Alternatives to Perpetual Chemotherapy for Metastatic Pancreatic Cancer

With the advent over the last decade of effective, multi-agent chemotherapy regimens for patients with pancreatic cancer (1), subgroups of those with metastatic disease who respond durably to treatment are emerging. In the last 15 years, the rate of long-term survivors of pancreatic cancer has nearly doubled from 5% to 10% largely because of these new treatments, such as gemcitabine/nab-paclitaxel or FOLFIRINOX. Between this development and a steadily rising incidence of the disease, a growing population of patients now stay on chemotherapy for many months or even years with sustained disease control. These gains are remarkable for a universally lethal illness, but such “perpetual chemotherapy” can be very difficult and unsustainable for patients. It fuels eventual therapy resistance and leads to cumulative toxicities that reliably degrade quality of life. In current clinical practice, chemotherapy is often modified in patients who have durable responses, and indeed there is early randomized clinical data to support maintenance chemotherapy (2). However, it is possible that novel, noncytotoxic maintenance strategies may result in superior outcomes with less toxicity, and these should therefore be pursued. Here, based on promising early results that leverage newfound insights regarding pancreatic cancer biology, we propose a new strategy that can be evaluated in clinical trials, namely, after a period of chemotherapy to reduce disease burden and stabilize the patient (“induction”), explore innovative “maintenance” therapies (Fig. 1) with a goal of preserving or extending clinical response without the cumulative toxicity of chemotherapy.

The POLO trial of the PARP inhibitor olaparib (NCT02184195) paved the way for the development of maintenance therapy for metastatic pancreatic cancer. Leveraging prior observations that patients with pathogenic germline variants (PV) in BRCA1 or BRCA2 (gBRCA) often demonstrate durable and deep responses to platinum-based chemotherapies (3, 4), the POLO trial, published in mid-2019, showed that following at least 16 weeks of induction therapy with platinum, maintenance treatment with olaparib compared with placebo doubled progression-free survival for patients with germline BRCA PVs (5). A point of discussion for the POLO study design was the placebo-control strategy; moreover, not all patients in POLO benefited from stopping chemotherapy. However, POLO definitively showed that a non-cytotoxic maintenance strategy was feasible and active, and olaparib was approved by the FDA for this indication in December of 2019. However, with a lack of survival benefit compared with placebo (6), the use of olaparib maintenance in clinical practice remains optional.

The concept of maintenance therapy for an aggressive, systemic cancer is controversial because the choice is still fraught. Stopping active chemotherapy in a patient with an ongoing response poses a risk of rapid cancer progression with no assurance that resumption of cytotoxic chemotherapy will reestablish control. This consideration was foremost when we designed two pancreatic cancer maintenance trials at Penn's Abramson Cancer Center (NCT03140670 and NCT03404960): we did not know if patients on active, effective chemotherapy would be willing to enroll onto these trials or if this maintenance strategy would be safe. In the RUCAPANC2 trial (NCT03140670), patients with germline or somatic PVs in BRCA1, BRCA2, or PALB2 and platinum-sensitive disease receive rucaparib monotherapy as maintenance therapy (7). In the PARPVAX trial (NCT03404960), patients with platinum-sensitive disease receive niraparib and either ipilimumab or nivolumab after chemotherapy is discontinued (8); a known mutation is not required for this second trial. With more than 100 patients enrolled in these two trials since 2017, we have learned that many patients welcome a break from cytotoxic drugs. Importantly, it has become clear that in the vast majority of cases, the risk of rapid or explosive disease progression off chemotherapy is minimal. With careful symptom monitoring and frequent imaging, it has been rare to miss the opportunity to safely and effectively re-start chemotherapy. Instead, the rate of ongoing and deepening clinical responses upon the switch to a PARP inhibitor—coupled with marked improvement in quality of life—has been promising, although results remain preliminary (7).

The POLO study demonstrated that PARP inhibitors are active against gBRCA-related pancreatic cancer, as had been previously suggested by findings from single arm studies (9, 10). Additional studies have demonstrated activity in a broader group of patients with pancreatic cancer, including those with PALB2 PVs and somatic BRCA PVs (7, 9). The POLO data, our data, and those of others supports the development of the next generation of novel maintenance clinical trials for those with BRCA and PALB2 PVs, including trials of PARP inhibitor combination therapies and trials in low-disease burden states, such as in the adjuvant setting. Several such studies are in active development and are anticipated to open in 2021. For example, the APOLLO Study (principal investigator: K.A. Reiss; ECOG-ACRIN) will evaluate the utility of PARP inhibition versus placebo following curative intent resection and completion of standard chemotherapy in patients with early-stage disease and BRCA or PALB2 PVs.

It is plausible that maintenance therapies could also be studied for a larger group of patients with pancreatic cancer than just those with BRCA and PALB2 PVs. A variety of DNA damage repair (DDR) gene variants are found in low frequencies across patients with pancreatic cancer. Because of their rarity, patients with various DDR PVs have been historically grouped together in clinical trials. However, it has become clear that there is wide variability in the clinical impact of these alterations, and as such, important efficacy signals are likely to be diluted and missed by this “lumping” strategy. For example, it is increasingly clear that patients with PVs in ATM and CHEK2 do not carry a DDR phenotype (11). In addition, patients without an identified PV on panel sequencing may nevertheless still have a “BRCAness” signature based on a functional genomic evaluation. In fact, recently published data estimate that 7% to 10% of sporadic PDAC has a DDR signature by HRDetect that may result in heightened platinum or PARPi sensitivity (11). Validated and clinically practical assays to reliably predict functional HRD within the pancreatic tumor bed are needed and an area of active development. The refinement of these assays will be a critical step as we develop innovative maintenance trials for patients with DDR alterations.

An alternative approach to identifying a patient population appropriate for maintenance therapy clinical trials uses a response-to-treatment phenotype. We and other clinical investigators are testing the efficacy of maintenance therapy in patients with metastatic pancreatic cancer who respond clinically to first-line chemotherapy (5, 8, 12–14). The PARPVAX trial, for example, does not require an identified mutation for enrollment. Instead, patient selection is based on a clinically observed durable platinum-sensitive response. Trials that identify patients via clinical phenotype could be further expanded to include additional signals associated with DDR, such as strong family history of BRCA-related cancers (15), in the hopes to refine the selection of patients. One of the key goals of trials that select solely on clinical phenotype will be to perform in-depth molecular assays to understand why certain patients might tumor control during maintenance therapy while others will not.

An immunotherapy approach in the maintenance setting is also of interest. Wu and colleagues reported a randomized phase II study of GM-CSF–transfected pancreatic tumor vaccine (GVAX) with ipilimumab versus continued chemotherapy in patients who had at least stable disease following ≥6 months of FOLFIRINOX therapy. The primary endpoint was an improvement in overall survival in the experimental compared with the chemotherapy arm. The study was closed early for lack of efficacy, but the investigators detected evidence for posttreatment systemic and intratumoral immune activation, providing rationale for further development of immune-based maintenance therapy. Such experimental approaches include mutant KRAS–targeted immunotherapy (NCT03592888, NCT04117087) or immune agonists, such as agonist CD40 antibodies with or without anti-PD-1 nivolumab which has shown promise in patients with metastatic disease (16). It should be noted that to evaluate the true benefit of a maintenance strategy with noncytotoxic agents, a control arm using cytotoxic agents might be needed, particularly in light of the data from the PANOPTIMOX trial (2).

The wide adoption of multiagent chemotherapy combinations for metastatic pancreatic cancer has become a double-edged sword. For the first time, we can reliably stop progression or trigger tumor regressions in 50% or more of patients and notably halt the rapid clinical deterioration and death that was previously a hallmark of this disease. On the other hand, our new standard practice has become to use this therapeutic tool indefinitely and continue chemotherapy to preserve response or disease stabilization even as toxicities mount and underlying biological resistance to therapy takes hold. Maintenance treatment using PARP inhibitors after a response to upfront chemotherapy, although only an option right now in a subset of patients, suggests that maintenance therapy is an alternative to be explored in greater depth. This hypothesis can be tested in novel clinical trials of new agents, which would need a design that embraces this clinical scenario. As examples, disease progression as an eligibility criterion would likely need to be foregone; the appropriate biological and clinical setting for any given maintenance therapy under investigation might be highly selective; and primary endpoints based on objective tumor response might be confounded by potentially ongoing responses to induction chemotherapy even after discontinuation. In each case, we and others are testing and finding appropriate alternative designs.

In conclusion, for most patients with metastatic pancreatic cancer, the standard approach is still life-long chemotherapy. However, evidence now supports further clinical investigation of nonchemotherapy in a maintenance setting. Olaparib for a small subset of patients is a good example of what might be possible. We suggest a change in mindset to study whether maintenance therapy, rather than perpetual chemotherapy, could be applied in the appropriate biological and clinical settings.

K.A. Reiss reports grants from Clovis Oncology, BMS, and GSK, as well as grants from Lilly Oncology outside the submitted work. R.H. Vonderheide reports a patent for Cellular Immunotherapy issued, licensed, and with royalties paid from Novartis; in addition, R.H. Vonderheide receives royalties from Children's Hospital Boston for licensed research-only mAbs.

We thank Rosemarie Mick for biostatistical advice and Drs. Peter O'Dwyer, Mark O'Hara, and Adham Bear for helpful discussions.