Reproducibility of ductal lavage cytology and cellularity over a six month interval in high risk women

Background Ductal lavage (DL) allows repeat sampling of breast epithelium for serial observation in a chemoprevention setting; however, the reproducibility of duct cannulation, cell yield and cytology has not been addressed. Methods We conducted a Phase 2 trial, wherein high risk women chose tamoxifen treatment or observation following an entry DL procedure. We present data from the non-intervention arm of our study to assess the reproducibility of cannulation, cell yield, and cytologic diagnosis from DL of the same duct at two time-points. Inter-observer variability was assessed by a blinded review of Papanicoloau-stained slides by two cytopathologists. Results Sixty-five women had a successful lavage of 187 ducts at baseline and chose observation; 63/65 (97%) had a successful lavage 6 months later. Successful recannulation of the same duct was accomplished in 63 women (97%) and162 ducts (87%). Total epithelial cell yields ≥100 were obtained from 57/65 women (88%) and 129/187 ducts (69%) at baseline, and 46/63 women (73%) and 80/162 ducts (49%) at both time-points. Cytologic diagnosis was reproducible in 27/63 (43%) women and 77/162 (48%) ducts. Inter-observer variability for cytologic diagnosis between two observers showed good agreement (κ = 0.62). Conclusions Recannulation and lavage of the same duct after a 6 month interval can be achieved with high frequency; however, reproducibility of cell yield and cytologic findings from the same duct is sub-optimal, leading to significant attrition of evaluable subjects. The utility of DL for the serial monitoring of breast epithelium is therefore limited.