A Long Noncoding RNA-based Classifier for Identifying More Aggressive Low-grade Ta Bladder Cancer with Elevated FGFR3...

Area of Science:

• Oncology
• Genomics
• Molecular Biology

Background:

• Non-muscle-invasive bladder cancer (NMIBC) classification requires improved risk stratification.
• Low-grade, stage Ta (TaLG) NMIBC can exhibit aggressive behavior, leading to recurrence or progression.
• Previous research identified long noncoding RNA cluster 2 (LC2) as a marker of aggressive TaLG NMIBC.

Purpose of the Study:

• To develop and validate a single-sample transcriptome classifier for identifying the aggressive LC2 subgroup in TaLG NMIBC.
• To assess the prognostic significance of the LC2 classifier in predicting recurrence-free survival.
• To understand the biological pathways associated with the LC2 subgroup.

Main Methods:

• Utilized the UROMOL cohort (n=276) for training and the Hurst cohort (n=72) for validation.
• Employed feature selection via median absolute deviation and nested cross-validation.
• Trained an elastic net regression model with ten-fold cross-validation and evaluated performance using pathway analysis, Kaplan-Meier, and Cox regression.

Main Results:

• The classifier successfully identified LC2 tumors, with 7/72 cases detected in the Hurst cohort.
• LC2-TaLG patients demonstrated significantly worse recurrence-free survival (log-rank p < 0.001).
• LC2 status was a strong predictor of recurrence (HR 4.52, p=0.001), and LC2-predicted tumors showed consistent biological patterns.

Conclusions:

• A reproducible single-sample transcriptomic classifier for aggressive TaLG NMIBC has been developed.
• The classifier demonstrates potential for identifying patients at higher risk of recurrence.
• Further validation could support clinical decision-making for NMIBC management.