1
The use of low-coverage sequencing of cell-free DNA for monitoring response to immune checkpoint inhibitors throughout treatment Taylor J. Jensen 1 , Aaron M. Goodman 2,3 , Shumei Kato 2,4 , Mina Nikanjam 2 , Christopher K. Ellison 1 , Gregory A. Daniels 2 , Lisa Kim 2 , Kimberly Kelly 1 , Kerry Fitzgerald 1 , Erin McCarthy 1 , Prachi Nakashe 1 , Amin R. Mazloom 1 , Eyad Almasri 1 , Graham McLennan 1 , Daniel S. Grosu 1 , Mathias Ehrich 1 , Razelle Kurzrock 2 1 Sequenom Inc., a wholly owned subsidiary of Laboratory Corporation of America® Holdings, San Diego, California, 92121; 2 Department of Medicine, Division of Hematology/Oncology, and Center for Personalized Cancer Therapy, University of California, San Diego, Moores Cancer Center; 3 Department of Medicine, Division of Blood and Marrow Transplantation, University of California, San Diego, Moores Cancer Center; 4 Department of Medicine, Division of Precision Medicine, University of California, San Diego, Moores Cancer Center III. Conclusion These data suggest that low coverage, genome-wide sequencing of cfDNA may have utility for monitoring response to immunotherapy in cancer patients. II. Methods IV. References 1. Ellison, CK et al., Using Targeted Sequencing of Paralogous Sequences for Noninvasive Detection of Selected Fetal Aneuploidies. Clin Chem, 2016. 62(12): p. 1621-1629. 2. Tynan, JA et al., Application of risk score analysis to low-coverage whole genome sequencing data for the noninvasive detection of trisomy 21, trisomy 18, and trisomy 13. Prenat Diagn, 2016. 36(1): p. 56-62. 3. Lefkowitz, RB et al., Clinical validation of a noninvasive prenatal test for genomewide detection of fetal copy number variants. Am J Obstet Gynecol, 2016. 215(2): p. 227 e1-227 e16. 4. Zhao, C et al., Detection of fetal subchromosomal abnormalities by sequencing circulating cell-free DNA from maternal plasma. Clin Chem, 2015. 61(4): p. 608-16. 5. Eisenhauer, EA et al., New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer, 2009. 45(2): p. 228-47. 6. Team, R.D.C., R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2008 I. Introduction Immune checkpoint inhibitors continue to revolutionize the cancer treatment paradigm. It has been observed that even some patients with advanced, refractory malignancies achieve durable responses; however, only a minority of patients benefit, demonstrating the importance of developing new biomarkers to predict and/or monitor patient outcome. While markers including PD-1/PD-L1 expression, microsatellite instability, and tumor mutational burden have been shown to have varying degrees of predictive power, they are not the ideal markers to monitor and differentiate response during treatment. Interrogating cell-free DNA (cfDNA) isolated from plasma (liquid biopsy) provides a promising noninvasive method for monitoring response. ©2018 Laboratory Corporation of America® Holdings. All rights reserved. onc-1021-v1-1018 Poster presented at the SITC Annual Meeting; 2018 November 7-11; Washington, D.C. Patient selection: We prospectively enrolled 76 patients with cancer who were going to initiate treatment with an immunotherapy at UC San Diego Moores Cancer Center between September, 2015 and November, 2017. Of the 76 patients enrolled, 68 had a baseline sample collected prior to the start of immunotherapy and could therefore be used for analysis. Data was collected by chart and imaging review. Patient treatment: Patients were treated with immunotherapy that included a checkpoint inhibitor per approved indication or on an experimental protocol. This study was performed and consents were obtained in accordance with UCSD Institutional Review Board guidelines for specimen collection and data analysis (NCT02478931) and for any investigational treatments. Sample collection and processing: Whole blood (~10 mL) was collected in Streck BCT tubes (Streck, Omaha, NE) and processed to plasma as previously described. 1 For a subset of patients, serial samples were collected throughout clinical treatment. Samples were deemed to be the baseline for each patient if the blood collection occurred prior to the initiation of immunotherapy treatment. If more than one sample was collected during that time, the baseline sample was the sample collected closest to the start of treatment. cfDNA extraction: cfDNA from the plasma of each sample was extracted using a bead-based method as previously described. 1 Sequencing library preparation: Libraries for genome-wide sequencing were created from cfDNA as previously described. 2 Mixture model preparation: DNA from four cell lines was obtained from the American Type Tissue Culture (ATCC) and sonicated to resemble cfDNA profiles (average size of ~170bp, verified by BioAnalyzer fragment size evaluation). Three uniquely barcoded library replicates were generated for each cell line, and libraries were quantified by electrophoresis. Cell line library replicates were each mixed with twelve uniquely barcoded libraries from healthy donors so that the cell line library DNA represented 1%, 2%, 3%, 5%, 10%, and 25% of the total mixture. The mixtures were pooled in 6-plex per lane with 48 mixtures per flowcell and sequenced on an Illumina HiSeq2500 sequencer. Genome-wide next generation sequencing: Normalized sequencing libraries were pooled and sequenced using HiSeq2500 (Illumina) instruments as previously described. 3 A mean of 33.9 million sequencing reads (~0.3X genomic coverage) were aligned to the human reference genome for each sample. Sequencing data analysis: Sequencing data were processed as previously described. 3 Briefly, sequencing reads were mapped to the human reference genome (hg19) and partitioned into 50 kbp non-overlapping segments. Regions were selected and data were normalized as previously performed for noninvasive detection of fetal copy number variants 4 and the resultant normalized values were used to calculate a genome instability number (GIN). The GIN is a metric intended to capture genome-wide autosomal deviation from empirically derived euploid dosage of the genome in circulation. The GIN is a non-negative, continuous value calculated as the absolute deviation of observed normalized sequencing read coverage from expected normalized read coverage summed across 50,034 autosomal segments. Observed normalized read coverage is defined for each genomic segment by an autosome-specific loess fit of the normalized data. Increasing values of GIN were observed to be indicative of increasing deviation relative to an expected normal genomic profile. Statistical Analysis and Outcome Evaluation: Clinical responses were assessed based on physician notation and radiograph review using RECIST criteria. 5 Progression-free survival (PFS) was calculated by the method of Kaplan and Meier (P values by log-rank (Mantel-Cox)) test. Patients were censored at date of last follow up for PFS, if they had not progressed. For disease status prediction using the GIN, patients were evaluated at the time point nearest +42 days relative to treatment start site. Statistical analyses were carried out using either custom scripts in an R programming language 6 or Graph-Pad Prism version 7.0 (San Diego, CA, USA). Figure 4. Overview of Patient 110 where the GIN prediction was discordant with RECIST criteria (stable disease for 4 months (hence labelled as progressor)). A) Genome-wide cfDNA profiles from each of the time points described above each plot. Sequencing reads were assigned to 50 kilobase (kb) non-overlapping segments of the human reference genome. The normalized read counts from each segment of the genome are shown with alternating colors delineating chromosomes. A LOESS regression was performed for each chromosome (white lines). Deviations above and below the median value (dashed black line) indicate amplifications and deletions, respectively. B) GIN values for each collected time point relative to treatment initiation. C) Relative change in CNA-specific z-scores over time, consistent with clonal evolution. Each line corresponds with the genomic locus described by like colored “*” in panel A. Note that the region on Chromosome 9 denoted with the purple * appears in all samples and reflects the GIN; however, the region on Chromosome 14 (green *) decreases over time while the region on Chromosome 12 (pink *) increases in the same samples. Collectively, these data are consistent with clonal selection in the tumor measured in the cfDNA. Figure 1. GIN values for training and cancer samples included in this study. GIN values for 27,754 samples (green) submitted for noninvasive prenatal testing (NIPT) for which no CNAs were detected using NIPT algorithms were used to identify a threshold. Using a threshold of GIN=170, specificity among the 27,754 samples without known cancer was 99.7%. Using this same threshold, 4/12 (33%) samples at 1% tumor DNA and all samples (60/60) with 2% or greater tumor DNA were detected. Figure 3. A: Progression-free survival (PFS) for 44 patients with sufficient data for analysis. The median PFS (range) was 2.8 (0.1-18.4+) months. Calculated using the method of Kaplan and Meier. B: Kaplan and Meier PFS for GIN-predicted responders (N=7; blue line) versus non- responders (N=10; red line). GIN prediction based on cfDNA profiles at ~6 weeks compared to baseline. Median PFS for GIN-predicted responders versus non- responders=12.0 versus 2.05 months (P=0.001); Hazard ratio (HR) (95% confidence interval (CI)) for PFS for GIN-predicted progressors versus responders was 5.74 (1.9 to 17.7) (p=0.001). Figure 2. Representative case studies for distinct clinical responses to immunotherapy. Collectively, these data suggest that it may be possible to differentiate response as well as pseudoprogression and hyperprogression from progressive disease. Top, Left) Patient 40 demonstrated a clinical response with CT scan from the first scan (A) and day +356 (B) shown. C) GIN for all samples collected across this study for Patient 40 shown. Each measured cfDNA sample is indicated by an open circle. Green rectangle is representative of treatment window. Top, Right) Patient 30 demonstrated progressive disease with CT scans from the first scan (D) and day +177 (E) shown. F) GIN for all samples collected across this study for Patient 30 shown. Each measured cfDNA sample is indicated by an open circle. Green and red rectangles are representative of treatment windows. Bottom, Left) Patient 40B demonstrated accelerated progression or hyperprogressive disease (HPD) with PET scans from the initial scan (G) and day +57 (H) shown. I) GIN for all samples collected across this study for Patient 40B shown. Each measured cfDNA sample is indicated by an open circle. Green rectangle is representative of treatment window. Bottom, Right) Patient 125 demonstrated pseudoprogression with CT scans from the initial scan (J), day +14 (K), and day +55 (L) shown. M) GIN for all samples collected across this study for Patient 125 shown. Each measured cfDNA sample is indicated by an open circle. Green rectangle is representative of treatment window. Abbreviations: CT CAP=computerized tomographic scan of chest, abdomen and pelvis 0 100 200 300 0 2000 4000 6000 8000 Day Relative to Immunotherapy Start GIN Pembrolizumab CT CT 0 50 100 150 200 0 500 1000 1500 2000 2500 3000 3500 Day Relative to Immunotherapy Start GIN Nivolumab CT 0 20 40 60 0 500 1000 1500 2000 2500 3000 Day Relative to Immunotherapy Start GIN PET Atezolizumab PET 0 10 20 30 40 50 60 0 1000 2000 3000 4000 5000 Day Relative to Immunotherapy Start GIN CT CAP P embrolizumab CT CAP C I F M Patient 40 (Complete Remission) Patient 30 (Progressive Disease) Patient 40B (Hyperprogression) Patient 125 (Partial Remission after Pseudoprogression Prenatal (n=27,754) 1 percent 2 percent 3 percent 5 percent 10 percent 25 percent 0 1000 2000 3000 Genome Instability Number (GIN)
