Prostate-specific membrane antigen (PSMA) based DNA/RNA aptamers enhance the efficacy of the deep learning model for the precise diagnosis to prostate cancer 

Haiying Liu¹, Yichen Yang¹, Jianshe Yang^2,3*

1. Medical College of Hexi University, Zhangye734000, China

2. Shanghai Tenth People’s Hospital, Tongji University School of Medicine, Shanghai 200072, China

3. Shanghai Clinical Nuclear Medicine Center, Shanghai 200072, China

Haiying Liu, associate professor of Hexi University. E-mail:  303803760@qq.com

Yichen Yang, undergraduate student of Hexi University. E-mail: 15002533955@163.com

*Corresponding author: Jianshe Yang, PhD, professor of Tongji University School of Medicine. No. 301 Yanchang Road (Middle), Shanghai 200072, China. Tel/Fax: +86-21-66302721, E-mail: 2305499@tongji.edu.cn

Abstract:

A recent study published in European Journal of Nuclear Medicine and Molecular Imaging furthers our understanding of prostate cancer precision diagnosis. However, diagnostic analysis based on clinical outcomes or downstream biomarkers leads to inaccurate disease diagnosis owing to poor differentiation of malignancy caused by other prostate diseases. Therefore, we suggest replacing prostate-specific antigen (PSA) with targeted prostate-specific membrane antigen (PSMA)-aptamers in the deep learning (DL) model to effectively overcome the limitation of non-specificity of the biomarker in the differential diagnosis of prostate cancer from other prostate diseases, thereby increasing its efficacy as a diagnostic method for prostate cancer.

Key words: PSMA; DNA/RNA aptamers; deep learning model; precision diagnosis; prostate cancer

Recently, an article published in the European Journal of Nuclear Medicine and Molecular Imaging details a novel deep learning (DL) model for the diagnosis of prostate cancer [1]. Based on clinically significant outcomes, the study highlighted the model as a superior diagnostic method for prostate cancer. However, diagnostic analysis based on clinical outcomes or downstream biomarkers frequently leads to inaccurate diagnosis of the disease because of poor differentiation between malignancy and prostate diseases. 

Accordingly, the DL model presents several limitations in being used as a diagnostic tool for prostate cancer. Firstly, the model has limited use for the differential diagnosis of benign and malignant lesions (DL-BM), and cs prostate cancer (PCa) and non-csPCa (DL-CS), compared to that of the traditional prostate imaging reporting and data system (PI-RADS) model. Secondly, reconstruction of targeted pictures for distinguishing benign from malignant tissues is challenging. Thirdly, the levels of prostate-specific antigen (PSA) in the training and validation patient cohorts were significantly different and unstandardised. Further, the ResNet3D alone, but not the other three sub-units had priority to PI-RADS. Although the DL model has excellent operational capability, its efficacy is compromised because of the restricted ability of logical decision-making and the stated limitations, implying lack of significant advantage over PI-RADS in terms of diagnostic sensitivity and specificity.

Although PSA exhibits tissue-specific expression restricted to prostate acini and ductal epithelial cells [2], its use as a tumour-specific marker is limited as the total PSA levels are enhanced (free- and compound-PSA) not only in PCa, but also in prostatitis and benign prostatic hyperplasia. By contrast, enhanced expression of prostate-specific membrane antigen (PSMA), a transmembrane glycoprotein expressed on the cell membrane of epithelial cells, persists post-castration and is independent of the degree of tumour cell differentiation, qualifying it as a promising biomarker in comparison to PSA for the evaluation of early diagnosis, recurrence, and progression of prostate cancer [3]. The PSMA expression is significantly enhanced in prostate cancer compared to that in the normal tissue and directly correlates with the aggressiveness of the disease, making it a classical molecular target of PCa. In addition, the large extracellular domain of PSMA facilitates the design of targeting molecules, making it an ideal target for diagnostic and therapeutic applications.

Precise anchoring of highly conserved PSMA genes by external components might aid specific diagnosis of diseases [4]. Aptamer, a DNA/RNA oligonucleotide fragment designed with reference to a specific gene sequence, is highly conserved, has enhanced stability, is easily radionuclide-labelled during synthesis, reflects reverse labelling of downstream products of gene coding, aids real-time visualisation of pathogenic gene state by radiography, and is capable of specifically detecting the occurrence and nature of the disease.

Thus, replacement of PSA with targeted PSMA-aptamers in the DL model would effectively overcome the limitation of non-specificity of the biomarker in the differential diagnosis of prostate cancer from other prostate diseases, thereby increasing its efficacy as a diagnostic method for prostate cancer.

Statements & Declarations

Funding

This work was supported by the National Natural Science Foundation of China (82071964), the Shanghai Municipal Health Commission (GWV-10.1-XK09), and the Shanghai Shenkang Centre (SHDC2020CR2054B).

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

References

1. Zhao L, Bao J, Qiao X, Jin P, Ji Y, Li Z, et al. Predicting clinically significant prostate cancer with a deep learning approach: a multicentre retrospective study. Eur J Nucl Med Mol Imaging. 2022 Nov 21. doi: 10.1007/s00259-022-06036-9. Epub ahead of print.

2. Guo T, Wang XX, Fu H, Tang YC, Meng BQ, Chen CH. Early diagnostic role of PSA combined miR-155 detection in prostate cancer. Eur Rev Med Pharmacol Sci. 2018;22(6):1615-1621. https://doi.org/10.26355/eurrev_201803_14568

3. Chikatamarla VA, Okano S, Jenvey P, Ansaldo A, Roberts MJ, Ramsay SC, et al. Risk of metastatic disease using [18F]PSMA-1007 PET/CT for primary prostate cancer staging. EJNMMI Res. 2021;11(1):128. https://doi.org/10.1186/s13550-021-00869-5

4. Zhong J, Ding J, Deng L, Xiang Y, Liu D, Zhang Y, et al. Selection of DNA Aptamers Recognizing EpCAM-Positive Prostate Cancer by Cell-SELEX for in vitro and in vivo MR Imaging. Drug Des Devel Ther. 2021;15:3985-3996. https://doi.org/10.2147/DDDT.S322854