NCCN Adds ArteraAI Post-RP Test for Biochemical Recurrence of Prostate Cancer

The NCCN has added the ArteraAI Prostate Test (Post-RP) to version 1.2027 of the Clinical Practice Guidelines in Oncology for Prostate Cancer as a risk stratification tool for patients with biochemical recurrence (BCR) after radical prostatectomy (RP), according to a news release from Artera.^1,2

The ArteraAI Prostate Test, a multimodal artificial intelligence (MMAI) test, was previously included in the guidelines for localized prostate cancer in 2024. In the validation study of the Post-RP version published in European Urology, the MMAI score was associated with distant metastasis (subdistribution HR per SD, 2.17; 95% CI, 1.65-2.85; P <.001) and remained prognostic after adjustment for clinical variables and treatment.³ The test found smaller benefit to adding androgen deprivation therapy (ADT) to RP in patients with low risk by MMAI than those with high-risk disease.²

“Artera’s technology supports treatment decisions at multiple points in prostate cancer care,” said Andre Esteva, PhD, chief executive officer and cofounder of Artera, in the news release.¹

Validation Study Design

The test applies a MMAI model to digitized hematoxylin and eosin–stained prostatectomy slides combined with 5 clinical variables: pathologic grade group, pathologic T stage, prostate-specific antigen (PSA) level before salvage radiotherapy, age, and surgical margin status.³ No additional tissue is needed, and the test has been commercially available since December 2025.¹

Investigators developed the model using data from 2 phase 3 trials of salvage radiotherapy with or without hormone therapy after RP: NRG/RTOG 9601 (NCT00002874), which tested long-term bicalutamide, and NRG/RTOG 0534 (NCT00567580), which tested short-term ADT with or without pelvic nodal radiotherapy. Of 1855 eligible patients with digitized slides and clinical data, 1322 were assigned to model development and 533 to a held-out validation set that was not accessible to the development team. The model was locked before validation, and distant metastasis was the primary end point.

In the validation cohort, the median age was 64 years, the median pre–salvage radiotherapy PSA level was 0.4 ng/mL, 9.2% of patients were Black, and median follow-up was 9.2 years for censored patients.

Prognostic and Predictive Findings

At 10 years, the time-dependent area under the curve for distant metastasis was 0.74 (95% CI, 0.65-0.81) with the MMAI model vs 0.68 (95% CI, 0.59-0.75) with a clinical nomogram. Using a prespecified cut point at the 60th percentile, the 10-year incidence of distant metastasis was 25% (95% CI, 19%-32%) in the MMAI high-risk group vs 8.8% (95% CI, 5.7%-13%) in the low-risk group, and 10-year prostate cancer–specific mortality was 13% vs 3.9%.

In the high-risk group, 10-year distant metastasis rates were 38% with salvage radiotherapy alone vs 17% with added hormone therapy (HR, 2.10; 95% CI, 1.22-3.61). In the low-risk group, the corresponding rates were 10% vs 7.8% (HR, 1.27; 95% CI, 0.59-2.72). The test for interaction between the biomarker and treatment was not statistically significant (P = .3).

According to Artera, separate external validation has been performed in institutional cohorts at Mayo Clinic and the University of California, San Francisco; those results were not part of the published study.¹

Limitations and Context

The authors noted that the analysis is a post hoc study of archived trial cohorts and represents internal validation. Patients were treated before PSMA PET/CT was routinely used and before current therapeutic options were available, and further validation in underrepresented racial and ethnic groups is needed. The authors also wrote that the model outputs a relative risk score rather than absolute probabilities.³ The NCCN cautioned that using these prognostic tools to influence treatment decisions should be done carefully in the absence of prospective trials.² 

The paper’s authors noted that external and prospective validation are ongoing.³ The ArteraAI Prostate Test is also being used to stratify patients in the phase 3 PROSTATE-IQ trial (NCT06274047), which began enrolling in 2025.¹ The trial is enrolling approximately 120 patients with BCR after prostatectomy who are eligible for salvage radiation plus androgen deprivation therapy (ADT). Patients classified as low risk are randomly assigned to 6 months of ADT or apalutamide (Erleada) monotherapy, and those classified as high risk to 24 months of ADT or 6 months of ADT plus apalutamide, with quality of life as the primary end point.⁴

“The addition of our test within the NCCN Guidelines for use in the post-RP salvage setting reinforces the role of AI-powered insights in helping clinicians make more personalized, evidence-based treatment decisions throughout a patient’s cancer journey,” Esteva stated in the news release.¹

REFERENCES

1. ArteraAI Prostate Test (Post-RP) Added to NCCN Guidelines. News release. Artera. November 11, 2026. Accessed November 12, 2026. https://tinyurl.com/4e3t5xde

2. NCCN. Clinical Practice Guidelines in Oncology. Prostate cancer, version 1.2027. Accessed November 12, 2026. https://tinyurl.com/3tfnmh9x

3. Morgan TM, Ren Y, Tang S, et al. Development and validation of a multimodal artificial intelligence–derived digital pathology–based biomarker predicting metastasis among patients with biochemical recurrence after radical prostatectomy in NRG/RTOG trials. Eur Urol. 2026;90:470-478. doi:10.1016/j.eururo.2025.12.007

4. PROSTATE-IQ: Parallel randOmized STudy of personalized apalutamide treatment and evaluation to improve quality of life in post-operative radiation with androgen axis suppression. a phase III multi-center study for men with detectable PSA after prostatectomy for prostate cancer. ClinicalTrials.gov. Last updated July 22, 2026. Accessed November 12, 2026. https://clinicaltrials.gov/study/NCT06274047