Interobserver Variability Between Expert Urologic Pathologists for Extraprostatic Extension and Surgical Margin Status in Radical Prostatectomy Specimens Andrew J. Evans, MD, PhD,* Pauline C. Henry, MD, PhD,* Theodorus H. Van der Kwast, MD,* Douglas C. Tkachuk, MD,*w Kemp Watson,w Gina A. Lockwood, MSc,z Neil E. Fleshner, MD,y Carol Cheung, MD,* Eric C. Belanger, MD,J Mahul B. Amin, MD,z Liliane Boccon-Gibod, MD,# David G. Bostwick, MD, MBA,** Lars Egevad, MD, ww Jonathan I. Epstein, MD,zz David J. Grignon, MD,yy Edward C. Jones, MD,JJ Rodolfo Montironi, MD,zz Madeleine Moussa, MD,## Joan M. Sweet, MD,* Kiril Trpkov, MD,*** Thomas M. Wheeler, MD,www and John R. Srigley, MDzzz Abstract: Accurate Gleason score, pathologic stage, and surgical margin (SM) information is critical for the planning of post- radical prostatectomy management in patients with prostate cancer. Although interobserver variability for Gleason score among urologic pathologists has been well documented, such data for pathologic stage and SM assessment are limited. We report the first study to address interobserver variability in a group of expert pathologists concerning extraprostatic soft tissue (EPE) and SM interpretation for radical prostatectomy specimens. A panel of 3 urologic pathologists selected 6 groups of 10 slides designated as being positive, negative, or equivocal for either EPE or SM based on unanimous agreement. Twelve expert urologic pathologists, who were blinded to the panel diagnoses, reviewed 40  whole-slide scans and provided diagnoses for EPE and SM on each slide. On the basis of panel diagnoses, as the gold standard, specificity, sensitivity, and accuracy values were high for both EPE (87.5%, 95.0%, and 91.2%) and SM (97.5%, 83.3%, and 90.4%). Overall k values for all 60 slides were 0.74 for SM and 0.63 for EPE. The k values were higher for slides with definitive gold standard EPE (k = 0.81) and SM (k = 0.73) diagnoses when compared with the EPE (k = 0.29) and SM (k = 0.62) equivocal slides. This difference was markedly pronounced for EPE. Urologic pathologists show good to excellent agreement when evaluating EPE and SM. Interobserver variability for EPE and SM interpretation was principally related to the lack of a clearly definable prostatic capsule and crush/thermal artifact along the edge of the gland, respectively. Key Words: prostate cancer, radical prostatectomy, surgical margins, extraprostatic extension, interobserver variability (Am J Surg Pathol 2008;32:1503–1512) R adical prostatectomy (RP) is a curative treatment option for conventional acinar-type prostate cancer that seems to be organ-confined at diagnosis. Accurate histopathologic examination of RP specimens is required to assess the risk of recurrence and to make decisions on additional therapy. The 3 histopathologic parameters of greatest prognostic importance are Gleason score, patho- logic stage, and surgical margin (SM) status, 11,16–18,23 where pathologic stage includes assessment for seminal vesicle invasion (SVI) and extension of tumor into extraprostatic soft tissue (EPE). Elements that contribute to the presence of EPE and/or positive SM include those inherent to the tumor (Gleason score, volume, and distribution), 11,16–18 surgical factors (surgeon experience and type of procedure), 4,7,20 and pathology-related factors (specimen inking, removal of sutures and surgical clips, and partial vs. complete submission of the speci- men). 13,21 In addition, the interpretation of the RP slides by the pathologist represents another critical variable in this process. Copyright r 2008 by Lippincott Williams & Wilkins From the Departments of *Pathology and Laboratory Medicine; zBiostatistics; ySurgical Oncology, University Health Network; wObjective Pathology Services, Toronto; JDepartment of Pathology and Laboratory Medicine, University of Ottawa, Ottawa; ##Depart- ment of Pathology, London Laboratory Services Group, London; zzzThe Credit Valley Hospital, Mississauga, Ontario; JJDepartment of Pathology and Laboratory Medicine, Vancouver General Hospital, Vancouver, British Columbia; ***Department of Pathol- ogy and Laboratory Medicine, Calgary Laboratory Services, Calgary, Alberta, Canada; zDepartment of Pathology, Cedars-Sinai Medical Center, Los Angeles, CA; **Bostwick Laboratories, Glen Allen, VA; zzDepartment of Pathology, The Johns Hopkins Medical Institutions, Baltimore, MD; yyDepartment of Pathology and Labora- tory Medicine, Indiana University, Indianapolis, IN; wwwDepart- ment of Pathology, Baylor College of Medicine, Houston, TX; #Pathology Department, Hospital Armand Trousseau, Paris; wwInternational Agency for Research on Cancer, Lyon, France; and zzSection of Pathological Anatomy, Polytechnic University of the Marche Region, Ancona, Italy. Correspondence: Dr Andrew J. Evans, MD, PhD, Toronto General Hospital, University Health Network, 200 Elizabeth Street, Toronto, Ontario M5G 2C6, Canada (e-mail: andrew.evans@uhn.on.ca). ORIGINAL ARTICLE Am J Surg Pathol  Volume 32, Number 10, October 2008 1503