1. Research Infrastructure Building.
- BSRR: Since November 2010, the BSRR has collected residual blood samples from 100,314 patients. These individuals are patients who sign the MD Anderson front door consent. We have extended this approach to develop additional banking efforts and research cohorts with minimal additional funds. This tremendous infrastructure provides strong research support allowing Center members to better compete for extramural grant funding. This has been most evident by our success in obtaining five new next generation sequencing (NGS) grants funded for pancreatic, colorectal, head and neck, and ovarian cancer as well as melanoma for a total cost of $20 million. This success is not limited to NGS grants - many other avenues of research including several SPORE projects and moonshot projects are also benefiting from the establishment of this cohort.
- Data and Biospeciemen Bank Building and Management: In addition to the development of our banked blood biospecimens, the center also provides support for the governance of BSRR, PHDB, TexGen data and approved request for use of these data that cover both investigator-driven research projects and infrastructure/clinical projects.
- MD Anderson Cancer Patients and Survivors Cohort (MDA-CPSC): This cohort is led by Dr. Xifeng Wu and supported by the faculty funds. In the era of personalized medicine, it is important to better predict clinical outcomes and guide individualized treatment paradigms while recognizing the outcome for cancer patients differ widely even with similar clinical characteristics. To support well-powered studies of clinical outcomes and survivorship, we have established MDA-CPSC which links the PHDB with the institutional EHR, BSRR, medical testing results in Laboratory medicine, and tumor registry data. The cohort currently enrolled 157,151 newly diagnosed, MD Anderson treated patients. Among these patients, SF12 QoL data was collected from 86,077 participants and 67,280 participants has banked biospecimens.
- Long-term Survivorship Cohort: This cohort is led by Drs. Xifeng Wu and Alma Rodriguez. We have recruited and banked biospecimens from 4,744 survivors. This long-term survivorship cohort will enable research addressing health concerns in this unique and growing population. This would include development of late effects of treatment, second primary tumors, toxicities, reductions in quality of life, symptoms, and chronic diseases. With the increasing number of long-term survivors, the need for focused research in this population is urgently needed.
- Pediatric Cancer Cohort: This cohort is being led and developed by Dr. Michelle Hildebrandt with support from her start-up funds. The Pediatric Cancer Cohort is a division-wide collaboration with Pediatrics to recruit all new patients coming to the Pediatrics Clinic as well as healthy sibling transplant donors.
2. Blood-based Biomarker Discovery, Development, and Validation.
- The CTPHG strives to be on the forefront of advances that allow for identification of novel genetic and phenotypic biomarkers for cancer. Many cutting-edge “-omic” platforms are available in the laboratory including the Fluidigm Biomark HD system, Life Technologies Ion Torrent sequencer, GE In Cell Analyzer HCA system, Beckman Biomeck FXp Laboratory Automation Workstation, Fluxion’s Isoflux apparatus for isolating circulating tumor cells (CTCs). Newly added this year is the Covaris high-throughput sonicator to support multiple next generation sequencing (NGS) projects. In addition, improved NGS sample preparation procedures are established in the laboratory to reduce costs, increase throughput, and minimize lab errors for these large-scale NGS studies. Together, these platforms and procedures enable CTPHG to conduct large scale population-based NGS studies and facilicate the funded NIH R01 NGS grants.
- The center is at forefront in designing and conducting candidate gene, pathway-based, and genome-wide association genotyping studies. The genotyping studies included the application of iSelect and OncoArray Beadchips on existing Illumina platform. The CTPHG has developed a customized global SNP panel with comprehensive collection of genetic variants that are known or predicted to affect miRNA binding sites which enables investigator to study miRNA-related variants more extensively. This chip has been applied to many cancer sites including bladder, lung, renal cell carcinoma to support the NIH funded SPORE and R01 projects.
- The center developed and established a number of assays to advance the discovery of non-invasive markers in circulation. These biomarkers could be used in risk stratification and may have translational potential. One of the major efforts is focused on circulating microRNA profiles in serum, urine, and exosome. As an example, we identified a panel of circulating serum microRNAs predictive of early stage lung cancer recurrence and survival using a multistage study design with a discovery and validation populations. Follow-up functional characterization of these miRNAs in primary tumor and cell lines together with extensive bioinformatics analysis found miR-150 has oncogenic effect by direct targeting SRCIN1signaling pathway which led to the identification of SRCIN1 as a potential therapeutic target. Other examples for the application of these assays include detection of KRAS mutations in exosome DNA as well as circulating DNA and showed the association of cfDNA KRAS mutation load with survival in lung cancer patients, identification of global DNA methylation with the risk of renal cell carcinoma. A new development is the immune signatures in peripheral blood mononuclear cell. Discovery, development, and validation are linked - all biomarkers undergo extensive testing at each phase to increase the likelihood of clinical application in the future. Together, these developments significantly expand our capabilities in biomarker discovery, development, and validation.
- Center resources have been actively utilized and integrated with several Moonshot programs including prostate, lung and colorectal cancers to identify biomarkers and develop prediction models to enable better risk stratification at diagnosis, personalized medicine, and better prediction of malignant progression.
3. Personalized Risk Prediction Models & Computational Epidemiology.
- Lung cancer risk model: The Centers for Medicare and Medicaid (CMS) recently approved Medicare coverage for screening for individuals at high risk for lung cancer with low dose computed tomography based on age of 55-77 and at least 30 pack years smoking history. We used a prospective cohort of 395,875 participants enrolled in MJ Cohort with a median follow-up of 7.3 years to develop integrative risk prediction models for lung cancer in heavy, light, and never smokers by incorporating epidemiological, laboratory test, and medical evaluation data. We have developed markedly improved risk models identifying light and never smokers, who account for the majority of lung cancer cases, at highest risk of developing lung cancer within this large Asian cohort. The risk model for heavy smokers can allow us to stratify heavy smokers into subgroups with distinct risks of developing lung cancer, which, if applied to LDCT screening, may greatly reduce false positives.
- Lung cancer risk model validation: Although many lung cancer risk prediction models have been published, these models used different set of risk factors and was developed using distinct study populations. A comprehensive review and extensive assessment of these published models is lacking. We collaborated with the international lung cancer consortium and analyzed data from a total of 466,016 subjects including two case control studies from the American Cancer Society (ACS) and MD Anderson Cancer Center (MDACC) Lung Study, and three cohort studies from the Nurse Health Study (NHS), Multi-Ethnic Cohort (MEC), and Prostate, Lung and Ovarian Cancer Screening Trial (PLCO) to validate and extend the lung cancer risk models.
- Liver cancer risk model: We leveraged a large prospective cohort of over 420,000 individuals with a median follow up of 8.5 years to develop a risk prediction model for the general population with average or unknown risk of liver cancer. Integration of epidemiologic and limited clinical data resulted in an AUC of 0.93. In collaboration with Dr. Lidong Qin from Methodist Research Institute, we have received funding to translate this model into the clinic by creating a portable point-of-care diagnostics V-Chip.
- Prediction of radiation-induced toxicity in lung cancer: Pneumonitis and esophagitis are two major dose-limiting acute toxicities for lung cancer patients receiving primary radiation therapy, which negatively impact patients’ quality of life. Current clinical-based prediction approaches are insufficient to predict the occurrence of these toxicities prior to treatment. We developed a risk model that integrated a polygenic risk score from key SNPs located in inflammatory-related pathway genes to predict acute radiation toxicity in patients with lung cancer with an AUC of 97%. Further research identified that the most promising genetic markers were also correlated with response to radiation and gene expression profiles in a lymphoblastoid cell line model system.
- Prediction of second primary tumor in head and neck patients: Chemoprevention clinical trials to prevent second primary tumors or recurrence following curative treatment in head and neck patients tested 13-cis-retinoic acid (13-cRA) with inconclusive results. In this study, we identified several promising biomarkers that can predict the risk of second primary tumor/recurrence with AUC of 0.79, the population of patients who are predicted to receive the most benefit from retinoid chemoprevention. This could be used in the future to select candidates for 13-cRA chemoprevention trial.
- Prediction of incident diabetes in Mexican Americans: In addition to prediction models for cancer risk and clinical outcomes, we further developed risk prediction model for chronic disease. Mexican Americans have high prevalence of diabetes compared to non-hispanic whites. In this study, we integrated epidemiological, exposure, and lifestyle factors to develop an improved risk prediction model for incident diabetes using a large prospective cohort of 8,566 participants.
4. Building Research Networks.
- The center has expanded the support and maintenance to a total of 6 programs across the institution in the Collaborative Biospecimen Banks. These biobanks have collected and banked a total of over 5,000 samples for the Tobacco Treatment Program (TTP, 1,830 samples), Project Health (Low SES; 614 samples), Community Network Program (167 samples), Project STEPS (192 samples), and Project CHURCH (1,260 samples), and Project LBJ (1,285 samples). This TTP resource has been leveraged in the Lung Cancer Moonshot. We continue the effort in scientific and public health education to foster collaborations and establish expanded research networks. These activities included the Distinguished Seminar Series which invited 31 high-profile scientists in the fields of translational science, epidemiology, and public health and a One-Day Symposium with 147 attendees from 25 different departments/offices. Working with the Global Academic Programs, we have successfully recruited two post-docs from Mexico through the International Training Program with the goal to promote global collaboration and help build resources in low- to middle-income countries.