A Phase I/II Study of Veliparib (ABT-888) in Combination with 5-Fluorouracil and Oxaliplatin in Patients with Metastatic Pancreatic Cancer

This article is featured in Highlights of This Issue, p. 5053

Pancreatic ductal adenocarcinoma (PDAC) is a deadly disease which, with a 5-year survival of only 10%, is poised to become the second leading cause of cancer-related death in the United States by 2030 (1). Treatment for metastatic PDAC (mPDAC) has improved, but the median overall survival (OS) remains less than 1 year (2, 3). However, for the subgroup of approximately 17% of patients with PDAC whose tumors harbor defects in the homologous recombination-DNA damage response and repair (HR-DDR) pathway (4–7), such as BRCA1/2 and PALB2 mutations, preliminary data suggest that treatment with platinum-based chemotherapy and/or PARP inhibitors offers significant improvements in patient outcomes (8–14). In fact, based on the phase III POLO trial, PARP inhibitors are now an FDA-approved standard treatment option as maintenance therapy for patients with mPDAC who harbor pathogenic germline BRCA1/2 mutations (15). PARP inhibitors have also been demonstrated to be active in patients with PDAC whose tumors harbor pathogenic PALB2 mutations (16). There is also evidence in other cancer types of the efficacy of PARP inhibitors in tumors with other (non-BRCA1/2/PALB2) HR-DDR mutations (17).

Moreover, there exists the potential for enhanced benefit when a PARP inhibitor is combined with DNA damaging chemotherapy. PARP plays a critical role in facilitating the repair of both single- and double-stranded DNA breaks (18–20), and PARP inhibition results in less efficient DNA repair. Thus, PARP inhibitors act as sensitizing agents for DNA-damaging chemotherapies (21). Veliparib (ABT-888, Abbvie) is a PARP inhibitor that has proven in vivo activity (22), and increases tumor cell sensitivity to chemotherapy and radiation (23). In humans, veliparib is safe and demonstrates inhibition of PARP activity in tumor biopsies (22). In phase III trials, veliparib has been used in combination with platinum-based chemotherapy at a dose of 150 mg orally twice a day; and at 400 mg twice a day as maintenance therapy without chemotherapy (24).

We previously demonstrated in PDAC cells that genetically disrupting the DNA-binding domain of PARP1, or treating with veliparib synergizes with oxaliplatin (21), particularly in HR-DDR–deficient cell lines. Therefore, we designed a trial combining veliparib with FOLFOX-based chemotherapy for patients with mPDAC. At the time of study design, FOLFIRINOX was not yet standard-of-care, and may present toxicity challenges in combination with a PARP inhibitor due to overlapping myelosuppression. Thus, herein, we present a phase I/II clinical trial of veliparib plus 5-fluorouracil (5FU) and oxaliplatin (FOLFOX) for patients with mPDAC.

Patients with mPDAC with measurable disease [as per RECIST 1.1 (25)] were eligible. Patients were ages ≥18 years, had an Eastern Cooperative Oncology Group performance status score of ≤2, and had adequate organ and bone marrow function [hemoglobin ≥9.5 g/dL, absolute neutrophil count ≥1.5 × 10⁹/L, platelet count ≥75 × 10⁹/L, serum creatinine level <1.5 mg/dL, bilirubin level ≤2.5 × upper limit of normal (ULN), and ALT/AST levels ≤3 × ULN]. For the phase I portion of the study, patients were not selected on the basis of prior therapy, family history (FH), nor germline or tumor HR-DDR mutational status.

For the phase II portion, there were two cohorts. Patients in the untreated cohort had not had any prior systemic therapy for mPDAC, though adjuvant chemotherapy was allowed if completed >6 months prior; and prior palliative radiation to the primary mass was allowable. Patients in the previously treated cohort may have had any number of prior therapies, including platinum-based regimens such as FOLFIRINOX, FOLFOX, or gemcitabine and cisplatin, but they may not have received a prior PARP inhibitor. Also, for the phase II portion (both cohorts), we preselected patients who either had a known pathogenic germline or somatic mutation (though testing was not required) in one of the HR-DDR genes [e.g., BRCA1/2, PALB2, ATM (expanded list in the Supplementary Materials)]; and/or patients who had an FH suggestive of a breast or ovarian cancer syndrome, as detailed in the NCCN guidelines (summarized in the Supplementary Materials; ref. 26). For the post hoc analyses detailed in the Results, outcomes for different patient subgroups were compared. Subgroups included: (i) prior platinum exposure: patients were placed into one of three categories as to whether their disease had:

• Never been exposed to prior platinum.

• Been exposed, but not progressed on prior platinum (defined as no disease progression within 3 months of stopping platinum).

• Progressed while on prior platinum.

(ii) the presence or absence of an FH (as defined previously); (iii) the presence of a mutation in an HR-DDR gene; patients who were tested but did not harbor a pathogenic germline or somatic HR-DDR mutation were also identified and highlighted. The study protocol, amendments, and informed consent forms were approved by the Institutional Review Board at Georgetown University (Washington, DC). Investigators obtained written informed consent from each participant or participant's guardian prior to screening. The research was conducted in accordance with recognized ethical guidelines including the Declaration of Helsinki, CIOMS, Belmont Report, and U.S. Common Rule, as described during training in Good Clinical Practice guidelines (CITI Training).

This was a single-center, phase I/II, open-label study. Initially, the dose and schedule for modified FOLFOX6 (27) were used (5FU bolus 400 mg/m², day 1; leucovorin 400 mg/m², day 1; FOLFOX 85 mg/m², day 1; and 5FU 2,400 mg/m² continuous infusion over 46 hours, days 1–3). Each cycle was 14 days. However, the first 6 patients dosed at 40 mg of veliparib demonstrated prolonged grade 2 or 3 myelosuppression. Thus, the 5FU bolus was dropped for all subsequent patients. For the phase I portion, the dose of veliparib was escalated in a standard 3+3 design from 40 to 60, 80, 100, 150, 200, and 250 mg twice a day, days 1–7 (Supplementary Table S1). As the primary endpoint of the phase I portion was to determine the recommended phase II dose (RP2D) as well as the maximally tolerated dose (MTD), dose-limiting toxicities (DLTs) were defined as any of the following that were definitely, possibly, or probably related to therapy (veliparib + FOLFOX) that occurred during the first cycle of therapy:

• (i) Grade 4 neutropenia lasting greater than 5 days or complicated by fever or infection.

• (ii) Grade 4 anemia or thrombocytopenia.

• (iii) Grade 3 thrombocytopenia associated with bleeding for which a transfusion was required.

• (iv) Grade 3 or 4 nonhematologic toxicity not manageable with routine supportive care (e.g., over-the-counter antidiarrheals for diarrhea).

• (v) Any toxicity, regardless of grade, which resulted in withholding of therapy for greater than 3 weeks.

For the phase II portion, the RP2D was 200 mg of veliparib, although the protocol did allow for stepwise deescalation to 150 mg and then to 100 mg for toxicity after the first cycle. Safety assessments were performed every 2 weeks for the first 4 cycles, then every 4 weeks thereafter. Tumor response was assessed radiographically every 8 to 12 weeks using RECIST 1.1 (25). Study treatment was continued without interruption in the absence of unacceptable toxicity or progressive disease. The protocol did allow for patients to stop the oxaliplatin for persistent neuropathy, at which time they were maintained on the RP2D of veliparib plus the standard dose of continuous infusion 5FU.

For the phase I portion, plasma samples were obtained according to the schedule in Supplementary Fig. S1 for pharmacokinetic (PK) assessment of veliparib performed at Abbvie. Results were compared with historical controls to identify any effect on veliparib PKs by FOLFOX. In addition, when samples were available, next-generation sequencing of cancer-related genes was performed commercially by Foundation Medicine, Caris Life Sciences, or Tempus laboratories on patient tumor samples. Several patients also had germline testing through Myriad or Ambry. Patients were defined as harboring HR-DDR mutations if a pathogenic mutation in an HR-DDR gene was identified in a blood sample (germline) or tumor sample (somatic). Pathogenic mutations for each patient are included in Supplementary Table S4. The study team centrally validated pathogenic mutations in the ClinVar and Cosmic databases.

The primary objective of the phase I portion was to determine the RP2D, with the primary endpoint being adverse events (AEs), as measured by NIH CTC version 4.03. The efficacy assessments included the objective response rate (ORR), disease control rate (DCR; stable disease through at least four cycles, partial response, or complete response), progression-free survival (PFS), and OS. For the phase II portion, the primary endpoint was the ORR, and each phase II cohort followed a Simon two-stage optimal design (28). For each cohort, nine patients were accrued in the first stage, and if ≥1 patient demonstrated a response, 15 additional patients were to be accrued in the second stage for a total of 24 patients. If ≥3 patients among 24 patients demonstrated a response, then the treatment was considered sufficiently promising to warrant further testing. At the time this protocol was designed, the ORR for first-line standard-of-care gemcitabine was only 7% (29), and there was no standard second-line therapy. Each cohort was designed to differentiate a 5% ORR from a 25% ORR at a one-sided 10% significance level with 90% power. Patient characteristics, medical features at study entry, and AEs at least possibly related to study therapy were tabulated. Differences in ORR and DCR among subgroups were compared using Fisher exact tests. OS was defined as the number of months from enrollment until death or last contact. Patients who were alive at the time of analysis were censored at their last contact. PFS was defined as the number of months from enrollment to progression or death, whichever occurred first. Patients who were alive and progression-free at the time of analysis were censored at their last tumor assessment time (as of May 3, 2019). The Kaplan–Meier methodology was used for analyzing OS and PFS. Analyses were performed in SAS software Version 9.4 (SAS Institute Inc.) and figures were created using STATA 12.1 (StatCorp LP).

The trial opened in January 2011 and the phase I portion accrued rapidly. However, when we restricted enrollment in the phase II portion, as defined above, the accrual rate slowed considerably, and the trial was closed to enrollment in 2019 due to slow accrual, although, as discussed below, the primary endpoint had been met in the two phase II cohorts.

Between January 2011 and December 2018, 75 patients were consented and 64 patients initiated treatment. Figure 1 depicts the screen failure and enrollment into the different cohorts. Thirty-one patients initiated treatment in the phase I portion; 15 patients initiated treatment in the untreated phase II cohort; and 18 patients initiated treatment in the previously treated phase II cohort. Patient characteristics are listed in Table 1. The median age for all 64 treated patients was 64 years (range, 40–84); most patients had an ECOG score of 0 or 1 (95%); and 56% of patients were male. In the phase I, and previously treated phase II cohort, patients had a median of 1 and 2 lines of prior therapy, respectively (range, 1–7).

For the phase I portion, 27/31 patients were evaluable for DLTs, with four patients withdrawing consent after one cycle (not for toxicity). In the 40-mg cohort, three of six patients required significant (>2 week) treatment delays for grade 2 or 3 myelosuppression, but the only protocol-defined DLT was a treatment delay of >3 weeks. Therefore, the protocol was amended to drop the 5FU bolus. Only one other DLT occurred, which was at the 250-mg cohort. Four of the six patients at 250 mg experienced significant grade 3 or 4 myelosuppression. Therefore, 200 mg was selected as the RP2D, though the protocol allowed for stepwise deescalation to 150 mg and then to 100 mg for toxicity after the first cycle.

PK samples were available for 14 patients in five dosing cohorts. The veliparib PK data suggested that coadministration of FOLFOX had no apparent impact on veliparib PKs (Supplementary Tables S2 and S3).

All 64 patients who received treatment were evaluable for AEs. Table 2 provides the number of patients experiencing AEs by category and cohort that are at least possibly related to treatment. No grade 5 events occurred. Overall, the combination of veliparib and FOLFOX was well tolerated, although minor treatment adjustments were required over the course of therapy for most patients. The primary toxicity of concern was myelosuppression, and 16% of patients experienced grade 3 or 4 neutropenia. One patient experienced grade 3 thrombocytopenia, and two patients had grade 3 or 4 anemia. Fifty-two and 64%, respectively, of patients experienced mild fatigue and nausea, but only 2% and 6%, respectively, were grade 3 or 4. Other notable grade 3 or 4 nonhematologic AEs included one patient each with a rash, diarrhea, and peripheral neuropathy. All patients required a reduction in the dose of veliparib, 5FU, and/or FOLFOX at some point in the trial, either due to myelosuppression or nausea prior to the first restaging imaging, or beyond four cycles typically for neuropathy.

Of the 64 patients who received study treatment, six patients in the phase I portion came off study prior to response evaluation (for reasons other than progression of disease), and thus were not evaluable for response. Table 3 presents the responses and survival times for the 58 response-evaluable patients. For the entire group, the ORR was 26%, which included 11 PRs and four CRs. The waterfall plot in Fig. 2A demonstrates the responses for each patient. The DCR, PFS, and OS were 52%, 4.0 months, and 7.8 months, respectively. The swimmer's plot in Fig. 2B demonstrates the treatment duration for each patient. The protocol did allow for patients to stop the oxaliplatin for persistent neuropathy, at which time patients were maintained on the RP2D of veliparib plus the standard dose of continuous infusion 5FU (n = 18, none of whom received more than 12 cycles of FOLFOX). One patient (patient 44) was granted an exception and was allowed to discontinue the 5FU as well, after having remained in a complete response for nearly 4 years.

For the phase I portion (n = 25), patients were not preselected on the basis of FH and known HR-DDR mutational status, and the ORR was 20%. For the two phase II cohorts combined (n = 33), with patients selected on the basis of FH or HR-DDR mutational status, the ORR was 31%. The ORR was 40% for patients who received no prior therapy (n = 15), with a DCR, PFS, and OS of 87%, 6.5 months, and 13.0 months, respectively. The ORR was 22% for the previously treated patients (N = 18), with a DCR, PFS, and OS of 28%, 1.6 months, and 4.5 months, respectively. Although the study was closed early due to slow accrual, the primary endpoint had been met in each of the two phase II cohorts.

We examined efficacy in several patient subgroups (Table 3; Fig. 3). There is evidence to suggest that the mechanisms of resistance to platinums overlap the mechanisms of resistance to PARP inhibitors. Correspondingly, the ORR for patients who received prior platinum was only 7%, with a DCR of only 14% (N = 14). This is compared with an ORR of 32% and a DCR of 64% for those who did not receive prior platinum (N = 44; Table 3; Fig. 3A and B). We further assessed outcomes for patients who had been previously exposed to platinums (Table 3; patient numbers were too small for corresponding Kaplan–Meier curves). Prior exposure to platinum was broken down into exposure, but nonprogression on prior platinum (as defined in the Patients and Methods); and frank progression while on prior platinum. The ORR for patients who had been exposed to, but not progressed on prior platinum was 25%, with a DCR of 50% (N = 4). However, none of the patients whose disease had progressed on prior platinum had a response, nor any reduction in tumor burden (Fig. 2A), nor even disease control with the combination of veliparib and FOLFOX (N = 10). Of note, patient 61 received neoadjuvant FOLFIRINOX for nonmetastatic disease and then was treated on study but did not meet the definition of prior platinum exposure in the metastatic setting. Unfortunately, her disease progressed on the combination of veliparib and FOLFOX.

Forty-three patients had a positive FH and the ORR for these patients was 30%, with a DCR of 53% (N = 43). This is compared with an ORR of 13% and a DCR of 47% for those with no FH (N = 15; Table 3; Fig. 3C and D).

Finally, 29 of 58 evaluable patients had NGS testing for HR-DDR mutations. Of these, 14 patients had a BRCA1/2 mutation; two patients had an ATM mutation; one patient had a PALB2 mutation; and one patient had a FANCG mutation (Supplementary Table S4). The ORR of the HR-DDR–mutated patients was 44%, with a DCR of 61% (N = 18). The highest ORR overall was observed in HR-DDR–mutated patients who had not previously received platinum at 57% with a DCR of 79% which was irrespective of the line of therapy (N = 14; Table 3; Fig. 3E and F). As visualized in Fig. 2, all of the responses were observed in patients with germline or somatic BRCA1/2 or PALB2 mutations. The one patient with a FANCG mutation experienced stable disease (PFS of 4.7, and ongoing at the time of data cut-off); but neither of the patients with ATM mutations (both of whom were also platinum-naïve) achieved even stable disease.

Patients with mPDAC are in desperate need of additional therapies. The modern chemotherapy regimens of FOLFIRNOX and gemcitabine + nab-paclitaxel have improved outcomes, but response rates are only 31% and 23%, respectively (2, 3). However, multiple sequencing efforts have revealed that a significant portion of PDACs harbor HR-DDR gene mutations, most commonly BRCA1/2 and ATM. Emerging data have revealed that patients with HR-DDR–mutated mPDAC, specifically those with BRCA1/2 or PALB2 mutations can respond to PARP inhibitors, and even CRs can be achieved (8–13, 15). We and others have published that in patients with pancreatic cancer whose tumors harbor HR-DDR mutations (ATM, ATRX, BAP1, BRIP1, FANCG, RAD50/51, etc.) have an improved survival when treated with platinum-based therapy (14, 30). In addition, there exist preclinical published data suggesting a benefit to PARP inhibitors, or PARP inhibitor combinations in pancreatic cancer models, including cell lines and patient-derived models of tumors with other HR-DDR mutations (31). While we have had personal communication with colleagues who have shared anecdotal stories of patients with other HR-DDR mutations benefitting from off-label PARP inhibitor therapy, there have not as yet been any published case reports of PARP inhibitor activity in such patients with pancreatic cancer.

In patients with pancreatic cancer with germline or somatic BRCA1/2/PALB2 mutations, the activity of single-agent PARP inhibitors has been limited, ranging from 16% to 21% (8, 13). Shroff and colleagues detailed that, similar to our findings, the majority of the responders were patients whose disease had not progressed on prior platinum (13). Yet, there may be enhanced benefit with the combination of a PARP inhibitor and a platinum, and preliminary results in a phase Ib trial from O'Reilly and colleagues demonstrated a remarkable 78% ORR, and 23 months OS with the combination of gemcitabine, cisplatin, and veliparib in patients with PDAC harboring germline BRCA1/2 mutations (11).

We present here the results of our phase I/II trial of FOLFOX + veliparib, and we demonstrate that patients with mPDAC can respond to this combination (ORR = 26%). For the two phase II cohorts, we chose to close the study early due to slow accrual, but we did meet our protocol-defined primary endpoint, with at least 3 responders in each cohort. In fact, with 40% of untreated and 22% of previously treated patients achieving a PR or CR, this combination is worthy of further investigation. When patients were selected on the basis of FH and/or the presence of HR-DDR mutations, the ORR increased to 45%, and the highest responses were observed in patients whose tumors harbored HR-DDR mutations, and who had not been exposed to prior platinum (ORR = 57%, N = 14). Our results also highlight the fact that patients whose disease has progressed on prior platinum are very unlikely to benefit from a PARP inhibitor combination. In our study, none of the 10 patients who had experienced disease progression on prior platinum had even a minor reduction in tumor burden (Fig. 2), and the ORR and DCR for these patients was 0% (Table 3). Our results are in line with the detailed results provided in the rucaparib study by Shroff and colleagues (13), and emphasize the notion that future trials of PARP inhibitors for patients with PDAC should exclude patients whose disease has progressed while on prior platinum. The combination was well tolerated, and, for the patients with long-term (>4–6 months) disease control, a strategy of “maintenance” therapy without the FOLFOX was able to maintain disease control for a prolonged period, as demonstrated by the swimmer's plot (Fig. 2B).

What is uncertain is the contribution of the veliparib to FOLFOX. It is possible that patients with HR-DDR mutations would respond to FOLFOX alone. However, two patients whose disease was only stabilized with prior FOLFIRINOX (i.e., no reduction in tumor burden), did have a response to FOLFOX + veliparib. Similarly, Shroff and colleagues highlighted two patients who did not respond to platinum, but did respond to rucaparib (13). Yet, the results of the recently reported randomized phase II trial of gemcitabine + cisplatin ± veliparib in patients with BRCA1/2 or PALB2 mutations failed to demonstrate any benefit to the addition of veliparib (32). In this trial, patients treated with gemcitabine + cisplatin alone had an ORR of 65%, and a median OS of 16.4 months, which was not different in the group of patients who received veliparib as well. While veliparib was not of any benefit when added to gemcitabine and cisplatin, the PARP inhibitor may play a different role when combined with FOLFOX -based chemotherapy, wherein patients can only remain on the platinum for approximately 4–6 months. In our study, 18 patients remained on therapy for >6 months, all of whom had gone off the oxaliplatin, due to neuropathy. Thus, there may be a role for transitioning patients off FOLFOX while maintaining the PARP inhibitor as long-term maintenance therapy, as was done in the POLO trial with olaparib (33). Such a strategy appeared effective in the phase III trial of carboplatin and paclitaxel ± veliparib in ovarian cancer (24). A similar trial of induction FOLFOX or even FOLFIRINOX, followed by maintenance veliparib versus, for example, maintenance 5FU (or oral capecitabine), would be an important study, and would address an important concern (the placebo control arm) of the POLO trial.

Furthermore, there is an ongoing debate on the functional role of PARP inhibition in the treatment of HR-DDR–mutated cancers. Originally, PARP inhibitors were demonstrated to inhibit the catalytic activity of the PARP-1 enzyme, thus inhibiting single-strand repair, particularly after cotreatment with a DNA-damaging chemotherapy (34). This mechanism was the foundation for exploring the combination of veliparib and chemotherapy (23). However, a second critical role of some PARP inhibitors involves the trapping of the PARP enzyme at the site of DNA damage (35). The trapped PARP enzyme complex results in replication fork arrest, leading to mitotic catastrophe and apoptotic cell death. Several PARP inhibitors such as olaparib, niraparib, rucaparib, and talazoparib can achieve PARP trapping, and thus are active as single agents (8, 13). Veliparib at tolerable doses can only achieve catalytic inhibition of the PARP enzyme. While veliparib may not be effective as a single agent, the tradeoff may be that the limited spectrum of activity of veliparib may also allow for the safe combination with DNA-damaging agents, such as radiation and chemotherapy. In contrast, many of the PARP-trapping inhibitors have been too toxic to use in combination with DNA-damaging chemotherapies (36–38). Ultimately, a trial that compares, for example, FOLFOX + veliparib with a PARP trapping inhibitor would have to be done to compare outcomes of the two DNA damage targeting strategies.

Finally, our study emphasizes the need to identify patients whose disease harbors HR-DDR mutations. The NCCN has recently recommended that all patients with PDAC undergo germline testing for the presence of an inherited mutation. In addition, there has been more widespread testing for somatic mutations. Thus, we are optimistic that more patients will be identified, and a strategy to steer such patients toward PARP inhibitor–based clinical trials will be critical, and may lead to the proof that we can improve outcomes for this subgroup of patients.

M.J. Pishvaian reports grants from AbbVie (institutional support only) during the conduct of the study; grants, personal fees, and nonfinancial support from Merck (consultant/advisor, institution), AstraZeneca/Medimmune (consultant/advisor, institution), Halozyme (consultant/advisor, institution); personal fees and nonfinancial support from Sirtex Medical (consultant/advisor, speaker), Caris Life Sciences (consultant/advisor, speaker); personal fees from Rafael (consultant/advisor), RenovoRx (consultant/advisor); personal fees, nonfinancial support, and other from Perthera (consultant/advisor/stock); grants from Bavarian Nordic (institutional support only), Celldex (institutional support only), Pfizer (institutional support only), Novartis (institutional support only), Boston Biomedical (institutional support only), Tesaro (institutional support only), Bristol-Myers Squibb (institutional support only), Genentech (institutional support only), ARMO Biosciences (institutional support only), Bayer (institutional support only), Calithera (institutional support only), Curegenix (institutional support only), Fibrogen (institutional support only), Gilead (institutional support only), GlaxoSmithKline (institutional support only), Karyopharm (institutional support only), Regeneron (institutional support only), Pharmacyclics (institutional support only); grants and personal fees from Celgene (consultant/advisor, institution); and personal fees from Foundation Medicine (consultant/advisor) outside the submitted work; and is listed as a coinventor on a pending patent regarding the use of FOLFOX + veliparib in pancreatic cancer. J.J. Hwang reports personal fees from Genentech/Roche (speaker bureau/consultant), Bristol-Myers Squibb (speaker bureau/consultant), Amgen (speaker bureau/consultant), Ipsen (speaker bureau/consultant), Celgene (consultant), Deciphera (consultant), Eisai (consultant), Bayer (consultant), Taiho (consultant), and Caris (consultant) outside the submitted work. B.A. Weinberg reports personal fees from Lilly, Bayer, Sirtex, and Taiho outside the submitted work. L.M. Weiner reports personal fees from Celldex Therapeutics, Cytomx Therapeutics, Jounce Therapeutics, Immunome, Klus Pharmaceuticals, Tessa Therapeutics, and Samyang Biopharm USA; other from Targeted Diagnostics & Therapeutics (shareholder), grants from Bioexcel Therapeutics outside the submitted work; and performs occasional consulting on patent cases in areas unrelated to the subject of this paper. J.L. Marshall reports personal fees from Caris Life Sciences (interim CMO), Indivumed (interim CMO), Amgen, Taiho, Merck, and Bayer outside the submitted work. J.R. Brody is listed as a coinventor on a pending patent regarding methods of treating gastrointestinal cancers and tumors thereof using combination therapy. No potential conflicts of interest were disclosed by the other authors.

M.J. Pishvaian: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing. H. Wang: Data curation, formal analysis, validation, methodology, writing-review and editing. A.R. He: Resources, investigation, writing-review and editing. J.J. Hwang: Resources, investigation, writing-review and editing. B.G. Smaglo: Data curation, investigation, writing-review and editing. S.S. Kim: Data curation, investigation, writing-review and editing. B.A. Weinberg: Data curation, investigation, writing-review and editing. L.M. Weiner: Data curation, investigation, writing-review and editing. J.L. Marshall: Resources, data curation, funding acquisition, investigation, writing-review and editing. J.R. Brody: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing-original draft, writing-review and editing.

We would like to thank Meeta Jaiswal, PhD for scientific guidance throughout protocol development and implementation. This work was funded by the Otto J. Ruesch Center for the Cure of GI Cancers, Lombardi Comprehensive Cancer Center. Abbvie, Inc. has provided veliparib and partial research funding for the correlative science. M.J. Pishvaian and J.R. Brody are supported by 1R01CA212600-01 (NCI, NIH); a 2015 Pancreatic Cancer Action Network American Association for Cancer Research Acceleration Network Grant (15-90-25-BROD); and M.J. Pishvaian and J.R. Brody are also supported in part by U01CA224012 (NCI, NIH). J.R. Brody was supported in part by the NCI of the NIH Award Number P30CA056036 SKCC Core Grant (Thomas Jefferson University). We also would like to acknowledge The Sarah Parvin Foundation for their generous support.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.