Abstract LB113: Genomic classification to refine prognosis in clear cell renal cell carcinoma
Bibliographic record
Abstract
Abstract Renal cell carcinomas (RCC) are characterized by their heterogenous clinical outcomes, and due to their indeterminate behavior and the absence of routine biomarkers, it is difficult to identify patients who are at high-risk for relapse after curative nephrectomy. To identify genomic biomarkers for clear cell RCC (ccRCC) risk-stratification we interrogated somatic mutation status of 12 RCC-relevant genes using next-generation sequencing (NGS) in tumor-normal pairs from 943 patients with matched follow up data from the Cancer Genomics of the Kidney (CAGEKID) study. We examined associations between genomically-defined patient groups, explained below, and disease-free as well as RCC-specific survival independently in two cohorts of patients (N=469 for cohort 1; 474 for cohort 2). We used the Kaplan-Meier method with log-rank tests to compare survival functions, and Cox proportional hazards models to stratify for patient stage and age to estimate association of each group with survival. RCC-specific survival was assessed with a competing-risks method to include deaths from other causes. Within these cohorts, 76.4% of patients harbored somatic mutations in VHL, the most common driver gene in ccRCC. The most commonly mutated genes within VHL-mutated tumors were PBRM1 (39.7%), SETD2 (19%), BAP1 (14.3%), and KDM5C (8.3%). Less frequently mutated genes included ATM, COL11A1, DMD, TP53, and TRRAP (~3-5%).Among VHL-driven tumors, we identified a new genomic classifier on the basis of the number of mutations in additional RCC driver genes in the panel examined. Patients were classified based on the presence of mutations only in VHL (VHL+0), those with mutations in VHL and one other driver gene (VHL+1), two other driver genes (VHL+2), and 3 or more other driver genes (VHL≥3). We observed within both cohorts that both the risk of disease recurrence as well as RCC-specific death were associated with an increased number of mutations within this classification. When stratified for patient stage and age, the hazard-ratio for 5-year disease-free survival for VHL≥3 patients was 6.69 (p=0.000212), 4.31 for VHL+2 (p=0.000862), and 2.43 for VHL+1 (p=0.035662), compared to patients with only mutations in VHL. These observations were replicated in the second patient cohort, with hazards ratios of 4.55, 2.49, and 1.40, for VHL≥3, VHL+2, and VHL+1 classified patients respectively, indicating that risk of disease recurrence increases with the number of driver mutations. Notably, tumor mutational burden (TMB) was not significantly different between the aforementioned groups, demonstrating that our classifier is independent of TMB. We created a model based on a set of 12 RCC-relevant genes, which can predict risk of relapse for the ~80% of patients with ccRCC that are VHL-driven. This classification can be defined based on a small panel of genes, making it easily applicable to the clinic, in the context of tumor or liquid biopsy analysis. Citation Format: Kate I. Glennon, Naveen S. Vasudev, Ghislaine Scelo, Michelle Wilson, Louis Letourneau, Robert Eveleigh, Nazanin Nourbehesht, Madeleine Arseneault, Antoine Paccard, Lars Egevad, Juris Viksna, Edgars Celms, Sharon M. Jackson, Behnoush Abedi-Ardekani, Anne Y. Warren, Peter J. Selby, Sebastian Trainor, Michael Kimuli, Naeem Soomro, Adebanji Adeyoju, Poulam Patel, Magdalena B. Wozniak, Ivana Holcatova, Antonin Brisuda, Vladimir Janout, Estelle Chanudet, David Zaridze, Anush Moukeria, Oxana Shangina, Lenka Foretova, Marie Navratilova, Dana Mates, Viorel Jinga, Ljiljana Bogdanovic, Bozidar Kovacevic, Anne Cambon-Thomsen, Guillaume Bourque, Alvis Brazma, Jörg Tost, Paul Brennan, Mark Lathrop, Yasser Riazalhosseini, Rosamonde E. Banks. Genomic classification to refine prognosis in clear cell renal cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr LB113.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".