Someday, we may find that we can design treatments based on sequencing, but we are not there yet. I don’t believe tumor genome sequencing is ready for prime time.
—Hope S. Rugo, MD
The role of next-generation sequencing (high-throughput technologies that allow DNA and RNA to be analyzed more quickly and inexpensively than earlier techniques) in breast cancer remains unclear and at present is primarily a research tool. Therefore, clinicians should be cautious in using genetic profiling to guide treatment, according to Hope S. Rugo, MD, Professor of Medicine and Director of Breast Oncology and Clinical Trials Education at the Helen Diller Family Comprehensive Cancer Center, San Francisco.
For the majority of patients, she emphasized, targeting specific markers is of uncertain clinical value. “Individual case studies are encouraging, but they are few and far between,” she noted.
“What we are asking from tumor genome profiling is whether we can look at specific genomic aberrations and find a treatment that is more likely to benefit the patient vs something else. Someday, we may find that we can design treatments based on sequencing, but we are not there yet. I don’t believe tumor genome sequencing is ready for prime time,” Dr. Rugo told attendees at the 32nd Annual Miami Breast Cancer Conference.
Growth of Next-Generation Sequencing
The most common approaches to genomic profiling today are whole-genome sequencing, which determines the complete DNA sequence of an organism’s genome at a single time, and whole-exome sequencing, which selectively sequences only the coding areas of the genome.
Falling prices are fueling the growth of sequencing services. Revenue from these services (for all health care, not just cancer) totaled $160 million in 2010 and should increase to $550 million by 2016.
“It’s now cheaper to do it, it can be done quickly, and it can be done in many locations,” Dr. Rugo said. “As technology improves, it’s becoming more available and therefore more of an issue for clinicians.”
Limited Candidate Alterations
The list of candidate “actionable” genomic alterations in breast cancer numbers nearly 20. These aberrations are theoretically targetable by specific agents or those in development, but the reality is that most occur in fewer than 10% of patients—and most affect even less.
Additionally, even with more common mutations, a direct link between mutant status and response to a targeted agent “remains to be seen” for most so-called actionable mutations, she added.
The rare frequency of these mutations makes for huge challenges in accrual to clinical trials in metastatic breast cancer, she pointed out, and this begs the question, Why do screening? “It is unclear whether this will be a rational approach in the future,” she noted.
Prognostic Power but Not Predictive Ability
Although theoretically attractive, using mutations to guide treatment selection has not been successful in the clinical trial setting. A prime example comes from the BOLERO-2 trial, which evaluated the mTOR inhibitor everolimus (Afinitor) given with exemestane in patients refractory to endocrine treatment.
The combination yielded a progression-free survival benefit, but no correlation was found between the presence of activating mutations (PIK3CA/PTEN/CCND1 or FGR1/2) and response to treatment. Patients with multiple different mutations had somewhat better outcomes than those without mutations (hazard ratio = 0.78), but the numbers in these subsets were small.1
“We know if your cancer is very disordered, it’s bad-acting, so this was not surprising,” she commented.
Similarly, in the CLEOPATRA study, which evaluated the addition of pertuzumab (Perjeta) to trastuzumab (Herceptin) in HER2-positive disease, biomarker analysis revealed no relative benefit according to PI3KCA mutational status.2 These trials only confirm “what we have known from many studies,” which is that mutations have prognostic power but not predictive ability. The addition of a novel treatment improves outcomes, but patients with a genotypically poor prognosis do not benefit more than other patients.
Numerous clinical trials are now screening for mutations to select patients for experimental drugs, but, for the individual oncologist or patient, this is not yielding abundant fruit, she indicated.
“For instance, in studies of the fibroblast growth factor receptor (FGFR) and the vascular endothelial growth factor inhibitor lucitanib, our center has screened a number of patients and has found only one with the FGFR mutation,” she reported. Furthermore, responses observed in patients with mutations actually do not prove selective efficacy of a drug. “You don’t know how patients without the mutation would do,” she noted. “In BOLERO-2, they did as well.”
An examination of genomic alterations in 1,445 invasive breast cancers by Foundation One paints a clear picture of the challenge. It showed TP53 to be the most common mutation, occurring in 60% of samples, followed by PIK3CA mutations in 30%.
“The problem is that we don’t have targeted agents for these mutations,” she pointed out. “This finding may spur further research, but it doesn’t help us manage our patients today.”
‘N of 1’ Driving Personalized Therapy
Most of the data guiding personalized therapy come from “N of 1 case reports,” such as that of the 53-year-old patient with progressive HER2-positive breast cancer, found on genomic profiling to have activating ERBB2 and epidermal growth factor receptor (EGFR) mutations. She was treated with trastuzumab and metronomic chemotherapy, then erlotinib (stopping lapatinib [Tykerb]). She was still responding at 4 months.
Dr. Rugo also described a patient with triple-negative disease who entered a trial of erlotinib plus bevacizumab (Avastin) 12 years ago. Hers was the only response among 42 patients, and it continued for 11 years, until she stopped treatment for reasons other than disease progression. Interestingly, no EGFR mutation was identified in the patient’s tumor.
Although “encouraging,” cases such as this, she said, “do not tell us what to do with the patient sitting across the table from us with triple-negative breast cancer.”
Unfortunately, for most patients whose tumors are sequenced, she added, a number of genomic alterations may be identified, but none is actionable.
‘An Ethical Dilemma’
At this point, Dr. Rugo concluded, the use of next-generation sequencing creates “an ethical dilemma” by adding toxicity and cost without a clear, predictable benefit from the chosen treatment. “To get somewhere, we will need the ability to look at panels of genes correlated to clinical phenotypes and outcomes,” she suggested.
Completed studies and ongoing or planned trials are providing information, but in the meantime, she said, “Treatment with unapproved drugs should be considered only in the setting of clinical trials, and screening should only be used to determine eligibility for marker-driven studies.” ■
Disclosure: Dr. Rugo reported no potential conflicts of interest.
1. Hortobagyi GN, Piccart-Gebhart MJ, Rugo HS, et al: Correlation of
molecular alterations with efficacy of everolimus in hormone receptor-positive, HER2-negative advanced breast cancer: Results from BOLERO-2. 2013 ASCO Annual Meeting. Abstract LBA509.
2. Baselga J, Cortés J, Im S-A, et al: Biomarker analyses in CLEOPATRA: A phase III, placebo-controlled study of pertuzumab in HER2-positive, first-line metastatic breast cancer. 2012 San Antonio Breast Cancer Symposium. Abstract S5-1. Presented December 7, 2012.