Convolutional neural networks for segmentation of FIB-SEM nanotomography data from porous polymer films for controlled drug release.

Phase-separated polymer films are commonly used as coatings around pharmaceutical oral dosage forms (tablets or pellets) to facilitate controlled drug release. A typical choice is to use ethyl cellulose and hydroxypropyl cellulose (EC/HPC) polymer blends. When an EC/HPC film is in contact with water, the leaching out of the water-soluble HPC phase produces an EC film with a porous network through which the drug is transported. The drug release can be tailored by controlling the structure of this porous network. Imaging and characterization of such EC porous films facilitates understanding of how to control and tailor film formation and ultimately drug release. Combined focused ion beam and scanning electron microscope (FIB-SEM) tomography is a well-established technique for high-resolution imaging, and suitable for this application. However, for segmenting image data, in this case to correctly identify the porous network, FIB-SEM is a challenging technique to work with. In this work, we implement convolutional neural networks for segmentation of FIB-SEM image data. The data are acquired from three EC porous films where the HPC phases have been leached out. The three data sets have varying porosities in a range of interest for controlled drug release applications. We demonstrate very good agreement with manual segmentations. In particular, we demonstrate an improvement in comparison to previous work on the same data sets that utilized a random forest classifier trained on Gaussian scale-space features. Finally, we facilitate further development of FIB-SEM segmentation methods by making the data and software used open access.

Drug release from pharmaceutical tablets or pellets is often controlled by applying a phase-separated polymer film coating. Ethyl cellulose and hydroxypropyl cellulose (EC/HPC) polymer blends are commonly used. The HPC phase leaches out when in contact with water and the result is a porous EC matrix coating, with mass transport properties that can be controlled by tailoring the structure of the porous network. High-resolution 3D imaging is necessary to characterize such materials, and the resolution of e.g. X-ray computed tomography is simply insufficient. Combined focused ion beam and scanning electron microscope (FIB-SEM) tomography on the other hand is a suitable technique, but segmentation of FIB-SEM data, in this case to separate the solid matrix and the porous network, is challenging. In this work, we develop a method for segmentation of FIB-SEM image data acquired from three different EC porous films where the HPC phases have been leached out. The segmentation is based on convolutional neural networks (CNNs). CNNs is a well-established machine learning paradigm and has demonstrated state-of-the-art performance in many image analysis and segmentation tasks. CNNs are inspired from biological processes in the visual cortex and act similarly, at least conceptually. In contrast to most conventional machine learning algorithms, CNNs learn by themselves which features to extract from the images. The features are extracted at different spatial scales and may constitute e.g. edge and contrast detectors. These features are subsequently used for classification. In this work, CNNs are used for image segmentation. The goal is to identify which regions in the images that contain either pore (empty space) or solid (material), hence a binary classification task. For the CNN to learn how to perform such a task, a ground truth is needed. This is achieved by letting an expert manually segment parts of the data. This is a very time-consuming endeavor, hence only a small random subset of the full dataset is manually segmented. The CNN is trained for the task using the manually segmented data, after which automatic segmentation of the full dataset is performed. We obtain very good agreement with manual segmentations in terms of accuracy and porosity, and a clear improvement in comparison to an earlier developed random forest classifier trained on Gaussian scale-space features on the same data. The development of accurate segmentation methods is a crucial step toward better understanding and tailoring of coatings for controlled drug release.

Quantification of surface composition and surface structure of inhalation powders using TOF-SIMS.

A multivariate TOF-SIMS methodology has been developed and applied to quantify surface composition and chemical distribution for dry powder blends. Surface properties are often critical to the behavior of powder formulations, especially in the case of dry powders for inhalation, as surface properties directly affect inter-particulate forces and, hence, the dispersibility of the formulation. The mass spectrum at each pixel was fit to a linear combination of reference spectra obtained by non-negatively constrained alternating least squares. From the pixel compositions, average surface coverage and a range of other image features were calculated. Two kinds of systems have been examined: 1) binary blends of lactose particles and coating agents, and 2) blends of different inhalation drugs with carrier lactose. For both kinds of systems, detailed insight into the surface composition and structure could be derived. For the former study, TOF-SIMS results were compared with a complementary surface analysis technique, XPS.

Characterization of acoustic emission analysis in applications for inhalation device performance assessment.

Acoustic Emission (AE) measurement technology has gained wide appreciation in material sciences and process monitoring. In inhalation research, AE has been used for adherence indicating applications in clinical studies. Promising results from feasibility studies using AE combined with multivariate data analysis (AE-MVDA) in the analysis of devices for inhalation have prompted a broader study reported in this paper. This work presents the novel application of AE-MVDA for assessment of the combined inhalation device and formulation performance. The purpose is to evaluate the benefits that this technology can provide to inhalation product development programs. The work was carried out using two different dry powder inhaler device model systems while investigating different performance features. The devices were filled with dry powder formulations with both placebo and with active pharmaceutical ingredient (API). The acquired AE data was analyzed using multivariate data analysis tools such as Principal component analysis (PCA) and orthogonal projections to latent structures (OPLS). The AE profiles were indicative for device and formulation performance. Normal and deviating performances were readily picked up in the AE data. Moreover, performance trends between doses withdrawn from the inhalers were also observable. Lastly, differences in the AE profile between the formulations could be detected. The overall conclusion from the AE-MVDA measurement approach evaluation is that it has the potential to add value as a cost-effective, non-invasive quality and performance monitoring technology both in development and in production of inhaled medicines.

New features of arabinoxylan ethers revealed by using multivariate analysis.

Hemicelluloses are a relatively unused renewable resource. One reason is their broad structure variety that makes it hard to understand structure-property relations. In this study arbinoxylan, extracted from barley husk, was chemically modified into hydroxypropyl methyl-, hydroxypropyl- and methyl arabinoxylan. The relationship between structure and phase behavior was investigated by using multivariate analysis. The arabinoxylan ethers were characterized using mid-infrared FTIR spectroscopy and from principal components analysis, PCA, structural or physical variations between samples were visualized. With orthogonal projections to latent structures, OPLS, vibrations specific for arabinoxylan hydroxypropyl and methyl substitutions was assigned. Among the observed differences between chemical derivatives was an intensity change in the water vibration. The differences in hydration were related to clouding phase behavior of the arabinoxylan ethers. This study shows that multivariate analysis methods are useful for finding unexpected and/or hidden features in the polysaccharide structure.

Imaging and Identification of Marine Algal Bioactive Compounds by Surface Enhanced Raman Spectroscopy (SERS).

Surface-enhanced Raman spectroscopy (SERS) is a highly selective technique that can be used for imaging of single algae cells. In contrast to normal Raman spectroscopy, SERS utilizes light interaction with colloidal gold or silver particles working as antennas to match the sensitivity of fluorescence measurements. Furthermore, SERS enables a more profound picture of not only the analyte of interest but also the present biological matrix without the need for additional fluorescence labelling. The introduction of an internal standard in the form of a thiol self-assembled monolayer (SAM) on the colloidal gold or silver particles can be used to normalize the SERS response that otherwise would also depend on the locations of the colloid particles in the microscope image.In light of the vast amounts of data that is generated in each spectrum and the large variance in enhancement signal, multivariate analysis is necessary for accurate evaluation. This can be done by the use of transposed orthogonal projections to latent structures (T-OPLS), where the variations of properties in the reference spectra, Y table, and the variation in spectra, X table, are correlated.

Near-infrared chemical imaging (NIR-CI) on roll compacted ribbons and tablets--multivariate mapping of physical and chemical properties.

Near-infrared chemical imaging (NIR-CI) is an attractive technique within the pharmaceutical industry, where tools are continuously in demand to assess the quality of the intermediate and final products. The present paper demonstrates how NIR-CI in combination with multivariate methods was utilized to spatially map physical properties and content of roll compacted ribbons and tablets. Additionally, extracted textural parameters from tablet images were correlated to the design parameters of the roll compaction process as well as to the physical properties of the granules. The results established the use of NIR-CI as a complementary nondestructive tool to determine the ribbon density and map the density distribution across the width and along the length of the ribbons. For the tablets, the compaction pressure developed during compression increased with the lateral distance from the center. Therefore, NIR-CI can be an effective tool to provide information about the spatial distribution of the compaction pressures on the surface of the tablet. Moreover, low roll compaction roll force correlated to a heterogeneous type of texture in the API chemical image. Overall, texture analysis of the tablets enabled efficient investigation of the spatial variation and could be used to advance process understanding. Finally, orthogonal projections to latent structures (O2PLS) model facilitated the understanding of the interrelationships between textural features, design parameters and physical properties data by separately joint and unique variations.

Roll compaction process modeling: transfer between equipment and impact of process parameters.

In this study, the roll compaction of an intermediate drug load formulation was performed using horizontally and vertically force fed roll compactors. The horizontally fed roll compactor was equipped with an instrumented roll technology allowing the direct measurement of normal stress at the roll surface, while the vertically fed roll compactor was equipped with a force gauge between the roll axes. Furthermore, characterization of ribbons, granules and tablets was also performed. Ribbon porosity was primarily found to be a function of normal stress, exhibiting a quadratic relationship thereof. A similar quadratic relationship was also observed between roll force and ribbon porosity of the vertically fed roll compactor. The predicted peak pressure (Pmax) using the Johanson model was found to be higher than the measured normal stress, however, the predicted Pmax correlated well with the ribbon relative density/porosity and the majority of downstream properties of granules and tablets, demonstrating its use as a scale-independent parameter. A latent variable model was developed for both the horizontal and vertical fed roll compactors to express ribbon porosity as a function of geometric and process parameters. The model validation, performed with new data, resulted in overall good predictions. This study successfully demonstrated the scale up/transfer between two different roll compactors and revealed that the combined use of design of experiments, latent variable models and in silico predictions result in better understanding of the critical process parameters in roll compaction.

Comparison of exhaled endogenous particles from smokers and non-smokers using multivariate analysis.

BACKGROUND: Smoking, along with many respiratory diseases, has been shown to induce airway inflammation and alter the composition of the respiratory tract lining fluid (RTLF). We have previously shown that the phospholipid and protein composition of particles in exhaled air (PEx) reflects that of RTLF. In this study, we hypothesized that the composition of PEx differs between smokers and non-smokers, reflecting inflammation in the airways. OBJECTIVE: It was the aim of this study to identify differences in the phospholipid composition of PEx from smokers and non-smokers. METHODS: PEx from 12 smokers and 13 non-smokers was collected using a system developed in-house. PEx was analysed using time-of-flight secondary ion mass spectrometry, and the mass spectral data were evaluated using multivariate analysis. Orthogonal partial least squares (OPLS) was used to relate smoking status, lung function and pack years to the chemical composition of RTLF. The discriminating ions identified by OPLS were then used as explanatory variables in traditional regression analysis. RESULTS: There was a clear discrimination between smokers and non-smokers according to the chemical composition, where phospholipids from smokers were protonated and sodiated to a larger extent. Poor lung function showed a strong association with higher response from all molecular phosphatidylcholine species in the samples. Furthermore, the accumulated amount of tobacco consumed was associated with variations in mass spectra, indicating a dose-response relationship. CONCLUSION: The chemical composition of PEx differs between smokers and non-smokers, reflecting differences in the RTLF. The results from this study may suggest that the composition of RTLF is affected by smoking and may be of importance for lung function.

A quality by design approach to investigate the effect of mannitol and dicalcium phosphate qualities on roll compaction.

Roll compaction is a continuous process for solid dosage form manufacturing increasingly popular within pharmaceutical industry. Although roll compaction has become an established technique for dry granulation, the influence of material properties is still not fully understood. In this study, a quality by design (QbD) approach was utilized, not only to understand the influence of different qualities of mannitol and dicalcium phosphate (DCP), but also to predict critical quality attributes of the drug product based solely on the material properties of that filler. By describing each filler quality in terms of several representative physical properties, orthogonal projections to latent structures (OPLS) was used to understand and predict how those properties affected drug product intermediates as well as critical quality attributes of the final drug product. These models were then validated by predicting product attributes for filler qualities not used in the model construction. The results of this study confirmed that the tensile strength reduction, known to affect plastic materials when roll compacted, is not prominent when using brittle materials. Some qualities of these fillers actually demonstrated improved compactability following roll compaction. While direct compression qualities are frequently used for roll compacted drug products because of their excellent flowability and good compaction properties, this study revealed that granules from these qualities were more poor flowing than the corresponding powder blends, which was not seen for granules from traditional qualities. The QbD approach used in this study could be extended beyond fillers. Thus any new compound/ingredient would first be characterized and then suitable formulation characteristics could be determined in silico, without running any additional experiments.

Chemical images of marine bio-active compounds by surface enhanced Raman spectroscopy and transposed orthogonal partial least squares (T-OPLS).

Surface enhanced Raman spectroscopy combined with transposed Orthogonal Partial Least Squares (T-OPLS) was shown to produce chemical images of the natural antibacterial surface-active compound 1,1,3,3-tetrabromo-2-heptanone (TBH) on Bonnemaisonia hamifera. The use of gold colloids functionalised with the internal standard 4-mercapto-benzonitrile (MBN) made it possible to create images of the relative concentration of TBH over the surfaces. A gradient of TBH could be mapped over and in the close vicinity of the B. hamifera algal vesicles at the attomol/pixel level. T-OPLS produced a measure of the spectral correlation for each pixel of the hyperspectral images whilst not including spectral variation that was linearly independent of the target spectrum. In this paper we show the possibility to retrieve specific spectral information with a low magnitude in a complex matrix.

Combining experimental design and orthogonal projections to latent structures to study the influence of microcrystalline cellulose properties on roll compaction.

Roll compaction is gaining importance in pharmaceutical industry for the dry granulation of heat or moisture sensitive powder blends with poor flowing properties prior to tabletting. We studied the influence of microcrystalline cellulose (MCC) properties on the roll compaction process and the consecutive steps in tablet manufacturing. Four dissimilar MCC grades, selected by subjecting their physical characteristics to principal components analysis, and three speed ratios, i.e. the ratio of the feed screw speed and the roll speed of the roll compactor, were included in a full factorial design. Orthogonal projection to latent structures was then used to model the properties of the resulting roll compacted products (ribbons, granules and tablets) as a function of the physical MCC properties and the speed ratio. This modified version of partial least squares regression separates variation in the design correlated to the considered response from the variation orthogonal to that response. The contributions of the MCC properties and the speed ratio to the predictive and orthogonal components of the models were used to evaluate the effect of the design variation. The models indicated that several MCC properties, e.g. bulk density and compressibility, affected all granule and tablet properties, but only one studied ribbon property: porosity. After roll compaction, Ceolus KG 1000 resulted in tablets with obvious higher tensile strength and lower disintegration time compared to the other MCC grades. This study confirmed that the particle size increase caused by roll compaction is highly responsible for the tensile strength decrease of the tablets.

Multivariate analysis of the relation between immune dysfunction and treatment intensity in children with acute lymphoblastic leukemia.

BACKGROUND: Immunoreconstitution following childhood acute lymphoblastic leukemia (ALL) is a complex process during which various immune functions recover differentially. This process is difficult to elucidate since variables are interrelated and require simultaneous evaluation, rendering conventional statistical methods inappropriate. PROCEDURE: We used principal components analysis (PCA) and projection of latent structures (PLS) to evaluate immune competence in 32 children treated for ALL. One or 6 months after completion of therapy, the relation between lymphocyte subpopulations, lymphocyte function and response to vaccination with tetanus, diphtheria and hemophilus influenzae, was investigated. RESULTS: PCA demonstrated that increasing treatment intensity correlated with progressive immune dysfunction. Children treated with high intensity had poor response to vaccination associated with loss of humoral memory, decreased CD4(+) 45RA(+) T-lymphocytes and increased CD5+ B-lymphocytes. Patients treated with intermediate intensity had better preservation of humoral memory but decreased CD4(+) 45RA(+) T-cells. Patients with a low intensity regimen had similar vaccination response and lymphocyte levels as controls. CONCLUSIONS: Our findings demonstrate the utility of PCA and PLS in detecting hidden structures in complex data and suggest that, even 6 months after therapy, patients treated with intermediate and high intensity have attenuated responses to de novo antigens whereas those with high intensity also respond poorly to recall antigens.

Transfer of NIR calibrations for pharmaceutical formulations between different instruments.

In order to evaluate how well existing techniques for transferring NIR calibrations perform for solid pharmaceutical formulations, a study on four assays of active ingredients was undertaken. The study included two configurations of dispersive NIR instruments and one Fourier transform (FT) instrument. Three methods for calibration transfer: slope/bias correction, local centring and piecewise direct standardisation (PDS), were tested and evaluated. Our conclusions are that the calibration transfer methods tested can perform equally well. It was shown that it is possible to transfer calibrations between instruments of different configurations or even of different types, without loosing the prediction ability of the calibration. To achieve a good calibration transfer, a larger variation in the content of the active ingredient in the samples and more samples are needed for the slope and bias correction method compared to the local centring method. For PDS to be a successful calibration transfer method, an optimisation of the number of transfer samples and how they are selected together with various factors specific for this method is needed. Local centring is the preferred transfer method as its performance is excellent yet it is simple to perform, no optimisation is needed, only a few transfer samples are required and the transfer samples do not have to vary in their content of the active ingredient.

Evaluation of basic algorithms for transferring quantitative multivariate calibrations between scanning grating and FT NIR spectrometers.

A key issue in near infrared spectroscopy (NIR) is the possibility to use calibrations generated on one instrument for predictions on others. A number of methods for calibration transfer have been proposed, but which method to choose is typically not straightforward. An evaluation of a number of methods for transferring quantitative calibrations between different instruments was carried out on near infrared diffuse-reflectance data from a pharmaceutical formulation. Six instruments were included in the study, five of which were scanning grating instruments, both with and without fibre-optic probe configuration, and one of which was a Fourier-transform instrument, equipped with a fibre-optic probe. The results show that it is possible to transfer calibrations between different instruments, provided that a structured procedure is used. Simple techniques for calibration transfer, such as slope/bias correction on the predicted results, as well as standard normal variate transformation and local centring of the raw spectra, gave considerably lower prediction errors on transfer than did standardisation with a certified diffuse-reflectance standard, or direct transfer without any transfer function. Notably, including more than one instrument in the calibration also improved the prediction ability of the models on calibration transfer. No significant differences in wavelength scale were found when a certified diffuse-reflectance wavelength standard was measured on the instruments studied. Nor did simulated wavelength scale differences below +/-0.3 nm cause any significant change in the prediction errors.

Surface-enhanced Raman scattering imaging of single living lymphocytes with multivariate evaluation.

This paper is aimed to show the possibility to determine individual organic compounds introduced into single living cells with surface-enhanced Raman spectroscopy (SERS). Surface enhancement was achieved with gold colloids that were allowed to diffuse into lymphocytes. An introduced analyte, rhodamine 6G, could be imaged together with for example nucleotides and amino acids of the cell. Multivariate evaluation of surface-enhanced Raman images proved to be a powerful tool for the separation of spectral information of various intracellular components. The principal component analysis (PCA) enabled identification of spectra containing different chemical information and separation of the spectral contribution of rhodamine 6G from the complex cellular matrix.