Lajos Pusztai, MD, DPhil
It is unlikely that any single gene can predict response to targeted therapy for patients with HER2-positive breast cancer, new findings from a study suggest. Instead, gene networks— specifically those involving PI3 kinase (PI3K)—may provide a clearer picture of patient outcomes.
Researchers from Yale Cancer Center analyzed 203 biopsies from the NeoALTTO trial, which examined patients with HER2-positive breast cancer who received preoperative therapy with paclitaxel and lapatinib (Tykerb), paclitaxel and trastuzumab (Herceptin), or paclitaxel plus both anti-HER2 drugs.
In the analysis, researchers performed whole-exome sequencing of pretreatment biopsies and examined whether genome-wide metrics of overall mutational load, clonal heterogeneity, or alterations at variant, gene, and pathway levels were associated with treatment response and survival.
They found that no recurrent single gene mutations were significantly associated with pathologic complete response (pCR), except PIK3CA
[odds ratio, 0.42; P
= .0185]. Mutations in 33 of 714 pathways were significantly associated with response, but different genes were affected in different individuals. PIK3CA
was present in 23 of these pathways, defining what researchers call the “trastuzumab-resistance network” of 459 genes. Cases with mutations in this network had low pCR rates to trastuzumab compared with cases with no mutations.
To learn more about the significance of these findings, OncLive
spoke with study author Lajos Pusztai, MD, DPhil, professor of Medicine, chief of Breast Medical Oncology, co-director Yale Cancer Center Genetics, Genomics and Epigenetics, Yale Cancer Center, Yale School of Medicine. In the interview, Pusztai explains how these findings could impact biomarker research and the future treatment paradigm for HER2-positive breast cancer.
OncLive: What was the goal of this research?
: The goal of this research was to see if we could find predictive markers that could define the patient population that needs trastuzumab, lapatinib, the combination of both, or those for whom it doesn’t matter which of these therapies they get.
It is a part of a large attempt to find biomarkers. The clinical trial group has already published results of single marker-like mutations in PI3K
and on the value of immune cells in the tumor microenvironment. This was an additional piece of this puzzle; we looked at DNA alterations that could predict response or resistance to these drugs. We looked at the entire coding region of all the human genes that are expressed or are present in the human genome. We did whole exome sequencing on these pretreatment biopsies.
The goal was to see if there were any DNA-level alterations—either a single mutation in a gene or the overall mutational burden, or neoantigens that the mutations generate—that would be predictive of response, or if any pathway level alterations were predicative. We organized the genes in the human genome into various biological pathways—groups of genes that provide a given biological function.
What were the most significant findings?
What we found was that there is no single individual gene that would be very powerful as a predictor, because every cancer is sort of sensitive or resistant in its own way. However, we could identify pathway-level alterations that were associated with better or worse response.
Here, in this study, we really classified patients as having a pCR or not having one. If patients did not have a pCR, we referred to them having residual disease or being resistant to therapy. If they had a complete response (CR), we called them sensitive. A CR is really extreme chemotherapy sensitivity, because the cancer is completely gone from the breast by the end of the treatment.
What we found is that patients who had not achieved a pCR tended to have mutations in genes that were involved in the PI3K signaling network. In the past, we learned that if patients have a change in the PI3K gene, then they tend to have a lesser response.
However, what this adds to this literature is that it is really not just PI3K alone, but many of its partners that carry the exact same functional impact. Any kind of break in the PI3K network leads to the same phenotypic effect, and that effect is that these tumors are less sensitive to HER2-targeted therapy.
How could this understanding be applied to treating patients with HER2-positive breast cancer?
Understanding the real biology behind a particular disease or a response of treatment is always helpful in the long run, because there could be new ways to exploit this knowledge to develop therapies.