Quantitative MRI was evaluated for monitoring post-radiation therapy (RT) response.

Hypoxia score (HS) from DWI and Ktrans from DCE-MRI were found to be most reliable.

The optimal measurement timeframe was 18-months post-RT timepoint.

Early detection of locally recurring prostate cancer (PCa) after definitive radiation therapy (RT) offers the opportunity to deliver targeted salvage therapies, thereby reducing the risk of disease progression. Quantitative MRI (qMRI) parameters show promise as imaging biomarkers for early detection of local recurrence. However, the feasibility of using qMRI for response monitoring in patients undergoing RT combined with androgen deprivation therapy (ADT) remains uncertain. Here, we identified the qMRI parameters with potential to reliably detect post-RT response in PCa and compared the response in patients receiving RT combined with ADT versus those receiving RT alone.

qMRI scans were acquired before and at 6-, 12-, and 18-months after standard definitive RT in sixteen patients with localised PCa. Patients undergoing neoadjuvant ADT were also scanned pre-ADT. Mean value of ADC, D, f, HS, R2*, T1, Ktrans, ve within the tumour were calculated at each imaging timepoint. Statistical significance of treatment-related changes was assessed using rANOVA and post hoc two-tailed t-test. Changes relative to the baseline scan exceeding the parameter uncertainty were classified as “detectable”.

Ktrans and HS measured at 18-months post-RT were found to be most reliable for detecting treatment response regardless of ADT use. Significant post-treatment changes were observed in other qMRI parameters but were unreliable due to large measurement uncertainties.

Quantitative MRI show promise for reliably detecting treatment response within 18-months post-RT. Future clinical trials should validate the potential of Ktrans and HS by correlating these parameters with treatment outcomes.

Prostate cancer

Radiation therapy

Androgen deprivation therapy

Treatment response monitoring

Quantitative MRI

Longitudinal imaging

© 2025 The Authors. Published by Elsevier Inc.