GRAIL’s approaches to overcome these obstacles include using high-intensity sequencing of unprecedented breadth, depth and scale to detect cell-free nuclei acid; conducting one of the largest clinical trial programs ever to generate scientifically rigorous evidence of clinical utility; using rigorous data science to classify patients by presence, type and severity of cancer; developing classifiers to determine cancer status and type and collaborating with leading community and academic medical centers.
Dr. Aravanis then briefly introduced GRAIL’s assay development pipeline and noted that the company’s early detection test for nasopharyngeal carcinoma is going to launch in 2018.
Ms. Victoria Raymond, director of Medical Affairs at Guardant Health followed Dr. Aravanis with her introduction of Guardant Health’s approach for liquid biopsy. Guardant Health was founded in 2012 and its slogan is to “Conquer Cancer with Data”. In contrast to GRAIL, which focuses on patients with early stage cancer, Guardant Health’s product line starts from the advanced cancer patients.
Its Guardant 360 test can help therapy selection for the 1.65 million patients who are already diagnosed with cancer. Therapy selection is very important for patients’ clinical outcome as targeted therapy based on genetic biomarkers can greatly improve patients’ prognosis. Liquid biopsy can be used to repeatedly measure mutations in tumors and guide therapy selection.
Using the data generated from the late stage cancer patients, Guardant is developing its LUNAR assays to detect early stage cancer or monitor relapse from residual cancer cells. It’s LUNAR-I assay is aimed at 13.8 million cancer survivors to help monitoring their relapse. On the other hand, the LUNAR-II assay is aimed at 50 million high risk individuals to help detecting early stage cancers. The key for developing these assay is to generate large amount of data to validate any assays in development. Guardant is working on collecting and accumulating data to design and validate their molecular diagnostic assays.
Dr. Wendy Winckler, executive director of Novartis, showcased the potential of liquid biopsy tests to guide personalized medicine. Since tumors continue to evolve through different therapies, they need to be retested after each relapse and liquid biopsy tests can help guide personalized therapy selection for each patient based on the molecular signatures of their tumors. Novartis has developed a NGDx cfDNA panCancer assay which has a 600-gene panel.
Dr. Winckler also noted that circulating tumor DNA (ctDNA) can be used to monitor tumor response and resistance to a given therapy because ctDNA is quantitative index of tumor cells. It will decrease as tumors shrink and increase when they acquire resistance and grow again.
She also added that beyond identifying point mutations, measuring other molecular signatures may also aid cancer treatment. For example, mutation burden has been shown to be related to immunotherapy response. Detecting non-human sequences may shed light on the viral origin of certain tumors.
The future directions to detect and characterize cancer will involve using multi-omics non-invasive tests that include analyses of total nuclei acid, exosomes, circulating tumor cells, peripheral immune cell profiling, proteins and metabolites.
The liquid biopsy panel was moderated by Dr. Hatim Husain from Moores Cancer Center. The panelists include Dr. Aravanis from GRAIL, Dr. Lyle Arnold, CSO of Biocept; Dr. Dongliang Ge, CEO of Apostle; Dr. Shidong Jia, CEO of Predicine and Dr. Traci Pawlowski, senior director of product development, oncology at Illumina. The panelists shared with the audience on the expectations, goals and challenges of liquid biopsy. On the challenges for adoption of liquid biopsy, the panelists noted that the standardization of the liquid biopsy assays, education for doctors and reimbursement issues are several obstacles that need to overcome. Discordance between liquid biopsy result and the result from cancer tissue biopsy raise concerns for assay quality. However, panelists also stressed that because liquid biopsy is collecting information from tumor cells that are different from localized tumor tissue. Some discordance is expected and also represents the opportunities for liquid biopsy.
Panelists also offered different solutions to overcome the challenge of limited amount of sample and improve assay quality. Dr. Dongliang Ge from Apostle said his company is working on improving the extraction efficiency from blood samples so that more DNA fragments can be captured for sequencing and analysis. Dr. Aravanis noted that specificity, not sensitivity, is key to a successful test. Dr. Jia suggested that bringing RNA signature to complement DNA based detection and strategically use selected bioliquid instead of blood (for example, urine for bladder cancer) may increase assay sensitivity and accuracy.
The afternoon session was chaired by Dr. Howard Pan, co-chair of the FTD organizing committee. What is the impact of deep learning for genomics and medicine? Profession Xiaohui Xie from UC Irvine shared with the audience his first-hand experience. Dr. Xie noted that from google translate to self-driving cars, AI has permeated our daily life. In the image processing field, AI has achieved tremendous success, its error rate has dropped rapidly since 2010. Currently, the accuracy of AI imaging processing has passed the limit for human. In the Genomics field, AI can provide insights for genome wide association studies. Because we have large amount of genomic data, we can use them to train deep learning algorithm to find key functional units in DNA sequences.
Dr. Michael White, VP and CSO of Pfizer Oncology R&D, delivered the keynote speech of the afternoon session. Dr. White noted that one major challenge for treating cancer is the large amount of somatic mutations. One tumor can harbor over 200 mutations. How to address this problem? One potential solution is to categorize different mutations and develop drug candidates that are specific for each subgroup of mutations. Using machine learning, we can find patterns from treating a large number of cancer cell lines, correlate drug potency with types of mutations, and find potential molecules that will work. Then we can explore the biological mechanisms of these molecules. In ideal situation, we can find uncharacterized mutations, or discover synergy from different small molecule drug and develop better therapy.
Dr. Tony Ho from CRISPR Therapeutics introduced how to translate CRISPR technology into clinical development. CRISPR/Cas9 technology is a novel gene editing technology. Dr. Ho showed that in sickle cell disease and b-thalassemia patients, if their fetal hemoglobin level is elevated, these patients will be protected from the diseases. Based on this observation, CRISPR Therapeutics developed a therapy that using the CRISPR/Cas9 technology to upregulate the level of fetal hemoglobin level. Dr. Ho also noted that just like small molecule drug, CRISPR therapies also have off-target effects. Using a proprietary platform, CRISPR Therapeutics first evaluates the safety of guide RNA through computer models, then validates using biological models, finally select one candidate to conduct clinical trials. Dr. Ho also introduced the application of CRISPR technology in making allogeneic CAR-T therapy.
The afternoon panel discussion was led by Dr. Lawrence Lum from Pfizer Oncology R&D. The panelists include Dr. Patrick D. Hsu, principal investigator of the Salk Institute; Dr. Alexis C Komor and Dr. Prashant Mali from UCSD; Dr. Prasun Mishra, CEO of Agility Pharmaceuticals and Dr. Marco Weinberg from Vertex Pharmaceuticals. They share their opinions on the application of novel gene editing technologies in drug discovery and development.
There are several technical challenges for the CRISPR technology: recent papers from Nature Medicine and PNAS showed that the polymorphism in human genes may impact the specificity and off-target effects of CRISPR/Cas9 technology. In addition, human immune system may respond to Cas9 protein and limit its action in vivo.
The panelists suggested that some of these challenges also presents opportunities. For example, polymorphism could increase risk for off-target effects, however, it could also bring potential opportunities for personalized medicine.
For the immune response issue, Dr. Komor and Dr. Hsu indicated that historically, this issue is not new. People already know that if we inject certain enzymes into the body, they will be neutralized by antibodies. The solution is to find more enzymes. Currently, there are new papers on Cas9 every week and introduce homologous proteins for Cas9. These proteins may be the solution. In addition, biochemists are also engineering these proteins to decrease their immunogenicity.
Dr. Malia added that maybe we should focus more on the immune response induced by the viral vectors that encode the Cas9 system. We know that many viruses can cause the immune system to be over active. Dr. Mishra agreed that we should monitor the immune response carefully in early clinical trials because an overactive immune system could potentially cause patient death.
The panelists also agreed that although CRISPR/Cas9 technology is very hot right now, we should not forget other technologies such as TALEN and RNAi. Dr. Hsu said the TALEN protein is smaller than Cas9, it may offer some advantage in the drug delivery process. Dr. Komor noted that the CRISPR technology that can make single base editing is a great choice to make single based mutation. Dr. Feng Zhang’s CRISPR/Cas13 RNA editing system is also a very promising technology.
Dr. Peter Huang, executive vice president of SABPA closed the meeting by thanking the speakers, sponsors and volunteers for making this event a huge success.