Vascular Lesion–Specific Drug Delivery Systems

Abstract

Drug delivery is central to modern cardiovascular care, where drug-eluting stents, bioresorbable scaffolds, and drug-coated balloons all aim to restore perfusion while inhibiting exuberant healing. The promise and enthusiasm of these devices has in some cases exceeded demonstration of efficacy and even understanding of driving mechanisms. The authors review the means of drug delivery in each device, outlining how the technologies affect vascular behavior. They focus on how drug retention and response are governed by lesion morphology: lipid displacing drug-specific binding sites, calcium inhibiting diffusion, blocking thrombi or promoting luminal washout, and vascular healing steering hyperplastic developments. In this regard, the authors outline the fundamental impact of vascular structure on drug delivery and review the development of contemporary and future devices for coronary and peripheral intervention. They look toward a future where incorporating information on lesion distribution is central to therapeutic success and envision a transition toward lesion-specific treatment for improved interventional outcomes.

Percutaneous and endovascular interventions have revolutionized the treatment of atherosclerosis, particularly when mechanical devices such as stents are combined with local drug release. The incorporation of drugs in the form of drug-eluting stents (DESs) and drug-coated balloons (DCBs)—both developed to tackle the occurrence of excessive hyperplastic growth and in-stent restenosis associated with noncoated equivalents—have enabled clinicians to treat de novo disease and mitigate recurrent postinterventional restenosis with concentrated local drug effect, minimizing systemic toxicity. What has become clear is that although devices differ in their mode of delivery, it is drug uptake and retention that determine therapeutic efficacy. There the interaction between device and apparent lesion morphology is dominant. In this review, we discuss the impact of atherosclerotic morphology on drug retention, clarify how complex lesion characteristics directly affect the efficacy of drug delivery devices—contemporary and future—and envision a transition toward lesion-specific treatment for improved future care.

ATHEROSCLEROTIC DISEASE AND DRUG DELIVERY

The diffusion and retention of drugs into the vascular wall—a prerequisite for clinical effect—is inherently coupled to vascular morphology. It is then imperative to understand the basics of atherosclerotic biology from a perspective of pharmacokinetics and dynamics, and in particular how lesion morphology influences the retention of drug, before reviewing specific techniques for drug delivery.

ATHEROSCLEROTIC DISEASE AND HYPERPLASTIC VASCULAR RESPONSE.

Atherosclerosis has both a passive, structural barrier that constructs luminal flow and as inhibits transmural transport, as well as an active, metabolic element that bears profound drug metabolism potential. The inflammatory aspect of disease is critical, and it is within this vantage point that drug delivery requires an understanding of different plaque phenotypes (1,2), and characteristic stages of lesion progression (2,3). Endothelial dysfunction, uptake of circulating blood cholesterol, creation of subintimal fatty streaks of oxidized low-density lipoprotein crystals, and macrophage-induced foam cell formation (3) all take on new meaning when their impact on drug penetration and retention are considered. Drugs will be handled differently by lesions dominated by collagenous-rich fibrotic caps, lipid-rich necrotic cores, agglomerates of calcium, and areas of intraplaque hemorrhage or active intraplaque neovascularization (3). The distribution of each entity varies greatly among individual plaques, and the possible progression into a late-stage vulnerable lesion is dependent on a multitude of factors (1).

Local delivery is dependent then on de novo lesion formation and the vascular response associated with device implantation (4). In the case of an angioplasty balloon, the inflated device abrades the endothelium, exposes intraplaque dissection planes, and triggers acute thrombosis and hyperplastic response (5). The process is even more complex when permanent devices are implanted, where stent struts induce local thrombosis, inflammation, smooth muscle cell proliferation, and vascular remodeling (6). It is also important to stress the role of the rheologic and hemodynamic environment (7), where modified wall shear stresses at an interventional site accelerates pathologic restenosis (8). Here, adding local drug delivery is imperative in avoiding hyperplastic response.

THE IMPACT OF VASCULAR AND ATHEROSCLEROTIC STRUCTURE ON DRUG RETENTION.

The idea that retention and delivery of drug is directly affected by vascular structure and lesion composition has evolved over time (Table 1, Central Illustration, Figure 1). In seminal experimental work, Hwang and Edelman (9) evaluated the distribution of hydrophilic dextran in the healthy arterial wall, showing that: 1) the drug favors binding to connective tissue elastin; 2) the drug favors transport along vascular fiber directions; and thus 3) the arterial ultrastructure modulates local accumulation of the drug. For hydrophobic drugs, vascular tissue structure is equally important. Paclitaxel, a commonly used antiproliferative drug inhibiting cell replication by means of microtubule stabilization (10), deposits primarily within the intima and adventitia (11), a fact explained by drug-specific binding to intracellular tubulin (12).

TABLE 1.

Vascular, Atherosclerotic, and Device-Induced Aspects Affecting Local Drug Delivery

Component	Influence on Drug Uptake	Influence on Drug Retention
Vascular wall
Intima	Modulating uptake and retention and acting as an initial permeation barrier	-
Media	High in nonspecific binding sites following high elastin content, also favoring drug transport along elastin and collagen fibers	-
Adventitia	Final permeation barrier, but also rich in microtubule binding-sites for paclitaxel	-
Atherosclerotic plaque
Lipid	Displacing drug-specific binding sites by the formation of foam cells and fatty streaks	Decreased uptake of drug
Calcium	Drug permeation barrier and inert binding nature	Decreased uptake of drug, but could also trap drug and create local concentration extrema
Fibrous	Elastin-rich nature favors nonspecific binding; similarly, collagen fibers act as transportation routes for drug	Increased uptake of drug*
Hemorrhage	Typically present only in actively remodeling plaques, increasing permeation and promoting transluminal transport	Increased uptake of drug*
Device-induced aspects
Endothelial denudation	Denudation removes a convection barrier and exposes medial binding sites for elastin, but removes subintimal microtubular binding	Increased uptake of drug (rapamycin) or decreased uptake of drug (paclitaxel)
Rupture of elastic lamina	Rupture again removes barriers for drug transport, but might induce detrimental healing processes	Increased uptake of drug
Thrombus formation	Potentially induced by stent placement, acting as an inert diffusion barrier; adluminal thrombus (between stent and lumen) will restrict washout and increase retention, whereas abluminal thrombus (between stent and wall) will decrease retention	Increased uptake of drug (adluminal) or decreased uptake of drug (abluminal)
Mechanical compression	The inflation of an angioplasty balloon reduces diffusivity by packing vascular tissue layers together; vascular compliance is also modified, negatively influencing acute coating transfer	Decreased uptake of drug*

^*Requires further experimental validation.

CENTRAL ILLUSTRATION. Influence of Vascular Biology and Intervention on Drug Delivery.

Vascular, atherosclerotic, and device-induced aspects affecting local drug delivery. (A) In the healthy artery, the intima, media, and adventitia present with different levels of permeation blockage, as well as specific and nonspecific binding sites for delivered drug. In the diseased lesion, calcium, lipid, fibrous, and hemorrhagic entities all modify uptake and retention. (B) Device-induced aspects such as endothelial denudation, elastic lamina rupture, mechanical compression, and thrombus also influence the delivery of drug. mTOR = mammalian target of rapamycin; SMC = smooth muscle cell.

FIGURE 1. Impact of Atherosclerotic Characteristics on Drug Delivery.

Extracts from seminal work, showcasing the impact of vascular and atherosclerotic structure on drug retention. (A) En face microscopy of paclitaxel (PTX) delivery after drug-eluting stent (DES) implantation in ex vivo bovine carotids, clearly showing how drug delivery is governed by stent position and design (115) (reprinted with permission from the American Heart Association). (B) Equilibrium distribution of dextran (DXT), sirolimus (SRL), and PTX in ex vivo calf carotids, showing variations in transluminal tissue-binding (12) (reprinted with permission from the National Academy of Sciences). (C) Transluminal distribution of PTX in atheromatous vs. healthy control rabbits, showing definite effect of lesion complexity (19). (D) Transluminal distribution of SRL in atheromatous vs. healthy control rabbits, showing less effect of lesion complexity (19). (E) Vascular drug retention as a function of lipid content (evaluated in a rabbit model), showing how increasing lipid content reduces drug uptake (19) (C to E reprinted with permission from Elsevier). (F) Drug absorption as a function of calcium content (evaluated in cadaveric postmortem specimens), showing reduced permeation with increasing calcium (24) (reprinted with permission from Elsevier). (G) The presence of stent-induced thrombi (immediately adjacent to the stent strut) blocks transluminal diffusion and reduces uptake compared with control vessel (31) (reprinted with permission from the American Heart Association).

Tissue binding and clearance also dictate the success of vascular drug delivery devices. Hydrophilic drugs, even those exhibiting excellent antiproliferative abilities and high diffusivity, are not well suited for modulating neointimal hyperplasia, where a lack of tissue-binding capacity (12) leads to tissue washout within hours of application. In contrast, hydrophobic drugs remain resident for days at concentrations significantly higher than the applied bulk load, all due to tissue-specific binding capacity: paclitaxel showing affinity for microtubule binding and rapamycin (another antiproliferative drug inhibiting cell cycle–dependent kinases and delaying phosphorylation of retinoblastoma protein [13]) binding to proteins FKBP12 (FK506-binding protein 12) or FRAP (FKBP12-ramapycin–associated protein). As such, rapamycin exhibits fairly homogeneous transmural distribution, compared with the aforementioned favoring of intimal and adventitial spaces by paclitaxel (12). Different formulations are used interchangeably within clinical practice, with only recent clinical data supporting sirolimus-based devices in a coronary setting (14,15), potentially corroborating the mechanisms of different drug-specific binding uncovered in preclinical work. However, a remaining issue with sirolimus-based coatings is the comparably low lipophilicity and difficulty in release control (16), representing decreased long-term efficacy (something now addressed through polymer-encapsulated sirolimus coatings [17,18]).

The pathologic changes imposed by atherosclerosis or hyperplasia also significantly affects drug affinity. Nonhuman animal models of postinterventional neointimal response have existed for decades, but the heterogeneity of lesions complicates the isolated study of intraplaque components and drug uptake. Nevertheless, in an influential work Tzafriri et al. (19) evaluated steady-state arterial distribution of paclitaxel, sirolimus, and everolimus (a 40–0-[2- hydroxyethyl] derivative of sirolimus [20]) in atherosclerotic human and rabbit tissues, relating uptake to lesion complexity. Lipid-rich arteries exhibited up to 3-fold lower affinity for the evaluated drug formulations, meaning that hydrophobic drug was effectively repelled by the lipid environment. Albeit counterintuitive, the explanation is again that drug deposition is highly regulated by the intracellular binding targets. As atherosclerotic fatty streaks form, adjacent cells expressing tubulin and FKBP12 are displaced, removing drug-specific binding targets from the lipid-rich environment, decreasing drug affinity, and dominating over ultrastructural dependence (9). Other in vitro work indicated how a lipid environment directly influences drug partitioning (21), and in clinical coronary stenting, lipid-rich lesions have been linked to increased neointimal growth (22,23), supporting the experimental findings of decreased drug deposition with lipid presence.

The vascular injury imposed by an interventional procedure also affects uptake and retention of drug (19,24,25). In healthy arteries, the endothelium acts as a control barrier, modulating the flux of particles in and out of the vascular wall (26). This modulating monolayer is disrupted with the inflation of an angioplasty balloon or the implantation of a stent. Loss of endothelium facilitates drug diffusion and convection into the tissue, increasing tissue penetration and transmural drug effect, a fact corroborated by experimental findings of increased uptake of heparin following endothelial denudation (25). For hydrophobic drugs, endothelial denudation modulates the access to drug-specific binding sites. With the elastinrich medial layer exposed by endothelial denudation, the nonspecific and specific binding of rapamycin—a drug distributing fairly evenly throughout the media (12)—might be assumed to increase. Whereas the nonspecific binding of paclitaxel increases with medial exposure, the drug-specific binding–being predominantly subintimal for paclitaxel (12)—might instead decrease following the same procedure. The clinical effect of excessive injury during intervention has been indicated as detrimental to DES performance (27); however, clinical work remains to verify this mechanism.

In atherosclerotic vessels, this reasoning is compounded by lesion complexity, and when no explicit assessment of lesion morphology has been included, contradictory results have been reported. In the setting of DCBs, Fernández-Parra et al. (28) observed a 4-fold increase in paclitaxel uptake in diseased rabbit vessels, whereas others have reported decreased uptake in similar settings (29). Although the increase could be attributed to the aforementioned changes in diffusion and access to tissue binding sites, the compression of the underlying tissue layer (30) may limit transmural transport (9). The possible thrombus formed around an implanted stent might also affect uptake (31), with both promoting and inhibiting effects present: clot positioned between struts and the vascular wall blocking diffusion and clot positioned around struts promoting vascular uptake through prohibited washout.

Vascular calcification also affects drug permeability, being especially dominant in peripherals, where extensive calcifications are commonly reported (32). Calcium is inherently inert, inhibiting tissue penetration by posing an impenetrable structural diffusion barrier (33). Tzafriri et al. (24) evaluated paclitaxel uptake in calcified human femoropopliteals, showing that calcium removal increased diffusivity by 70%, along with a log-linear increase in bulk absorption rate, again with the effect dominating over native tissue binding alone (9). Numerous clinical studies also have linked decreased treatment efficacy with increasing calcium (34,35). Similarly, the morphology of calcified deposits has recently been proposed as influential for interventional outcome (36,37), where superficial or sheet-like calcium might inhibit drug uptake to a larger extent than what deep or regionally nodular calcium might do. Validating experimental work, however, remains to be performed.

DEVICES FOR VASCULAR DRUG DELIVERY

For the catheter-based interventional treatment of vascular disease, contemporary devices typically incorporate local drug delivery. Although early attempts were made with the use of drug-eluting balloons, it is through DESs that drug delivery became a bedrock cardiovascular tool. The knowledge gathered from the use of DESs spurred the development of contemporary DCBs, so a review of techniques naturally begins with DESs. Below is an overview of the evolution of drug delivery devices, highlighting contemporary developments, emphasizing the relationships among device, delivery, and vascular environment, as well as coupling interventional success to lesion complexity. To highlight the wide selection of devices currently on the market, Tables 2 to 4 list some relevant commercially available devices within separated device classes.

TABLE 2.

Overview of Commercially Available Drug-Eluting Stents

Drug and Device	Company	Polymer Coating	Base Material	Drug Load, mg/mm2	Vessel Territory
Paclitaxel
Eluvia	Boston Scientific	Fluoropolymer	Nitinol	0.2	Peripheral
Zilver	Cook Medical	Polymer-free	Nitinol	3.0	Peripheral
Sirolimus
Orsiro	Biotronik	PLLA	CoCr	1.4	Coronary
Ultimaster	Terumo	PDLLA-PCL	CoCr	3.9	Coronary
MiStent	Micell Technologies	PLGA	CoCr	2.4	Coronary
Combo	OrbusNeich	PLA-copolymer	Stainless steel	5.0	Coronary/peripheral
Everolimus
Synergy	Boston Scientific	PLGA	PtCr	1.0	Coronary
Xience Alpine/Sierra	Abbott	Fluoropolymer	CoCr	1.0	Coronary
Promus Premier/Elite	Boston Scientific	PVDF-HFP	PtCr	1.0	Coronary
Promus Element	Boston Scientific	PLGA	PtCr	1.0	Peripheral
Novolimus
DESyne	Elixir Medical	PBMA	CoCr	5.0	Coronary
Zotarolimus
Resolute Integrity/Onyx	Medtronic	BioLinx	Co-alloy	1.6	Coronary
Umirolimus/Biolimus
BioMatrix Nobori/Flex	Biosensors	PDLLA	Stainless steel	*	Coronary/peripheral
BioFreedom	Biosensors	Polymer-free	Stainless steel	*	Coronary/peripheral
Ridaforolimus
EluNIR	Medinol	Elastomer-based (eDES)	CoCr	1.1	Coronary

^*Data unavailable.

CoCr = cobalt-chromium; eDES = elastomeric drug-eluting stent; HFP = hexafluoropropylene; PBMA = poly(butyl methacrylate); PCL = polycaprolactone; PDLLA = poly(D,L-lactic acid); PLA = polylactic acid; PLGA = poly-L-glutamic acid; PLLA = poly-L-lactic acid; PtCr = platinum-chromium; PVDF-HFP = poly(vinylidene fluoride)-co-hexafluoropropylene.

TABLE 4.

Overview of Commercially Available Drug-Coated Balloons

Drug and Device	Company	Polymer Coating	Drug load, mg/mm2	Vessel Territory
Paclitaxel
Agent	Boston Scientific	Acetyl tributyl citrate	2.0	Coronary
Danubio	Minvasys	n-Butyryl tri-n-hexyl citrate	2.5	Coronary
Restore	Cardionovum	Shellac	3.0	Coronary
AngioSculptX	Spectranetics	Nordihydroguaiaretic acid	3.0	Coronary
Essential	iVascular	Organic ester	3.0	Coronary
ElutaX	Aachen Resonance	Polymer-free	2.2	Coronary/peripheral
Ranger	Boston Scientific	Acetyl tributyl citrate	2.0	Peripheral
Lutonix	BD Interventional	Polysorbate and sorbitol	2.0	Peripheral
IN.PACT	Medtronic	Urea	3.5	Peripheral
StellarX	Philips	PEG	2.0	Peripheral
Chocolate Touch	QT Vascular	*	3.0	Peripheral
Dior I/II/Biostream	Eurocor Biosensors	Shellac	3.0	Peripheral
Pantera Lux	Biotronik	n-butyryl tri-n-hexyl citrate	3.0	Peripheral
Sirolimus
Selution	MedAlliance	Cell-adherence technology	*	Peripheral
Virtue	Caliber Therapeutics	Biodegradable polyester	*	Peripheral
Magic Touch	Concept Medical	Phospholipid	*	Peripheral
Sequent Please Neo/SCB	Braun	Iopromide	4.0	Peripheral

^*Data unavailable.

PEG = polyethylene glycol.

DRUG-ELUTING STENTS

The basic components of DESs include a mechanical mesh structure enforcing long-term patency, and drug bound to a coating controlling drug release into the vascular wall. To date, a wide variety of DES devices exist; however, development has more often than not been dictated by coating modifications, controlling pharmaceutical release and long-term elution. It is therefore worth reviewing mechanisms of controlled released and coating technology before detailing specific designs.

MECHANISMS OF CONTROLLED RELEASE.

Controlled release falls under 2 main classes: physical and chemical (38,39). The former includes diffusion of drug through a durable polymer coating, release of drug through a dissolving or eroding coating matrix, drug-specific binding by means of ion exchange or immobilized tissue antibodies, or utilizing osmotic pressure differences to drive drug into the vascular wall (39). Physical control mechanisms have become the mainstay of DESs, where manufacturing aspects such as polymer formulation, coating thickness, or exposed stent surfaces are directly linked to predictable release rates. Chemical release control relies on the splitting of bonds that bind the drug to its carrier coating to release the drug. Because this requires chemical modification of the drug to bind to its carrier (further resulting in the release of drug in so-called prodrug form [40]), the technique has not gained any prominence in the DES field.

COATINGS AND EXCIPIENTS FOR CONTROLLED RELEASE.

Variations in coating and excipient formulations have postulated roles in the resulting tissue response (41), and excellent reviews of durable, biodegradable, or deployable coating technologies have been published elsewhere (39,42). Durable coatings are the most common and rely on diffusion of drug through a monolithic or multilayered coating. Although they allow for high-degree of control, issues have been raised regarding hypersensitivity to the long-term residence of foreign polymeric material (41). Although recent durable amphiphilic polymers, silicon carbide, or even antibody-based coatings have been touted as suppressing detrimental inflammatory response (42), significant efforts have been put into developing biodegradable copolymers such as poly-Llactic acid or poly-L-glutamic acid (PLGA), where the coating is resorbed over the course of implantation. The optimal balance between degradation and release, and their impact on therapeutic versus toxic effects, remains to be clarified and is still an active area of research.

The last class of deployable coatings is more exploratory and clinical data are scarcer. Here, an absorbable coating is programmed to detach and spread into the neointimal through tissue remodeling (used, e.g., in the MiStent DES with PLGA [43]), and as neointimal growth encapsulates more and more of the coating, the area of drug delivery increases dynamically and sustained effect is envisioned. Specific release kinetics and transluminal retention patterns of these devices, however, remain to be clarified. Still, deployable coatings with delivery of drug in microcrystalline form are directly translatable to the example of DCBs, where complete delivery is achieved during acute inflation.

Worth noting with respect to drug-binding coatings is the need for dual antiplatelet therapy (DAPT), a pharmaceutical strategy currently part of recommended guidelines following DES implantation, but for which clear contraindicated patients groups exist. With the use of absorbable coatings, a shorter period of DAPT could be envisioned, and exploratory polymer-free stents have been proposed (44), circumventing the need for DAPT.

FIRST-GENERATION DES.

In early first-generation DESs, relatively thick (>100 mm) metallic struts were covered by a durable polymer coating, releasing drug in a diffusion-controlled manner (with drug traveling into the vascular wall along local concentration gradients). The sirolimus-based Cypher and the paclitaxel-based Taxus DESs were among the first such devices to gain U.S. Food and Drug Administration (FDA) approval, showing promising clinical results with reduced rates of in-stent restenosis (45,46). Notably, Taxus and Cypher used different types of coating morphology, Taxus being covered by a single layer of drug-embedded polymer and Cypher using a bilayered design with drug-embedded polymer covered by a diffusion-limiting top layer (38). As such, Cypher in theory allowed for an additional level of release control, even though practical challenges in manufacturing obstructed some of this control (47). In fact, even the single-layer design of the Taxus stent allows for a fairly high-degree of diffusion control: as demonstrated by Acharaya and Park (38), release rates over a 10-day period can be modified by up to a factor of 6 by simply modifying the polymer content. Although successful in early-phase trials against bare-metal stent equivalents, first-generation DESs became associated with an increased risk of late stent thrombosis (48). Some postulated a direct toxic effect of drug on endothelium (49) and others the combined effects of device length and strut thickness (50). Combined results forced continued evolution of device design, specifically targeting strut size to reduce vascular injury, as well as more potent drug delivery techniques and adjunctive DAPT.

SECOND-GENERATION AND CONTEMPORARY DES.

Second-generation DESs used materials with greater radial strength, allowing for thinner designs. The zotarolimus-eluting Endeavour, or the everolimus-eluting Xience V were both made of cobaltchromium with a strut thickness of w80 to 90 mm; some 30% thinner than older devices. This decrease in thickness could also be directly associated with increased reendothelialization and accelerated vascular healing (51), as well as improved outcome in clinical trials (52,53).

Second-generation DESs also represented a modification in drug formulation and coating morphology, allowing for improved release control and sustained effect (15,39). With results from first-generation DESs favoring sirolimus over paclitaxel (54,55), a range of other sirolimus analogues were introduced. Everolimus and zotarolimus (a highly lipophilic immunosuppressant with shorter tissue half-life [56]) both indicated improved re-endothelialization (51,57). The clinical trials of novolimus (inhibiting smooth muscle cell proliferation [58]), umirolimus (exhibiting extreme levels of hydrophobicity [59]), and ridaforolimus (with high therapeutic-to-toxicity margin [60]) are all expected in the near future.

Coating manipulation adds further regulation of delivered drug. Biodegradable coatings offer control by adding erosiveness to diffusive-mediated release (39) but may also produce acidic byproducts during degradation (61). In the case of deployable coatings, the coating itself is designed to spread and create encapsulated drug depots inside the vascular tissue, ensuring prolonged delivery and homogeneous spread (62). Contemporary polymer-free techniques have been suggested as effective in patients with high bleeding risk following decreased thrombogenicity, where drug is released directly from the metallic DES backbone (44). Owing to the absence of any diffusion-controlling polymer embedding, polymer-free DESs exhibit a significant release of up to 90% of the total drug load only during the first few days of implantation (39,44).

How vascular anatomy and lesion morphology affects performance and uptake from newer-generation DESs remains scarcely studied. Loh et al. (63) compared the clinical outcome of firstversus second-generation DESs in complex coronary lesions (including American College of Cardiology (ACC)/ American Heart Association (AHA) lesion class type C, or thrombotic presence), indicating slightly reduced rates of major adverse cardiac events and significantly reduced numbers of thrombosis in the second-generation arm. Stefanini et al. (64) evaluated both zotarolimus and everolimus-eluting DESs as a function of lesion complexity, showing similar outcomes between the two, although without quantifying lesion morphology. Regarding intraplaque composition, lipid-rich plaques have been attributed as an independent risk factor for nontarget ischemic events and in-stent remodeling (23) after DES implantation, indicating how a lipid-rich environment could cause suboptimal uptake and outcome. New-generation DESs promise suppressed inflammatory response and faster reendothelialization, which could be beneficial in inert calcific or necrotic lesions. More prominently, plaque calcification reduces interventional success after DES implantation (66), and minor clinical evaluations have coupled its presence to mechanical underexpansion (36), malapposition (67), and fracture (68).

BIORESORBABLE SCAFFOLDS

In bioresorbable scaffolds, or bioresorbable vascular scaffolds (BVSs), the device itself is made up of a drug-loaded bioresorbable polymer, dissolving completely over time and theoretically allowing the vessel to return to its native state after intervention. From a drug delivery standpoint, the release mechanism consists of a combination of diffusion and dissolution control, being similar to biodegradable coatings used in DESs. Pharmacokinetic studies of BVS release rates also showed similar release behavior between BVSs and analogous DESs (69), but challenges remained in balancing the rate of resorption with desired rate of drug release. Instead, the benefits of BVS were envisaged on the mechanical side, where the absence of a metallic implant would allow the vessel to restore native vasomotive, hemodynamic, and constitutive behavior.

However, early BVSs were compromised by the need for thicker struts and increased surface area to maintain mechanical stability, along with potential deleterious effects of significant local inflammation associated with active mass erosion. Initial clinical experiences were also disappointing, including premature loss of mechanical support, device shrinkage, fracture, and increased rates of in-stent restenosis (70). With delayed healing, reduced reendothelialization, and increased rates of in-stent restenosis and thrombosis (71), the first FDA-approved BVS—the everolimus-eluting ABSORB stent—was taken off the market, followed by many of its contemporaries. Revised second- and third-generation BVSs exist, designed with thinner struts, novel magnesiumbased scaffolds, or with recommended sizing and pre-procedural evaluations (72,73). Nevertheless, to date, BVSs occupy only a minuscule part of clinically utilized devices, and though mentioned in revascularization guidelines, their use is primarily limited to controlled clinical studies (74) evaluating iterated improved designs.

Having been launched and having received considerable setbacks over the course of only a decade, no extensive studies have been performed relating BVSs to lesion morphology. However, with drug formulation and polymer base similar to that of second-generation biodegradable-coated DESs, BVS behavior could be likened to that of those stents: inhibited delivery in the presence of lipid-rich or thrombotic environments, and particularly decreased efficacy in calcified lesions. Calcific environments poses a specific problem for BVSs, where issues relating to mechanical stability are enhanced in the presence of a rigid calcified lesion (75). Intriguing signs of neoatherosclerosis and calcium formation around BVS struts were indicated in part of the Absorb BVS trial, but larger inferences were not possible from those preliminary datasets (76).

DRUG-COATED BALLOONS

As an alternative to DESs, local drug delivery is also possible by DCB. Here, a drug-coated angioplasty balloon is used to reopen a stenosed vessel, exclusively delivering drug during inflation, and leaving the vessel without any remaining implant. In theory, a range of advantages can be associated with DCBs, including broader surface area with higher degree of homogeneous drug-to-tissue transfer, and avoidance of delayed arterial healing owing to the absence of implanted stent struts (77). Similarly, DCBs offer a drug delivery alternative where DESs are deemed to be unfeasible, in narrow peripheral vessels with a high degree of mechanical flexure, where long-term usage of DAPT is contraindicated, or in the setting of in-stent restenosis (note, however, that DCBs have yet to receive FDA approval for treatment of coronary in-stent restenosis). Yet, DCBs come with a set of challenges: delivery being restricted to the time of inflation, tissue-retained dose and residence time more difficult to control, and requiring high initial loading (78). Clinical study results also remain elusive regarding benefits of DCBs versus DESs in different settings (79,80), and currently DCBs have FDA approval only for use in peripheral vessels.

While DESs offer long-term control from permanent stent struts, DCBs require complete transfer during balloon inflation. As such, delivery is not primarily diffusion or dissolution driven, but rather achieved by mechanically forcing drug or drugcarrying coating into the vascular wall in the acute phase (81). That mechanical contact forces are imperative for acute DCB delivery has been experimentally shown: Stolzenburg et al. (29) observed increasing acute drug transfer with higher DCB inflation pressure, and Chang et al. (82) described similar phenomena ex vivo with the use of 2 different coatings. The need for mechanical adhesion is also limiting: because <10% of the DCB coating is transferred into the wall during routine inflation (83), up to 90% of the administered drug dose gets lost into the systemic circulation. Amplified vascular injury (as induced by excessive balloon inflation or by pre-procedural vessel preparation) may improve the retention and efficacy of DCBs (84), but the fact remains that DCBs are associated with low acute transfer efficiency. Once transferred, however, transluminal distribution is still governed by diffusion kinetics and tissue-specific binding, just as outlined for DES.

To date, a range of DCBs exist on the clinical market. For their lipophilic characteristics and beneficial protein binding, devices have predominantly used formulations of paclitaxel (77), but zotarolimus (85), and sirolimus (17) DCBs also exist, where the drug is coated in a crystalline-bound form. Decisive results showing beneficial efficacy and safety of limus-based over paclitaxel-based DCBs, however, remain to be reported. Promising clinical trial results have been achieved using DCBs (77,86), but large-scale outcome data remain elusive, again possibly owing to a lack of appreciation of lesion morphology. Noninferior DCB behavior has been inferred in a peripheral setting (77), whereas recent data on coronary in-stent restenosis—a setting with vastly different lesion morphology compared with the periphery—indicate everolimus-eluting DESs as superior to DCBs (80). Worrying indications of increased mortality were recently reported for paclitaxel-coated devices (primarily DCBs) used to treat peripheral arteries (87), but contradictory results have since been published (88). Nevertheless, the DCB field is in rapid development. With recent DES data favoring sirolimus analogue–based formulations (14,15), the translation of that into a DCB setting was initially obstructed by low acute absorption and elution owing to the drug’s relatively low lipophilicity. Contemporary sirolimus DCBs mainly use polymer-encapsulated coatings, where agglomerates of embedded sirolimus are acutely transferred into the tissue. Once transferred, these embedded nanosized carriers elute drug into the tissue over time, acting like local drug depots and creating a scenario of controlled sustained release (17,18). Though long-term follow-up data have yet to be published, registry-based outcomes show promising results in a coronary setting (89).

Just as with stents, DCB performance is affected by vascular environment. Numerous studies have identified calcium as a hindrance for optimal DCB performance (35,90). Specifically, with DCBs relying on acute mechanical deposition, stiff superficial calcium will be particularly detrimental for acute delivery and intraprocedural lesion preparation is often recommended for DCBs (84). Once coating is transferred into the wall, retention after DCBs will, just as for DESs, be affected by lesion environment: the lipidic displacement of drug-specific binding sites obstructing drug-specific binding, and lesions with an active remodeling core exhibiting similar obstructive behavior. Tada et al. (91) evaluated the relationships between lesion morphology and DCB outcomes in treating in-stent restenosis. There, the treatment of lesions with heterogeneous structures did not differ from noncoated equivalents, again highlighting how acute DCB delivery (being reliant on mechanical penetration) is possibly obstructed under such complicating conditions. Lastly, early angioplasty studies associated intraplaque thrombus with increased risk (92), and similar concerns could be raised for DCBs. Likewise, even though acute delivery is modulated by inflation pressure, the compression of tissue during balloon expansion might hinder optimal long-term delivery: tightly packed vascular layers will reduce diffusivity, possibly decreasing transluminal retention.

FUTURE DIRECTIONS: TRANSITION TOWARD LESION-SPECIFIC DRUG DELIVERY

As evident from the above, a wide variety of devices exist, using local drug delivery to tackle vascular lesion development. Controlled delivery is achieved through a multitude of ways: using diffusion-controlling coatings, exploiting biodegradable drug-loaded devices, or embedding agglomerates of drug directly into the vascular wall. Contemporary devices promise refined control mechanisms, and ideal properties would include controlled delivery over the entire lifetime of the device, a predictable endpoint to the delivery, optimal stimulation of device-covering re-endothelialization, and homogeneous transluminal deposition of drug. In all instances, refined interventional outcome can only be achieved by appreciating the inseparable relationship between delivery mode, device specifics, and vascular biology. In particular, lesion morphology is a deciding factor that dictates retention and response to local drug delivery. Consequently, it should also be viewed as dictating future strategies for interventional success. Before large-scale implementation, any clinical device has to be reviewed in a randomized control trial (RCT) setting, evaluating performance either against the absence of the device or against clinical routine intervention. Deriving mechanistic understanding from RCTs, however, is cumbersome, and observational data seldom capture the complex vascular response imposed by local drug delivery devices. Conflicting RCT outcomes are frequently reported. First-generation DESs versus bare-metal stents showed interchanging superiority (93), whereas second-generation DESs seem to demonstrate favorable outcomes (94). Findings also vary with drug formulation, and a recent meta-analysis of RCTs indicated increased risk of death with paclitaxel-coated devices in the periphery (87), albeit more recent assessments not inferring the same (88). With this review highlighting how preclinical results show a direct relationship between retention and intraplaque composition (19,24,28), we think that some of these conflicting results could be better understood and potentially rectified if lesion morphology had been taken into account, providing a more direct appreciation of how device-vascular interplay dictates interventional outcome. Further-more, we think that above all, the reviewed material highlights and suggest a needed transition toward lesion-specific intervention, where plaque morphology and distribution dictate the choice of technique and delivery (Tables 5 and 6, Figure 2).

TABLE 5.

Characteristic Plaque Morphologies, Associated Interventional Obstacles, and Potential Implications for a Lesion-Specific Approach

Plaque Characteristic	Obstacle	Potential Implications for a Lesion-Specific Approach*
Lipid pool	Displacement of nonspecific and specific drug-binding sites causing decreased drug retention.	Sirolimus-based delivery, showing less sensitivity to lipid entities. No rigidity barrier for DCB or DES.
Calcium
Superficial	Permeation barriers and inert binding nature of calcium decreasing drug retention. Superficial calcium blocking medial and adventitial binding. Mechanical rigidity at luminal wall hinders acute coating transfer (DCB) or appropriate stent apposition (DES).	Vessel preparation to either remove or crack superficial nodules (exposing deeper tissue and exposing binding sites). Sirolimus-based delivery once treated. DES possibly more rigid for long-term delivery.*
Deep	Permeation barrier and inert binding nature decreasing drug retention. Deep calcium blocks adventitial binding and might make vessel preparation less effective.	Vessel preparation for deep modification, while still opting for paclitaxel delivery (binding predominantly subintimal). DCB an option with luminal surface still giving way for acute coating transfer. DES also possible to ensure long-term delivery.*
Fibrous cap
Thin	Lesion instability following from stress concentrations in the thin cap. Increased rupture risk at mechanical stress.	Instability might require implanted rigidity, favoring DES. Unclear preference regarding paclitaxel (predominantly subintimal) vs. sirolimus-based (homogeneous distribution) delivery.
Thick	Surface rigidity hindering acute coating transfer (DCB) or appropriate stent apposition (DES).	Mechanical rigidity obstructing acute coating transfer (DCB), instead favoring DES. Unclear preference regarding paclitaxel (predominantly subintimal) vs. sirolimus-based (homogenous distribution) delivery.
Hemorrhage	Hemorrhage and neovascularization signs of active remodeling and lesion instability. Increased permeation and local drug uptake can lead to concentration extrema in hemorrhagic regions.	If superficial, instability might require DES rigidity. If deep, DCB a viable option leaving no implant behind. Possible risk of concentration extrema favors cytostatic sirolimus over cytotoxic paclitaxel.
Thrombus	Lesion instability, and thrombus entity causing permeation blockage of penetrating drug.	If superficial, instability might require DES rigidity. If deep, DCB a viable option leaving no implant behind.

Note that these are posed as future perspectives and have to be validated in randomized controlled trials.

^*Requires further validation in clinical trial.

DCB = drug-containing balloon; DES = drug-eluting stent.

TABLE 6.

Plaque Guideline Classifications and Their Implications for a Lesion-Specific Approach

Type Class	Characteristics	Obstacle	Potential Implications for a Lesion-Specific Approach*
Modified American Heart Association classification
Type I–II	Near-normal wall thickness, no calcification.	None (rarely treated by interventional procedure). Only hindrance might be minor displacement of drug-specific binding by initial foam cell formation.	Predominantly therapeutic. If interventional, sirolimus-based delivery by DCB, leaving no implant behind.
Type III	Diffuse intimal thickening or small eccentric plaque, no calcification.	Lipid entities displacing drug-specific binding. Pronounced eccentricity might obstruct stent apposition.	Sirolimus-based delivery less affected by lesion complexity. Soft plaque nature could favor DCB delivery (leaving no implant behind); however, in the presence of excessive eccentricity DES stability could be preferred.*
Type IV–V	Lipid necrotic core covered by fibrotic tissue, including possible calcifications.	Lipid entities displacing drug-specific binding. Thick fibrous coverage creates surface rigidity hindering effective acute coating transfer (DCB). Thin fibrous coverage creates plaque instability and increased rupture risk. Presence of calcium creates permeation barriers.	Sirolimus-based delivery less affected by lesion complexity. Thick fibrous coverage gives preference to DES over DCB. Thin fibrous coverage also favors DES, creating mechanical rigidity to tackle plaque instability. In the presence of surface calcium, vessel preparation could be advised,* followed by DES (if heterogeneous deep tissue) or DCB (if homogeneous soft deep tissue). In the presence of deep calcium, vessel preparation followed by paclitaxel delivery (binding subintimally) could be preferred.*
Type VI	Complex phenotype with possible surface defects, hemorrhage, or thrombus presence.	Lipid entities displacing drug-specific binding. Hemorrhage, thrombus, and surface defects create plaque instability and increased rupture risk.	Unstable plaque nature might favor rigidity of DES. Subintimal binding of paclitaxel might be favored to dampen surface remodeling; however, sirolimus-based delivery less affected by lesion complexity. Further data are needed.*
Type VII	Moderate to heavy calcifications.	Calcium represents inert calcium barrier. Heavy surface calcium increases surface rigidity, obstructing acute coating transfer (DCB).	Vessel preparation.* If deep calcium, paclitaxel preferred for its subintimal affinity. If surface calcium, vessel preparation might be followed by sirolimus-delivery.
Type VIII	Fibrotic without lipid core, including possible minor calcifications.	Thick fibrous coverage increases surface rigidity, obstructing acute coating transfer (DCB).	DES preferred over DCB. Sirolimus vs. paclitaxel still undecided.*
American College of Cardiology/American Heart Association classification of coronary lesions
Type A	Discrete, concentric, nonangulated, little to no calcium, without thrombus or hemorrhage.	Only minor, with the phenotype rarely treated by interventional procedure. Only hindrance might be minor displacement of drug-specific binding by initial foam cell formation, and— In the presence of excessively thick fibrous coverage in the presence of a thick fibrous coverage—obstruction of acute coating transfer (DCB).	Predominantly therapeutic. If interventional, sirolimus-based delivery by DCB, leaving no implant behind. DES could be preferred.*
Type B1-B2 (B1 having 1 of Eccentric, moderate the characteristics, B2 having 2 or more)	tortuosity, irregular, moderate-to-heavy calcifications, thrombus presence.	Several, varying as a function of characteristics. Lipid entities displacing drug-specific binding. Thick fibrous coverage creates surface rigidity hindering effective acute coating transfer (DCB). Thin fibrous coverage creates plaque instability and increased rupture risk, as does thrombus presence. Presence of calcium creates permeation barriers. Pronounced eccentricity might obstruct stent apposition.	ACC/AHA classification not sufficient to define one strategy for the B1–B2 class. Instead, depending on morphology, different strategies will prevail. In complex soft lipidic plaques, sirolimus-based delivery is preferred. Thick fibrous coverage gives preference to DES over DCB. Thin fibrous coverage also favors DES, creating mechanical rigidity to tackle plaque instability. In the presence of calcium, vessel preparation could be advised,* followed by DES (if heterogeneous deep tissue) or DCB (if homogeneous soft deep tissue).
Type C	Diffuse, excessive tortuosity, angulation of >90°, surface defects.	Surface defects associated with plaque instability. Angulation and tortuosity detrimental for stent apposition and underexpansion.	ACC/AHA classification not sufficient to define one strategy for C class. Instead, depending on morphology different strategies will prevail, as above (type B1–B2 recommendations). In the presence of heavy angulations and surface defects, vessel preparation could homogenize response,* followed by either DES or DCB.*

Plaque morphologies characterized by the modified AHA score and the ACC/AHA classification score, with each coupled to specific interventional obstacles, and potential implications for a lesion-specific approach (note that these are posed as future perspectives and have to be validated in randomized controlled trials). Note that the modified AHA score takes lesion morphology into account, whereas the ACC/AHA classification score is less optimal and cannot sufficiently distinguish between morphologies with different lesion-specific treatment strategies. However, it is included following common clinical usage.

^*Requires further validation in clinical trial.

ACC = American College of Cardiology; AHA = American Heart Association; other abbreviations as in Table 5

FIGURE 2. Plaque Morphologies and Related Lesion-Specific Approaches.

Characteristic plaque morphologies defined according to the modified American Heart Association (AHA) classification, associated interventional obstacles relating to drug delivery, and potential implications of lesion-specific approach (note that these are posed as future perspectives and would have to be validated in randomized controlled trials). DCB = drug-containing balloon; DES = drug-eluting stent.

CLINICAL IMPLICATIONS FOR LESION-SPECIFIC INTERVENTION.

A possible clinical shift toward lesion-specific intervention is not achieved by the introduction of a single device or by modifying a single interventional aspect. Instead, concomitant modification of lesion characterization methods, device choice, and drug formulation are envisioned, along with critical and purposely designed clinical trials to validate the role of lesion-specific characterization in clinical practice. Only by doing so, and only by including lesion characteristics as a demographic entity in clinical trials, can our understanding of the interaction between atherosclerotic biology and drug retention be truly incorporated into our evaluation of acquired clinical data and aid in a possible transition toward lesion-specific treatment.

Lesion-specific c haracterization: the role of me dica l ima ging.

Central to the transition toward lesion-specific intervention is the characterization of lesion-specific aspects. Here, medical imaging plays a key role, where advances in intravascular imaging allow for unprecedented insights into in vivo plaque morphology. Virtual histology by intravascular ultrasound (VH-IVUS) utilizes spectral signal decomposition to classify tissue into fibrotic, fibro-fatty, calcific, and necrotic core classes (95), and has been successfully validated in coronary lesions (96). VH-IVUS has also been used in the study of coronary DESs (97), highlighting relationships between lesion phenotype and hyperplastic development (35,98). The inclusion of near-infrared spectroscopy in existing IVUS devices also promises effective lipid detection (99) in addition to the aforementioned VH-IVUS utility. Optical coherence tomography is another intravascular modality, offering mm-resolution through optical wave scattering (100). The modality is increasingly used in conjunction with vascular drug delivery, and has been particularly successful in relating calcific developments to obstructed delivery (36,91). Virtual histology–like features have also recently been added to optical coherence tomography (101). Developing similar micromorphology characterization from noninvasive biplanar or 3-dimensional angiography would be desired as that modality still dominates the field, but as yet, merely qualitative scores focusing on calcium burden exist in clinical usage (102).

In all cases, imaging offers the ability to quantify lesion morphology, and would as such play a key role in an envisioned lesion-specific era. Numerous examples also exist where imaging has been effectively incorporated into clinical practice to identify vulnerable plaques (37,99,103) or to quantify both acute (stent apposition, dissection, and so on) and long-term treatment efficacy (restenosis growth, plaque remodeling, and so on) (103,104). A recent meta-analysis even highlighted improved patient survival when incorporating intravascular imaging and assessment of regional lesion state before DES implantation (105). For the sake of a lesion-specific transition, quantification of lesion morphology—by means of medical imaging—could also help steer the choice of device and drug for interventional treatment: lipid-rich plaques implying possible sirolimus-based delivery following increased displacement of tissue-specific paclitaxel binding, hemorrhagic or thin fibrotic structures implying structural instability and the need for rigid stent-based delivery, and so forth (detailed explanations on this provided in the subsequent section and in Tables 5 and 6). Still, to conclusively determine the added value of morphology-resolving imaging, it is paramount that image-derived lesion morphology is included as a studied entity in clinical evaluations of drug delivery devices, monitoring, for example, volumetric content, anatomic positioning (superficial vs. deep), and morphologic regionality (local vs. circumferential) as a function of clinical outcome.

Along the same lines, image-based lesion characterization could also aid in selecting appropriate lesion preparation strategies, possibly improving subsequent retention of delivered drug. As highlighted in recent review work (37,106), image-based detection of highly calcific structures could merit pre-procedural removal or modification of such permeation-limiting calcium, and, similarly, image-based identification of lipid-rich areas could provide insights into mechanical stability, displacement of imperative tissue-binding sites, or even uncover actual lesion length in complicated phenotypes (107). In general, lesion preparation could aid in homogenizing vascular response and drug retention (by means of changing morphologic structure), and incorporating imaging here is central to uncovering definite vascular mechanisms steering such behavior. Again, large-scale clinical data remain to be collected regarding the effect of lesion preparation on adjunctive drug-eluting technologies, although ongoing trials such as Disrupt CAD/PAD I–III (Shockwave Coronary Rx Lithoplasty Study/Safety and Performance Study of the Shockwave Lithoplasty System) and ECLIPSE (Evaluation of Treatment Strategies for Severe Calcific Coronary Arteries) all promise such data in the near future. Likewise, we again think it is paramount to properly acknowledge and include imaging and micromorphologic quantification in future clinical trial designs to effectively evaluate performance of future drug delivery devices.

Lesion-specific intervention: choice of drug and device.

As noted above, with imaging providing insight into morphologic state, lesion-specific aspects could be incorporated into the choice of intervention. First, an area where lesion-specific intervention is of relevance is the choice of drug, a topic of constant debate in the interventional field. As reviewed by Levin et al. (12), transluminal retention rely on drug-specific binding sites; mammalian target of rapamycin receptors (for sirolimus-based drugs) are more homogeneously distributed through the wall, whereas tubulin (for paclitaxel) exhibit predominantly subintimal and adventitial residence. Device-induced intimal disruption, or atherosclerotic displacement of binding sites, might thus be more detrimental to paclitaxel-based delivery. Preclinical work has indeed shown how sirolimus binding is less sensitive to lesion complexity, and how paclitaxel-based partition coefficients decline more rapidly with increasing lipid content (19). As such, complex lesions with an active core and definite lipidic entities could benefit from sirolimus-based delivery, and long-term retention could be further bolstered by giving priority to sirolimus formulations with high partition coefficients, as represented by the more hydrophobic zotarolimus or everolimus versus sirolimus. The only area where paclitaxel delivery could still be preferred is in instances of high medial-to-deep calcification, where such deep permeation barriers shift binding to subintimal spaces, and where—assuming no major subintimal disruption from the implanted or inflated device—the higher partition coefficient of paclitaxel could result in a more efficient antiproliferative effect (in contemporary practice, paclitaxel delivery is also seemingly favored in calcified peripheral lesions [108]). However, in such instances, lesion preparation could be a more viable alternative to tackle to calcific environment before intervention, and experimental and clinical validation of paclitaxel versus sirolimus in such settings remains to be performed. For all of the above, a lesion-specific shift would also have to be coupled to conclusive clinical evidence, guided by purposely designed clinical trials.

In addition to the impact that lesion complexity has on the choice of drug formulation, apparent lesion morphology could also steer the choice of interventional device. In particular, the difference in delivery mode between DCBs (short-term mechanical deposition during inflation) and DESs/BVSs (long-term diffusion/convection from implanted struts) is important to contrast with lesion complexity. For DCBs, acute delivery is influenced by the physical forcing of coating particles into the vascular wall, meaning that delivery into soft, hematoma-rich, fatty lesions (e.g., in-stent restenosis tissue) would be facilitated, whereas delivery into fibro-calcific or calcific lesions would by the same logic be obstructed. The choice of coating micromorphology might help to overcome more rigid intimal entities (microneedle configurations inducing higher contact pressure and consequently higher degrees of coating transfer, compared with amorphous coating equivalents [81]); however, in highly calcific lesions, intervention should instead be stent based and possibly include appropriate vessel preparation. For the latter, atherectomy has been proposed to reduce superficial calcium by means of physical grinding (109), and for deep calcium, contemporary lithotripsy-based techniques promise exciting refinements. As such, divergent results spurring from heavily heterogeneous lesion phenotypes could possibly by rectified by means of vessel preparation; however, definitive large-scale clinical outcome results remain to be established in this arena.

With the heterogeneity of atherosclerotic lesions in mind, we envision lesion-specific intervention as a shift toward lesion-tailored vessel preparation techniques and lesion-guided delivery devices rather than continuing toward a one-size-fits-all scenario. We see interventional sessions guided first and foremost by lesion characterization—effectively quantified by modern imaging techniques—with subsequent interventional choice steered thereafter (Table 5 presents a summary relating to lesion-specific treatment, and Table 6 provides similar overviews summarized as a function of both the modified AHA score [2], and the more commonly used ACC/AHA score [being less relevant for morphologic classification, but included following clinical usage]). To review: softer lesions with high lipid content would possibly benefit from sirolimus-based DCB delivery; heavily calcified lesions from vessel preparation and more rigid DES treatment; and actively remodeling phenotypes with surface defect and erosion-prone nature from less impactful intimal disruption, possibly achieved by gentler stent-based expansion versus the mechanically forced coating transfer necessitated to achieve effective DCB delivery.

Lesion-specific mechanisms: enlightened understanding by computational and nonclinical advances.

Finally, transition toward lesion-specific drug delivery intervention will be continuously updated and refined be increasingly powerful computational, nonclinical, and scientifically based evaluations. In the computational domain, the inclusion of high-fidelity imaging enables patient-specific modeling, thus allowing for predictive assessment of interventional success (50), and the simulated environment is particularly effective in parametrically evaluating aspects governing hemodynamic, structural, and drug-retentive response. The incorporation of pharmacokinetic and pharmacodynamic descriptions of drug uptake has enabled elevated understanding of drug delivery (110), and recent developments also attempt to include lesion-specific components in the simulated assessments (111), illustrating how computational modeling will serve as a continuing effective tool in establishing lesion-specific guidelines.

To a similar extent, nonhuman animal modeling will also be central in establishing the frameworks of lesion-specific intervention. As with computational modeling, animal models allow for a controlled study environment. However, in this instance, isolated evaluation can be performed while still keeping the full complexity of the cardiovascular system in place. As such, animal modeling has been central to the development of local drug delivery devices, and lesion-specific aspects have been included in this nonclinical arena (112). Importantly, animal and computational modeling should not be viewed as preceding clinical evaluations. Instead, development of local drug delivery devices should be developed using iterative evaluations of theoretical, computational, animal, and clinical assessments.

LESION-SPECIFIC INTERVENTION AND FUTURE DRUG DELIVERY: LOOKING TOWARD THE HORIZON.

A transition toward lesion-specific intervention and improved clinical outcome is not necessarily bound to novel coating designs, new drug formulations, or updated device specifications. Instead, we postulate that performance of existing devices could very well be increased by acknowledging in which morphologic scenario they are most effectively used, and that such morphologic assessment can in fact be effectively achieved by available high-fidelity imaging techniques. In addition, some promising trends are approaching with respect to new device designs. Novel materials such as tantalum- or nitinol-based alloys promise flexible yet durable low-profile struts, promoting early re-endothelialization and reducing vascular injury (42), and would thus minimize disruptions on subintimal binding. Contemporary results favor sirolimus-based formulations (54,55): the drug performs better in complex phenotypes (19), and by using polymeric capsules within deployable coatings, sirolimus-based delivery now has specific promise in contemporary DCB use. With lesion-specific treatment in mind, vessel preparation is another important incorporation into future drug delivery procedures, where homogenization of vessel response could be achieved. Significant advancements in the field of molecular biology and genetics also come with a promise of cell-selective drug formulations (113) or even gene-eluting stents (114), but significant work remains in translating them to any clinical implementation and understanding their role in lesion-specific intervention. Nonetheless, the incorporation of any of the above techniques will have to be closely coupled to assessment of lesion morphology in future clinical and nonclinical studies, acknowledging its governing influence on local drug delivery, vascular behavior, and interventional outcome.

CONCLUSIONS

Vascular drug delivery devices have become a key component of cardiovascular medicine, with a wide variety of devices in clinical practice including DESs, BVSs, and DCBs. Different devices come with unique ways of achieving controlled and desired drug delivery, but in all instances, lesion morphology is the overruling aspect dictating retention and response. As such, improved interventional outcome can only be achieved by shifting toward lesion-specific intervention, where therapy is first and foremost guided by lesion phenotype and vascular characteristics. Contemporary image developments, predictive modeling tools, and novel delivery techniques will aid in this transition, allowing for improved and refined endovascular therapy through future lesion-specific diagnostics.

TABLE 3.

Overview of Commercially Available BVS

Drug and Device	Company	Polymer Base	Drug Load, mg/mm2	Vessel Territory
Sirolimus
MeRes	Merik	PDLLA with Pt-Ir	1.3	Coronary/peripheral
Magmaris	Biotronik	PLLA with Mg-Ta	1.4	Coronary
Everolimus
Absorb GT1*	Abbott	PLLA with Pt	1.1	Coronary
Novolimus
DESolve	Elixir	PLLA with Pt-Ir	5.0	Coronary/peripheral

^*Currently discontinued.

Abbreviations as in Table 2.

HIGHLIGHTS.

• Local drug delivery is used increasingly to prevent restenosis after intervention for obstructive atherosclerosis.

• Enthusiasm for this technology in some cases exceeds understanding of therapeutic mechanisms.

• Retention and efficacy of locally administered drugs are largely dependent on vessel anatomy and lesion morphology.

• Clinical application of local drug delivery devices may expand with the continued evolution of lesion-specific therapies.

ABBREVIATIONS AND ACRONYMS

BVS

bioresorbable vascular scaffold

DAPT

dual antiplatelet therapy

DCB

drug-coated balloon

DES

drug-eluting stent