging the reported scores of two independent pathologists) for independent observer pairs was estimated. The total scores gave higher reliability figures than the final grade both for individual pathologists and double reported scores. The double reported total scores and grade also gave significantly higher reliability figures than the individual scores and grade (average k(w)=0.65 and 0.78 for the individual and double reported total scores, respectively; 0.61 and 0.66 for the individual and double reported grades).Therefore, use of total scores in addition to grade as well as double reporting of the same significantly increases the interobserver reliability of breast cancer grading.