Skip to main content
Premium Trial:

Request an Annual Quote

Genentech, 23andMe Seek Participants for Study on Consumer-Driven GWAS Model


This article has been updated to clarify the enrollment challenges with the InVite Study.

Genentech is conducting a pilot project to assess the feasibility of using 23andMe's direct-to-consumer genomics service to collect patient-reported genetic data and outcomes information for genome-wide association studies.

The clinical trial, called InVite, is currently recruiting metastatic breast cancer patients treated with Genentech's Avastin (bevacizumab). The partners hope to enroll 1,000 patients and recruitment is taking place through 23andMe's online customer base as well as referrals from patient advocacy organizations.

According to Genentech, interim data from the trial suggests that this non-traditional research model may be more effective in recruiting patients for clinical trials than are traditional paradigms.

"At this point we are actively recruiting [patients] and … we've had good uptake in it," Genentech spokesperson Holli Dickson told PGx Reporter this week. "To date, we've seen a more rapid uptake of enrollment than we would see in a traditional trial."

In March, 23andMe and Genentech launched the InVite Study, an observational, web-based prospective trial in which researchers will collect self-reported, treatment-related outcomes data and genetic data from metastatic breast cancer patients who received Avastin between Jan.1, 2010, and Dec. 31, 2011, or are currently on Avastin. Women interested in participating in the study can enroll through a website dedicated to the InVite Study.

The 1,000-patient enrollment goal is ambitious for InVite, since the US Food and Drug Administration revoked Avastin's metastatic breast cancer indication last year after studies found that the drug didn't extend survival in patients compared to standard treatments. This shrinks the pool of metastatic breast cancer patients who can join the trial, because only women who received Avastin before Jan. 1 can participate. In an effort to encourage enrollment among patients previously treated with Avastin, 23andMe in the next few days is planning to post stories from patients and advocates on its website in order to encourage enrollment.

23andMe is also counting on a number of breast cancer-focused patient groups, such as the Avon Foundation for Women, Facing Our Risk for Cancer Empowered, and the Triple Negative Breast Cancer Foundation, to refer patients to the trial. According to Dickson, the majority of patients recruited for InVite so far have come through advocacy organizations that have referred patients to 23andMe's trial enrollment site. During a public hearing last year to discuss FDA's decision to revoke Avastin's breast cancer indication, patient advocacy groups such as FORCE had urged for more research to identify which breast cancer patients respond to the drug with the help of biomarkers (PGx Reporter 6/29/2011).

"The goal of this study is really just to understand if this research model works, and if women on Avastin or patients in general are willing to … provide saliva samples and potentially blood specimens independently from a formal clinical trial setting, if they're willing to go online and answer these questions and be actively involved in research," Dickson said.

23andMe's enrollment website for InVite instructs participants to order its DNA collection kit, to "spit" a sample, and to take surveys. As an added incentive for participation, those who enroll in InVite will receive free access to 23andMe's Personal Genome Service, which reports on genes associated with 242 diseases and conditions, including genetic associations linked with carrier status, disease risk, drug response, and physical traits. Outside the scope of this study, 23andMe charges customers $299 for its Personal Genome Service.

After providing informed consent, InVite participants will fill out baseline information about their breast cancer history through an online survey. The investigators will contact participants via e-mail to fill out additional questionnaires for up to one year. In the meantime, 23andMe will analyze patient saliva samples for genes associated with breast cancer and response to Avastin.

Although researchers are primarily trying to gauge whether patients will use 23andMe's consumer genomics services to participate in a clinical trial and provide useful data for analysis, they also hope to identify gene markers associated with Avastin response. According to Dickson, any response markers discovered through the InVite study will be solely for the use of hypothesis generation, and will have to be further validated by Genentech and parent company Roche.

"While [InVite] … may generate important scientific findings, participants should not expect to personally benefit from it, as the findings will not change how they are or have been treated for their disease," 23andMe cautioned in a blog post this week urging metastatic breast cancer patients to participate in the study.

PGx Trials and Tribulations

Researchers in industry and academia often decry the difficulties of meeting target recruitment numbers when conducting prospectively designed pharmacogenetic investigations of the efficacy or safety of personalized drugs. Often the molecular markers associated with a disease or drug response are already rare in the overall population, meaning that large sample sizes are required in order to gather data of statistical significance. Additional study recruitment criteria, such as patients' clinical factors and prior treatments, further reduce the number of eligible participants. As a result, recruiting patients and genetically testing them in real time for prospectively designed genome-wide association studies becomes a costly and time-consuming process.

Given these limitations, drug developers and other researchers usually resort to retrospective analyses or meta-analyses of previous studies in order to identify statistically significant gene associations. Such study designs, of course, are largely considered by the scientific community to produce lower grade evidence than the gold standard: the prospective, randomized-controlled trial.

However, because the unique challenges associated with pharmacogenomic trials often make it unfeasible or unethical to conduct prospective, randomized-controlled investigations, regulatory agencies have been flexible in terms of the types of data sponsors can submit to gain approval for a personalized medicine product.

For example, the FDA asked the sponsors of the colorectal cancer treatments Vectibix and Erbitux — marketed by Amgen and Bristol-Myers Squibb/Merck/Lilly, respectively — to submit data from retrospective analyses to establish whether patients with KRAS mutations are poor responders to these drugs. Furthermore, Qiagen recently received FDA approval for its KRAS companion test for Erbitux. This approval was based on retrospective analysis of tumor samples from the clinical trial to support Erbitux's initial approval in 2004 (PGx Reporter 7/11/2012).

Through the InVite study, Genentech is evaluating whether the direct-to-consumer model is a more effective way to incorporate genetic information into clinical trials.

With the majority of genetic testing, "right now, you have the physician in between the patient, and not all information is collected," Dickson said. "So the thought that we can hear directly from the patients about their experiences could potentially provide a much richer body of information in addition to all of the other data that's collected from traditional research."

When 23andMe launched its online genomic testing service in 2007, one of the company's main goals was to change the way clinical trials were conducted. The company believes that a web-based, consumer-driven research model can allow researchers to conduct trials more efficiently and cost effectively than they are done today.

23andMe claims that its online model provides certain advantages over traditional research methodologies, such as enabling seamless integration of genetic data into clinical trial results to study differences in patient outcomes; eliminating the need for physical research locations, resulting in reduced costs and better accessibility to potential study participants; and enhancing patient engagement in the research. Furthermore, "returning 23andMe results to individuals provides an element of partnership with the patient, making patients more vested in the research and its outcomes," a 23andMe spokesperson highlighted.

23andMe markets its Personal Genome Service directly to consumers. Customers of the service can get genetically tested without the need for a doctor's prescription and receive "health reports" on hundreds of diseases and conditions, including information on genes associated with drug response. 23andMe currently has more than 150,000 customers, who, as part of the company's online community, can participate in research surveys through which the company collects qualitative data about various traits and conditions.

This growing database of phenotypic data, in combination with the genotypic data that 23andMe already has on its customers, enables the firm to do studies to uncover new gene associations or gather more data on previously reported associations. For example, in April, 23andMe reported data in PLoS ONE on five genes significantly associated with hypothyroidism that it identified through the web-based GWAS model.

CEO Anne Wojcicki noted in a statement announcing the discovery of the hypothyroidism gene associations that nearly 90 percent of 23andMe's customers participate in online research, "making crowd-sourced science a reality."

The company has published the results from several other web-based GWAS, including a proof-of-concept trial published in 2010 in PLoS Genetics that reported gene associations for unusual traits such as asparagus anosmia and photic sneeze reflex and replicated associations for other common genetic traits; another PLoS Genetics study published in June 2011 on two genetic associations for Parkinson's disease; and a major replication study of over 180 genetic associations published in August 2011 in PLoS ONE.

InVite's Unique Challenges

Despite the efficiencies promised by 23andMe's online community and testing services, the company acknowledges that there are still barriers to recruiting patients for the InVite study. "Due to the rare and aggressive nature of metastatic breast cancer, it can be challenging to find individuals who have the disease and are physically able to participate for the full duration of the study," a 23andMe spokesperson told PGx Reporter. Additionally, because the FDA removed the breast cancer indication from Avastin's label last year, "it can be challenging to find individuals who used the treatment during the designated time frames to be eligible for the study."

The FDA pulled the metastatic breast cancer indication for Avastin last November due to its limited efficacy in prolonging patients' lives (PGx Reporter 11/30/2011). During a hearing to publicly discuss the available data on Avastin's efficacy and safety in advanced breast cancer patients, Genentech had proposed keeping the treatment on the market with the help of a companion molecular diagnostic to identify best responders. However, based on the data Genentech submitted at the time, the FDA was not convinced that Genentech had any promising molecular markers upon which it could deploy such a strategy in a reasonable time frame.

The InVite study and other biomarker investigations ongoing at Genentech suggest that the loss of the breast cancer indication for Avastin – a setback that amounts to lost revenue of around $1 billion for the company – hasn't dissuaded the drug developer from continuing to search for biomarkers that could characterize which patients will respond to the drug.

For example, Genentech is planning to initiate the MERiDiAN trial later this year. In this Phase III trial, researchers will evaluate Avastin in combination with paclitaxel in people with HER2-negative metastatic breast cancer and will also evaluate VEGF-A as a potential predictive biomarker for efficacy with Avastin. Roche is developing a VEGF-A assay that researchers will use to pick out patients expressing this marker.