By George W. Sledge, Jr., MD
San Antonio is the largest breast-only cancer meeting every year, and arguably the most impactful. The meeting was started 4 decades ago by my mentor, Bill McGuire, and rapidly became the premiere venue for presenting both clinical and laboratory research. This year has been a good one both for breast cancer biology and treatment, with data that is immediately applicable to the clinic. I'll focus on just a few key take-home messages, involving treatment advances in HER2-positive, ER-positive, and triple-negative breast cancer, as well as the emerging science of genomics.
T-DM1: Best in Show
The highlight of the Symposium, from a clinical standpoint, was the presentation of the KATHERINE trial (please don't ask me what the acronym stands for). The trial started with a well-known observation: if you treat a patient with neoadjuvant therapy, and her surgery specimen demonstrates significant residual cancer burden for HER2-positive disease, she is at significant risk for early recurrence of breast cancer. This setting is one in which it is rational to apply novel therapeutic interventions in hopes of diminishing recurrence of disease. The high event rate with short follow-up times makes this an ideal study population.
KATHERINE, presented by Charles Guyer, MD, on behalf of an international group of investigators, randomized HER2-positive patients with significant residual cancer burden after neoadjuvant HER2-targeted therapy to either trastuzumab or to T-DM1 out to a year of total HER2-targeted therapy. The results, published simultaneously with the oral presentation in the New England Journal of Medicine, are striking (2018; doi:10.1056/NEJMoa1814017). T-DM1 improved invasive disease-free survival (the study's primary endpoint) by an impressive 11.3% at a median 41 months of follow-up; similar improvements were seen for distant disease-free survival, and overall survival—though early—is also trending in a positive way.
This represents a major win in HER2-positive breast cancer, and T-DM1 now joins the list of HER2-targeted therapies that improve outcome in the setting of micrometastatic disease. I suspect that this approach will receive rapid approval by the FDA and become the new standard of care in the post-neoadjuvant population. It will likely obliterate the use of post-neoadjuvant pertuzumab and/or neratinib, as pointed out by discussant Dr. Eric Winer, MD, and may well reduce the use of doxorubicin-based therapy as well.
All therapies have side effects. T-DM1 has both financial and therapeutic toxicity. In the study, T-DM1 patients had a significantly increased number of severe adverse events compared to the control trastuzumab arm, though the most important of these were reversible and relatively rare thrombocytopenia and transaminitis. But the benefits clearly outweigh the harms for the average patient considering such therapy.
The PHARE trial presented a long-term update on duration of therapy. PHARE tested duration of trastuzumab with a non-inferiority design, and the new 10-year results mirror the previous 6-year results in not rejecting the possibility of inferiority. A similar trial did reject, though the results are similar enough to PHARE and the major difference is the allowed confidence intervals for noninferiority. The T-DM1 results will likely render both trials irrelevant for high-risk populations, at least in resource-rich countries.
Stellar Results From SOLAR-1
PI3 kinase inhibitors have had a rough go of it. Though PI3K is a central node for several growth factor receptor pathways, and though PIK3CA is frequently mutated in both ER-positive and HER2-positive breast cancers, trials of PI3K inhibitors have largely been a bust, with minimal improvement in progression-free survival accompanied by significant toxicity, particularly in the form of hyperglycemia.
This meeting saw the presentation of the SOLAR-1 trial, a phase III randomized controlled trial of the selective PI3K alpha inhibitor alpelisib, comparing alpelisib plus fulvestrant to fulvestrant alone in metastatic ER-positive breast cancer. Though the trial recruited both PIK3CA-mutant and wild-type patients, the presentation focused on the patients with mutated tumors. In patients without prior CDK 4/6 inhibitor therapy, the addition of the study drug improved progression-free survival from 6.8 to 11 months. Overall survival trended positive but is immature for a truly confident analysis.
As with other members of this class of drugs, hyperglycemia represents a significant class effect, and a common management problem for patients and clinicians. I suspect real life, with older and less healthy patients, will offer up even greater hyperglycemia challenges for this agent. Nevertheless, this is the cheeriest trial presentation to date for a PI3K inhibitor, and the drug appears headed for the FDA.
An interesting side experiment in this trial involved the measurement of the mutation using ctDNA. Plasma ctDNA measures of PIK3CA mutations were highly concordant with tumor measures, and correlated well with PFS, suggesting that a simple blood test might allow us to "PIK" patients for this drug. Indeed, ctDNA was superior to tissue biopsy in predicting benefit.
In a related manner, this year's meeting saw a surge in ctDNA and circulating tumor cell presentations. None of these were definitive, though the gestalt from them was that this is a field that is rapidly heating up, both for surveillance for recurrence and progression, as well as prediction of therapeutic benefit for targeted therapy. The genomics field is rapidly progressing, and as prices fall and knowledge accumulates we are edging towards a day where liquid biopsies will guide treatment selection.
Immuno-Oncology for Triple-Negative Breast Cancer
Earlier this year, at the European Society of Medical Oncology meetings, we saw the first positive phase III trial of a checkpoint inhibitor in triple-negative breast cancer. The IMpassion 130 trial was updated (or cloned) for San Antonio, and though not much has changed, the results are worth describing. IMpassion 130 compared nab-paclitaxel alone to the same plus the PD-L1 inhibitor atezolizumab in first-line metastatic breast cancer. The trial enrolled patients with both PD-L1 positive and negative tumors.
Quite simply, there was no benefit seen for PD-L1 negative patients, and a statistically significant progression-free survival benefit for PD-L1 positive patients, from 5.0 months for the placebo arm to 7.5 months for the atezolizumab arm. Other biomarkers, such as CD8 and stromal TILs, added nothing to PD-L1 as a predictor of benefit. Though the presentation suggested an impressive overall survival benefit, with an increase from 15.5 to 25 months, these results are immature and were not supposed to be formally tested per the original hierarchical study design.
One concern arising from the initial data presentations is that the progression-free survival results do not, to date, look like what we would hope for: a true plateau in the PFS curves, such as seen in melanoma. This no doubt reflects the fact that, while we think of TNBC as the "heavily mutated" form of breast cancer, its mutational burden pales in comparison to melanoma, lung cancer, and MSI-high colorectal or endometrial cancer. Just not enough funny epitopes on the cell surface. Nevertheless, this is the first compelling evidence for therapeutic benefit of a checkpoint inhibitor in breast cancer. I cannot wait to see further follow-up.
There were several wonderful presentations looking at the genomics of metastatic breast cancer. The genomic landscape of metastatic disease is now beginning to come into focus as data from hundreds of patients (many involving paired samples) become available. To summarize what this year's meeting suggested, comparison of primaries with metastases shows an increase (though not across the board) in tumor mutational burden and in specific mutations that represent potential therapeutic targets. The driver mutations identified represent a tractable number from a therapeutic standpoint, not dozens or hundreds, which is reassuring for drug developers. One presentation by Dr. Angus suggested that roughly a quarter of analyzed metastatic genomes had alterations for which FDA-approved drugs are available.
Within the past year, we saw the FDA approval of broad genomic panels, which also carry the Medicare imprimatur. As mentioned above, we are also beginning to see evidence that ctDNA is applicable to evolving breast cancers. These point to a number of fairly straightforward conclusions: first, that we will be able to measure tumor evolution in more-or-less real time, particularly that Darwinian selection induced by drug therapies; second, that we should be able to determine from such studies what treatments are best suited to keeping the patient's tumor in check.
Evolutionary biologists speak of the "Red Queen Hypothesis," derived from Lewis Carroll's Through the Looking Glass, in which Alice says "Now, here, you see, it takes all the running you can do, to keep in the same place." The Red Queen Hypothesis describes the evolutionary situation in which predator and prey co-evolve. Our "hunted" patients would like to stay several steps ahead of their vicious predator. Increasing knowledge of the genomics of metastasis, linked (for certainly it must be) to well-done therapeutic trials, should allow us to co-evolve, and perhaps get a bit ahead, of the pursuing beasts.
My Stanford colleague Christina Curtis, PhD, presented an updated analysis from the METABRIC consortium. METABRIC, published several years ago in Nature, used genomics to divide breast cancers into several integrative subtypes with prognostic significance, but never found its way into routine clinical use. At this year's meeting, Curtis and colleagues presented a detailed, clinically well-annotated update focused on several ER-positive integrative subtypes with increased risk of distant metastasis. These represent a substantial portion (roughly a quarter) of ER-positive patients and are associated with specific oncogenic drivers that could well serve as therapeutic targets, such as the fibroblast growth factor receptor and S6K1. Trials are in development to evaluate whether these findings will translate to clinical practice.
On a sadder note, we learned at San Antonio of the death of Charles Coltman, MD. Chuck led the clinical program where I trained in San Antonio, and had a distinguished career as a clinical trialist, primarily in the lymphoma field. A somewhat gruff, hard-charging ex-military man (we called him "COL. T Man, though never to his face), Chuck was a passionate advocate for clinical trials, serving for many years as head of the Southwest Oncology Group. He also co-chaired the San Antonio Breast Cancer Symposium for many years, first with Bill McGuire and then with Kent Osborne. I learned a lot about being an oncologist from him, and I mourn the passing of one of our field's giants.