Scholarlink

Objective <p>Accurate exclusion of cribriform growth (GP4Crib+), an adverse histologic feature in prostate cancer, remains a challenge in selecting intermediate-risk patients for active surveillance (AS). This study evaluates whether an MRI-based radiomics model predicting GP4Crib+ can support AS inclusion decisions under the assumption that intermediate-risk men without GP4Crib+ could be safely managed with AS.</p> Materials and methods <p>This single-center retrospective study was approved by the institutional review board (IRBd21-108) and included men with Cambridge Prognostic Group (CPG) -1, CPG-2, or CPG-3 (Gleason grade (GG) 2) prostate cancer who underwent MRI and radical prostatectomy. In this cohort, a previously trained radiomics model was applied across the whole prostate to generate voxel-wise GP4Crib+ probability maps, to ultimately identify men with GP4Crib. This model was tested in two scenarios: (1) CPG-1+2, and (2) CPG-1+2+3(GG2) patients on biopsy. The reference scenario was CPG-1 men to AS (guideline recommendations).</p> Results <p>We included 127 patients (median age 66 years, interquartile range 47‒78). Standalone radiomics performance was moderate (area under the receiver operating characteristic curve (AUROC): 0.68 overall; 0.60 for CPG-2; and 0.70 for CPG-3(GG2)). At a 0.60 probability threshold, the model reduced overtreatment by 9% in CPG-1 + 2 men (with 1% increase in undertreatment). In CPG-1+2+3(GG2) men, the model reduced overtreatment by 8%, maintaining low undertreatment (3%). Most false negatives involved smaller than 1.5-mm cribriform foci.</p> Conclusion <p>Our MRI-based GP4Crib+ radiomics model is able to support AS decisions in men with CPG-1+2+3(GG2). Despite modest standalone diagnostic performance, its integration into the clinical work-up may help safely expand AS selection to intermediate-risk men, ultimately reducing overtreatment.</p> Relevance statement <p>MRI-based radiomics may support noninvasive exclusion of cribriform growth to guide active surveillance eligibility in intermediate-risk prostate cancer, enabling safer management decisions and reducing unnecessary treatment.</p> Key Points <p><UnorderedList Mark="Bullet"> <ItemContent> <p>Prostate cancer cribriform growth is difficult to detect but crucial for treatment decisions.</p> </ItemContent> <ItemContent> <p>MRI radiomics helps exclude cribriform growth in intermediate-risk prostate cancer patients.</p> </ItemContent> <ItemContent> <p>Radiomics-informed decisions reduced overtreatment under structured AS scenarios.</p> </ItemContent> </UnorderedList></p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

MRI radiomics predicting cribriform growth informs active surveillance decision in intermediate-risk prostate cancer

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月13日

查看全文
  • Mar Fernandez Salamanca,
  • Rita Simões,
  • Malgorzata Deręgowska-Cylke,
  • Eduardo Pais Pooch,
  • Pim J. van Leeuwen,
  • Henk G. van der Poel,
  • Marcos A. S. Guimaraes,
  • Uulke A. van der Heide,
  • Ivo G. Schoots

摘要

Objective

Accurate exclusion of cribriform growth (GP4Crib+), an adverse histologic feature in prostate cancer, remains a challenge in selecting intermediate-risk patients for active surveillance (AS). This study evaluates whether an MRI-based radiomics model predicting GP4Crib+ can support AS inclusion decisions under the assumption that intermediate-risk men without GP4Crib+ could be safely managed with AS.

Materials and methods

This single-center retrospective study was approved by the institutional review board (IRBd21-108) and included men with Cambridge Prognostic Group (CPG) -1, CPG-2, or CPG-3 (Gleason grade (GG) 2) prostate cancer who underwent MRI and radical prostatectomy. In this cohort, a previously trained radiomics model was applied across the whole prostate to generate voxel-wise GP4Crib+ probability maps, to ultimately identify men with GP4Crib. This model was tested in two scenarios: (1) CPG-1+2, and (2) CPG-1+2+3(GG2) patients on biopsy. The reference scenario was CPG-1 men to AS (guideline recommendations).

Results

We included 127 patients (median age 66 years, interquartile range 47‒78). Standalone radiomics performance was moderate (area under the receiver operating characteristic curve (AUROC): 0.68 overall; 0.60 for CPG-2; and 0.70 for CPG-3(GG2)). At a 0.60 probability threshold, the model reduced overtreatment by 9% in CPG-1 + 2 men (with 1% increase in undertreatment). In CPG-1+2+3(GG2) men, the model reduced overtreatment by 8%, maintaining low undertreatment (3%). Most false negatives involved smaller than 1.5-mm cribriform foci.

Conclusion

Our MRI-based GP4Crib+ radiomics model is able to support AS decisions in men with CPG-1+2+3(GG2). Despite modest standalone diagnostic performance, its integration into the clinical work-up may help safely expand AS selection to intermediate-risk men, ultimately reducing overtreatment.

Relevance statement

MRI-based radiomics may support noninvasive exclusion of cribriform growth to guide active surveillance eligibility in intermediate-risk prostate cancer, enabling safer management decisions and reducing unnecessary treatment.

Key Points

Prostate cancer cribriform growth is difficult to detect but crucial for treatment decisions.

MRI radiomics helps exclude cribriform growth in intermediate-risk prostate cancer patients.

Radiomics-informed decisions reduced overtreatment under structured AS scenarios.

Graphical Abstract