Researcher reviews progression of genetic profiling during plenary lecture


Mitchell Dowsett, PhD, FMedSci
Mitchell Dowsett, PhD, FMedSci

Genetic profiling has been an active area of breast cancer research for more than 20 years. And while genetic profiling has become a useful clinical tool, it remains an inexact science.

“That first paper did not describe the subtypes we have become very used to using ― the luminal subtypes, the basal subtype, ErbB2+ subtypes,” said Mitchell Dowsett, PhD, FMedSci, Head of the Centre for Molecular Pathology, the Ralph Lauren Centre for Breast Cancer Research, and the Academic Department of Biochemistry at the Royal Marsden Hospital and The Institute of Cancer Research, London, UK. “The critical point, which came out later, was that some subtypes, particularly Luminal A, had better outcomes. And a few years later we saw that many tumors, up to 30%, did not seem to relate to any subtype.”

It took nearly a decade to see the first clinically helpful risk predictors based on subtype in estrogen receptor positive (ER+) breast cancers, Dr. Dowsett continued. Using molecular subtyping by PAM50, it was found that Luminal A tumors had lower risk and better outcomes than basal-like tumors and HER2-enriched tumors, while Luminal B tumors showed the worst outcomes.

Dr. Dowsett explored the progression of genetic profiling from early research to clinical utility and looked ahead to clinically critical questions that have yet to be answered during the Plenary Lecture Genomic Profiling in Early State ER Positive Breast Cancers/Precision Medicine on Thursday, December 9. The presentation will be available on-demand to registered SABCS participants until March 10, 2022.

It soon became clear that gene expression is not an on/off switch, but a continuum of activity. Different biopsies from the same tumor can yield different profiles. Tumor phenotypes emerged as a useful marker for clinical management, but clinical utility was complicated because different commercial screens can produce different results. OncotypeDX is dominated by estrogen signaling, for instance, and Prosigna by proliferation.

The relationship between estrogen levels and expression of PGR, TFF1, PDZK1, and GREB1 genes in breast cancers had been noted as early as 2010, Dr. Dowsett continued. Later work confirmed that the estrogenic environment affects gene transcription, but estrogen levels seem to have little impact on the results of commercial gene signatures and the resulting risk scores.

TP53 mutations emerged as one of the most important prognostic markers in primary ER+ breast cancers. A mutational analysis of cancers treated with anastrozole or tamoxifen found that TP53 mutation carries a hazard ratio of 2.23 for distant recurrence compared to wild type TP53.

Transcriptional profiling can be useful in the clinic, Dr. Dowsett added, but results are not always what they seem.

“One of the things we have become increasingly aware of is the presence of artefacts that can occur during transcriptional profiling,” he cautioned.

Whole genome expression arrays conducted during the POETIC (Perioperative Endocrine Therapy for Individualised Care) Trial in ER+ primary breast cancer found that patients treated with aromatase inhibitors showed 3,269 genes changed compared to 110 changed genes in untreated patients. But a detailed comparison of the top-ranked genes in AI-treated and untreated tumors found very similar results.

“Most of the largest gene changes in the aromatase inhibitor group are completely artefactual,” Dr. Dowsett said. “We will be focusing on other genes for an understanding of what is going on in these tumors.”

POETIC also provided clues to mechanisms of endocrine resistance, but few answers. There are clear differences in Ki67 expression between poor responders and good responders. ESR1-low tumors respond poorly. But there are two groups of ESR1-high tumors, some that respond well and some that respond poorly. Gene expression comparisons found that most of the differences are related to immune signaling, but multiple types of immune cells appear to be involved.

“Immune signaling in ER+ tumors is related to poor outcome and poor response to endocrine treatment rather than better response to treatment,” Dr. Dowsett said. “The exact mechanisms behind this still need to be disentangled.”