Patient-Specific, Mechanistic Models of Tumor Growth Incorporating Artificial Intelligence and Big Data.

Despite the remarkable advances in cancer diagnosis, treatment, and management over the past decade, malignant tumors remain a major public health problem. Further progress in combating cancer may be enabled by personalizing the delivery of therapies according to the predicted response for each individual patient. The design of personalized therapies requires the integration of patient-specific information with an appropriate mathematical model of tumor response. A fundamental barrier to realizing this paradigm is the current lack of a rigorous yet practical mathematical theory of tumor initiation, development, invasion, and response to therapy. We begin this review with an overview of different approaches to modeling tumor growth and treatment, including mechanistic as well as data-driven models based on big data and artificial intelligence. We then present illustrative examples of mathematical models manifesting their utility and discuss the limitations of stand-alone mechanistic and data-driven models. We then discuss the potential of mechanistic models for not only predicting but also optimizing response to therapy on a patient-specific basis. We describe current efforts and future possibilities to integrate mechanistic and data-driven models. We conclude by proposing five fundamental challenges that must be addressed to fully realize personalized care for cancer patients driven by computational models.

Multidimensional opioid abuse deterrence using a nanoparticle-polymer hybrid formulation.

Misuse of prescription opioid drugs is the leading cause of the opioid crisis and overdose-related death. Abuse deterrent formulations (ADFs) have been developed to discourage attempts to tamper with the formulation and alter the ingestion methods. However, abusers develop complex extraction strategies to circumvent the ADF technologies. For comprehensive deterrence of drug abuse, we develop tannic acid nanoparticles (NPs) that protect encapsulated opioids from solvent extraction and thermal challenge (crisping), complementing the existing formulation strategy to deter injection abuse. Here, we develop a hybrid ADF tablet (NP-Tab), consisting of iron-crosslinked tannic acid NPs encapsulating thebaine (model opioid compound), xanthan gum, and chitosan (gel-forming polymers), and evaluate its performance in common abuse conditions. NP-Tab tampered by crushing and suspended in aqueous solvents forms an instantaneous gel, which is difficult to pull or push through a 21-gauge needle. NPs insulate the drug from organic solvents, deterring solvent extraction. NPs also promote thermal destruction of the drug to make crisping less rewarding. However, NP-Tab releases thebaine in the simulated gastric fluid without delay, suggesting that its analgesic effect may be unaffected if consumed orally as prescribed. These results demonstrate that NP-Tab can provide comprehensive drug abuse deterrence, resisting aqueous/organic solvent extraction, injection, and crisping, while retaining its therapeutic effect upon regular usage.

Inverse resolution of spatially varying diffusion coefficient using Physics-Informed neural networks.

Resolving the diffusion coefficient is a key element in many biological and engineering systems, including pharmacological drug transport and fluid mechanics analyses. Additionally, these systems often have spatial variation in the diffusion coefficient which must be determined, such as for injectable drug-eluting implants into heterogeneous tissues. Unfortunately, obtaining the diffusion coefficient from images in such cases is an inverse problem with only discrete data points. The development of a robust method that can work with such noisy and ill-posed datasets to accurately determine spatially-varying diffusion coefficients is of great value across a large range of disciplines. Here, we developed an inverse solver that uses physics informed neural networks (PINNs) to calculate spatially-varying diffusion coefficients from numerical and experimental image data in varying biological and engineering applications. The residual of the transient diffusion equation for a concentration field is minimized to find the diffusion coefficient. The robustness of the method as an inverse solver was tested using both numerical and experimental datasets. The predictions show good agreement with both the numerical and experimental benchmarks; an error of less than 6.31% was obtained against all numerical benchmarks, while the diffusion coefficient calculated in experimental datasets matches the appropriate ranges of other reported literature values. Our work demonstrates the potential of using PINNs to resolve spatially-varying diffusion coefficients, which may aid a wide-range of applications, such as enabling better-designed drug-eluting implants for regenerative medicine or oncology fields.

Geometry of adipocyte packing in subcutaneous tissue contributes to nonlinear tissue properties captured through a Gaussian process surrogate model.

Subcutaneous tissue mechanical response is governed by the geometry and mechanical properties at the microscale and drives physiological and clinical processes such as drug delivery. Even though adipocyte packing is known to change with age, disease, and from one individual to another, the link between the geometry of the packing and the overall mechanical response of adipose tissue remains poorly understood. Here we create 1200 periodic representative volume elements (RVEs) that sample the possible space of Laguerre packings describing adipose tissue. RVE mechanics are modeled under tri-axial loading. Equilibrium configuration of RVEs is solved by minimizing an energetic potential that includes volume change contributions from adipocyte expansion, and area change contributions from collagen foam stretching. The resulting mechanical response across all RVE samples is interpolated with the aid of a Gaussian process (GP), revealing how the microscale geometry dictates the overall RVE mechanics. For example, increase in adipocyte size and increase in sphericity lead to adipose tissue softening. We showcase the use of the homogenized model in finite element simulations of drug injection by implementing a Blatz-Ko model, informed by the GP, as a custom material in the popular open-source package FEBio. These simulations show how microscale geometry can lead to vastly different injection dynamics even if the constituent parameters are held constant, highlighting the importance of characterizing individual's adipose tissue structure in the development of personalized therapies.

Poroelastic Characterization and Modeling of Subcutaneous Tissue Under Confined Compression.

Subcutaneous tissue mechanics are important for drug delivery. Yet, even though this material is poroelastic, its mechanical characterization has focused on its hyperelastic response. Moreover, advancement in subcutaneous drug delivery requires effective tissue mimics such as hydrogels for which similar gaps of poroelastic data exist. Porcine subcutaneous samples and gelatin hydrogels were tested under confined compression at physiological conditions and strain rates of 0.01%/s in 5% strain steps with 2600 s of stress relaxation between loading steps. Force-time data were used in an inverse finite element approach to obtain material parameters. Tissues and gels were modeled as porous neo-Hookean materials with properties specified via shear modulus, effective solid volume fraction, initial hydraulic permeability, permeability exponent, and normalized viscous relaxation moduli. The constitutive model was implemented into an isogeometric analysis (IGA) framework to study subcutaneous injection. Subcutaneous tissue exhibited an initial spike in stress due to compression of the solid and fluid pressure buildup, with rapid relaxation explained by fluid drainage, and longer time-scale relaxation explained by viscous dissipation. The inferred parameters aligned with the ranges reported in the literature. Hydraulic permeability, the most important parameter for drug delivery, was in the range k 0 ∈ [ 0.142 , 0.203 ] mm 4 /(N s). With these parameters, IGA simulations showed peak stresses due to a 1-mL injection to reach 48.8 kPa at the site of injection, decaying after drug volume disperses into the tissue. The poro-hyper-viscoelastic neo-Hookean model captures the confined compression response of subcutaneous tissue and gelatin hydrogels. IGA implementation enables predictive simulations of drug delivery.

α-Gal Nanoparticles in CNS Trauma: II. Immunomodulation Following Spinal Cord Injury (SCI) Improves Functional Outcomes.

BACKGROUND: Previous investigations have shown that local application of nanoparticles presenting the carbohydrate moiety galactose-α-1,3-galactose (α-gal epitopes) enhance wound healing by activating the complement system and recruiting pro-healing macrophages to the injury site. Our companion in vitro paper suggest α-gal epitopes can similarly recruit and polarize human microglia toward a pro-healing phenotype. In this continuation study, we investigate the in vivo implications of α-gal nanoparticle administration directly to the injured spinal cord. METHODS: α-Gal knock-out (KO) mice subjected to spinal cord crush were injected either with saline (control) or with α-gal nanoparticles immediately following injury. Animals were assessed longitudinally with neurobehavioral and histological endpoints. RESULTS: Mice injected with α-gal nanoparticles showed increased recruitment of anti-inflammatory macrophages to the injection site in conjunction with increased production of anti-inflammatory markers and a reduction in apoptosis. Further, the treated group showed increased axonal infiltration into the lesion, a reduction in reactive astrocyte populations and increased angiogenesis. These results translated into improved sensorimotor metrics versus the control group. CONCLUSIONS: Application of α-gal nanoparticles after spinal cord injury (SCI) induces a pro-healing inflammatory response resulting in neuroprotection, improved axonal ingrowth into the lesion and enhanced sensorimotor recovery. The data shows α-gal nanoparticles may be a promising avenue for further study in CNS trauma.

Rapid Self-Healing Hydrogel with Ultralow Electrical Hysteresis for Wearable Sensing.

Self-healing hydrogels are in high demand for wearable sensing applications due to their remarkable deformability, high ionic and electrical conductivity, self-adhesiveness to human skin, as well as resilience to both mechanical and electrical damage. However, these hydrogels face challenges such as delayed healing times and unavoidable electrical hysteresis, which limit their practical effectiveness. Here, we introduce a self-healing hydrogel that exhibits exceptionally rapid healing with a recovery time of less than 0.12 s and an ultralow electrical hysteresis of less than 0.64% under cyclic strains of up to 500%. This hydrogel strikes an ideal balance, without notable trade-offs, between properties such as softness, deformability, ionic and electrical conductivity, self-adhesiveness, response and recovery times, durability, overshoot behavior, and resistance to nonaxial deformations such as twisting, bending, and pressing. Owing to this unique combination of features, the hydrogel is highly suitable for long-term, durable use in wearable sensing applications, including monitoring body movements and electrophysiological activities on the skin.

α-Gal Nanoparticles in CNS Trauma: I. In Vitro Activation of Microglia Towards a Pro-Healing State.

BACKGROUND: Macrophages and microglia play critical roles after spinal cord injury (SCI), with the pro-healing, anti-inflammatory (M2) subtype being implicated in tissue repair. We hypothesize that promoting this phenotype within the post-injured cord microenvironment may provide beneficial effects for mitigating tissue damage. As a proof of concept, we propose the use of nanoparticles incorporating the carbohydrate antigen, galactose-α-1,3-galactose (α-gal epitope) as an immunomodulator to transition human microglia (HMC3) cells toward a pro-healing state. METHODS: Quiescent HMC3 cells were acutely exposed to α-gal nanoparticles in the presence of human serum and subsequently characterized for changes in cell shape, expression of anti or pro-inflammatory markers, and secretion of phenotype-specific cytokines. RESULTS: HMC3 cells treated with serum activated α-gal nanoparticles exhibited rapid enlargement and shape change in addition to expressing CD68. Moreover, these activated cells showed increased expression of anti-inflammatory markers like Arginase-1 and CD206 without increasing production of pro-inflammatory cytokines TNF-α or IL-6. CONCLUSION: This study is the first to show that resting human microglia exposed to a complex of α-gal nanoparticles and anti-Gal (from human serum) can be activated and polarized toward a putative M2 state. The data suggests that α-gal nanoparticles may have therapeutic relevance to the CNS microenvironment, in both recruiting and polarizing macrophages/microglia at the application site. The immunomodulatory activity of these α-gal nanoparticles post-SCI is further described in the companion work (Part II).

Mechanical damage in porcine dermis: Micro-mechanical model and experimental characterization.

Skin is subjected to extreme mechanical loading during needle insertion and drug delivery to the subcutaneous space. There is a rich literature on the characterization of porcine skin biomechanics as the preeminent animal model for human skin, but the emphasis has been on the elastic response and specific anatomical locations such as the dorsal and the ventral regions. During drug delivery, however, energy dissipation in the form of damage, softening, and fracture, is expected. Similarly, reports on experimental characterization are complemented by modeling efforts, but with similar gaps in microstructure-driven modeling of dissipative mechanisms. Here we contribute to the bridging of these gaps by testing porcine skin from belly and breast regions, in two different orientation with respect to anatomical axes, and to progressively higher stretches in order to show damage accumulation and stiffness degradation. We complement the mechanical test with imaging of the collagen structure and a micro-mechanics modeling framework. We found that skin from the belly is stiffer with respect to the breast region when comparing the calf stiffness of the J-shaped stress-stretch response observed in most collagenous tissues. No significant direction dependent properties were found in either anatomical location. Both locations showed energy dissipation due to damage, evident though a softening of the stress-stretch response. The microstructure model was able to capture the elastic and damage progression with a small set of parameters, some of which were determined directly from imaging. We anticipate that data and model fits can help in predictive simulations for device design in situations where skin is subject to supra-physiological deformation such as in subcutaneous drug delivery.

Patient-specific, mechanistic models of tumor growth incorporating artificial intelligence and big data.

Despite the remarkable advances in cancer diagnosis, treatment, and management that have occurred over the past decade, malignant tumors remain a major public health problem. Further progress in combating cancer may be enabled by personalizing the delivery of therapies according to the predicted response for each individual patient. The design of personalized therapies requires patient-specific information integrated into an appropriate mathematical model of tumor response. A fundamental barrier to realizing this paradigm is the current lack of a rigorous, yet practical, mathematical theory of tumor initiation, development, invasion, and response to therapy. In this review, we begin by providing an overview of different approaches to modeling tumor growth and treatment, including mechanistic as well as data-driven models based on ``big data" and artificial intelligence. Next, we present illustrative examples of mathematical models manifesting their utility and discussing the limitations of stand-alone mechanistic and data-driven models. We further discuss the potential of mechanistic models for not only predicting, but also optimizing response to therapy on a patient-specific basis. We then discuss current efforts and future possibilities to integrate mechanistic and data-driven models. We conclude by proposing five fundamental challenges that must be addressed to fully realize personalized care for cancer patients driven by computational models.

Interpenetrating Networks of Collagen and Hyaluronic Acid That Serve as In Vitro Tissue Models for Assessing Macromolecular Transport.

High-fidelity preclinical in vitro tissue models can reduce the failure rate of drugs entering clinical trials. Collagen and hyaluronic acid (HA) are major components of the extracellular matrix of many native tissues and affect therapeutic macromolecule diffusion and recovery through tissues. Although collagen and HA are commonly used in tissue engineering, the physical and mechanical properties of these materials are variable and depend highly on processing conditions. In this study, HA was chemically modified and crosslinked via hydrazone bonds to form interpenetrating networks of crosslinked HA (HAX) with collagen (Col). These networks enabled a wide range of mechanical properties, including stiffness and swellability, and microstructures, such as pore morphology and size, that can better recapitulate diverse tissues. We utilized these interpenetrating ColHAX hydrogels as in vitro tissue models to examine macromolecular transport and recovery for early-stage drug screening. Hydrogel formulations with varying collagen and HAX concentrations imparted different gel properties based on the ratio of collagen to HAX. These gels were stable and swelled up to 170% of their original mass, and the storage moduli of the ColHAX gels increased over an order of magnitude by increasing collagen and HA concentration. Interestingly, when HAX concentration was constant and collagen concentration increased, both the pore size and spatial colocalization of collagen and HA increased. HA in the system dominated the ζ-potentials of the gels. The hydrogel and macromolecule properties impacted the mass transport and recovery of lysozyme, ß-lactoglobulin, and bovine serum albumin (BSA) from the ColHAX gels─large molecules were largely impacted by mesh size, whereas small molecules were influenced primarily by electrostatic forces. Overall, the tunable properties demonstrated by the ColHAX hydrogels can be used to mimic different tissues for early-stage assays to understand drug transport and its relationship to matrix properties.

Damage and Fracture Mechanics of Porcine Subcutaneous Tissue Under Tensile Loading.

Subcutaneous injection, which is a preferred delivery method for many drugs, causes deformation, damage, and fracture of the subcutaneous tissue. Yet, experimental data and constitutive modeling of these dissipation mechanisms in subcutaneous tissue remain limited. Here we show that subcutaneous tissue from the belly and breast anatomical regions in the swine show nonlinear stress-strain response with the characteristic J-shaped behavior of collagenous tissue. Additionally, subcutaneous tissue experiences damage, defined as a decrease in the strain energy capacity, as a function of the previously experienced maximum deformation. The elastic and damage response of the tissue are accurately described by a microstructure-driven constitutive model that relies on the convolution of a neo-Hookean material of individual fibers with a fiber orientation distribution and a fiber recruitment distribution. The model fit revealed that subcutaneous tissue can be treated as initially isotropic, and that changes in the fiber recruitment distribution with loading are enough to explain the dissipation of energy due to damage. When tested until failure, subcutaneous tissue that has undergone damage fails at the same peak stress as virgin samples, but at a much larger stretch, overall increasing the tissue toughness. Together with a finite element implementation, these data and constitutive model may enable improved drug delivery strategies and other applications for which subcutaneous tissue biomechanics are relevant.

Mechanistic Computational Modeling of Implantable, Bioresorbable Drug Release Systems.

Implantable, bioresorbable drug delivery systems offer an alternative to current drug administration techniques; allowing for patient-tailored drug dosage, while also increasing patient compliance. Mechanistic mathematical modeling allows for the acceleration of the design of the release systems, and for prediction of physical anomalies that are not intuitive and may otherwise elude discovery. This study investigates short-term drug release as a function of water-mediated polymer phase inversion into a solid depot within hours to days, as well as long-term hydrolysis-mediated degradation and erosion of the implant over the next few weeks. Finite difference methods are used to model spatial and temporal changes in polymer phase inversion, solidification, and hydrolysis. Modeling reveals the impact of non-uniform drug distribution, production and transport of H+ ions, and localized polymer degradation on the diffusion of water, drug, and hydrolyzed polymer byproducts. Compared to experimental data, the computational model accurately predicts the drug release during the solidification of implants over days and drug release profiles over weeks from microspheres and implants. This work offers new insight into the impact of various parameters on drug release profiles, and is a new tool to accelerate the design process for release systems to meet a patient specific clinical need.

Transcriptomic alterations in hypertrophy of the ligamentum flavum: interactions of Rho GTPases, RTK, PIK3, and FGF.

PURPOSE: To analyze the differential transcriptome expression in hypertrophic ligaments flavum (HLF) compared to normal ligaments. METHODS: A case-control study was conducted that included 15 patients with hypertrophy of LF and 15 controls. Samples of LF were obtained through a lumbar laminectomy and analyzed by DNA microarrays and histology. The dysregulated biological processes, signaling pathways, and pathological markers in the HLF were identified using bioinformatics tools. RESULTS: The HLF had notable histological alterations, including hyalinosis, leukocyte infiltration, and disarrangement of collagen fibers. Transcriptomic analysis showed that up-regulated genes were associated with the signaling pathways of Rho GTPases, receptor tyrosine kinases (RTK), fibroblast growth factors (FGF), WNT, vascular endothelial growth factor, phosphoinositide 3-kinase (PIK3), mitogen-activated protein kinases, and immune system. The genes PIK3R1, RHOA, RPS27A, CDC42, VAV1, and FGF5, 9, 18, and 19 were highlighted as crucial markers in HLF. The down-expressed genes in the HLF had associations with the metabolism of RNA and proteins. CONCLUSION: Our results suggest that abnormal processes in hypertrophied LF are mediated by the interaction of the Rho GTPase, RTK, and PI3K pathways, which have not been previously described in the HLF, but for which there are currently therapeutic proposals. More studies are required to confirm the therapeutic potential of the pathways and mediators described in our results.

Investigation of macromolecular transport through tunable collagen hyaluronic acid matrices.

Therapeutic macromolecules possess properties such as size and electrostatic charge that will dictate their transport through subcutaneous (SC) tissue and ultimate bioavailability and efficacy. To improve therapeutic design, platforms that systematically measure the transport of macromolecules as a function of both drug and tissue properties are needed. We utilize a Transwell chamber with tunable collagen-hyaluronic acid (ColHA) hydrogels as an in vitro model to determine mass transport of macromolecules using non-invasive UV spectroscopy. Increasing hyaluronic acid (HA) concentration from 0 to 2 mg/mL within collagen gels decreases the mass transport of five macromolecules independent of size and charge and results in a maximum decrease in recovery of 23.3% in the case of bovine immunoglobulin G (IgG). However, in a pure 10 mg/mL HA solution, negatively-charged macromolecules bovine serum albumin (BSA), ß-lactoglobulin (BLg), dextran (Dex), and IgG had drastically increased recovery by 20-40% compared to their performance in ColHA matrices. This result was different from the positively-charged macromolecule Lysozyme (Lys), which, despite its small size, showed reduced recovery by 3% in pure HA. These results demonstrate two distinct regimes of mass transport within our tissue model. In the presence of both collagen and HA, increasing HA concentrations decrease mass transport; however, in the absence of collagen, the high negative charge of HA sequesters and increases residence time of positively-charged macromolecules and decreases residence time of negatively-charged macromolecules. Through our approach, ColHA hydrogels serve as a platform for the systematic evaluation of therapeutic macromolecule transport as a function of molecular characteristics.

Automated Layer Identification Method for Skin Tissue Histology Images.

We present a novel automated tissue layer identification method for histology images. The method requires a single user input: the number of layers to be identified. The method incorporates a coarse boundary identification step followed by a refinement step. The coarse identification segments the image into 125 × 125 pixel sub-tiles, computes the histogram of each sub-tile, implements K-means clustering to label each sub-tile, and uses Dijkstra's algorithm to form the layer boundary. The refinement step identifies hair follicles, improves the detail and accuracy of the boundary, and segments the epidermis. The method only uses one color channel (blue). We test our proposed method using eight excised porcine tissue samples taken at different anatomical locations. The layer segmentations demonstrated that the dermis thickness increased, and the subcutaneous thickness decreased moving from breast to belly. Minimal variation in the thickness of the epidermis layer across anatomical locations was observed. Overall, these results highlight the importance of quantifying and assessing the tissue environment. Moreover, we demonstrate that our proposed method was robust across different histology stains and did not depend on color-specific information.

Basic Salt Additives Modulate the Acidic Microenvironment Around In Situ Forming Implants.

There is a growing number of protein drugs, yet their limited oral bioavailability requires that patients receive frequent, high dose injections. In situ forming implants (ISFIs) for controlled release of biotherapeutics have the potential to greatly reduce the injection frequency and improve patient compliance. However, protein release from ISFIs is a challenge due to their proclivity for instability. Specifically, factors such as the acidic microclimate within ISFIs can lead to protein aggregation and denaturation. Basic salts have been shown in PLGA microparticle and microcylinder formulations to significantly reduce protein instability by neutralizing this acidic environment. The overall objective of the study was to demonstrate that basic salts can be used with an ISFI system to neutralize the implant acidification. To this end, the basic salts MgCO3 and Mg(OH)2 were added to a protein-releasing ISFI and the effect on drug release, pH, implant swelling, implant diffusivity, and implant erosion were evaluated. Either salt added at 3 wt% neutralized the acidic environment surrounding the implants, keeping the pH at 6.64 ± 0.03 (MgCO3) and 6.46 ± 0.11 (Mg(OH)2) after 28 day compared to 3.72 ± 0.05 with no salts added. The salts initially increased solution uptake into the implants but delayed implant degradation and erosion. The 3 wt% Mg(OH)2 formulation also showed slightly improved drug release with a lower burst and increased slope. We showed that salt additives can be an effective way to modulate the pH in the ISFI environment, which can improve protein stability and ultimately improve the capacity of ISFIs for delivering pH-sensitive biomolecules. Such a platform represents a low-cost method of improving overall patient compliance and reducing the overall healthcare burden.

Multi-feature-Based Robust Cell Tracking.

Cell tracking algorithms have been used to extract cell counts and motility information from time-lapse images of migrating cells. However, these algorithms often fail when the collected images have cells with spatially and temporally varying features, such as morphology, position, and signal-to-noise ratio. Consequently, state-of-the-art algorithms are not robust or reliable because they require manual inputs to overcome the cell feature changes. To address these issues, we present a fully automated, adaptive, and robust feature-based cell tracking algorithm for the accurate detection and tracking of cells in time-lapse images. Our algorithm tackles measurement limitations twofold. First, we use Hessian filtering and adaptive thresholding to detect the cells in images, overcoming spatial feature variations among the existing cells without manually changing the input thresholds. Second, cell feature parameters are measured, including position, diameter, mean intensity, area, and orientation, and these parameters are simultaneously used to accurately track the cells between subsequent frames, even under poor temporal resolution. Our technique achieved a minimum of 92% detection and tracking accuracy, compared to 16% from Mosaic and Trackmate. Our improved method allows for extended tracking and characterization of heterogeneous cell behavior that are of particular interest for intravital imaging users.

PLA-PCL microsphere formulation to deter abuse of prescription opioids by smoking.

Opioids are commonly prescribed across the United States (US) for pain relief, despite their highly addictive nature that often leads to abuse and overdose deaths. Abuse deterrent formulations (ADFs) for prescription opioids make the non-therapeutic use of these drugs more difficult and less satisfying. Although approximately one-third of surveyed abusers in the US reported smoking opioids, to our knowledge, no commercialized ADF effectively prevents opioid smoking. Here, we report a novel approach to deter smoking of a model prescription opioid drug, thebaine (THB), by using polymer blend microspheres (MS) comprising polylactic acid (PLA) and polycaprolactone (PCL). We utilized high-performance liquid chromatography (HPLC) and thermogravimetric analysis (TGA) to test the ability of PLA-PCL MS to limit the escape of vaporized THB. Additionally, we compared the abuse-deterrent potential of PLA-PCL MS to that of activated carbon (AC) and mesoporous silica (MPS), two materials with excellent drug-adsorbing properties. Our MS formulation was effective in reducing the amount of both active drug and thermal degradation products in the vapor generated upon heating of THB. These results support that PLA-PCL microspheres can be co-formulated in a tablet with common prescription opioids to deter their abuse via the smoking route.