In vivo detection of oral precancer using a fluorescence-based, in-house-fabricated device: a Mahalanobis distance-based classification

A smartphone-based standalone fluorescence spectroscopy tool for cervical precancer diagnosis in clinical conditions

Real-time prediction about the severity of noncommunicable diseases like cancers is a boon for early diagnosis and timely cure. Optical techniques due to their minimally invasive nature provide better alternatives in this context than the conventional techniques. The present study talks about a standalone, field portable smartphone-based device which can classify different grades of cervical cancer on the basis of the spectral differences captured in their intrinsic fluorescence spectra with the help of AI/ML technique. In this study, a total number of 75 patients and volunteers, from hospitals at different geographical locations of India, have been tested and classified with this device. A classification approach employing a hybrid mutual information long short-term memory model has been applied to categorize various subject groups, resulting in an average accuracy, specificity, and sensitivity of 96.56%, 96.76%, and 94.37%, respectively using 10-fold cross-validation. This exploratory study demonstrates the potential of combining smartphone-based technology with fluorescence spectroscopy and artificial intelligence as a diagnostic screening approach which could enhance the detection and screening of cervical cancer.

Cervical pre-cancer classification using entropic features and CNN: In vivo validation with a handheld fluorescence probe

Cervical cancer is one of the most prevalent forms of cancer, with a lengthy latent period and a gradual onset phase. Conventional techniques are found to be severely lacking in real time detection of disease progression which can greatly enhance the cure rate. Due to their high sensitivity and specificity, optical techniques are emerging as reliable tools, particularly in case of cancer. It has been seen that biochemical changes are better highlighted through intrinsic fluorescence devoid of interference from absorption and scattering. Its effectiveness in in-vivo conditions is affected by the fact that the intrinsic spectral signatures vary from patient to patient, as well as in different population groups. Here, we overcome this limitation by collectively enumerating the subtle changes in the spectral profiles and correlations through an information theory based entropic approach, which significantly amplifies the minute spectral variations. In conjunction with artificial intelligence (AI)/machine learning (ML) tools, it yields high specificity and sensitivity with a small dataset from patients in clinical conditions, without artificial augmentation. We have used an in-house developed handheld probe (i-HHP) for extracting intrinsic fluorescence spectra of human cervix from 110 different subjects drawn from diverse population groups. The average classification accuracy of the proposed methodology using 10-fold cross validation is 93.17%. A combination of polarised fluorescence spectra from i-HHP and the proposed classifier is proven to be minimally invasive with the ability to diagnose patients in real time. This paves the way for effective use of relatively smaller sized sensitive fluorescence data with advanced AI/ML tools for early cervical cancer detection in clinics.

Detection of oral mucosal lesions by the fluorescence spectroscopy and classification of cancerous stages by support vector machine

Detection of oral mucosal lesions has been performed by an in-house developed fluorescence-based portable device in the present study. A laser diode of 405 nm wavelength and a UV-visible spectrometer are utilized in the portable device as excitation and detection sources. At the 405 nm excitation wavelength, the flavin adenine dinucleotide (FAD) band at 500 nm and three porphyrin bands at 634, 676, and 703 nm are observed in the fluorescence spectrum of the oral cavity tissue. We have conducted this clinical study on a total of 189 tissue sites of 36 oral squamous cell carcinoma (OSCC) patients, 18 dysplastic (precancerous) patients, and 34 volunteers. Analysis of the fluorescence data has been performed by using the principal component analysis (PCA) method and support vector machine (SVM) classifier. PCA is applied first in the spectral data to reduce the dimension, and then classification among the three groups has been executed by employing the SVM. The SVM classifier includes linear, radial basis function (RBF), polynomial, and sigmoid kernels, and their classification efficacies are computed. Linear and RBF kernels on the testing data sets differentiated OSCC and dysplasia to normal with an accuracy of 100% and OSCC to dysplasia with an accuracy of 95% and 97%, respectively. Polynomial and sigmoid kernels showed less accuracy values among the groups ranging from 48 to 88% and 51 to 100%, respectively. The result indicates that fluorescence spectroscopy and the SVM classifier can help to identify early oral mucosal lesions with significant high accuracy.

Multimodal fluorescence imaging and spectroscopic techniques for oral cancer screening: a real-time approach

The survival rate of oral squamous cell carcinoma (OSCC) patients is very poor, but it can be improved using highly sensitive, specific, and accurate techniques. Autofluorescence and fluorescence techniques are very sensitive and helpful in cancer screening; being directly linked with the molecular levels of human tissue, they can be used as a quantitative tool for cancer detection. Here, we report the development of multi-modal autofluorescence and fluorescence imaging and spectroscopic (MAF-IS) smartphone-based systems for fast and real-time oral cancer screening. MAF-IS system is indigenously developed and offers the advantages of being a low-cost, handy, non-contact, non-invasive, and easily operable device that can be employed in hospitals, including low-resource settings. In this study, we report the results of 43 individuals with 28 OSCC and 15 oral potentially malignant disorders (OPMDs), i.e., epithelial dysplasia and oral submucous fibrosis, using the developed devices. We observed a red shift in fluorescence emission spectrain vivo. We found red-shift of 7.72 ± 6 nm, 3 ± 4.36 nm, and 1.33 ± 0.47 nm in the case of OSCC, epithelial dysplasia, and oral submucous fibrosis, respectively, compared to normal. The results were compared with histopathology and found to be consistent. Further, the MAF-IS system provides results in real-time with higher accuracy and sensitivity compared to devices using a single modality. Our system can achieve an accuracy of 97% with sensitivity and specificity of 100% and 94.7%, respectively, even with a smaller number of patients (28 patients of OSCC). The proposed MAF-IS device has great potential for fast screening and diagnosis of oral cancer in the future.

Deep-Learning-Based Automated Identification and Visualization of Oral Cancer in Optical Coherence Tomography Images

Early detection and diagnosis of oral cancer are critical for a better prognosis, but accurate and automatic identification is difficult using the available technologies. Optical coherence tomography (OCT) can be used as diagnostic aid due to the advantages of high resolution and non-invasion. We aim to evaluate deep-learning-based algorithms for OCT images to assist clinicians in oral cancer screening and diagnosis. An OCT data set was first established, including normal mucosa, precancerous lesion, and oral squamous cell carcinoma. Then, three kinds of convolutional neural networks (CNNs) were trained and evaluated by using four metrics (accuracy, precision, sensitivity, and specificity). Moreover, the CNN-based methods were compared against machine learning approaches through the same dataset. The results show the performance of CNNs, with a classification accuracy of up to 96.76%, is better than the machine-learning-based method with an accuracy of 92.52%. Moreover, visualization of lesions in OCT images was performed and the rationality and interpretability of the model for distinguishing different oral tissues were evaluated. It is proved that the automatic identification algorithm of OCT images based on deep learning has the potential to provide decision support for the effective screening and diagnosis of oral cancer.

In-vivo Testing of Oral Mucosal Lesions with an In-house Developed Portable Imaging Device and Comparison with Spectroscopy Results

Progression of oral mucosal lesions is generally marked by changes in the concentration of the intrinsic fluorophores such as collagen, nicotinamide adenine dinucleotide (NADH), flavin adenine dinucleotide (FAD) and porphyrin present in the human oral tissue. In this study, we have probed the changes in FAD and porphyrin by exciting with 405 nm laser light on different sites (tongue, buccal mucosa, lip etc.) of the oral cavity. Testing has been done by an in-house developed fluorescence-based portable imaging device on oral squamous cell carcinoma (OSCC) patients, dysplastic patients and control (normal) group. Fluorescence images recorded from OSCC and dysplastic patients have displayed an enhancement in the red band (porphyrin) as compared to those from the normal volunteers. Porphyrin to FAD intensity ratio (I_Porphyrin/I_FAD), referred to red to green ratio (I_red/I_green) has been taken as the diagnostic marker for classification among the groups. Receiver operating characteristic (ROC) analysis applied on I_Porphyrin/I_FAD is able to discriminate OSCC to normal, dysplasia to normal and OSCC to dysplasia with sensitivities of 100%, 81%, 92% and specificities of 100%, 93% and 92% respectively. Fluorescence imaging probe can capture a large area of oral lesions in a single scan and hence would be useful for initial scanning. On comparison with spectroscopy studies performed by our group, it is found that combining both spectroscopy and imaging as a device may be effective for the early detection of oral lesions. This clinical study was registered on the date 13/10/2017 in the clinical trials registry-India (CTRI) with registration number CTRI/2017/10/010102.

Hippocampus as a sorter and reverberatory integrator of sensory inputs

The hippocampus must be capable of sorting and integrating multiple sensory inputs separately but simultaneously. However, it remains to be elucidated how the hippocampus executes these processes simultaneously during learning. Here we found that synchrony between conditioned stimulus (CS)-, unconditioned stimulus (US)- and future retrieval-responsible cells occurs in the CA1 during the reverberatory phase that emerges after sensory inputs have ceased, but not during CS and US inputs. Mutant mice lacking N-methyl-D-aspartate receptors (NRs) in CA3 showed a cued-fear memory impairment and a decrease in synchronized reverberatory activities between CS- and US-responsive CA1 cells. Optogenetic CA3 silencing at the reverberatory phase during learning impaired cued-fear memory. Thus, the hippocampus uses reverberatory activity to link CS and US inputs, and avoid crosstalk during sensory inputs.

Machine learning in point-of-care automated classification of oral potentially malignant and malignant disorders: a systematic review and meta-analysis

Machine learning (ML) algorithms are becoming increasingly pervasive in the domains of medical diagnostics and prognostication, afforded by complex deep learning architectures that overcome the limitations of manual feature extraction. In this systematic review and meta-analysis, we provide an update on current progress of ML algorithms in point-of-care (POC) automated diagnostic classification systems for lesions of the oral cavity. Studies reporting performance metrics on ML algorithms used in automatic classification of oral regions of interest were identified and screened by 2 independent reviewers from 4 databases. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. 35 studies were suitable for qualitative synthesis, and 31 for quantitative analysis. Outcomes were assessed using a bivariate random-effects model following an assessment of bias and heterogeneity. 4 distinct methodologies were identified for POC diagnosis: (1) clinical photography; (2) optical imaging; (3) thermal imaging; (4) analysis of volatile organic compounds. Estimated AUROC across all studies was 0.935, and no difference in performance was identified between methodologies. We discuss the various classical and modern approaches to ML employed within identified studies, and highlight issues that will need to be addressed for implementation of automated classification systems in screening and early detection.

Detection of inaccessible head and neck lesions using human saliva and fluorescence spectroscopy

Head and neck cancer detection using fluorescence spectroscopy from human saliva is reported here. This study has been conducted on squamous cell carcinoma (SCC), and dysplastic (precancer) and control (normal) groups using an in-house developed compact set-up. Fluorescence set-up consists of a 375-nm laser diode and optical components. Spectral bands of flavin adenine dinucleotide (FAD), porphyrins, and Raman are observed in the spectral range of 400 to 800 nm. Presence of FAD and porphyrin bands in human saliva is confirmed by the liquid phantoms of FAD and porphyrin. Significant differences in fluorescence intensities among all the three groups are observed. Three spectral ranges from 455 to 600, 605 to 770, and 400 to 800 nm are selected for each group and area values under each spectral range are computed. To differentiate among the groups, receiver operating characteristic (ROC) analysis is employed on the area values. ROC differentiates among the groups with accuracies of 98%, 92.85%, and 81.13% respectively in the spectral ranges of 400 to 800 nm. However, in other two spectral ranges (455 to 600 and 605 to 770 nm), low accuracy values are found. Obtained accuracy values indicate that selection of human saliva for head and neck cancer detection may be a good alternative.

Machine-Learning Assisted Discrimination of Precancerous and Cancerous from Healthy Oral Tissue Based on Multispectral Autofluorescence Lifetime Imaging Endoscopy

Multispectral autofluorescence lifetime imaging (maFLIM) can be used to clinically image a plurality of metabolic and biochemical autofluorescence biomarkers of oral epithelial dysplasia and cancer. This study tested the hypothesis that maFLIM-derived autofluorescence biomarkers can be used in machine-learning (ML) models to discriminate dysplastic and cancerous from healthy oral tissue. Clinical widefield maFLIM endoscopy imaging of cancerous and dysplastic oral lesions was performed at two clinical centers. Endoscopic maFLIM images from 34 patients acquired at one of the clinical centers were used to optimize ML models for automated discrimination of dysplastic and cancerous from healthy oral tissue. A computer-aided detection system was developed and applied to a set of endoscopic maFLIM images from 23 patients acquired at the other clinical center, and its performance was quantified in terms of the area under the receiver operating characteristic curve (ROC-AUC). Discrimination of dysplastic and cancerous from healthy oral tissue was achieved with an ROC-AUC of 0.81. This study demonstrates the capabilities of widefield maFLIM endoscopy to clinically image autofluorescence biomarkers that can be used in ML models to discriminate dysplastic and cancerous from healthy oral tissue. Widefield maFLIM endoscopy thus holds potential for automated in situ detection of oral dysplasia and cancer.

Rab6c is a new target of miR‑218 that can promote the progression of bladder cancer

Bladder cancer has high morbidity and mortality rates among the male genitourinary system tumor types. MicroRNA‑218 (miR‑218) is associated with the development of a variety of cancer types, including bladder cancer. Rab6c is a member of the Rab family and is involved in drug resistance in MCF7 cells. The aim of the present study was to clarify the relationship between Rab6c and miR‑218 in bladder cancer cell lines. In this study, the expression levels of miR‑218 and Rab6c were evaluated via reverse transcription‑quantitative PCR and western blotting, respectively. The association between Rab6c and miR‑218 was recognized via TargetScan analysis and dual luciferase reporter gene detection. Cell proliferation was analyzed using Cell Counting Kit‑8 and colony formation assays, and the invasive ability was measured via Transwell assays. Rab6c was highly expressed in bladder cancer, while miR‑218 had abnormally low expression in bladder cancer. In addition, there was a mutual regulation between Rab6c and miR‑218 in bladder cancer. It was found that overexpression of Rab6c significantly enhanced the proliferation, colony formation and invasion of T24 and EJ cells. Furthermore, miR‑218 overexpression blocked the promoting effects of Rab6c on the malignant behavior of bladder cancer cells. Thus, Rab6c promotes the proliferation and invasion of bladder cancer cells, while miR‑218 has the opposite effect, which may provide a novel insight for the treatment of bladder cancer.

Classification of oral salivary gland tumors based on texture features in optical coherence tomography images

Currently, the diagnoses of oral diseases primarily depend on the visual recognition of experienced clinicians. It has been proven that automatic recognition based on images can support clinical decision-making by extracting and analyzing objective hidden information. In recent years, optical coherence tomography (OCT) has become a powerful optical imaging technique with the advantages of high resolution and non-invasion. In our study, a dataset composed of four kinds of oral salivary gland tumors (SGTs) was obtained from a homemade swept-source OCT, including two benign and two malignant tumors. Seventy-six texture features were extracted from OCT images to create computational models of diseases. It was demonstrated that the artificial neural network (ANN) based on principal component analysis (PCA) can obtain high diagnostic sensitivity and specificity (higher than 99%) for these four kinds of tumors. The classification accuracy of each tumor is larger than 99%. In addition, the performances of two classifiers (ANN and support vector machine) were quantitatively evaluated based on SGTs. It was proven that the texture features in OCT images provided objective information to classify oral tumors.

Studying the Degree of Tooth Enamel Mineralization through Raman Spectroscopy in Various Spectral Ranges

In vitro and in vivo methods of Raman spectroscopy have been developed to assess the degree of mineralization of the enamel of different functional groups. This article presents comparative studies that were carried out using scanning Raman microspectroscopy with various sources of laser excitation with wavelengths of 532, 785, and 1064 nm. It is shown that the intensity of Raman scattering of enamel can be a measure of its thickness. The obtained dependence of the Raman scattering intensity on the distance from the incisal edge is in good agreement with the literature data, where two independent methods (computer tomography and electron microscopy) are used to determine the enamel thickness values. The proposed methods can be considered as potential quantitative methods for express diagnostics of the state of tooth enamel in vivo.

Use of autofluorescence and fluorescent probes as a potential diagnostic tool for oral cancer: A systematic review

INTRODUCTION

The prognosis of patients with Oral squamous cell carcinoma (OSCC) are directly related to the stage of development of the tumor at the time of diagnosis, but it is estimated an average delay in diagnosis of 2-5 months. New non-invasive techniques for the early diagnosis of OSCC are being developed, such as methodologies to detect spectral changes of tumor cells. We conducted a systematic review to analyze the potential use of autofluorescence and/or fluorescent probes for OSCC diagnosis.

MATERIAL AND METHODS

Four databases (PubMed, Scopus, Embase and Web of Science) were used as research sources. Protocol was registered with PROSPERO. It was included studies that evaluated tissue autofluorescence and/or used fluorescent probes as a method of diagnosing and/or treatment of oral cancer in humans.

RESULTS

Forty-five studies were selected for this systematic review, of which 28 dealt only with autofluorescence, 18 on fluorescent probes and 1 evaluated both methods. The VELscope® was the most used device for autofluorescence, exhibiting sensitivity (33%-100%) and specificity (12%-88.6%). 5-Aminolevulinic acid (5-ALA) was the most used fluorescent probe, exhibiting high sensitivity (90%-100%) and specificity (51.3%-96%). Hypericin, rhodamine 6 G, rhodamine 610, porphyrin and γ-glutamyl hydroxymethyl rhodamine green have also been reported.

CONCLUSION

Thus, the autofluorescence and fluorescent probes can provide an accurate diagnosis of oral cancer, assisting the dentist during daily clinical activity, but it is not yet possible to suggest that this method may replace histopathological examination.