Immunohistochemical correlation of steroid receptors and disease-free interval in 206 consecutive cases of breast cancer: validation of telequantification based on global scene segmentation

Abstract

Objective algorithms to evaluate immunocytochemistry in cytology and in histological sections are still under investigation. Clinical value, simplicity, robustness and adaptation to new technologies such as telemedicine and telequantification have to be taken into consideration. The present paper summarizes the experience of the immunoquantification of 206 cases diagnosed from 1990 to 1991, 36 clinical stage I cases, 110 stage II cases and 60 stage III cases. Initial clinical stage IV cases were excluded from the series. Quantification was done on frozen sections incubated with monoclonal estrogen and progesterone receptors (ER, PgR) by the peroxidase antiperoxidase (PAP) technique revealed with DAB. Automatic computerized measurements with a 550-nm filter were based on global scene segmentation, obtaining the parameters of percentage of positive area (pA) and immunoscore (ISC = sum i x pi), being the staining intensity (i) automatically evaluated in 5 grades multiple of the standard deviation (SD) of the control specimen (i x SD) stained with progressive haemathoxyline. Cut point of positive and negative specimens regarding ISC were obtained by semilogarithmic plot with biochemical receptor values (r = 0.73). The whole series showed a significant increase of negative specimens according to clinical stage and a higher number of PgR negative cases (44%) than ER negative cases (32%). The clinical relevance of the quantified parameters were evaluated by free interval disease survival curves being significant in the positive cutting point of ISC = 100 (ER P < 0.06; PgR P < 0.02) and the 40% negative area for the pA (ER P < 0.001; PgR P < 0.004). This quantification technique, which does not require individual cell segmentation, was reproducible at a distance (telequantification through high resolution videophones at 9600 bauds transmission rate; 55 cases). Coefficient of variation (CV) of the parameters in the original and transmitted cases were in the range of 0-5% in specimens with positive areas over 70%; less positive specimens require lower segmentation thresholds to get acceptable coefficients of variation. According to our results, immunohistochemical steroid receptor contents in breast cancer pathology, quantified by the technique of pixel integration, demonstrated the added value of the percentage of positive area in patient follow-up. Those having more than 40% of negative areas have recurrences mainly within the first year.