Nancy K. Gillis

Clinical Implementation of Germ Line Cancer Pharmacogenetic Variants During the Next‐Generation Sequencing Era

More than 100 medications approved by the US Food and Drug Administration include pharmacogenetic biomarkers in the drug label, many with cancer indications referencing germ line DNA variations. With the advent of next‐generation sequencing (NGS) and its rapidly increasing uptake into cancer research and clinical practice, an enormous amount of data to inform documented gene–drug associations will be collected that must be exploited to optimize patient benefit. This review focuses on the implementation of germ line cancer pharmacogenetics in clinical practice. Specifically, it discusses the importance of germ line variation in cancer and the role of NGS in pharmacogenetic discovery and implementation. In the context of a scenario in which massive amounts of NGS‐based genetic information will be increasingly available to health stakeholders, this review explores the ongoing debate regarding the threshold of evidence necessary for implementation, provides an overview of recommendations in cancer by professional organizations and regulatory bodies, and discusses limitations of current guidelines and strategies to improve third‐party coverage.

Clonal haemopoiesis and therapy-related myeloid malignancies in elderly patients: a proof-of-concept, case-control study

BACKGROUND Clonal haemopoiesis of indeterminate potential (CHIP) is an age-associated genetic event linked to increased risk of primary haematological malignancies and increased all-cause mortality, but the prevalence of CHIP in patients who develop therapy-related myeloid neoplasms is unknown. We did this study to investigate whether chemotherapy-treated patients with cancer who have CHIP are at increased risk of developing therapy-related myeloid neoplasms. METHODS We did a nested, case-control, proof-of-concept study to compare the prevalence of CHIP between patients with cancer who later developed therapy-related myeloid neoplasms (cases) and patients who did not develop these neoplasms (controls). We identified cases from our internal biorepository of 123 357 patients who consented to participate in the Total Cancer Care biobanking protocol at Moffitt Cancer Center (Tampa, FL, USA) between Jan 1, 2006, and June 1, 2016. We included all individuals who were diagnosed with a primary malignancy, were treated with chemotherapy, subsequently developed a therapy-related myeloid neoplasm, and were 70 years or older at either diagnosis. For inclusion in this study, individuals must have had a peripheral blood or mononuclear cell sample collected before the diagnosis of therapy-related myeloid neoplasm. Controls were individuals who were diagnosed with a primary malignancy at age 70 years or older and were treated with chemotherapy but did not develop therapy-related myeloid neoplasms. Controls were matched to cases in at least a 4:1 ratio on the basis of sex, primary tumour type, age at diagnosis, smoking status, chemotherapy drug class, and duration of follow-up. We used sequential targeted and whole-exome sequencing and described clonal evolution in cases for whom paired CHIP and therapy-related myeloid neoplasm samples were available. The primary endpoint of this study was the development of therapy-related myeloid neoplasm and the primary exposure was CHIP. FINDINGS We identified 13 cases and 56 case-matched controls. The prevalence of CHIP in all patients (23 [33%] of 69 patients) was higher than has previously been reported in elderly individuals without cancer (about 10%). Cases had a significantly higher prevalence of CHIP than did matched controls (eight [62%] of 13 cases vs 15 [27%] of 56 controls, p=0·024; odds ratio 5·75, 95% CI 1·52-25·09, p=0·013). The most commonly mutated genes in cases with CHIP were TET2 (three [38%] of eight patients) and TP53(three [38%] of eight patients), whereas controls most often had TET2 mutations (six [40%] of 15 patients). In most (four [67%] of six patients) cases for whom paired CHIP and therapy-related myeloid neoplasm samples were available, the mean allele frequency of CHIP mutations had expanded by the time of the therapy-related myeloid neoplasm diagnosis. However, a subset of paired samples (two [33%] of six patients) had CHIP mutations that decreased in allele frequency, giving way to expansion of a distinct mutant clone. INTERPRETATION Patients with cancer who have CHIP are at increased risk of developing therapy-related myeloid neoplasms. The distribution of CHIP-related gene mutations differs between individuals with therapy-related myeloid neoplasm and those without, suggesting that mutation-specific differences might exist in therapy-related myeloid neoplasm risk. FUNDING Moffitt Cancer Center.

Evidence Required to Demonstrate Clinical Utility of Pharmacogenetic Testing: The Debate Continues

Pharmacogenetics is an area of research that has potential to greatly benefit patients. However, the routine use of diagnostic pharmacogenetic testing to inform treatment decisions is limited. Here we discuss the determination of clinical utility of pharmacogenetic testing and the level of evidence required to support translation into clinical practice.

Key Lessons Learned from Moffitt's Molecular Tumor Board: The Clinical Genomics Action Committee Experience

BACKGROUND The increasing practicality of genomic sequencing technology has led to its incorporation into routine clinical practice. Successful identification and targeting of driver genomic alterations that provide proliferative and survival advantages to tumor cells have led to approval and ongoing development of several targeted cancer therapies. Within many major cancer centers, molecular tumor boards are constituted to shepherd precision medicine into clinical practice. MATERIALS AND METHODS In July 2014, the Clinical Genomics Action Committee (CGAC) was established as the molecular tumor board companion to the Personalized Medicine Clinical Service (PMCS) at Moffitt Cancer Center in Tampa, Florida. The processes and outcomes of the program were assessed in order to help others move into the practice of precision medicine. RESULTS Through the establishment and initial 1,400 patients of the PMCS and its associated molecular tumor board at a major cancer center, five practical lessons of broad applicability have been learned: transdisciplinary engagement, the use of the molecular report as an aid to clinical management, clinical actionability, getting therapeutic options to patients, and financial considerations. Value to patients includes access to cutting-edge practice merged with individualized preferences in treatment and care. CONCLUSIONS Genomic-driven cancer medicine is increasingly becoming a part of routine clinical practice. For successful implementation of precision cancer medicine, strategically organized molecular tumor boards are critical to provide objective evidence-based translation of observed molecular alterations into patient-centered clinical action. Molecular tumor board implementation models along with clinical and economic outcomes will define future treatment standards. The Oncologist 2017;22:144-151Implications for Practice: It is clear that the increasing practicality of genetic tumor sequencing technology has led to its incorporation as part of routine clinical practice. Subsequently, many cancer centers are seeking to develop a personalized medicine services and/or molecular tumor board to shepherd precision medicine into clinical practice. This article discusses the key lessons learned through the establishment and development of a molecular tumor board and personalized medicine clinical service. This article highlights practical issues and can serve as an important guide to other centers as they conceive and develop their own personalized medicine services and molecular tumor boards.

The pharmacogenomics of drug resistance to protein kinase inhibitors

Dysregulation of growth factor cell signaling is a major driver of most human cancers. This has led to development of numerous drugs targeting protein kinases, with demonstrated efficacy in the treatment of a wide spectrum of cancers. Despite their high initial response rates and survival benefits, the majority of patients eventually develop resistance to these targeted therapies. This review article discusses examples of established mechanisms of drug resistance to anticancer therapies, including drug target mutations or gene amplifications, emergence of alternate signaling pathways, and pharmacokinetic variation. This reveals a role for pharmacogenomic analysis to identify and monitor for resistance, with possible therapeutic strategies to combat chemoresistance.

Chipping in on clonal hematopoiesis

Myeloid neoplasms are clonal diseases of hematopoietic stem or progenitor cells that result from molecular alterations that perturb cellular self-renewal, proliferation, and differentiation. As classified by the World Health Organization, myeloid neoplasms include myeloproliferative neoplasms (MPN), myelodysplastic syndromes (MDS), MPN/MDS overlap, and acute myeloid leukemia (AML), which can occur de novo, secondary to MDS/MPN, or after treatment with chemotherapy or radiation (therapy-related, T-MN). Epidemiological features of myeloid neoplasms include higher incidence in men, Caucasians, and increased frequency with age. Aside from these demographic criteria, there are currently no clear biomarkers or risk factors for predisposition to myeloid malignancies. The recent discovery of age-related clonal hematopoiesis [1, 2], commonly termed clonal hematopoiesis of indeterminate potential (CHIP), may narrow this critical knowledge gap. CHIP is an idiopathic genetic event in which individuals harbor somatic mutations, primarily in genes associated with myeloid neoplasms (e.g., DNMT3A, TET2, and ASXL1), without overt signs of hematologic malignancy. Similar to myeloid neoplasms, CHIP mutations are most frequent in older individuals, men, and Caucasians [1-3]. The presence of CHIP is associated with poor outcomes, including a significantly increased risk of hematologic malignancies (HR 12.9), all-cause mortality (HR 1.4), and cardiovascular disease (HR 2.0) [1, 2]. We, and others, have demonstrated that individuals with CHIP mutations at the time of primary cancer are also at a significantly increased risk for T-MNs [4, 5]. Taken together, these findings suggest that CHIP mutations may represent an important biomarker for the development of myeloid neoplasms. While we understand the basic demographics of individuals who have CHIP mutations, a complete understanding of who acquires CHIP has yet to be elucidated. One study reported a modest association (OR 1.37) with germline mutations in TERT (telomerase reverse transcriptase) [6]. Environmental factors, such as smoking and radiation, have also been associated with CHIP [1, 3]. However, it is also possible that CHIP occurs through stochastic errors in DNA replication. It has been hypothesized that cancer incidence rates correlate with stem cell turnover within the tissue of origin, especially for cancers without strong hereditary or environmental components [7]. In fact, a strong correlation (r2 = 0.80) was observed between total stem cell divisions during an average lifetime and lifetime risk for cancer of the corresponding tissue type [7]. Notably, cancers without known environmental risk factors, such as MDS and AML, had the strongest correlation with stem cell turnover. Perhaps this model of association between seemingly spontaneous cancers and stem cell division rate can help explain at least a portion of acquired CHIP mutations. If this hypothesis proves true for CHIP, efforts should be refocused toward secondary prevention (i.e., monitoring for early detection of progression to disease and early intervention), as primary prevention (i.e., avoidance of risk factors) would not impact stochastically acquired mutations. A subsequent critical knowledge gap relates to what drives progression from CHIP to overt hematologic malignancy. Is CHIP a precursor for de novo MDS/AML in all cases? While the incidence of myeloid malignancies is significantly higher in individuals with CHIP, noncancer patients with CHIP have only an approximately five percent absolute risk of developing a hematologic malignancy [2]. Contributors to progression could include hereditary, environmental, gene-specific factors (i.e., which “CHIP genes” are mutated), or stochastic events similar to that responsible for CHIP acquisition. Therefore, do the factors those drive CHIP development also drive progression? The observation that cancer patients of all Editorial

Identification of clonal hematopoiesis mutations in solid tumor patients undergoing unpaired next-generation sequencing assays

Purpose: In this era of precision-based medicine, for optimal patient care, results reported from commercial next-generation sequencing (NGS) assays should adequately reflect the burden of somatic mutations in the tumor being sequenced. Here, we sought to determine the prevalence of clonal hematopoiesis leading to possible misattribution of tumor mutation calls on unpaired Foundation Medicine NGS assays. Experimental Design: This was a retrospective cohort study of individuals undergoing NGS of solid tumors from two large cancer centers. We identified and quantified mutations in genes known to be frequently altered in clonal hematopoiesis (DNMT3A, TET2, ASXL1, TP53, ATM, CHEK2, SF3B1, CBL, JAK2) that were returned to physicians on clinical Foundation Medicine reports. For a subset of patients, we explored the frequency of true clonal hematopoiesis by comparing mutations on Foundation Medicine reports with matched blood sequencing. Results: Mutations in genes that are frequently altered in clonal hematopoiesis were identified in 65% (1,139/1,757) of patients undergoing NGS. When excluding TP53, which is often mutated in solid tumors, these events were still seen in 35% (619/1,757) of patients. Utilizing paired blood specimens, we were able to confirm that 8% (18/226) of mutations reported in these genes were true clonal hematopoiesis events. The majority of DNMT3A mutations (64%, 7/11) and minority of TP53 mutations (4%, 2/50) were clonal hematopoiesis. Conclusions: Clonal hematopoiesis mutations are commonly reported on unpaired NGS testing. It is important to recognize clonal hematopoiesis as a possible cause of misattribution of mutation origin when applying NGS findings to a patients care. See related commentary by Pollyea, p. 5790

Quantitation of Targetable Somatic Mutations among Patients Evaluated by a Personalized Medicine Clinical Service: Considerations for Off-Label Drug Use

Moffitt Cancer Centers Personalized Medicine Clinical Service (PMCS) reviews somatic next‐generation sequencing (NGS) assay results, provides interpretations, and identifies potential therapeutic options. The number of individuals reviewed by our clinical service who are eligible for on‐label or off‐label drug therapy based on genetic test results has previously not been quantitated. We determined the number of patients harboring an actionable mutation that would qualify a patient for an on‐label drug or consideration for off‐label drug treatment.

Tumor exome sequencing and copy number alterations reveal potential predictors of intrinsic resistance to multi-targeted tyrosine kinase inhibitors

Multi-targeted tyrosine kinase inhibitors (TKIs) have broad efficacy and similar FDA-approved indications, suggesting shared molecular drug targets across cancer types. Irrespective of tumor type, 20-30% of patients treated with multi-targeted TKIs demonstrate intrinsic resistance, with progressive disease as a best response. We conducted a retrospective cohort study to identify tumor (somatic) point mutations, insertion/deletions, and copy number alterations (CNA) associated with intrinsic resistance to multi-targeted TKIs. Using a candidate gene approach (n=243), tumor next-generation sequencing and CNA data was associated with resistant and non-resistant outcomes. Resistant individuals (n=11) more commonly harbored somatic point mutations in NTRK1, KDR, TGFBR2, and PTPN11 and CNA in CDK4, CDKN2B, and ERBB2 compared to non-resistant (n=26, p<0.01). Using a random forest classification model for variable reduction and a decision tree classification model, we were able to differentiate intrinsically resistant from non-resistant patients. CNA in CDK4 and CDKN2B were the most important analytical features, implicating the cyclin D pathway as a potentially important factor in resistance to multi-targeted TKIs. Replication of these results in a larger, independent patient cohort has potential to inform personalized prescribing of these widely utilized agents.

Abstract 4304: Prevalence and triggers of drug-induced Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) in a cancer patient cohort

SJS and TEN are extremely rare (approximately 2-7 cases per million per year) hypersensitivity reactions most commonly attributed to medications. The life-threatening nature of SJS/TEN necessitates early diagnosis and immediate identification and withdrawal of causative agents. Some of the most commonly associated culprits include sulfonamide antibiotics, aromatic anticonvulsants (phenytoin, carbamazepine, lamotrigine), -oxicam NSAIDs (e.g., meloxicam), allopurinol, and nevirapine. Individual case reports of SJS/TEN have been associated with anti-cancer agents, but a comprehensive assessment has not been performed. A large-scale retrospective analysis of the prevalence of SJS/TEN in cancer patients treated at Moffitt Cancer Center (MCC) between Jan 1, 2002 - Dec 31, 2014 was conducted. A total of 104,917 cancer patients were screened for SJS/TEN-related ICD-9 codes in the MCC health research informatics database. SJS/TEN diagnoses were identified in 121 patients. Manual chart review of physician, dermatology, and pathology notes confirmed SJS/TEN in 47 patients (in-patient + historical) and possible SJS/TEN in an additional 17 patients, corresponding to an overall prevalence of 0.06%. Confirmed in-patient cases of SJS/TEN were more common in hematologic malignancies compared to solid tumors (n = 12 vs. n = 7, respectively). Notably, 5 of the 19 (26.3%) confirmed cases were observed in patients with acute myeloid leukemia. Physician-reported culprits for SJS/TEN diagnoses included antibiotics, immunomodulators, anticonvulsants, and anti-cancer agents. The observed prevalence of SJS/TEN was higher in cancer patients than previous reports from the general USA population. There were 19 confirmed in-patient diagnoses of SJS/TEN and an additional 45 historical and possible cases of SJS/TEN in 104,917 cancer patients, corresponding to an overall prevalence of 0.018% to 0.06%, an approximately 90-fold increase when compared to the general population. Possible explanations for increased risk in cancer patients include increased exposure to culprit medications, cancer disease process, immunocompromised state, or synergy between risk factors. A thorough understanding of the factors that increase risk to SJS/TEN in cancer patients is critical to facilitate culprit withdrawal and maximize patient outcomes and survival. Citation Format: Nancy K. Gillis, Gillian C. Bell, Howard L. McLeod, Amy J. Brandt. Prevalence and triggers of drug-induced Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) in a cancer patient cohort. [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 4304.