PIPAC-OX: A Phase I Study of Oxaliplatin-Based Pressurized Intraperitoneal Aerosol Chemotherapy in Patients with Peritoneal Metastases

Peritoneal metastases are associated with poor prognosis (1, 2) and dismal quality of life (3). Systemic chemotherapy in these patients has suboptimal response (2, 4). These unique features have prompted a paradigm shift in recent years, with increased understanding of the pathophysiology (5), paired with the concept of peritoneal metastases being considered a locoregional disease in some patients, as well as a gradual acceptance that the optimal treatment of these patients requires a multimodal approach that may include some form of peritoneal-directed therapy.

Pressurized intraperitoneal aerosol chemotherapy (PIPAC) is an innovative peritoneal-directed chemotherapy concept for treating peritoneal metastases that has been purported to enhance efficacy through superior distribution and depth of penetration of drugs (6, 7). To date, PIPAC is performed with only two regimens (8)—cisplatin in combination with doxorubicin or oxaliplatin monotherapy. While the dose used for PIPAC with cisplatin and doxorubicin is based on the results of a phase I dose-escalation study (9), the dose for PIPAC oxaliplatin is currently at 92 mg/m²—arbitrarily derived as 20% of the dose administered in a commonly used hyperthermic intraperitoneal chemotherapy (HIPEC) regimen (10). Furthermore, in clinical studies on PIPAC oxaliplatin, a significant number of patients underwent recent or concurrent administration of systemic chemotherapy, confounding the interpretation of the side effects and safety profile (11–13).

We aim to determine the safety and tolerability of escalating doses of oxaliplatin delivered via PIPAC, and in so doing, derive the recommended phase II dose (RP2D). The secondary aim is to evaluate the short-term clinical and pathologic response of PIPAC with oxaliplatin, as well as identify the pharmacokinetic profile of oxaliplatin administered via PIPAC.

This is a prospective, single-arm, phase I trial in a 3+3 dose-escalation design to evaluate the safety and tolerability of oxaliplatin-based PIPAC in patients with peritoneal metastases. Ethics approval was obtained from the National Healthcare Group Domain Specific Review Board (2016/01088). This trial is also registered on ClinicalTrials.gov. (NCT03172416). This study was conducted in accordance with the Declaration of Helsinki. Informed written consent was obtained from all study subjects.

Details and rationale of the protocol have been described previously (14). In brief, patients with unresectable peritoneal metastases of gastrointestinal cancer origin were recruited from May 2017 to September 2019. Recruitment was subsequently expanded to include patients with peritoneal metastases from primary gallbladder, pancreatic, and appendiceal cancers. All patients were recruited only if they had disease progression on, or were intolerant to first-line systemic chemotherapy. Patients were excluded if there were predominant extraperitoneal metastases, poor performance status, or an expected survival of less than 3 months. The study timeline and inclusion and exclusion criteria are listed in Supplementary Fig. S1.

PIPAC was administered as described by Solass and colleagues (15), using oxaliplatin reconstituted in 150 mL of dextrose solution administered at a flow rate of 30 mL/minute with a maximal upstream pressure of 200 psi. The system was kept at a steady state for a total of 30 minutes at intraabdominal pressure of 12 mmHg.

In total, five cohorts were planned at doses of 45, 60, 90, 120, and 150 mg/m². Patients were planned to undergo a minimum of two cycles of PIPAC, 6 weeks apart. Surgical complications were graded according to the updated Clavien–Dindo Classification (16). Toxicity was monitored for and graded according to the NCI—Common Terminology Criteria for Adverse Events Version 4.03 (17). Dose-limiting toxicities (DLT) were defined as grade 3 or higher toxicity which may be related to the study treatment during the first 28 days after therapy. Adverse events (AE) were considered treatment related only if they had developed or, if preexisting, worsened after PIPAC. Clinicopathologic response was assessed according to the intraoperative peritoneal cancer index (PCI; ref. 18) and peritoneal tumor biopsies scored according to the mean peritoneal regression grade score (PRGS; ref. 19) during each PIPAC procedure, and with cross-sectional imaging using RECIST (20). During each PIPAC procedure, four-quadrant biopsies would be taken where possible. PRGS for each patient was derived from the mean of individual quadrant PRGSs. Analysis of clinical data was performed using Microsoft Excel and STATA version 15.1.

Blood samples for pharmacokinetic studies were collected by venepuncture or via an indwelling peripheral line, followed by transfer into appropriately labeled purple top ethylenediaminetetraacetic acid vacutainer tubes at predefined sampling time points. The first 5 mL blood withdrawn from an indwelling vascular line were routinely discarded. One 5 mL whole blood sample was collected at each time point—0 (prior to PIPAC), 0.5, 0.75, 1, 2, 4, 8, 24, and 30 hours after the start of PIPAC. Within 15 minutes from collection, blood samples were centrifuged at 3,000 g for 10 minutes at 4°C to separate the plasma and 1 mL of plasma was further processed to obtain ultrafiltrates (4,000 g, 60 minutes, 25°C using Corning Spin-X UF6 mL Concentrators, cut-off: 10 kD; Corning Incorporated Life Sciences). Plasma samples and ultrafiltrates were stored at −80°C until analysis. Plasma and ultrafiltrate samples were shipped with inventory sheet in a Styrofoam container and packaged in dry ice to ensure that they remain frozen. Pharmacokinetic assays were performed at Drug Analysis and Pharmacokinetic Core facility of the Cancer Science Institute of Singapore, National University of Singapore, Singapore. The quantitative analysis of oxaliplatin was performed in inductively coupled plasma-mass spectrometry (ICP-MS). Methodology development and validation was done based on the assay published by Morrison and colleagues (21). The pharmacokinetic parameters of PIPAC oxaliplatin were calculated via Phoenix WinNonlin 64.

Funding for this study was used to pay for the costs of treatment for patients. This is an investigator-initiated trial and the study team had full access to all data in the study and take final responsibility for the decision to submit this study for publication.

A total of 17 patients were recruited for this study (see Fig. 1 for study flow chart). One patient initially recruited into the second dose cohort was excluded from the study as PIPAC could not be administered after safe laparoscopic access into the abdomen was limited by extensive adhesions. A summary of the patient demographics and characteristics are presented in Table 1.

No surgical morbidity occurred during this trial. Treatment-related AEs and all AEs are presented in Table 2 and Supplementary Table S1, respectively. Six (37.5%) patients experienced treatment-related AEs. Abdominal pain (12.5%), fever (6.3%), fatigue (6.3%), and vomiting (6.3%) occurred within 1 week after PIPAC in these patients. Three patients developed acute pancreatitis (grade 2 and 3) after PIPAC, which all resolved with conservative treatment. Two patients were from the 45 mg/m² dose cohort and one was from the 90 mg/m² cohort. The first patient, with metastatic poorly differentiated adenocarcinoma of the colon, had disease progression on three lines of chemotherapy prior to PIPAC. Seven days after the first cycle of PIPAC, he developed self-limiting mild abdominal pain (grade 1). However, he developed acute pancreatitis (grade 3) 6 days after the second cycle of PIPAC. No alternate risk factors for acute pancreatitis were identified and he was treated conservatively. However, his clinical condition deteriorated because of rapid disease progression and he passed away 51 days after the initial PIPAC. The second patient had peritoneal recurrence after radical gastrectomy for poorly differentiated adenocarcinoma of the stomach and adjuvant chemotherapy, and then disease progression on first-line palliative chemotherapy. She developed acute pancreatitis (grade 2) 9 days after the first cycle of PIPAC which resolved with conservative management. However, she subsequently developed subacute intestinal obstruction due to disease progression and did not undergo the second cycle of PIPAC. The third patient (from the 90 mg/m²cohort) had asymptomatic elevation of amylase and lipase one day after the first cycle of PIPAC, which later normalized. He underwent a second PIPAC subsequently with no AEs noted. No significant hematologic, hepatic, or renal toxicity occurred in all study subjects. However, a transient rise in the leucocyte (neutrophil dominant) counts was noted on days 1 and 2 after PIPAC, with rapid return to near baseline counts subsequently.

The median overall (OS) and progression-free survival (PFS) were 124 (95% CI, 57–250) and 42 (95% CI, 28–75) days, respectively. Three patients included in this study had extraperitoneal metastases prior to the study (Table 1). There was no statistical difference in median PFS (36 vs. 63 days, P = 0.39) and median OS (73 vs. 124 days, P = 0.86) in these patients compared with patients with only peritoneal metastases. Clinical response (based on cross-sectional imaging and histopathology) are summarized in Table 3; Supplementary Table S2. Of note, 6 patients (37.5%) developed new or worsening extraperitoneal metastases during the study period. These metastases were in the lung for 3 patients, liver for 2 patients, and bone in 1 patient.

Eight patients (50%) underwent only one PIPAC procedure. Of these patients, 2 patients dropped out due to disease progression per RECIST criteria (worsening peritoneal and liver metastases in 1 patient and worsening peritoneal and new lung metastases in the other), 3 patients (one of whom developed grade 3 pancreatitis after initial PIPAC) dropped out due to rapid clinical disease progression despite stable disease per RECIST on CT scans, and 3 patients chose to withdraw from the study (1 opted to enrol in another clinical trial; 2 were for personal reasons).

Of the 8 patients who underwent two PIPAC procedures, 5 patients (62.5%; 2 patients at 90 mg/m², 1 patient each at 45, 60, and 120 mg/m²) had an improvement in PCI score during laparoscopy at the second cycle of PIPAC. Median PCI score of these 8 patients at the first PIPAC and second PIPAC were 15 and 12, respectively. Four patients (57.1%; 2 patients at 90 mg/m², 1 patient each at 45 and 60 mg/m²) had an improvement in median PRGS with an overall decrease from 2.5 to 2.0. Five patients had stable disease per RECIST criteria. For the 3 patients who had progressive disease, 1 patient had worsening peritoneal disease only, another had worsening lung and peritoneal disease, while the last patient developed new bone metastases despite having stable peritoneal disease.

Pharmacokinetic analysis was performed on 14 of 16 patients based on plasma samples during/after the first cycle of PIPAC using noncompartmental analysis. Two patients were excluded, one patient from the 45 mg/m² dose cohort due to sample contamination, another from the 90 mg/m² dose cohort as she did not receive the full dose due to technical error during PIPAC administration. The results are summarized in Fig. 2; Supplementary Table S3; Supplementary Figs. S2 and S3.

The aforementioned advantages of PIPAC in peritoneal metastases have been a driving factor in the growing adaptation of this drug delivery method (8, 22–25). While several studies (11–13, 26) have been published on PIPAC oxaliplatin for peritoneal metastases, none of them provides a scientific rationale for dosing. This study remains the first phase I study published for PIPAC oxaliplatin, providing a scientific basis for the RP2D. Our results show that PIPAC oxaliplatin can be administered safely at doses up to 120 mg/m², with no PIPAC oxaliplatin–related AEs seen at this dose cohort. The transient rise in leucocyte and neutrophil counts on days 1 and 2 after PIPAC may be secondary to the inflammatory response, which is also seen after HIPEC (27). This inflammatory response is also consistent with existing literature on PIPAC (28).

This study identified acute pancreatitis as a previously unreported AE in relation to PIPAC oxaliplatin. Even with intravenous oxaliplatin, pancreatitis is a rare AE reported mostly in postmarketing reports (29). We hypothesize that this relatively high incidence (3 of 24 PIPAC procedures, 12.5%) may be due to the increased tissue (peritoneal) penetrance and distribution of oxaliplatin via PIPAC. As the pancreas is a retroperitoneal organ with its anterior capsule only exposed to the lesser sac, the vaunted advantages of PIPAC may be a double-edged sword with regards to pancreatitis. While two pancreatitis events occurred at the dose cohorts of 45 mg/m², the number of patients may be too small to detect a dose-dependent effect. Abdominal pain after PIPAC oxaliplatin is a commonly reported AE, with a recent multicentre cohort study from Europe reporting an incidence rate of 22.8% (26). No mention of routine serum amylase/lipase testing is found in the literature on PIPAC oxaliplatin. As such, we opine that the true incidence of pancreatitis and/or elevated amylase/lipase after PIPAC oxaliplatin may be underreported. As pancreatitis occurred 1, 7, and 9 days after PIPAC administration in the 3 patients in our study, we recommend that routine serum amylase/lipase testing should be performed repeatedly in the first 2 weeks after PIPAC oxaliplatin administration, to facilitate early diagnosis and institution of supportive treatment. Furthermore, centres performing PIPAC should be vigilant and have a low index of suspicion for acute pancreatitis.

This is the first pharmacokinetic report of PIPAC oxaliplatin in the literature. The pharmacokinetic analyses on platinum in plasma total (plasma^tot), which reflects both bound and ultrafiltered platinum, demonstrate excellent linear relationship between dose and both maximum concentration (C_max) and exposure/AUC. This linearity was also demonstrated in the ultrafiltrate samples (Fig. 2). Importantly, the platinum exposure in ultrafiltrate was discovered to be about one-tenth of plasma^tot (Supplementary Fig. S3). These results suggest that systemic platinum exposure will increase proportionately and predictably with PIPAC dosing. Systemic platinum exposure in our highest dose cohort of 120 mg/m² was 3.4%–10.8% of that reported for intravenous infusion of oxaliplatin at 130 mg/m² over 2 hours (30), and therefore unlikely to contribute significantly to systemic AEs. Thus, it may be feasible to combine PIPAC oxaliplatin with systemic chemotherapy safely, with predictable and reduced (compared with systemic oxaliplatin infusion) systemic toxicity. In this study, the pharmacokinetic results demonstrated that plasma protein–bound platinum accounts for approximately nine-tenth of total plasma platinum content, with the remaining 10.9% of active platinum detected in plasma ultrafiltrate. Interestingly, this distribution ratio between plasma and ultrafiltrate was relatively consistent across the dose ranges of PIPAC oxaliplatin from 45 to 120 mg/m², leading to proportional active platinum exposure in the circulation system with increasing dose of PIPAC oxaliplatin.

In terms of efficacy, data in this study are based on patients with high peritoneal disease burden (median PCI score of 17), many of whom have exhausted multiple lines (43.7%) of systemic palliative chemotherapy. Hence, it is unsurprising that the median PFS and OS in our study are relatively short, as patients in this study were likely to have poor tumor biology. While the 3 patients who had extraperitoneal metastases prior to PIPAC did not have significant differences in PFS and OS as compared with the remaining 13 patients who had peritoneal-only metastases, the trend suggests that a larger sample size may show differences. Therefore, it should be taken into consideration whether patients with synchronous extraperitoneal metastases should be included in such future studies.

On the basis of RECIST criteria on cross-sectional imaging, 62.5% of patients in this study had stable disease after the first PIPAC, with 50% of the patients who underwent two PIPAC procedures noted to have stable disease. As cross-sectional imaging is notoriously limited for the evaluation of peritoneal disease (31–33), we used PCI score and four-quadrant peritoneal tumor biopsies as surrogate measures of response and saw increased efficacy at the higher dose cohorts of 90 and 120 mg/m². It is also notable that half of the patients in this study did not undergo the second PIPAC procedure, the main reason being disease progression (5 of 8 patients). This, coupled with the development of new or progression of known extraperitoneal metastases in some patients in this study, suggests that monodirectional therapy with PIPAC only is insufficient in the treatment of metastatic disease, and lends support to the concept of bidirectional therapy, combining systemic therapy with PIPAC. Several such trials are currently ongoing (NCT03294252, NCT03287375, NCT03246321, NCT03280511; refs. 34–37). While the results of these studies are eagerly awaited, this trial demonstrates an additional advantage of PIPAC—the ease of access to the peritoneum to evaluate for response via laparoscopy and histopathologic analysis. Future studies on PIPAC oxaliplatin should exploit the benefits of laparoscopic access to the peritoneum during PIPAC for translational research to guide treatment, potentially via cell culture based in vitro drug screening, and to improve our understanding of peritoneal metastases.

The study was terminated after the dose cohort of 120 mg/m². As such, we did not establish the MTD, which is a limitation of this study. The reason for not escalating to the last preplanned dose cohort at 150 mg/m² was due to data from a French phase I study (38) that was similarly evaluating PIPAC oxaliplatin. They had established the RP2D at 90 mg/m², with DLTs (grade 4 allergic reaction and grade 3 neutropenia) seen at the higher dose of 140 mg/m². Furthermore, we have seen an improvement in PCI score and PRGS at all dose cohorts tested from 45 mg/m² to 120 mg/m². Hence, a decision was made to terminate the study to avoid potential AEs that patients may suffer had they undergone PIPAC oxaliplatin at the dose cohort of 150 mg/m², especially since some efficacy was seen at the lower dose levels.

As a phase I study in the conventional 3+3 dose-escalation design, it is important to note that the small number of patients at the RP2D cohort of 120 mg/m² warrants a larger phase II study to further evaluate safety and efficacy at this dose level. Also, this study is not designed to evaluate for AEs related to repeated/recurrent PIPAC oxaliplatin treatments, as the patients were only evaluated for a maximum of two PIPAC procedures, with half of them only undergoing one PIPAC treatment. As such, AEs such as severe peritoneal sclerosis (39) and cumulative neuropathy will not be seen in this study.

Our study has established the RP2D for PIPAC oxaliplatin at 120 mg/m². PIPAC oxaliplatin is well tolerated because of limited systemic exposure compared with intravenous infusion, a drug delivery approach routinely used in clinical settings. It has also provided first-in-human pharmacokinetic parameters on PIPAC oxaliplatin with preliminary evidence of efficacy. Future studies should build on the results of this phase I trial to further delineate the efficacy and role of PIPAC oxaliplatin in the treatment of peritoneal metastasis.

R. Sundar reports honoraria and/or research funding from Bristol Myers-Squibb, Merck, Eisai, Bayer, Taiho, MSD, Eli Lilly, Roche, AstraZeneca, and Paxman Coolers outside the submitted work. W.P. Yong reports personal fees from Eisai, Bayer, BMS, MSD, Amgen, and Lilly outside the submitted work. No disclosures were reported by the other authors.

G. Kim: Conceptualization, formal analysis, validation, investigation, methodology, writing–original draft, project administration. H.L. Tan: Conceptualization, data curation, formal analysis, validation, investigation, methodology, writing–original draft, project administration. R. Sundar: Data curation, visualization, writing–review and editing. B. Lieske: Data curation, visualization, writing–review and editing. C.E. Chee: Conceptualization, data curation, visualization, methodology, writing–review and editing. J. Ho: Data curation, visualization, writing–review and editing. A. Shabbir: Data curation, visualization, writing–review and editing. M.V. Babak: Data curation, formal analysis, writing–review and editing. W.H. Ang: Data curation, formal analysis, writing–review and editing. B.C. Goh: Validation, writing–review and editing. W.P. Yong: Conceptualization, resources, data curation, supervision, funding acquisition, visualization, methodology, writing–review and editing. L. Wang: Conceptualization, resources, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft. J.B.Y. So: Conceptualization, resources, data curation, supervision, funding acquisition, visualization, methodology, writing–review and editing.

This study was funded by the National Medical Research Council (NMRC) Open Fund - Large Collaborative Grant (MOH-OFLCG18May-0003) and by the NMRC Centre Grant Programme (CG12Aug12). ICP-MS analysis of oxaliplatin was performed by Huiqing Xie from IMRE-A*Star, Singapore. The authors thank the patients and their families, as well as the study team coordinators (Elya, Clarisse Jang, Siok Chin Teo, Corissa Chee, Huey Ying Lim, and Jana Chan), for making this trial possible.

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.