Established and novel methods and biomarkers in breast cancer

Abstract

The present thesis is based on three published papers. The work arises from three cohorts of women from Trøndelag county in Norway. These were women born between 1886 and 1977, that were diagnosed with breast cancer (BC). After diagnosis they were followed until the end of 2015 or until time of death from BC or death by other causes. Follow-up data was made accessible from national registries. The main aim of this thesis was to study established biomarkers in BC assessment with new approaches, and to study the validity of a new biomarker. In the first study we assessed expression of the proliferation marker Ki-67 in 248 invasive carcinomas (NST) using both the conventional method (visually in a light microscope) and digital image analysis (QuPath software). We compared the results from these two methods and found that assessment of Ki-67 in breast tumours using digital image analysis identified a greater proportion of cases with high Ki-67 levels compared to visual assessment of the same tumours. We concluded that diagnostic cut-off levels should be recalibrated on the introduction of new methodology. In the second study we assessed PAK1 copy number (CN) in 512 breast tumours using fluorescence in situ hybridization (FISH) on tissue microarray (TMA) slides. Copy numbers were estimated by counting the number of fluorescent signals for PAK1 and the chromosome enumeration probe for chromosome 11 (CEP11) in 20 tumour cell nuclei/case. We assessed associations between PAK1 CN and tumour features, and PAK1 and CCND1 CNs. CCND1 is located close to PAK1 on chromosome 11. We found that PAK1 CN increase was associated with high proliferation and high histopathological grade, but not with prognosis. PAK1 CN increase was most frequent in the HER2- and Luminal B (HER2-) subtypes of BC. PAK1 CN increase was associated with CN increase of CCND1. We found no significant difference in CN increase and risk of death from BC between cases with increased CN of PAK1 alone, CCND1 alone, or cases with increased CN for both genes. In the third study, we assessed associations between levels of Oestrogen Receptor (ER) expression and tumour characteristics, and prognosis in 1955 cases of BC. All cases were stratified into patients unlikely to have received adjuvant therapy according to treatment guidelines at the time of diagnosis (before 1995), and those who could have received adjuvant therapy (diagnosed in 1995 or later). ER status was divided into three categories: <1%, ≥1<10%, and ≥10% positive tumour cell nuclei. Histopathological grade, proliferation status, and molecular subtypes were correlated with ER-status within each time period, and across time periods. The highest proportion of ER Low Positive tumours (ER≥1<10%) were found in the Luminal B (HER2+) subtype and grade 3 tumours. Risk of death from BC was lower in ER Low Positive and ER≥10% compared to ER negative BCs. Women diagnosed in 1995 or later had a higher proportion of ER Low Positive BC, and the tumours were of smaller size, lower grade and lower proliferation than tumours in women diagnosed before 1995.