MOISES OLIVEIRA DOS SANTOS

MOISES OLIVEIRA DOS SANTOS

Página do item completo

Resultados de Busca

Agora exibindo 1 - 10 de 42

Evaluation of calcified mitral valves after Er,Cr:YSGG irradiation using Optical Coherence Tomography
2022 - DEL-VALLE, MATHEUS; CARVALHO, MARCELO; SANTOS, MOISES; PINTO, NATHALI; JATENE, FABIO; POMERANTZEFF, PABLO; BRANDAO, CARLOS; ZEZELL, DENISE
Mitral valve is responsible to control the left atrium-ventricle blood flux. Mitral stenosis is a disease that occurs in consequence of calcification and fibrosis on the cuspids of the valve. Diagnosis can be performed using echocardiography.Many treatments are possible, and one of them is commissurotomy (surgical approach).High intensity laser irradiation may be a new strategy for this surgical technique[1], and the optical coherence tomography (OCT) may contribute to the valve evaluation[2], asit provides higherspatialresolutionin exchange of lower penetrationthan ultrasonography. In this way, the aim of this study is to evaluate laser irradiation effectsincalcified mitral valvesusing OCTand digital processing.To that, it was conducted an ex-vivostudywith four human mitral valvessamples,obtained from valve replacement surgeries in the Heart Institute.The samples were splitin four groups: scalpel cut, laser cut, scalpel debridement and laser debridement.Cutting and debridement procedures were performed in calcified regions of the valves, usinga disposable scalpelbladeand anEr,Cr:YSGG laser(Waterlase; Biolase Inc., CA, USA), emitting at 2780 nm. The laser parameters were set at power = 1.6W, frequency = 20 Hz, energy density = 28.3J/cm2,pulse duration = 700 μs, 15% of water and 15% of air.The imaging was performed using a spectraldomain OCT system(Callisto110C1;ThorLabs Inc., NJ, USA).It was acquired10 B-scans per sample, 5 inprocedures regions and 5 in sound regions. The Optical Attenuation Coefficient (OAC) was calculated by comparing a beer-lambert like equation to exponential fittings of the A-scans[3].The distribution and normality of variances were tested using Shapiro-Wilk test,and statistical comparison was performed using one-way ANOVA and Tukey’s post hoc. All tests considered a level of significance of 5%.The FigureAshows a representative B-scan of a visibly calcified region, where a pattern of higher intensities can be observed.Thispattern is related tomorphological and optical changes, mainly a refractive index change, due to calcium presence in the valve tissue.This B-scan was acquired only to understand the calcified tissue aspect, as the procedures regions does notpresent visibly largecalcium stones.The Figure Bshowsthe statisticalanalysis, where the sound OAC values, as a mean of all sound regions, presented a significant statistical difference in comparison to scalpel groups, while no difference waspresentedin relation to laser groups. Higher OAC values are related to anaugmentation of the light backscatteringdue to calcium refractive index, leading to a change of lightpropagation in tissue-calcium interfaces.This finding indicates thatthe laser procedures promoted a better removal of calcified tissue than the scalpelmethods, which can be related to tissue-ablation interaction.Furthermore, the statistical difference between scalpel cut group and both laser groups suggests that the scalpel needs more wear interaction with the tissue, such as in the debridement procedure, being unable to significatively remove the calcification in a single cut.This study points the Er,Cr:YSGG and the OCT as potential techniques for the calcified tissue removal and evaluation,respectively, duringmitral valvessurgeries, although further studieswith higher sample numbermust be performed.
FTIR imaging on glass substrates evaluation of histological skin burn injuries specimens treated by femtosecond laser pulses
2022 - ZEZELL, DENISE; CASTRO, PEDRO; DEL-VALLE, MATHEUS; CAMILLO-SILVA, CARLOS; SAMAD, RICARDO; DE ROSSI, WAGNER; SANTOS, MOISES
Burn injuries continue to be one of the leading causes of unintentional death and injury in low- and middle-income countries [1]. Burns are considered an important public health problem, because in addition to physical problems that can lead the patient to death, they cause psychological and social damage. An estimated 180,000 deaths every year are caused by burns [2]. The use of infrared (IR) spectroscopy for studying biological specimens is nowadays a wide and active area of research. The IR microspectroscopy has proved to be an ideal tool for investigating the biochemical composition of biological samples at the microscopic scale, as well as its fast, sensitive, and label-free nature [3]. IR image spectral histopathology has shown great promise as an important diagnostic tool, with the potential to complement current pathological methods, reducing subjectivity in biopsy samples analysis. However, the use of IR transmissive substrates which are both fragile and prohibitively very expensive, hinder the clinical translation. The goal of this study is to evaluate the potential of discriminating healing process, in burned skin specimens treated with ultrashort pulses laser 3 days after the burn. This study is considering a previous paper [4], in which it analyzed only micro-ATR-FTIR spectra of a frozen sample point. The specimens were obtained from third degree burn wound. The wounds treatment were performed three days after the burn, and the animals were sacrificed 3 and 14 days post-treatment. Using coverslipped H&E stained tissue on glass from previous histopathological analysis and applying the analytical techniques PCA and K-means on N−H, O−H, and C−H stretching regions occurring at 2500−3800 cm−1 (high wavenumber region), were possible to discriminate burned epidermal and dermal regions from irradiated in same regions on sample. In the figures is shown the average spectrum at (a) day 3 and (b) day 14. , in both there were increase of burned+laser treated bands. The great potential of this study was to analyse coverslipped H&E stained tissue on glass, without compromising the histopathologist practices and contribute for clinical translation.
Breast cancer subtypes diagnostic via high performance supervised machine learning
2022 - FAROOQ, SAJID; DEL-VALLE, MATHEUS; SANTOS, MOISES O. dos; NASCIMENTO, SOFIA; BERNARDES, EMERSON S.; ZEZELL, DENISE M.
Aim: Breast cancer molecular subtypes are being used to improve clinical decision. The Fourier transform infrared (FTIR) spectroscopic imaging, which is a powerful and non-destructive technique, allows performing a non-perturbative and labelling free extraction of biochemical information towards diagnosis and evaluation for cell functionality. However, methods of measurements of large areas of cells demand a long time to achieve high quality images, making its clinical use impractical because of speed of data acquisition and dearth of optimized computational procedures. In order to cope with these challenges, Machine learning (ML) technologies can facilitate to obtain accurate prognosis of Breast Cancer (BC) subtypes with high action ability and accuracy. Methods: Here we propose a ML algorithm based method to distinguish computationally BC cell lines. The method is developed by coupling K neighbors Classifier (KNN) with Neighborhood Component Analysis (NCA) and NCA-KNN methods enables to identify BC subtypes without increasing model size as well additional parameters. Results: By incorporating FTIR imaging data, we show that using NCA-KNN method, the classification accuracies, specificities and sensitivities improve up to 97%, even at very low co-added scan (S_4). Moreover, a clear distinctive accuracy difference of our proposed method was obtained in comparison with other ML supervised models. Conclusion: For confirming our model results performance, the cross validation (k fold = 10) and receiver operation characteristics (ROC) curve were used and found in great agreement, suggest a potential diagnostic method for BC subtypes, even with small co-added scan < 8 at low spectral resolution (4 cm-1).
Breast cancer subtype classification using a one-dimensional convolutional neural network in hyperspectral images
2022 - DEL-VALLE, MATHEUS; SANTOS, MOISES O. dos; BERNARDES, EMERSON S.; ZEZELL, DENISE M.
FTIR spectroscopy imaging in addition to deep learning is a potential tool for breast cancer subtype classification, where accuracies higher than 86% can be achieved to predict among all subtypes.
Breast tissue diagnosis using artificial intelligence applied to FTIR spectroscopy images
2021 - VALLE, MATHEUS del; SANTOS, MOISES O. dos
The estimative of new breast cancer cases was of 2.1 million of new breast cancer cases in 2018, hence being the most incident type of cancer in women. The improvement of its diagnosis has been the aim of many researchers, including vibrational spectroscopy teams. With the advancement of the artificial intelligence, a field of computer science to enhance intelligence into computer systems, specially of the deep learning, big data acquired from spectroscopy image has entered a new era. Therefore, the proposal of this work was to diagnose breast tissue samples as malignant (cancer) or benign (adenosis) using deep learning techniques. Micro-FTIR spectroscopy images were acquired from BR804b human breast tissue microarray (Biomax, USA), resulting in more than 100 thousand spectra for each group. A k-means approach was established to separate spectra into three clusters: tissue, paraffin and slide. A preprocessing step was applied by the following pipeline: outlier removal; biofingerprint truncation; Savitzky–Golay filter to smooth and to obtain the second derivative; extended multiplicative signal correction to correct spectra and remove the paraffin contribution. The deep learning algorithm was built using two-layers of one-dimensional convolutional neural network (CNN) connected to a two-layers (100 and 50 neurons) feedforward network (FFN). Both networks used dropout layers of 50% and rectified linear unit activations. CNN kernel size was set to 5. The output neuron used a sigmoid activation. Adam optimizer was applied to train the networks, using a binary cross-entropy loss to improve the weights. A 4-fold cross-validation of 20 epochs and batch size of 250 was performed. The networks exhibited an accuracy of (97.8 ± 0.4)% during the training stage, and (96.9 ± 0.8)% during the testing stage, demonstrating a generalized classification. Accuracies of almost 100% indicates this approach as a potential technique for the breast diagnosis.
A deep learning approach for breast tissue malignancy diagnosis using micro-FTIR hyperspectral imaging
2021 - DEL-VALLE, MATHEUS; SANTOS, MOISES O. dos; SANTOS, SOFIA N. dos; BERNARDES, EMERSON S.; ZEZELL, DENISE M.
The breast cancer is the most incident cancer in women with an estimative of 2.1 million new cases in 2018. With the grown of deep learning techniques, several approaches in vibrational spectroscopy have been studied. In this way, this work aimed to classify breast samples as breast cancer or adenosis using a deep learning model. It was used the human breast cancer microarray BR804b (Biomax, Inc., USA), where one core of each group, cancer and adenosis, was imaged by a Cary Series 600 micro-FTIR imaging system (Agilent Technologies, USA). The system has a spatial resolution of 5.5 μm and about 100 thousand spectra were acquired for each group. The regions of interest were selected by two k-means clustering using amide I/II (1700 to 1500 cm-1) and highest paraffin intensity (1480 to 1450 cm-1) bands. Spectra were preprocessed by five steps: outlier removal using Hotelling’s T2 versus Q residuals; biofingerprint truncation; Savitzky–Golay filtering for smoothing and second derivative; Extended multiplicative signal correction (EMSC) with digital de-waxing; another outlier removal. The deep learning model was a convolutional neural network (CNN) fused with a fully connected neural network (FCNN). The CNN was built with 2 Conv1D-ReLU-MaxPooling1D-Dropout layers. The kernel size was set to 5 and dropout of 0.5. Dense layers were built by two layers of neurons-BatchNorm-ReLU-Dropout, with 100 and 50 neurons, dropout of 0.2. The output was a single neuron with sigmoid activation. Binary cross-entropy loss function was adopted with Adam optimizer. Accuracy metric was calculated during the training, where a threshold of 0.5 was applied on the output predictions. Model was trained by a 4-fold cross-validation by 20 epochs and using a batch size of 250. The train accuracy was 0.978/0.004 (mean/std), while the testing accuracy was 0.969/0.008, demonstrating a generalized model without overfitting. Accuracies near one indicate the proposed model as a potential technique for the breast cancer vs adenosis classification, where hyperparameters and the architecture should be optimized along higher sample number acquisition.
The impact of scan number and its preprocessing in micro-FTIR imaging when applying machine learning for breast cancer subtypes classification
2021 - DEL-VALLE, MATHEUS; SANTOS, MOISES O. dos; SANTOS, SOFIA N. dos; CASTRO, PEDRO A.A. de; BERNARDES, EMERSON S.; ZEZELL, DENISE M.
The breast cancer molecular subtype is an important classification to outline the prognostic. Gold-standard assessing using immunohistochemistry adds subjectivity due to interlaboratory and interobserver variations. In order to increase the diagnosis confidence, other techniques need to be examined, where the FTIR spectroscopy imaging allied with machine learning techniques may provide additional and quantitative information regarding the molecular composition. However, the impact of co-added scans acquisition parameter into machine learning classifications still needs better evaluation. In this study, FTIR images of Luminal B and HER2 subtypes were acquired varying the scan number and preprocessing techniques. It was demonstrated a spectral quality improvement when the scan number was increased, decreasing the standard deviation and outliers. Six machine learning models were used to classify the subtypes: Linear Discriminant Analysis, Partial Least Squares Discriminant Analysis, K-Nearest Neighbors, Support Vector Machine, Random Forest and Extreme Gradient Boosting. Best mean accuracy of 0.995 was achieved by Extreme Gradient Boosting model. It was found that all models achieved similar high accuracies with groups b256_064 (256 background and 064 scans), b256_128 and b128_128. Besides assessing the performance of different models, the b256_064 was established as the optimal group due to the minimum acquisition time. Therefore, this work indicates b256_064 for breast cancer subtype classification and also as a basis for other studies using machine learning for cancer evaluation.
Breast cancer estrogen and progesterone receptors evaluation using FTlR spectroscopy imaging
2020 - DEL-VALLE, M.; SANTOS, M.O.; SANTOS, S.N. dos; BERNARDES, E.S.; ZEZELL, D.M.
INTRODUCTION. Breast cancer is the second leading cause of cancer death in woman worldwide with an incidence of 2.09 million and 627 thousand deaths in 2018. Histopathology is the gold standard method for cancer diagnosis and identification of therapeutic targets, however it still presents interpretation difficulties, especially when comparing different cancer subtypes. OBJETICVE: The aim of this study was to evaluate Fourier transform infrared (FTIR) spectroscopy in the diagnose and differentiation of molecular differences between two different breast cancer subtypes: positive and negative for estrogen (ER) and progesterone (PR) receptors. METHODS: Two human breast cell lines, BT474 (ER and PR positive) and SKBR3 (ER and PR negative), were inoculated in Balb/c nude mice. Tumors were collected when reached 0.5 cm3, processed by formalin fixation and paraffin embedding. 5μm thick tissue cuts were fixed in low-e slides (MirrIR, Kevley Technologies). Spectral images were performed in a micro-FTIR (Cary 660, Agilent Technologies) with 32 x 32 FPA of 5.5 μm pixel size. Scattering correction (RMieS-EMSC) was performed using MATLAB and remaining processing using Python. Groups differentiation were evaluated by PCA from 1350 to 1000 cm-1 second derivatives. RESULTS AND DISCUSSION: Groups were split in two clusters, separated by PC-1 with a 99 % accuracy in both groups and 45 % of explained variance. The absorptions in the selected region for the PCA were mainly related to DNA, RNA and protein content. The main contribution was presented by the 1238 cm-1 peak, which was correlated with nucleic acids symmetrical stretching. Hyperspectral image built from this peak presented a spatial correlation with the microscope white light imaging, indicating that possible region for histopathological correlation might be present. CONCLUSAO: Our pilot study shows that FTIR spectroscopy imaging can distinguish ER/PR positive from negative breast cancer subtypes.
Machine Learning methods for micro-FTIR imaging classification of human skin tumors
2021 - DEL VALLE, MATHEUS; STANCARI, KLEBER; CASTRO, PEDRO A.A. de; SANTOS, MOISES O. dos; ZEZELL, DENISE M.
This review presents some methods applied to micro-FTIR imaging for classification of human skin tumors. It is a collection of the pre-processing pipeline and machine learning classification models. The aim of this review is to update and summaiize the current methods which an applied in our skin tumor research.
Evaluation of machine learning models for the classification of breast cancer hormone receptors using micro-FTIR images
2021 - VALLE, MATHEUS del; SANTOS, MOISES O. dos; SANTOS, SOFIA N. dos; BERNARDES, EMERSON S.; ZEZELL, DENISE M.
The breast cancer is the most incident cancer in women. Evaluation of hormone receptors expression plays an important role to outline treatment strategies. FTIR spectroscopy imaging may be employed as an additional technique, providing extra information to help physicians. In this work, estrogen and progesterone receptors expression were evaluated using tumors biopsies from human cell lines inoculated in mice. FTIR images were collect from histological sections, and six machine learning models were applied and assessed. Xtreme gradient boost and Linear Discriminant Analysis presented the best accuracies results, indicating to be potential models for breast cancer classification tasks.