Improvement of diagnostic agreement among pathologists in resolving an “atypical glands suspicious for cancer” diagnosis in prostate biopsies using a novel “Disease-Focused Diagnostic Review” quality improvement process

SummaryOne of the major goals of an anatomic pathology laboratory quality program is to minimize unwarranted diagnostic variability and equivocal reporting. This study evaluated the utility of Miraca Life Sciences' “Disease-Focused Diagnostic Review” (DFDR) quality program in improving interobserver diagnostic reproducibility associated with classification of “atypical glands suspicious for adenocarcinoma” (ATYP) in prostate biopsies. Seventy-one selected prostate biopsies with a focus of ATYP were reviewed by 8 pathologists. Participants were blinded to the original diagnosis and were first asked to classify the ATYP as benign, atypical, or limited adenocarcinoma. DFDR comprised a “theoretical consensus” (in which pathologists first reached consensus on the morphological features they considered relevant for the diagnosis of limited prostatic adenocarcinoma), a didactic review including relevant literature, and “practical consensus” (pathologists performed joint microscopic sessions, reconciling each other's observations and positions evaluating a separate unique slide set). Participants were finally asked to reclassify the original 71 ATYP cases based on knowledge gleaned from DFDR. Pre- and post-DFDR interobserver reproducibility of overall diagnostic agreement was assessed. Interobserver reproducibility measured by Fleiss κ values of pre- and post-DFDR was 0.36 and 0.59, respectively (P = .006). Post-DFDR, there were significant improvement for “100% concordance” (P = .011) and reduction for “no consensus” (P = .0004) categories. Despite a lower pre-DFDR reproducibility for non–uropathology fellowship–trained (n = 3, κ = 0.38) versus uropathology fellowship–trained (n = 5, κ = 0.43) pathologists, both groups achieved similarly high post-DFDR κ levels (κ = 0.58 and 0.56, respectively). DFDR represents an effective tool to formally achieve diagnostic consensus and reduce variability associated with critical diagnoses in an anatomic pathology practice.