Validation of a Morphology-Based Risk Taxonomy (Favorable, Borderline, Unfavorable Histology) for Prostate Cancer Using PLCO Trial Data
Principal Investigator
Name
Enrico Munari
Degrees
M.D., Ph.D.
Institution
University Hospital Trust Verona
Position Title
PI
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCOI-2019
Initial CDAS Request Approval
Feb 4, 2026
Title
Validation of a Morphology-Based Risk Taxonomy (Favorable, Borderline, Unfavorable Histology) for Prostate Cancer Using PLCO Trial Data
Summary
Background and Rationale
Histopathological grading remains the cornerstone of prostate cancer risk stratification, yet the conventional Gleason/ISUP Grade Group system incompletely captures the biological heterogeneity of disease, particularly within Gleason pattern 4. Over the past decade, outcome-driven morphologic studies have identified specific architectural patterns—most notably cribriform and intraductal carcinoma—as reproducible and powerful predictors of metastasis and prostate cancer–specific mortality. Building on this evidence, McKenney and colleagues demonstrated that architectural phenotype can outperform traditional Gleason grading in predicting outcome, while Nguyen et al. recently proposed a refined taxonomy categorizing prostate cancer into favorable, borderline, and unfavorable histology, independent of Grade Group and pathological stage. This taxonomy defines unfavorable histology by the presence of large cribriform or intraductal carcinoma (>0.25 mm), classic Gleason pattern 5, complex papillary or anastomosing architecture, or high-grade stromogenic carcinoma. Borderline histology encompasses complex Gleason pattern 4 subtypes (e.g., small cribriform or glomeruloid glands ≤0.25 mm, predominant poorly formed glands), whereas favorable histology lacks these adverse architectural features. Early single-institution studies suggest that this framework more accurately identifies tumors with true metastatic potential and, critically, distinguishes indolent from aggressive disease within Grade Group 2–3 cancers. However, large-scale validation in a population-based cohort with long-term follow-up is lacking. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial offers a unique opportunity to validate this morphology-based risk taxonomy due to its well-annotated clinical data, long follow-up, and availability of diagnostic and surgical pathology material linked to oncologic outcomes.
Objectives and Study Design
We propose a retrospective pathology-outcomes study leveraging PLCO prostate cancer cases with digitized available histologic material and follow-up data. Histologic slides will be centrally reviewed by expert uropathologists and classified according to the favorable/borderline/unfavorable histology framework as defined by McKenney and Nguyen.
The primary endpoint will be overall survival, with secondary endpoints including biochemical recurrence and prostate cancer–specific mortality. Multivariable Cox regression models will compare the prognostic performance of the histology-based taxonomy against ISUP Grade Groups, adjusting for established clinicopathologic variables. Model discrimination will be assessed using concordance indices and likelihood-ratio testing.
Significance
Validation of this simplified, architecture-driven taxonomy in PLCO dataset would provide high-level evidence that prostate cancer risk can be more accurately stratified using biologically grounded histologic features rather than numeric grade alone. This approach has direct implications for clinical decision-making, particularly in identifying patients with intermediate-grade disease who may be safely managed with active surveillance versus those harboring high-risk biology.
Aims
-Aim 1: To validate the prognostic value of favorable, borderline, and unfavorable histology in prostate cancer within the PLCO cohort, by correlating this morphology-based classification with long-term clinical outcomes. The primary endpoint will be overall survival, with secondary analyses including biochemical recurrence and prostate cancer–specific mortality. This aim seeks to establish whether a simplified, architecture-driven histologic taxonomy provides robust and reproducible risk stratification in a large, population-based dataset with extended follow-up.
-Aim 2: To compare the predictive performance of this morphology-based risk taxonomy with conventional ISUP Grade Groups and established clinicopathologic risk models, including multivariable models incorporating age, PSA, clinical stage, and treatment variables. Model performance will be evaluated using hazard ratios, concordance indices, and likelihood-based metrics to determine whether the histology-based framework offers incremental prognostic value beyond standard grading systems.
-Aim 3: To assess whether unfavorable histology identifies a distinct subset of biologically aggressive disease within intermediate Grade Groups (GG2–3), independent of traditional grading parameters and tumor burden measures. This aim specifically addresses whether unfavorable architectural features can refine risk stratification in patients otherwise classified as intermediate risk, with potential implications for treatment selection and active surveillance eligibility.
Collaborators
Enrico Munari University and Hospital Trust Verona
Rita Polati University Hospital Trust Verona
Luca Cima University Hospital Trust Verona
Pietro Antonini University Hospital Trust Verona