Image-guided intratumoral immunotherapy: Developing a clinically practical technology

Abstract

Immunotherapy has revolutionized the contemporary oncology landscape, with durable responses possible across a range of cancer types. However, the majority of cancer patients do not respond to immunotherapy due to numerous immunosuppressive barriers. Efforts to overcome these barriers and increase systemic immunotherapy efficacy have sparked interest in the local intratumoral delivery of immune stimulants to activate the local immune response and subsequently drive systemic tumor immunity. While clinical evaluation of many therapeutic candidates is ongoing, development is hindered by a lack of imaging confirmation of local delivery, insufficient intratumoral drug distribution, and a need for repeated injections. The use of polymeric drug delivery systems, which have been widely used as platforms for both image guidance and controlled drug release, holds promise for delivery of intratumoral immunoadjuvants and the development of an in situ cancer vaccine for patients with metastatic cancer. In this review, we explore the current state of the field for intratumoral delivery and methods for optimizing controlled drug release, as well as practical considerations for drug delivery design to be optimized for clinical image guided delivery particularly by CT and ultrasound.

1. Introduction

1.1. Overview of immune checkpoint inhibitor therapy

Since the turn of the century, systemic immunotherapy with immune checkpoint inhibitors (ICI) has offered an exciting and powerful route for targeting cancerous cells by stimulating immune response through the modulation of immune checkpoints. [1,2] Under normal physiologic conditions, activation of the adaptive immune system, by for example the detection of a pathogen, concomitantly activates a series of checkpoint molecules on the immune cells. These “off switches” ensure an eventual denouement of the immune response and prevent a runaway inflammatory response. Tumors, however, capitalize on this checkpoint system by inappropriately activating these signaling pathways before the immune cells can mount an anti-tumor response. By abrogating this inhibitory mechanism, ICIs unfetter cytotoxic lymphocytes and allow for an anti-tumor response to be mounted and sustained. Multiple reviews have gone in depth through the mechanisms of ICI therapy, and a summary of the mechanism of checkpoint blockade is presented in Figure 1. [1–3]

Figure 1.

Mechanisms of immunotherapy. From left to right, in a recruitment phase cytokines, TLR-9 agonists, STING agonists, etc, induce IFN-y release locally leading to dendritic cell recruitment and activation in the tumor region. In the middle, priming phase, these dendritic cells activate T-cells, and where anti- CTLA-4 antibodies block T-cell and dendritic cell inhibitory signals, allow activation. Subsequently PD-1 agonists inhibit inhibitory signals from the tumor cells. Created with BioRender.com.

Unfortunately, most patients will not respond to ICI treatment; the efficacy rate for ICI across the spectrum of cancer types is less than 20% (Table 1).[4] Multiple resistance mechanisms to ICI have been elucidated. For example, one potential biomarker is the tumor mutational burden; tumors with a low rate of mutations are less likely to generate unique proteins that will yield neoantigens for which a tumor-specific adaptive immune response can be generated. Furthermore, it is also clear that the pre-existing tumor immune microenvironment plays an important prognostic role in ICI efficacy. Tumors with an “immune inflamed” phenotype at baseline, characterized by T cell and myeloid cell infiltration within the tumor, are much more likely to respond to systemic immunotherapies compared to tumors that are “immune excluded” or “immune deserts” in which immune cells are marginalized at the tumor-parenchyma border or not present at all, respectively. [5,6]Adjuvant interventions that can convert the latter two phenotypes (also known as “cold” tumors) into the former (also known as “hot” tumors) have the potential to convert ICI non-responders into responders. This potential comprises the rationale behind intratumoral immunotherapy studies. [7,8]

Table 1:

Currently available anti-PD-1 and Anti-CTLA-4 immunotherapies, table adapted from Seidel et. al. “Anti-PD-1 and Anti-CTLA-4 Therapies in Cancer: Mechanisms of Action Efficacy, and Limitations. Creative Commons License [3]

Target	Drug	Condition	Objective Response Rate	Complete Response Rate	Overall Survival (months)	Participants	Reference
PD-1	Nivolumab	Melanoma	43.70%	8.90%	n/a	316	[133]
		RCC (metastatic)	25% (4% Control)	1% (<1% control)	25.0 (19 control)	406 (397 control)	[134]
		Hodgkin’s Lymphoma	87%	17%	n/a	23	[135]
		SCC - Head and Neck	13.3% (5.8%)	2.5% (0.8% control)	36%/ 1 year (16.6% control)	240 (121 control)	[136]
		NSCLC	19% (12% control)	1% (<1% control)	12.2 (9.4 control)	292 (290 control)	[137]
		Ovarian Cancer	15%	10%	20	20	[138]
	Pembrolizumab	Melanoma	33.7–32.9%	5–60.1%	n/a	279–277	[139]
		Merkel cell carcinoma	56%	16%	n/a	26	[140]
		Non-small cell lung cancer	19.40%	n/a	12	495	[141]
		Progressive metastatic colorectal carcinoma	40%	0	>5 months/ 5	10 (18 control)	[142]
	Pidlizumab	B cell lymphoma	51%	34%	85% at 16 months	66	[143]
		Follicular lymphoma	66%	52%	n/a	29	[144]
Anti-CTLA4	Ipilimumab	Melanoma (stage III/IV)	15.2% (10.3% control)	1.6% (0.8% control)	11.2 (9.1 control)	250 (252 control)	[145]
	Tremelimumab	Melanoma (stage III/IV)	10.7% (9.8% control)	3% (2% control)	12.6 (10.7 control)	326 (327 control)	[146]
Combination Therapy	Nivolumab + Ipilimumab	Melanoma (stage III/IV)	57.60%	11.50%	n/a	314	[133]
		Small cell lung cancer (recurrent)	23%	2%	7.7	61	[147]

1.2. Intratumoral delivery as a method to improve immunotherapy efficacy

A wide variety of approaches to increase the response rate and efficacy of checkpoint inhibitors have been tried; one of the most promising is intratumoral injection of immunotherapy. [6–9] Under this framework (Figure 2), localized injection of agents into a ‘cold’ metastasis causes an inflammatory response that modulates the baseline immunosuppressive tumor microenvironment towards an immunologically permissive state. Example intratumoral therapies include immunoadjuvant agents that activate the innate immune system, such as toll-like receptor (TLR) or stimulator of interferon gene (STING) agonists. [7] These produce an inflammatory signaling cascade that recruits dendritic cells and other antigen presenting cells to the tumor. These cells in turn present tumor antigen to cytotoxic immune cells and activate them against the tumor. Cytotoxic immune cells then migrate throughout the body to attack the tumor both at the site of immunoadjuvant injection and at distant metastases.

Figure 2.

(Graphical abstract). Mechanism of image guided intra-tumoral immunotherapy and abscopal effect. From left to right, CT or ultrasound are clinically practical ways to inject deep lesions with an immunotherapy controlled release depot that contains a contrast agent for visualization, the resultant treated lesion leads to cytokine release, dendritic cell recruitment and antigen that is picked up by dendritic and macrophage tissues. This material is taken to a draining lymph node where (middle figure) an antigen presenting cell presents to a T – cell to be activated. Right figure, an activated T-cell then subsequently induces apoptosis in distant cancer such as in the lung. Created with BioRender.com.

1.3. Clinical trial evaluation of intratumoral immunotherapy delivery

As of September 2021, there are 113 clinical trials of intratumoral immunotherapy, of which 68 are phase I and 57 are phase II studies (Figure 3). These studies encompass a wide range of immunotherapy-resistant tumor types and include local delivery of immunoadjuvants such as TLR agonists, STING agonists, cytokines such as IL-2, oncolytic viruses (including engineered adenovirus, vaccinia, polio, dengue, HSV-1 among others), intratumoral injection of enriched immune cells (CAR T cell, Dendritic cell, or NK Cell therapy), often in combination with systemic or intratumoral ICI for a synergistic effect. [10] (Figure 1, left panel) Several of these trials have shown significant success. For example, the oncolytic virus talimogene laherparepvec (T-VEC), now approved for the treatment of melanoma, demonstrated a 16% durable response rate as a single agent and a 52% response rate in combination with ipilimumab compared to 23% for ipilimumab alone.[11] A phase 1 trial of intratumoral CpG oligonucleotide, a TLR9 agonist, in combination with anti-PD-1 immunotherapy demonstrated promising response rates up to 78% in 9 patients who were immunotherapy naïve. [12] Additional promising studies have included delivery of clostridium spores [13,14], and injection of dendritic cells pre-loaded with tumor cell lysate. [15,16]

Figure 3.

Clinical trials of intra-tumoral immunotherapies. a) Stages of ongoing clinical trials related to the injection-based delivery of intra-tumoral immunotherapy agents. Data from clinicaltrials.gov as of July 2021. b) Tumor types addressed in these clinical trials.

1.4. Challenges in intratumoral injection: controlling release and visualizing delivery

Effective intratumoral injections require precise delivery of the drug into the target lesion and may also depend on the distribution of the drug throughout tumor. The fundamental underlying assumption with intratumoral drug delivery is that by injecting the drug through a needle positioned within the tumor, the drug will be deposited within the tumor. However, we have found that this assumption is profoundly false.[17] Variations in injection technique and injection needle design, as well as tumor microenvironmental factors such as tumor interstitial pressure, can result in dramatic alterations in intratumoral drug distribution. Moreover, spillage of the drug into off-target tissues and blood vessels can result in significant systemic toxicities. In the subsequent sections, we detail specific challenges and potential solutions for intratumoral immunotherapy interventions.

1.4.1. Limitations of imaging in intratumoral injections

While superficial lesions can be injected by visual inspection, the majority of lesions are non-palpable and require image guidance for injection in the interventional radiology suite. The mostly widely utilized imaging modalities for intratumoral injection are CT and ultrasound, with ultrasound being favored for anything < 5 cm deep, including subcutaneous, thyroid, or breast lesions, and CT for deeper lesions and those in visceral organs. [18–23] The technique for a CT or ultrasound-guided therapeutic injection is similar to that of biopsy, which is routinely performed and accurate within a few mm of the target virtually anywhere in the body. Imaging is used to guide needle placement into tumor, but the distribution of injected intratumoral therapy within (or outside of) tumor is often unevaluable since the therapy itself cannot by visualized by the imaging modality.

Recent studies have demonstrated that image guidance is necessary to ensure accurate delivery and immune efficacy. [24,25] For example, an intratumoral drug study evaluating an iodinated compound showed CT evidence of off-target injection despite proper needle placement. [24] The majority of intratumoral drugs coming to trial cannot be visualized using standard clinical imaging techniques. This can at times impede the interpretation of clinical trial results, when leakage and diffusion away from target can contribute to non-efficacy of a drug. [10].

1.4.1. Inadequate drug distribution

Drug distribution is dependent on the method of administration (intravascular versus intratumoral, intratumoral injection technique, type of injection needle), the tumor environment (degree of interstitial pressure, vascular in-growth and tumor necrosis), and drug characteristics (charge, hydrophobicity, and size). These constraints have been addressed extensively to modulate chemotherapy release from controlled release formulations, and the same principles can be applied to optimizing intratumoral immunotherapy delivery. [26]

1.4.2. Frequent need for weekly injections

Intratumoral therapies approved or in trials often require multiple weekly injections. The requirement for multiple injections arises in part because small molecules rapidly diffuse away from the injection site on the order of minutes to hours, with variable penetration from either intravascular or intratumoral approaches.[27] To compensate and reach the therapeutic window of the agent in the desired environment, excessive levels of drug may be injected which can lead to undesired off-target side effects. Multiple weekly procedures increase the risk of bleeding and infection, and for deeper tumors requires conscious sedation with associated risks and inconvenience for the patient. They are also taxing on patients and their caregivers, as well as for the busy interventional radiology practices where most of these procedures are performed. These logistical challenges have the potential to severly limit therapeutic adoption. [28] As such, mechanisms to reduce the number of repeated injections are necessary. Extended-release technologies that have already been used for delivery of chemotherapy, such as microparticles, hydrogels or other polymeric drug carriers, may similarly enable single-administration immunotherapy treatments.[29]

2. Controlled release systems for intratumoral immunotherapy delivery

2.1. Current controlled release technology

To overcome these limitations of traditional immediate release injections, controlled release delivery mechanisms such as slow-releasing and retentive polymeric drug vehicles can help increase drug deposition and targeting as well as tracing capabilities. Figure 4 outlines different classes of injectable drug delivery formulations for chemo- and immunotherapy drugs, from micro- to nanoscale and increasing in functionality from left to right.

Figure 4.

Examples of commercially relevant injectable polymer-based drug release formulations for chemotherapy (lower half) and immunotherapy (upper half). First injectable formulations in the 1980’s were mostly based on degradable and non-degradable microparticle suspensions (1^st block from left). On the nanoscale, functional groups on the polymer backbone (such as polycarboxylic acid polymers) or at the end groups (typically PEG) can be used to make drug-polymer-conjugates that can reduce systemic toxicity and increase targe accumulation (2^nd block). Amphiphilic block-copolymers such as PEG-PLA’s will self-assemble around hydrophobic drugs, enabling improved target uptake through decoration of target-specific functionalities on the resulting corona surface (3^rd block). Hydrogels are another platform technology to facilitate encapsulation, injectability, local retention and extended release of chemo and immunotherapy agents, through either medium- to long-chain viscosity increasing polymers, such as PEG or hyaluronic acid, or stimuli-responsive gelling co-polymers such as PLGA-PEG-PLGA. Towards the paradigm of personalized medicine, implementing imaging modalities and thus the live trackability of an injected payload is a desired feature of the future for which not many examples exist yet. Adapted with permission from Kamaly et al. Copyright 2016 and modified. American Chemical Society [29]

Going back in history, the foundation for extended-release systems was laid by macroscopic formulations that were too large to be injected, but were typically implanted, swallowed or had other administration routes. These first-generation drug release formulations relied either on diffusion from a non-degradable polymer matrix, or more commonly on the biodegradation of the excipient to liberate the drug, for which commonly degradable polymer structures such as polyesters and carbohydrates are used. For example, polylactic-co-glycolic acid (PLGA), polyglycolic acid (PGA), polycaprolactone (PCL), polyanhydrides, hydroxypropyl methylcellulose (HPMC), and cellulose acetate have been used in resorbable sutures, long-acting implants and extended-release oral formulations, and continue to be used for these applications. [29] Drug release from degradable copolymer systems can be controlled by polymer backbone properties such as the monomer chemistry, ratio of hydrophobic/hydrophilic groups, molecular weight, and crystallinity, which all affect degradation rate and encapsulation efficiency of the drug. For example, PLGA degradation speed can be tuned from weeks to months by increasing the ratio of hydrophobic lactic acid versus the more hydrophilic glycolic acid, while polycaprolactone is more hydrophobic so it can last up to 2–3 years. [29–31]

The same or similar polymers were then later on used for microscale and injectable systems, such as PLGA- or Polyglutamic acid (PGA)-based microparticle suspensions, albeit almost exclusively for chemotherapy agents and hormones (Figure 4, left). Particulate formulations have been investigated for TLR7/8 agonists, where delivery through sub-micron particles (~300 nm) showed improved CD8 T-cell activation compared with other polymer configurations such as micelles (~10 nm), or random coils (~4 nm), but have seen little translation or clinical trials. [32,33]More recently, nano-scale linear and self-assembling polymer conjugates (Figure 4, 2^nd block from left) have been used for systemic chemotherapy delivery. Conjugates are polymer backbones with functional repeat units or functional end groups that can be linked (conjugated) to drugs bearing complementary functional groups. While conjugated, the drug is inactive and becomes a “prodrug”, until the bond is cleaved (e.g., by hydrolysis in physiological setting) and the parent drug is reactivated. Such covalent linkages bear certain advantages over the free drug encapsulation described in previous technologies, such as reduced burst release, solubilization of hydrophobic drugs, high drug loadings, prolonged circulation, and enhanced tumor retention due to the large macromolecular size (EPR effect). Another benefit is a reduction in free drug levels and concomitant systemic toxicity, which otherwise in cases such as neocarzinostatin have had limited clinical translation. [34] Several polymers with functional backbones are in clinical or pre-clinical development. Examples are Opaxio / Polyglumex, a Paclitaxel-bearing backbone polyglutamic acid formulation, and DEP Docetaxel, a Docetaxel-bearing dendrimer polymer.[35–37] In general, polymer backbones with functional groups on every repeat unit open the door to co-conjugation of multiple species, for example drug combinations, or addition of targeting and imaging agents.

The majority of marketed conjugates, however, are based on PEG, which has only two end-groups available for conjugation and is non-degradable at molecular weights larger than a few thousand Daltons.[38] Here, PEG is not used as a drug-carrying backbone, but rather for its “stealth” effect that reduces immune system recognition of the agent by steric repulsion of the encapsulating shell, while increasing blood stream circulation by its increased size that evades renal clearance systems. [37] PEGylated conjugates for immunotherapy in clinical development are Pegilodecakin (Interleukin 10, anti-inflammatory cytokine), which has shown promising clinical activity when delivered systemically in combination with anti-PD-1 inhibitors [39] and NKTR-262 (intra-tumoral TLR7/8 agonist delivered with Bempegaldesleukin, a CD122 preferential IL-2 pathway agonist), that has shown good safety data and robust TLR-7/8 engagement (Figure 4).[40]

Linear polymers can also be synthesized as copolymer blocks of differing hydrophilicity, often involving PEG as the hydrophilic part and degradable polyesters such as PLGA as the hydrophobic portion, thus creating amphiphilic structures that can readily self-assemble around drug cargoes. These polymers form encapsulating nanoparticles around the active agents, thereby increasing solubility of hydrophobic agents in aqueous formulations, limiting their systemic toxicity compared with free drugs and improving their target specificity by featuring target-binding moieties on the surface of the formed corona of the nanosphere. An example for such system to improve chemotherapy is BIND-014, which is a docetaxel-encapsulating nanoparticle decorated with functionalities that preferentially bind to prostate-cancer cell transmembrane proteins. An example for nanoparticle delivery of immunotherapy agents is SEL-068, a PLA-PEG formulation encapsulating nicotine antigen and TLR agonists as vaccine against smoking addiction (Figure 4, 3^rd column).[29]

Another popular class of polymeric systems that elicit macroscale effects based on nanoscale chemistries, are hydrogels. Hydrogels are soft materials with extremely high-water content whose crosslinked polymer backbone structure allows for structural integrity while also allowing for controllable degradation, stimuli-responsiveness, biocompatibility, and controlled drug release. [29,41] Hydrogels have become a platform technology as biocompatible drug delivery systems and, unlike earlier technologies creating covalent crosslinks through toxic reagents, typically form in situ or before administration through the formation of non-covalent linkages. The gelling may be a result of molecular self-assembly and molecular rearrangement of polymer molecules following physiological cues such as body temperature or pH, or may be achieved by the addition of charged species, as in alginate crosslinking with Ca²⁺ ions. Other polymer systems able to form hydrogels are certain PEG-copolymers, polyacrylamides, and polysaccharides such as alginates and carboxymethyl cellulose., such as thermosensitive An example for stimulus-induced thermogelation are low-molecular weight PLGA-PEG-PLGA copolymers, which are liquid at room temperature, allow for solubilization of hydrophobic drugs through micellar encapsulation, and form a gel at body temperatures through the increase of hydrophobic interactions (Figure 4, third panel from left). [42]

Hydrogels have been used extensively for intratumoral chemotherapy delivery where varying tumor microenvironments may affect the dwell time of the drug. [27] By producing a local depot, hydrogel injections allow for controlled sustained release of the underlying therapeutic agent to increase efficacy. [43] This technology allows significantly higher local concentrations of drug and has been further applied to intratumoral immunotherapy, for example intratumoral multi-domain peptide (MDP) hydrogel injected through a 29-gauge needle in a rat flank tumor model produced a fourfold increase in STING agonist levels locally compared to free drug where it demonstrated stimulation of the local immune environment. [44] The TransCon platform for intratumoral immunostimulatory injection uses a PEGylated hydrogel depot with a pH and temperature sensitive linker to release pro-drugs including TLR7/8 agonist resiquimod. [45–47]Hyaluronic acid (30 kDa) modified with tocopherol to facilitate nanospherical encapsulation of tocopherol-conjugated resiquimod prodrugs is another approach developed to provide long-acting intratumoral gel-like depot formation that has successfully induced tumor remission in dogs. [48] Indeed, similar to the results seen with chemotherapy release, the efficacy of intratumoral immunotherapy can be augmented by increasing the duration tumor exposure, such as by increasing the molecule size or increasing matrix binding. [49] As such, the structure of an ideal hydrogel depot would include, injectability through at least an 18 and ideally 22-gauge needle (reduces injury), controlled tuned release on the order of at least a week tuned to the specific immunoadjuvant in question, and the ability to visualize the hydrogel during delivery.

Besides the capacity of the various polymeric carriers to encapsulate and release immunotherapy drugs, some polymers and nanoparticles have been associated with immunostimulatory potential themselves. In several studies, researchers found an immune response, such as CD8(+) and CD4(+) T-cell activation and increased presence of antibodies, in their vehicle controls similar to the adjuvanted version. Specifically, drug-free formulations involving liposomal micelles, pluronic-stabilized polypropylene sulfide NPs, Eudragit polymers, and polyanhydrides were found to elicit immunostimulatory activity. [50] These effects should be considered when designing cancer vaccines and other immunotherapies, as they have the potential to amplify but also to interfere with the a priori expected effect of the drug of choice. In fact, drug delivery vehicle design often involves the choice of polymers that are considered non-immunogenic, such as hyaluronic acid polymers, or acrylamide-based polymers with certain side chain functionalization. [29,51]PEG has been widely used to increase bioavailability and actually reduce immunogenicity of drugs and biologics, but concerns around PEGylated formulations being immunogenic have increased, which is a topic under investigation given the vast usage of PEG in modern drug delivery systems. [52–54]

An emerging non-polymeric approach to enable long-acting release of immunostimulatory agents utilize RNA vectors to build extended release of cytokines such as self-replicating IL-12 RNA into the cell machinery itself. [55] Additional methods to bind the delivery of the immunoadjuvants to the extracellular environment may also be a solution for prolonging release as seen by groups that have used collagen or alum-anchored immunotherapies.[56,57] Also non-aqueous but rather lipophilic formulations of TLR7/8 agonists conjugated to lipid molecules are a strategy developed to increase retention and local delivery following intratumoral injection, e.g. in sesame oil with 7.5% ethanol.[58]

2.2. Pre-clinical studies of hydrogels for immunotherapy

As described, there are multiple methodologies to generate controlled release intratumoral immunotherapy modalities. Hydrogels, in particular, have a long history of use as controlled release systems given their biocompatibility, injectability and a structural variety that allows for tuning of critical drug loading and releasing properties. They are amenable to being combined for both imaging, delivery and therapy with lessons that can expand to many of the other modalities. For the purpose of this review, we will focus on the hydrogel example as a platform for immunotherapy delivery and imaging. Hydrogel use for immunotherapy delivery has been recently reviewed intensively given excitement in the field, such as Huang et. al.’s review on nano, micro- and macroscale drug delivery. [59] An adaptation of their review of different gel types can be seen in Table 2. Overall, gels have used variations of intratumoral [60–63], post-surgical (tumor resection beds) [64,65], intravenous, [66–69] and subcutaneous [70–78] approaches. Subcutaneous approaches have generally attempted vaccination of tumor antigen, with multiple groups using ovalbumin as a prototype molecule. [74–78] Surgical resection allows for harvesting of unique tumor antigen which can then be used to prime the immune response. One group delivered tumor lysate, TLR3 agonist, and gemcitabine in an implanted 5 mm disc that released 40% of the drug by 6 days [64], which increased local immunity in a breast cancer model. Another group injected CpG ODN and anti-PD-1 antibody in a nano-cocoon formed from the CpG-ODN itself with direct injection in the tumor bed. [79] These post-surgical approaches are interesting as adjuvant approaches.

Table 2.

Studies investigating controlled release of Immunotherapy and Immunoadjuvants. Adapted with copyright permission of tables originally presented by Huang, et. al. Nano, micro-, and macroscale drug delivery systems for cancer immunotherapy. Copyright Elsevier. [59]

Therapy Route	Scale	Immunotherapy	Formulation	Duration of Release	Animal Outcomes	Imaging	Ref
Surgical Implantation into the tumor resection bed	Macroscale	Tumor lysate + TLR3 agonists, gemcitabine	Cross-linked collagen and hyaluronic acid scaffold - implanted as a 5 mm disc	40% plateau release in 6 days of the nano-gel, and release of 4T1 lysate over 30 days,	increase DC, enhance T- cells, IFN-y, resolution of lung metastasis compared to control surgery, vaccine alone, or gemcitabine	Not applicable	[64]
Direct injection in the tumor bed	Nanoscale	CpG ODN, Anti-PD-1 antibody	CpG Nano-Cocoon	Activation of macrophages by 6 hours.	Increased survival time of mice in post- surgery, with direct injection in the tumor bed	None	[79]
Intratumoral injection	Microscale	anti-CTLA4	Ethiodized oil and poly-lactic-co-glycolic acid nanoparticles (PEEP)	30-day release to plateau	Improvement in survival in CT-26 tumors, similar with direct anti-CTLA4 and with PEEP.	Radio-opaque, ethiodiozed oil	[60]
Intratumoral injection	Macroscale	Anti-PD-1 antibody, IDO Inhibitor	P(Me-D-1MT)-PEG-P(Me-D-1MT) Hydrogel	25% release at 4 days prerelease, and near 100% release with ROS by 24 hours. aPD-L1 seen up to 7 days in vivo as opposed to free aPD-L1(3–5 days)	Decreased systemic toxicity and enhanced anti-tumor t-cell immature response, reduced tumor growth in vivo	Fluorescence	[61]
Intratumoral injection	Nano-macroscale	STING agonist + Campthothecin	Peptide drug moiety iRGD+camptothecin - supramolecular nanotubes + STING agonist (c-di-AMP)	Retention drug release out to 15 days in vivo	Reduced tumor growth and regression in GL-261 brain tumors.	Fluorescence - FITC	[62]
intratumoral injection	Macroscale	OX40 and CD40	Nanofluidic membrane with multiple channels from patterned tungsten layer created via physical vapor deposition. Lyophilized antibody solubilizes in the fluid and pushes out by diffusion	30-day release	Inhibition of tumor growth (4T1 orthotopic murine tumor)	None	[63]
IV injection	Nanoscale	IDO inhibitor, DPPA-1	Self-assembling D-peptide programmed cell death ligand 1	80% release in 4 hours in tumor mimicking environment (pH 6.8 + MMP), 20% release > 48 hours in pH 7.4 without MMP-2	sensitive to MMP-2, slowed melanoma growth and inc'd CD8+ T cells	None	[66]
IV injection	Nanoscale	TGF-B inhibitor, IL-2	Nanoscale liposomal polymeric gel	7 days	IV injection- Increased efficacy compared to solubilized drug alone. Treatment every 7 days for 40 days.	Fluorescence - rhodamine	[67]
IV injection	Micro/Nanoscale	CpG oligonucleotide with TRP2 peptide	Porous silicon microparticles (500–1000 nm)	Majority release in 20 hours and remainder by 100 hours, cleared in vivo by 72 hours, primarily in liver and lungs	Co-delivery of two different TLR Agonists (CPG oligo nucleotide and MPLA increased CD8 response against B 16 melanoma). Mildly increased survival in B16 melanoma (1/8) to 60 days	Fluorescence	[148]
IV injection	Nanoscale	Imiquimod (TLR - 7 agonist)	PLGA-ICG-Imiquimod	N/A (Biological effect within 7 days)	Photoacoustic + immunotherapy effect inhibitors growth	Fluorescence	[69]
Intradermal injection	Macroscale	OVA, CpG	CpG DNA hydrogel	20–50% release of OVA by 24 hours in vitro, and near 100% release by 24 hours, as compared to less than 6 in vivo	Cationized antigen reduced tumor growth in vaccinated subgroups compared to control without the gel on tumor challenge	Fluorescence	[70]
Subcutaneous injection in the hind footpads of mice	Nanoscale	OVA, TLR-7/8a	N-(2-hydroxypropyl) methacrylamide (HPMA) and N-isopropylacrylamide (NIPAM) covalently conjugated to TLR-7/8a	Persisted up to 20 days	Increase cytokine production in lymph node with particles, compared to free drug	Fluorescence	[71]
Subcutaneous implantation	Microscale	Anti-CD137, anti-OX40 mAb in a hybridoma	Alginate-Ca microparticles	Majority in 1-week, complete loss by 14 days	Alginate-Ca microparticles encapsulated with hybridomas result in increased in cytotoxic T-cell lymphocyte response with 4–6/7 complete eradication of CT-26 models	N/A	[72]
Subcutaneous injection	Macroscale	Tumor vaccine CpG - ODN, GM-CSF	Mesoporous silica rod scaffold	66% release of GM-CSF by day 40, OVA 45% release in 5 days, sig increased from free OVA	Scaffold induces dendritic cell recruitment and increases cytotoxic t- cells, and allows adjuvant release	Fluorescence	[73]
Subcutaneous injection	Microscale	PLGA particle containing CpG-ODN and Ovalbumin	PLGA microparticles	Highest peak IgG titer at day 56, and IFN-y persisting until day 42, PLGA fluorescence down to near zero by 20 days, majority by day 9	Increased cytotoxic T cell response and interfered with tumor growth.	Fluorescence	[74]
Subcutaneous injection, tail base	Nanoscale	CpG Neo-antigen	Lipoprotein-mimicking nanodisc	Peak 28 hours, and reduced by 48 hours (compared to < 10 hours control) - cell culture	47-fold greater neo-antigen-specific CTLs than soluble vaccines, eliminating MC-38 and B16F10 tumors, 30 % Glioma	Fluorescence	[75]
Subcutaneous injection, tail base	Microscale	OVA or TCL	Yeast-derived microparticles	N/A	Increased DC triggered antigen delivery. detected OVA antibody post injection, followed by subsequent tumor (lymphoma) challenge showing improvement with YS-OVA vs either alone or with alum	Fluorescence-FITC	[76]
Subcutaneous injection, tail base	Macroscale	Tumor cell lysates, TLR 3 agonist, poly(I:C)	PEG-poly(L-Valine) hydrogel	Cumulative release to 80 % by 120 hours. Burst release to 40% in 1 day	Modulate antigen specific immunity	Fluorescence	[77]
Subcutaneous, intrascapular region	Macroscale	E7 peptide, B16F10 or CT-26 neoantigens	Mesoporous silica rod scaffold-Polyethyleneimine	Not studied.	Neoantigens- eliminated established lung metastasis in TC-1 B16F10 and CT -26 mice synergized with anti-CTLA4 therapy	Not applicable	[78]

However, patients with metastatic disease are unlikely to have tumor resected. Thus, several groups have evaluated controlled release of intratumorally injected immunoadjuvants (Table 2).[28] For example, one group combined anti-PD-1 antibody with Indoleamine 2,3-dioxygenase (IDO) inhibitor chemotherapy, with a release over 24 hours, which allowed for chemotherapy-induced cell death to release tumor antigen in a pro-inflammatory environment. [61] Other groups have injected immunoadjuvants, including STING agonists or OX-40, with controlled release over 15 to 30 days using supramolecular nanotubes and a nano-fluidic membrane. [62,63] Both of these approaches showed successful synergistic effect with distant metastasis reduction in brain tumors and breast cancer models.

In general, intratumorally injected immunoadjuvants were microscaled demonstrating the capability to produce a depot effect, and likely best able to be adapted for deeper percutaneous image guided approaches. In contrast, therapies dependent on intravenous systemic injection system were typically nanoscaled, taking advantage of enhanced permeation and retention or ligand targeting, and surgically implanted systems were macroscale systems requiring an incision or built for placement in the tumor bed. The wide range of tumor delivery times under controlled release with varying efficacy suggests that the ideal time period of delivery is not yet known. As discussed earlier, the minimum clinically practical time period is at least one week before requiring repeat injection, and longer time intervals are far preferable for patients and practitioners alike.

2.3. Challenges in intratumoral hydrogel injection: visualizing delivery

Few hydrogels delivering immunotherapy use clinically applicable imaging, with a notable exception being one group that created a Pickering emulsion of PLGA nanoparticles, anti-CTLA-4 antibodies, and ethiodiozed oil to create a radio-opaque injectable emulsion with the average droplet measuring 42 +/−5 μm. [60] Instead, fluorescence models have been used in multiple animal models to assess persistence of the drug for up to 20 days. While fluorescence is commonly used in animal models because their small size means tumors can be visualized through the skin (up to 1 cm deep), fluorophore labelled therapeutics are generally not translatable to humans where most tumors are typically far deeper.

Confirmation of intratumoral therapy delivery is key to evaluating the clinical success of clinical compounds. As a result, most clinical trials and applications have focused on superficial lesions amenable to direct optical visualization such as melanoma skin metastasis, peritoneal implants by laparoscopic surgery, or polyps/mass accessible by endoscopy, colonoscopy, or cystoscopy, or direct exam in the oral cavity, or cervix for example.

Deeper lesions not easily accessible by natural orifices has been more difficult. Based on CT imaging of the few agents that had iodinated radio-opaque compounds, several demonstrated off-target spread, with the suggestion that the apparent efficacy of many investigational agents be compromised due to unknown leakage (Figure 5). Confirming injection led to improved distribution and intravasation of drug, dendritic cell uptake, and immune response in these trials, suggesting accurate drug delivery has important systemic implications. [44] Given the powerful local and systemic consequences of intratumoral delivery of immunotherapy agents, supporting their delivery with imaging agents is an exciting step necessary to bring these drugs to clinic. Multiple groups have begun to explore combining clinically translatable contrast agents with controlled drug delivery vehicles.

Figure 5.

Original figure a-f adapted from Sheth, et. al. JAMA Open, Creative Commons, used with permission. [17] a) Demonstration of injection distribution in a patient throughout the tumor. b-d) Demonstration of injection leaking out of the tumor. e) Injection using a multi-prong injection needle f) contrast demonstrating better injection distribution from the multi-prong injection. g) Side by side comparison of a multi-side hole versus end hole needle injection in a kidney animal model. h) Injection side by side shows more uniform distribution with the multi-side hole. i) Better retention is seen post injection with the multi-side hole. Images in g,h,i reproduced with permission from the work by Munoz et. al. 2021, Creative Commons License. [44]

3. Imaging modalities available for local drug delivery

3.1. Clinical Imaging Tools in Clinical Practice: cross-sectional tomography (CT), ultrasound (US), and x-ray fluoroscopy

Given the importance of being able to “see” where intratumoral delivery occurs clinically, future drug design should occur with an aim to allow visualization. As of 2021, clinically the most widespread imaging modalities in clinical use remain ultrasound[80] and CT[81] both in the United States and globally. Fluoroscopy is still widely used for angiographic and spine interventions and widely available by multiple specialties including interventional radiology, orthopedic and neurosurgery. [82–84] These modalities are relatively quick to use and available for clinicians. As such any targeted drug should ideally design for these modalities for wide-spread utilization.

Preclinical research often attaches fluorophores, but fluorescent imaging is significantly depth dependent and despite much preclinical utility over the years, has only limited clinical applications.[85,86] The vast majority of image-guided clinical intratumoral procedures use ultrasound and CT guidance, with rare specialized imaging using MRI or PET. (Figure 6) While these technologies can be difficult to scale down to the size of mice, and often more expensive than existing fluorescence options, microCT[87], small animal MRI, PET [88], and ultrasound[89] do exist and are available to the pre-clinical researcher. As the inclusion of contrast agents for each of these modalities can change pharmacokinetics of intratumorally delivered agents, factoring them into the delivery agent design is important for guiding drug delivery and successful trial outcome. For example, drug release itself is controlled by modulating hydrophobicity and hydrophilicity of the agents, porosity of hydrogels, and distribution and surface area (more in nanoparticles versus microparticles), [90] hence contrast agents may interfere with the carrier’s capability to retain or release drug. Contrast agents can vary in hydrophilicity, bulkiness, and solubility, exhibiting variable methods of clearance and safety in different compartments. Understanding these interactions, the required duration of imaging contrast, and efficacy, is integral during design itself. Below is a summary of how each tool is used clinically framed for the pre-clinical engineer.

Figure 6:

Practical clinical imaging modalities and example clinical image of a procedure. a) CT apparatus and image of a microwave probe in a hepatocellular lesion. b) Ultrasound device, and ultrasound image of an ultrasound guided biopsy. Contrast enhanced ultrasound with biopsy of a proven melanoma metastasis. Bottom two images reproduced with permission from Kopf et. al. Copyright European Journal of Ultrasound. [161] c) Fluoroscopy suite with angiographic image demonstrating selective catheterization of the celiac artery during a trans arterial chemoembolization.

CT scans use ionizing radiation, X-rays, to generate a tomographic image based on tissue-induced signal attenuation. The x-ray beam is absorbed (attenuated) more by structures with higher atomic numbers/electron density and increased mass density (for example calcium in bone, or iodine in iodinated contrast/ thyroid), and less so by water, or air. [91] Densities under CT scan are characterized by Hounsfield units (HU, water = 0 HU, air = −1000 HU, bone = ~1000 HU), which are often used to identify cancer locations showing hypo- or hyperdense signals compared with surrounding tissue. [92]

CT guided interventions are performed using non-contrast exams with the patient in a recumbent position. Patients are typically under moderate sedation, although procedures may be done either with local anesthetic only and/or under anesthesia care. An initial pre-intervention axial CT image is acquired, and intravenous iodinated contrast can be given, if necessary to identify critical structures. A needle entry location is identified, the location sterilized, and local anesthetic applied. A needle is inserted and advanced under CT-imaging guidance until it is in the reached the target. (Figure 6a). Size of needle depends on the intervention to be performed, with fine needle aspirations performed through 18–22-gauge needles, core biopsies using 17–19 gauge biopsy systems, and thermal ablations using 14–16 gauge probes. Multi-side hole and multi-prong needles have also been developed but are not in routine clinical use in 2022, though they have been previously used in ethanol injection and demonstrate greater diffusion of drug delivery (Figure 6). [44]

Ultrasound provides real time and high-quality soft tissue contrast (Figure 6b). Ultrasound-guided procedures tend to be quicker than CT-guided procedures given real-time imaging feedback. Additionally, ultrasound equipment is less expensive to purchase and maintain, and does not require a dedicated technologist to operate. It uses differences in sonographic reflections to distinguish tissues. For example, water is generally hypoechoic, whereas fat is hyperechoic. Ultrasound cannot visualize through air or bone. Intravascular contrast using microbubbles can be used to allow contrast enhancement temporarily as well over the same time duration as seen using CT visible iodinated contrast (Figure 6d). [93]

For ultrasound guided interventions, the patient is oriented to allow for the shortest distance to the target. After a similar sterile preparation and delivery of local anesthetic, the needle is followed to the target under active real-time ultrasound guidance.

Many lesions within the kidney, liver, peritoneum or pleura may be amenable to both ultrasound and CT depending on the exact imaging characteristics of the lesions. [94–96] Of note, contrast agents delivered to the site, generally should clear prior to subsequent staging exams as hyperdense objects such as metal or barium have significant streak artifact that limits response assessment. [97,98]

In the context of cancer therapy, fluoroscopy is most commonly used for intra-arterial therapies, most commonly liver-directed delivery of radiation (Yttrium 90) or chemotherapy-saturated beads. [99,100] Figure 6c. Fluoroscopy heavily uses contrast injections to allow real time assessments of catheter locations and interventions through the vasculature, which can be used to assess therapeutic delivery. In these methods, combinations with ethiodol within microbeads such as the Lumi© beads from Boston Scientific have been used in the past to confirm target delivery, or in mixtures with contrast during injection at the bedside table. [101,102]

3.2. Uncommon clinical interventional techniques: MRI, PET, and Optical Imaging

Magnetic resonance imaging-guided procedures are time-intensive and thus not considered a cost-effective use of an MRI scanner by most radiology practices. MRI interventions are used in select instances where CT or US is deemed insufficient for adequate guidance and predominantly by research institutions. Where used MRI has been applied for both drug delivery but also in the use of MR thermometry to evaluate ablation. [103,104] MR-guided focal ultrasound for prostate ablation is an example that has begun to be used for prostate ablations and may provide future promise as the field grows (Figure 7 a). [105] Similarly PET imaging, has been used more as a planning study with fusion as opposed to live imaging. [106,107] Access to an FDG-PET guided intervention is not yet widespread, though has been reported in case series to increase biopsy yield (Figure 7b). [106–108] Labelling intratumoral therapeutics with radioactivity for purposes of imaging adds significant logistical issues of production, limiting operator exposure, and overall radiation safety that significantly limits adoption. Near-infrared fluorescent imaging has not had wide-spread adoption due to limited depth penetration. Applications are primarily surgical, for example for laparoscopic surgery or in the optical fields of cystoscopy or endoscopy, often currently looking at vascularization of the lesions, the common bile duct, or lymphatics. [85,86] These applications remain limited significantly by depth of visualization and can be combined with use of ultrasound (such as endoscopic ultrasound or laparoscopic ultrasound), or with fluoroscopy as is done with fiducial guided lung resections. Applications in fluorescence imaging to aid percutaneous intervention radiology techniques remain experimental (Figure 7c). [109]

Figure 7.

Uncommon clinical imaging modalities. a) MRI machine, and MR guided focal ultrasound ablation showing MR thermometry and area of ablation. Adapted with permission from Napoli et. al. Copyright European Urology[105] b) PET- CT, PET guided biopsy of a liver lesion, Adapted from Govindarajan, et. al. Creative Commons License [108]. c) Use of near infrared imaging to confirm biopsy specimen subpart A shows lesion on MRI, subpart B shows on ultrasound, subpart C shows FDG avidity, sub-part D shows no fluorescence on normal liver, sub-part E shows fluorescence during active guidance in sub-part E, and fluorescence on external imaging, sub-part F, and corollary visuals in sub-part G. Images reproduced with permission from Sheth et. al. Radiology 2014 Copyright Radiology.[109]

3.3. Materials for enabling contrast on imaging.

3.3.1. Non-permanent contrast for CT imaging

Iodinated contrast is the mainstay for increased contrast visualization particularly to look at enhancement of lesions. Concentrations for radio-opacity are typically in the 100 mgI/mL range and can be mixed in at 10–50% v/v intravenously. The structure of the iodinated components is shown in Figure 8a. [110] For combination into polymers, most iodinated compounds have a hydroxyl group that can be used for chemical conjugation or be combined using hydrophobic/hydrophilic characteristics for steric combination into polymer conjugates. For intratumoral injection, non-permanent contrast is recommended, as permanent contrast agents can complicate future response assessment. The retention of iodinated contrast depends on both the iodinated contrast used and how it is delivered. For example, groups have shown an end-hole injection into a flank model has dispersion of iohexol within 5 minutes while a multi-side hole injection offers greater retention. [44] In a study of hysterosalpingograms, injection of lipiodol (oil-soluble contrast) had a half-life of 50 days whereas iotrolan, a water-soluble agent, was fully excreted in 2 days.[111] For confirmation of target delivery during the procedure, retention within tumor of approximately an hour is necessary. However, for purposes of understanding drug-release kinetics, it is much better for the imaging contrast release profile to match the expected release rate drug.

Figure 8:

a) Various iodinated contrast agents and their structure, reproduced from Solomon, et. al. Creative Commons License [110] b) lipid microbubbles, reprinted with permission from Elsevier, adapted from Omata, et. al. 2020 [112]

3.3.3. Ultrasound agents for imaging

Ultrasound examination have used microbubbles to increase echogenicity, thereby providing increased contrast within injected structures upon administration. Lipid encasing microbubbles have been used to add functionality and even enable drug delivery. [112] (Figure 8b) Applications can often be multimodality and be both radio-opaque and echogenic as the modalities tend to be orthogonal to each other. Interval echogenicity assessment allows follow-up of gel degradation. [112,113] Given the convenience of ultrasound globally, building in this capability for polymeric drug delivery systems dramatically increases the reach for appropriate injection and therapy broadly around the world.

3.4. Hydrogels containing imaging contrast

Multiple groups have taken advantage of the above knowledge to create contrast-containing hydrogels that can deliver therapeutic compounds under image guidance. Several reviews have delved deeply into possible techniques. [114,115] A review of various polymeric formulations for imaging via an investigational MRI technique, chemical exchange saturation transfer (CEST-MRI), fluorescence, and CT has been adapted here with a particular focus on drug release and method of utilizing imaging contrast as examples of what could possibly be done with immunotherapy. (Table 3)

Table 3.

Summary of imaging modalities used in a variety of hydrogel delivery agents. Adapted with copyright permission from Dong, et. al. “Detecting and monitoring hydrogels with medical imaging”. Copyright 2021 American Chemical Society [114]

Imaging modality	Imaging functionalization	Polymer	Drug Delivery	Drug Release Duration	Contrast Duration	Ref
CEST-MRI	Intrinsic	Alginate/Liposome	--	--	Duration of polymer	[116]
CEST-MRI	Intrinsic	Hyaluronic acid-SH/gelatin-SH (Can then be labeled with NIR days as well).	--	--	Duration of polymer	[117]
CEST-MRI	Intrinsic	Peptide-conjugate (glutamic acid + phenylalanine	--	--	Duration of polymer	[118]
CT	Gold Nanoparticles	Gelatin-Tyramine	N-acetyl cysteine			[127]
CT	Gold Nanoparticles	Alginate	Cisplatin	minimum 7 days (unsure)	--	[149]
CT	Gold Nanoparticles	PNAGA-PAAm copolymer N-acryloyl glycinamide	Doxorubicin	24 hrs	minimum 48 hr	[128]
CT	iopamidol and tantalum-oxide particles	gelatin-tetrazine/gelatin -norbornene	--	--	minimum 12 days (Ta-gel)	[150]
CT	Platinum nanoparticles	alginate	--	--	minimum 5 days (experiment did not continue beyond this; 80% Pt loss in DOHEPts at this time point)	[65]
CT	Gold Nanoparticles	Selenocystamine - poly(bis(4-carboxyphenoxy) phosphazene)) (Se PCPP)	doxorubicin, quisinostat, etc.)	70% release in 7 days with XR		[151]
CT	Iodinated trimethylene carbonate	PLGA-PEG-PLGA with co-polymer of the iodinated trimethylene carbonate	--	--	24 days for 100% released	[129]
CT	Ethidiozed Oil	Ethiodized oil and poly-lactic-co-glycolic acid nanoparticles (PEEP)	anti-CTLA-4	30-day release to plateau	Not tested	[60]
CT	Barium sulfate suspensions in hydrogel microcoils	Polyacrylonitrile	--	--	minimum 12 weeks (experiment did not continue beyond this)	[152]
CT	2,3,4-triiodobenzoic acid	Poly (ethylene glycol) and aliphatic polyester coupling the hydroxyl end of the diblock copolymer with 2,3,5-triiodobenzoic acid (TIB)	--	--	5–7 days (little left at in vivo in mouse abdomen; no gel visible at 7 days)	[153]
CT & Optical	Ta2O5 India Ink	Gelatin functionalized with tetrazine/norbornene	--	--	Stability for up to 12 days in vivo	[154]
Fluorescence	Intrinsic	Polyacrylamide	--	--	Duration of polymer	[119]
Fluorescence	Intrinsic	BSA/HAS	--	--	Duration of polymer	[120]
Fluorescence	Carbon Nanodots	N-methacryloyl chitosan	--	--	minimum 500 hr (experiment did not continue beyond this)	[122]
Fluorescence	Indocyanine Green	RADA16-I peptide	Melittin	2 days (at least; experiment did not continue beyond this)	2 days (ICG contrast had slower release than melittin, reaching 20% release rate at 2 days, while melittin reached 40%; at least; experiment did not continue beyond this)	[155]
Fluorescence	Quantum dots	Fmoc-diphenylalanine	--	--	minimum 5 hr (experiment did not continue beyond this)	[125]
Fluorescence/ photoacoustic	Ag2S quantum Dots	polypeptide PC10	Paclitaxel	6 days	35 days	[121]
fluorescence/MRI/ultrasound	rhodamine B and CoFe2O4	PLGA-PEG-PLGA	Luciferin (as a model for small molecule delivery)	7 days; liposomes remain stable in vivo for 3 weeks and release drug upon heating triggered by laser light (42 C)	20 mins (can be reactivated upon reheating)	[126]
Fluoroscopy	iodinated contrast agents	chitosan	--	--	at 4 h, decreased 80% [Conray], 56% [Visipaque] and 68% [Isovue]	[130]
Fluoroscopy	iohexol	PEGDA/ Alginate	--	--	--	[131]
MRI	M-Ferrite nanoparticles	AM PEG550	Paclitaxel	28 days (at least; at which 71.16 ±13.34% of initial amount put in gel was released)	28 days (93.68 ± 1.06% of initial amount put in gel released)	[156]
MRI	DOTA-GD(III)	Polyethylene - Glycol	--	--	minimum 2 wks (only 23% of Upy-Gd was released at this point)	[157]
MRI	ferumoxytol	Chitosan	Doxorubicin	16 hr	40 hr	[158]
SPECT/CT	Ba2+In3+Zr4+	Alginate	--	--	8 days in nose (function of mucus clearance; different duration when hydrogel is injected into different body parts)	[159]
SPECT/CT	99mTc	Carboxymethyl-cellulose	Pheonix WinNonlin (1-compartmental model)	--	--	[160]
Ultrasound	SonoVue microbubbles	silk fibrin	--	--	minimum 7 days (experiment did not continue beyond this)	[132]

Hydrogels can be built where the polymer itself provides contrast. Examples include alginate and hyaluronic acid in the setting of CEST-MRI. [116–118] However, CEST-MRI is not in current clinical use outside of focused MR spectroscopy. The lack of widescale use reduces the utility of this technique for drug delivery and monitoring, particularly on a visualization scale. Similarly, fluorescence-based methods [119–126] are difficult to be applied for anything but surgical applications with direct camera visualization.

CT, as a fast and widespread modality, provides the highest translatability. Among examples of already developed CT-imageable hydrogels, groups have used physical loading of gold nanoparticle into hydrogels made of gelatin-tyramine or alginate [127], or a PNAGA-PAAm copolymer [128]. These conjugates were evaluated for the intratumoral delivery of chemotherapies and would be interesting to explore for immunoadjuvants as well, with the caveat that gold would be retained as a permanent contrast agent, leading to potential treatment assessment problems from streak artifact.

Approved iodinated contrast agents are designed to be highly water soluble and thus difficult to load directly into hydrogels without washing out. Groups have tackled this by incorporating iodine into the hydrogel backbone [129]. Microbeads with iodinated contrast have been used as intravascular contrast agents during chemoembolization delivery [130,131]. Duration of contrast ranges from a few hours to 12 days for degradable systems, roughly enough time to determine localization, but unlikely to cause issues with follow-up imaging, where contrast retention may complicate response assessment. Ultrasound microbubbles take advantage of intrinsic contrast from air in the bubble which makes the gel hyperechogenic and have been used to visualize degradation [132].

Imageable hydrogels developed to date have generally been used to deliver hydrophobic chemotherapies, such as doxorubicin or paclitaxel, however the underlying technologies are directly applicable for checkpoint inhibitors and immunotherapy adjuvants. For example, the combination Pickering emulsion of ethiodiozed oil and PLGA nanoparticles delivering anti-CTLA-4 is a clear example of an intratumoral delivery agent combining both mechanisms [60]. In particular, the CT technologies discussed previously, including adding iodinated compounds to polymer backbone or high Z particles like gold into the depots show promise and likely applicability to immunoadjuvant delivery as already described that had otherwise used primarily fluorescence to show delivery in pre-clinical animal models. Of note, few of the delivery mechanisms have optimized for aligning visualization duration with drug release curve and as such primarily confirm delivery. This is potentially useful for avoiding long term contrast related streak artifacts. However, being able to see the drug’s duration and pharmacokinetics may be useful for therapy planning. In an assessment of currently running clinical trials of intratumoral immunotherapy agent, only one compound, Rose Bengal Disodium, which has iodine in its chemical composition, was able to be tracked. Future studies creating image guided controlled release technologies would be ideal for confirming both delivery and visualizing real time pharmacokinetics of therapy. [17]

Summary

Immunotherapy, the technology that has brought the hope of a cure for a subset of patients with metastatic disease, continues to spur significant interest in developing methods for overcoming resistance. Intratumoral delivery of immunoadjuvants is an area of active and fertile clinical investigation, and clinical experience to date has taught us that this nascent fields’ success will increase with both controlled drug release profiles that reduce the number of procedures necessary as well as the ability to visually confirm target-delivery and drug-release kinetics. For the practitioner and engineer, intratumoral delivery requires development of delivery platforms that can be seen easily in near real-time, particularly under CT, ultrasound, or X-ray fluoroscopy. Biocompatible polymers provide a toolbox to build suitable delivery vehicles for intratumoral imunoadjuvants, ranging from particle suspensions and polymer-drug-conjugates to amphiphilic block-polymers that self-assemble into nanoparticles. These polymers can allow the solubilization of hydrophobic compounds, encapsulation of imaging agents, increased drug load, specific tissue-targeting capabilities and the control of release patterns over time. Some of these technologies, such as nano-encapsulating hydrogels, can provide platform technologies, however, given the complexity of immunotherapy agent chemistries, their optimal range of efficacy, and the targeted organs, an individual formulation for each scenario is still required. Especially in conjunction with the incorporation of imaging agents, a better understanding of the change in delivery performance as a function of formulating certain drugs and imaging agents into polymeric systems is required to speed up formulation efforts. In this review, we have underlined some of the existing technologies that have been utilized for image-guided delivery as well as controlled release of immunotherapy.

Highlights.

• Intratumoral immunotherapy is a rapidly growing clinical field bridging clinical image guided delivery by interventional radiology and controlled release drug delivery.

• Real-time imaging feedback is necessary to ensure proper intratumoral delivery.

• A variety of imaging modalities are under scientific exploration, yet only a few have the potential for near-future application in clinical settings, particularly CT and ultrasound.

• Eventual clearance of the imaging contrast agent is ideal to minimize impact on interpretation of subsequent staging studies.

• Controlled-release formulations will improve clinical adoption by minimizing the need for multiple procedures and may enhance efficacy of intratumoral immunoadjuvants.

Abbreviation

CT

computed tomography

MRI

magnetic resonance imaging