Observer variability in reporting of breast lesions.

The consistency of histological diagnosis of breast lesions by members of the panel was studied using a simplified six point classification system that covered the range from normal tissue to invasive carcinoma. In a representative set of 40 sections of the range of breast disease included in the Trial for Early Detection of Breast Carcinoma (the consecutive series), there was, overall, 82% agreement between the diagnoses submitted by members of the panel: diagnosis was virtually consistent (99% agreement) for frankly invasive carcinoma, but there was much greater variability in the diagnosis of borderline lesions. Diagnostic consistency was greatly improved when two categories only (benign and malignant) were considered (94% agreement). The diagnostic consistency of individual pathologists was studied by recirculating the sections: the overall agreement was 78% and, again, most of the inconsistencies occurred when borderline lesions were diagnosed. Additional studies were undertaken on borderline lesions that had been flagged during the first two years of the trial (the borderline series). Unsurprisingly, there was less agreement among the pathologists when all six diagnostic categories were used (70% and 77% for the specimens from the first and second years, respectively), but consistency was greatly improved when classification was simplified to either benign or malignant (86% and 91%, respectively). The diagnoses submitted by individual pathologists were found to deviate from the "majority" diagnosis to a relatively minor extent: each pathologist was generally consistent in either under-diagnosing or overdiagnosing in both the consecutive and borderline series.

[1]  J. R. Landis,et al.  The measurement of observer agreement for categorical data. , 1977, Biometrics.

[2]  E. B. Butler,et al.  Combined external quality assessment of cytology and histology opinions: a pilot scheme for a cluster of five laboratories. , 1984, Journal of clinical pathology.

[3]  N. Day,et al.  EVALUATION OF SCREENING FOR BREAST CANCER IN A NON-RANDOMISED STUDY (THE DOM PROJECT) BY MEANS OF A CASE-CONTROL STUDY , 1984, The Lancet.

[4]  S. Shapiro,et al.  Ten- to fourteen-year effect of screening on breast cancer mortality. , 1982, Journal of the National Cancer Institute.

[5]  D. Ransohoff,et al.  Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. , 1983, Human pathology.

[6]  Jacob Cohen A Coefficient of Agreement for Nominal Scales , 1960 .

[7]  J. Richard Landis,et al.  Large sample variance of kappa in the case of different sets of raters. , 1979 .

[8]  A. Patchefsky,et al.  Increased detection of occult breast carcinoma after more thorough histologic examination of breast biopsies. , 1973, American journal of clinical pathology.

[9]  M F Dixon,et al.  Observer variation in the histological grading of rectal carcinoma. , 1983, Journal of clinical pathology.

[10]  B Stenkvist,et al.  Histopathological systems of breast cancer classification: reproducibility and clinical significance. , 1983, Journal of clinical pathology.

[11]  L. Tabár,et al.  REDUCTION IN MORTALITY FROM BREAST CANCER AFTER MASS SCREENING WITH MAMMOGRAPHY Randomised Trial from the Breast Cancer Screening Working Group of the Swedish National Board of Health and Welfare , 1985, The Lancet.

[12]  W. Grove Statistical Methods for Rates and Proportions, 2nd ed , 1981 .