Search

Uveal melanomas are rare tumors arising from melanocytes that reside in the eye. Despite surgical or radiation treatment, approximately 50% of patients with uveal melanoma will progress to metastatic disease, most often to the liver. Cell-free DNA (cfDNA) sequencing is a promising technology due to the minimally invasive sample collection and ability to infer multiple aspects of tumor response. We analyzed 46 serial cfDNA samples from 11 patients with uveal melanoma over a 1-year period following enucleation or brachytherapy (n = ∼4/patient) using targeted panel, shallow whole genome, and cell-free methylated DNA immunoprecipitation sequencing. We found detection of relapse was highly variable using independent analyses (P = 0.06–0.46), whereas a logistic regression model integrating all cfDNA profiles significantly improved relapse detection (P = 0.02), with greatest power derived from fragmentomic profiles. This work provides support for the use of integrated analyses to improve the sensitivity of circulating tumor DNA detection using multi-modal cfDNA sequencing.Significance:Here, we demonstrate integrated, longitudinal cfDNA sequencing using multi-omic approaches is more effective than unimodal analysis. This approach supports the use of frequent blood testing using comprehensive genomic, fragmentomic, and epigenomic techniques.

Supplemental Figure 3

Uveal melanomas are rare tumors arising from melanocytes that reside in the eye. Despite surgical or radiation treatment, approximately 50% of patients with uveal melanoma will progress to metastatic disease, most often to the liver. Cell-free DNA (cfDNA) sequencing is a promising technology due to the minimally invasive sample collection and ability to infer multiple aspects of tumor response. We analyzed 46 serial cfDNA samples from 11 patients with uveal melanoma over a 1-year period following enucleation or brachytherapy (n = ∼4/patient) using targeted panel, shallow whole genome, and cell-free methylated DNA immunoprecipitation sequencing. We found detection of relapse was highly variable using independent analyses (P = 0.06–0.46), whereas a logistic regression model integrating all cfDNA profiles significantly improved relapse detection (P = 0.02), with greatest power derived from fragmentomic profiles. This work provides support for the use of integrated analyses to improve the sensitivity of circulating tumor DNA detection using multi-modal cfDNA sequencing.Significance:Here, we demonstrate integrated, longitudinal cfDNA sequencing using multi-omic approaches is more effective than unimodal analysis. This approach supports the use of frequent blood testing using comprehensive genomic, fragmentomic, and epigenomic techniques.

differentially methylated probes

Background In high grade serous ovarian cancer (HGSOC), there is a spectrum of sensitivity to first line platinum‐based chemotherapy. This study molecularly characterizes HGSOC patients from two distinct groups of chemotherapy responders (good vs. poor). Methods Following primary debulking surgery and intravenous carboplatin/paclitaxel, women with stage III–IV HGSOC were grouped by response. Patients in the good response (GR) and poor response (PR) groups respectively had a progression‐free intervals (PFI) of ≥12 and ≤6 months. Analysis of surgical specimens interrogated genomic and immunologic features using whole exome sequencing. RNA‐sequencing detected gene expression outliers and inference of immune infiltrate, with validation by targeted NanoString arrays. PD‐L1 expression was scored by immunohistochemistry (IHC). Results A total of 39 patient samples were analyzed (GR = 20; PR = 19). Median PFI for GR and PR patient cohorts was 32 and 3 months, respectively. GR tumors were enriched for loss‐of‐function BRCA2 mutations and had a significantly higher nonsynonymous mutation rate compared to PR tumors (p = 0.001). Samples from the PR cohort were characterized by mutations in MGA and RAD51B and trended towards a greater rate of amplification of PIK3CA, MECOM, and ATR in comparison to GR tumors. Gene expression analysis by NanoString correlated increased PARP4 with PR and increased PD‐L1 and EMSY with GR. There was greater tumor immune cell infiltration and higher immune cell PD‐L1 protein expression in the GR group. Conclusions Our research demonstrates that tumors from HGSOC patients responding poorly to first line chemotherapy have a distinct molecular profile characterized by actionable drug targets including PARP4., In patients with suboptimally‐debulked, advanced high grade serous ovarian cancer (HGSOC), the spectrum of responses to first line platinum‐based chemotherapy (PtC) vary from durable to non‐existent, highlighting an unmet medical need for those with poor responses. Our study molecularly characterizes a cohort of HGSOC patients with distinct responses, either good or poor, to PtC. The good and poor responders have median progression‐free intervals of 32 and 3 months, respectively, and molecular analysis of tumors identifies differences in DNA damage response pathways and immunity, namely, tumors from poor responders to first line PtC have a distinct molecular profile characterized by actionable drug targets including PARP4.

Background The correlation between blood-based tumor mutation burden (bTMB) and tissue-based tumor mutation burden(tTMB) has not been broadly tested in a multicancer cohort. Here, we assess the correlation between bTMB with tTMB in phase I trial patients treated with immunotherapy. As an exploratory analysis, we evaluated circulating tumor DNA (ctDNA) dynamics in responders. Methods Patients treated with immunotherapy at the Princess Margaret phase I trials unit were enrolled. Pretreatment plasma ctDNA and matched normal blood controls were collected. Available archival tissue formalin-fixed paraffin-embedded (FFPE) samples were analyzed. A 425-gene panel was used to sequence both ctDNA and FFPE samples. Samples with TMB within the highest tertile were considered as high TMB. Results Thirty-eight patients were accrued from 25 different trials, 86.8% of which involved an anti-PD-1/PD-L1 agent. Thirty patients (78.9%) had detectable mutations in ctDNA, of which the median (range) bTMB was 5 (1-53) mutations per megabase (mut/Mb). Of the 22 patients with available FFPE samples, mutations were detected in 21 (95.4%); the median (range) tTMB was 6 (2-124) mut/Mb. Among the 16 patients with detectable mutations in both FFPE and ctDNA, a statistically significant correlation between bTMB and tTMB was observed (ρ = 0.71; P = .002). High TMB was not associated with better survival. All 3 responders had a decrease in the variant allele frequency of mutations detected in ctDNA at a second timepoint relative to baseline, indicating a potential early marker of response. Conclusions In this small series, bTMB correlated with tTMB. An on-treatment decrease in VAF of mutations detected in ctDNA at baseline was observed in responders. Larger studies to verify our findings are warranted.

Background High tissue-tumor mutation burden (tTMB) is a predictor of response to immunotherapy (IO). Tissue availability and tumor heterogeneity are barriers to tTMB use in clinical practice. Plasma-based blood TMB (bTMB) is a convenient alternative that strongly correlates with tTMB in non-small cell lung cancer. Whether this correlation holds true in other cancers is unknown. Here, we examined the correlation between bTMB and tTMB as well as the clinical utility of TMB as a predictive marker of response in a heterogeneous Phase I IO cohort. Methods Advanced cancer patients (pts) treated with mono- or combination IO therapy at the Princess Margaret phase I unit were enrolled. Pre-treatment plasma ctDNA and matched normal blood controls were collected via an institutional liquid biopsy program (LIBERATE, NCT03702309). Available archival tissue FFPE samples were analyzed. The GeneseeqPrime 425 gene panel was used to sequence both ctDNA and FFPE samples. Results From December 2017 to July 2018, 39 pts with 19 tumor types were accrued from 25 different trials, 87% of which involved a PD-1/PD-L1 inhibitor. The median age was 59y (21 – 77) and 52% were female. The most frequent cancers were colorectal, head and neck and breast, each with 5 cases. Thirty-one patients (79%) had detectable mutations in plasma ctDNA. The median bTMB was 5 (1 - 53) mut/Mb. Twenty-one pts had available FFPE samples. Of those, mutations were detected in 20 (95%) samples. The median tTMB was 6 (2 - 124) mut/Mb. Among the 16 pts with detectable mutations in both FFPE and plasma samples, a significant correlation between bTMB and tTMB was observed (r = +0.67; p Conclusions In a typical heterogeneous phase I IO cohort, bTMB was correlated with tTMB. In this small series, neither bTMB nor tTMB were associated with survival. However, 2/3 PRs had high bTMB. Further studies in larger cohorts are warranted. Legal entity responsible for the study Lillian Siu. Funding Princess Margaret Cancer Centre; BMO Chair in Precision Genomics; Geneseeq Technology Inc. Disclosure A. Wang: Full / Part-time employment: Geneseeq Technology Inc. J. Huang: Full / Part-time employment: Geneseeq Technology Inc. A. Spreafico: Advisory / Consultancy, Research grant / Funding (self): Novartis; Advisory / Consultancy, Research grant / Funding (self), Travel / Accommodation / Expenses: Merck; Advisory / Consultancy, Research grant / Funding (self), Travel / Accommodation / Expenses: Bristol-Myers Squibb; Advisory / Consultancy: Oncorus; Travel / Accommodation / Expenses: Idera; Research grant / Funding (institution), Travel / Accommodation / Expenses: Bayer; Research grant / Funding (self): Symphogen; Research grant / Funding (self): AstraZeneca/MedImmune; Research grant / Funding (self): Surface Oncology; Research grant / Funding (self): Jansseen Oncology; Research grant / Funding (self): Northern Biologics. A.R. Hansen: Advisory / Consultancy: Genentech/Roche; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): Merck; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): GlaxoSmithKline; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): Bristol-Myers Squibb; Advisory / Consultancy, Research grant / Funding (institution): Novartis; Advisory / Consultancy: Boston Biomedical; Advisory / Consultancy, Research grant / Funding (self): Boehringer Ingelheim; Honoraria (self): AstraZeneca/MedImmune; Honoraria (self): Pfizer; Research grant / Funding (institution): Karyopharm Therapeutics. A.A. Razak: Advisory / Consultancy, Research grant / Funding (self): Lilly; Advisory / Consultancy, Research grant / Funding (self): Merck; Honoraria (self), Advisory / Consultancy, Research grant / Funding (self): Boehringer Ingelheim; Research grant / Funding (self): CASI Pharmaceuticals; Research grant / Funding (self): Novartis; Research grant / Funding (self): Deciphera; Research grant / Funding (self): Karyopharm Therapeutics; Research grant / Funding (self): Pfizer; Research grant / Funding (self): Roche/Genentech; Research grant / Funding (self): Boston Biomedical; Research grant / Funding (self): Bristol-Myers Squibb; Research grant / Funding (self): MedImmune; Research grant / Funding (self): Amgen; Research grant / Funding (self): GlaxoSmithKline; Research grant / Funding (self): Blueprint Medicines; Research grant / Funding (self): AbbVie; Research grant / Funding (self): Adaptimmune. P. Bedard: Research grant / Funding (institution): Bristol-Myers Squibb; Research grant / Funding (institution): Sanofi; Research grant / Funding (institution): AstraZeneca; Research grant / Funding (institution): Genentech/Roche; Research grant / Funding (institution): Servir; Research grant / Funding (institution): GlaxoSmithKline; Research grant / Funding (institution): Novartis; Research grant / Funding (institution): SignalChem; Research grant / Funding (institution): PTC Therapeutics; Research grant / Funding (institution): Nektar; Research grant / Funding (institution): Merck; Research grant / Funding (institution): Seattle Genetics; Research grant / Funding (institution): Mersana; Research grant / Funding (institution): Immunomedics; Research grant / Funding (institution): Lilly. H. Bao: Full / Part-time employment: Geneseeq Technology Inc. X. Wu: Leadership role, Full / Part-time employment: Geneseeq Technology Inc. T.J. Pugh: Advisory / Consultancy: DynaCare; Licensing / Royalties: Hybrid-capture sequencing for determining immune cell clonality; Licensing / Royalties: Combined hybrid-capture DNA sequencing for disease detection; Honoraria (self): Merck; Honoraria (self): Prosigna; Honoraria (self): Chrysalis Biomedical Advisors; Research grant / Funding (institution): Boehringer Ingelheim. L.L. Siu: Advisory / Consultancy, Research grant / Funding (institution): Merck; Advisory / Consultancy, Research grant / Funding (institution): AstraZeneca/MedImmune; Advisory / Consultancy: MorphoSys; Advisory / Consultancy, Research grant / Funding (institution): Symphony Evolution; Advisory / Consultancy, Research grant / Funding (institution): Genentech/Roche; Advisory / Consultancy: Loxo; Shareholder / Stockholder / Stock options, Spouse / Financial dependant, Immediate Family Member - Leadership and Stock/Owenership: Angios; Research grant / Funding (institution): Bristol-Myers Squibb; Research grant / Funding (institution): GlaxoSmithKline; Research grant / Funding (institution): Novartis; Research grant / Funding (institution): Pfizer; Research grant / Funding (institution): Boehringer Ingelheim; Research grant / Funding (institution): Bayer; Research grant / Funding (institution): Amgen; Research grant / Funding (institution): Astellas Pharma. All other authors have declared no conflicts of interest.