By Davendra P. S. Sohal, MD, MPH
- The two frontline options for metastatic pancreatic cancer management remain FOLFIRINOX and gemcitabine/nab-paclitaxel.
- Second-line therapy options are extrapolated from studies conducted in patients who received gemcitabine as first-line therapy, including the NAPOLI-1 and the CONKO-003 trials.
- An important addition to our armamentarium has been the evaluation and approval of pembrolizumab for tumors with MSI-H or dMMR.
- Integrated genomic analyses have identified distinct molecular subtypes of pancreatic cancer, which will hopefully allow precision oncology trials in this arena.
Metastatic pancreatic cancer continues to carry a dismal prognosis and is predicted to be the second leading cause of cancer-related death in the United States by 2020.1 ASCO issued guidelines in 2016 for management of metastatic pancreatic cancer2; recommendations therein, advances since then, and future directions are discussed here. A quick note on staging: this discussion applies to metastatic disease. Although locally advanced (unresectable) cancer is also incurable, treatment implications are somewhat different.
The two frontline options for management remain FOLFIRINOX (5-fluorouracil, irinotecan, oxaliplatin) and gem/nab-P (gemcitabine/nab-paclitaxel).3,4 Since there is no head-to-head comparison of these two regimens, we have to evaluate the nuances of each trial to determine which regimen may be best for a given patient, albeit with the caveat that a formal comparison across two different trials is not feasible.
The first step is study design and eligibility criteria. The FOLFIRINOX trial was conducted in France from 2005 to 2009; the gem/nab-P study was global and conducted from 2009 to 2012. Both used gemcitabine—the previous de facto standard of care—as the control arm. Eligibility criteria for both trials were very similar: notable items include ECOG Performance Status (ECOG PS) greater than or equal to 1 for FOLFIRINOX and Karnofsky Performance Status greater than or equal to 70 for gem/nab-P; upper age limit of 75 years for FOLFIRINOX but none for gem/nab-P; bilirubin up to 1.5 times upper limit of normal for FOLFIRINOX but no greater than upper limit of normal for gem/nab-P. The primary outcome for each study was overall survival (OS). The FOLFIRINOX study enrolled 342 patients; the gem/nab-P study had 861 patients.
The next step is evaluating the actual enrolled population. The median age in the FOLFIRINOX trial was 61 years, with the imposed upper limit of 75 years; the median was 63 years for the gem/nab-P study, with patients up to age 88. In the FOLFIRINOX trial, 99% had ECOG PS greater than or equal to 1; in the gem/nab-P trial, 92% had a Karnofsky Performance Status greater than or equal to 80, which is equivalent to ECOG PS greater than or equal to 1. The rest of the characteristics were remarkably similar as well. For the FOLFIRINOX and gem/nab-P studies: biliary stents were present in 14% and 17% of patients, 88% and 84% had liver metastases, and 42% and 52% had CA19.9 values greater than or equal to 5X ULN, respectively. Notably, only 38% and 43% of patients, respectively, had head of pancreas tumors. The normal distribution of pancreatic adenocarcinoma is approximately 70% head tumors, with the rest being body/tail tumors. The lower proportion of head tumors in these trials is because of the stipulation of normal or near-normal bilirubin, which is often not the case in head tumors because they frequently lead to biliary obstruction.
The final comparison lies in the performance of the control arm. Here, too, the survival curves are almost superimposable: gemcitabine led to a median OS of 6.8 and 6.7 months in the FOLFIRINOX and gem/nab-P trials, respectively. Median OS was 11.1 months in the FOLFIRINOX arm, and 8.5 months in the gem/nab-P cohort. Objective response was 32% (compared with 9% in the gemcitabine arm) in the FOLFIRINOX trial and 23% (compared with 7% in the gemcitabine arm) in the gem/nab-P trial. In terms of patients with grade 3 or higher toxicities, comparing FOLFIRINOX and gem/nab-P arms of the respective studies, 46% and 38% had neutropenia, 5% and 3% had febrile neutropenia, 24% and 17% had fatigue, 13% and 6% had diarrhea, and 9% and 17% had peripheral neuropathy, respectively.
This evaluation helps put into perspective our decisions on which regimen to pick. The consensus of the ASCO guideline panel was that FOLFIRINOX is perhaps more efficacious, but comes with the associated cost of higher toxicity burden. The administration schedules are, of course, somewhat different—FOLFIRINOX necessitates a chemotherapy port, while gem/nab-P requires weekly visits. The guideline recommends FOLIRINOX, therefore, for patients with metastatic pancreatic cancer and “ECOG PS 0 to 1, favorable comorbidity profile, patient preference and support system for aggressive medical therapy, and access to chemotherapy port and infusion pump management services.”2 A favorable comorbidity profile is defined empirically and includes normal bone marrow, hepatic, and renal function, as well as the absence of comorbid conditions requiring active medical care. For patients with an ECOG PS of 0 or 1 and a “relatively favorable comorbidity profile,” again, defined empirically as normal organ function and comorbid conditions that are well-controlled, gem/nab-P is the appropriate choice.2 For patients with an ECOG PS of 2 or a comorbidity profile that make frontline choices too onerous, gemcitabine as a single agent is appropriate. For patients with an ECOG PS of 3 or 4, cancer-directed therapy is not recommended.2
Second-line therapy options are extrapolated from studies conducted in patients who received gemcitabine as first-line therapy. These include the NAPOLI-1 and the CONKO-003 trials5,6; the former tested 5-fluorouracil and nanoliposomal irinotecan, and the latter used 5-fluoruracil and oxaliplatin. The comparison arm in each study was 5-
fluorouracil—there has been no placebo comparison. Each regimen improved median OS compared with 5-fluorouracil alone; NAPOLI-1 showed 6.1 months versus 4.2 months, and CONKO-003 showed 5.9 versus 3.3 months. Interestingly, the PANCREOX study tested FOLFOX versus 5-fluorouracil and actually showed detriment from adding oxaliplatin (median OS 6.1 vs. 9.9 months).7
All of this makes it difficult to make a good recommendation. The consensus appears to be that we have five chemotherapy agents shown to provide clinical benefit in metastatic pancreatic cancer: 5-fluorouracil (substitution with capecitabine has not been studied rigorously in the metastatic setting, but is probably acceptable, given data in the adjuvant setting), irinotecan (or nanoliposomal irinotecan), oxaliplatin, gemcitabine, and nab-paclitaxel. Depending on which agents were used in the first-line setting, the patient’s comorbidity profile, and any residual toxicities, nonoverlapping agents can be selected for second-line therapy.
Advent of Immunotherapy
An important addition to our armamentarium has been the evaluation and approval of pembrolizumab, a PD-1 monoclonal antibody, for tumors with high microsatellite instability (MSI-H) or deficient mismatch repair (dMMR). Pembrolizumab’s approval is remarkable as the first site-agnostic agent based on molecular markers and driven by a pivotal single-arm study.8 This trial enrolled 86 patients with metastatic cancers that had progressed on prior therapy; tumors had to test positive for either MSI-H by polymerase chain reaction (PCR) or dMMR by immunohistochemistry. All patients received pembrolizumab; 53% had an objective response and another 23% had stable disease; the 2-year OS proportion was 64%; median OS was not reached. There were eight patients with metastatic pancreatic cancer in the study; two had a complete response, three had a partial response, and one had stable disease—or 75% with disease control.
These are very exciting results for refractory metastatic pancreatic cancer. Therefore, MSI/MMR testing should be performed in every case where second-line therapy is being considered. The proportion of pancreatic cancers with MSI-H/dMMR tumors has not been well-studied; however, small series report 17% to 22% of cases with resected cancers,9,10 which is perhaps an overestimation of what might be seen in the metastatic setting, akin to colon cancer. More routine testing will help define population-based prevalence.
Patients with metastatic pancreatic cancer have a high symptom burden. Key problems causing distress include pain, anorexia, nausea and vomiting, wasting and fatigue, depression and anxiety, biliary obstruction, ascites, and thromboembolic events. Therefore, early institution of palliative care is strongly recommended. Each institution and practice setting has its own method of delivering supportive care. All methods are acceptable as long as minimizing distress is a component of care parallel to cancer-directed therapy. At our institution, for example, patients with metastatic pancreatic cancer are referred to a palliative medicine physician as soon as therapy starts, in order to maximize quality of life. Eventually, if needed, palliative care teams can help with the transition to end-of-life care, including home or facility hospice provision.
Opioids remain the mainstay of cancer-related pain control. Neurolytic agents and celiac plexus block can be used if opioids fail to control pain or cause intolerable side effects. For anorexia and nausea/vomiting, olanzapine can provide multifaceted relief. Assistance from psychiatrists and psychologists can be sought for management of emotional distress. Aggressive management of ascites, with indwelling peritoneal catheters if indicated, can provide relief and minimize office visits for paracenteses. Edema may accompany ascites, although it is important to suspect venous thrombosis as a potential cause of edema and intervene accordingly.
There is certainly a lot of room for improvement in the treatment of metastatic pancreatic cancer. Clinical trials evaluating newer therapies are ongoing. There is a long list of studies of targeted (noncytotoxic) therapies that have not shown benefit so far. A growing understanding of the molecular underpinnings of pancreatic cancer is driving our efforts now. Integrated genomic analyses have identified distinct molecular subtypes of pancreatic cancer,11-13 which can hopefully allow precision oncology trials in this arena. Key approaches include sequencing tumor genomes to identify therapeutic targets (used principally in basket studies such as NCI-MATCH, the ASCO TAPUR Study, Strata Oncology, My Pathway, NCT02967770, etc.), and using immunologic and molecular therapies targeting specific cellular pathways (e.g., NCT02562898, ibrutinib; NCT02340117, p53 gene therapy; NCT01489865 and NCT02890355, PARP inhibitors; NCT02077881, IDO inhibitor; and NCT02847000, THU/decitabine).
To continue to do better for our patients, it behooves us to remember the overarching recommendation in the ASCO guideline: “Every patient with pancreatic cancer should be offered information about clinical trials, which include therapeutic trials in all lines of treatment, as well as palliative care, biorepository/biomarker, and observational studies.”2
About the Author: Dr. Sohal is an assistant professor of medicine of the Cleveland Clinic Taussig Cancer Institute, where he also serves as the director of the Clinical Genomics Program.