Engineering the subcutaneous delivery of antibody-drug conjugates

Abstract

Subcutaneous (SC) delivery has evolved beyond a convenience formulation to become an operational re-architecture of cancer care, with the potential to convert infusion-centered oncology into a distributed model without compromising exposure, safety, or product quality. Monoclonal antibodies have already completed this transition, establishing that the SC route delivers clinical equivalence while improving patient experience and system efficiency. Antibody-drug conjugates (ADCs) represent the next frontier, yet they remain almost exclusively intravenous (IV). This gap reflects not an incompatibility with SC delivery but rather the specific points at which ADCs stress the delivery stack: tissue poroelasticity and backpressure, extreme concentration requirements, non-Newtonian injectability, route-dependent linker degradation, and payload-driven interfacial instability. Here we argue that these barriers are tractable design targets amenable to advanced materials and integrated engineering. In this perspective, we outline the conceptual and technical approaches required to enable subcutaneous ADC delivery. We frame the SC compartment as a biophysical reactor governed by interstitial osmotic pressure, convection-dominated transport, and immune surveillance. We then outline two enabling pillars already in motion, namely transient matrix modulation (for example, hyaluronidase-mediated hyaluronic acid depolymerisation) and high-concentration formulation science (>100 mg mL⁻¹), and show why ADCs require a fundamentally different playbook than antibodies alone. Finally, we position injectable, biodegradable, stimulus-responsive polymers and hydrogels as the enabling layer that can co-control injectability, depot protection, linker integrity, and release kinetics under SC constraints (pH ∼6.8, volume-limited compliance). SC ADC formulations are inevitable because they align clinical need, patient preference, and healthcare capacity. However, we propose that the systems most likely to succeed will be those that co-design molecule, formulation, material, process, and device as a single integrated platform.

Graphical abstract

Subcutaneous delivery of antibody-drug conjugates (ADCs) demands integrated engineering beyond conventional antibody formulation. The SC compartment imposes route-specific liabilities on linker stability, payload-driven colloidal behavior, and DAR-dependent rheology that distinguish ADCs from monoclonal antibodies. Transient matrix modulation, high-concentration formulation enabled by ionic liquids, and injectable hydrogel depots constitute complementary pillars that must be co-designed with device and conjugation chemistry to transition oncology from IV infusion toward distributed ambulatory care.

Highlights

• •Vesicant ADC payloads may preclude conventional dispersion-based SC strategies.
• •Depot confinement could offer a safer alternative to hyaluronidase co-formulation.
• •Injectable hydrogels can co-control injectability, depot protection, and release.
• •SC Cmax attenuation may reduce peak-driven ADC toxicities while preserving AUC.
• •Molecule, formulation, material, and device should be co-designed as one platform.

1. Introduction

Oncology drug development has historically privileged intravenous (IV) administration. In early-phase trials, IV delivery reduces exposure variability, limits confounding factors, and simplifies pharmacokinetic interpretation. This rationale was defensible when biologics were administered intermittently and infusion capacity was abundant. It is increasingly misaligned with current practice, in which chronic regimens, combination backbones, and expanding biotherapeutic portfolios compete for finite infusion resources [1]. Meanwhile, the subcutaneous (SC) route has matured from a niche alternative into a clinically validated and commercially strategic standard for monoclonal antibodies (mAbs). The SC formulations of trastuzumab, rituximab, and the pertuzumab-trastuzumab fixed-dose combination (Phesgo) have clarified the system-level value proposition with higher patient preference, reduced chair time, operational efficiency, and the potential to shift administration away from hospital-based infusion centres [2], [3], [4].

Antibody-drug conjugates (ADC) represent one of the most transformative drug classes in oncology, pairing antibody selectivity to small-molecule cytotoxic potency [5], [6], [7], [8], [9]. As ADC development expand, the IV-only constraint becomes both a throughput bottleneck and an access limitation [10]. The relevant question is therefore not whether ADCs will transition to SC delivery, but whether the field will continue to treat SC as a late-stage reformulation exercise or will instead recognize it as an integrated engineering challenge that must be addressed upstream, at the level of conjugation chemistry, formulation physics, manufacturing, and device co-design.

Two enabling pillars have already emerged. The first expands the effective capacity of the SC space by transiently reducing extracellular matrix (ECM) resistance to fluid flow [1], [11], [12]. The second compresses the therapeutic dose into feasible injection volumes by designing stable high-concentration formulations that can exceed 100 mg mL⁻¹ [13], [14], [15]. Neither pillar is a workaround and together they provide a foundation for advanced materials that treat the SC compartment as a controlled transport environment rather than a passive reservoir.

2. The SC compartment as a biophysical reactor

From a delivery standpoint, the tissue represents an interface where injection dynamics, formulation viscosity, and molecular stability interact to define local pressure and transport kinetics. SC delivery is governed by a distinct set of transport and reaction phenomena [1], [16]. The hypodermis is a hydrated, viscoelastic composite network composed of collagen fibrils, elastin, and a dominant hyaluronic acid (HA) component. HA is a high-molecular-weight polymer (roughly 50 kDa to >2000 kDa) that forms noncovalent hydrogels in situ [17]. At physiological concentrations (approximately 200–400 μg mL⁻¹), HA generates apparent tissue viscosities exceeding 1000 cP (mPa·s), several orders of magnitude higher than water. Collagen introduces additional structure through organized fibrillar architectures and tortuous microchannels that compartmentalize transport. The combined HA-collagen network therefore functions not simply as a diffusion barrier but as a tunable poroelastic medium whose hydraulic conductivity, compliance, and viscoelastic relaxation determine injectability and early post-injection dispersion (Fig. 1).

Fig. 1.

The subcutaneous delivery landscape for antibody-drug conjugates. Left, anatomical layers and physiological constraints relevant to SC injection, including the dense, viscoelastic extracellular matrix of the hypodermis and local immune surveillance that can shape distribution and immunogenicity. Right, enabling subcutaneous strategies. Clinically established approaches include transient matrix modulation via hyaluronidase and high-concentration formulations supported by excipients that mitigate aggregation, interfacial instability, and injectability challenges. Emerging approaches include biodegradable hydrogels and stimulus-responsive polymers that provide route-specific protection and controlled release, as well as nanoscale carriers and engineered matrices designed to stabilize ADC and reduce local payload exposure.

Three mechanistic concepts translate this biology into engineering constraints. First, injection is a coupled pressure problem. The rapid introduction of volume into a confined, compliant tissue raises interstitial pressure; by contrast, slow subcutaneous infusion (e.g., immunoglobulin SC therapy) can accommodate equivalent volumes without significant pressure elevation [16]. Injection related pain arises from a combination of mechanical distension, local shear, and fluid shifts influenced by the interstitial Donnan osmotic pressure set by polyanionic glycosaminoglycans. The practical 2–3 mL SC volume ceiling reflects this coupled mechanics rather than an arbitrary convention [1].

Second, large proteins and ADCs diffuse slowly through the ECM, and early distribution is dominated by pressure-driven convection through interstitial pathways, followed by lymphatic uptake [18], [19]. This convection transport is regulated by the local interstitial pressure field, which is itself shaped by the device-imposed flow rate (for example, an autoinjector delivery profile) and needle geometry. Faster delivery increases transient backpressure and steepens the convective pressure gradient that governs initial dispersion. Injection pressure profiles, local tissue permeability, and transient changes in hydraulic conductivity are therefore first-order determinants of bioavailability and local exposure.

Third, the SC space is biologically active. Fibroblasts remodel ECM, macrophages and dendritic cells migrate, bind and process proteins, and vascular and lymphatic networks mediate clearance and shape systemic exposure [19]. For ADCs, this biological activity is not background noise but rather a potential amplifier of local toxicity risk, as any premature payload release at the injection site occurs within a tissue compartment densely populated by immune and stromal cells [20].

Importantly, SC tissue is mildly acidic relative to plasma (approximately pH 6.8–7.2 versus 7.4), and contains active local proteases and higher reducing thiol availability (for example, glutathione and cysteine) [1]. This represents precisely the chemical microenvironment in which linker stability, conjugation chemistry, and payload hydrophobicity become route-dependent variables [7], [21].

3. Expanding SC capacity through transient matrix modulation

The most clinically validated approach to enlarge SC capacity is enzymatic remodeling of HA networks. Recombinant human hyaluronidase (rHuPH20, and next-generation variants) cleaves β-1,4 glycosidic bonds in HA, shortening polymer chain lengths, decreasing matrix resistance, and increasing hydraulic conductivity [11], [12]. Functionally, the tissue becomes transiently more permissive to convection, allowing injection volumes exceeding 10 mL for antibodies. rHuPH20 has reported optimal activity near pH 5.0 and Michaelis-Menten behavior characterized by Kₘ ≈ 0.9 mg mL⁻¹ for high-molecular-weight substrates [12]. Peak enzymatic activity occurs during injection and immediately thereafter, with a local tissue half-life of approximately 30 min, creating a defined window of reduced resistance. As the enzyme is cleared and the substrate becomes depleted, HA is restored via endogenous synthesis, typically within 24–48 h, restoring tissue integrity and enabling repeated dosing without cumulative structural damage [11].

For ADC, hyaluronidase provides more than increased volume tolerance: it changes the balance between dose compression and tissue expansion. Larger permissible volumes reduce the required formulation concentration for a given dose, easing molecular crowding and rheology constraints [14]. However, even with expanded volume windows (for example, <15 mL), many ADC regimens still require high-concentrations formulation to achieve the required dosage [15]. Tissue modulation therefore does not eliminate the need for high-concentration science but rather reframes it.

A key underleveraged insight is that rHuPH20 operates in a microenvironment that is not optimized for its catalytic preference. SC tissue pH is typically 6.8–7.2, making catalytic enzymatic efficiency inherently submaximal at the injection site [1], [12]. This pH mismatch represents an engineering opportunity rather than a fixed limitation. Nevertheless, at current dosing levels (<200 μg for a 10 mL injection volume), it achieves sufficient HA depolymerization to enable high-volume SC delivery. This apparent disadvantage is further mitigated by the co-injected formulation buffer, typically adjusted to pH 5.5–6.5 to meet antibody stability requirements, which transiently shifts the local microenvironment toward more favorable enzymatic conditions. For instance, enzyme variants engineered for shifted pH-activity profiles, enhanced thermal stability, or reduced immunogenicity could extend effective modulation without increased dosing. Likewise, formulation microenvironments that locally tune ionic strength or buffer capacity, while respecting SC tolerability, could increase effective activity within the safe window.

Beyond enzymes, the broader direction is clear: reversible, bounded modulation of the ECM state variables that govern hydraulic conductivity and backpressure. Biophysical approaches that combine device-controlled flow profiles, pressure shaping, and locally acting permeability modifiers could complement enzymatic strategies. The goal is not maximal disruption of the interstitium but rather reproducible control with rapid recovery.

4. High-concentration formulation is where ADCs diverge from mAbs

SC feasibility ultimately reduces to a dose-volume inequality. For many ADCs, clinically relevant doses will not fit within a 1–2 mL injection window unless the drug product is routinely formulated above 100 mg mL⁻¹ and, for some, approaches the 200 mg mL⁻¹ regime [14], [15], [22].

At these concentrations, two coupled phenomena dominate. First, molecular crowding drives non-ideal solution behavior. Close molecular proximity amplifies attractive and repulsive protein-protein interactions (PPIs), increasing the probability of reversible self-association, liquid-liquid phase separation, and irreversible aggregation [14], [23]. Second, the fluid is rarely Newtonian. High-concentration biologics often exhibit shear-thinning, yield-stress behavior, and time-dependent viscosity. Injectability is therefore defined by the full rheological profile across device-relevant shear rates rather than by a single viscosity value [22]. This is where ADCs diverge mechanistically from mAbs. Conjugation introduces hydrophobic microdomains and steric protrusions that raise effective surface free energy, reshape the interaction landscape, and steepen the concentration dependence of both viscosity and aggregation kinetics [24], [25]. A formulation that is acceptable for a mAb at 100–150 mg mL⁻¹ may become fragile for an ADC once drug-to-antibody ratio (DAR) heterogeneity, payload hydrophobicity, and linker chemistry are introduced [26], [27].

Overcoming this concentration-viscosity coupling requires strategies that go beyond conventional excipient optimization. As discussed in the advanced materials section below, thermostable ionic liquids offer a promising route to decouple dose compression from injectability by suppressing the nonspecific protein-protein interactions that drive non-Newtonian viscosity growth at ultra-high concentration [13]. Before considering such emerging materials, we examine the classical excipient toolkit and its limitations for ADC-specific formulation challenges.

Excipients as interaction modifiers rather than generic stabilizers. A rational excipient strategy for SC ADCs should be anchored in interaction energetics. Arginine and related amino-acid excipients can weaken attractive PPIs [28], [29], while hydroxypropyl-β-cyclodextrin (HP-β-CD) and related cyclodextrins can form inclusion complexes with hydrophobic regions to reduce interfacial adsorption and agitation-induced aggregation [30]. Surfactant selection is higher stakes for ADCs because payload-driven hydrophobicity amplifies surface adsorption, and surfactant degradation products can act as destabilizers [15]. Buffer choice must maintain pH near SC boundary conditions while minimizing chemical liabilities for linkers and payloads [7], [31].

Process engineering is formulation engineering. High-concentration ADC formulation cannot be separated from manufacturing. Every processing step imposes shear, interfacial exposure, and residence time, all of which directly influence aggregation kinetics [23]. Mixing, filtration, and filling become nontrivial in non-Newtonian fluids, and the formulation's rheological profile must be evaluated across the shear rates that correspond to pump heads, tubing geometries, and device-driven injection rates. Emerging viscosity-modulating strategies, including the ionic liquid approaches described later, may simplify processing by reducing yield stress and improving flow properties without requiring aggressive dilution [13]. These approaches, however, introduce new considerations for compatibility dimensions that must be controlled across container closure systems, single-use materials, and downstream analytics.

5. ADC-specific liabilities in the SC microenvironment

A common but consequential error in SC ADC development is to treat ADCs as “antibodies plus potency.” Conjugation changes both the thermodynamics and kinetics of formulation behavior in ways that diverge fundamentally from unconjugated antibodies [7], [8].

Linker stability becomes route dependent. The SC tissue imposes chemical stresses absent in plasma: mildly acidic pH, local protease activity, and reducing thiols [8]. These conditions can compromise linker integrity during SC residence, potentially leading to premature payload release and injection-site toxicity [20].

Cleavable linkers, including valine-citrulline motifs, may be susceptible to dermal cathepsins present in the SC environment [7], [31]. Premature cleavage reduces the fraction of intact ADC reaching systemic circulation and increases the probability of local payload exposure. Non-cleavable linkers avoid premature cleavage but can alter release and activation dynamics, with implications for efficacy depending on payload and mechanism [8], [32].

First-generation maleimide conjugation introduces an additional SC-specific liability. In thiol-rich environments, retro-Michael deconjugation can proceed, scrambling drug attachment sites and reducing the intact ADC fraction before systemic uptake [21], [33]. This provides a direct rationale for next-generation conjugation strategies, including engineered cysteines and bioorthogonal chemistries designed to resist thiol exchange during SC residence while preserving intracellular processing [27], [34].

Payload hydrophobicity rewires colloidal stability. Hydrophobic auristatin derivatives (MMAE, MMAF) can increase aggregation propensity as concentration rises [24], [25]. More hydrophilic payloads (for example, exatecan or SN-38 derivatives) present different risks, including premature dissociation in aqueous microenvironments. Ultra-potent, highly hydrophobic classes such as pyrrolobenzodiazepine dimers combine extreme hydrophobicity with severe consequence for any local release, necessitating sophisticated solubilization strategies such as cyclodextrin inclusion at high molar ratios [30], [35].

Across payload classes, the unifying mechanistic variable is surface free energy. Conjugation creates hydrophobic microdomains that increase interfacial adsorption and lower the energy barrier for aggregation under shear or at interfaces [24]. This explains why surfactant and interfacial protection strategies that suffice for antibodies can fail for ADCs [10], [16], [26], [36].

DAR is not only a potency variable but also a rheology variable. Clinical ADCs span a wide range of drug-to-antibody ratios (DAR), reflecting payload-specific optimization. High-DAR designs can maximize payload delivery and bystander effects, but they also increase hydrophobicity and shift colloidal phase behavior [9], [35]. In the concentration regime required for SC delivery, high-DAR species can preferentially self-associate and aggregate, potentially altering DAR distribution during storage and complicating batch consistency [24]. Site-specific conjugation and narrower DAR distributions, including ones that use a uniform DAR near 2, illustrate how controlled conjugation can facilitate high-concentration formulation while maintaining activity through improved pharmacokinetics and reduced aggregation propensity [27], [34].

Finally, efficacy depends on rate processes downstream of delivery, which includes binding, internalization, intracellular trafficking, linker cleavage, and payload activation. These processes can be framed as coupled kinetic steps, including endo-lysosomal processing and, for some mechanisms, downstream proteasomal degradation kinetics that shape antigen turnover and cell fate [8], [32]. Importantly, the material design strategy must be stratified by payload toxicity. While severe vesicants like auristatins (e.g., MMAE) and maytansinoids (e.g., DM1/DM4) demand strict physical sequestration to prevent catastrophic subcutaneous necrosis upon linker failure [20], [25], the emerging class of topoisomerase I inhibitors (e.g., DXd, SN-38) present a lower necrotic risk but still require kinetic control to mitigate local inflammation and immune recruitment. Thus, the biophysical container must be matched to the payload's specific liability profile.

6. Advanced materials as the enabling layer

While ionic liquids address the concentration bottleneck, controlling the spatiotemporal fate of the injected dose requires a complementary architectural approach. If rHuPH20 represents the archetype of first-generation extracellular matrix modulation [11], [12], advanced materials constitute the second-generation control layer. These technologies are orthogonal rather than competitive: ionic liquids primarily compress and stabilize the dose in the vial and inside the device, whereas injectable depots control post-deposition kinetics and local protection in tissue [37]. Together, they can co-engineer injectability, local residence, and release kinetics while remaining compatible with the biophysical boundary conditions of the SC compartment, including mildly acidic pH, high ionic strength, interstitial backpressure, and local protease activity.

Zwitterionic and thixotropic depots decouple injection mechanics from release kinetics. The key concept is not only reduced viscosity but independent tunability of injection viscosity and release kinetics. Shear-thinning, thixotropic hydrogels can flow through fine-gauge needles under high shear yet recover rapidly structure after deposition, thereby limiting backflow and improving depot integrity [38], [39]. The principal mechanistic handle is the network mesh size, distribution, and stress-responsive remodeling. Mesh size sets an upper bound on diffusive transport, and governs early leakage in diffusion-dominated (Fickian) regimes. Although true size-exclusion of smaller proteases while permitting intact IgG-class remains unlikely, mesh engineering can be paired with chemical defences, such as protease-resilient linker designs, local inhibitors, or zwitterionic chemistries that reduce nonspecific adsorption and enzymatic colocalization.

Injectable polymeric depots illustrate how materials design can be translated into clinically compatible handling polymer-nanoparticle (PNP) hydrogels [38], [39] and moldable dynamic covalent networks built around reversible boronate esters [40]. Both architectures are inherently shear-thinning and self-healing, enabling low-force injection through narrow-gauge needles while rapidly recovering a yield-stress network in situ. For ADCs, their relaxation spectrum and mesh size can be tuned by adjusting particle-polymer affinity, reversible bond kinetics, and polymer topology, creating a route to admit the intact conjugate while suppressing the ingress of exogenous proteases and inflammatory mediators that can accelerate linker cleavage [37], [38], [39], [40].

Thermostable ionic liquids as an orthogonal lever for ultra-high concentration delivery. Thermostable ionic liquids (ILs) provide a complementary lever that acts upstream of the depot. Rather than enlarging the permissive SC volume by enzymatic matrix degradation, ILs suppress the viscosity rise associated with dose compression. A choline-based IL candidate, CMP1:2 (choline:4-methyl-2-pentenoic acid), enabled antibody formulations up to 230 mg mL⁻¹ while maintaining viscosity below 20 mPa.s under injection-relevant shear, with preserved structure and function after 4 months at room temperature and more than five-fold higher serum absorption after SC dosing compared with saline [13]. This result is decisive for autoinjector engineering, because it shifts feasibility from a brute-force pressure limitation to a controllable pressure-flow and power-budget problem.

Mechanistically, IL ion pairs provide molecular shielding combining electrostatic and hydrophobic solvation to reshape the protein hydration shell, attenuate long-range electrostatics, and dampen hydrophobic patch contacts. The net effect is suppression of nonspecific PPIs that otherwise drive clustering, yield-stress behavior, and non-Newtonian viscosity increases at ultra-high concentration. The observed shifts in colloidal descriptors such as zeta potential, together with favorable non-Newtonian flow behavior, are consistent with this interaction-damping model [13]. This interaction engineering also reduces reliance on classical surfactants such as polysorbates, which are vulnerable to oxidative and enzymatic degradation and represent a recurrent root cause of particle formation in biologic drug products [36].

For SC ADCs, ILs do not replace classic excipients but rather expand the feasible design space. Cyclodextrins can still be used to complex exposed hydrophobic payload motifs [30], amino-acid excipients can still modulate short-range attraction [26], [28], and buffer systems must still control pH in a regime where the protein contributes significant buffering capacity. However, ILs uniquely target the coupled problem of dose compression and injectability, making them particularly attractive for hydrophobic, higher-DAR ADCs that hit a viscosity wall at target concentration [24].

Integrated platforms should combine dose compression with programmable release. For ADCs, the most compelling path is a hybrid platform in which IL-enabled concentrates are paired with zwitterionic, thixotropic depots. In this architecture, the IL addresses concentration and injectability [13], while the hydrogel governs local residence and attenuates the acute tissue exposure that drives inflammation and payload-related local toxicity [37]. Such platforms also create an opportunity to protect the biologic through the hostile SC microenvironment by reducing nonspecific adsorption to extracellular matrix components and, plausibly, limiting early macrophage engagement at the depot interface. More broadly, protein-avoidant IL coatings have been shown to resist protein adsorption and reduce opsonization in vivo, extending circulation and redirecting biodistribution, which supports the premise that IL-derived shielding can modulate immune recognition [41].

Biodegradability and immunological neutrality must be proven at primate scale. Any depot material intended for repeated administration must demonstrate complete biodegradation, absence of chronic inflammation, and preservation of biologic integrity at translational scale. A zwitterionic, two-component chitosan hydrogel depot illustrates this standard. Mixed directly with sterile IV monoclonal antibody formulations, validated in mice and nonhuman primates, and delivering tunable controlled release with pharmacokinetic parameters comparable to rHuPH20 plus mAb [37]. In vivo degradation was rapid and near-complete, with roughly 90% of the hydrogel mass resorbed by day 27 in a murine implantation model [37]. Importantly, the degradation kinetics of such hydrogel depots are tunable: adjusting polymer concentration, degree of functionalization, and molecular weight can shorten the resorption half-life to accommodate dosing intervals shorter than 28 days. Injection site rotation, a standard practice in subcutaneous biologic administration, provides an additional degree of freedom by allowing local tissue recovery between doses at any given site.

7. Practical considerations for high-viscosity subcutaneous injection

Translating high-concentration ADC formulations into clinical practice requires co-optimization of formulation rheology and delivery device engineering. Formulations exceeding 50 mPa.s at injection-relevant shear rates (approximately 10³–10⁴ s⁻¹) challenge conventional prefilled syringe and autoinjector platforms, where glide force is constrained by spring energy budgets and patient-acceptable injection times (typically <15 s for ≤2 mL volumes [42]). Three complementary strategies address this bottleneck. First, viscosity-reducing excipients and ionic liquid formulation approaches, as discussed above, can suppress non-Newtonian viscosity growth at ultra-high concentrations, keeping glide forces within device specifications. Second, device selection must be matched to the formulation's rheological profile: spring-driven autoinjectors accommodate moderate viscosities (<30 mPa.s), while electromechanical or gas-pressurized large-volume on-body delivery systems (e.g., patch pumps) extend feasibility to higher viscosities and volumes exceeding 2 mL. Third, needle gauge selection involves a direct trade-off between patient comfort (favoring 27-29G) and injection force (favoring larger bore), a tension that viscosity reduction at the formulation level can partly resolve [43]. For ADCs specifically, the shear sensitivity of conjugate integrity must also be evaluated: extrusion through fine-gauge needles at high flow rates imposes transient shear stresses that can promote aggregation or, in extreme cases, affect linker integrity, particularly for high-DAR species with increased hydrophobic surface area. Device-formulation co-design should therefore include shear-stress characterization across the relevant injection parameter space.

8. Bioequivalence strategy for the IV-to-SC transition

Transitioning an ADC from intravenous to subcutaneous delivery introduces predictable pharmacokinetic differences, including a lower peak concentration (C_max) and a delayed time to peak (T_max), that reflect absorption-limited kinetics of depot-mediated systemic uptake rather than a deficiency in total drug exposure. Comparable systemic exposure is therefore typically assessed despite these known route-dependent differences. For ADCs, clinical efficacy is not driven by C_max or trough concentrations alone; rather, the pharmacokinetic-pharmacodynamic relationship is primarily exposure-based, making AUC (AUC_last or AUC_∞) the most pharmacologically relevant bioequivalence criterion. This is consistent with the precedent established by subcutaneous trastuzumab, where non-inferior efficacy was demonstrated at AUC-matched exposures despite substantially lower Cmax relative to IV dosing. Importantly, bioequivalence assessments for ADCs should consider integrated exposure across all key analytes, including total antibody, intact conjugate (conjugated payload), and unconjugated antibody rather than relying on any single pharmacokinetic metric. The C_max attenuation inherent to SC dosing may in fact confer a safety advantage for ADCs whose dose-limiting toxicities are peak-driven. This principle has been demonstrated clinically for non-ADC biologics: subcutaneous daratumumab and subcutaneous amivantamab both show reduced infusion-related reaction rates relative to their IV counterparts while maintaining therapeutic exposure, establishing a clinical precedent for route-dependent safety benefits [44]. For ADCs specifically, peak-related toxicities such as hepatotoxicity and thrombocytopenia (T-DM1) or interstitial lung disease (T-DXd) could plausibly benefit from Cmax attenuation, provided that total exposure remains within the therapeutic window. Route-specific pharmacokinetic bridging studies, guided by exposure-response modeling and informed by catabolite-level analytics, will be essential to establish the therapeutic equivalence of subcutaneous ADC formulations.

9. Outlook toward distributed oncology and industrial sustainability

The transition from IV to SC delivery for ADCs is a predictable pharmaceutical evolution driven by patient preference, healthcare capacity, and the rapid growth of ADC portfolios, and, as discussed above, the C_max attenuation inherent to SC dosing may confer additional safety benefits for ADCs with peak-driven dose-limiting toxicities [5], [9], [45]. What has delayed it is not the absence of motivation but rather the reality that ADCs expose the limits of conventional antibody formulation science when forced into the SC biophysical regime. Volume constraints, viscosity, and linker liability are not impasses; they are the boundary conditions guiding next generation of advanced materials and integrated bioprocessing.

As SC ADCs reach operational scale, the impact will extend beyond convenience. It will relieve infusion centres, reduce the invasiveness of cancer care, and enable ambulatory oncology models that shift treatment from hospitals to outpatient settings and, for selected regimens, potentially to home based administration. This decentralization is not a cosmetic improvement but a structural change that can expand access in regions where infusion infrastructure limits care, accelerate trial execution by simplifying visit logistics, and deliver measurable economic gains through reduced chair time, staffing intensity, and scheduling friction [10]. Importantly, the economic case is already clinically grounded with the PrefHer randomized crossover study which reported an overwhelming patient preference for SC administration of trastuzumab [4], and time-and-motion analyses quantified substantial reductions in treatment-chair occupancy and active healthcare professional time per session, directly translating route choice into usable capacity [10]. For ADCs, route-specific materials and formulation engineering can further reduce injection-site risk by preserving conjugate integrity during SC residence. This protection is enabled when platforms decouple injection viscosity (set by high-shear flow during injection) from post-deposition kinetics (set by mesh size, partitioning, and erosion), thereby attenuating burst leakage of free payload and limiting local exposure [37]. Industrially, SC-compatible platforms support portfolio-wide standardization of device-compatible fill-finish and supply chains. Over the next decade, the systems that successfully co-design conjugation chemistry, formulation physics, advanced materials, manufacturing, and delivery devices will not merely enable SC ADCs but could contribute to reshaping the delivery model of oncology toward ambulatory and decentralized care.

CRediT authorship contribution statement

Alexandre Detappe: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. María Gutiérrez-Blanco: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Xavier Pivot: Writing – review & editing, Writing – original draft, Conceptualization.

Declaration of competing interest

A.D. and X.P. are co-founders of Recobia Tx who seeks to develop a hydrogel formulation for subcutaneous administration of large proteins.

Data availability

Data will be made available on request.