The prostate is a gland in the male reproductive system with the main function to produce the seminal fluid that nourishes and transports sperm. Most prostate cancers are slow-growing and are confined to the prostate gland, where they may not cause serious harm. However, while some types of prostate cancer grow slowly and may need minimal or even no treatment, other types are aggressive and can spread quickly to other areas of the body, particularly the bones and lymph nodes. If prostate cancer is detected early, when it is still confined to the prostate gland, the chance for successful treatment is high. Prostate cancer is the fourth most common type of cancer and the second most common in men. The incidence and mortality rates are strongly related to the age with the highest incidence being seen in elderly men above 65 years of age.
Figure 1. The volcano plot shows the adjusted p-value compared to the difference in average protein expression (NPX) for all proteins in prostate cancer compared to all other cancers. The lollipop plot shows the top 10 most important proteins resulting from the cancer prediction model with importance scores ranging between 0 to 100.
Pan-cancer protein panel
14 proteins were selected for identification of prostate cancer (Table 1). Notably, PSA was not included in the analysis and the identified proteins are therefore potentially attractive alternatives or complements to this frequently used and somewhat controversial blood biomarker. The top protein, DNER, is a notch receptor with low tissue specificity, but annotated as related to cancer (Wang Z et al. (2020)). The next three proteins (IL20, FAP and CXCL6) are all secreted to blood of various tissue origin, with angiogenic potential (IL20), enhance tumor growth (FAP) and chemotactic for neutrophils (CXCL6). Many of the other proteins selected by the model to predict prostate cancer are according to literature related to cancer, but the involvement in prostate cancer must be further explored. An earlier study on prostate cancer (Liu S et al. (2021)) has identified S100A4 as being upregulated and this protein is also found to have elevated plasma levels in our study, but was not used in the panel selected by the AI-based prediction model. The 14 selected panel proteins are able to predict patients with prostate cancers with relatively high confidence.