Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a form of exocrine pancreatic cancer that accounts for more than 90% of all pancreatic cancer diagnoses and is the 12th most common cancer worldwide (Sung et al. 2021). The 5-year-survival rate is approximately 10% (Park et al. 2021). PDAC is usually diagnosed at advanced stage, with 30–35% of patients presenting with locally advanced disease and 50–55% of patients with metastatic disease (Park et al. 2021).

Pancreatic intraepithelial neoplasms acquire cumulative genetic insults which instigate oncogenes, such as KRAS, CDKN21, TP53 and SMAD4 that promote cancer development (Park et al. 2021). These mutated genes prevent innate tumour suppression and result in altered growth signalling and inflammation (Park et al. 2021). They also provide targets for oncolytic virotherapy, as they allow viruses to selectively replicate inside cells with such faulty pathways.

Surgical resection is the only curative alternative, although most patients present with unresectable disease. Current standard-of-care chemotherapy includes various combinations of gemcitabine, FOLFIRINOX, nab-paclitaxel, capecitabine and erlotinib (Principe et al. 2021) and has improved 5-year survival, with rates increasing over the past several decades, from 3.2% in 1987 to 10.7% in 2016 (Welfare 2022). This rate is significantly lower than the average 5-year survival rate achieved for all cancers, which was 69% in 2016 (Welfare 2022).

This high mortality rate is posited to be due to both the late stage at presentation, and the high resistance of PDAC to conventional chemotherapy. This resistance is, at least, partly due to the dense extracellular matrix that characterises PDAC. This environment distorts vessels, leading to hypoxic tumours that are inaccessible to chemotherapy which is typically delivered through the bloodstream. New approaches are needed to overcome the hostile PDAC microenvironment that may result in more effective treatments to almost 500,000 people diagnosed with pancreatic cancer every year (Sung et al. 2021).

Various forms of alternative/complementary treatments are being actively investigated. Radiation is being explored as an adjunct to chemotherapy option for locally advanced PDAC (Katz et al. 2017; Loehrer et al. 2011; Versteijne et al. 2020). Immunotherapies, including immune checkpoint inhibitors, immune agonists, cytokines and vaccines are also being explored as novel methods to overcome the immunosuppressive PDAC tumour microenvironment. These approaches have been described in recent reviews (Timmer et al. 2021; Zhu et al. 2022).

Oncolytic viruses offer an alternative approach to overcoming the challenges associated with treating PDAC. Clinical trials of oncolytic adenoviruses offer hope in improving outcomes for patients with this challenging cancer. Until recently, this approach was significantly limited by the very cumbersome and difficult intratumoral route of administration. Recent clinical trials using the intravenous route now facilitate the evaluation of this unique approach. This paper explores the use of oncolytic adenoviruses in clinical trials to date, with a brief summary of the preclinical trials.

Oncolytic adenoviruses

Oncolytic viruses, or oncoviruses, preferentially kill tumour cells. They can be created through genetic modifications of existing viruses to ensure they selectively bind to and replicate in cancer cells. Several oncolytic viruses have been trialled in preclinical and clinical models of pancreatic cancer (Hamidi-Sofiani et al. 2022). Their biology has been utilised in two distinct ways.

Replication-incompetent viruses fail in one or more functions required for their replication, synthesis or the assembly of viral particles, and can be used as vaccines or vaccine vectors. They may be used to deliver cytokines or other cytotoxic molecules into the cancer cells. Alternatively, replication-competent viruses, capable of replicating, synthesising and assembling viral progeny can infect and reproduce effectively in cancer cells, leading them to elimination.

To optimise the use of oncoviruses for cancer therapy, two focal points are being improved upon in current research: selectivity and efficiency. The selectivity for cancer cells is improved by deleting the genes required to replicate in non-cancerous cells, by transcriptional or transductional targeting. Transcriptional targeting involves the use of tissue specific promoters on the virus to control replication. Transductional targeting is the retargeting of the virus specifically to tumour cells. These targeting strategies prevent oncoviruses from replicating in healthy cells, thereby making the virus selective for tumour cells. The efficiency of oncoviruses relies upon 3 factors; the ability to directly lyse cancer cells, the ability to migrate to surrounding cells and the ability to indirectly lyse cancer cells by promoting immune-mediated tumour cell death.

One family of viruses being investigated for use in PDAC are adenoviruses. These double-stranded DNA viruses are encompassed by an icosahedral protein capsid and can undergo mutations in order to improve their selectivity and efficiency as oncoviruses. Such recombinant adenoviruses enter cells when the fiber protein, which forms spikes on the corners of the capsid, binds to the coxsackie adenovirus receptor (CAR) with higher affinity than the receptor binds to itself, therefore hindering cell–cell adhesion and disrupting the tight junction. Inside the cell, the adenovirus genome is released and replicates until the cell disintegrates, hence lysing the cell and allowing progeny viruses to infect and lyse further cells.

Both deletional and transcriptional mutations have been used to create adenoviruses that are both selective for PDAC cells, and efficient at oncolysis, as have been previously reviewed (Nattress and Hallden 2018).

This review explores the modified adenoviruses that have been used in preclinical studies, as well as compiling and comparing reported clinical trials that have used oncolytic adenoviruses in PDAC.

Summary of preclinical studies

Oncolytic adenoviruses have been studied in numerous preclinical models of pancreatic cancer. Of particular relevance are those modifications which have conferred increased tumour-selectivity and oncolytic efficiency, prompting their use in the design of clinical trials. Supplementary Table S1 shows a comprehensive list of the modifications explored in pre-clinical trials which have been reviewed further in recent papers (Nattress and Hallden 2018; Sato-Dahlman and Yamamoto 2018).

Of the various preclinical findings, some have been used in clinical trials. For example, deletions in the E1B region of the adenoviral genome improved tumour-selectively, as in ONYX-015 (Hecht et al. 2003; Mulvihill et al. 2001). The inclusion of the adenoviral death protein increases oncolytic activity and has been used in a clinical trial with Ad5-DS (Lee et al. 2020). Ad5-DS and a similar adenovirus, Ad5-yCD/mutTKsr39rep-hIL-12 have a yeast cytosine deaminase (yCD)/mutant sr39 herpes simplex virus thymidine kinase fusion gene which allows conversion of co-administered prodrugs to their active form within tumour cells, improving tumour cell lysis (Barton et al. 2021; Lee et al. 2020). In addition, the success of arming viruses with various cytokines has led their use in clinical trials. One example is Ad5-yCD/mutTKsr39rep-hIL-12 which harbours the immunostimulatory IL-12 (Barton et al. 2021).

E1A deletions have also been used in two clinical trials of VCN-01, as this deletion prevents viral replication in cells with functional pRb, therefore improving tumour-selectivity (Bazan-Peregrino et al. 2021; Garcia-Carbonero et al. 2022). VCN-01 also harbours RGD to allow binding and entry of the virus into cells with down-regulated CAR, such as tumour cells, therefore improving infectivity. Moreover, VCN-01 harbours E2F1 promoters in the E1A region to increase viral replication in tumour cells. Finally, human sperm hyaluronidase degrades the dense extracellular matrix of the tumour and was included in VCN-01 in recent clinical trials (Bazan-Peregrino et al. 2021; Garcia-Carbonero et al. 2022).

Ongoing clinical trials are also exploring the use of immunostimulatory modifications, such as CD40 ligand, 4-1BB ligand and interferon beta.

Summary of clinical data

The great number of preclinical studies of adenovirus in models of pancreatic cancer have led to six clinical trials published to date, the majority of which in phase 1. Table 1 summarises the main features of each of these studies.

Table 1 Clinical trials exploring the use of oncolytic adenoviruses in PDAC
Full size table

Characteristics of the clinical trials

Patient demographics

The cohort size in the studies ranged from 8 to 23 patients, and the median ages were similar, ranging from 60 to 68. All studies included both male and female patients with a male to female ratio ranging between 1:3 to 3:1. The functional status (expressed as the mean Eastern Cooperative Oncology Group (ECOG) score) varied from 0.33 to 0.67 between the studies. Arguably the most significant difference in patient demographics was in the tumour burden at the time of enrolment in the study. Barton et al. (2021) only included patients with distant metastases at baseline, whereas Lee et al. (2020) excluded all patients with distant metastases. The other 4 studies included patients with both locally advanced and metastatic disease (Table 2).

Table 2 Patient demographics
Full size table

Treatment

Previous treatments

Two studies (Lee et al. 2020; Barton et al. 2021) only included previously untreated patients, whereas all other studies included patients who had received previous chemotherapy, radiation, surgery or a combination of these therapies. The number of patients who had received no prior treatment ranged from 26 to 100% (Table 3).

Table 3 Previous treatments
Full size table

Adenovirus

The two earliest studies (Mulvihill et al. 2001; Hecht et al. 2003) investigated ONYX-015, which is an adenovirus with a deletion of the E1B-55 kDa region which allows it to replicate selectively inside cancer cells with abnormalities in the p53 pathway, including pancreatic ductal adenocarcinoma cells. This virus has been shown to be effective in preclinical models of PDAC (Bischoff et al. 1996; Heise et al. 1997).

The subsequent two studies (Lee et al. 2020; Barton et al. 2021) used similar adenoviruses. Lee et al. (2020) investigated Ad5-DS, which is an adenovirus with two suicide genes in the E1 domain: yeast cytosine deaminase and herpes simplex virus 1 thymidine kinase. These genes convert the prodrugs 5-FC and valganciclovir to 5-fluorouracil and valganciclovir-5-monophosphate respectively. These active metabolites lyse tumour cells. The Ad5 adenoviral death protein gene in the E3 region enhances the virus’s oncolytic capacity. Barton et al. (2021) uses Ad5-yCD/mutTKSR39rep-hIL-12, which expresses the two suicide genes as used in Lee et al. (2020), as well as human IL-12. The IL-12 acts to stimulate the immune system to improve oncolysis.

Finally, the most recent two studies (Bazan-Peregrino et al. 2021; Garcia-Carbonero et al. 2022) investigated VCN-01, another type 5 adenovirus that selectively replicates inside cells with retinoblastoma protein pathway deregulation, such as pancreatic ductal adenocarcinoma cells. The integrin-binding motif RGDK replaces the heparin sulfate glycosaminoglycan putative-binding site KKTK of the adenovirus 5 fiber, which increases tumour selectively and decreases liver tropism. Furthermore, VCN-01 expresses human sperm hyaluronidase that degrades the dense extracellular matrix that characterises pancreatic tumours to facilitate viral dissemination as well as chemotherapy infiltration.

Administration route

Mulvihill et al. (2001) used intratumoural administration, either under CT guidance or intra-operative injection. The subsequent 4 trials (Barton et al. 2021; Bazan-Peregrino et al. 2021; Hecht et al. 2003; Lee et al. 2020) used endoscopic ultrasound-guided intratumoural administration. The first trial to use this approach was Hecht et al. (2003). In this study both the trans-gastric and trans-duodenal approaches were utilised for virus administration, though two duodenal perforations occurred and subsequently the protocol was changed to mandate the trans-gastric approach. Finally, Garcia-Carbonero et al. (2022) was the first study to use systemic administration of an adenovirus for pancreatic cancer, via the intravenous route. Though there is a greater potential for systemic adverse events, this approach is less invasive and is more practical, especially if multiple administrations are to be used.

Dose

The first study of ONYX-015 in 2001 used doses between 2 × 109 and 2 × 1012 viral particles and the maximum tolerated dose was not reached. The following study of ONYX-015 in 2003 used similar doses of 2 × 1010 and 2 × 1011 viral particles, with most patients (18 out of 21) receiving the higher dose. The maximum tolerated dose was also not reached.

In 2020, Ad5-DS was used at doses between 1011 and 1012 viral particles in Lee et al. (2020). The maximum tolerated dose was not reached. A similar adenovirus was studied in Barton et al. (2021) at the same doses as above and again the maximum tolerated dose was not reached.

The final adenovirus, VCN-01 was used at relatively lower doses, 1010 and 1011 viral particles in Bazan-Peregrino et al. (2021), and in this case the maximum tolerated dose was deemed to be 1011 viral particles. This suggests that VCN-01 causes more severe side effects at a lower dose than previously studied adenoviruses.

However, Garcia-Carbonero et al. (2022) studied VCN-01 at higher doses of 3.3 × 1012 and 1013 viral particles and the maximum tolerated dose was 100 times greater than in Bazan-Peregrino et al. (2021), at 1013 viral particles (Fig. 1).

Fig. 1
figure 1

Doses of adenovirus used. Formatted using SigmaPlot version 15. The doses of adenovirus used in each clinical trial are shown on a logarithmic scale and in chronological order of the trial’s publication date. Most trials used two or three doses, whereas the first trial used six. Garcia-Carbonero et al. 2022 uses the largest doses of adenovirus of all of the trials and was the only trial to utilise the intravenous route

Full size image

Concomitant treatment

The protocol in Mulvihill et al. (2001) did not include concomitant treatments. The subsequent 5 papers included gemcitabine alone (Hecht et al. 2003; Bazan-Peregrino et al. 2021), gemcitabine, 5-FC and valganciclovir (Lee et al. 2020), 5-FC and FOLFIRINOX (Barton et al. 2021), 5-FC, gemcitabine and nab-paclitaxel (Barton et al. 2021) or gemcitabine with nab-paclitaxel (Bazan-Peregrino et al. 2021; Garcia-Carbonero et al. 2022). Lee et al. (2020) also reported that eight out of nine patients received stereotactic body radiation therapy during the study.

Barton et al. (2021) and Bazan-Peregrino et al. (2021) were the only studies that investigated the use of different concomitant chemotherapeutic drugs. Some patients in Barton et al. (2021) received FOLFIRINOX in place of gemcitabine and nab-paclitaxel, with both groups receiving 5-FC as well. In Bazan-Peregrino et al. (2021), some patients received nab-paclitaxel in addition to gemcitabine, whilst the others received gemcitabine as a single concomitant treatment.

Interestingly, Garcia-Carbonero et al. (2022) was the only study that investigated whether the timing of the concomitant chemotherapy in relation to the administration of the adenovirus would affect the success of the treatment or the adverse events. In this study, one arm (Part 2) received nab-paclitaxel and gemcitabine on the same day as the intravenous adenovirus infusion. In the other arm (Part 3), the adenovirus infusion occurred 1 week prior to the administration of the chemotherapy. This delay resulted in significantly fewer grade 3 or greater adverse events in Part 3 compared to Part 2.

Treatment schedule

Treatment schedules differed greatly between the six studies. In Mulvihill et al. (2001), ONYX-015 was administered every 4 weeks without concomitant chemotherapy. In Hecht et al. (2003), ONYX-015 was administered weekly as a single agent for 4 weeks and then in combination with gemcitabine for 4 weeks.

Patients in Lee et al. (2020) received treatment until disease progression or unacceptable toxicity. Each cycle consisted of a single Ad5-DS administration, then gemcitabine infusions two days, 9 days and 16 days later. Prodrugs 5-FC and valganciclovir were administered daily starting 2 days after adenovirus administration. The groups receiving the two lower viral doses had 7 days of prodrugs, whilst the group receiving the higher dose had 14 days.

In Barton et al. (2021), all patients received 7 days of 5-FC, beginning 48 h after adenovirus injection. In the 21 days following adenovirus injection, patients began receiving standard-of-care chemotherapy, such as FOLFIRINOX or gemcitabine with albumin-bound paclitaxel, at the discretion of the treating physician.

In Bazan-Peregrino et al. (2021), the lower adenoviral dose group received three injections of VCN-01 on days 1, 22 and 43 and gemcitabine on days 1, 8, 15, 22, 29, 35 and 43. Standard-of-care chemotherapy had changed to gemcitabine with nab-paclitaxel for the higher dose group, who received three injections of VCN-01 on days 1, 29 and 57, with gemcitabine and nab-paclitaxel on days 1, 8, 15 and then every 4 weeks. After day 56, patients continued to receive standard-of-care chemotherapy until disease progression, unacceptable toxicity, withdrawal of consent or at the discretion of the investigator.

Garcia-Carbonero et al. (2022) was the first study to explore the effects of a delayed administration of chemotherapy. In Part 2, the group received VCN-01 on the same day as chemotherapy, whereas the Part 3 group received the chemotherapy 7 days after the VCN-01 infusion. The schedule for Part 2 was a single injection of VCN-01 on day 1 with infusions of gemcitabine and nab-paclitaxel on the same day, followed by infusions of gemcitabine and nab-paclitaxel on days 8 and 15 of a 28-day cycle. Patients in Part 3 received VCN-01 on day 1, with infusions of gemcitabine and nab-paclitaxel on days 8, 15 and 22 of a 35-day cycle. Subsequent cycles for both groups were 28 days, with nab-paclitaxel and gemcitabine on days 1, 8 and 15.

Adverse events

The oncolytic treatment was generally well tolerated with a low incidence of severe adverse events across the trials. It is difficult to analyse the trend in adverse events due to the different systems used to define adverse events and the differences in the reporting of adverse events in the six trials.

The system used to define treatment adverse events was not reported in the first two studies (Mulvihill et al. 2001; Hecht et al. 2003). Moreover, these two studies did not report the total number of adverse events but did explain that the majority of adverse events were not severe.

In the 4 subsequent papers, the Common Terminology Criteria for Adverse Events (CTCAE) grading was used.

In Lee et al. (2020), Bazan-Peregrino et al. (2021) and Garcia-Carbonero et al. (2022), the adverse events were reported according to the dose of adenovirus that the patients received. The other three papers did not classify the adverse events according to which dose level the patients had received (Table 4).

Table 4 Total adverse events (AEs) reported
Full size table

In Lee et al. (2020), there were no adverse events reported in the patients who received the lowest viral dose. In the two higher doses, there were approximately 0.66 adverse events per patient on average, all of which were grade 1 or 2 in severity.

Barton et al. (2021) reported the greatest number of adverse events, with an average of approximately 10.08 events per patient. The vast majority of these were grade 1 or 2, with only 0.58 grade 3–4 adverse events per patient.

Bazan-Peregrino et al. (2021) interestingly reported a greater proportion of adverse events in the group who used the lower dose of the adenovirus compared to the higher dose (8.00 vs. 3.17 adverse events per patient). The proportion of grade 3 + severity events was greater in the group who received the higher dose of adenovirus (0.84 vs. 0.50 events per patient). Importantly, the group using the lower dose was comprised of only 2 patients, compared with the larger cohort of 6 patients in the higher dose group. Three dose-limiting toxicities occurred and so the maximum tolerated dose was deemed to be 1011, despite all other studies either not reaching the maximum tolerated dose or having a significantly higher maximum tolerated dose.

In Garcia-Carbonero et al. (2022), there were more adverse events reported in the group who received the adenovirus and the chemotherapy concurrently, compared to the group in which the adenovirus preceded the chemotherapy by 1 week (3.67 vs. 2.07 adverse events per patient). This trend was consistent across all grades of adverse events. Of particular importance, the group with the delayed chemotherapy regimen had significantly fewer reported grade 3 + adverse events (0.21 vs. 0.83 grade 3 + adverse events per patient).

The most commonly reported adverse events can be found in Table 5.

Table 5 Frequency of commonly reported adverse events
Full size table

Though these adverse events were common, very few patients experienced grade 3 + events.

It is important to note that Garcia-Carbonero et al. (2022) is the only study in which the adenovirus was administered systemically. Despite this, the proportion of total adverse events per patient, and the proportion of grade 3 + adverse events were less than those reported in the previous two studies, Bazan-Peregrino et al. (2021) and Barton et al. (2021).

Clinical outcomes

The earlier studies in 2001 and 2003 investigated ONYX-015 and showed no objective responses observed when administered as a sole treatment. It was noted in Mulvihill et al. (2001) that whilst systemic dissemination of the virus occurred rapidly after administration, viral replication was not detectable thereafter. Perhaps this was due to the inflammatory and fibrotic PDAC microenvironment inhibiting viral spread, or due to hepatic or immunological clearance. In Hecht et al. (2003), gemcitabine was given alongside the final four administrations of adenovirus, and 9.5% of patients experienced partial regressions after this combination therapy, suggesting that dual chemotherapy-adenovirus treatments are more effective than adenovirus treatments alone. The response rate for gemcitabine alone, however, is 7% (Von Hoff et al. 2013), showing an inconsequential difference in response rates between chemotherapy alone and with adenovirus. All subsequent studies explored adenoviruses in combination with chemotherapy (Tables 6, 7, 8).

Table 6 Overall survival in the published trials, compared to standard treatments
Full size table
Table 7 Patient responses in the clinical trials that used RECIST V.1.1 criteria
Full size table
Table 8 Comparison of the patient responses in the clinical trials that used criteria other than RECIST v.1.1
Full size table

The more standardised evaluation of clinical outcomes (RECIST) has been applied to four of the trials and they yielded some encouraging data. A different adenovirus (Ad5-DS) was explored in Lee et al. (2020) and 11.1% of patients experienced a partial response, which where the most promising results yet. This study did, however, exclude patients with metastatic disease and therefore the participants had better prognoses than participants in the other studies. Adenoviral DNA was detected in serum at 8 weeks in 44.4% of patients, suggesting viral replication had occurred.

In contrast, Barton et al. (2021) investigated a similar adenovirus to Lee et al. (2020), but exclusively recruited patients with distant metastases. The poor prognosis of the patients included likely contributed to the poor survival and response rates seen, with no objective responses reported.

Bazan-Peregrino et al. (2021) and Garcia-Carbonero et al. (2022) both explored another different adenovirus, VCN-01. When administered via the intratumoural route in Bazan-Peregrino et al. (2021), 12.5% of patients had a minor response and the mean time to progression was 8.4 months. In contrast, Garcia-Carbonero et al. (2022) was the first study to investigate the less invasive intravenous route as well as the effect of administering the adenovirus on the same or different day as the chemotherapy. In the group that received VCN-01 and chemotherapy on the same day, a higher objective response rate (10% compared to 0%) and a longer median time to progression (9.9 months compared to 6.7 months) were observed. The objective response rates of both groups in Garcia-Carbonero et al. (2022) were less than that of Bazan-Peregrino et al. (2021), and so perhaps the intravenous route of administration is less effective than the intratumoural route. Despite this, a second peak in plasma viral concentration as well as increased serum hyaluronidase was seen in Garcia-Carbonero et al. (2022), suggesting viral replication occurred in all patients.

While more difficult to compare, there are encouraging data arising from trials where criteria other than RECIST were used for the evaluation of clinical outcomes. In Mulvihill et al. (2001), a minor response is defined as a 25–49% decrease in the cross-sectional area of the tumour. Stable disease is defined as an increase or decrease of less than 25% in the cross-sectional area of the tumour. Progressive disease is defined as an increase of equal to or greater than 25% in the cross-sectional area of the tumour.

Hecht et al. (2003) reports that responses were evaluated using standard WHO criteria. Using these criteria, a partial response involves a decrease of greater than or equal to 50% in the sum of products of the two longest diameters in perpendicular dimensions (Park et al. 2003). This is a significantly greater response than is defined as a minor response in Mulvihill et al. (2001).

Discussion

Preclinical studies of oncolytic adenoviruses have shown promise of these agents in models of pancreatic cancer. Clinical trials offer encouraging results, including the favourable safety profiles of the four adenoviruses used. The most common adverse events included fever, nausea and asthenia, which are events commonly associated with the current standard treatment for pancreatic cancer, gemcitabine. Moreover, the majority of the reported adverse events were grade 1–2, suggesting the severity of such events was mild and patients tolerated the adenoviruses well. Of significance, the intravenous route utilised in Garcia-Carbonero et al. (2022) appears to be safe and well-tolerated. This trial did not report the highest number of adverse events, as might be expected with systemic administration of the adenovirus. Additionally, the higher dose used in the trial was up to 5000 times greater than doses used previously, yet the number and severity of adverse events was in keeping with the previous trials. The patient responses were also encouraging with the majority of patients achieving at least stable disease throughout the surveillance period. Survival rates varied significantly even between patients receiving the same viral dose and treatment schedule, perhaps owing to the small sample sizes of the trials, individual variation in disease severity or other unreported factors. Larger studies are needed to explore the effectiveness of this therapy in more detail now that the safety and tolerability of intratumoural and intravenous adenovirus administrations have been established.

Despite the promising results shown in the published trials, there are challenges of the human PDAC microenvironment that pose barriers to treatment with oncoviruses and have limited the effectiveness of this treatment. Firstly, viral selectivity is reduced when there is widespread distribution of primary cellular receptors as systemic toxicity limits the viral dose that may be tolerated.

The route of administration poses another challenge, as the commonly used intratumoural administration is relatively invasive and less suitable to target distant metastases than the intravenous route. This method itself poses the theoretical significant risk of liver tropism, which could cause considerable adverse effects and reduce the availability of adenovirus in the circulation. However, Garcia-Carbonero et al. (2022) has now shown that an oncolytic adenovirus may be administered via the intravenous route at doses significantly greater than previously used, whilst remaining tolerable with comparable adverse effects to the intratumoural delivery.

The significant host immune response triggered by adenoviruses can enhance oncolysis in the local tumour environment but can also reduce efficiency and safety when a systemic reaction occurs. Modifying the adenovirus such that is largely or exclusively delivered into tumour cells is necessary to prevent such adverse reactions.

Furthermore, the dense fibrotic microenvironment that characterises PDAC prevents traditional chemotherapeutic agents from penetrating the tumour, and similarly prevents efficient adenovirus replication and spread. This issue is difficult to overcome, but Bazan-Peregrino et al. (2021) and Garcia-Carbonero et al. (2022) have attempted to overcome this through expression of hyaluronidase on the oncolytic adenovirus to degrade the extracellular matrix and facilitate the spread of viral progeny and chemotherapies throughout the tumour. A phase III trial of pegvorhyaluronidase, however, showed no improvement in overall survival or progression free survival in patients with metastatic pancreatic cancer, posing doubts about the utility of viral hyaluronidase expression (Van Cutsem et al. 2020).

Whilst there are many challenges to overcome, conditionally-replicative adenoviruses offer an innovative therapy for a cancer that is difficult to treat.

Further advances in the construction and administration of adenoviruses aim to build upon the promising results seen in the clinical trials. For example, the adenovirus library approach, which involves the display of random peptides on a fiber knob, allows generation of adenovirus that can target specific cell types, allowing researchers to synthesise adenoviruses with affinities for specific cancers (Yamamoto et al. 2014).

Arming adenoviruses with cytokines has also been shown to increase the oncolytic ability of oncoviruses, without compromising safety.

Ongoing clinical trials are summarised in supplementary table 2. These trials include adenoviruses which have been trialled before, such as Ad5-yCD/mutTKSR39rep-ADP which was studied in Lee et al. (2020), and adenoviruses which have not yet been explored in clinical trials of pancreatic cancer.

Many other strategies are also being explored to improve the efficacy of oncolytic adenoviruses without modifying the viral genome. For example, administering the virus within mesenchymal stromal carrier cells, monocytes and lymphocytes encourages viral production by reducing the immune response to the virus (Bunuales et al. 2012; Hammer et al. 2015; Kaczorowski et al. 2016; Na et al. 2019; Santos et al. 2021). Creating a complex of oncolytic adenovirus and cationic nanoparticles also enhances infection and spread (Man et al. 2022). Oncolytic adenoviruses can be chemically conjugated with neurotensin-conjugated PEF to reduce immunogenicity and improve tumour selectivity through the neurotensin receptor (Na et al. 2015). Hydrogel-based delivery of oncolytic adenoviruses is being explored as an additional method to overcome the immune reaction to systemic administration of adenovirus (Du et al. 2022).

Finally, alternative therapies, such as domperidone, have been shown to enhance the replicative ability of oncolytic adenoviruses in PDAC (Nishimae et al. 2022). At the preclinical stage, there are a myriad of avenues by which the oncolytic efficacy of adenoviruses is being improved in order for these viruses to offer a practical and effective option in the treatment of pancreatic cancer. The developments already tested in the clinical trials reported here have yielded very encouraging results whereby significant clinical objective responses have been achieved in a disease which progression is normally linear. A combination of novel viral modifications and the evaluation of safer and effective routes of administration to be tested in the ongoing clinical trials is likely to further enhance the potential of this therapeutic approach.