Abstract
The deliberate occlusion of blood flow through transarterial embolization is currently being used to treat conditions ranging from hemorrhages to hypervascular tumors. Degradable, imageable high borate glass microspheres (BRS2) were developed and tested to improve lesion targeting and promote a temporary vascular occlusion which is sufficient for most embolization procedure. A 48 hour pilot study, in a swine renal model, was conducted to assess the embolization effectiveness and potential risks of this new embolic agent. Bilateral embolization of the caudal branch of the renal arteries using test and control particles were performed in 4 pigs. Embolization efficacy, recanalization and resulting ischemia were evaluated at different time frame (0, 24 and 48 hours). The primary outcomes for this study were the assessment of: (i) embolization effectiveness, and (ii) vessel recanalization. The test article was found to occlude vessels as effectively as the control microspheres, with the use of a smaller volume of microspheres. At the 24 hour time point, over 95% of the material was found to be completely degraded, although little to no recanalization was observed. This data suggests that BRS2 is an effective embolic agent, however further investigations into the method of delivery are required prior to clinical implementation.
Introduction
There is increasing emphasis on the development and use of degradable microspheres for transarterial embolization (TAE) indications. 1 Motivated by the improved patient experience which can arise as a result of transient ischemia of a target tissue(s), 2 it is contended that with suitable engineering, degradable microspheres may one day deliver the idealized performance requirements, 3 while minimizing the potential for long-term sequelae caused by permanent alterations in vascular capacity, histological architecture, and/or injury to both ‘on target’ and ‘off target’ locations. 4 Secondarily, the advantages of degradable embolic microspheres are thought to also include, but are not limited to, a reduction in duration of exposure to foreign materials for patients, and the ability to a repeat treatment(s) where necessary.2, 3 However, in order to attain safe and effective transient embolotherapy, appropriate user needs and design inputs must be balanced. 5 This is especially important when considering a degradable embolic; as the degradation timeline, and level of occlusion achievable, are critical parameters which much be considered during the clinical process.
Currently, there are a small number of transient embolization materials options available to the physician, each exhibiting functional limitations with respect to desired clinical attributes. 3 For example, the embolization effectiveness of gelatin sponge relies heavily on preparation technique (i.e. cutting and pumping slurries between syringes), which creates a broad range of particle sizes and unpredictable degradation timeframes 1 ; gelatin sponge particles (non-spherical) reported in the clinical literature exhibit particle size distributions from <500 µm to >2000 µm and degrade between 7 days and 3 months.6–8 Conversely, starch and PLGA microspheres which are available in calibrated particle sizes exhibit degradation timeframes which are not optimized for several transient embolization indications. For example, starch microspheres degrade < 40 minutes, which may contraindicate their use in UAE and many other TAE procedures,9, 10 while PLGA microspheres degrade over several months, and may provoke long-lasting inflammatory and fibrotic reactions. 11
In addition to the increasing emphasis on degradable materials for TAE, there has been a parallel effort to standardize, optimize, and personalize such procedures. 12 To deliver on these requirements, the literature has focused on the development of multi-modal imageable microsphere technologies, which would permit temporal and spatial assessments of microsphere flow and distribution respectively.4,12, 13 The use of intrinsically imageable microsphere technologies has delivered on this promise under preclinical conditions, where specific microspheres have been verified and validated to offer physicians an assessment of true spatial distribution of microspheres in a target tissue, whilst also providing real-time intraprocedural monitoring. However, to date, no commercially available product allows for controlled degradation within a microsphere engineered for multimodal imageability. 3
Recent developments in bioglass research demonstrate that silicate glasses exhibit much of the desirable characteristics of an imageable and permanent microspheres for TAE, 14 while borate glass microspheres may be ideal candidates for use as transient embolic microspheres. 15 It has been shown that borate glasses provide the flexibility to alter key performance attributes, such as degradation time and radiopacity, through systematic compositional design. 4 In addition, glass materials may be readily milled into tightly calibrated particle sizes to fit a variety of indications. These features provide the basis upon which glass particle technologies for TAE may provide standardization, optimization, and personalization for patients and physicians. The authors have recently investigated the feasibility and composition-structure-property relationships of a tertiary glass series comprising 70B-(30-X)Rb-XSr (where X = 246,810 mol%) as the basis of establishing the potential utility of borate glasses for TAE applications. 4 One glass composition, BRS2 (70B-28Rb-2Sr), demonstrated radiopacity and degradation timeframes commensurate with the intended performance of a transient embolic device. 4 In particular, this glass network has been show to hydrolytically degrade by 93% in 48 hours using a solution of 10% fetal bovine serum and DMEM media. 4 The objective of this study was to examine the feasibility of this borate glass in microsphere form, as transient embolic devices, and to test, in a pilot in vivo model, the safety and efficacy of borate glasses against recognized controls of performance, safety and efficacy in a manned commensurate with the relevant guidance documents produced by the food and drug administration (FDA) relating to the evaluation of vascular embolic devices. 16
Materials and methods
Degradable embolic microspheres
The glass composition, BRS2 (70 mol% B2O3, 28 mol% Rb2O and 2 mol% SrO) was synthesized and processed to microspheres as reported elsewhere.4, 15 These test articles where processed through sieves with 100 and 300 µm apertures. Particle size characteristics are provided in Table 1. The test article microspheres had a density of 2.86 g/cm3, and an onset glass transition temperature of 427.9°C. The microspheres were free from identifiable crystalline species as confirmed using X-Ray Diffraction (XRD) and the sphericity of the particles was confirmed to be 92.0%, with a 94.6% symmetry. 15 Microspheres were mounted on Scanning Electron Microscopy (SEM) stubs using carbon paste and coated with 20 nm of gold-palladium to ensure that no contamination of the product was evident which would confound any clinical observations. Multiple sites of each sample were then inspected using a model No. S-4700 SEM (Hitachi, Chula Vista, CA) operating at an accelerating voltage of 15 kV, a working distance of 12.3 mm, and using magnifications up to 3500 (Figure 3). The microspheres were sterilized using 30 kGy gamma irradiation with ancillary devices being sterilized using ethylene oxide gas sterilization. One control package comprising 100–300 µm BeadBlock® and a traditional delivery stopcock where also prepared for the pilot study.
Spherical data and particle sizes of BRS2 microspheres.
Selected microbiology values immediately prior to embolization and immediately prior to sacrifice, and reference ranges. 18
*Creatinine and urea are reported in units of µmol/L.
**Creatine kinase is reported in units of U/L.
***Leukocyte count is in absolute values, in units of 109/L.
Animal model and environmental conditions
A bilateral swine renal model was reviewed and approved by both Dalhousie’s and CHUM’s Institutional Animal Care and Use Committee to ensure compliance with Canadian Council on Animal Care (CCAC) regulations per Animal Ethics protocol ’Dalhousie 18–008’. Three cohorts were used to analyze embolization effectiveness and distribution/migration at time equal to t = 0, 24, 48 h, as well as one 48-hour control. Four female pigs (75–85 kg, Ferme Triporc, Saine-Elizabeth, QC) were acclimatized for at least 6 days prior to interventional procedures under the supervision of a veterinarian. All housing conditions met or exceed the standards set forth in the Guide for the Care and Use of Laboratory Animals.
Pre-interventional procedure
Each animal was assigned a cohort (t = 00, 24, 48 hours, or control) with one animal per cohort. Animals were labeled with their timepoint and T for test article or C for control (48-hour timepoint only). Animals were initially anesthetized with 25 mg/kg ketamine and 2 mg/kg xylazine administered intramuscularly (IM). Anesthesia induction for tracheal intubation was achieved with 1.66 mg/kg propofol injected intravenously (IV). Upon induction of anesthesia, the subject animal was intubated and supported with mechanical ventilation. Isoflurane in oxygen was administered to maintain a surgical plane of anesthesia. Animals received one dose of antibiotic Cefazolin® (22 mg/kg) at the beginning of the procedure to prevent postoperative infections (excluding the t = 0 cohort). To prevent pain sensitization and minimize postoperative pain, a starting IV bolus of fentanyl (0.005 mg/kg), ketamine (0.2 mg/kg), and lidocaine (1 mg/kg) were administered followed by a continuous infusion of a dose of 0.005 mg/kg/h fentanyl, 0.3 mg/kg/h ketamine, and 3 mg/kg/h lidocaine. All procedures were performed under general anesthesia with cardiac blood pressure and oxygen saturation monitoring.
Embolization procedure
The pigs underwent a bilateral renal artery (caudal branch) embolization to determine the safety and efficacy of the test article. After disinfection and sterile draping, the right femoral artery was accessed under ultrasound guidance to insert a 5 French introducer perfused with heparinized saline at 2 IU/mL. Before any contrast was injected, a C-Arm CT was performed to evaluate the initial size and morphology of both kidneys.
Prior to particle injection, a digital subtraction angiography (DSA) with a selective opacification of the target renal artery was performed. All bifurcations downstream to the caudal branches were recorded as D1, D2, D3, where D1 referrers to the immediate and largest branches of the caudal artery, the following subsequent branches from this artery were labelled D2, and any further and/or terminal branches were all labelled as D3, as demonstrated in Figure 1. After D3, the resolution of the arteries was too variable to be assessed systematically, thus this was the most distal division assessed. It should be noted that the diameter of the branches and their presumed location in 3-D space were also taken into consideration to determine whether a branch was D1, D2 or D3; thus, a major branch (e.g. D2) could give off multiple D3 branches before finally bifurcating into the terminal D3 branches. 17
Division System Method: D1 refers to the immediate and largest branches of the caudal artery, subsequent branches of this artery are labelled D2, further branches labelled D3, etc. 13
The initial protocol planned to use a 2.4 Fr microcatheter advanced coaxially inside a 4 French catheter to perform selective embolization, but effective delivery of the particles was not possible. On multiple occasions the microspheres became aggregated in the hub of the catheter. Thus particles were injected through a 4 French cobra catheter (Terumo, Tokyo, Japan) advanced selectively into the proximal portion of the right and left caudal renal arteries, respectively, under fluoroscopic guidance.
The initial protocol planned to inject the radiopaque test particles with saline while relying only on particle radiopacity for particle visualization. Since real time radiopacity under fluoroscopy during embolization was suboptimal to prevent non target embolization. The protocol was modified and subsequent embolizations were performed using a 50% contrast/saline injection.
Fluoroscopy loops and shots were acquired during particle injection. Microspheres were injected until effective stasis was reached (stasis of contrast media in the target vessel for 5 or more cardiac cycles). 13 The total volume of TA delivered was recorded for each kidney, as well as total duration of the each procedure, and the ease of use of the material. One shot films and C-Arm CT (in the angiography suite) images were acquired to evaluate the distribution of the radiopaque microspheres in each animal. Post-embolization DSA was performed to evaluate the degree of vessel occlusion and document embolization efficacy. All catheters and sheaths were removed after embolization and final angiography and hemostasis obtained by manual compression.
Post procedure recovery and euthanasia
Meloxicam (0.3 mg/kg) and buprenorphine (0.005 mg/kg) were given as post-operative analgesia. Odansetron (0.1 mg/kg) was administered on animals 48 C, 48 T and 24 T for post-operative nausea and vomiting. Following the procedure, animals were allowed to recover and resume eating. Animals were monitored daily for signs of post embolization syndrome (PES), respiratory distress, signs of stroke/neurological damage, and bleeding. At their designated timepoints, animals (in deep anesthesia) were euthanized by a lethal injection of sodium pentobarbital (euthanyl, rapid IV bolus, 108 mg/kg). All treated animals were subjected to necropsy, defined as gross examination of the embolized arteries and kidneys, whole body (external surface), all orifices, thoracic and abdominal cavities, and brain.
In vivo follow-up imaging
An outline of the order of medical imaging per procedure can be seen in Figure 2. Pig 00 T was euthanized just after completion of the post embolization control fluoroshots, C-arm CT and DSA acquisitions. For pig 24 and 48 T fluoroshots, c-arm CT (250 µ) and DSA was repeated just before sacrifice using the methods previously detailed.
Order of Imaging Techniques Used.
The following grade scale was used for qualitative evaluation of recanalization on DSA:
0 = No angiographic visible signs of arterial occlusion.
1 = Reduction of parenchymal staining of the dependent territory.
2 = Reduction of the parenchymal staining and occlusion of the supplying arcuate artery.
3 = Reduction of parenchymal staining, occlusion of the supplying arcuate artery and occlusion of the feeding artery downstream of the catheter tip.
The lungs and kidneys were scanned by C-Arm CT after the necropsy to evaluate potential particle migration.
Ex vivo imaging
The explanted kidneys underwent additional imaging via Dental Cone-Beam CT (DCBCT) with higher resolution (75 µm) to give a better visualization of the remaining particles, and the area of ischemia.
Quantitative evaluation of particle volume
The particle were segmented on the DCBCT using the program MITK Workbench 2016.11.0 thought the threshold tool, to separate the particles from the surrounding kidney tissue by comparing their hyper-density and creating a polygonal model. After that, the volume of each polygonal structure was measured using MeshLab v1.3.4BETA to calculate the total volume of residual particles.
Histological analysis
Each kidney for each animal was harvested free of fat and the cranial pole was identified in a consistent manner. To allow good fixation, the kidney were sliced longitudinally in two sections from the lateral aspect to the hilus. Both kidneys from an animal were placed in neutral-buffered formalin (NBF). Selected organs (brain and samples of right middle pole and left caudal pole of the lung) were placed into appropriately labeled containers filled with NBF. Lesions were documented and collected when feasible, immersion-fixed in NBF and processed for histology. All collected samples underwent macroscopic examination prior to sectioning. After necropsy, ex vivo imaging, and macroscopic examination, three renal blocks of 1.0 cm3 were sectioned per kidney (two from the caudal pole and one from the cranial pole) opting for places with a higher concentration of particles. Each block was submerged in formalin for 6 hours, then refrigerated at 4°C for 48 hours (changing the formalin before refrigeration and again at 24 hours). Blocks were removed from formalin and frozen at −20°C before being embedded in paraffin. Blocks were then sectioned with a microtome to yield 3 series of 3 section, 4 µm in thickness, separated by ∼200 µm. All the slides were stored at 4°C and stained in Hematoxylin and Eosin (H&E) to allow observation of occluded vessels and the locations of identified microspheres. 17 Slides were examined for remaining TA particles, inflammatory cell infiltration, and recanalization.
Results
Particle size analysis
Characterization data for BRS2 was reported previously. 4 Particle size distribution and spherical analysis data is listed in Table 1. The columns labeled D10, D50, and D90 illustrate that 10%, 50% and 90% respectively, of particles in suspension are below the stated particle size in µm. The true particle size distribution is higher than the standard sieves used. Symmetry analysis of the microspheres confirmed perfect spherical symmetry of over 90% of the particles.
Scanning electron microscopy
SEM images (Figure 3) display spherical particles with a size range of ca. 100–300 µm prior to injection. Their surfaces appear to be smooth with no chipping or cracking visible on the surfaces of the spheres.
SEM images of microspherical BRS2 particles (100x).
Animal recovery
Following embolization, many animals experienced post anesthesia nausea and vomiting, however no animal exhibited any signs of post embolization syndrome (PES), respiratory distress, or signs of stroke/neurological damage. Oxygen saturation levels above 90% were observed at all time points. Animal 48 C developed a hematoma approx. 2.5 cm by 10 cm at the right femoral surgical site the morning after the procedure, however the hematoma did not grow in size, and began to resolve by time of sacrifice. No additional signs of bleeding were observed in any of the animals.
Biochemistry report
Blood serum levels of creatinine, urea, and creatine kinase all increased significantly following embolization, as seen in Table 2. Unfortunately, the pre-sacrifice values for animal 48 C were lost, and are thus not reported here.
Embolization procedure
The visibility of microspheres under fluoroscopy when injected only with saline was deemed suboptimal. Thus we decided to use a 50:50 contrast/saline dilution for adequate visualization of particles during embolization and prevent reflux in non-target vessels. The right kidney of animal 48 T was injected both without and with contrast, resulting in a higher volume of particle injection. The total volume of TA particles injected per kidney was very similar, averaging 279.7 mm3 (0.81 g). A higher volume of control particles was injected, with an average of 1500 mm3 (1.55 g). Total volumes and procedure times per kidney can be found in Table 3. The 00 T pig had two branches on the right caudal pole, therefore the injection was performed twice to ensure complete occlusion was achieved. The density of BRS2 is 2.89 g/cm3 and the Bead Block has a density of 1.03 g/cm3. The times described in the table correspond to the beginning of the injection of particles to the time at which effective stasis was achieved. In addition, the total time refers to the time between the initial incision and closure of the entry site at the end of the procedure.
Procedure time and volume of particles injected.
* Procedure times for the right kidney of Pig 00 T were reported separately to reflect the dual supply to the caudal lobe.
Imaging findings
Fluoroshots of all kidneys (Figure 4) show retention of contrast in the caudal pole due to effective vascular occlusion. Radiopaque particles can be seen in all kidneys embolized with test agent, (A-E) Figure 4(e) exhibits substantial contrast retention combined with the presence of radiopaque particle both in the caudal and cranial pole of the right kidney of animal 48 T due to particle reflux in non-target vessel when injecting without contrast agent. In Figure 4(f) and (g) (animal 48 C) only contrast retention is observed in since radiolucent particles were used.
Fluoroscopy image immediately following embolization. Red circles indicate retained contrast and injected particles. (a) – right 00 T kidney. (b) – left 00 T kidney. (c) – right 24 T kidney. (d) – left 24 T kidney. (e) – right 48 T kidney. (f) – left 48 T kidney. (g) – right 48 C kidney. (h) – left 48 C kidney.
Computed tomography
On CT, the particles were well delineated without inducing significant artifact at T0. On T24 and 48, radiopaque particles were still detected but significant degradation was observed when compared to T0 (Figures 5 to 7).
Figure 5 demonstrates that particles can be visualized in the caudal poles of both kidneys of animal 00 T, and not the cranial poles confirming the absence of non-target embolization. The test article particles are denser than the contrast agent and can be can be distinguished inside the embolized vessels.
Fluoroscopy, C-Arm CT, and Dental Cone Beam CT of the right and left kidneys of 00 T pig, sliced longitudinally after necropsy. Red arrows indicate BRS2 particles. Red solid line circles and yellow arrows indicate retained contrast. Red dashed lines mark the caudal pole, showing hypodense areas corresponding to the areas of ischemia.
Few non-target particles were visible in the cranial portion of the left kidney of animal 24 T immediately after the procedure (Figure 6(a)), however there was no evidence of vascular occlusion. 24 hours after the procedure, very few particles were visualized, both in the in vivo images and the explanted tissues. Only a small number of particles were visualized in the caudal portion of both kidneys using C-arm CT and DCBCT. A small amount of retained contrast can be seen throughout the entire kidney 24 hours post procedure. The darkened tissue colour (outlined in red) of the explanted kidneys upon CT indicates the area of ischemia.
Computed tomography images of the right and left kidney of animal 24 T. (a) – in vivo scans of the right and left kidney (respectively) showing retained contrast (red circle) and radiopaque beads (red arrows) immediately after injection. (b) – C arm CT images of the right and left kidneys 24 hours after embolization. (c) C-Arm CT of right and left kidney (sliced after necropsy), 24 h after the embolization, showing retained contract (yellow arrows), remaining particles (red arrows), and areas of ischemia (red line). (d) – Dental Cone-Beam CT (500HU) of right and left kidney (sliced after necropsy), 24 h after the embolization showing retained contract (yellow arrows), remaining particles (red arrows), and areas of ischemia (red line).
Immediately following embolization, it is obvious that substantial non-target embolization of the right cranial pole of animal 48 T had occurred (Figure 7(a)). 48 hours after the procedure, heterogeneous hypodense areas of ischemia are visible on both renal poles of the right kidney, and only at the caudal pole of the left kidney (Figure 7(c) and (d)). Remaining particles can be visualized in the cranial and caudal poles of the right kidney in Figure 7(d), as well as in the caudal pole of the left kidney.
Computed tomography images of the right and left kidney of animal 48 T. (a) – in vivo scans of the right and left kidney (respectively) showing retained contrast (red circle) and radiopaque beads (red arrows) immediately after injection. (b) – C arm CT images of the right and left kidneys 48 hours after embolization. (c) C-Arm CT of right and left kidney (sliced after necropsy), 48 h after the embolization, showing remaining particles (red circles), and areas of ischemia (red arrows). (d) – Dental Cone-Beam CT (500HU) of right and left kidney (sliced after necropsy), 24 h after the embolization showing remaining particles (red circles), and areas of ischemia (red arrows).
Particles are not visible in animal 48 C, as they are radiolucent. The darkened tissue colour in the caudal portion of the explanted kidneys upon CT indicates the area of ischemia. No non-target areas of ischemia are visible.
The volume of particles remaining at time of sacrifice, calculated via DCBCT, is listed in Table 4. Values for 00 T were assumed to be the same as the volume injected, as contrast confounded the imaging technology. The DCBCT data indicates that 98.8% and 97.2% of the particles injected into the right and left kidney, respectively, were degraded at the 24-hour timepoint. Further, over 96% (96.6% and 99.9% for right and left respectively) of the test article was completely degraded after 48 hours.
Volume of test article particles in tissue after embolization and at time of sacrifice.
DSA
All vessels feeding the caudal poles were effectively embolized in animal 48 C. The right kidney shows a small amount of recanalization after 48 hours, without significant revascularization. Significant non-target embolization of the cranial pole of the right kidney of animal 48 T was visible, accompanied by a large volume of visible particles and retained contrast. All distal branches of the caudal pole of the left kidney were properly embolized.
The right kidney of animal 24 T exhibits patent D2 and D3 vessels post embolization coming from a segmental artery arising proximal to the caudal artery which was not embolized (Figure 9(o) and (q)), All vessels coming from the embolized polar artery were effectively occluded. Effective occlusion of all caudal vessel of the left kidney was achieved. No non-target vessels appear embolized. Animal 00 T received effective embolization of the right caudal pole, occluding all branches from D1 to D3. The left kidney of animal 00 T (Figure 9(t) and (v)) was less effectively embolized, leaving a small number of D2 and D3 vessels partially patent. No non-target vessels appear embolized.
Recanalization was assessed via angiography and given a score based on the level of occlusion (Table 5). No significant recanalization was visible in any animal. After 48 hours, test article and control occlusion levels were essentially the same. Further details were assessed in histology.
Angiographic assessment of recanalization.
Macroscopic necropsy findings
Both kidneys of pig 48 C displayed a distinct demarcation between the pale cranial pole and the dark caudal pole suggesting ischemia and necrosis in this portion of the organs. The longitudinal cut of the kidney revealed areas of infarcts in the renal cortex. The necropsy findings are consistent to the radiology images. No non-target organs were infarcted. The surface of both kidneys in pig 48 T presented as a mosaic of large irregular areas of white, tan and dark lesions suggesting infarcts. The white lesions were more abundant in the caudal pole of the kidney. The cranial pole of the right kidney was uniformly dark ischemia with a few white irregular areas, whereas the cranial pole of the left kidney displayed only a few dark areas. The longitudinal cut of the kidney revealed areas of pale infarcts in the caudal pole of the renal cortex. The necropsy findings are consistent with the radiology images. No non-target organs were infarcted.
Pig 24 T showed marked dark areas of necrosis in the both caudal poles, as well as small areas of necrosis in the cranial lobes, especially in the left kidney. The left caudal pole was uniformly dark, suggesting ischemia; whereas the right caudal pole demonstrated a dark ischemic surface with a few large, irregular, white areas, suggesting infarcts. The necropsy findings are consistent with the radiology images. No non-target organs were infarcted. Pig 00 T displayed ischemic areas predominating in the caudal pole of both kidneys. Some areas of infarct, presenting as areas of pale and tan lesions were visible in both the cranial and caudal lobes. The ischemic area was smaller than those seen in other animals, however all necropsy findings were consistent with radiology images. No non-target organs were infarcted.
Histological analysis
Animal 00 T showed no significant areas of necrosis. Irregular and fragmented particles were observed in the lumen of the vessels, as seen in Figure 11, and in greater numbers than seen in animals 24T and 48T, in greater numbers than seen in animals 24 T and 48 T. A very small number of intact particles were also visible.
Histological analysis of the caudal poles of pig 24 T showed areas of necrosis accompanied by moderate polymorphic inflammatory infiltrate as seen in Figure 12. No intact particles were observed in the lumen, however more particle fragments were observed than in animal 48 T. The beginning of recanalization was also observed.
The renal parenchyma of the caudal pole of pig 48 T shows large areas of advanced necrosis accompanied by polymorphic inflammatory infiltrate. In the lumen of the vessel, a minimum number of fragmented particles were found, and the adjacent renal parenchyma demonstrates satisfactory ischemia. The level of recanalization was difficult to assess as many vessel lumens were filled with fibrin (Figure 13(a)). The beginning of recanalization was visible in some vessels. The cranial (non-embolized) pole of the right kidney showed obstruction of a renal artery beside a patent vessel, surrounded by necrosis tissue. This corresponds to an area of non-target embolization, as seen in Figure 13(d).
48 hours after the embolization with Bead Block®, extensive areas of advanced necrosis and vessels obstruction was observed in the control animals. The microspheres are still completely visible in the vessel lumen on histology. No non-target ischemia or necrosis was visible in the cranial lobe.
Discussion
There is substantial interest in the clinical and biomaterials literature relating to the development of degradable microsphere technologies for applications in trauma, fibroid embolization, prostatic artery embolization, and across the broader field of embolotherapy.1,3,5,11,19–21 The authors contend that borate glass networks may provide a means to engineer safe and effective embolic microspheres which exhibit congruent hydrolytic degradation coupled with multi-modal imageability to support clinical determinations of spatial distribution of such microspheres in a target tissue. 4 However, the pre-clinical performance of borate glass microspheres in vascular embolization models is absent from the literature. Accordingly, this work was performed to assess, in a pilot model, the performance of a well characterized borate glass and to establish critical performance attributes for the potential clinical use of borate glass microspheres in embolotherapy. To frame this preclinical investigation, the authors utilized existing guidance documents from FDA; specifically, the authors used the class II special controls guidance document for vascular and neurovascular embolization devices. 16 This document identifies special controls for establishing the preclinical safety and efficacy of new embolic microspheres, and in particular, underscores that new devices under investigation should consider (i) ease of deliverability (from a friction and tortuosity standpoint), (ii) acute complications, (iii) local and systemic foreign body reactions, (iv) recanalization, (v) embolization effectiveness, and (vi) device migration in order to provide a reasonable assurance of the safety and effectiveness of a new vascular and neurovascular embolization devices.3, 16 It is important to note that these special controls are in addition to all normal controls that would be expected to be utilized in the design and development of a new embolic therapy.
In addition to these special controls, the authors also considered the toxicological risk associated with borate glass networks. It is noted that acute exposure to high levels of boron have demonstrated toxicity, although the variability in the dose estimation and human response is considerably large 18 ; in children and adults exposed to greater than 84 mg boron/kg, systemic and nervous system effects—as well as death—have been observed. 18 Clinical signs of boron toxicity include gastrointestinal distress, cardiopulmonary hypotension, renal and hepatic failure, and muscular effects such as weakness and lethargy. Long-term exposure in animal studies have shown toxicity to the male reproductive system, but human data from epidemiological studies have not been able to demonstrate a significant link between boron and reproductive toxicity. Boron does not pose a risk in respect of carcinogenic potential or developmental and genetic toxicity.18, 22 Based on the procedures performed in this work, animals were subject to an average boron exposure of 118.85 mg (based on the chemistry of the degradable microspheres). All animals were monitored for cardiovascular, respiratory, and neuromuscular abnormalities following the embolization procedure, with no remarkable findings relevant to boron toxicity being noted in this study.
To investigate the special controls criteria, a bilateral renal artery embolization model was deployed per the literature.13,17, 19 It is widely accepted that this type of study be conducted in vivo, since no in vitro model currently exists that can mimic the complex biological responses to embolization. The non-atherosclerotic swine model utilized in this work is well established for angiographic/embolization studies, and there exists a large volume of data on its vascular response properties and their correlation to human vascular response.19,20, 23 In addition, important anatomical and physiological characteristics are shared between porcine and human anatomy23, 24 that further support the use of this model.
Prior to examining the host-material interactions, this research paid close attention to the ease of delivery of the borate microspheres under simulated clinical conditions. This consideration was particularly important given that variations in microsphere composition are associated with changes in overall function in the clinical literature. 1 In particular the borate microspheres utilized in this study are non-compressible, a feature commonly associated with ease of delivery through a microcatheter. Animal 48 T was the first to receive an embolization treatment and as noted (Table 3) an excess of particles was delivered during the embolization procedure. This excess delivery was related to two factors; (i) insufficient radio-opacity for particle injection without contrast, and (ii) given the pilot nature of the work there were operator familiarity issues in the clinical setting with the device utilized to introduce the microspheres into the catheters. These issues were only manifest at the early portion of the embolization procedure for animal 48 T. Of note, during the clinical deployment of the test articles, although the particles were imageable on fluoroscopy, their temporal resolution was not sufficient to monitor embolization efficacy when injected with just saline, and as such, the procedure was altered to add a mixture of 50:50 contrast/saline to permit temporal visualization of embolization. Additionally, while the protocol considered the use of microcatheters for all embolization procedures, a clinical decision was made, during embolization of animal 48 T, to opt for selective catheterization of target tissue using 4 F catheters rather than microcatheters. This deviation resulted in the injection of a larger volume of particles to achieve effective embolization of the right caudal pole. After addressing these issues there were no further issues relating to the delivery of microspheres into the catheters. Despite these two events, which artificially increased the volume of particles injected to animal 48 T, all test article kidneys received a smaller volume and weight of material than the control to achieve effective stasis of blood flow. This is encouraging and indicates that less material may be required to achieve the desired endpoint when using the borate glass microspheres.
The occlusion of the microcatheters was not related to the size of the particle (100–300µm) but rather to the nature of the dedicated 3-way stopcock utilized (as a prototype) to deliver the microspheres; succinctly, the particles had a propensity to accumulate in the delivery chamber of the stopcock and occlude the hub of the microcatheter. This is a future design consideration which needs to be considered to advance these materials to clinical use. The time to achieve effective embolization was monitored as an objective measure of ease of delivery. While each animal has unique vasculature, making direct comparison of procedural times somewhat challenging, the injection time for the test article was clinically acceptable and was noted as ca. 8 minutes, versus ca. 6 minutes for the control. These differences are likely due to operator experience with a commercially well-known product, versus a prototype product with limited clinical use. In addition, animal 00 T had two branches feeding the caudal pole of the right kidney (a double right renal artery is seen in about 1% of porcine renal vasculature 25 ) and as both branches required embolization, the procedure time for this animal was longer than expected. Collectively, these issues impacted the average time to complete embolization with the test articles. Embolization procedures on animals 00 T, 24 T and 48 C were non-eventful and completed in a clinically acceptable manner.
As previously noted, less of the test article was required to provide effective embolization versus the control. However, the reduction in volume of material did not compromise embolization effectiveness versus the control. For animals 00 T, 24 T, and 48 C, DSA and CT images confirmed that the divisions of the caudal branch (D1–D3) were effectively embolized with no non-target embolization observed, and only a small number of non-target particle visible in animal 24 T (Figures 5, 6, 8 and 9). Necropsy further confirmed that both caudal poles were effectively embolized for 00 T, 24 T and 48 C (Figure 10). Regarding the animal 24 T, all branch divisions arising from the embolized right caudal branch were effectively embolized. Patent D2 and D3 branches observed in the lower pole of the right kidney were fed by a segmental branch arising proximally from the main renal artery which was not embolized In contrast, animal 48 T which, as noted previously, received an excess of particles, exhibited discrete areas of non-target embolization, especially in the cranial pole of the right kidney (Figure 8(k)) however the divisions of the caudal branch (D1-D3) were confirmed as effectively embolized with DSA and CT. At necropsy, heterogeneous areas of ischemia in the right kidney of 48 T and were correlated with the presence of non-target embolization (Figure 10(e)).
Computed tomography images of the right and left kidney of animal 48 C. (a) – in vivo scans of the right and left kidney (respectively) showing retained contrast (red circle) immediately after injection. (b) – C arm CT images of the right and left kidneys 48 hours after embolization. (d) C-Arm CT of right and left kidney (sliced after necropsy), 48 h after the embolization, showing areas of ischemia (red arrow). (d) – Dental Cone-Beam CT (500HU) of right and left kidney (sliced after necropsy), 48 h after the embolization showing areas of ischemia (red arrow).
Renal Arteriograms of animals pre-embolization (a,b,g,h,m,n,s, and t), immediately post embolization (c,d,i,j,o,p,u, and v) and immediately prior to sacrifice (e,f,k,l,q, and r).
Macroscopic images of all test kidneys. (a) – 00 T right kidney. (b) – 00 T left kidney. (c) – 24 T right kidney. (d) – 24 T left kidney. (e) – 48 T right kidney. (f) – 48 T left kidney. (g) – 48 C right kidney. (h) – 48 C left kidney. Dark coloring indicates necrotic areas.
Histologic aspects of renal tissue that received embolization with TA particles immediately following embolization. Magnification 10x. (a) – Right kidney with an intact (3.9 µm) particle, (b) – Left kidney; (Arrows) TA particle, both intact and fractured; (RA) Renal Artery.
Histology of renal tissue that received embolization with TA particles after 24 hours. Magnification 10x. (A) – Right kidney, (B) – Left kidney; (*) Remaining degraded particles; (RA) Renal Artery; (RC) Recanalization; (T) Thrombus; (N) Necrotic area.
Given that the test articles were engineered to undergo hydrolytic degradation, the study also examined the persistence of embolic agent within the vascular anatomy using diagnostic imaging. To measure the volume of particles remaining in the vasculature after sacrifice, DCBCT was utilized, as it offers enhanced resolution versus C-Arm CT (75 µm vs 250 µm). Animal 00 T presented the largest volume of particles, followed by 24 T, and finally 48 T had the smallest volume of particles remaining. Since animal 00 T was sacrificed immediately following embolization, there was insufficient time for the contrast media to dissipate; this resulted in confounded CT images. As such, due to the extremely short time between injection and sacrifice, it was assumed that the volume of particles injected was the same as the volume of particles remaining at sacrifice. While the predicted degradation timeframe of these particles in vitro is ca. 48 hours, 4 DCBCT measurements show that the actual degradation timeframe is <24 hours under physiological conditions. The DCBCT data indicates that 98.8% and 97.2% of the particles injected into the right and left kidney, respectively, were degraded at the 24-hour timepoint (Table 4). Further, over 96% (96.6% and 99.9% for right and left respectively) of the test article was completely degraded after 48 hours. Additionally, due to the use of contrast, there is a chance that upon final imaging some of the retained contrast artificially increased the area of hyperdensity used to measure the remaining particles at sacrifice, thereby overestimating the volume of particles remaining. Thus a larger portion of particles may have degraded than what is reported in this study. Regardless, these glass microspheres are essentially completely absent 24 hours following injection.
With respect to the durability of the occlusion, the DSA data shows that little to no recanalization has occurred at 24 or 48 hours after embolization with the test article, despite their degradation at these timepoints. In fact, the scores seen at 48 hours were very similar to those evaluated immediately after embolization for both the control and the test article (Table 4). It is likely that occlusion in the absence of the test article is due to natural thrombus and fibrin formation within the vessel which has not yet resolved following the degradation of the test article. Further investigations comprising longer timeframes for follow-up should be considered to properly assess the resolution of natural thrombotic effects. Additionally, significant increases in both urea and creatinine were seen at the time of sacrifice, indicating that although the kidneys were still functioning (due to the retained cranial function), they had sustained significant injury from embolization. Animal 48 T showed the largest increase in both creatinine and creatine kinase, most likely due to the substantial non-target embolization of the cranial pole of its right kidney.
Although there is a marked increase in leukocyte values for animals pre-sacrifice compared to pre-injection, these values are still within normal limits and do not indicate the presence of a systemic infection. Most likely they represent a localized response to the embolic particles that have temporarily increased inflammatory cell production. However, due to the short nature of this study, this increase in leukocytes may alternatively be the beginnings of a more wide scale inflammatory cascade. Additional studies with longer time frames are required to assess the systemic resolution of the inflammatory effects.
From a histopathological standpoint, animal 48 C demonstrated complete occlusion of the vessels, leading to advanced areas of necrosis in both caudal poles. Additionally, a large volume of fragmented control particles was visualized in the vessel lumen (Figure 14). Although no particles were visualized during histological evaluation in animal 48 T, similar levels of tissue necrosis versus the control were observed, further supporting the evidence that suggests the transient nature of the test article did not compromise the effective and complete embolization of targeted tissues. For animal 48 T, it was observed that the renal arteries were either completely occluded by fibrin and/or thrombi or had begun to display early stages of recanalization (Figure 13). No significant number of inflammatory cells were visualized in any slides harvested from the animals treated with the test articles, most likely due to the short nature of the study. Interestingly, very few particles were found in the histology slides of animals 00 T and 24 T. Animal 00 T displayed no significant areas of necrosis as there was not enough time after embolization for ischemic necrosis to occur. Significantly more particles were visualized in the vessels of 00 T versus 24 T or 48 T. However, residual particles visualized during histopathological evaluations for animal 00 T had diameter of ca. 4 µm, indicating that significant degradation may have occurred in a short period of time (initial diameter was 100–300 µm) post implantation. There are two possible explanations for this: 1) these particles degrade much more quickly than anticipated, and 2) the particles continue to degrade during the histological processing. No migration of the test or control articles was observed; all examined organs appeared normal after macroscopic evaluation, and CT of the lung and brain also displayed normal physiology. The caudal portion of diaphragmatic lobes of all animals presented with atelectasis and congestion, consistent with prolonged dorsal recumbency and anesthesia and was not related to the presence of embolic materials.
Histology of renal tissue that received embolization with TA particles after 48 hours. Magnification 10x. (a) – Right kidney caudal pole, (b,c) – Left kidney caudal pole; (d) – Right kidney cranial pole; (Arrows) fragmented particles; (RA) Renal Artery; (RC) Recanalization; (f) Fibrin; (N) Necrotic area; (T) Thrombus.
Histology of renal tissue that received embolization with Bead Block®. Magnification 10x. (a) – Right kidney caudal pole, (b) – Left kidney caudal pole; (*) Control particles, both spherical and irregular shapes; (RA) Renal Artery; (N) Necrotic area.
Conclusion
The development of degradable, radiopaque particles for transarterial embolization may potentially overcome existing limitations for TAE materials, leading to a standardized, optimized and personalized procedures. Borate glass provides both the required radiopacity and tailorable degradation rates desired from such a material. Medical imaging and histopathological evaluations have revealed that borate-rubidium-strontium glass microspheres can provide durable occlusion of renal arteries resulting in significant ischemia of the desired target. Additionally, these particles display an in vivo degradation time of ca. 24 hours, and do not migrate to non-target areas (e.g. brain, lung). Although further studies are needed, the data collected here suggests that borate glass microspheres may prove suitable for applications in transient vascular embolization.
Footnotes
Acknowledgements
The authors thank the personnel of the CRCHUM Animal core facility for their expert collaboration in the project and their skilled technical support.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors gratefully acknowledge the financial assistance of NSERC, and the Idea to Innovation fund [award number 505848–17] for supporting this work.
