Pharmaceutical Particles Research Articles

Low-cost, autonomous microscopy using deep learning and robotics: A crystal morphology case study

Data availability is a major obstacle to the successful application of artificial intelligence. In chemistry, pharmaceutical and materials science, generating labelled data on scale relies on expensive screening approaches that require major human and/or economic investments. Although industry uses sophisticated autonomous systems for such tasks, tight intellectual property concerns and strict confidentiality agreements limit data accessibility, thus restricting innovation and hindering the concept of ”open science”. Without automation, data collection is laborious and sample characterization often suffers from subjectivity introduced by the operator. This study tackles the challenge of pharmaceutical particle characterization by modifying a 3D printer to enable rapid, autonomous sample characterization using light microscopy. The system uses low-cost hardware and open-source software in a platform that is accessible and reproducible for the scientific community. As well as increasing throughput, this system uses deep learning to assign sample labels in real-time overcoming bias from subjective labelling, which hinders model performance. Different neural architectures were applied, including convolution and attention blocks, to maximize performance and demonstrate the transferability of the system. This autonomous system effectively characterized pharmaceutical crystal morphology, an area where subjective labelling and limited data have hindered applying machine learning models. By releasing this platform, researchers without access to sophisticated automation platforms can carry out larger in-house screening efforts to generate datasets for developing data-driven, intelligent models.

Controlling Microparticle Morphology in Melt-Jet Printing of Active Pharmaceutical Ingredients through Surface Phenomena.

Achieving homogeneity and reproducibility in the size, shape, and morphology of active pharmaceutical ingredient (API) particles is crucial for their successful manufacturing and performance. Herein, we describe a new method for API particle engineering using melt-jet printing technology as an alternative to the current solvent-based particle engineering methods. Paracetamol, a widely used API, was melted and jetted as droplets onto various surfaces to solidify and form microparticles. The influence of different surfaces (glass, aluminum, polytetrafluoroethylene, and polyethylene) on particle shape was investigated, revealing a correlation between substrate properties (heat conduction, surface energy, and roughness) and particle sphericity. Higher thermal conductivity, surface roughness, and decreased surface energy contributed to larger contact angles and increased sphericity, reaching a near-perfect micro-spherical shape on an aluminum substrate. The integrity and polymorphic form of the printed particles were confirmed through differential scanning calorimetry and X-ray diffraction. Additionally, high-performance liquid chromatography analysis revealed minimal degradation products. The applicability of the printing process to other APIs was demonstrated by printing carbamazepine and indomethacin on aluminum surfaces, resulting in spherical microparticles. This study emphasizes the potential of melt-jet printing as a promising approach for the precise engineering of pharmaceutical particles, enabling effective control over their physiochemical properties.

Digital process design to define and deliver pharmaceutical particle attributes

A digital-first approach to produce quality particles of an active pharmaceutical ingredient across crystallisation, washing and drying is presented, minimising material requirements and experimental burden during development. To demonstrate current predictive modelling capabilities, the production of two particle sizes (D90 = 42 and 120 µm) via crystallisation was targeted to deliver a predicted, measurable difference in in vitro dissolution performance. A parameterised population balance model considering primary nucleation, secondary nucleation and crystal growth was used to select the modes of production for the different particle size batches. Solubility prediction aided solvent selection steps which also considered manufacturability and safety selection criteria. A wet milling model was parameterised and used to successfully produce a 90 g product batch with a particle size D90 of 49.3 µm, which was then used as the seeds for cooling crystallisation. A rigorous approach to minimising physical phenomena observed experimentally was implemented, and successfully predicted the required conditions to produce material satisfying the particle size design objective of D90 of 120 µm in a seeded cooling crystallisation using a 5-stage MSMPR cascade. Product material was isolated using the filtration and washing processes designed, producing 71.2 g of agglomerated product with a primary particle D90 of 128 µm. Based on experimental observations, the population balance model was reparametrised to increase accuracy by inclusion of an agglomeration term for the continuous cooling crystallisation. The dissolution performance for the two crystallised products is also demonstrated, and after 45 min 104.0 mg of the D90 of 49.3 µm material had dissolved, compared with 90.5 mg of the agglomerated material with D90 of 128 µm. Overall, 1513 g of the model compound was used to develop and demonstrate two laboratory scale manufacturing processes with specific particle size targets. This work highlights the challenges associated with a digital-first approach and limitations in current first-principles models are discussed that include dealing ab initio with encrustation, fouling or factors that affect dissolution other than particle size.

Computational investigation of flow characteristics and particle deposition patterns in a realistic human airway model under different breathing conditions

Airborne particle pollution causes a range of respiratory and cardiovascular disorders by entering the human respiratory system through the breathing process. The administration of pharmaceutical particles by inhalation is another effective way to treat pulmonary illnesses. Studying particle deposition in the respiratory system during human breathing is crucial to maintaining human health. This necessity served as the impetus for this work, which aims to investigate how the airflow and particles' deposition are influenced by constant inhalation and circulatory breathing, particle diameter, and changes in airflow rate. The focus of this paper is to compare the particle deposition results of circulatory respiration with constant respiration. Based on computed tomography (CT) scan pictures, a precise human airway model from the mouth cavity to the fifth-generation bronchi was created. Flow fields and particle deposition inside the respiratory tract were examined at varied breathing rates (30, 60, and 90 L/min of constant and circulatory breathing) and varying haled particle sizes (5 and 10 µm). The results showed that the oropharyngeal area is often where the majority of particles are deposited. The particle distribution fraction is more significant in the bronchial area than the oropharyngeal region due to lower inhalation velocities and smaller particle sizes. For particles with a diameter of 5 µm, constant respiration and circulatory respiration have virtually identical particle distribution fractions in each region. For particles with a diameter of 10 µm, the particle distribution fraction for circulatory respiration is slightly higher than for constant respiration in the bronchial region as the flow rate increases. For both constant and circulatory respiration, particles are deposited more in the right lung and less in the left. These results contribute to further research on respiratory diseases caused by inhaled particles and guide inhalation therapy for better treatment outcomes.

AI-based analysis of in-line process endoscope images for real-time particle size measurement in a continuous pharmaceutical milling process

This paper presents a case study on the first in-line application of AI-based image analysis for real-time pharmaceutical particle size measurement in a continuous milling process. An AI-based imaging system, which utilises a rigid endoscope, was tested for the real-time particle size measurement of solid NaCl powder used as a model API in the range of 200–1000 µm. After creating a dataset containing annotated images of NaCl particles, it was used to train an AI model for detecting particles and measuring their size. The developed system could analyse overlapping particles without dispersing air, thus broadening its applicability. The performance of the system was evaluated by measuring pre-sifted NaCl samples with the imaging tool, after which it was installed into a continuous mill for in-line particle size measurement of a milling process. By analysing ∼100 particles/s, the system was able to accurately measure the particle size of sifted NaCl samples and detect particle size reduction when applied in the milling process. The Dv50 values and PSDs measured real-time with the AI-based system correlated well with the reference laser diffraction measurements (<6% mean absolute difference over the measured samples). The AI-based imaging system shows great potential for in-line particle size analysis, which, in line with the latest pharmaceutical QC trends, can provide valuable information for process development and control.

A spherical harmonic-random field coupled method for efficient reconstruction of CT-image based 3D aggregates with controllable multiscale morphology

The physical and mechanical performance of concrete is inherently dependent on aggregate morphology including the general shape at coarse-scale, the local roundness at medium-scale and the surface texture at fine-scale. This study develops a computational method for highly efficient generation of realistic 3D aggregates using micro X-ray Computed Tomography (μXCT) images, the spherical harmonic (SH) analysis and a random-field reconstruction algorithm. In this method, the real aggregate surface segmented from CT images is first decomposed and mapped by the SH function to radius fields at different scales, whose statistical data are then extracted to reconstruct new aggregates through a spectral representation method and an innovative folding algorithm. The proposed method is verified by comparison of four morphology indices (sphericity, convexity, roundness and roughness) of the real aggregate and the new ones, and its flexibility in reconstructing new sets of random aggregates with specified morphology indices at selected scales is also demonstrated. As an application, eight sets of Fuller-grade random aggregates with target morphology indices are constructed and packed into a cube container to generate digital concrete specimens. The aggregate sphericity and convexity are found to be linearly correlated with the maximum aggregate volume fraction that can be packed into the cube. The developed method can be also applied to other granular materials, such as pharmaceutical particles, colloids, ceramics, soils and coal.

Spatial aerosol deposition correlated to anatomic feature development in 6-year-old upper airway computational models

The upper airways of children undergo developmental changes around age 6, yielding differences between adult and pediatric anatomies. These differences include the cricoid ring area shape, the location of narrowest constriction, and the angle of the epiglottis, all of which are expected to alter local fluid dynamic profiles and subsequent upper airway deposition and downstream aerosol delivery of inhaled therapeutics. In this work, we quantify “pediatric”-like and “adult”-like geometric and fluid dynamic features of two computed tomography (CT)-scan derived models of 6-year-old upper airways in healthy subjects and compare to an idealized model. The two CT-scan models had a mixture of “adult”- and “pediatric”-like anatomic features, with Subject B exhibiting more “pediatric”-like features than Subject A, while the idealized model exhibited entirely “adult”-like features. By computational fluid-particle dynamics, these differences in anatomical features yielded distinct local fluid profiles with altered aerosol deposition between models. Notably, the idealized model better predicted deposition characteristics of Subject A, the more “adult”-like model, including the relationship between the impaction parameter, dp2Q and the fraction of deposition across a range of flow rates and particle diameters, as well as deposition of an approximate pharmaceutical particle size distribution model. Our results with even this limited dataset suggest that there are key personalized metrics that are influenced by anatomical development, which should be considered when developing pediatric inhalable therapeutics. Quantifying anatomical development and correlating to aerosol deposition has the potential for high-throughput developmental characterization and informing desired aerosol characteristics for pediatric applications.

Experimental analysis and thermodynamic modelling of lenalidomide solubility in supercritical carbon dioxide

• Lenalidomide (LND) solubility in supercritical CO 2 (sc-CO 2 ) was measured for the first time. • Solubility data were correlated by different density-based models. • Experimental data was also correlated by Modeified Wilson and SRK models. • The accuracy of the models was evaluated by statistical criteria (AARD %), ( Radj ). • Thermodynamic enthalpies of Lenalidomide were estimated. Reduction of the size of pharmaceutical particles to micro/nano scale is an approved strategy to enhance their dissolution rate at decremented side effects. Production of drug micro/nanoparticles via a supercritical carbon dioxide (sc-CO 2 )- based process requires accurate data on the drug solubility in sc-CO 2 . In this study, the solubility of Lenalidomide (LND), an anti-cancer, was experimentally determined in sc-CO 2 at various temperatures (308–338 K) and pressures (120–300 bar). Furthermore, the obtained solubility data were correlated by different theoretical methods. LND solubility, in terms of mole fraction, was obtained in the range of 0.02 × 10 −4 –1.08 × 10 −4 depending on the conditions. The empirical models with different adjustable parameters, Soave–Redlich–Kwong equation of state (SRK-EoS) with two parameters van der Waals (vdW2) mixing rule, and expanded liquid theory (modified Wilson’s model) were applied for experimental data correlation. According to the results, modified Wilson’s model can correlate LND solubility in sc-CO 2 with high accuracy. However, the SRK-EoS did not show acceptable accuracy in the correlation of LND solubility in sc-CO 2 . Among the empirical models, Alwi and Garlapati, Sung and Shim, Ch and Madras, Hozhabr et al. , Garlapati and Madras, Keshmiri et al. , Sparks et al. , Reddy and Garlapati, Bian et al. (2011) and Belghait et al. showed R adj values higher than 0.99 and produced the best correlation with 3,4, 5, 6 and 8 adjustable parameters, respectively. Moreover, the approximate values of total mixing heat ( Δ H total ) as well as vaporization ( Δ H vap ), and solvation ( Δ H sol ) enthalpies were computed.

Defect and satellite characteristics of additive manufacturing metal powders

Metal additive manufacturing (AM) requires high-quality metal powders to three-dimensionally (3D) print metallic components with complex and customizable geometries. The lack of quantification of AM metal powders creates quality control challenges for 3D printed components, increases the uncertainty of printing reliability and net cost of inspected and certified printed components, and reduces the recyclability of used powders. However, critical characteristics of AM metal powders that are decisive factors for the 3D printing process, such as internal porosity, contamination, and satellite feature, remain ambiguous. Here, we developed a novel approach to 3D quantify key characteristics of AM metal powders down to individual particles by using high-resolution synchrotron x-ray computed tomography. Empowered by the penetrative capability of high-energy x-ray, internal porosity and contamination within as-atomized metal powders from high-entropy alloys to nickel-based superalloys were evaluated. Additionally, the newly-developed dispersion method enables the homogeneous separation of individual particles, and consequently, results in the implementation of 3D particle shape analysis. To resolve a major challenge of identification and quantification of satellite-feature particles in as-atomized AM metal powders, the satellite features were quantitated by modeling and analyzing the shape parameter of local thickness variance. Furthermore, the 3D analytical methods of particle assessment in this study can be applied to other materials systems like rock, food, and pharmaceutical particles, and provide insights for process optimization across powder metallurgy, concrete, food and pharmaceutical manufacturing, and AM industries.

Change in spray behavior of pharmaceutical particles by creating bypass hole in dry powder inhaler

This study presents the effects of adding bypass holes in a dry powder inhaler on aerosol particle characteristics. A commercially available inhaler device was modified with bypass holes, and the resulting drug powder spray characteristics were investigated in terms of how the direction, size, and position of the bypass holes in the inhaler passage affect the spray output to find the optimal design. All three factors influenced the aerosol behavior, and in particular, it is most significantly affected by the direction of bypass hole faces. Adding bypass holes to the inhaler passage was found to decrease the mean penetration length of the drug particles by 12%; however, at the same time, the spray angle is increased by 35%, which leads to a significant decrease in forward particle momentum.

Newly designed mouthpiece to improve spray characteristics of pharmaceutical particles in dry powder inhaler

This study presents the effects of mouthpiece designs in carrier-based dry powder inhalation on aerosol particle characteristics. Simulation models are developed to investigate drug powder spray characteristics from mouthpieces of four different-shaped flow channels. The simulation results agree well with the results of two experiments, one involving particle image velocimetry and the other the Anderson cascade impactor. A mouthpiece with a spiral channel and mixing chamber is discovered to be effective for low axial-flow velocities in the forward direction and yielded a more uniform dispersion of particles in the radial direction as the drug particles are ejected through the mouthpiece. This results in more effective delivery of the drug particles to the peripheral target regions of the respiratory system.

Grey-box model calibration and validation for a continuous horizontal fluidized bed dryer

Fully continuous drying can improve the production flexibility and throughput of tablets manufactured by wet granulation lines. A recently developed prototype incorporates a screw conveyor to a horizontal fluidized bed dryer for pharmaceutical particles. Research on advanced process control strategies led to the publication of a two-phase tanks-in-series framework to describe its dynamics. This paper calibrates the model to replicate the prototype behavior. A grey-box identification methodology estimates the empirical and uncertain parameters using experimental input-output data. The underlying assumptions are cross-validated with time-domain comparisons, output root mean square errors and confidence intervals. Results show that the model reproduces the observed trends in general and the discrepancies result from mechanical issues on the equipment.

Particle Synthesis: Continuous Embedded Droplet Printing in Yield‐Stress Fluids for Pharmaceutical Drug Particle Synthesis (Adv. Mater. Technol. 4/2021)

In article number 2001245, Saif A. Khan, Patrick S. Doyle, and co-workers develop a platform that enables the continuous production of droplets embedded in yield-stress fluids. This environment allows for the production of highly flowable pharmaceutical particles, which are persistently difficult to achieve conventionally. Similar platforms are envisioned to enable precise, rapid, customized, and distributed drug particles with superior attributes.

Continuous Embedded Droplet Printing in Yield‐Stress Fluids for Pharmaceutical Drug Particle Synthesis

AbstractEmbedded droplet printing is a recent mode of generating and processing droplets within yield‐stress fluids. This technique has shown promise for performing sensitive processes like pharmaceutical crystallization, as well as chemical synthesis and biological experimentation due to the unique ability to process droplets that are quiescently suspended. Despite improving on conventional microfluidic technologies in numerous ways, current embedded droplet printing methods are limited to batch processes, severely hampering their overall utility. A new platform that enables continuous production of embedded droplets is presented and characterized. This platform expands the capabilities of embedded droplet printing and allows for its application to areas of continuous materials discovery, screening, and manufacturing. Here, the platform is used for the rapid production of pharmaceutical particles that are highly spherical and uniform, key targets for flowability, and ultimately manufacturability of pharmaceutical drug products. The presented platform achieves a maximum throughput of over 100 g per day, enabling characterization of the superior powder flow properties. The available operating space of this platform is demonstrated for an antisolvent crystallization process with an anti‐malarial drug. This understanding provides design guidelines for how similar platforms may be engineered for precise, rapid, customized, and distributed manufacturing of drug particles with superior flowability.

Utilization of supercritical CO2 gas antisolvent (GAS) for production of Capecitabine nanoparticles as anti-cancer drug: Analysis and optimization of the process conditions

Reducing the size of pharmaceutical particles is an approved way to enhance solubility and bioavailability of a drug, leading to a decrease in the drug dose and its side effects. In this research, nanoparticles of Capecitabine (CPT), an anti-cancer drug, were produced using the gas anti-solvent supercritical (GAS) process. Also, Box–Behnken design (BBD) method was applied to optimize the process condition. Accordingly, the GAS process was performed at different pressures of 120, 140 and 160 bar, temperatures of 308, 318 and 328 K, and solute concentrations of 15, 45 and 75 mg/mL. Based on the size of the precipitated CPT particles, the optimum condition was specified as the pressure of 160 bar, the temperature of 318 K and the initial CPT concentration of 45 mg/mL. Physical identification analysis including differential scanning calorimetry (DSC), X-ray diffraction (XRD), scanning electron microscopy (SEM), dynamic light scattering (DLS), Fourier transform infrared spectrometry (FTIR) and high pressure liquid chromatography (HPLC) analyses were employed to characterize the produced CPT nanoparticles. Significant size reduction of drug particles from 25 μm on average (original) to about 243.3 nm and uniform structure of obtained nanoparticles with a narrow size distribution indicate the effectiveness of the GAS process. Results of the physical analysis confirm the purity of CPT structure with no changes during the GAS process. Also, according to the results of DSC and XRD, the crystallinity of CPT nanoparticles was lower than that of the original sample, leading to higher solubility. Improved CPT solubility was also confirmed by comparison of the dissolution rate of the original and obtained CPT samples in optimum conditions.

Computational analysis of triboelectrification due to aerodynamic powder dispersion

Triboelectric charging can strongly influence bulk powder flow behaviour, and hence its characterization is of great interest for safe manufacturing operations. In a recent development, the use of an aerodynamic disperser, employing a pressure pulse to disperse a small powder quantity, shows a great potential for inducing triboelectric charge transfer. We analyse this process by coupled Discrete Element Method and Computational Fluid Dynamics (DEM-CFD) simulations, incorporating triboelectric charge transfer. The simulations are based on property data of glass ballotini as model particles, together with those of α-lactose monohydrate (α-LM) and aspirin, as powders of practical interest. The characteristics of particle-particle and particle-wall collisions are analysed in detail. The analysis shows that pharmaceutical particles charge significantly more than glass ballotini. The charge-to-surface area ratio is remarkably constant and close to its equilibrium value for each test material. Overall, the analysis provides a great insight on the triboelectric charging by aerodynamic dispersion.

Characterization of Drug Delivery Particles in Pharmaceutical Disperse Systems: A Review

This review focuses on micromerties which is the science and technology of small particles. Pharmaceutical particle and powder characterization is key to pharmaceutical manufacturing processes and products as they often involve bulk powders in pharmaceutical dispersed systems. Particle size and powder properties will influence physical properties, such as the flowability of a powder, and in turn, will impact processing, blending, dissolution, delivery and bioavailability. Moreover, the appropriate selection of methods and particle engineering technologies of their production, characterization, are critical issues need considerations. Furthermore, optimization of their production process along with the drug/excipient particles, can provide insight understanding of their in vivo behaviour. In particular, characterization techniques for particle size distribution and morphology, drug loading structure of matrix components, porosity, particle surface chemistry, surface charge, are discussed in this review.

Quantification of swelling characteristics of pharmaceutical particles.

Particle swelling is a crucial component in the disintegration of a pharmaceutical tablet. The swelling of particles in a tablet creates stress inside the tablet and thereby pushes apart adjoining particles, eventually causing the tablet to break-up. This work focused on quantifying the swelling of single particles to identify the swelling-limited mechanisms in a particle, i.e. diffusion- or absorption capacity-limited. This was studied for three different disintegrants (sodium starch glycolate/SSG, croscarmellose sodium/CCS, and low-substituted hydroxypropyl cellulose/L-HPC) and five grades of microcrystalline cellulose (MCC) using an optical microscope coupled with a bespoke flow cell and utilising a single particle swelling model. Fundamental swelling characteristics, such as diffusion coefficient, maximum liquid absorption ratio and swelling capacity (maximum swelling of a particle) were determined for each material. The results clearly highlighted the different swelling behaviour for the various materials, where CCS has the highest diffusion coefficient with 739.70 μm2/s and SSG has the highest maximum absorption ratio of 10.04g/g. For the disintegrants, the swelling performance of SSG is diffusion-limited, whereas it is absorption capacity-limited for CCS. L-HPC is both diffusion- and absorption capacity-limited. This work also reveals an anisotropic, particle facet dependant, swelling behaviour, which is particularly strong for the liquid uptake ability of two MCC grades (PH101 and PH102) and for the absorption capacity of CCS. Having a better understanding of swelling characteristics of single particles will contribute to improving the rational design of a formulation for oral solid dosage forms.

Impact of powder composition on processing-relevant properties of pharmaceutical materials: An experimental study

In this work, we studied the influence of powder composition on packing density and other processing-relevant properties of binary mixtures, including powder flowability. Binary mixtures of pharmaceutical powders with different particle size ratios, α and varying fractions of large and small particles were analyzed systematically. Mixtures of three excipients and one API with different composition (2, 5, 10, 30, 50, 70, 90, 95 and 98 wt%) were prepared in a Turbula mixer. Powders with different properties and particle size distribution were chosen, in order to obtain three binary mixtures with different size ratios. Then, macroscopic powder properties including bulk (poured) and tapped density (BD and TD) were measured. A powder rheometer was used to measure the flow function coefficient (ffc), cohesion, compressibility and permeability of the binary mixtures. We considered experimentally three classes of binary mixtures, which are characterized by two critical ratios of particle diameter: the critical size ratio of entrance (αc) and the critical size ratio of replacement (αr), where αc = 0.154 and αr = 0.741. Below the critical size ratio of entrance (αc), the particle asymmetry (ratio between large and small particle diameters) is high and small particles can fill the voids between larger ones. Between αc and the critical size ratio of replacement (αr), the smaller particles are too large to fit in the voids between larger particles (packing structure changes). Above αr, the particles are more or less symmetric in size and overall packing structure does not change by mixing the particles. Our experiments show that there is a non-linear and non-monotonic dependence of all relevant properties on composition for powder mixtures that have an α < αr. This non-linear behavior is even more significant for strongly asymmetric binary mixtures with α < αc. We argue that this behavior is related to the composition dependence of random packing of particulate systems. Our results have relevance to pharmaceutical particle processing operations where constant powder mixture properties are needed to ensure quality standards are met; such operations include capsule or die filling during tableting, and the continuous feeding of powders via screw feeders. Our results suggest that for pharmaceutical particle processing operations, where constant powder mixture properties are a prerequisite for process robustness, the size ratio of API and excipient particles, α should not be smaller than αr = 0.741.

Predicting Pharmaceutical Particle Size Distributions Using Kernel Mean Embedding.

In the pharmaceutical industry, the transition to continuous manufacturing of solid dosage forms is adopted by more and more companies. For these continuous processes, high-quality process models are needed. In pharmaceutical wet granulation, a unit operation in the ConsiGma-25 continuous powder-to-tablet system (GEA Pharma systems, Collette, Wommelgem, Belgium), the product under study presents itself as a collection of particles that differ in shape and size. The measurement of this collection results in a particle size distribution. However, the theoretical basis to describe the physical phenomena leading to changes in this particle size distribution is lacking. It is essential to understand how the particle size distribution changes as a function of the unit operation’s process settings, as it has a profound effect on the behavior of the fluid bed dryer. Therefore, we suggest a data-driven modeling framework that links the machine settings of the wet granulation unit operation and the output distribution of granules. We do this without making any assumptions on the nature of the distributions under study. A simulation of the granule size distribution could act as a soft sensor when in-line measurements are challenging to perform. The method of this work is a two-step procedure: first, the measured distributions are transformed into a high-dimensional feature space, where the relation between the machine settings and the distributions can be learnt. Second, the inverse transformation is performed, allowing an interpretation of the results in the original measurement space. Further, a comparison is made with previous work, which employs a more mechanistic framework for describing the granules. A reliable prediction of the granule size is vital in the assurance of quality in the production line, and is needed in the assessment of upstream (feeding) and downstream (drying, milling, and tableting) issues. Now that a validated data-driven framework for predicting pharmaceutical particle size distributions is available, it can be applied in settings such as model-based experimental design and, due to its fast computation, there is potential in real-time model predictive control.