NEW YORK – New research suggests a 23-gene clonal expression biomarker known as the "Outcome Risk Associated Lung Expression" or ORACLE test can provide prognostic insights for individuals with early-stage lung adenocarcinoma, a form of non-small cell lung cancer (NSCLC).
"The typical single-site needle biopsy samples less than 1 percent of the primary tumor mass, failing to capture the full extent of genetic and transcriptomic [intratumor heterogeneity] within individual tumors," the researchers wrote in a paper published in Nature Cancer on Thursday. The team was led by Charles Swanton of University College London and the Francis Crick Institute and Nicolai Birkbak of University College London, the Francis Crick Institute, and Aarhus University.
"To address this sampling bias problem," the authors explained, "we previously reported the development of a clonal expression biomarker (ORACLE), which is associated with [overall survival] outcomes in retrospective cohorts."
As part of their effort to prospectively validate the ORACLE approach, the investigators profiled mutation and clonal gene expression patterns across 450 tumor regions from 184 treatment-naïve TRACERx study participants with stage I to stage III lung adenocarcinoma.
Specifically, they looked at clonal mutation frequencies in several cancer driver genes in tumor regions classified as high-risk or low-risk by the ORACLE test as outlined in a study that Swanton and other TRACERx Consortium members published in Nature Medicine in 2019, identifying relationships between the genetic alterations and risky expression profiles revealed by ORACLE.
From there, the team considered the potential consequences of tumor sampling bias in an ORACLE validation cohort comprised of 158 TRACERx participants with lung adenocarcinoma who had not been tested in the past, focusing on 122 stage I to III lung adenocarcinoma patients with available multiregion RNA-seq data spanning 333 tumor regions.
By comparing the ORACLE scores for each tumor region with classifications developed for the corresponding tumor overall, the researchers found that all but 19 of the 122 tumor regions had high- or low-risk classifications that were concordant with the risk estimates established for the broader tumor.
Across all 158 lung adenocarcinoma patients in the cohort, meanwhile, the team's overall survival analyses linked age-, sex-, smoking-, tumor stage-, and treatment-adjusted overall survival times to risk estimates based on average ORACLE scores for distinct regions from the same tumor.
Similarly, when the team analyzed data from random tumor region samples to simulate the availability of single-biopsy samples in clinical settings, it continued to see significant ties between overall survival and ORACLE scores — findings that were validated further using data for another 70 stage I cases of lung adenocarcinoma from the TRACERx study.
"The association between ORACLE and clinical outcomes was significant for lung cancer-specific survival and [disease-free survival]," the authors reported. "As an RNA marker, ORACLE complemented the use of liquid biopsy (ctDNA) and pathology (STAS) markers to predict 5-year survival outcomes."
The analyses further pointed to apparent ties between ORACLE classifications and adjuvant chemotherapy response and lung cancer evolution patterns. In addition, the investigators showed that the clonal expression approach can help to identify lung adenocarcinoma cases with increased potential for invasive or metastatic disease.
Together, the authors explained, the current findings suggest that expression-based markers and clinical-grade assays developed in the future may hold the potential for guiding treatment decisions and patient risk stratification in lung cancer, similar to the existing Oncotype DX or Agendia MammaPrint assays in breast cancer.
"Breast cancer trials have prospectively evaluated the use of RNA markers to refine risk stratification for chemotherapy, thereby reducing overtreatment," they noted. "A similar approach, designing a randomized Phase III trial comparing observation versus chemotherapy or closer surveillance for ORACLE high-risk tumors, may similarly move the needle for precision diagnostics in lung cancer."