Emerging agents for boron neutron capture therapy: Biomolecular carriers paving the way for precision radiotherapy Authors: Chang Chen, Maosong Yang, Jiheng Wang, Ling Zhang, Shaobo Huang, Xiao Xu, Yang Liu Published: 2026 ABSTRACT AbstractBoron neutron capture therapy (BNCT) is an emerging targeted radiotherapy technology that enables precise eradication of tumor cells. The development of efficient boron carriers is crucial for the advancement of BNCT technology. Notably, biomolecular carriers currently represent the only class of drug carriers that has achieved market approval (e.g., boronophenylalanine [BPA] in Japan), underscoring their pivotal role in clinical applications. Their outstanding biocompatibility, high payload capacity, and tumor‐targeting capabilities align closely with the stringent requirements of BNCT drug delivery. For instance, amino acid‐based carriers (such as BPA) utilize L‐type amino acid transporter 1 (LAT1) for tumor‐specific accumulation; nucleoside prodrugs are activated by thymidine kinase 1 (TK1) and incorporated into tumor DNA; carbohydrate‐based carriers improve solubility and targeting through glucose transporters (GLUTs); peptide carriers facilitate cellular uptake via receptor‐mediated endocytosis or cell‐penetrating peptides; and liposomes leverage the enhanced permeability and retention (EPR) effect together with active‐targeting modifications for enhanced tumor delivery. Nevertheless, delivering ultrahigh doses of boron remains a major challenge. Biomacromolecules, particularly albumin‐based systems, have garnered increasing attention because of their excellent biosafety and high boron‐loading potential, offering promising avenues for overcoming dose delivery barriers in BNCT. This review systematically summarizes the recent advances in biomolecular carriers for BNCT and highlights emerging applications of protein‐based boron carriers, addressing a notable gap in the comprehensive review of biomolecular strategies within the BNCT drug delivery field.Keywords: Biomolecular Carriers, Boron Neutron Capture Therapy, Precision Radiotherapy INTRODUCTION Boron neutron‐capture therapy (BNCT) has emerged as a cutting‐edge technology in precision oncology. Its mechanism hinges on the nuclear reaction between non‐radioactive boron‐10 (10B) and low‐energy neutrons (<0.5 eV). Upon neutron capture, 10B undergoes fission to generate α particles (4He2 +) and recoil lithium‐7 nuclei (7Li3 +), which release high energy over an extremely short range of <10 µm—equivalent to the diameter of a single cell 1 , 2 , 3 , 4 (Figure 1). This mechanism enables the selective destruction of tumor cells with minimal damage to adjacent normal tissues, establishing BNCT as a highly promising cancer treatment strategy. 5 Principle and developmental timeline of boron neutron capture therapy A review of their history reveals that boron carriers have played a pivotal role in the advancement of BNCT. The discovery of neutrons in 1932 6 and the proposal of the BNCT principle in 1936 7 pioneered explorations in this field. The first clinical trial in 1951 provided initial evidence of tumor regression effects. 8 , 9 , 10 However, severe adverse reactions caused by insufficient neutron penetration and poor selectivity of boron compounds led to a temporary suspension of research efforts. 11 Breakthroughs by Japanese researchers revitalized BNCT: Hatanaka employed sodium borocaptate (BSH) in astrocytoma trials, extending median overall survival (MOS) to 21.3 months 12 —surpassing the 14.6 months achieved with the current standard Stupp protocol. 13 The development of boronophenylalanine (BPA) and its fructose conjugate further expanded the spectrum of tumors amenable to BNCT. 14 , 15 , 16 Notably, following the world's first successful BNCT treatment for head and neck cancer by Kato et al., 17 BPA—an amino acid‐based biomolecular carrier—became the first BNCT drug to receive market approval (Steboronine®) in 2020, indicated for the treatment of unresectable or recurrent head and neck cancer 18 (Figure 1). However, delivering ultrahigh doses of boron remains a critical challenge. Against this backdrop, next‐generation boron carriers based on biomolecular scaffolds have emerged, integrating precision targeting, high boron‐loading capacity, and excellent biocompatibility to achieve tumor‐specific delivery and controlled release of boron drugs. In particular, albumin‐based systems have garnered increasing attention owing to their outstanding biosafety and high boron‐loading potential, offering a promising avenue for overcoming dose delivery barriers associated with BNCT. 19 While the clinical translation of BNCT has been systematically addressed in review articles in the literature, 20 the current review focuses on recent advances in biomolecular carriers, including amino acids, peptides, nucleosides, sugars, proteins, and lipids; highlighting on their design concepts, optimization strategies, and application potential; and fills a notable gap in comprehensive reviews of biomolecular strategies for BNCT drug delivery. BIO‐MOLECULAR CARRIERS Amino‐acids Boron carriers based on amino acids have shown significant potential in BNCT for the treatment of intractable tumors such as brain tumors. Research in this field has primarily focused on optimizing tumor targeting, boron‐loading capacity, water solubility, and cellular uptake efficiency. Early efforts began with the synthesis of 4‐dihydroxyborylphenylalanine (BPA) by Snyder et al. in 1958. 21 BPA selectively accumulates in melanoma and brain tumors in vivo through L‐type amino acid transporter 1 (LAT1)‐mediated uptake. However, BPA contains only one boron atom, exhibits poor water solubility, and achieves insufficient tumor‐to‐normal tissue (T/N) ratio. Consequently, subsequent work centered on two major goals: achieving a higher boron content and enhancing its targeting capability. In the 1970s, researchers attempted to replace the phenyl ring with a closododecaborate cluster to synthesize ortho‐carboranylalanine (o‐CBA) 22 by theoretically increasing the number of boron atoms from 1 to 12. However, cellular studies revealed reduced affinity for LAT1 and boron accumulation levels less than half of those achieved with BPA. These findings indicated that merely augmenting boron clusters was insufficient to overcome membrane transport limitations. Since the 1990s, three major strategies were adopted: a) Structural hydrophilization—the introduction of hydroxyl groups 23 or carbohydrate appendages 24 (e.g., BPA‐fructose and borophenylserine syn‐22) has been shown to increase solubility from 0.008 M in BPA to up to 1.2 M. However, this was accompanied by reduced cellular uptake, suggesting that enhanced hydrophilicity may impair recognition by LAT1. b) Stereochemical optimization: Asymmetric synthesis was performed to obtain pure L‐enantiomers. Studies confirmed that LAT1 exhibits significantly higher affinity for L‐BPA than for D‐BPA, providing a basis for subsequent drug design. 25 c) Non‐natural cyclic scaffold compounds developed after the 2000s, such as cis‐1‐amino‐3‐boronocyclopentane‐carboxylic acid (cis‐ABCPC) (Figure 2), demonstrated a superior T/N boron distribution ratio compared to that of BPA. In vitro studies of its derivative, 1‐amino‐3‐fluorocyclobutane‐1‐carboxylic acid conjugated with BSH (ACBC‐BSH), showed higher boron accumulation in LAT1‐expressing brain tumor‐derived cell lines (C6, A172, and U81) than L‐BPA. 26 Structures of representative biological carriers: amino‐acids, nucleosides, and carbohydrates. Recent advances have primarily focused on three aspects. First, highly soluble derivatives—represented by 3‐borono‐L‐tyrosine (BTS; 250 mg•mL−1; Table 1)—have enabled a linear increase in tumor boron accumulation with elevated administration doses, while maintaining LAT1‐dependent uptake, albeit accompanied by increased non‐target deposition in the kidneys and pancreas. 27 Second, triBoctryptophan compounds (TB‐6‐BT), designed using a Boc‐protecting group strategy, combine indole rings with boron clusters to provide a novel template for blood‐brain barrier penetration. 28 Third, real‐time imaging with fluorine‐18–labeled fluoroboronophenylalanine (18F‐FBPA) positron emission tomography (PET) enables noninvasive quantification of LAT1‐mediated boron distribution, allowing precise patient selection, dose prediction, and therapeutic response evaluation for BNCT. 29 TABLE 1Key properties and clinical progress of biomolecular carriers for boron neutron capture therapyBiomolecular carrierBoron loading capacityTargeting efficiencyClinical stageSafety assessmentReferencesBPA (p‐boronophenylalanine)Each molecule contains one boron atom.Poor retention in tumorsClinical approvalClinically safe with low intrinsic toxicity, with minor boron accumulation in normal tissues. 14 , 18 BTS (3‐borono‐L‐tyrosine)Each molecule contains one boron atom. (Boron delivery in vivo is 2–3 times greater than BPA)Tumor/Blood > 3PreclinicalNo significant inherent toxicity, rapid blood clearance 27 3‐(carboranylalkyl) thymidinesEach molecule contains 10 boron atoms/PreclinicalGood biocompatibility, no significant toxicity 33 Thymidine‐metallacarboraneEach molecule contains 18 boron atoms (approximately 1.5 times that of BSH)accumulate in V79 Chinese hamster cells (boron content: 60 pg per 106 cells)PreclinicalGood biocompatibility, no significant toxicity 35 D‐glucoheptose derivative (C‐7)Each molecule contains one boron atom (Boronic acid)6 ng B/mg cells boron accumulation in glioma cellsPreclinicalLow toxicity, metabolically stable, high biocompatibility of sugar carriers, weak off‐target effects. 38 D‐glucal derivatives (2‐boryl‐1,2‐dideoxy‐D‐glucose)Each molecule contains one boron atom. (Boronic acid ester)Higher uptake in tumor tissuePreclinicalNo significant inherent toxicity 39 β‐D‐glucopyranose (Glucose‐BSH)Each molecule contains 12 boron atoms. (BSH conjugated)T/B: 2.2, T/N(skin): 3.1, T/N(pancreas):5.8.PreclinicalNo significant inherent toxicity 40 GRPR (Carborane‐functionalized peptide agonist)Each molecule contains 40–80 boron atoms (polycarborane).Higher uptake in tumor tissuePreclinicalNo intrinsic cytotoxicity, low immunogenicity and excellent biocompatibility 45 ANG‐B (Angiopep‐2 conjugate)Each molecule contains multiple boron atoms (Peptide conjugate)High tumor‐to‐blood ratioPreclinicalNo obvious brain tissue damage, no significant systemic toxicity. 50 PSMA (Carborane‐loaded Nanoparticles)High boron loading (Carborane loaded in PLGA‐PEG core)Tumor/Blood ratio = 25;The accumulation in PSMA(+) tumors is twice that in PSMA(‐) tumors.PreclinicalGood biocompatibility, no significant immune response, low toxicity risks. 53 Liposomes (Water‐soluble boranes)Encapsulation of high‐concentration water‐soluble ionic boron compoundsi‐B20H18 2 − group achieved a T/B ratio of 12 at 48 h, with blood boron at only 1.2 ppmPreclinicalNo significant toxicity 57 Liposomes (PEGylated, nido‐carborane)Encapsulated water‐soluble nido‐carborane (47.5% loading efficiency)Tumor/Blood ratio = 4 ± 0.3;Tumor/Muscle = 37;Tumor boron accumulation reached 93.3 ppm.PreclinicalPEG modification reduces immunogenicity, high loading capacity minimizes administration volume, with no significant systemic toxicity observed. 62 Liposomes (Boronsome)Boron uptake after 12 h is 182 µg/106 cells—61 % above BPA.tumor/muscle = 37 ± 1; tumor/brain = 62 ± 2; tumor/bone = 17 ± 1; tumor/blood = 4 ± 0.3PreclinicalNo systemic toxicity and minimal risk of immune response or inflammation. 65 Transferrin, BSA (MID‐conjugates)MID‐TF conjugates (MID: TF ¼ 10:1, 3000 ppm B)MID‐TF (20 ppm) (12 hours after administration)PreclinicalExcellent biocompatibility, no apparent toxicity in vivo 66 HemoglobinHigh boron loading (Simultaneous loading of multiple carborane molecules)/Theoretical researchHigh biocompatibility 67 Key properties and clinical progress of biomolecular carriers for boron neutron capture therapy Nucleosides Cancer remains a major public health concern that severely endangers human life and health. 30 , 31 The ability of cancer cells to incorporate nucleotide analogs into newly synthesized DNA strands during replication underlies the mechanism of action of many anticancer chemotherapeutic drugs. Consequently, boronic nucleotides (BN) and their precursors appear to be ideal agents for the selective delivery 10B into cancer cells, where they can be stably integrated into DNA. 32 This precise targeting strategy may significantly reduce the boron dosage required to induce tumor cell death. The earliest example, N3‐carboranyl deoxythymidine (Table 1), first reported by Lunato et al., demonstrated that exogenous nucleosides can be recognized and phosphorylated by human thymidine kinase 1 (TK1), thereby introducing 10B into rapidly proliferating tumor cells. 33 To address its low solubility under physiological conditions owing to its strong hydrophobicity, the same team introduced an additional dihydroxypropyl substituent. This compound not only retained its affinity for TK1/2 but also significantly extended survival in the F98 rat glioma model, establishing the feasibility of the “nucleoside‐kinase‐DNA” delivery route. 34 Subsequently, Leśnikowski et al. introduced metallacarboranes into nucleosides to form nucleoside‐metallacarborane conjugates, increasing the boron load per molecule to 1.5 times that of monocarboranes and 18 times that of BPA. These conjugates also demonstrated preliminary capability to incorporate into oligonucleotides, highlighting their potential for enhancing dose density in BNCT. 35 Luvino et al. 36 synthesized the first 3′‐boronated thymidine monophosphate analog using a 7‐step route. They confirmed its ability to assemble into “phosphate‐free” dinucleotides through reversible boronate ester bonds with cis‐diol‐containing nucleosides/nucleotides in aqueous solution. The formation of these bonds depends on the C3′‐endo sugar conformation and base stacking, with a distinct preference for pyrimidine‐based substrates. Their work provided a new chemical tool for constructing dynamic nucleic acid backbones and developing high‐boron carriers for BNCT that can be incorporated into DNA or RNA. Carbohydrates The 1995 study by Kirihata et al. first demonstrated that the formation of a stable 2:1 coordination complex between D‐fructose and BPA could overcome two major limitations: the poor aqueous solubility of BPA (<2 mg•mL− 1) and its insufficient cellular uptake. This “sugar‐coating” strategy, facilitated by GLUTs, increased the intratumoral boron concentration of BPA by approximately threefold, and established a fundamental paradigm of “using sugars for solubilization and targeting” in all subsequent “carbohydrate‐boron” conjugates. 37 Over the past five years, research on the new BNCT strategy of “glycosylated boron agents” has progressed in stages, from “small molecule‐monosaccharide” and “oligosaccharide‐polysaccharide” systems to “polysaccharide‐nanocomposite” systems. Muz et al. 38 introduced a boronic acid group at the C‐7 position of a glucoheptose via a two‐step olefin metathesis and deprotection strategy to achieve a total yield of 43%. The resulting compound exhibited negligible toxicity in normal fibroblasts but achieved a boron accumulation of 6 ng•mg− 1 in D54 glioma cells after 24 h—approximately 40 times higher than in untreated cells—providing a concise and efficient platform for “carbohydrate‐boron” small molecule BNCT research. Itoh et al. 39 further advanced the small‐molecule strategy by directly synthesizing 2‐boryl‐1,2‐dideoxyglucose derivatives at the C2 position through regio‐ and stereoselective hydroboration of D‐glucal (Table 1). This approach achieved direct C–B bond formation for the first time. Uptake was significantly enhanced under hypoxic conditions when glucose transporter 1 (GLUT1) expression is upregulated, and intracellular phosphorylation by hexokinase enabled metabolic “trapping,” suggesting potential for combined delivery with BPA and sodium borocaptate (BSH) via multiple transporters. Fujimoto et al. 40 translated the “sugar‐boron” concept a step closer to clinical application by covalently conjugating glucose to BSH to produce glucose‐BSH. This agent leveraged the high expression of GLUT1/3 in CA19‐9‐positive pancreatic cancer cells to overcome the limitations of LAT1 dependency. The tumor‐to‐pancreas ratio reached 11 and a single BNCT session reduced the tumor volume from 779 to 14 mm. Based on these results, they proposed a “Precision BNCT” strategy for preoperative downstaging. The carrier scale then advanced from molecular to nanoscale, with Yamana et al. 41 fabricating a composite of high‐boron‐density BNNT and a β‐1,3‐glucan‐Protein A mimetic via high‐speed ball milling, followed by conjugation with HER‐2 antibodies to construct an 80–130 nm ternary complex. Antibody‐antigen mediation enhanced cellular uptake by 12‐fold, achieving a T/B ratio of 390. The therapeutic efficacy was 30 times higher than that of L‐BPA/fructose, marking the first realization of polysaccharide‐boron nitride nanotube‐antibody synergistic BNCT. Peptides Peptide‐conjugated BNCT takes advantage of the specific binding affinity of certain peptides for proteins (e.g., receptors or transporters) overexpressed on the surface of cancer cells. 42 By linking boron‐containing compounds to these peptides, boron can be selectively delivered to cancer cells, sparing healthy tissues. 43 Early research focused on utilizing single receptors that are overexpressed in tumors to achieve selective boron enrichment. Nakase et al. conjugated the Z33 peptide (Figure 3), which recognizes the Fc domains, with BSH, and demonstrated that Fc receptor‐mediated pinocytosis significantly enhanced boron uptake in A431 cells. 44 This approach resulted in stronger cancer cell‐killing efficacy in subsequent In vitro BNCT experiments. Hoppenz et al. developed a carborane‐functionalized peptide agonist that targeted the gastrin‐releasing peptide receptor (GRPR). They demonstrated that both selectivity and boron load could be simultaneously optimized in GRPR‐overexpressing HEK293 cells. 45 Worm et al. further conjugated the human Y1 receptor (hY1R, a type of G protein‐coupled receptor [GPCR]) ligand neuropeptide Y (NPY) with bis‐deoxygalactosylcarborane, and showed that adjusting the linker length could maximize the number of boron atoms per peptide molecule while maintaining selectivity. 46 In vitro and in vivo studies revealed that the optimized NPY conjugate exhibited high selectivity toward NPY receptor‐overexpressing cancer cells. This led to increased uptake of boron‐containing compounds and enhanced the therapeutic efficacy in mouse cancer models. Delivery strategy of boron drug carrier‐peptides: ① Antibody Fragment crystallizable (Fc) fragment for immune synergy; ② Carborane‐functionalized peptide for targeting the gastrin‐releasing peptide receptor (GRPR); ③ Defensin B‐derived peptide with 8 arginine residues(DB‐r8) and Defensin B‐derived peptide with RLA motif (DB‐RLA) peptides for subcellular organelle targeting; ④ 1,4,7,10‐Tetraazacyclododecane‐1,4,7,10‐tetraacetic acid (DOTA‐Cu64) radionuclide for Positron emission tomography (PET) imaging. ⑤ Angiopep‐2 peptide to enhance blood‐brain barrier penetration. BBB, Blood‐brain barrier; Z33, Z domain variant 33. Subsequently, researchers introduced cell‐penetrating peptides (CPPs) such as polyarginine (R8/R11) and angiopep‐2 into boron delivery systems. Michiue et al. developed two BSH‐CPP derivatives. First, they constructed an octameric BSH‐11R dendrimeric peptide, which increased intracellular boron concentration in glioma cells by 350‐fold via macropinocytosis, significantly enhancing BNCT efficacy. 47 Subsequently, they developed a short BSH‐3R peptide derivative that balanced efficient cellular uptake with 64Cu‐DOTA PET visualization, providing a new tool for individualized pharmacokinetic monitoring in BNCT. 48 Nakase et al. reported that DB‐r8 and DB‐RLA achieved cytoplasmic and mitochondrial localization, respectively, demonstrating that CPP‐boron conjugates are capable of targeting subcellular organelles. Compared to unmodified BSH, ELISA quantification revealed a 4‐fold increase in the cellular uptake of arginine‐lysine amide (RLA) conjugates, whereas the uptake of r8 conjugates was lower than that of RLA but still higher than that of BSH. WST‐1 assays confirmed no significant changes in cell viability after 30 min of treatment with the CPP‐dodecaborate conjugates. 49 Xiang et al. developed an angiopep‐2‐boronic acid conjugate (ANG‐B) that reduced the tumor volume by 62.9% in a mouse model of intracranial glioma and significantly outperformed BPA (23.0%) (Table 1). This further validated the clinical translational potential of the “angiopep‐2‐CPP” strategy for BNCT in brain tumors. 50 Unlike earlier receptor‐dependent targeting strategies, CPP‐based delivery systems do not rely on specific cell surface receptors, offering broader applicability across diverse tumor types. However, this may increase the risk of off‐target toxicity resulting from non‐specific uptake by normal tissues. Mitigation strategies include activation by the tumor microenvironment (e.g., pH or enzymes) or combining boron‐drug conjugates with focused neutron beams to reduce toxicity. Recent studies have integrated the aforementioned strategies with nanocarriers or self‐assembly technologies to develop a “peptide‐nanoparticle‐boron” tripartite delivery platform. Kawai et al. synthesized cyclic RGD‐functionalized closo‐dodecaborate albumin conjugates (cRGD‐MID‐BSA), which actively target integrin αvβ3. Compared with unmodified albumin conjugates and free boron compounds, this conjugate demonstrated significantly enhanced therapeutic efficacy. In vivo fluorescence imaging revealed a markedly higher tumor‐to‐liver signal ratio than that in non‐targeted groups. 51 Mishiro et al. combined RGD with Ga‐DOTA to achieve integrated BNCT theranostics. 52 Additionally, Meher et al. reported a prostate‐specific membrane antigen (PSMA)‐targeted PLGA‐b‐PEG nanoparticle with excellent stability suitable for both PET imaging and boron delivery. In vitro experiments confirmed its PSMA‐specific binding capability. 53 Sharma et al. developed a pH‐low insertion peptide (pHLIP)‐targeted zinc gallogermanate long‐persistent‐luminescence nanoprobe (SZGO:Cr‐10B‐NF) that enabled efficient targeting of the acidic tumor microenvironment. In animal models, it reduced tumor volume by 75–90% without systemic toxicity. 54 Isono et al. substantially enhanced the cellular uptake efficiency of BSH and its cytotoxicity under neutron irradiation using a lipopeptide modification strategy. 55 These studies provide diverse insights into the design of efficient and low‐toxicity boron carriers for BNCT. Lipids In recent years, liposomes have emerged as highly promising and versatile boron delivery platforms for BNCT owing to their high loading efficiency, clinical safety, passive tumor accumulation (EPR effect), 56 and ease of targeted modification (Figure 4). Early pioneering work by the Hawthorne group (UCLA) demonstrated that liposomes can simultaneously encapsulate hydrophilic polyhedral boranes (e.g., Na3[B20H17NH3]) in the aqueous core and embed hydrophobic nido‐carboranes [nido‐7‐CH3(CH2)15‐7,8‐C2B9H11] within the lipid bilayer. This approach enables the delivery of 3–5 wt% boron in a single administration, achieving intratumoral concentrations of >30 ppm. 57 , 58 Subsequent systematic optimization by the International Institute of Nano and Molecular Medicine (MU‐I2NM2) at the University of Missouri further demonstrated that small unilamellar liposomes (MAC‐TAC liposomes), 100–130 nm in size, could selectively accumulate in various rodent tumor models (e.g., EMT6 and RG2) within 30–48 h via the EPR effect. The peak tumor boron level reached 60 ppm, with T/B and T/N ratios both exceeding 5.5. Notably, no significant toxicity was observed in mouse, hamster, or canine head and neck cancer models. 59 , 60 Schematic illustration of cancer‐targeting drug delivery strategies based on liposomes carriers. Meanwhile, research groups led by Yanagie and Lee have advanced the strategy from “passive EPR” to “active targeting.” Yanagie encapsulated BSH in anti‐CEA immunoliposomes and demonstrated a synergistic effect between antibody density and 10B dose on cell killing. 61 Lee et al. achieved a boron‐loading rate of 47.5% using PEGylated nido‐carborane liposomes (Table 1), which significantly outperformed free BSH in suppressing the growth of CT26 tumors in mice. 62 Takeuchi et al. further compared “post‐insertion” and “pre‐insertion” PEG modification strategies, reporting that the former required only half the amount of PEG to maintain equivalent particle size, zeta potential, and encapsulation efficiency. Both liposomal formulations delivered 10B to tumor cells at a concentration of 109 atoms per cell, meeting the requirements for BNCT. 63 Kullberg et al. co‐delivered EGF receptor‐targeted liposomes with boronated acridine and confirmed enhanced BNCT‐induced killing of glioma cells. 64 Recently, Liu et al. developed “boronsomes”—liposomes capable of co‐loading boron agents and chemotherapeutic drugs. By monitoring 64Cu‐labeled tracers via PET imaging, they achieved precise delivery with a T/B ratio of 4.0 and validated the synergistic tumor‐suppressive effects of combined BNCT and chemotherapy in mouse models. 65 Proteins In recent years, utilizing blood transport proteins as “nanocarriers” for hydrophobic carboranes has emerged as a delivery strategy that combines biocompatibility with tumor‐targeting potential (Figure 5). Kikuchi et al. (2020) 66 conducted a systematic comparison of covalent conjugates formed by maleimide‐functionalized closo‐dodecaborane (MID) with transferrin (TF) and bovine serum albumin (BSA), using colon 26 tumor‐bearing mice model (Table 1). Their results demonstrated that 68 kDa MID‐BSA achieved higher tumor accumulation (62 ppm at 12 h) via EPR effect. In contrast, the 48 kDa MID‐TF relied on transferrin receptor‐mediated endocytosis and achieved lower tumor concentration (20 ppm). Cellular experiments confirmed the time‐ and dose‐dependent uptake of MID‐TF, and competitive inhibition suggested the involvement of non‐receptor pathways. Marforio et al. 67 employed reverse docking and 100 ns molecular dynamics simulations to systematically screen and quantitatively compare the binding energies of 15 plasma proteins with ortho‐carborane. The results revealed the following order: hemoglobin (−27.5 kcal mol− 1) > transthyretin (−25.3 kcal mol− 1) > human serum albumin (−23.5 kcal mol− 1). The remaining eight proteins exhibited binding energies comparable to those of human serum albumin (HSA) and transthyretin (TTR), ranging from approximately ‐20 to ‐23 kcal·mol− 1. This study not only validated earlier experimental findings confirming reversible carborane binding to HSA and TTR, 68 , 69 but also provided the first systematic atomic‐level characterization of binding sites and interaction fingerprints: the FA5 pocket of HSA (Phe554/551 BH‐π/CH‐π interactions), the T4 site of TTR (hydrophobic pocket of Leu110 and dual hydrogen bonds with Thr119), and the heme cavity of Hb (BH‐π interaction with Phe98 and dual hydrogen bonds with Ser131). The authors further highlighted that hydrophobic interactions, BH‐π/CH‐π interactions, dual hydrogen bonds, surfactant‐like effects, and classical hydrogen bonds are common driving forces across proteins. This suggests the potential to enhance EPR effects or receptor‐mediated tumor uptake through protein engineering or chemical modifications (e.g., folic acid conjugation or near‐infrared fluorescent labeling). By synergizing methodologies, this work complements earlier studies on HSA‐based solubilization of metallacarboranes 70 , 71 and TTR‐carborane inhibitors 67 through methodological synergy. While earlier experimental studies focused on binding validation and pharmacokinetics, Marforio et al. established an atomic‐level structural basis and a virtual screening framework, laying a theoretical foundation for the targeted optimization and multifunctional theranostic integration of “protein‐carborane” complexes in BNCT. Schematic illustration of boron drug delivery strategies based on protein carriers. DISCUSSION AND CONCLUSIONS BNCT is a refined cancer radiotherapy modality that has been developed over decades of research (Table 1). The core challenge in its clinical application lies in achieving tumor‐selective targeting by delivering boron agents at therapeutically effective concentrations to tumors, while minimizing boron uptake by normal tissues. Although candidate delivery vectors, including boronated nucleosides, peptides, and proteins, have demonstrated substantial progress in cellular and animal models, the mechanism of action of BNCT remains fundamentally distinct from that of conventional intravenously administered antineoplastic drugs. BNCT requires boron accumulation in tumor tissues to exceed 20 µg·g− 1, which necessitates a single clinical dose of boron carriers (e.g., BPA) as high as 500 mg·kg− 1—approximately 100 times higher than conventional chemotherapeutics dosages—posing a tremendous challenge to current delivery systems. As the most abundant endogenous macromolecular protein in human plasma, albumin exhibits excellent biocompatibility and safety: a single clinical infusion dose can reach 20–50 g (equivalent to 400–1000 mg/kg) with favorable tolerability, and more than ten albumin‐based drugs (e.g., nab‐paclitaxel Abraxane®) have received regulatory approval for clinical application. Albumin enables passive tumor targeting via EPR effect, and its plasma half‐life is extended to 21 days through FcRn receptor‐mediated recycling, thus conferring significant advantages in tumor‐targeted delivery and sustained release. In the future, boron‐albumin conjugates with high boron loading capacity can be developed via site‐specific conjugation technology or nanoparticle encapsulation strategies. Modifying the albumin surface with tumor‐specific ligands, such as HER2 antibody fragments, can further improve targeting precision. Additionally, various biological carriers possess unique merits, including high loading capacity and favorable biocompatibility. Integrating these advantages through a synergistic combination of carrier systems (e.g., peptide‐functionalized lipid nanoparticles and glyco‐modified albumin conjugates) is expected to resolve the dose‐selectivity dilemma. Nevertheless, clinical translation requires the establishment of scalable manufacturing processes that balance cost control, product consistency, purity, and bioactivity. For small‐molecule carriers, it is essential to optimize synthetic routes featuring high yields and simplified purification steps while complying with good manufacturing practice (GMP) standards. Microfluidic and continuous‐flow synthesis technologies can be adopted for macromolecular and nanoscale carriers to ensure reproducibility and batch‐to‐batch consistency. Quality control parameters must be established to monitor critical attributes such as boron loading efficiency and particle size distribution. Only through interdisciplinary collaboration to develop customized and scalable manufacturing processes can BNCT carriers advance from preclinical research to clinical evaluation and commercialization. CONFLICT OF INTEREST STATEMENT The authors declare that they have no competing interests. ETHICAL STATEMENT Not applicable. ACKNOWLEDGMENTS This study was supported by the National Natural Science Foundation of China (Grant Nos. 52303212), the Institute of Energy, Hefei Comprehensive National Science Center (Grant No. 21KZS202 and 26KZS101),China Postdoctoral Science Foundation (Grant Number 2023M743351), and the National Key Research and Development Program of China (2024YFE03250600). Contributor Information Shaobo Huang, Email: bob2022@smu.edu.cn. Xiao Xu, Email: xiaoxu721@smu.edu.cn. Yang Liu, Email: yangliu2022@ustc.edu.cn. DATA AVAILABILITY STATEMENT Data will be made available upon reasonable request. REFERENCES Associated Data Data Availability Statement Data will be made available upon reasonable request.