EUVIMEDOdkrijte EUVIMED
Open access · CC BY 4.0

Cancer Risk and Temporal Sequence Prediction of Prostate-Specific Antigen by Long Short-Term Memory Network

Alex H. Lin, Hoi Wai Chan, Ka Man Cheung, Amy M. K. Chu, Sharon C. L. Ho, Chin Pan Kong, Bryan C. W. Li, Joanna K. M. Ng, Hei Ming Lai, Chun Yan So, Gabriel C. H. Wong, Rong Na and 2 more

Machine Learning and Knowledge Extraction · 2026-07-06 · 8 (7) · 198

DOI: 10.3390/make8070198 ↗

At a glance

Journal article in Machine Learning and Knowledge Extraction, published online on 2026-07-06. Written by Alex H. Lin and 13 co-authors. 39 references. The full text is freely available (open access, CC BY 4.0).

PovzetekBibliografski podatkiCiteRelated studies

Abstract (original language)

Prostate-specific antigen (PSA) is a well-established marker for prostate cancer screening, but current ≥4 ng/mL cutoff suffers from low specificity. This study aims to demonstrate the use of a long short-term memory (LSTM) network for accurate prediction of prostate cancer risk and next sequential PSA value. Hong Kong-wide PSA test data over a 25-year period, including PSA values, time difference between PSA tests, and PSA velocity change, were retrieved for model training with variable PSA cutoffs and sequence length. A total of 1,158,915 PSA tests from 499,342 patients (including 18,629 patients with prostate cancer) were included. Models predicting the next PSA level performed well (accuracy 0.724–0.910, AUROC 0.751–0.892). For a ≥4 ng/mL cutoff, the best model was at sequence length of 4 (AUROC 0.864). Temporal prediction of PSA performance was lower (accuracy 0.739–0.811, AUROC 0.740–0.849). Prostate cancer prediction performed excellent (AUROC 0.888–0.973), the sensitivity (0.734), and specificity (0.962) were high even at the shortest sequence length and a ≥4 ng/mL cutoff (4). Utilizing PSA levels only without additional markers or clinical data, LSTM-based models accurately predicted the next PSA level with modest temporal predictions while significantly improving the specificity of prostate cancer risk prediction, demonstrating their clinical utility.

Cite

APA

Lin, A. H., Chan, H. W., Cheung, K. M., Chu, A. M. K., Ho, S. C. L., Kong, C. P., Li, B. C. W., Ng, J. K. M., Lai, H. M., So, C. Y., Wong, G. C. H., Na, R., Chiu, M. K. L., & Li, J. J. X. (2026). Cancer Risk and Temporal Sequence Prediction of Prostate-Specific Antigen by Long Short-Term Memory Network. Machine Learning and Knowledge Extraction, 8(7), 198. https://doi.org/10.3390/make8070198

Vancouver

Lin AH, Chan HW, Cheung KM, Chu AMK, Ho SCL, Kong CP, et al. Cancer Risk and Temporal Sequence Prediction of Prostate-Specific Antigen by Long Short-Term Memory Network. Machine Learning and Knowledge Extraction. 2026;8(7):198. doi:10.3390/make8070198

BibTeX
@article{Lin2026,
  author = {Lin, Alex H. and Chan, Hoi Wai and Cheung, Ka Man and Chu, Amy M. K. and Ho, Sharon C. L. and Kong, Chin Pan and Li, Bryan C. W. and Ng, Joanna K. M. and Lai, Hei Ming and So, Chun Yan and Wong, Gabriel C. H. and Na, Rong and Chiu, Matthew K. L. and Li, Joshua J. X.},
  title = {Cancer Risk and Temporal Sequence Prediction of Prostate-Specific Antigen by Long Short-Term Memory Network},
  journal = {Machine Learning and Knowledge Extraction},
  year = {2026},
  volume = {8},
  number = {7},
  pages = {198},
  doi = {10.3390/make8070198}
}
Start for free →