Diagnosis of early gastric cancer based on fluorescence hyperspectral imaging technology combined with partial-least-square discriminant analysis and support vector machine

A two-step method of cyclolinopeptide (linusorb) preparation from flaxseed cake via dry-screening

Linusorbs (LO), cyclolinopeptides, are a group of cyclic hydrophobic peptides and considered a valuable by-product of flaxseed oil due to numerous health benefits. Currently applied acetone or methanol extraction could contaminate the feedstocks for further food-grade application. Using flaxseed cake as feedstock, this study established a practical method for preparing LO from pressed cake. Firstly, LO composition of 15 flaxseed cultivars was analyzed. Next, cold-pressed cake was milled and screened mechanically. The kernel and hull fractions were separated based on the disparity of their mechanical strength. Monitored by hyperspectral fluorescence, the LO-enriched kernel fraction separated from cold-pressed flaxseed cake was further used as feedstock for LO production. After ethanol extraction, partition, and precipitation, LOs were extracted from cold-pressed flaxseed cake with a purity of 91.4%. The proposed method could serve as feasible flaxseed cake valorization strategy and enable the preparation of other polar compounds such as flax lignan and mucilage.

Accurate recognition of rice plants based on visual and tactile sensing

BACKGROUND

Crop recognition is the basis of intelligent agricultural machine operations. Visual perception methods have achieved high recognition accuracy. However, the reliability of such methods is difficult to guarantee because of the complex environment of paddy fields. Tactile sensing methods are not affected by background or environmental interference, and have high reliability. However, in an ideal environment, the recognition accuracy is not as high as that of the visual method.

RESULTS

To balance the accuracy and reliability of rice plant recognition, a combined visual-tactile method was proposed in this study. A rice plant recognition device was developed with a poly(vinylidene fluoride) sensor embedded inside the device as a tactile perceptron and a graphic designed as a visual perceptron. The primary role of the tactile perceptron is to initially recognize rice plants and provide a time point for image capture for visual perception. The main role of the visual perceptron is to extract features from the captured images and recognize rice plants again. The results of tactile and visual recognition were eventually fused to achieve accurate recognition of rice plants.

CONCLUSION

The contact speed between the recognition perceptron and rice-weed was selected for the field performance test based on the real situation of paddy field operation. The results showed that the accuracy and reliability of rice plant recognition decreased as the travelling speed of the paddy field operation machine increased. The results of this study provide a basis for intelligent farm machinery operations in rice fields. © 2024 Society of Chemical Industry.

CrossU-Net: Dual-modality cross-attention U-Net for segmentation of precancerous lesions in gastric cancer

Gastric precancerous lesions (GPL) significantly elevate the risk of gastric cancer, and precise diagnosis and timely intervention are critical for patient survival. Due to the elusive pathological features of precancerous lesions, the early detection rate is less than 10%, which hinders lesion localization and diagnosis. In this paper, we provide a GPL pathological dataset and propose a novel method for improving the segmentation accuracy on a limited-scale dataset, namely RGB and Hyperspectral dual-modal pathological image Cross-attention U-Net (CrossU-Net). Specifically, we present a self-supervised pre-training model for hyperspectral images to serve downstream segmentation tasks. Secondly, we design a dual-stream U-Net-based network to extract features from different modal images. To promote information exchange between spatial information in RGB images and spectral information in hyperspectral images, we customize the cross-attention mechanism between the two networks. Furthermore, we use an intermediate agent in this mechanism to improve computational efficiency. Finally, we add a distillation loss to align predicted results for both branches, improving network generalization. Experimental results show that our CrossU-Net achieves accuracy and Dice of 96.53% and 91.62%, respectively, for GPL lesion segmentation, providing a promising spectral research approach for the localization and subsequent quantitative analysis of pathological features in early diagnosis.

The value of machine learning approaches in the diagnosis of early gastric cancer: a systematic review and meta-analysis

BACKGROUND

The application of machine learning (ML) for identifying early gastric cancer (EGC) has drawn increasing attention. However, there lacks evidence-based support for its specific diagnostic performance. Hence, this systematic review and meta-analysis was implemented to assess the performance of image-based ML in EGC diagnosis.

METHODS

We performed a comprehensive electronic search in PubMed, Embase, Cochrane Library, and Web of Science up to September 25, 2022. QUADAS-2 was selected to judge the risk of bias of included articles. We did the meta-analysis using a bivariant mixed-effect model. Sensitivity analysis and heterogeneity test were performed.

RESULTS

Twenty-one articles were enrolled. The sensitivity (SEN), specificity (SPE), and SROC of ML-based models were 0.91 (95% CI: 0.87-0.94), 0.85 (95% CI: 0.81-0.89), and 0.94 (95% CI: 0.39-1.00) in the training set and 0.90 (95% CI: 0.86-0.93), 0.90 (95% CI: 0.86-0.92), and 0.96 (95% CI: 0.19-1.00) in the validation set. The SEN, SPE, and SROC of EGC diagnosis by non-specialist clinicians were 0.64 (95% CI: 0.56-0.71), 0.84 (95% CI: 0.77-0.89), and 0.80 (95% CI: 0.29-0.97), and those by specialist clinicians were 0.80 (95% CI: 0.74-0.85), 0.88 (95% CI: 0.85-0.91), and 0.91 (95% CI: 0.37-0.99). With the assistance of ML models, the SEN of non-specialist physicians in the diagnosis of EGC was significantly improved (0.76 vs 0.64).

CONCLUSION

ML-based diagnostic models have greater performance in the identification of EGC. The diagnostic accuracy of non-specialist clinicians can be improved to the level of the specialists with the assistance of ML models. The results suggest that ML models can better assist less experienced clinicians in diagnosing EGC under endoscopy and have broad clinical application value.

Combined measurement of serum zinc with PSA ameliorates prostate cancer screening efficiency via support vector machine algorithms

Background

Early screening of prostate cancer (PCa) is pivotal but challenging in the clinical scenario due to the phenomena of false positivity or false negativity of some serological evaluations, e.g. PSA testing. Decline of serum Zn2+ levels in PCa patients reportedly plays a crucial role in early screening of PCa. Accordingly, we combined 4 indices comprising the serum levels of total PSA (tPSA), free PSA (fPSA), Zn2+ and demographic information (especially age) in order to ameliorate the efficacies of PCa screening with support vector machine (SVM) algorithms.

Methods

A total of 858 male patients with prostate disorders and 345 healthy male controls were enrolled. Patients' data included 4 variables and serum Zn2+ was quantified via a self-invented Zn2+ responsive AIE-based fluorescent probe as previously published. tPSA and fPSA were routinely determined by a chemiluminescent method. Mathematical simulations were conducted to establish a SVM model for the combined diagnostics with the four variables. Moreover, ROC and its characteristic AUC were also employed to evaluate the classification efficacy of the model. Sigmoid function was utilized to estimate corresponding probabilities of classifying the clinical subjects as per 5 grades, which were incorporated into our established prostate index (PI) stratification system.

Results

In SVM model, the mean AUC of the ROC with the quartet of variables was approximately 84% for PCa diagnosis, whereas the mean AUC of the ROCs with tPSA, fPSA, [Zn2+] or age alone was 64%, 62%, 55% and 59%, respectively. We further established an integrated prostate index (PI) stratification system with 5 grades and a software package to support clinicians in predicting PCa, with the accuracy of our risk stratification system being 83.3%, 91.6% and 83.3% in predicting normal, benign and PCa cases in corresponding groups. Follow-up findings especially MRI results and PI-RADS scores supported the reliability of this stratification platform as well.

Conclusion

Findings from our present study demonstrated that index combination via SVM algorithms may well facilitate clinicians in early differential screening of PCa. Meanwhile, our established PI stratification system based on SVM model and Sigmoid function provided substantial accuracy in preclinical risk prediction of developing prostate cancer.

Comparison of the diagnostic efficacy of mathematical models in distinguishing ultrasound imaging of breast nodules

This study compared the diagnostic efficiency of benign and malignant breast nodules using ultrasonographic characteristics coupled with several machine-learning models, including logistic regression (Logistics), partial least squares discriminant analysis (PLS-DA), linear support vector machine (Linear SVM), linear discriminant analysis (LDA), K-nearest neighbor (KNN), artificial neural network (ANN) and random forest (RF). The clinical information and ultrasonographic characteristics of 926 female patients undergoing breast nodule surgery were collected and their relationships were analyzed using Pearson's correlation. The stepwise regression method was used for variable selection and the Monte Carlo cross-validation method was used to randomly divide these nodule cases into training and prediction sets. Our results showed that six independent variables could be used for building models, including age, background echotexture, shape, calcification, resistance index, and axillary lymph node. In the prediction set, Linear SVM had the highest diagnosis rate of benign nodules (0.881), and Logistics, ANN and LDA had the highest diagnosis rate of malignant nodules (0.910~0.912). The area under the ROC curve (AUC) of Linear SVM was the highest (0.890), followed by ANN (0.883), LDA (0.880), Logistics (0.878), RF (0.874), PLS-DA (0.866), and KNN (0.855), all of which were better than that of individual variances. On the whole, the diagnostic efficacy of Linear SVM was better than other methods.

Emerging clinical applications in oncology for non-invasive multi- and hyperspectral imaging of cell and tissue autofluorescence

Hyperspectral and multispectral imaging of cell and tissue autofluorescence is an emerging technology in which fluorescence imaging is applied to biological materials across multiple spectral channels. This produces a stack of images where each matched pixel contains information about the sample's spectral properties at that location. This allows precise collection of molecularly specific data from a broad range of native fluorophores. Importantly, complex information, directly reflective of biological status, is collected without staining and tissues can be characterised in situ, without biopsy. For oncology, this can spare the collection of biopsies from sensitive regions and enable accurate tumour mapping. For in vivo tumour analysis, the greatest focus has been on oral cancer, whereas for ex vivo assessment head-and-neck cancers along with colon cancer have been the most studied, followed by oral and eye cancer. This review details the scope and progress of research undertaken towards clinical translation in oncology.

Hyperspectral imaging of lipids in biological tissues using near-infrared and shortwave infrared transmission mode: A pilot study

Label-free hyperspectral imaging (HSI) of lipids was demonstrated in the near-infrared (NIR) and shortwave infrared (SWIR) regions (950-1800 nm) using porcine tissue. HSI was performed in the transmission light-pass configuration, using a NIR-SWIR camera coupled with a liquid crystal tunable filter. The transmittance spectra of the regions of interest (ROIs), which correspond to the lipid and muscle areas in the specimen, were utilized for the spectrum unmixing. The transmittance spectra in ROIs were compared with those recorded by a spectrophotometer using samples of adipose and muscle. The lipid optical absorption bands at 1210 and 1730 nm were first used for the unmixing and mapping. Then, we performed the continuous multiband unmixing over the entire available spectral range, thereby, considering a combination of characteristic absorption bands of lipids, proteins, and water. The enhanced protocol demonstrates the ability to visualize small adipose inclusions of 1-10 μm size.

Machine Learning Enhanced Optical Spectroscopy for Disease Detection

Optical spectroscopy plays an important role in disease detection. Improving the sensitivity and specificity of spectral detection has great importance in the development of accurate diagnosis. The development of artificial intelligence technology provides a great opportunity to improve the detection accuracy through machine learning methods. In this Perspective, we focus on the combination of machine learning methods with the optical spectroscopy methods widely used for disease detection, including absorbance, fluorescence, scattering, FTIR, terahertz, etc. By comparing the spectral analysis with different machine learning methods, we illustrate that the support vector machine and convolutional neural network are most effective, which have potential to further improve the classification accuracy to distinguish disease subtypes if these machine learning methods are used. This Perspective broadens the scope of optical spectroscopy enhanced by machine learning and will be useful for the development of disease detection.

Assessment of angle-dependent spectral distortion to develop accurate hyperspectral endoscopy

Hyperspectral endoscopy has shown its potential to improve disease diagnosis in gastrointestinal tracts. Recent approaches in developing hyperspectral endoscopy are mainly focusing on enhancing image speed and quality of spectral information under a clinical environment, but there are many issues in obtaining consistent spectral information due to complicated imaging conditions, including imaging angle, non-uniform illumination, working distance, and low reflected signal. We quantitatively investigated the effect of imaging angle on the distortion of spectral information by exploiting a bifurcated fiber, spectrometer, and tissue-mimicking phantom. Spectral distortion becomes severe as increasing the angle of the imaging fiber or shortening camera exposure time for fast image acquisition. Moreover, spectral ranges from 450 to 550 nm are more susceptible to the angle-dependent spectral distortion than longer spectral ranges. Therefore, imaging angles close to normal and longer target spectral ranges with enough detector exposure time could minimize spectral distortion in hyperspectral endoscopy. These findings will help implement clinical HSI endoscopy for the robust and accurate measurement of spectral information from patients in vivo.

Review on the Application of Hyperspectral Imaging Technology of the Exposed Cortex in Cerebral Surgery

The study of brain science is vital to human health. The application of hyperspectral imaging in biomedical fields has grown dramatically in recent years due to their unique optical imaging method and multidimensional information acquisition. Hyperspectral imaging technology can acquire two-dimensional spatial information and one-dimensional spectral information of biological samples simultaneously, covering the ultraviolet, visible and infrared spectral ranges with high spectral resolution, which can provide diagnostic information about the physiological, morphological and biochemical components of tissues and organs. This technology also presents finer spectral features for brain imaging studies, and further provides more auxiliary information for cerebral disease research. This paper reviews the recent advance of hyperspectral imaging in cerebral diagnosis. Firstly, the experimental setup, image acquisition and pre-processing, and analysis methods of hyperspectral technology were introduced. Secondly, the latest research progress and applications of hyperspectral imaging in brain tissue metabolism, hemodynamics, and brain cancer diagnosis in recent years were summarized briefly. Finally, the limitations of the application of hyperspectral imaging in cerebral disease diagnosis field were analyzed, and the future development direction was proposed.

Automatic Recognition of Colon and Esophagogastric Cancer with Machine Learning and Hyperspectral Imaging

There are approximately 1.8 million diagnoses of colorectal cancer, 1 million diagnoses of stomach cancer, and 0.6 million diagnoses of esophageal cancer each year globally. An automatic computer-assisted diagnostic (CAD) tool to rapidly detect colorectal and esophagogastric cancer tissue in optical images would be hugely valuable to a surgeon during an intervention. Based on a colon dataset with 12 patients and an esophagogastric dataset of 10 patients, several state-of-the-art machine learning methods have been trained to detect cancer tissue using hyperspectral imaging (HSI), including Support Vector Machines (SVM) with radial basis function kernels, Multi-Layer Perceptrons (MLP) and 3D Convolutional Neural Networks (3DCNN). A leave-one-patient-out cross-validation (LOPOCV) with and without combining these sets was performed. The ROC-AUC score of the 3DCNN was slightly higher than the MLP and SVM with a difference of 0.04 AUC. The best performance was achieved with the 3DCNN for colon cancer and esophagogastric cancer detection with a high ROC-AUC of 0.93. The 3DCNN also achieved the best DICE scores of 0.49 and 0.41 on the colon and esophagogastric datasets, respectively. These scores were significantly improved using a patient-specific decision threshold to 0.58 and 0.51, respectively. This indicates that, in practical use, an HSI-based CAD system using an interactive decision threshold is likely to be valuable. Experiments were also performed to measure the benefits of combining the colorectal and esophagogastric datasets (22 patients), and this yielded significantly better results with the MLP and SVM models.

New evidence from hyperspectral imaging analysis on the effect of photobiomodulation therapy on normal skin oxygenation

The aim of this study was to assess the changes induced by photobiomodulation therapy in oxygenation of normal skin and underlying tissue using hyperspectral imaging combined with a chemometric regression approach. Eleven healthy adult volunteers were enrolled in this study. The dorsal side of the left hand of each subject was exposed to photobiomodulation therapy, while the correspondent side of the right hand was used as a control (placebo effect). Laser irradiation was performed with a laser diode system (635 nm, 15mW, 9 J/cm²) for 900 s. Changes in skin oxygenation were assessed before and after applying the photobiomodulation therapy and placebo using the hyperspectral imaging. Hyperspectral data analysis showed that variations of oxyhemoglobin and deoxyhemoglobin concentrations had no statistical significance in both groups. In conclusion, photobiomodulation therapy does not induce changes in oxyhemoglobin and deoxyhemoglobin concentrations in the normal skin measured from hyperspectral images, at least at λ = 635 nm and 900-s exposure time.

Feedforward Artificial Neural Network-Based Colorectal Cancer Detection Using Hyperspectral Imaging: A Step towards Automatic Optical Biopsy

Simple Summary

Detection of colorectal carcinoma is performed visually by investigators and is confirmed pathologically. With hyperspectral imaging, an expanded spectral range of optical information is now available for analysis. The acquired recordings were analyzed with a neural network, and it was possible to differentiate tumor from healthy mucosa in colorectal carcinoma by automatic classification with high reliability. Classification and visualization were performed based on a four-layer perceptron neural network. Based on a neural network, the classification of CA or AD resulted in a sensitivity of 86% and a specificity of 95%, by means of leave-one-patient-out cross-validation. Additionally, significant differences in terms of perfusion parameters (e.g., oxygen saturation) related to tumor staging and neoadjuvant therapy were observed. This is a step towards optical biopsy.

Abstract

Currently, colorectal cancer (CRC) is mainly identified via a visual assessment during colonoscopy, increasingly used artificial intelligence algorithms, or surgery. Subsequently, CRC is confirmed through a histopathological examination by a pathologist. Hyperspectral imaging (HSI), a non-invasive optical imaging technology, has shown promising results in the medical field. In the current study, we combined HSI with several artificial intelligence algorithms to discriminate CRC. Between July 2019 and May 2020, 54 consecutive patients undergoing colorectal resections for CRC were included. The tumor was imaged from the mucosal side with a hyperspectral camera. The image annotations were classified into three groups (cancer, CA; adenomatous margin around the central tumor, AD; and healthy mucosa, HM). Classification and visualization were performed based on a four-layer perceptron neural network. Based on a neural network, the classification of CA or AD resulted in a sensitivity of 86% and a specificity of 95%, by means of leave-one-patient-out cross-validation. Additionally, significant differences in terms of perfusion parameters (e.g., oxygen saturation) related to tumor staging and neoadjuvant therapy were observed. Hyperspectral imaging combined with automatic classification can be used to differentiate between CRC and healthy mucosa. Additionally, the biological changes induced by chemotherapy to the tissue are detectable with HSI.

Early Diagnosis of Type 2 Diabetes Based on Near-Infrared Spectroscopy Combined With Machine Learning and Aquaphotomics

Early diagnosis is important to reduce the incidence and mortality rate of diabetes. The feasibility of early diagnosis of diabetes was studied via near-infrared spectra (NIRS) combined with a support vector machine (SVM) and aquaphotomics. Firstly, the NIRS of entire blood samples from the population of healthy, pre-diabetic, and diabetic patients were obtained. The spectral data of the entire spectra in the visible and near-infrared region (400–2,500 nm) were used as the research object of the qualitative analysis. Secondly, several preprocessing steps including multiple scattering correction, variable standardization, and first derivative and second derivative steps were performed and the best pretreatment method was selected. Finally, for the early diagnosis of diabetes, models were established using SVM. The first overtone of water (1,300–1,600 nm) was used as the research object for an aquaphotomics model, and the aquagram of the healthy group, pre-diabetes, and diabetes groups were drawn using 12 water absorption patterns for the early diagnosis of diabetes. The results of SVM showed that the highest accuracy was 97.22% and the specificity and sensitivity were 95.65 and 100%, respectively when the pretreatment method of the first derivative was used, and the best model parameters were c = 18.76 and g = 0.008583.The results of the aquaphotomics model showed clear differences in the 1,400–1,500 nm region, and the number of hydrogen bonds in water species (1,408, 1,416, 1,462, and 1,522 nm) was evidently correlated with the occurrence and development of diabetes. The number of hydrogen bonds was the smallest in the healthy group and the largest in the diabetes group. The suggested reason is that the water matrix of blood changes with the worsening of blood glucose metabolic dysfunction. The number of hydrogen bonds could be used as biomarkers for the early diagnosis of diabetes. The result show that it is effective and feasible to establish an accurate and rapid early diagnosis model of diabetes via NIRS combined with SVM and aquaphotomics.

Pre-Cancerous Stomach Lesion Detections with Multispectral-Augmented Endoscopic Prototype

In this paper, we are interested in the in vivo detection of pre-cancerous stomach lesions. Pre-cancerous lesions are unfortunately rarely explored in research papers as most of them are focused on cancer detection or conducted ex-vivo. For this purpose, a novel prototype is introduced. It consists of a standard endoscope with multispectral cameras, an optical setup, a fiberscope, and an external light source. Reflectance spectra are acquired in vivo on 16 patients with a healthy stomach, chronic gastritis, or intestinal metaplasia. A specific pipeline has been designed for the classification of spectra between healthy mucosa and different pathologies. The pipeline includes a wavelength clustering algorithm, spectral features computation, and the training of a classifier in a “leave one patient out” manner. Good classification results, around 80%, have been obtained, and two attractive wavelength ranges were found in the red and near-infrared ranges: [745, 755 nm] and [780, 840 nm]. The new prototype and the associated results give good arguments in favor of future common use in operating rooms, during upper gastrointestinal exploration of the stomach for the detection of stomach diseases.

DA-Based Parameter Optimization of Combined Kernel Support Vector Machine for Cancer Diagnosis

As is well known, the correct diagnosis for cancer is critical to save patients’ lives. Support vector machine (SVM) has already made an important contribution to the field of cancer classification. However, different kernel function configurations and their parameters will significantly affect the performance of SVM classifier. To improve the classification accuracy of SVM classifier for cancer diagnosis, this paper proposed a novel cancer classification algorithm based on the dragonfly algorithm and SVM with a combined kernel function (DA-CKSVM) which was constructed from a radial basis function (RBF) kernel and a polynomial kernel. Experiments were performed on six cancer data sets from University of California, Irvine (UCI) machine learning repository and two cancer data sets from Cancer Program Legacy Publication Resources to evaluate the validity of the proposed algorithm. Compared with four well-known algorithms: dragonfly algorithm-SVM (DA-SVM), particle swarm optimization-SVM (PSO-SVM), bat algorithm-SVM (BA-SVM), and genetic algorithm-SVM (GA-SVM), the proposed algorithm was able to find the optimal parameters of SVM classifier and achieved better classification accuracy on cancer datasets.