Generation and evaluation of protein microarrays for prostate cancer diagnostics

Abstract

Prostate cancer is the most prevalent type of cancer in men.<br />Current diagnostic screening methods are error prone due to low specificity. This leads to a high number of painful, unnecessary and expensive biopsies. Tumor-associated antigens (TAAs) are highly promising biomarkers that are capable of detecting early stages of cancer, even months or years before they become detectable by current clinical diagnostic methods. TAAs give an insight into disease progression and enable customized treatment. The earlier cancer can be detected, the bigger is the chance for cure.<br />In this thesis a TAA protein microarray was developed. Preliminary tests were performed on a protein microarray containing 15,456 proteins. On these arrays a screening of purified serum IgG from 50 prostate cancer patients and 50 patients with biopsy-confirmed benign prostatic hyperplasia was performed. Statistical analyses performed on these data correctly predicted 91% of the sample classes with an ROC curve AUC of 0.924. From the array a set of 718 proteins was defined and a new targeted protein microarray was produced. Validation tests were performed on the targeted array using the same samples as before.<br />Statistical analyses could not reproduce previous results. The best achievable prediction was 61% correct class prediction with an ROC curve AUC of 0.613. Diverse measures were taken to re-establish good classification. This aim could not be met. It was revealed that proteins printed on the targeted array strongly differed in IgG binding capacities from identical ones on the preliminary array.