Originally published March 25.
Researchers from 18 institutions in France tested their ability to benefit metastatic breast cancer patients with personalized treatment strategies in a study, and in the process found that scientific and logistical challenges enabled very few participants to receive precision medicine and that even fewer actually benefitted from the interventions they received.
The study, published in Lancet Oncology and involving hundreds of patients, highlights the fact that even with the availability molecular diagnostic tools that can rapidly screen and identify genomic markers involved in driving cancer, the delivery and therapeutic impact of the recommended individualized approaches are often hindered by the way clinical trials are conducted and by limited scientific understanding of the complex molecular tumor environment.
Researchers led by Fabrice Andre, director of the INSERM Unit U981, a lab focused on developing personalized treatments, recruited more than 400 patients with metastatic breast cancer, obtained their biopsy samples, and used CGH array and Sanger sequencing to analyze the molecular characteristics of 67 percent and 70 percent of the patients' tumors, respectively. Surprisingly, recruiting patients in to the study was no problem. Andre and his group had anticipated that it would take them three years to enroll more than 400 metastatic breast cancer patients into the trial. It took them only 13 months. "This … shows the high level of expectation from patients and doctors about personalized medicine," they wrote in the paper.
Using CGH to pick up on DNA copy number changes that could be driving the metastases and Sanger sequencing to home in on AKT1 and PIK3CA genes, Andre and his colleagues picked out a genomic alteration in 195 patients, or 46 percent, that they could conceivably target with therapy. The most common markers they identified in patients were in PIK3CA, CCND1, and FGFR1 genes, while more uncommon alterations showed up in AKT1, EGFR, MDM2, FGFR2, AKT2, IGF1R, and MET.
Despite identifying these markers associated with patients' disease, researchers were able to recommend personalized strategies for only 55 patients, or 13 percent of 423 participants. Only 43 women could actually be assessed and receive the proposed treatments, and more than half of this subset received treatments as part of Phase I or Phase II trials. Out of these 43 patients, four saw an objective response and nine had stable disease for more than four months. Only 1 percent of enrolled patients experienced a serious adverse event as a result of being biopsied.
The study did not reach its primary endpoint of matching 30 percent of molecularly tested patients to targeted treatments based on their oncogenic characteristics. Andre and his team proposed precision treatment options for 28 percent of the 195 patients who had targetable genomic alterations. "This finding emphasizes the need to increase access to drugs for patients shown to have a genomic alteration on screening," Andre and colleagues wrote. Furthermore, they acknowledged that the low number of patients who had an objective response to the recommended precision strategy "seems disappointing."
"Progress in cancer genomics has raised hopes of increased precision in the identification of patients suitable for targeted therapies tailored to their genotypes," Charles Swanton of Cancer Research UK wrote in an accompanying editorial in Lancet Oncology. The analysis by Andre and colleagues "provides important insights into the logistical, scientific, and clinical challenges of implementing national cancer genomic assessment," Swanton added.
The French National Cancer Institute has implemented a program in which it tests melanoma, lung, and colorectal cancer patients routinely for eight biomarkers associated with these diseases, in an effort to predict their ability to respond to various therapies. Only with a better understanding of the key markers and pathways implicated in metastatic breast cancer will a similar nationally supported personalized approach be possible for this patient subset.
Although genomic understanding of early-stage breast cancer has contributed to improving disease prognosis and treatment, metastatic disease still remains the leading cause of cancer-related deaths globally, with nearly 40,000 women succumbing to the illness each year. In the last decade, Andre and his group note that only eribulin has shown to improve survival among the available chemotherapeutics or endocrine treatments for metastatic breast cancer patients. When it comes to personalized treatment approaches, therapies targeting the 15 percent to 30 percent of breast cancer patients with HER2 overexpression (i.e. Herceptin, Perjeta, etc.) have shown the most promise in terms of impacting outcomes.
"Besides HER2 amplification, breast cancers include many genomic alterations located in oncogenes or tumor suppressor genes," Andre and colleagues wrote in the paper. "Each specific oncogenic event could potentially be blocked by a targeted therapy."
Given the low objective responses seen in the latest trial, the researchers suggest a need to improve algorithms that interpret genomic data and identify the most relevant pathways that should be targeted with therapies, as well as a need for "highly bioactive" drugs against the identified molecular targets. At a personalized medicine conference in January, Eric Lai, head of pharmacogenomics at Takeda Pharmaceuticals, explained that despite technological advances enabling cheaper and more in depth sequencing of patients, the field of personalized medicine is being hindered by the fact that there aren't enough agents that target the molecular underpinning of diseases. "Even if we have the technology to do personalized medicine," Lai said, "we don't have enough drugs to give you personalized medicine. I don't care how good your technology is."
Similarly, as Andre and colleagues note in their paper, one patient's tumor can have many oncogenic drivers, which in turn, don't occur that often across all those with metastatic disease. This makes it a challenge for companies to develop drugs for such rare pharmacogenomic patient subgroups. "Each [oncogenic marker] may occur in only a few cases, which makes the investigation of each driver in a dedicated study almost impossible," they wrote. "Alternative strategies are needed, therefore, to investigate which targets deserve further investigation."
Researchers are employing novel study designs in an effort to mitigate this issue. For example, adaptive design trials, such as the I-SPY studies, offer an opportunity for drug makers to quickly gain insights into molecularly-defined patient subgroups for their therapies. For example, in the I-SPY2 study, early stage breast cancer patients with certain molecular characteristics are placed on one of seven experimental arms and depending on their response to a particular investigational drug either they stay on the arm or is placed in a different arm.
At the San Antonio Breast Cancer Symposium, researchers from the I-SPY2 trial presented data from more than 100 HER2-negative patients showing that if drug developer AbbVie performs a Phase III trial investigating its PARP inhibitor veliparib in combination with carboplatin and standard neoadjuvant chemotherapy in triple-negative patients, it would have a 90 percent chance of success.
Meanwhile, the state of the science in metastatic breast cancer was a major hindrance to advancing personalized treatment approaches in Andre et al.'s investigation. While there has been a lot of work in the life sciences community to molecularly characterize early stage-breast cancer patients, leading to prognostic and predictive tests such as Oncotype DX, little data is available for patients whose disease has spread beyond the primary tumor.
In the latest study, the researchers looked at the genomic alterations that appeared to be implicated in metastatic disease, but they did not find any markers disproportionately overexpressed or particularly unique to the setting compared to early stage patients. However, some patients whose primary tumors were HER2-negative were found to have HER2-amplified disease once their breast cancer metastasized, demonstrating the molecular evolution that cancer tumors undergo as they progress. The researchers noted that next-generation sequencing approaches might have identified additional markers of interest.
A major scientific challenge highlighted by Andre and his team, as well as Swanton, was the confounding effect of intra-tumor heterogeneity – when the same tumor has different disease-associated molecular alterations at different locations – on precision medicine approaches. This study "provides a stark reminder that understanding of biological drivers of metastatic disease progression is basic," Swanton wrote in his editorial. "Defining drug-targetable drivers in the context of intra-tumor heterogeneity will require greater knowledge to distinguish somatic events occurring early in tumor evolution … which are present at every metastatic site of disease, from later events that occur in some subclones but not others."
In addition to the scientific challenges, Andre and his colleagues also came up against logistical difficulties. When patients had multiple genomic alterations, the review committee favored targeting those markers for which drugs were being studied in Phase I, or if available, drugs that targeted the most amplified gene, or both. It is worth noting that 17 patients in whom researchers identified treatable alterations and recommended for Phase I trials were not enrolled in the studies. This was at least partly because of their metastatic disease burden, observed Swanton in his editorial.
Additionally, the low clinical responses could also be due to the fact that 63 of the 195 patients with targetable genomic markers had already received multiple lines of therapy, and response rates in heavily pretreated breast cancer patients are usually low.
Ultimately, if a personalized medicine approach with the help of a molecular screening strategy were to be attempted in a real-world setting, the outcome could be even more disappointing, Andre and his group state. "The population of [this study] … is not representative of the overall population of patients with metastatic breast cancer," they noted. "We recruited only patients with good performance status and metastases accessible for biopsy. The success rate of a personalized approach in the context of daily practice, therefore, might be lower than that reported."