Automatic recognition of breast invasive ductal carcinoma based on terahertz spectroscopy with wavelet packet transform and machine learning

Statistical Analysis of Gastric Cancer Cells Response to Broadband Terahertz Radiation with and without Contrast Nanoparticles

The paper describes the statistical analysis of the response of gastric cancer cells and normal cells to broadband terahertz radiation up to 4 THz, both with and without the use of nanostructured contrast agents. The THz spectroscopy analysis was comparatively performed under the ATR procedure and transmission measurement procedure. The statistical analysis was conducted towards multiple pairwise comparisons, including a support medium (without cells) versus a support medium with nanoparticles, normal cells versus normal cells with nanoparticles, and, respectively, tumor cells versus tumor cells with nanoparticles. When generally comparing the ATR procedure and transmission measurement procedure for a broader frequency domain, the differentiation between normal and tumor cells in the presence of contrast agents is superior when using the ATR procedure. THz contrast enhancement by using contrast agents derived from MRI-related contrast agents leads to only limited benefits and only for narrow THz frequency ranges, a disadvantage for THz medical imaging.

Biomedical application of terahertz imaging technology: a narrative review

Background and Objective

Terahertz (THz) imaging has wide applications in biomedical research due to its properties, such as non-ionizing, non-invasive and distinctive spectral fingerprints. Over the past 6 years, the application of THz imaging in tumor tissue has made encouraging progress. However, due to the strong absorption of THz by water, the large size, high cost, and low sensitivity of THz devices, it is still difficult to be widely used in clinical practice. This paper provides ideas for researchers and promotes the development of THz imaging in clinical research.

Methods

The literature search was conducted in the Web of Science and PubMed databases using the keywords "Terahertz imaging", "Breast", "Brain", "Skin" and "Cancer". A total of 94 English language articles from 1 January, 2017 to 30 December, 2022 were reviewed.

Key Content and Findings

In this review, we briefly introduced the recent advances in THz near-field imaging, single-pixel imaging and real-time imaging, the applications of THz imaging for detecting breast, brain and skin tissues in the last 6 years were reviewed, and the advantages and existing challenges were identified. It is necessary to combine machine learning and metamaterials to develop real-time THz devices with small size, low cost and high sensitivity that can be widely used in clinical practice. More powerful THz detectors can be developed by combining graphene, designing structures and other methods to improve the sensitivity of the devices and obtain more accurate information. Establishing a THz database is one of the important methods to improve the repeatability and accuracy of imaging results.

Conclusions

THz technology is an effective method for tumor imaging. We believe that with the joint efforts of researchers and clinicians, accurate, real-time, and safe THz imaging will be widely applied in clinical practice in the future.

Revolutionary approaches for cancer diagnosis by terahertz-based spectroscopy and imaging

Most tumors are easily missed and misdiagnosed due to the lack of specific clinical signs and symptoms in the early stage. Thus, an accurate, rapid and reliable early tumor detection method is highly desirable. The application of terahertz (THz) spectroscopy and imaging in biomedicine has made remarkable progress in the past two decades, which addresses the shortcomings of existing technologies and provides an alternative for early tumor diagnosis. Although issues such as size mismatch and strong absorption of THz waves by water have set hurdles for cancer diagnosis by THz technology, innovative materials and biosensors in recent years have led to possibilities for new THz biosensing and imaging methods. In this article, we reviewed the issues that need to be solved before THz technology is used for tumor-related biological sample detection and clinical auxiliary diagnosis. We focused on the recent research progress of THz technology, with an emphasis on biosensing and imaging. Finally, the application of THz spectroscopy and imaging for tumor diagnosis in clinical practice and the main challenges in this process were also mentioned. Collectively, THz-based spectroscopy and imaging reviewed here is envisioned as a cutting-edge approach for cancer diagnosis.

Identifying plastics with photoluminescence spectroscopy and machine learning

A quantitative understanding of the worldwide plastics distribution is required not only to assess the extent and possible impact of plastic litter on the environment but also to identify possible counter measures. A systematic collection of data characterizing amount and composition of plastics has to be based on two crucial components: (i) An experimental approach that is simple enough to be accessible worldwide and sensible enough to capture the diversity of plastics; (ii) An analysis pipeline that is able to extract the relevant parameters from the vast amount of experimental data. In this study, we demonstrate that such an approach could be realized by a combination of photoluminescence spectroscopy and a machine learning-based theoretical analysis. We show that appropriate combinations of classifiers with dimensional reduction algorithms are able to identify specific material properties from the spectroscopic data. The best combination is based on an unsupervised learning technique making our approach robust to alternations of the input data.

Accurate and early prediction of the wound healing outcome of burn injuries using the wavelet Shannon entropy of terahertz time-domain waveforms

Significance

Severe burn injuries cause significant hypermetabolic alterations that are highly dynamic, hard to predict, and require acute and critical care. The clinical assessments of the severity of burn injuries are highly subjective and have consistently been reported to be inaccurate. Therefore, the utilization of other imaging modalities is crucial to reaching an objective and accurate burn assessment modality.

Aim

We describe a non-invasive technique using terahertz time-domain spectroscopy (THz-TDS) and the wavelet packet Shannon entropy to automatically estimate the burn depth and predict the wound healing outcome of thermal burn injuries.

Approach

We created 40 burn injuries of different severity grades in two porcine models using scald and contact methods of infliction. We used our THz portable handheld spectral reflection (PHASR) scanner to obtain the in vivo THz-TDS images. We used the energy to Shannon entropy ratio of the wavelet packet coefficients of the THz-TDS waveforms on day 0 to create supervised support vector machine (SVM) classification models. Histological assessments of the burn biopsies serve as the ground truth.

Results

We achieved an accuracy rate of 94.7% in predicting the wound healing outcome, as determined by histological measurement of the re-epithelialization rate on day 28 post-burn induction, using the THz-TDS measurements obtained on day 0. Furthermore, we report the accuracy rates of 89%, 87.1%, and 87.6% in automatic diagnosis of the superficial partial-thickness, deep partial-thickness, and full-thickness burns, respectively, using a multiclass SVM model.

Conclusions

The THz PHASR scanner promises a robust, high-speed, and accurate diagnostic modality to improve the clinical triage of burns and their management.

A General Neural Network Model for Complex Refractive Index Extraction of Low-Loss Materials in the Transmission-Mode THz-TDS

The complex refractive index for low-loss materials is conventionally extracted by either approximate analytical formula or numerical iterative algorithm (such as Nelder-Mead and Newton-Raphson) based on the transmission-mode terahertz time domain spectroscopy (THz-TDS). A novel 4-layer neural network model is proposed to obtain optical parameters of low-loss materials with high accuracy in a wide range of parameters (frequency and thickness). Three materials (TPX, z-cut crystal quartz and 6H SiC) with different dispersions and thicknesses are used to validate the robustness of the general model. Without problems of proper initial values and non-convergence, the neural network method shows even smaller errors than the iterative algorithm. Once trained and tested, the proposed method owns both high accuracy and wide generality, which will find application in the multi-class object detection and high-precision characterization of THz materials.

Federated Learning-Based Detection of Invasive Carcinoma of No Special Type with Histopathological Images

Invasive carcinoma of no special type (IC-NST) is known to be one of the most prevalent kinds of breast cancer, hence the growing research interest in studying automated systems that can detect the presence of breast tumors and appropriately classify them into subtypes. Machine learning (ML) and, more specifically, deep learning (DL) techniques have been used to approach this problem. However, such techniques usually require massive amounts of data to obtain competitive results. This requirement makes their application in specific areas such as health problematic as privacy concerns regarding the release of patients’ data publicly result in a limited number of publicly available datasets for the research community. This paper proposes an approach that leverages federated learning (FL) to securely train mathematical models over multiple clients with local IC-NST images partitioned from the breast histopathology image (BHI) dataset to obtain a global model. First, we used residual neural networks for automatic feature extraction. Then, we proposed a second network consisting of Gabor kernels to extract another set of features from the IC-NST dataset. After that, we performed a late fusion of the two sets of features and passed the output through a custom classifier. Experiments were conducted for the federated learning (FL) and centralized learning (CL) scenarios, and the results were compared. Competitive results were obtained, indicating the positive prospects of adopting FL for IC-NST detection. Additionally, fusing the Gabor features with the residual neural network features resulted in the best performance in terms of accuracy, F1 score, and area under the receiver operation curve (AUC-ROC). The models show good generalization by performing well on another domain dataset, the breast cancer histopathological (BreakHis) image dataset. Our method also outperformed other methods from the literature.

A novel strategy regarding geometric product for liquids discrimination based on THz reflection spectroscopy

A novel expression of geometric product that associated with the geometric relationship from geometric algebra constructed by the vectorized refractive index and absorption coefficient in THz region is proposed, which could provide a new insight into the THz properties of materials. From the novel expression, the candidate characteristic parameters are extracted for liquids discrimination and present the abundant second order correlation information of optical parameters with the consideration of dimension rising. Three groups of liquids, containing C-reactive protein calibrators and alpha fetoprotein calibrators, were selected as examples to validate the feasibility of the proposed strategy. Comparing with the traditional THz parameters, including refractive index, absorption coefficient, and complex permittivity, the novel approach exhibits notable superiority for differentiation with the evaluation of statistical differences and effect sizes. The proposed geometric product expression could have a large potential on promoting the substance identification in some applications of THz technology.

Artificial neural networks for material parameter extraction in terahertz time-domain spectroscopy

Terahertz time-domain spectroscopy (THz-TDS) is a proven technique whereby the complex refractive indices of materials can be obtained without requiring the use of the Kramers-Kronig relations, as phase and amplitude information can be extracted from the measurement. However, manual pre-processing of the data is still required and the material parameters require iterative fitting, resulting in complexity, loss of accuracy and inconsistencies between measurements. Alternatively approximations can be used to enable analytical extraction but with a considerable sacrifice of accuracy. We investigate the use of machine learning techniques for interpreting spectroscopic THz-TDS data by training with large data sets of simulated light-matter interactions, resulting in a computationally efficient artificial neural network for material parameter extraction. The trained model improves on the accuracy of analytical methods that need approximations while being easier to implement and faster to run than iterative root-finding methods. We envisage neural networks can alleviate many of the common hurdles involved in analyzing THz-TDS data such as phase unwrapping, time domain windowing, slow computation times, and extraction accuracy at the low frequency range.

Deep neural network classification of in vivo burn injuries with different etiologies using terahertz time-domain spectral imaging

Thermal injuries can occur due to direct exposure to hot objects or liquids, flames, electricity, solar energy and several other sources. If the resulting injury is a deep partial thickness burn, the accuracy of a physician's clinical assessment is as low as 50-76% in determining the healing outcome. In this study, we show that the Terahertz Portable Handheld Spectral Reflection (THz-PHASR) Scanner combined with a deep neural network classification algorithm can accurately differentiate between partial-, deep partial-, and full-thickness burns 1-hour post injury, regardless of the etiology, scanner geometry, or THz spectroscopy sampling method (ROC-AUC = 91%, 88%, and 86%, respectively). The neural network diagnostic method simplifies the classification process by directly using the pre-processed THz spectra and removing the need for any hyperspectral feature extraction. Our results show that deep learning methods based on THz time-domain spectroscopy (THz-TDS) measurements can be used to guide clinical treatment plans based on objective and accurate classification of burn injuries.

Supervised machine learning for automatic classification of in vivo scald and contact burn injuries using the terahertz Portable Handheld Spectral Reflection (PHASR) Scanner

We present an automatic classification strategy for early and accurate assessment of burn injuries using terahertz (THz) time-domain spectroscopic imaging. Burn injuries of different severity grades, representing superficial partial-thickness (SPT), deep partial-thickness (DPT), and full-thickness (FT) wounds, were created by a standardized porcine scald model. THz spectroscopic imaging was performed using our new fiber-coupled Portable HAndheld Spectral Reflection Scanner, incorporating a telecentric beam steering configuration and an f-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\theta$$\end{document}θ scanning lens. ASynchronous Optical Sampling in a dual-fiber-laser THz spectrometer with 100 MHz repetition rate enabled high-speed spectroscopic measurements. Given twenty-four different samples composed of ten scald and ten contact burns and four healthy samples, supervised machine learning algorithms using THz-TDS spectra achieved areas under the receiver operating characteristic curves of 0.88, 0.93, and 0.93 when differentiating between SPT, DPT, and FT burns, respectively, as determined by independent histological assessments. These results show the potential utility of our new broadband THz PHASR Scanner for early and accurate triage of burn injuries.

Translation-Invariant Zero-Phase Wavelet Methods for Feature Extraction in Terahertz Time-Domain Spectroscopy

Wavelet transform is an important tool in the computational signal processing of terahertz time-domain spectroscopy (THz-TDS) measurements. Despite its prevalence, the effects of using different forms of wavelet transforms in THz-TDS studies have not been investigated. In this paper, we explore the implications of using the maximal overlap discrete wavelet transform (MODWT) versus the well-known discrete wavelet transform (DWT). We demonstrate that the spectroscopic features extracted using DWT can vary over different overlapping frequency ranges. On the contrary, MODWT is translation-invariant and results in identical features, regardless of the spectral range used for its implementation.We also demonstrate that the details coefficients obtained by the multiresolution analysis (MRA) using MODWT are associated with zero-phase filters. In contrast, DWT details coefficients suffer from misalignments originated from the down- and upsampling operations in DWT pyramid algorithm. Such misalignments have adverse effects when it is critical to retain the exact location of the absorption lines. We study the differences of DWT and MODWT both analytically and experimentally, using reflection THz-TDS measurements of α-lactose monohydrate. This manuscript can guide the researchers to select the right wavelet analysis tool for their specific application of the THz spectroscopy.

Terahertz Imaging for Breast Cancer Detection

Terahertz (THz) imaging has the potential to detect breast tumors during breast-conserving surgery accurately. Over the past decade, many research groups have extensively studied THz imaging and spectroscopy techniques for identifying breast tumors. This manuscript presents the recent development of THz imaging techniques for breast cancer detection. The dielectric properties of breast tissues in the THz range, THz imaging and spectroscopy systems, THz radiation sources, and THz breast imaging studies are discussed. In addition, numerous chemometrics methods applied to improve THz image resolution and data collection processing are summarized. Finally, challenges and future research directions of THz breast imaging are presented.

Machine Learning Techniques for THz Imaging and Time-Domain Spectroscopy

Terahertz imaging and time-domain spectroscopy have been widely used to characterize the properties of test samples in various biomedical and engineering fields. Many of these tasks require the analysis of acquired terahertz signals to extract embedded information, which can be achieved using machine learning. Recently, machine learning techniques have developed rapidly, and many new learning models and learning algorithms have been investigated. Therefore, combined with state-of-the-art machine learning techniques, terahertz applications can be performed with high performance that cannot be achieved using modeling techniques that precede the machine learning era. In this review, we introduce the concept of machine learning and basic machine learning techniques and examine the methods for performance evaluation. We then summarize representative examples of terahertz imaging and time-domain spectroscopy that are conducted using machine learning.

Diagnosis of Glioma Molecular Markers by Terahertz Technologies

This review considers glioma molecular markers in brain tissues and body fluids, shows the pathways of their formation, and describes traditional methods of analysis. The most important optical properties of glioma markers in the terahertz (THz) frequency range are also presented. New metamaterial-based technologies for molecular marker detection at THz frequencies are discussed. A variety of machine learning methods, which allow the marker detection sensitivity and differentiation of healthy and tumor tissues to be improved with the aid of THz tools, are considered. The actual results on the application of THz techniques in the intraoperative diagnosis of brain gliomas are shown. THz technologies’ potential in molecular marker detection and defining the boundaries of the glioma’s tissue is discussed.

Terahertz Imaging and Spectroscopy in Cancer Diagnostics: A Technical Review

Terahertz (THz) waves are electromagnetic waves with frequency in the range from 0.1 to 10 THz. THz waves have great potential in the biomedical field, especially in cancer diagnosis, because they exhibit low ionization energy and can be used to discern most biomolecules based on their spectral fingerprints. In this paper, we review the recent progress in two applications of THz waves in cancer diagnosis: imaging and spectroscopy. THz imaging is expected to help researchers and doctors attain a direct intuitive understanding of a cancerous area. THz spectroscopy is an efficient tool for component analysis of tissue samples to identify cancer biomarkers. Additionally, the advantages and disadvantages of the developed technologies for cancer diagnosis are discussed. Furthermore, auxiliary techniques that have been used to enhance the spectral signal-to-noise ratio (SNR) are also reviewed.