In:
Cancer Research, American Association for Cancer Research (AACR), Vol. 83, No. 11_Supplement ( 2023-06-02), p. B048-B048
Abstract:
Introduction Many biomarker assays improve sensitivity in diagnosing high grade prostate cancer (PCa), but their performance during active surveillance, after a diagnosis of PCa, is not well understood. Procedures In the multicenter Canary PASS cohort, we explored the contribution of four distinct biomarker panels (blood-based: 4Kpanel, phi; urine-based: PCA3, MDxSelect RNA) and clinical variables in predicting 4-year extreme biopsy reclassification to Gleason Grade Group 3 (GG3) or above. Biomarker assays were performed between 2014-2020 on subcohorts including 727-1,176 participants; outcomes were collected through March 2020. Models were built using a) biomarker, b) biomarker + prostate volume, c) biomarker + biopsy variables d) biomarker + prostate volume + biopsy variables, and, for urine markers e) biomarker + prostate volume + biopsy variables + serum PSA. Clinical PSA was used as a reference. Partly conditional Cox proportional hazards regression models for residual time to event were constructed based on information available at each prediction time, accounting for competing risk. The probability of extreme reclassification within 4 years was calculated after diagnosis and after first follow-up biopsy (Bx1) and the accuracy was assessed with the cross-validated receiver operating characteristic (ROC) curve analysis. Results Cross validated Areas Under the Curve (AUCs) for predictions made with biomarkers ranged from 0.648 (95% CI: 0.581, 0.716) to 0.755 (95% CI: 0.687, 0.824) after diagnosis and from 0.588 (95% CI: 0.480, 0.696) to 0.669 (95% CI: 0.558, 0.780) after Bx1, and were consistently higher than AUCs for clinical PSA at both timepoints; the difference in AUC between biomarker and PSA was statistically significant only for 4Kpanel and phi. Adding clinical variables to predictive models improved AUCs to varying degrees; AUCs for full models with all variables were 0.702 (95% CI: 0.641, 0.764) to 0.776 (95% CI: 0.715, 0.837) after diagnosis and 0.740 (95% CI: 0.656, 0.824) to 0.763 (95% CI: 0.691, 0.836) after Bx1. Adding clinical variables to blood biomarkers resulted in less incremental improvement than for urine markers. A limitation of this study is that prostate MRI was not part of the PASS protocol, although MRI data collected suggests minimal predictive ability during active surveillance. Conclusions Our results suggest that early during prostate cancer active surveillance the blood-based biomarkers of the 4Kpanel and phi may predict future reclassification to higher grade cancer as well as or better than common clinical variables. Clinical variables improve the performance of the urine-based markers of PCA3 or MDxSelect. External validation studies are needed. Citation Format: Lisa F. Newcomb, Yingye Zheng, Menghan Liu, James D. Brooks, Peter R. Carroll, Atreya Dash, William J. Ellis, Christopher J. Filson, Martin E. Gleave, Michael A. Liss, Frances M. Martin, Todd M. Morgan, Peter S. Nelson, Andrew A. Wagner, Daniel W. Lin. Performance of diagnostic biomarkers in the Canary Prostate cancer Active Surveillance Study (PASS) [abstract]. In: Proceedings of the AACR Special Conference: Advances in Prostate Cancer Research; 2023 Mar 15-18; Denver, Colorado. Philadelphia (PA): AACR; Cancer Res 2023;83(11 Suppl):Abstract nr B048.
Type of Medium:
Online Resource
ISSN:
1538-7445
DOI:
10.1158/1538-7445.PRCA2023-B048
Language:
English
Publisher:
American Association for Cancer Research (AACR)
Publication Date:
2023
detail.hit.zdb_id:
2036785-5
detail.hit.zdb_id:
1432-1
detail.hit.zdb_id:
410466-3
Permalink
