Trial design and patient recruitment
PA.7 received Institutional Ethics Board approval and was conducted in accordance with the Declaration of Helsinki and the International Ethical Guidelines for Biomedical Research Involving Human Subjects (NCT02879318). The study protocol was approved by the REBs of participating centres, the Ontario Cancer Research Ethics Board and the UBC BC Cancer Research Ethics Board. Patients provided written informed consent. PA.7 was a phase II randomized clinical trial conducted at 28 centers in Canada. After a safety run-in that included 11 patients, 180 patients diagnosed with metastatic pancreatic ductal adenocarcinoma (mPDAC) were included in the study. The inclusion criteria for the trial were: patients must have histologically or cytologically confirmed PDAC that is metastatic, must have the presence of measurable or evaluable disease as defined by the response endpoints in the solid tumors (RECIST 1.1), must be considered suitable and capable candidates to receive first-line chemotherapy for metastatic disease with gemcitabine and nab-paclitaxel, must consent to the provision of a formalin-fixed paraffin block of tumor tissue and blood, serum and plasma specimens, must be at least 18 years of age with an ECOG performance status of zero or one with a life expectancy of at least 12 weeks, must not have received treatment anterior for metastatic disease, must have adequate normal organ and marrow function, have imaging including CT/MRI of the chest/abdomen/pelvis for document disease sites within 28 days of randomization, must be able to complete quality of life questionnaires, and must be accessible for treatment and follow-up. Exclusion criteria included: history of other malignancies, prior treatment with a PD1 or PD-L1 inhibitor, history of primary immunodeficiency, documented active or prior autoimmune or inflammatory disorders, and disease active or uncontrolled intercurrent. Patients were randomized to receive gemcitabine (1000 mg/m2 D1, 8, 15), nab-paclitaxel (125 mg/m2 D1, 8, 15), durvalumab (1500 mg D1 q 28 days) and tremelimumab (75 mg D1 for the first 4 cycles) versus gemcitabine and nab-paclitaxel alone in a 2:1 ratio (respectively). Randomization was dynamically balanced by ECOG performance status (0 versus 1) and receipt of prior adjuvant therapy (yes versus no) using the method of minimization. Randomization was performed centrally by the central office of the Canadian Cancer Trials Group (CCTG). Overall survival (OS) was the primary endpoint. Secondary endpoints included progression-free survival (PFS), safety, overall response rate (ORR), and quality of life as assessed by EORTC QLQ-C30. Patient recruitment began on August 22, 2016. The randomized phase II arm of this study was opened on April 10, 2017 after reviewing the results of 11 patients in the run-in phase and ended on July 28, 2018 after randomization of the last patient. . The CCTG’s Data and Safety Monitoring Committee regularly assessed the conduct and safety of the study.
cfDNA extraction, library preparation, probe capture and sequencing
Circulating cell-free DNA was extracted by the QIAamp Circulating Nucleic Acid Kit from plasma samples. The quantity and quality of purified cfDNA was checked using the Qubit Fluorometer and the Bioanalyzer 2100. For samples with severe genomic contamination from peripheral blood cells, bead-based size selection was performed to eliminate large genomic fragments. Five to 30 ng of extracted cfDNA was subjected to library construction, including dA-tailing end repair and adapter ligation. Library fragments ligated with appropriate adapters were amplified by PCR. The amplified DNA libraries were then verified using the Bioanalyzer 2100 and samples with sufficient yield are subjected to hybrid capture.
The PredicineATLAS with 600 genesMT A panel with biotin-labeled DNA probes was used for target enrichment. Briefly, the library was hybridized overnight with a Predicine NGS panel and paramagnetic beads. Unbound fragments were washed and enriched fragments were amplified via PCR amplifications. Similar to library preparation, purified product was verified on Bioanalyzer 2100 and then loaded into Illumina NovaSeq 6000 for NGS sequencing with 2x150bp paired sequencing kits.
Analyzes of cfDNA NGS data
NGS data was analyzed using the Predicine DeepSea NGS analysis pipeline, which starts from raw sequencing data (BCL files) and produces the final mutation calls. In short, the pipeline first performs adapter slicing, barcode verification, and remediation. Cleaned paired FASTQ files are aligned to the hg19 human reference genome construct using the BWA alignment tool. Consensus BAM files are then derived by merging paired reads from the same molecules (based on mapping location and unique molecular identifiers) as single-stranded fragments. Single-stranded fragments from the same double-stranded DNA molecules were then fused together as a double-strand. Using the error suppression method described by Newman and colleagues33sequencing and PCR errors were mostly corrected during this process.
Candidate variants were called by comparison with the local variant background (defined based on plasma samples from healthy donors and historical data). The variants were then filtered by the log-odds (LOD) threshold34base and mapping quality thresholds, repeated regions, and other quality metrics.
bTMB score estimation
Blood-based tumor mutational load (bTMB) was defined as the number of somatic single-nucleotide variants (SNVs), including synonymous and non-synonymous variants in the target regions of the panel. The bTMB score was then normalized by the total effective target panel size in the coding region35. Since there were no paired normal samples, such as peripheral blood mononuclear cells (PBMC), available for germline variant screening, germline variants were inferred based on variant annotation, frequency of variant alleles and other variant information such as variant copy number status. Variants of common clonal hematopoietic mutations of genes of undetermined potential (CHIP) (DNMT3A, TET2, ASXL1 and JAK2) were excluded from the bTMB estimate.
Estimation of the circulating tumor fraction
Circulating tumor fractions were estimated based on allelic fractions of autosomal somatic mutations described previously.36. Briefly, the mutant allele moiety (MAF) and the ctDNA moiety are linked as MAF = (ctDNA*1)/[(1-ctDNA)*2 + ctDNA *1], and therefore ctDNA = 2/((1 / MAF) + 1). Somatic mutations in genes with a detectable copy number change were omitted from the estimation of the circulating tumor fraction.
For somatic and germline variants, frameshifts, and inframe insertions/deletions (indels), missense and nonsense mutations were included in our analysis. For secondary survival analysis of kras wild-type samples, CTF was used as a metric to conservatively filter the samples. To determine the CTF threshold by which to exclude samples, we iteratively tested increasing CTF thresholds. At each iteration, samples with CTF values below the threshold were excluded and the frequency of kras wild type status has been recalculated. A kras a wild-type frequency of approximately 10%, while retaining a large proportion of patients, was reached at CTF = 1.68%, and this value was therefore chosen as the threshold for excluding samples in the secondary. kras wild-type survival analysis as the rate of kras wild-type status was more similar to previous studies12,13,14.
The primary OS endpoint was defined as the time from date of randomization to date of death from any cause. The PFS secondary endpoint was defined as the time from date of randomization to the first date that disease progression was objectively documented or date of death from any cause and ORR as the proportion of randomized patients with a documented complete response or partial response. This study was designed to detect, with 80% power and a 10% bilateral level, an HR of 0.65 between two treatment arms, which corresponded to an increase in median OS of 8.5 months for the chemo arm at 13.1 months for the chemo+ICI arm and required the observation of 150 deaths before the final analysis.
OS and PFS were summarized by the Kaplan-Meier method and compared by a log-rank test stratified by ECOG performance status and prior adjuvant therapy. Stratified Cox models were used to calculate relative risks (HR) and 90% confidence intervals (CI). ORR between treatment groups was compared by a Cochran-Mantel-Haenszel test stratified by ECOG performance status and prior adjuvant therapy. Fisher’s exact test was used to compare adverse event rates between treatment arms as well as predefined primary quality of life endpoints as proportions of patients with deterioration (defined as a score of change from baseline of -10 points or less) in physical function and global health status at 8 weeks and 16 weeks after randomization. The two-tailed Wilcoxon test was used to compare CTF values between KRAS wild-type and mutant groups. Primary analyzes included all randomized patients, while ctDNA analysis included patients with successful sequencing of the ctDNA sample. SAS statistical software (version 9.0; SAS Institute, Inc) and R v3.6.3 were used for analyses.
Summary of reports
Further information on the research design can be found in the summary of nature research reports linked to this article.