On This Page

← Back to Highlights

Research Highlight · JCO / ASCO Conference Abstract 2026

Transcriptomic signatures for treatment outcome prediction in renal cell carcinoma

June 2026 · 4 min read · ASCO 2026 Conference Abstract

Authors: Peng Li, Zuhaier Majeed, Sena Ozgul, M. I. H. Ali, Nikhil Single, Daniel G. Stover, Mark S. Makary, F. Bicer, A. Mortazavi, W. Kimryn Rathmell, Eric A. Singer, Mohammad K. K. Niazi, Rulong Wu, Milan Hasanov, and Elshad Hasanov.

#renal-cell-carcinoma#transcriptomics#machine-learning#immunotherapy

Can tumour RNA profiles add prognostic information beyond routine clinical risk grouping in advanced RCC? This ASCO 2026 poster evaluates transcriptomic signatures within sunitinib and avelumab plus axitinib treatment cohorts, covering survival and disease control.

Development data
JAVELIN Renal 101
Data layers
RNA expression + clinical records
Endpoints
PFS · OS · disease control

From RNA profiles to outcome models

The workflow starts with transcriptome-wide feature selection and builds separate models for survival and disease control. The poster distinguishes internal evaluation from external overall-survival validation, so each result can be read against the task it actually tests.

The ASCO poster workflow from JAVELIN RNA profiles through model development and validation
Study workflow extracted from the ASCO 2026 poster. It connects transcriptomic feature selection, survival and disease-control modeling, and internal and external evaluation. Clinical application is the intended direction of the work. Click to enlarge.

What RNA adds beyond IMDC

The most direct comparison is between RNA-only, IMDC-only, and combined models within the same treatment cohort. In the poster analyses, RNA models showed stronger survival discrimination than IMDC, while adding IMDC to RNA produced little further improvement.

Poster comparison of RNA-only, RNA plus IMDC, and IMDC-only survival models
RNA-only, RNA plus IMDC, and IMDC-only C-index comparisons from poster Fig. 4. Left: sunitinib; right: avelumab plus axitinib. Read validation results separately from the higher training estimates. Source: Li et al., ASCO 2026 poster. Click to enlarge.

This comparison asks whether clinical grouping adds information once the transcriptomic score is known. Training and validation estimates should be kept separate; performance in a new patient cohort is the relevant test of transfer.

Disease control and validation

PFS-derived classifiers were also evaluated for disease control versus progressive disease, using ROC curves, precision–recall curves, and confusion matrices in holdout data. External OS analyses used TCGA-KIRC and an OSU Total Cancer Care cohort, as shown in the poster workflow.

These analyses support further study of regimen-specific prognosis. Turning such scores into a treatment-selection rule would additionally require evaluating whether they identify differences in benefit between regimens.

The full poster contains the risk-group survival curves, disease-control results, and comparisons with published signatures. A short video presentation walks through the study.

Resources

ASCO Abstract · JCO DOI · Poster PDF · Video · OncoDaily