Classification of benign and malignant subtypes of breast cancer histopathology imaging using hybrid CNN-LSTM based transfer learning.

BACKGROUND: Grading of cancer histopathology slides requires more pathologists and expert clinicians as well as it is time consuming to look manually into whole-slide images. Hence, an automated classification of histopathological breast cancer sub-type is useful for clinical diagnosis and therapeutic responses. Recent deep learning methods for medical image analysis suggest the utility of automated radiologic imaging classification for relating disease characteristics or diagnosis and patient stratification. METHODS: To develop a hybrid model using the convolutional neural network (CNN) and the long short-term memory recurrent neural network (LSTM RNN) to classify four benign and four malignant breast cancer subtypes. The proposed CNN-LSTM leveraging on ImageNet uses a transfer learning approach in classifying and predicting four subtypes of each. The proposed model was evaluated on the BreakHis dataset comprises 2480 benign and 5429 malignant cancer images acquired at magnifications of 40×, 100×, 200× and 400×. RESULTS: The proposed hybrid CNN-LSTM model was compared with the existing CNN models used for breast histopathological image classification such as VGG-16, ResNet50, and Inception models. All the models were built using three different optimizers such as adaptive moment estimator (Adam), root mean square propagation (RMSProp), and stochastic gradient descent (SGD) optimizers by varying numbers of epochs. From the results, we noticed that the Adam optimizer was the best optimizer with maximum accuracy and minimum model loss for both the training and validation sets. The proposed hybrid CNN-LSTM model showed the highest overall accuracy of 99% for binary classification of benign and malignant cancer, and, whereas, 92.5% for multi-class classifier of benign and malignant cancer subtypes, respectively. CONCLUSION: To conclude, the proposed transfer learning approach outperformed the state-of-the-art machine and deep learning models in classifying benign and malignant cancer subtypes. The proposed method is feasible in classification of other cancers as well as diseases.

Identification of Novel Ribonucleotide Reductase Inhibitors for Therapeutic Application in Bile Tract Cancer: An Advanced Pharmacoinformatics Study.

Biliary tract cancer (BTC) is constituted by a heterogeneous group of malignant tumors that may develop in the biliary tract, and it is the second most common liver cancer. Human ribonucleotide reductase M1 (hRRM1) has already been proven to be a potential BTC target. In the current study, a de novo design approach was used to generate novel and effective chemical therapeutics for BTC. A set of comprehensive pharmacoinformatics approaches was implemented and, finally, seventeen potential molecules were found to be effective for the modulation of hRRM1 activity. Molecular docking, negative image-based ShaEP scoring, absolute binding free energy, in silico pharmacokinetics, and toxicity assessments corroborated the potentiality of the selected molecules. Almost all molecules showed higher affinity in comparison to gemcitabine and naphthyl salicylic acyl hydrazone (NSAH). On binding interaction analysis, a number of critical amino acids was found to hold the molecules at the active site cavity. The molecular dynamics (MD) simulation study also indicated the stability between protein and ligands. High negative MM-GBSA (molecular mechanics generalized Born and surface area) binding free energy indicated the potentiality of the molecules. Therefore, the proposed molecules might have the potential to be effective therapeutics for the management of BTC.

Identification of Potential Cytochrome P450 3A5 Inhibitors: An Extensive Virtual Screening through Molecular Docking, Negative Image-Based Screening, Machine Learning and Molecular Dynamics Simulation Studies.

Cytochrome P450 3A5 (CYP3A5) is one of the crucial CYP family members and has already proven to be an important drug target for cardiovascular diseases. In the current study, the PubChem database was screened through molecular docking and high-affinity molecules were adopted for further assessment. A negative image-based (NIB) model was used for a similarity search by considering the complementary shape and electrostatics of the target and small molecules. Further, the molecules were segregated into active and inactive groups through six machine learning (ML) matrices. The active molecules found in each ML model were used for in silico pharmacokinetics and toxicity assessments. A total of five molecules followed the acceptable pharmacokinetics and toxicity profiles. Several potential binding interactions between the proposed molecules and CYP3A5 were observed. The dynamic behavior of the selected molecules in the CYP3A5 was explored through a molecular dynamics (MD) simulation study. Several parameters obtained from the MD simulation trajectory explained the stability of the protein-ligand complexes in dynamic states. The high binding affinity of each molecule was revealed by the binding free energy calculation through the MM-GBSA methods. Therefore, it can be concluded that the proposed molecules might be potential CYP3A5 molecules for therapeutic application in cardiovascular diseases subjected to in vitro/in vivo validations.

Screening of ß1- and ß2-Adrenergic Receptor Modulators through Advanced Pharmacoinformatics and Machine Learning Approaches.

Cardiovascular diseases (CDs) are a major concern in the human race and one of the leading causes of death worldwide. ß-Adrenergic receptors (ß1-AR and ß2-AR) play a crucial role in the overall regulation of cardiac function. In the present study, structure-based virtual screening, machine learning (ML), and a ligand-based similarity search were conducted for the PubChem database against both ß1- and ß2-AR. Initially, all docked molecules were screened using the threshold binding energy value. Molecules with a better binding affinity were further used for segregation as active and inactive through ML. The pharmacokinetic assessment was carried out on molecules retained in the above step. Further, similarity searching of the ChEMBL and DrugBank databases was performed. From detailed analysis of the above data, four compounds for each of ß1- and ß2-AR were found to be promising in nature. A number of critical ligand-binding amino acids formed potential hydrogen bonds and hydrophobic interactions. Finally, a molecular dynamics (MD) simulation study of each molecule bound with the respective target was performed. A number of parameters obtained from the MD simulation trajectories were calculated and substantiated the stability between the protein-ligand complex. Hence, it can be postulated that the final molecules might be crucial for CDs subjected to experimental validation.

Patterning of corpus callosum integrity in glioma observed by MRI: Effect of 2D bi-axial lamellar brain architecture.

PURPOSE: Corpus callosum (CC) is a main channel histologically for glioma spreading, downgrading the prognosis, the infiltration occurring through cellular reaction-diffusion process. Preliminary clinical trial indicates that CC's surgical interruption appreciably enhances clinical outcome. We aim to find how high-grade glioma phenomenology is reflected in CC parameters, including various 3D diffusion eigenvalues differentially, whereby this information may be utilized for planning radiotherapy and surgical intervention. METHODS: Using 3 Tesla MRI diffusion-tensor imaging of glioma patients and matched controls, we formulated the callosal volume, fibre count, and 3D directional diffusivity eigenvalues (λ1-λ2-λ3), utilizing FDT/FMRIB-based analysis. RESULTS: In glioma, the callosal volume, fibre count and normalized volume decreases (p < 0.001), while axial diffusivity λ1 and radial diffusivity component λ2 significantly increase (p = 0.03, p = 0.04). Though not expected, the other radial diffusivity component λ3 remains unchanged (p = 0.11). Increase of λ1 and λ2 is due to gliomatous migration across the two directions (eigenvectors of λ1, λ2), which correlate respectively with medio-lateral commissural fibres and dorso-ventral perforating fibres in CC. These are corroborated by collateral radiological findings and immunohistological staining of those two fibre-systems in cat and human. CONCLUSION: In glioma, the two diffusivities (λ1, λ2), enhance due to fluidic edema permeation through CC's bi-axial lamina-type structural scaffold, formed by mediolateral commissural fibres and dorsoventral perforating cingulo-septal fibres. On other hand, the two radial diffusivities (λ2, λ3) are physiologically different and can be distinguished as lamellar diffusivity and focal diffusivity respectively. Lamellar diffusivity λ2 needs to be considered for MRI-assisted surgical intervention and radiotherapy planning in glioma.

Age-Related Differences in White Matter Integrity in Healthy Human Brain: Evidence from Structural MRI and Diffusion Tensor Imaging.

The aim is to investigate the relationship between microstructural white matter (WM) diffusivity indices and macrostructural WM volume (WMV) among healthy individuals (20-85 years). Whole-brain diffusion measures were calculated from diffusion tensor imaging using FMRIB software library while WMV was estimated through voxel-based morphometry, and voxel-based analysis was carried out using tract-based spatial statistics. Our results revealed that mean diffusivity, axial diffusivity, and radial diffusivity had shown good correlation with WMV but not for fractional anisotropy (FA). Voxel-wise tract-based spatial statistics analysis for FA showed a significant decrease in four regions for middle-aged group compared to young-aged group, in 22 regions for old-aged group compared to middle-aged group, and in 26 regions for old-aged group compared to young-aged group (P < 0.05). We found significantly lower WMV, FA, and mean diffusivity values in females than males and inverted-U trend for FA in males. We conclude differential age- and gender-related changes for structural WMV and WM diffusion indices.

A patient-specific therapeutic approach for tumour cell population extinction and drug toxicity reduction using control systems-based dose-profile design.

BACKGROUND: When anti-tumour therapy is administered to a tumour-host environment, an asymptotic tapering extremity of the tumour cell distribution is noticed. This extremity harbors a small number of residual tumour cells that later lead to secondary malignances. Thus, a method is needed that would enable the malignant population to be completely eliminated within a desired time-frame, negating the possibility of recurrence and drug-induced toxicity. METHODS: In this study, we delineate a computational procedure using the inverse input-reconstruction approach to calculate the unknown drug stimulus input, when one desires a known output tissue-response (full tumour cell elimination, no excess toxicity). The asymptotic extremity is taken care of using a bias shift of tumour-cell distribution and guided control of drug administration, with toxicity limits enforced, during mutually-synchronized chemotherapy (as Temozolomide) and immunotherapy (Interleukin-2 and Cytotoxic T-lymphocyte). RESULTS: Quantitative modeling is done using representative characteristics of rapidly and slowly-growing tumours. Both were fully eliminated within 2 months with checks for recurrence and toxicity over a two-year time-line. The dose-time profile of the therapeutic agents has similar features across tumours: biphasic (lymphocytes), monophasic (chemotherapy) and stationary (interleukin), with terminal pulses of the three agents together ensuring elimination of all malignant cells. The model is then justified with clinical case studies and animal models of different neurooncological tumours like glioma, meningioma and glioblastoma. CONCLUSION: The conflicting oncological objectives of tumour-cell extinction and host protection can be simultaneously accommodated using the techniques of drug input reconstruction by enforcing a bias shift and guided control over the drug dose-time profile. For translational applicability, the procedure can be adapted to accommodate varying patient parameters, and for corrective clinical monitoring, to implement full tumour extinction, while maintaining the health profile of the patient.

Magnetic resonance image enhancement using stochastic resonance in Fourier domain.

OBJECTIVE: In general, low-field MRI scanners such as the 0.5- and 1-T ones produce images that are poor in quality. The motivation of this study was to lessen the noise and enhance the signal such that the image quality is improved. Here, we propose a new approach using stochastic resonance (SR)-based transform in Fourier space for the enhancement of magnetic resonance images of brain lesions, by utilizing an optimized level of Gaussian fluctuation that maximizes signal-to-noise ratio (SNR). MATERIALS AND METHODS: We acquired the T1-weighted MR image of the brain in DICOM format. We processed the original MR image using the proposed SR procedure. We then tested our approach on about 60 patients of different age groups with different lesions, such as arteriovenous malformation, benign lesion and malignant tumor, and illustrated the image enhancement by using just-noticeable difference visually as well as by utilizing the relative enhancement factor quantitatively. RESULTS: Our method can restore the original image from noisy image and optimally enhance the edges or boundaries of the tissues, clarify indistinct structural brain lesions without producing ringing artifacts, as well as delineate the edematous area, active tumor zone, lesion heterogeneity or morphology, and vascular abnormality. The proposed technique improves the enhancement factor better than the conventional techniques like the Wiener- and wavelet-based procedures. CONCLUSIONS: The proposed method can readily enhance the image fusing a unique constructive interaction of noise and signal, and enables improved diagnosis over conventional methods. The approach well illustrates the novel potential of using a small amount of Gaussian noise to improve the image quality.

Enhancement of ultrasound images using stochastic resonance-based wavelet transform.

Ultrasound diagnostic imaging technique is used to visualize muscles and internal organs, their size, structures and possible pathologies or lesions. The limited soft tissue contrast of ultrasound may lead to problems in characterizing perivascular soft tissues. We develop a technique using stochastic resonance (SR)-based wavelet transform for the enhancement of unclear diagnostic ultrasound images. The proposed method enhances the edges more clearly. The advantages of this method are that it can simultaneously operate both as an enhancement process as well as a noise-reduction operation, and that the method can also optimally enhance an image even if the image noise level is considerable.

Stochastic resonance-based tomographic transform for computed tomographic image enhancement of brain lesions.

Motivated by statistical thermodynamics, we develop a technique using stochastic resonance-based tomographic transform for enhancement of noisy or indistinct computer-assisted tomographic images of the brain lesions for radiological diagnosis. The proposed method makes the edges of the lesion prominent, delineates the edematous zones more clearly, enhances the active zone in tumors, and clarifies the latent structure of the lesions, the mean enhancement index being 165%. The advantages of this method are that it can simultaneously operate both as an enhancement process and as a noise-reduction operation, and that the method can also optimally enhance an image even if the noise level is considerable. A general approach of thermodynamics-based image enhancement for computed tomographic diagnosis is outlined.