Blood mRNA Model Predicts Castration Resistance in Metastatic Prostate Cancer

A prognostic model built from 3 peripheral blood messenger RNA (mRNA) markers accurately predicted progression to metastatic castration-resistant prostate cancer (CRPC) among patients with newly diagnosed metastatic hormone-sensitive prostate cancer (HSPC) receiving first-line androgen deprivation therapy (ADT)–based treatment, according to results from a prospective, single-center cohort study published in Frontiers in Oncology.¹ The investigators also identified elevated SLC6A8 expression as an independent risk factor for progression and a potential driver of glycolysis in CRPC cells.

The model, which incorporated SLC25A39, SLC6A8, and YBX3 expression, achieved an area under the curve (AUC) for predicting 1-year metastatic CRPC–free survival of 0.89 (95% CI, 0.80-0.97) in the development cohort and 0.83 (95% CI, 0.62-1.00) in the validation cohort. Corresponding C-index values were 0.849 (95% CI, 0.795-0.899) and 0.783 (95% CI, 0.688-0.880), respectively, and Brier scores were 0.056 and 0.060.

Patients classified as high risk by the model experienced significantly faster progression to metastatic CRPC vs those classified as low risk (HR, 17.19; 95% CI, 6.55-45.11; P <.0001). The median metastatic CRPC–free survival was 466 days in the high-risk group and was not reached in the low-risk group. In multivariate Cox analysis, elevated peripheral blood expression of SLC25A39 and SLC6A8 emerged as independent predictors of progression.

Compared with high-risk criteria from the LATITUDE trial (NCT01715285), the 3-gene model demonstrated higher specificity at 365 days (82.1% vs 51.3%; adjusted P = .016) and higher accuracy (81.0% vs 52.4%; adjusted P = .018).² At 545 days, the model again showed greater specificity (94.7% vs 26.3%; adjusted P <.001), accuracy (81.1% vs 29.3%; adjusted P <.001), and balanced accuracy (75.0% vs 30.6%; adjusted P = .001). Sensitivity did not differ significantly between the approaches at 365 days (67.2% vs 66.4%; adjusted P = .988) or 545 days (55.3% vs 34.9%; adjusted P = .474). A decision curve analysis showed greater net benefit for the model than "treat-all" and "treat-none" strategies at threshold probabilities between approximately 0.2 and 0.6.

"Our transcriptomic signature offers a framework for precision risk stratification that could inform treatment decisions. First, for patients in whom clinical risk is equivocal—such as those with oligometastatic bone disease or borderline [prostate-specific antigen (PSA)] levels—the signature could guide intensification with triplet therapy or earlier combination with local treatment (eg, metastasis-directed radiotherapy) if the molecular profile predicts rapid progression," corresponding author Shengke Yang, of the Department of Urology at Sichuan Cancer Hospital and Institute in Chengdu, China, wrote in the publication with study coinvestigators.¹

The authors added that patients with high-volume disease but a favorable molecular signature might be candidates for omitting docetaxel, and that the signature could enrich clinical trials for patients with aggressive molecular phenotypes. They emphasized that these applications remain hypothetical pending prospective validation.

Investigators first analyzed single-cell transcriptomic data from the GSE288878 data set to identify genes associated with metastatic CRPC, then performed transcriptome sequencing on peripheral blood from a prospective cohort at Sichuan Cancer Hospital. Of 230 patients screened between February 1, 2024, and February 1, 2025, 157 were enrolled and randomly split 7:3 into development and validation cohorts; 147 patients (103 and 44, respectively) completed follow-up.

Eligible patients had de novo metastatic HSPC, at least 1 measurable lesion per RECIST 1.1, an ECOG performance status of 2 or lower, and received ADT plus an androgen receptor pathway inhibitor with or without docetaxel. Progression was adjudicated according to Prostate Cancer Clinical Trials Working Group 3 criteria. At a median follow-up of 611 days (95% CI, 560-641), 54 patients had progressed to metastatic CRPC.

The primary end point of the study was metastatic CRPC–free survival. Intersecting differentially expressed genes from the cohort with the single-cell data yielded 5 candidate genes (SLC25A39, GRB10, SELENBP1, SLC6A8, and YBX3), from which elastic net–regularized Cox regression selected the final 3.

In The Cancer Genome Atlas prostate cancer data set (n = 498), elevated SLC6A8 expression in tumor tissue was an independent risk factor for shorter progression-free survival (PFS). In C4-2B CRPC cells, the SLC6A8 inhibitor ompenaclid suppressed proliferation and promoted apoptosis in a dose-dependent manner, downregulated the glycolysis-associated genes HK2, PFKP, and ALDH1B1, and reduced glycolytic metabolites including lactic acid, pyruvic acid, and fructose-1,6-bisphosphate.

The authors noted several limitations, including the single-center design in an exclusively East Asian population, the lack of an external validation data set, a small validation cohort with limited events that produced wide CIs, and follow-up sufficient only for 1-year and 1.5-year projections. The mechanistic findings relied on pharmacologic inhibition in a single cell line. A prospective multicenter validation study across 3 to 5 tertiary centers in China is being planned.

References

1. Li S, Zhou H, Wu Y, et al. Peripheral blood mRNA-based prognostic model for progression to metastatic castration-resistant prostate cancer and identification of SLC6A8 as a glycolysis-promoting biomarker. Front Oncol. Published online October 14, 2026. doi:10.3389/fonc.2026.1942820
2. Fizazi K, Tran N, Fein L, et al. Abiraterone plus prednisone in metastatic, castration-sensitive prostate cancer. N Engl J Med. 2017;377(4):352-360. doi:10.1056/NEJMoa1704174