Analysis of the reliability of manual and automated immunohistochemical staining procedures A pilot study

OBJECTIVE: To study the variation in the number of stained cells and staining intensity comparing 2 immunostainers and manual staining for estrogen receptor (ER) expression in breast carcinoma. STUDY DESIGN: In 5 cases, 15 consecutive paraffin sections were investigated after simultaneous immunohistochemical ER staining. The slides were evaluated using a CM-2 TV image analysis system (Hund, Wetzlar, Germany). One viewing field, identified around a histologic structure present on all 15 sections, was analyzed. The percentage of immunoreactive cells (PP), mean grey values of the immunopositive (GV p o s . ) and immunonegative nuclei (GV n e g . ), and immunohistochemical staining intensity (SI, defined as GV n e g . -GV p o s . ) were calculated. RESULTS: The mean PP values were higher for immunostainers A (70.2 %) and B (53.8% ) than for manual staining (40.8%). The results were significantly different comparing the 2 immunostainers (P=.0143) or immunostainer A and manual staining (P<.0001). Also, the mean SI values were higher for immunostainers A (24.5 ′ 2.8% [CV]) and B (18.5 ′ 31.1%) than for manual staining (10.8 ′33.8%). These differences revealed statistical significance comparing the immunostainers with manual staining (.0001 < P =.0048). CONCLUSION: Our results underline the higher staining quality using immunostainers in comparison with manual staining.