OBJECTIVE: To determine interobserver and intraobserver reproducibility in the assessment of the HercepTest- and TAB250-immunostained slides. STUDY DESIGN: Three independent expert pathologists (two with and one without training in HercepTest assessment) evaluated the HercepTest and TAB250-immunostained slides of 108 infiltrating breast carcinomas with a triple-blind method. The evaluation was repeated, with the same method and sequence of view, after 60 days. RESULTS: Expert pathologists, after adequate training in HercepTest evaluation, could reach excellent interobserver (K=.911, P<.001) and intraobserver reproducibility (K of .863-.926; P <.001 for all). The percentage of disagreement in intraobserver reproducibility ranged from 0.9% to 3.7%. Interobserver and intraobserver reproducibility in the evaluation of TAB250-immunostained slides was good (K = .658, P < .001) and from good to excellent (K of .600-.895, P < .001 for all), respectively. CONCLUSION: Optimization of the level of accuracy in HercepTest evaluation is mandatory because the decision to initiate therapy with Herceptin depends on the result. Moreover, considering that the percentage of disagreement in intraobserver reproducibility ranges from 0.9% to 3.7%, it is advisable that two expert pathologists evaluate all HercepTest slides with a double-blind method. If there are discordant results, they must be discussed by the same pathologists.
OBJECTIVE: To determine interobserver and intraobserver reproducibility in the assessment of the HercepTest- and TAB250-immunostained slides. STUDY DESIGN: Three independent expert pathologists (two with and one without training in HercepTest assessment) evaluated the HercepTest and TAB250-immunostained slides of 108 infiltrating breast carcinomas with a triple-blind method. The evaluation was repeated, with the same method and sequence of view, after 60 days. RESULTS: Expert pathologists, after adequate training in HercepTest evaluation, could reach excellent interobserver (K=.911, P<.001) and intraobserver reproducibility (K of .863-.926; P <.001 for all). The percentage of disagreement in intraobserver reproducibility ranged from 0.9% to 3.7%. Interobserver and intraobserver reproducibility in the evaluation of TAB250-immunostained slides was good (K = .658, P < .001) and from good to excellent (K of .600-.895, P < .001 for all), respectively. CONCLUSION: Optimization of the level of accuracy in HercepTest evaluation is mandatory because the decision to initiate therapy with Herceptin depends on the result. Moreover, considering that the percentage of disagreement in intraobserver reproducibility ranges from 0.9% to 3.7%, it is advisable that two expert pathologists evaluate all HercepTest slides with a double-blind method. If there are discordant results, they must be discussed by the same pathologists.
Authors: Seon-Ah Ha; Youn Soo Lee; Seung Min Shin; Hyun Kee Kim; Sanghee Kim; Hong Namkoong; Hae Joo Kim; Sang Min Jung; Yu Sun Lee; Yeun Jun Chung; Sang Seol Jung; Jin Woo Kim Journal: BMC Cancer Date: 2009-02-11 Impact factor: 4.430