Early termination of NCT04617067, a phase II, open label, clinical trial of oral paricalcitol in combination with gemcitabine and NAB-paclitaxel therapy in advanced pancreatic cancer

Abstract

Background

Pancreatic cancer (PDAC) often responds poorly to chemotherapy, represents an unmet clinical need and new therapeutic approaches are urgently required. Desmoplasia is a hallmark of PDAC. Multiple preclinical studies suggest cancer associated fibroblasts (CAF) support cancer growth, and attention has recently turned towards inclusion of anti-stromal agents into chemotherapy trials. Our objective was to evaluate safety and tolerability of oral paricalcitol in combination with systemic chemotherapy in patients with advanced PDAC.

Methods

This was a phase II, single-arm study in patients with advanced PDAC who had received no prior systemic chemotherapy. Gemcitabine and NAB-paclitaxel were administered weekly for 3 of every 4 weeks (days 1, 8 and 15) and paricalcitol administered orally every day of a 28-day cycle. Patients were treated until disease progression with an interim analysis. The primary efficacy endpoint was progression free survival (PFS). Secondary efficacy endpoints were evaluation of overall survival (OS), time to treatment failure (TTF) and tumour response rate (TRR).

Results

Fifteen patients were enrolled. Median PFS was 14.6 weeks with 95% CI of (7.9–24.0). Estimated PFS rate at 24 weeks was 18.0% with 95% CI of (2.9–43.4). Five patients achieved stable disease; one achieved a partial response. Confirmed tumour response rate was 8.3% with 90% CI of (0.4–33.9). Mild hypercalcaemia, previously associated with vitamin D receptor agonists, occurred in nine (60%) patients, moderate (Grade 3) hypercalcaemia in 2 patients and there was no grade 4/5 hypercalcaemia. The study did not meet its primary objective and discontinued following interim analysis.

Conclusions

Oral paricalcitol was safely combined with chemotherapy. The prespecified efficacy threshold for 6-month progression free survival probability (≥ 70%) was not met. The study was stopped early after the planned interim analysis as the criterion pre-specified in order to move to the next stage of the study was not met. Efficacy of paricalcitol in advanced PDAC was lower than expected, with a non-significant trend towards decreased PFS. Our study has implications for interpretation and design of clinical trials incorporating paricalcitol for patients with advanced pancreatic cancer.

Trial registration

ClinicalTrials.gov identifier: NCT04617067, registration date 10/27/2020.

Background

Pancreatic ductal adenocarcinoma (PDAC) has one of the poorest prognoses out of all solid tumours and patients with metastatic disease have a median survival of only 3–6 months [1]. This likely reflects multiple factors, including innate features of tumour biology such as desmoplasia and immune evasion [2]. PDAC is the third most common cause of cancer related death in the USA with an increasing incidence [1]. Most patients present with metastatic disease at the time of diagnosis [3], and effective new therapeutic approaches are urgently required. Until recently, single-agent gemcitabine was standard first-line treatment in advanced PDAC. However, resistance occurs readily, and two regimens have largely superseded gemcitabine as first-line treatment to extend average survival of patients with advanced PDAC by several months: (1) FOLFIRINOX (5-fluorouracil/irinotecan/oxaliplatin), and (2) Gemcitabine combined with nanoparticle albumin bound (NAB)-paclitaxel [4].

An earlier study of Gemcitabine plus (NAB)-paclitaxel found an OS of 8.5 months, PSF (5.5 months) and Overall Response Rate (ORR) of 23% [4]. A more recent meta-analysis comparing these regimens found an increased PFS for FOLFIRINOX (7.3 months), versus Gemcitabine and Nab-Paclitaxel (5.7 months). However, there was no statistically significant difference in OS (11.7 versus 10.4 months), or ORR (31.6% versus 35.0%) between FOLFIRINOX and Gemcitabine plus (NAB)-paclitaxel respectively [5]. Although it currently remains unclear which regimen is the most effective, gemcitabine plus (NAB)-paclitaxel is the more widely used because of better tolerability and less toxicity [4]. Although chemotherapy is ineffective for most patients, responses occur in a minor proportion, motivating the search for new therapeutic approaches.

Desmoplasia is a hallmark of PDAC where stromal hyperplasia together with excessive extracellular matrix (ECM) formation comprise up to 90% of tumour volume [2]. This restricts vasculature, resulting in hypoxia, decreased drug delivery and limited benefits from chemotherapy. Much interest surrounds targeting this stromal reaction, with a focus upon cancer-associated fibroblasts (CAF), derived from resident fibroblasts and pancreatic stellate cells (PSC), responsible for desmoplasia [6]. The tumour microenvironment (TME) is also important in metastasis, although signalling pathways between stromal and cancer cells are less well characterised at distant sites [7].

Early studies found PSCs promoted tumour progression. PSCs supported growth of PDAC cell lines in xenografts and PSC-conditioned media increased growth and migration of PDAC cells [8–10]. Taken together, initial data supported approaches aimed at ablation of tumour-promoting stroma in patients with PDAC.

However, the role of stromal cells remains controversial, with some preclinical studies finding CAFs restrain tumour growth [11–13]. Consistent with this idea, Sonic Hedgehog pathway inhibitors have shown a lack of efficacy in clinical trials [11, 14, 15]. These findings have been interpreted to indicate heterogeneity among stromal cells with both tumour promoting and restraining subtypes present in PDAC. Evidence for this idea derives from gene expression studies. Moffitt et al., used RNA-Seq to describe normal and activated PSC subtypes [16], later extended to include ECM-rich and immune-rich stromal subtypes [17]. Tuveson’s group described inflammatory (iCAFs) and myofibroblast (myCAFs); hypothesised to promote and restrain tumour progression, respectively [18, 19]. On the other hand, myofibroblasts have also been described, including LRRC15⁺ and PRRX1⁺ cells, which exert pro-tumorigenic effects [20, 21].

Gene expression analysis has also been used to classify PDAC, with a consensus for two molecular subtypes [16]. A classical subtype expressing epithelial markers with a more favourable prognosis and a basal-like subtype with mesenchymal markers, displaying aggressive clinical behaviour. More recent studies indicate these two subtypes exist on a spectrum with intermediate phenotypes, and both subtypes may be present within an individual tumour [22].

Vitamin D agonists (VDAs) have attracted attention as anti-stromal agents. Calcitriol, the naturally occurring product of vitamin D metabolism and synthetic VDAs, inhibit tumour growth in xenografts of pancreatic cancer [23, 24]. Clinical studies with calcitriol remain largely disappointing, in part because of dose limiting hypercalcaemia [25]. Less hypercalcaemic VDAs such as paricalcitol, calcipotriol and inecalcitol have been developed to circumvent this problem [26]. Calcipotriol reduced desmoplasia in an orthotopic mouse model of PDAC, and enhanced efficacy of gemcitabine, increasing survival compared to gemcitabine alone [24]. Taken together, the data suggested that VDAs promote PSC differentiation and quiescence. This important preclinical study was an impetus for multiple clinical trials, testing the effects of paricalcitol in combination with chemotherapy in patients with PDAC.

A pilot study combined IV paricalcitol with gemcitabine and NAB-paclitaxel in the neoadjuvant setting (NCT0203086) and further studies have extended into phase II and include patients with metastatic disease [27, 28]. An emerging idea is that VDAs reprogram CAFs, counteracting their tumour-promoting activities, and improving delivery of chemotherapy [29, 30]. The hope is that this will translate into better responses in patients with metastatic PDAC. Here we describe the results of a phase II, single arm, multi-centre trial, using a Simon two-stage design, to evaluate anti-tumour efficacy of paricalcitol in combination with gemcitabine and NAB-paclitaxel in patients with advanced PDAC.

Methods

The study was approved by the Health Products Regulatory Authority and the National Research Ethics Committee. Cancer Trials Ireland ID number CTRIAL-IE 19–33. Our report adheres to CONSORT guidelines (https://www.equator-network.org/reporting-guidelines/consort/).

Patients

Patients with previously untreated advanced PDAC were enrolled into a single arm, Phase II, open label trial. Cancers were incurable recurrent, locally advanced, or metastatic, based upon biopsy-proven disease and radiological imaging as per Response Evaluation Criteria in Solid Tumours (RECIST). Baseline patient demographics and disease characteristics are contained within Table 1. Written informed consent was obtained prior to any study-related procedures. Patients attended for treatment at: Beaumont Hospital, (Dublin), St Vincent University Hospital, (Dublin), University Hospital Waterford, Cork University Hospital, and University Hospital Limerick, (Ireland), where data was collected for the study.

Table 1.

Baseline characteristics

Characteristic	Statistics	Total Patients (n = 15)
Sex	Male	8 (35%)
	Female	7 (47%)
Age (years)	Mean (SD)	66.9 (10.74)
	Median	72.0
	Min-Max	45–79
Race	Caucasian	14 (93%)
	Black	1 (7%)
ECOG*	0	5 (33.33%)
	1	9 (60.00%)
	2	1 (6.67%)
Histology	Adenocarcinoma	15 (100%)
Stage	III	3 (20%)
	IV	12 (80%)
Prior Pancreatic Surgery	No	13 (87%)
	Yes	2 (13%)
Lung metastases	Positive	4 (29%)
	Negative	10 (71%)
Liver metastases	Positive	11 (79%)
	Negative	3 (21%)

*ECOG status: 0-Fully active, able to carry on all pre-disease performance without restriction

1 - Restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, e.g., light housework, office work

2 - Ambulatory and capable of all selfcare but unable to carry out any work activities. Up and about more than 50% of waking hours

Main inclusion and exclusion criteria

Eligibility criteria included: Eastern Cooperative Oncology Group (ECOG) performance status 0–2, adequate haematological, renal, and hepatic function measured within 28 days prior to commencing study treatment (including a corrected serum calcium of ≤ 2.9 mmol/L (≤ grade 1), life expectancy of at least 12 weeks and RECIST 1.1 measurable disease. Major exclusion criteria included: known brain metastases, altered mental status and uncontrolled intercurrent illness. Additional criteria included: a known history of hypercalcaemia, disease associated with disordered calcium, (such as ongoing hyperparathyroidism and sarcoidosis), and symptomatic kidney stones within the last 5 years. Sensory or motor neuropathy (greater than or equal to grade 2) was a further exclusion factor.

Study design and treatment

Patients were treated with paricalcitol plus gemcitabine and NAB-paclitaxel in 28-day cycles. Gemcitabine (at 1,000 mg/m²) and NAB-paclitaxel (at 125 mg/m² of body-surface area) were administered weekly for 3 of every 4 weeks (on days 1, 8 and 15 only). Paricalcitol (12mcg) was administered orally on every day of a 28-day cycle.

Oral dose paricalcitol has a bioavailability of 72–86%, comparable with the IV route. Moreover, the pharmacokinetics of paricalcitol are similar for both routes. Paricalcitol plasma concentrations decrease log-linearly and approximately 87% is removed at 12–18 h (three half-lives) after injection. Hence, we anticipated that oral daily dosing of paricalcitol might provide a more constant steady state concentration than the three times per week dose span>previously used in IV trials [31].

Patients were treated up until disease progression (radiological or clinical progression, e.g., ECOG performance status ≥ 3), or when one of the following occurred: (a) unacceptable toxicity, (b) development of intercurrent illness or a situation which, might affect assessments of clinical status and study endpoints, or (c) in the investigator’s opinion, continuation of the study would be harmful to the patient’s well-being. The study design is summarized in Fig. 1.

Fig. 1.

Schematic of study design: a phase II, single arm, multi-centre trial, incorporating an interim analysis. A Simon two-stage design was used to evaluate anti-tumour efficacy of paricalcitol in combination with gemcitabine and NAB-paclitaxel

End points and definitions

The primary efficacy endpoint of the study was progression free survival (PFS); that is, the percentage of patients alive and free of progression at 24 weeks from enrolment into the study as determined by radiographic disease assessments as per RECIST v1.1 criteria. Secondary efficacy endpoints were evaluation of overall survival (OS), time to treatment failure (TTF), defined as time from registration to discontinuation of therapy or add-on of new anti-cancer therapy for any reason (including death, progression, and toxicity), and confirmed tumour response rate as per RECIST v1.1. The primary safety endpoint was to evaluate safety and tolerability of the regimen as measured by incidence of adverse events reported and toxicity evaluation as per the NCI Common Terminology Criteria for Adverse Events (CTCAE version 5.0). Secondary safety endpoints were assessed in an ongoing fashion. Each patient had targeted physical examinations on days 1, 8 and 15 of each cycle, and at other times if presenting with toxicity. Incidence of hypercalcaemia was an additional secondary endpoint. For the purposes of the primary endpoint, an evaluable patient was defined as one who received at least one dose of paricalcitol and an assessment of PFS (or died prior to assessment).

Assessments

Clinical and laboratory parameters were assessed to evaluate disease response and toxicity of study therapy. Patients were assessed with radiologic restaging by computerised tomography (CT) scanning of the thorax, abdomen, and pelvis every 12 weeks (+/- 14 days), up until disease progression. Data collected at baseline included ECOG performance status, height, weight, vital signs (temperature, pulse and blood pressure), Baseline Tumour Assessment, Complete chemistry panel, Serum vitamin D, Parathyroid Hormone, Corrected calcium and phosphate, Complete blood count and CA-19.9.

Statistics

We considered that a result finding 70% PFS at 24 weeks (i.e. that at least 70% of patients are progression-free at 24 weeks) would indicate that gemcitabine/NAB-paclitaxel/paricalcitol is worthy of further investigation in metastatic PDAC. The expected PFS of 70% was based on a previous phase 3 clinical trial of first-line gemcitabine and NAB-paclitaxel in patients with advanced PDAC. This study found 50% PFS at 5.5 months [4]. Testing for 70% vs. 50% PFS after 24 weeks using an alpha level of 5%, assuming the analysis was performed at 24 weeks after the last patient accrued, would require 43 evaluable patients to provide 80% power and this was the planned sample size for the present study. Taking the data together, we considered that a PFS of 70% at 6 months was a conservative estimate of the outcome that could be achieved in this study.

We used a Simon’s two stage design with an interim analysis planned at the end of the first stage of the study, when 15 evaluable patients had been assessed after 24 weeks, or earlier if the primary object could no longer be achieved [32]. Continuation into the second stage would have required at least nine of these fifteen patients to be progression-free at 24 weeks after registration. Statistical analysis was conducted using the software package, SAS Version 9.4.

Results

Treatment exposure

Fifteen patients were registered and received study treatment. All patients started study treatment within a week of registration, the majority within 2 days. The median number of cycles received was 3 and the maximum was 8 (one patient had a 9th cycle of gemcitabine alone), the duration of study treatment ranged from one to 36 weeks.

Interim analysis

Once 7 of 15 patients (53.3%) recruited to the first stage of the study failed to remain progression-free for up to 24 weeks, the study was stopped for futility. At the time of interim analysis, two patients were progression-free at 24 weeks and the 5 remaining patients had not yet reached 24 weeks after registration. At the point when the decision was made to stop the study, 5 patients had not yet reached week 24 and remained on treatment until they progressed or completed the study.

Full set analysis

Reasons for treatment discontinuation and study withdrawal are contained in Fig. 2. Four patients (26.7%) discontinued due to radiological disease progression. Two patients (13.3%) discontinued treatment due to clinical disease progression, one because of clinical deterioration, and the other because of poor performance status. Two patients died (13.3%); cause of death was progression for one patient and an adverse event unrelated to study treatment for the other. Three patients discontinued study treatment because of adverse events, one due to renal failure, a second due to malaise, and a third patient due to pulmonary embolism all of which were considered unlikely to be related to study treatment. Three patients discontinued study treatment for reason of investigator and one patient stopped when informed the trial would no longer proceed to second stage.

Fig. 2.

Swimmer plot for all patients enrolled in the study, indicating duration of treatment, response at week 12 and reason for discontinuation of study treatment. Patient 6 achieved stable disease at 12 weeks with a partial response after 24 weeks of treatment. Patient 11 had stable disease at 10 weeks and thereafter was unable to attend for further restaging scans due to malaise

The median PFS time was 14.6 weeks with 95% CI of (7.9–24.0). The estimated progression-free survival rate at 24 weeks was 18.0% with 95% CI of (2.9–43.4). Five patients achieved stable disease and one patient achieved a partial response. Hence the confirmed tumour response rate was 8.3% with 90% CI of (0.43–33.9). Three patients did not have any post-baseline RECIST assessment or any evidence of clinical progression and are excluded from the confirmed tumour response rate analysis. The median TTF time was 12.3 weeks with 95% CI of (6.1–24.1). The estimated treatment failure-free rate at 24 weeks was 28.6% with 95% CI of (8.8–52.4). The median OS time was 24.1 weeks with 95% CI of (14.6–42.4) and the estimated OS rate at 24 weeks was 51.3% with 95% CI of (21.8–74.7). Kaplan-Meier curves depicting the PFS, TTF and OS are shown in Figs. 3, 4 and 5 respectively. All eleven of the response-evaluable patients had an elevated CA-19.9 at baseline. Patient 5 had no baseline assessment and was excluded from analysis. At the end of treatment Carbohydrate antigen 19 − 9 (CA19-9) was decreased in 7 of these patients (63.63%), shown in Fig. 6.

Fig. 3.

Kaplan Meier plot. Progression Free Survival (PFS) for intention-to-treat patients (N = 13). Two patients were withdrawn before post baseline RECIST assessments (with no evidence of clinical progression) and were excluded from PFS analysis

Fig. 4.

Kaplan Meier plot. Time to Treatment Failure (TTF) for intention-to-treat patients (N = 14). One patient had no post-baseline RECIST assessment, continued on treatment for 4 weeks post study closure and was excluded from TTF analysis. One patient had no post-baseline RECIST assessment, with no clinical progression and discontinued study treatment due to an adverse event

Fig. 5.

Kaplan Meier plot. Overall Survival for intention-to-treat patients (N = 15). Patients with event (N = 12), Patients censored (N = 3)

Fig. 6.

Waterfall plot for evaluable patients with elevated CA19-9 at baseline, showing change of CA19-9 (between baseline and end of treatment). Patient 5 had no baseline assessment and was excluded

Safety

All adverse event relationships to study treatment were assessed and recorded by the treating investigator. The most common any-grade treatment-emergent adverse effects (TEAEs) were: hypercalcaemia (in 11 (73.3%) of patients and all cases related to paricalcitol), fatigue (10 patients, (66.7%), all related to both gemcitabine and nab-paclitaxel, and 5 also related to paricalcitol), anaemia (6 patients, (40.0%), all related to gemcitabine, all but one related to nab-paclitaxel and 2 also related to paricalcitol), diarrhoea (6 patients, 40.0%), all related to both gemcitabine and nab-paclitaxel and 2 related to paricalcitol), thrombocytopenia (5 patients, (33.3%), all related to both gemcitabine and nab-paclitaxel) and Nausea (5 patients (33.3%), all related to both gemcitabine and nab-paclitaxel with 1 related to paricalcitol). All other related TEAEs were experienced by four or less patients.

A total of 13 patients (86.7%) experienced TEAEs with grade ≥ 3. Of these, 2 patients had grade 5 general physical health deterioration and bacterial infection, unrelated to treatment. Three patients had grade 4 TEAEs: platelet count decreases (2 patients, 13.3%) and hypokalaemia (1 patient, 6.7%). The most common TEAEs with grade ≥ 3 were fatigue (5 patients, 33.3%) and anaemia (4 patients, 26.7%). TEAEs of grade 3 and above, attributed to gemcitabine, Nab-paclitaxel or paricalcitol are summarised in Table 2. Hypercalcaemia of grade 3 occurred in 2 patients, both related to paricalcitol only.

Table 2.

Summary of treatment emergent adverse events*, grade 3 and above, attributed to gemcitabine^G, Nab-Paclitaxel^N, or paricalcitol^P

	Grade 3	Grade 4	Grade 5	Total
Patients with any related TEAEs Grade 3 and above	7 (46.7%)	3 (20.0%)	0	10 (66.7%)
Haematological
AnaemiaGNP	4 (26.7%)	0	0	4 (26.7%)
NeutropeniaGNP	4 (26.7%)	0	0	4 (26.7%)
PancytopeniaGN	1 (6.7%)	0	0	1 (6.7%)
ThrombocytopeniaGNP	3 (20.0%)	2 (13.3%)	0	5 (33.3%)
Thrombotic microangiopathyG	1 (6.7%)	0	0	1 (6.7%)
LymphopeniaGN	1 (6.7%)	0	0	1 (6.7%)
Gastrointestinal disorders
Abdominal painN	1 (6.7%)	0	0	1 (6.7%)
DiarrhoeaGNP	2 (13.3%)	0	0	2 (13.3%)
NauseaGN	1 (6.7%)	0	0	1 (6.7%)
Constitutional
FatigueGNP	4 (26.7%)	0	0	4 (26.7%)
Peripheral oedemaGNP	1 (6.7%)	0	0	1 (6.7%)
Weight decreaseGN	1 (6.7%)	0	0	1 (6.7%)
Hepatobiliary
HyperbilirubinaemiaN	1 (6.7%)	0	0	1 (6.7%)
Infection
AbdominalGNP	1 (6.7%)	0	0	1 (6.7%)
InfectionGN	1 (6.7%)	0	0	1 (6.7%)
Rash pustularGNP	1 (6.7%)	0	0	1 (6.7%)
Skin infectionGNP	1 (6.7%)	0	0	1 (6.7%)
Urinary tract infectionGN	1 (6.7%)	0	0	1 (6.7%)
Liver Function Tests
ALT increaseGN	1 (6.7%)	0	0	1 (6.7%)
AST increaseGN	2 (13.3%)	0	0	2 (13.3%)
ALP increaseGN	1 (6.7%)	0	0	1 (6.7%)
GGT increaseGN	2 (13.3%)	0	0	2 (13.3%)
Electrolytes and Metabolism
HypercalcaemiaP	2 (13.3%)	0	0	2 (13.3%)
HypokalaemiaGNP	1 (6.7%)	1 (6.7%)	0	2 (13.3%)
Renal
Acute kidney injuryG	1 (6.7%)	0	0	1 (6.7%)
Skin and subcutaneous tissue
PruritusGNP	1 (6.7%)	0	0	1 (6.7%)

*Treatment Emergent Adverse Event (TEAE) are adverse events with start date on or after the first date of study treatment and up to one month after last date of therapy

Discussion

Pancreatic cancer represents an unmet clinical challenge. Chemotherapy remains largely ineffective and new therapeutic avenues are urgently required. Preclinical studies previously showed VDAs induce normalization of stromal cells, direct anti-proliferative effects on PDAC cells and exert complex effects upon the immune system [33, 34]. VDAs have entered clinical trials in combination with chemotherapy regimens [29]. The authors are aware of 12 clinical trials incorporating paricalcitol for patients with PDAC. Most studies have used IV delivery of paricalcitol, and our study is one of two trials utilizing the oral route. Here we demonstrate that oral paricalcitol (12 mg per day) can be used safely with the standard protocol for gemcitabine and NAB-paclitaxel. Hypercalcaemia did not present a limiting toxicity for treatment and symptoms were managed by rehydration.

Nine previous studies utilised paricalcitol together with gemcitabine and NAB-paclitaxel and several trials have added cisplatin, or immune checkpoint inhibitors into their protocols. Reports have been presented in meeting abstracts [28, 29, 35]. Borazanci and colleagues reported a phase II study [28] where patients with metastatic PDAC, received IV paricalcitol combined with gemcitabine and Nab-paclitaxel, cisplatin and nivolumab (NCT02754726). They found a high response rate with an ORR of 83%; median PFS was 8.17 months and OS 15.3 months, encouraging results compared to gemcitabine and Nab-paclitaxel alone [4]. Our trial is not directly comparable with this earlier study, which included additional therapeutic agents [28].

A more recent study combined paricalcitol with nanoliposomal irinotecan and 5-fluorouracil/ leucovorin (5-FU)/LV) chemotherapy in patients following progression on gemcitabine chemotherapy (NCT03883919) [36]. They found a disease control rate (DCR) of 82%. Median PFS was 3.22 months, close to that previously reported in the Phase III NAPOLI-1 study (3.1 months) comprising nanoliposomal irinotecan and 5-FU/LV without addition of paricalcitol [37]. However, in response-evaluable patients, median PFS was increased at 4.14 and 4.83 months for two dose levels of paricalcitol. Correlative studies found increased vascularity in post-treatment tumours consistent with changes in stromal density, likely due to anti-stromal activity, encouraging further clinical trials combining paricalcitol with 5-FU chemotherapy [36].

Finally, a report published during the preparation of our manuscript found addition of IV paricalcitol to pembrolizumab resulted in no significant differences in PFS or OS. The data remain sparse, and it is not yet possible to compare efficacy of oral versus IV dosing of paricalcitol [38].

The aim of our study was to find a signal for paricalcitol activity in combination with gemcitabine chemotherapy. This was an intention-to-treat (ITT) study. An interim analysis was included to decrease possible non-beneficial interventions, and this resulted in a smaller sample size with reduced power. Our study employs an historical control wherein median PFS was 22 weeks [4] which contrasts with the median survival of 14.6 weeks found in the current study. Hence there was a trend towards decreased median PFS upon addition of paricalcitol though no statistical analysis was performed due to early termination of the study.

This provides preliminary evidence that efficacy of paricalcitol may be lower than anticipated based upon preclinical studies. There are several possible explanations for this outcome, including: (a) subtypes of PDAC may exhibit differential responses to loss of stromal support, and (b) the TME within metastatic sites may differ from primary tumours resulting in a decreased response to VDAs.

Consistent with the first possibility, CAFs support the growth of epithelial PDAC cells, but not poorly differentiated mesenchymal cancers, where growth even may be restrained [39–41]. These authors suggested stromal targeting may result in distinct therapeutic outcomes according to PDAC subtype [41]. Moreover, a preclinical study suggested that direct anti-tumour effects of vitamin D may be restricted to the classical subtype with ineffective or even pro-tumour effects in basal cells [42], although this has not been reported in clinical studies.

A second possibility is that stromal cells in hepatic metastases are less responsive to paricalcitol than CAFs in primary tumours. Preclinical research validated PSCs as a sensitive target for VDAs such as calcipotriol [24]. Hepatic metastases exhibit extensive desmoplasia in PDAC [42], and the assumption has been that hepatic stromal cells, closely related to PSCs, would exhibit similar sensitivity to VDAs [43]. However, a recent study finds liver metastases contain stromal elements with a pericyte-like state in contrast to the fibroblast-like gene expression found in primary PDAC [7] and this may result in differential responses to therapeutic interventions. Both of these explanations remain unproven, and it is not clear whether PDAC or stromal subtype is related to response to anti-stromal agents in clinical studies [43, 44].

The rationale for incorporation of paricalcitol into clinical trials has been based upon well-established anti-stromal effects. Preclinical studies have demonstrated that multiple VDAs (including calcipotriol, VDR ligand I5 and paricalcitol) exert anti-tumour activity and enhance therapeutic efficacy of gemcitabine within xenograft and orthotopic mouse models [24, 44, 45]. Moreover, encouraging data are now beginning to emerge from clinical trials that have incorporated chemotherapeutic regimens in addition to or in place of gemcitabine, suggesting that a combination of paricalcitol with 5-FU may be a useful approach [28, 36].

Conclusions

The role of anti-stromal agents as cancer medicines remains controversial. Our study did not detect a signal for paricalcitol activity in combination with gemcitabine-based chemotherapy in patients with metastatic PDAC. If subtypes display differential sensitivities, then some clinical trials (including the current) may be underpowered to detect therapeutic effects. One approach to further understand the role of VDAs in PDAC might be a for a meta-analysis combining data from multiple trials involving paricalcitol. Where available, clinical samples could be used to define PDAC and CAF subtype and determine whether this is related to outcome of treatment with paricalcitol. Our study has several limitations. First, the number of patients was small and secondly, we used an historical control. Thus, our findings must be considered exploratory. Further, larger studies will be required to determine patient response to VDAs and data from multiple on-going clinical trials are eagerly awaited. The inclusion of post-treatment biopsies within protocols of some of these studies offers an opportunity to uncover prognostic molecular biomarkers. In the future, these could be used to stratify patients for therapy and may yield more powerful anti-tumour responses in studies incorporating stromal targeting agents.

Abbreviations

CA-19.9

Carbohydrate antigen 19-9

CAF

Cancer associated fibroblast

CI

Confidence interval

CT

Computed tomography

CTCAE

Common terminology criteria for adverse events

DCR

Disease control rate

ECM

Extracellular matrix

ECOG

Eastern cooperative oncology group

FOLFIRINOX

5-fluorouracil/ leucovorin/ irinotecan/ oxaloplatin

5-FU/LV

5-Fluorouracil/ leucovorin

ITT

Intention-to-treat

LRRC15

Leucine-rich repeat-containing 15

myCAF

Myofibroblast Cancer associated fibroblast

NAB

Nanoparticle albumin bound

NCI

National cancer institute

NAPOLI

Nanoliposomal irinotecan, fluorouracil and folinic acid

ORR

Overall response rate

OS

Overall survival

PDAC

Pancreatic ductal adenocarcinoma

PFS

Progression free survival

PRRX1

Paired related homeobox 1

PSC

Pancreatic stellate cells

RECIST

Response evaluation criteria in solid tumours

RNA-Seq

RNA sequencing

SAS

Statistical analysis system

TEAE

Treatment-emergent adverse effects

TME

Tumour microenvironment

TTF

Time to treatment failure

TRR

Tumour response rate

VDA

Vitamin D agonist

Authors’ contributions

Development of the concept and design: BH, DE; Development of methodology BH, DE, IP, VP; Writing, review and revision of manuscript, preparation of figures: BH, DE, VP, AS, MN; All authors reviewed the final manuscript.

Funding

This study was supported by grants from GATEWAY for cancer research, Illinois, USA (G-15-2200), the Pat Smullen Pancreatic Cancer Research Fund, Friends of Brian Lenihan Fund and Cancer Trials Ireland, Dublin, Ireland.

Data availability

Formal requests for data sharing are considered in line with Cancer Trials Ireland procedures with due regard given to patient consent. Data requests must include a description of the research proposal. Data sharing requests will be reviewed in terms of scientific merit and ethical considerations including patient consent. Data recipients are required to enter a formal data sharing agreement, which describes the conditions for release and requirements for data transfer, storage, archiving, publication, and intellectual property.

Declarations

Ethics approval and consent to participate

The study was approved by the Health Products Regulatory Authority and the National Research Ethics Committee. All patients provided informed consent to participate.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.