Search

Advances in infrared (IR) spectroscopic imaging instrumentation and data science now present unique opportunities for large validation studies of the concept of histopathology using spectral data. In this study, we examine the discrimination potential of IR metrics for different histologic classes to estimate the sample size needed for designing validation studies to achieve a given statistical power and statistical significance. Next, we present an automated annotation transfer tool that can allow large-scale training/validation, overcoming the limitations of sparse ground truth data with current manual approaches by providing a tool to transfer pathologist annotations from stained images to IR images across diagnostic categories. Finally, the results of a combination of supervised and unsupervised analysis provide a scheme to identify diagnostic groups/patterns and isolating pure chemical pixels for each class to better train complex histopathological models. Together, these methods provide essential tools to take advantage of the emerging capabilities to record and utilize large spectroscopic imaging datasets.

Atomic force microscopy-infrared (AFM-IR) spectroscopic imaging offers non-perturbative, molecular contrast for nanoscale characterization. The need to mitigate measurement artifacts and enhance sensitivity, however, requires narrowly-defined and strict sample preparation protocols. This limits reliable and facile characterization; for example, when using common substrates such as Silicon or glass. Here, we demonstrate a closed-loop (CL) piezo controller design for responsivity-corrected AFM-IR imaging. Instead of the usual mode of recording cantilever deflection driven by sample expansion, the principle of our approach is to maintain a zero amplitude harmonic cantilever deflection by CL control of a subsample piezo. We show that the piezo voltage used to maintain a null deflection provides a reliable measure of the local IR absorption with significantly reduced noise. A complete analytical description of the CL operation and characterization of the controller for achieving robust performance are presented. Accurate measurement of IR absorption of nanothin PMMA films on glass and Silicon validates the robust capability of CL AFM-IR in routine mapping of nanoscale molecular information., Atomic force microscopy-infrared (AFM-IR) spectroscopic imaging techniques offer a non-perturbative, molecular contrast for characterization of nanomaterials; however, data are often complicated by the measurement apparatus, sample preparation conditions and low signal-to-noise ratio. Here, the authors demonstrate a closed-loop controlled AFM-IR instrument design to address measurement artifacts and reduce noise up to 5x compared to previous methods.

Digital analysis of cancer specimens using spectroscopic imaging coupled to machine learning is an emerging area that links spatially localized spectral signatures to tissue structure and disease. In this study, we examine the role of spatial-spectral tradeoffs in infrared spectroscopic imaging configurations for probing tumors and the associated microenvironment profiles at different levels of model complexity. We image breast tissue using standard and high-definition Fourier Transform Infrared (FT-IR) imaging and systematically examine the localization, spectral origins, and utility of data for classification. Results demonstrate that higher spatial detail provides high sensitivity and specificity of tissue segmentation, despite the increased subcellular variability. High definition imaging also allows accurate analysis of complex, multiclass models of breast tissue without compromising accuracy. A comparison of results also highlights the key differences in the data distributions and classification performance across modalities to better guide decision making for acquiring and analyzing IR imaging data for specific histopathological models.

Spectroscopic imaging offers a potential path to all-digital molecular pathology by relating the spectral data to histopathological details of tissue. In the different statistical approaches applied to spectral data, there are gaps in the appropriate sample size estimation for a significant statistical power and accurate model assessment especially for multiclass problems. Underestimation of the sample size can lead to statistically insignificant diagnostic tests while an overestimation can greatly increase experimental costs and time frames.Since the receiver operating characteristic (ROC) curve is designed primarily for a binary test, there are no straightforward approaches to use it for multiple classes in a typical pathology image. In this study, we have described sample size estimation (power analysis) and multiclass diagnostic ROC curve generation for hyperspectral datasets.

Optical microscopy for biomedical samples requires expertise in staining to visualize structure and composition. Midinfrared (mid-IR) spectroscopic imaging offers label-free molecular recording and virtual staining by probing fundamental vibrational modes of molecular components. This quantitative signal can be combined with machine learning to enable microscopy in diverse fields from cancer diagnoses to forensics. However, absorption of IR light by common optical imaging components makes mid-IR light incompatible with modern optical microscopy and almost all biomedical research and clinical workflows. Here we conceptualize an IR-optical hybrid (IR-OH) approach that sensitively measures molecular composition based on an optical microscope with wide-field interferometric detection of absorption-induced sample expansion. We demonstrate that IR-OH exceeds state-of-the-art IR microscopy in coverage (10-fold), spatial resolution (fourfold), and spectral consistency (by mitigating the effects of scattering). The combined impact of these advances allows full slide infrared absorption images of unstained breast tissue sections on a visible microscope platform. We further show that automated histopathologic segmentation and generation of computationally stained (stainless) images is possible, resolving morphological features in both color and spatial detail comparable to current pathology protocols but without stains or human interpretation. IR-OH is compatible with clinical and research pathology practice and could make for a cost-effective alternative to conventional stain-based protocols for stainless, all-digital pathology.

Infrared-Optical Hybrid Microscopy (IR-OH) provides mid-IR imaging on the visible microscope platform based on photothermal expansion. Spectral IR imaging of full tissue slides is demonstrated for automated disease diagnosis and computational H&E staining.

The application of infrared (IR) spectroscopy to the systematic analysis of biological materials has been ongoing for more than 125 years. Recent technological developments and detailed theoretical understanding have resulted especially in instrumentation for imaging that has transformed analytical performance and enhanced the ability to extract useful information in shorter time frames. We provide an overview of IR spectroscopic imaging, with a focus on Fourier transform IR spectral acquisition, and practical aspects of making measurements for biomedical applications. The core aspects of the science of IR imaging and its subsequent use for histologic analysis are described. We discuss in-depth data acquisition, considerations for spectral fidelity, and spatial measures of performance that are important for biomedical analysis pipelines. The application of this technology and its potential is illustrated with a case study of application to extant problems in breast cancer pathology. Finally, a perspective on the emerging areas of progress for the technology is presented.

Infrared (IR) spectroscopic imaging provides both morphologic and chemical detail; however, obtaining this extensive spectral-spatial information requires the ability to rapidly record high-quality data. Discrete frequency infrared (DFIR) imaging using a point scanning microscope strikes a balance between data quality and acquisition speed that, in principle, can further be aided by computational methods. Here, we report a deep learning-based framework to complement the process of data acquisition and information extraction. First, we introduce a convolutional neural network (CNN) to leverage both spatial and spectral information for segmenting data into informative sub-classes, which we call the IR-SEG network. We show that this framework increases accuracy by using approximately half of the features used in the typical pixel-wise classification of IR data. Second, we present a generative adversarial network (GAN)-based approach to reconstruct full data sets with low loss in the information from incomplete spatial and spectral data recording. Termed IR-REC, this approach is shown to speed up data acquisition by up to 20-fold for typical biomedical samples. In addition to enhancing the speed and quality of data, we also propose a method to utilize complementary morphologic detail to estimate the spatial details of a single band IR image beyond the diffraction limit. Finally, we discuss potential pitfalls and new opportunities that can be addressed by developing these methods further. Together, these deep learning techniques provide new capabilities for IR imaging to extract better quality information faster.

In mid-infrared (IR) spectroscopy, the choice of substrate and corresponding illumination method are critical considerations. Transflectance measurements with reflective metallic substrates are commonly used as a cost effective approach. Reflective substrates, however, are widely known to lead to spectral distortions and variability making it difficult to interpret complex spectra. Understanding these effects is a topic of much recent interest. Previous studies have focused only on spectroscopy using incoherent light, which is likely to present distinct distortions from measurements that might use coherent light sources such as lasers. Hence, there is a need to specifically understand spectral data recorded with these emerging sources. Here, we first experimentally determine the structure of an inexpensive and commercially available substrate referred to as low-emissivity glass and then we compare it with a gold-coated glass, using analytical calculations to help understand spectral differences. Transflectance calculations reveal significant differences between the two substrates for both incoherent and coherent light propagation through the multilayer structure. These results highlight the importance of substrate choice for mid-infrared spectroscopy and illustrate that reflective substrates are not universally equivalent. Using a model based on the experimental data, we study the consequences of the low-e structure on the acquired spectral signal and validate the differences between gold and low-e substrates experimentally using polymer films.

Infrared (IR) spectroscopic imaging, utilizing both the molecular and structural disease signatures, enables extensive profiling of tumors and their microenvironments. Here, we examine the relative merits of using either the fingerprint or the high frequency regions of the IR spectrum for tissue histopathology. We selected a complex model as a test case, evaluating both stromal and epithelial segmentation for various breast pathologies. IR spectral classification in each of these spectral windows is quantitatively assessed by estimating area under the curve (AUC) of the receiver operating characteristic curve (ROC) for pixel level accuracy and images for diagnostic ability. We found only small differences, though some that may be sufficiently important in diagnostic tasks to be clinically significant, between the two regions with the fingerprint region-based classifiers consistently emerging as more accurate. The work provides added evidence and comparison with fingerprint region, complex models, and previously untested tissue type (breast) - that the use of restricted spectral regions can provide high accuracy. Our study indicates that the fingerprint region is ideal for epithelial and stromal models to obtain high pixel level accuracies. Glass slides provide a limited spectral feature set but provides accurate information at the patient level.

Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been extensively studied as a method for noise removal in HSI data. It too, however entails a manual speed-accuracy trade-off, namely the process of manually selecting the relevant bands in the MNF space. This process currently takes roughly around a month’s time for acquiring and pre-processing an entire TMA with acceptable signal to noise ratio. We present three approaches termed ‘Fast MNF’, ‘Approx MNF’ and ‘Rand MNF’ where the computational time of the algorithm is reduced, as well as the entire process of band selection is fully automated. This automated approach is shown to perform at the same level of accuracy as MNF with now large speedup factors, resulting in the same task to be accomplished in hours. The different approximations produced by the three algorithms, show the reconstruction accuracy vs storage (50×) and runtime speed (60×) trade-off. We apply the approach for automating the denoising of different tissue histology samples, in which the accuracy of classification (differentiating between the different histologic and pathologic classes) strongly depends on the SNR (signal to noise ratio) of recovered data. Therefore, we also compare the effect of the proposed denoising algorithms on classification accuracy. Since denoising HSI data is done unsupervised, we also use a metric that assesses the quality of denoising in the image domain between the noisy and denoised image in the absence of ground truth.

Significance Cancer alters both the morphological and the biochemical properties of multiple cell types in a tissue. Generally, the morphology of epithelial cells is practical for routine disease diagnoses. Here, infrared spectroscopic imaging biochemically characterizes breast cancer, both epithelial cells and the tumor-associated microenvironment. Unfortunately, conventional spectral analyses are slow. Hence, we designed and built a laser confocal microscope that demonstrates a high signal-to-noise ratio for confident diagnoses. The instrument cuts down imaging time from days to minutes, making the technology feasible for research and clinical translation. Finally, automated human breast cancer biopsy imaging is reported in ∼1 hour, paving the way for routine research into the total tumor (epithelial plus microenvironment) properties and rapid, label-free diagnoses., Histopathology based on spatial patterns of epithelial cells is the gold standard for clinical diagnoses and research in carcinomas; although known to be important, the tissue microenvironment is not readily used due to complex and subjective interpretation with existing tools. Here, we demonstrate accurate subtyping from molecular properties of epithelial cells using emerging high-definition Fourier transform infrared (HD FT-IR) spectroscopic imaging combined with machine learning algorithms. In addition to detecting four epithelial subtypes, we simultaneously delineate three stromal subtypes that characterize breast tumors. While FT-IR imaging data enable fully digital pathology with rich information content, the long spectral scanning times required for signal averaging and processing make the technology impractical for routine research or clinical use. Hence, we developed a confocal design in which refractive IR optics are designed to provide high-definition, rapid spatial scanning and discrete spectral tuning using a quantum cascade laser (QCL) source. This instrument provides simultaneously high resolving power (2-μm pixel size) and high signal-to-noise ratio (SNR) (>1,300), providing a speed increase of ∼50-fold for obtaining classified results compared with present imaging spectrometers. We demonstrate spectral fidelity and interinstrument operability of our developed instrument by accurate analysis of a 100-case breast tissue set that was analyzed in a day, considerably speeding research. Clinical breast biopsies typical of a patients’ caseload are analyzed in ∼1 hour. This study paves the way for comprehensive tumor-microenvironment analyses in feasible time periods, presenting a critical step in practical label-free molecular histopathology.

First, we thank Paraskevaidi et al. (1) for their kind comments and assessment of the potentially significant impact of our work (2). Their letter recognizes that our study could reduce the caseloads of pathologists/surgeons but that it may not directly benefit patients. Surgeons, pathologists, and technologists all strive to improve outcomes for patients; indeed, diagnosis is the gateway to therapy (and better outcomes). The speed and one-shot tumor and microenvironment measurement focus of our study are aimed precisely toward providing new information that is useful while being compatible with how diagnoses are actually made—via pathology. Second, Paraskevaidi et al. (1) propose that earlier detection and less-invasive tools … [↵][1]1To whom correspondence should be addressed. Email: rxb{at}illinois.edu. [1]: #xref-corresp-1-1

BACKGROUND Assessment of process comparability is often required across different phases of manufacturing (Phase I vs Phase II vs Phase III vs Commercial) as well as other key activities during product commercialization (process scale-up, technology transfer, process improvement). In this work, chemometrics was applied to compare two versions of a biotech process and identify process steps where significant differences exist as well as the parameters that cause these differences. RESULTS The dataset used in this analysis consists of data from 229 manufacturing batches. Partial least squares has been used for modeling the data. Scatter plots and variable importance plots have been used for evaluating comparability. Process parameters identified as significant by using chemometrics have been compared against those identified from process characterization using traditional lab scale experimental studies. The comparison has been followed by discussion on the pros and cons of the two approaches. To our knowledge this is the first time that such a comparison has been published for biotech processing. CONCLUSIONS The study further demonstrates the usefulness of chemometrics in defining process comparability and in gathering process understanding from analysis of manufacturing data to supplement traditional lab-scale experimentation. © 2014 Society of Chemical Industry

Biotech unit operations are often characterized by a large number of inputs (operational parameters) and outputs (performance parameters) along with complex correlations among them. A typical biotech process starts with the vial of the cell bank, ends with the final product, and has anywhere from 15 to 30 such unit operations in series. Besides the above-mentioned operational parameters, raw material attributes can also impact process performance and product quality as well as interact among each other. Multivariate data analysis (MVDA) offers an effective approach to gather process understanding from such complex datasets. Review of literature suggests that the use of MVDA is rapidly increasing, fuelled by the gradual acceptance of quality by design (QbD) and process analytical technology (PAT) among the regulators and the biotech industry. Implementation of QbD and PAT requires enhanced process and product understanding. In this article, we first discuss the most critical issues that a practitioner needs to be aware of while performing MVDA of bioprocessing data. Next, we present a step by step procedure for performing such analysis. Industrial case studies are used to elucidate the various underlying concepts. With the increasing usage of MVDA, we hope that this article would be a useful resource for present and future practitioners of MVDA.

Signal transducer and activator of transcription factor 3 (STAT-3) is known to be overexpressed in cancer stem cells. Poor solubility and variable drug absorption are linked to low bioavailability and decreased efficacy. Many of the drugs regulating STAT-3 expression lack aqueous solubility; hence hindering efficient bioavailability. A theranostics nanoplatform based on luminescent carbon particles decorated with cucurbit[6]uril is introduced for enhancing the solubility of niclosamide, a STAT-3 inhibitor. The host-guest chemistry between cucurbit[6]uril and niclosamide makes the delivery of the hydrophobic drug feasible while carbon nanoparticles enhance cellular internalization. Extensive physicochemical characterizations confirm successful synthesis. Subsequently, the host-guest chemistry of niclosamide and cucurbit[6]uril is studied experimentally and computationally. In vitro assessments in human breast cancer cells indicate approximately twofold enhancement in IC50 of drug. Fourier transform infrared and fluorescence imaging demonstrate efficient cellular internalization. Furthermore, the catalytic biodegradation of the nanoplatforms occur upon exposure to human myeloperoxidase in short time. In vivo studies on athymic mice with MCF-7 xenograft indicate the size of tumor in the treatment group is half of the controls after 40 d. Immunohistochemistry corroborates the downregulation of STAT-3 phosphorylation. Overall, the host-guest chemistry on nanocarbon acts as a novel arsenal for STAT-3 inhibition.

High-definition (HD) Fourier transform infrared (FT-IR) spectroscopic imaging is an emerging technique that not only enables chemistry-based visualization of tissue constituents, and label free extraction of biochemical information but its higher spatial detail makes it a potentially useful platform to conduct digital pathology. This methodology, along with fast and efficient data analysis, can enable both quantitative and automated pathology. Here we demonstrate a combination of HD FT-IR spectroscopic imaging of breast tissue microarrays (TMAs) with data analysis algorithms to perform histologic analysis. The samples comprise four tissue states, namely hyperplasia, dysplasia, cancerous and normal. We identify various cell types which would act as biomarkers for breast cancer detection and differentiate between them using statistical pattern recognition tools i.e. Random Forest (RF) and Bayesian algorithms. Feature optimization is integrally carried out for the RF algorithm, reducing computation time as well as redundant spectral features. We achieved an order of magnitude reduction in the number of features with comparable prediction accuracy to that of the original feature set. Together, the demonstration of histology and selection of features paves the way for future applications in more complex models and rapid data acquisition.

Current methods for cancer detection rely on clinical stains, often using immunohistochemistry techniques. Pathologists then evaluate the stained tissue in order to determine cancer stage treatment options. These methods are commonly used, however they are non-quantitative and it is difficult to control for staining quality. In this paper, we propose the use of mid-infrared spectroscopic imaging to classify tissue types in tumor biopsy samples. Our goal is to augment the data available to pathologists by providing them with quantitative chemical information to aid diagnostic activities in clinical and research activities related to breast cancer.

Separation media, in particular chromatography media, is typically one of the major contributors to the cost of goods for production of a biotechnology therapeutic. To be cost-effective, it is industry practice that media be reused over several cycles before being discarded. The traditional approach for estimating the number of cycles a particular media can be reused for involves performing laboratory scale experiments that monitor column performance and carryover. This dataset is then used to predict the number of cycles the media can be used at manufacturing scale (concurrent validation). Although, well accepted and widely practiced, there are challenges associated with extrapolating the laboratory scale data to manufacturing scale due to differences that may exist across scales. Factors that may be different include: level of impurities in the feed material, lot to lot variability in feedstock impurities, design of the column housing unit with respect to cleanability, and homogeneity of the column packing. In view of these challenges, there is a need for approaches that may be able to predict column underperformance at the manufacturing scale over the product lifecycle. In case such an underperformance is predicted, the operators can unpack and repack the chromatography column beforehand and thus avoid batch loss. Chemometrics offers one such solution. In this article, we present an application of chemometrics toward the analysis of a set of chromatography profiles with the intention of predicting the various events of column underperformance including the backpressure buildup and inefficient deoxyribonucleic acid clearance.