Interobserver variation in the evaluation of bone scans in newly diagnosed prostate cancer

H.D. ZACHO; J.A. BIURRUN MANRESA; J.C. MORTENSEN; H. BERTELSEN; L.J. PETERSEN

Lyon

Congreso; Annual Congress of the European Association of Nuclear Medicine; 2013

Aim: The objective of this study was to assess variability among trained observers in the interpretation of bone scans in an unselected population of patients with newly diagnosed prostate cancer. Materials and Methods: A total of 671 consecutive patients were referred to three sites from March 2008 to October 2009 for bone scintigraphy as a part of the staging workup for newly diagnosed prostate cancer. Standard whole body bone scans were independently reviewed by three experienced nuclear medicine physicians in a separate session. The three observers were blinded to clinical and laboratory data. Each scan was classified according to two grading systems, A) a four category scale (1: normal scan and/or benign changes, 2: equivocal, 3: most likely malignant findings, and 4; multiple bone metastases), and B) a dichotomous scale (bone metastasis present or absent). Results: Uniform classification of bone scans on the 4-point scale was observed in 440/671 (66%) patients. A difference of one category scale was observed in 225/671 (34 %) and only in 6 (0.9%) patients was the scale difference > 1. The linear weighted Cohens kappa ranges between the three observers is 0.71, 0.72 and 0.72, and the Intraclass Correlation Coefficient 0.835, 0.835 and 0.839. The agreement among individual observers and the median category scale was very similar (76% -78%). There was uniform agreement in 45% of patients with category 2 and 3 scans (representing 38% of all patients) with more than one scale of difference in 6/247 (2.4%) of these patients.Uniform agreement between the three reviewers was observed in 641/671 (96%) patients based on evaluation on a dichotomous scale. Conclusion: There is a low variation among trained observers for evaluation of bone scans in patients with newly diagnosed prostate cancer. The variation was very low using a dichotomous grading scale. Variation in classification of patients with equivocal and suspected malignant findings may have significant impact on subsequent imaging workup and therapeutic decisions.