Sara E. Patterson

The clinical trial landscape in oncology and connectivity of somatic mutational profiles to targeted therapies.

BackgroundPrecision medicine in oncology relies on rapid associations between patient-specific variations and targeted therapeutic efficacy. Due to the advancement of genomic analysis, a vast literature characterizing cancer-associated molecular aberrations and relative therapeutic relevance has been published. However, data are not uniformly reported or readily available, and accessing relevant information in a clinically acceptable time-frame is a daunting proposition, hampering connections between patients and appropriate therapeutic options. One important therapeutic avenue for oncology patients is through clinical trials. Accordingly, a global view into the availability of targeted clinical trials would provide insight into strengths and weaknesses and potentially enable research focus. However, data regarding the landscape of clinical trials in oncology is not readily available, and as a result, a comprehensive understanding of clinical trial availability is difficult.ResultsTo support clinical decision-making, we have developed a data loader and mapper that connects sequence information from oncology patients to data stored in an in-house database, the JAX Clinical Knowledgebase (JAX-CKB), which can be queried readily to access comprehensive data for clinical reporting via customized reporting queries. JAX-CKB functions as a repository to house expertly curated clinically relevant data surrounding our 358-gene panel, the JAX Cancer Treatment Profile (JAX CTP), and supports annotation of functional significance of molecular variants. Through queries of data housed in JAX-CKB, we have analyzed the landscape of clinical trials relevant to our 358-gene targeted sequencing panel to evaluate strengths and weaknesses in current molecular targeting in oncology. Through this analysis, we have identified patient indications, molecular aberrations, and targeted therapy classes that have strong or weak representation in clinical trials.ConclusionsHere, we describe the development and disseminate system methods for associating patient genomic sequence data with clinically relevant information, facilitating interpretation and providing a mechanism for informing therapeutic decision-making. Additionally, through customized queries, we have the capability to rapidly analyze the landscape of targeted therapies in clinical trials, enabling a unique view into current therapeutic availability in oncology.

mTOR Inhibitors in Castration-Resistant Prostate Cancer: A Systematic Review

AbstractBackgroundThe progression of prostate cancer to castration-resistant prostate cancer (CRPC) is often a result of somatic alterations in the PI3K/Akt/mTOR (mammalian target of rapamycin) pathway, suggesting that therapies targeting this pathway might lead to improved survival and efficacy. Here, we systematically evaluate the results of clinical trials investigating mTOR inhibition in CRPC and utilize preclinical data to predict clinical outcomes.MethodsTrials included in the study were identified through PubMed and via review of conference abstracts cited by relevant review articles. The eligibility of trials was independent of sample size, clinical setting, or date.ResultsA total of 14 studies were eligible for qualitative analysis. The clinical setting was variable among studies, and all utilized an allosteric mTOR inhibitor as either a monotherapy or in combination. Molecular criteria were evaluated in three trials. Among most studies, the prostate-specific antigen level declined during treatment, but often increased shortly thereafter. Partial responses to treatment were minimal, and no complete responses were reported. Two studies exploring therapy with an mTOR inhibitor in combination with bicalutamide resulted in minimal efficacy. Overall, allosteric mTOR inhibition was deemed to be inadequate for the treatment of CRPC.ConclusionPreclinical data suggest that a reciprocal feedback mechanism between PI3K and androgen receptor signaling is a potential mechanism behind the clinical inefficacy of mTOR inhibitors in CRPC, indicating combinatorial targeting of PI3K, mTORC1/2, and the androgen receptor might be more effective. Comprehensive analysis of preclinical data to assess clinical trial targets and efficacy may reduce the number of unproductive trials and identify potentially beneficial combinatorial therapies for resistant disease.

Molecular Genetic Analysis of Ovarian Brenner Tumors and Associated Mucinous Epithelial Neoplasms: High Variant Concordance and Identification of Mutually Exclusive RAS Driver Mutations and MYC Amplification

Benign ovarian Brenner tumors often are associated with mucinous cystic neoplasms, which are hypothesized to share a histogenic origin and progression, however, supporting molecular characterization is limited. Our goal was to identify molecular mechanisms linking these tumors. DNA from six Brenner tumors with paired mucinous tumors, two Brenner tumors not associated with a mucinous neoplasm, and two atypical proliferative (borderline) Brenner tumors was extracted from formalin-fixed, paraffin-embedded tumor samples and sequenced using a 358-gene next-generation sequencing assay. Variant calls were compared within tumor groups to assess somatic mutation profiles. There was high concordance of the variants between paired samples (40% to 75%; P < 0.0001). Four of the six tumor pairs showed KRAS hotspot driver mutations specifically in the mucinous tumor. In the two paired samples that lacked KRAS mutations, MYC amplification was detected in both of the mucinous and the Brenner components; MYC amplification also was detected in a third Brenner tumor. Five of the Brenner tumors had no reportable potential driver alterations. The two atypical proliferative (borderline) Brenner tumors both had RAS mutations. The high degree of coordinate variants between paired Brenner and mucinous tumors supports a shared origin or progression. Differences observed in affected genes and pathways, particularly involving RAS and MYC, may point to molecular drivers of a divergent phenotype and progression of these tumors.

A harmonized meta-knowledgebase of clinical interpretations of cancer genomic variants

Precision oncology relies on the accurate discovery and interpretation of genomic variants to enable individualized therapy selection, diagnosis, or prognosis. However, knowledgebases containing clinical interpretations of somatic cancer variants are highly disparate in interpretation content, structure, and supporting primary literature, reducing consistency and impeding consensus when evaluating variants and their relevance in a clinical setting. With the cooperation of experts of the Global Alliance for Genomics and Health (GA4GH) and of six prominent cancer variant knowledgebases, we developed a framework for aggregating and harmonizing variant interpretations to produce a meta-knowledgebase of 12,856 aggregate interpretations covering 3,437 unique variants in 415 genes, 357 diseases, and 791 drugs. We demonstrated large gains in overlapping terms between resources across variants, diseases, and drugs as a result of this harmonization. We subsequently demonstrated improved matching between patients of the GENIE cohort and harmonized interpretations of potential clinical significance, observing an increase from an average of 34% to 57% in aggregate. We developed an open and freely available web interface for exploring the harmonized interpretations from these six knowledgebases at search.cancervariants.org.

Barriers preventing the adoption of comprehensive cancer genomic profiling in the clinic

ABSTRACT Introduction: Comprehensive cancer genomic profiling provides the opportunity to expose the various molecular aberrations potentially driving tumor progression. Consequently, the identity of these genetic drivers can be utilized to match a patient to the most appropriate targeted therapy, thereby increasing the probability of improved clinical outcome. Despite its capability of informing patient care, the adoption of comprehensive cancer genomic profiling in the clinic has not been widespread. The barriers surrounding its universal acceptance are attributed to both physician and patient perspectives. Areas covered: The following report discusses the various obstacles in place, including those related to clinical utility, education, insurance coverage, and clinical trials, which can deter physicians and patients from utilizing genomic profiling for therapeutic decision-making. Expert commentary: The authors review the recent growth and potential of clinical utility studies over the last two years, provide a suggestive framework for educational support, and comment on the use of social media to enhance clinical trial recruitment.

Abstract 541: A comprehensive analysis delineating the immunotherapeutic terrain of cancer-related clinical trials

Technological innovations have facilitated a greater understanding of how the tumor microenvironment contributes to cancer, leading to rapid FDA approval of four immunotherapies. To assess how these therapies are being further investigated in combination with other therapies and in tumor types outside of the current FDA approval, we performed a comprehensive analysis of the curated clinical trials in the JAX Clinical Knowledgebase (JAX-CKB). In brief, clinical trials investigating Atezolizumab, Nivolumab, Pembrolizumab, and Ipilimumab, curated from clinicaltrials.gov, were queried in the JAX-CKB and then analyzed for comparison. Further analyses were executed to illustrate possible unmet needs within the field of cancer therapeutics. Of the four immunotherapies, Pembrolizumab was identified with the greatest number of clinical trials overall, with 305 compared to Atezolizumab, 79, Nivolumab, 183, and Ipilimumab, 126. Of these trials the number of trials investigating Pembrolizumab, Atezolizumab, Nivolumab, or Ipilimumab in combination with another therapy was higher than those investigating one of the four immunotherapies as a monotherapy. Phase II trials for both single therapy and combinatorial therapies were greater than both Phase I and Phase III for the same groups, regardless of therapy. On average, 12% ± 3.8% of the combined trials for all four drugs included any advanced solid tumor. Nivolumab combined with Ipilimumab demonstrated the greatest number of trials (61) investigating an immunotherapy in combination with another immunotherapy. Among those, 30 were Phase II trials, 16 were Phase I, and 14 were Phase III. Across five cancer indications (lung, pancreatic, ovarian, prostate, and colon), lung cancer was most commonly indicated in the trials, among all four drugs. Prostate was indicated in the least number of trials, with Ipilimumab ranking the highest (10). The recent success with immunotherapies has garnered significant interest in understanding how these therapies will perform in different tumor types and whether specific combinations will have a greater impact. Interrogation of the clinical trial terrain in the JAX-CKB provides a basis for determining additional investigations that might be warranted. Citation Format: Cara M. Statz, Sara E. Patterson, Taofei Yin, Susan M. Mockus. A comprehensive analysis delineating the immunotherapeutic terrain of cancer-related clinical trials [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 541. doi:10.1158/1538-7445.AM2017-541

Abstract 757: Development and validation of the ActionSeqTMtest system

Introduction In the constantly changing field of oncology precision medicine, it is exceedingly important to keep diagnostic and therapeutic assays clinically relevant. Next generation sequencing (NGS) panels in oncology are greatly impacted by new findings in clinical actionability. In order to ensure that cancer panels continue to provide the most beneficial results to patients, they must be regularly updated. In keeping with this idea, JAX has launched a new 212 oncology gene panel which focuses on genes and variants with documented actionability, referred to as ActionSeq. Methods Development of ActionSeq included the optimization of a new targeted capture assay. This process included running multiple batches of samples through the assay to determine appropriate DNA input, ligation times, PCR cycles, and pooling conditions. The fully optimized assay was then validated using 24 uncharacterized FFPE samples. The validation was executed in 5 phases: (1) confirm that assay optimizations yielded sufficient wet lab results; (2) LOD & sensitivity (3) inter-assay concordance; (4) intra-assay concordance; (5) specificity and accuracy. Results During development, the standard protocol was optimized using a 200ng input, 30 minute ligation period, 5 cycles of pre-PCR, and the pooling of 4 samples per hybridization reaction. Wet lab processing results of the first validation batch can be seen in Table 1. The inter- and intra-assay concordances were found to be ≥ 96% for variants and 100% for CNVs. The sensitivity was calculated to be 98.92% at a LOD of 3% for SNVs, 100% at a LOD of 8% for INDELs, 100% at a LOD of 6 copies for CNV amplifications, and 100% at a LOD of 0 copies for CNV deletions. The specificity and accuracy were found to be 100% for all mutation types. Conclusion Based on the success of this validation ActionSeq has been incorporated into the JAX clinical test menu. This addition accomplished the goal of providing a more clinically relevant (actionable) somatic tumor profiling assay to patients and clinicians. Citation Format: Samantha Helm, Vanessa Spotlow, Aleksandra Ras, Kevin Kelly, Guruprasad Ananda, Sara Patterson, Honey V. Reddi. Development and validation of the ActionSeq TM test system [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 757. doi:10.1158/1538-7445.AM2017-757

An open-source clinical knowledgebase featuring the ability to match patient somatic mutational profiles to eligible clinical trials.

e23242Background: Clinical trials in precision oncology generally require the presence or absence of specific mutations as part of the criteria for patient eligibility. Identification of these tria...

Abstract 4503: Actionable mutations and mutational burden in renal cell carcinoma

Clear cell renal cell carcinoma (CCRCC), the most common subtype of renal cell carcinoma, is a heterogeneous cancer with variable outcomes and molecular aberrations. Further elucidation of mutational profiles and actionable mutations may improve patient outcomes in CCRCC. Somatic mutational profiling via next-generation sequencing was conducted on 30 CCRCC whole section FFPE patient samples to determine the mutational burden as compared to other solid tumor types. The sequencing failure rate for these samples was 20% (6/30). A 358-gene panel was also run on an additional 57 non-CCRCC solid tumor FFPE patient samples (81 samples in total). FASTQ files generated from Illumina9s CASAVA software were submitted to the JAX Clinical Genome Analytics (CGA) data analysis pipeline to perform automated read quality assessment, alignment, and variant calling. Identified variants were then submitted for clinical curation using the JAX Clinical Knowledgebase (CKB) for actionable mutational analysis and overall mutational burden. Of the CCRCC sample set, 33% had one or more actionable mutations. This was low compared to actionable mutations in colon (16/16, 100% actionable), TNBC (18/20, 90% actionable), squamous lung (8/9, 89% actionable), and pancreatic (8/12, 67% actionable). Furthermore, the CCRCC set had a low overall mutational burden with 22 ± 6.7 nonsynonymous SNV9s as compared to colon (29 ± 11), squamous lung (43 ± 17), and TNBC (31.5 ± 12). However, the pancreatic set had even lower SNV9s at 12.7 ± 3.4. The number of indels was comparable across the five sets, except for colon, which averaged 6.9 ± 8.3. Significant copy number variations (CNV’s; amplification validated at greater than or equal to 6 copies) also had a wide range with squamous lung at the highest (11 ± 11), followed by TNBC (5.6 ± 5.0), and then colon (2.3 ± 3.0). Both the CCRCC and pancreatic sets had few CNV9s at 0.8 ± 1.5 and 0.17 ± 0.40, respectively. Common mutations identified in the CCRCC set included EPHB6 duplications near S166, which were identified in 25% of samples. Gene specific mutations in CCRCC were also identified in VHL (54%) and SETD2 (21%). These three genes, EPHB6, VHL, and SETD2 have been previously implicated in CCRCC pathogenesis. Six variants identified in VHL were C77*, P86R, Q132fs, A149D, R177fs, and L184fs. Due to the low mutational burden and/or actionable mutations in CCRCC, even large NGS panels may not be adequate to profile somatic mutations and whole exome or whole genome may be more conducive for molecular subtyping of renal cell carcinoma. Citation Format: Susan M. Mockus, Sara E. Patterson, Jason R. Pettus, Gregory J. Tsongalis. Actionable mutations and mutational burden in renal cell carcinoma. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4503.

Clinical knowledgebase to interprete genomic variations from targeted exome sequencing of patient tumor samples

Targeted exome sequencing of patient tumor samples has been applied in a CLIA-certified molecular diagnostic lab to connect genetic variants to personalized treatments. The Jackson Laboratory has established a Cancer Treatment Profile (JAX-CTP) of 358 genes. We have also built a Clinical Knowledgebase (CKB) that enables annotation of genomic variations as predictive markers of cancer treatment responses, including drug sensitivity and resistance. Clinical trials on targeted cancer therapies in the United States and Canada are also integrated into the knowledgebase. In this report, we are presenting the integrated informatics framework, including the adoption of standard ontologies and transformation of genomic variations using Human Genome Variation Society (HGVS) nomenclature in our data loading process. Our framework uses regular expressions (regex) as control and is able to support a scalable number of variation and expression types in the molecular profiles, such as amino acid changes, gene amplifications and gene fusions. This process automatically maps genomic variants from the JAX clinical analytic pipeline to the predictive markers in the Clinical Knowledgebase. It enables our clinical analysis team to quickly interpret genomic variations in patient samples and identify clinically actionable variants with associated cancer treatment responses in the knowledgebase. This automation greatly improves the capacity to produce clinical reports by querying our CKB database.