A Multi-Stage Approach to Breast Cancer Classification Using Histopathology Images.

Breast cancer is one of the deadliest diseases worldwide among women. Early diagnosis and proper treatment can save many lives. Breast image analysis is a popular method for detecting breast cancer. Computer-aided diagnosis of breast images helps radiologists do the task more efficiently and appropriately. Histopathological image analysis is an important diagnostic method for breast cancer, which is basically microscopic imaging of breast tissue. In this work, we developed a deep learning-based method to classify breast cancer using histopathological images. We propose a patch-classification model to classify the image patches, where we divide the images into patches and pre-process these patches with stain normalization, regularization, and augmentation methods. We use machine-learning-based classifiers and ensembling methods to classify the image patches into four categories: normal, benign, in situ, and invasive. Next, we use the patch information from this model to classify the images into two classes (cancerous and non-cancerous) and four other classes (normal, benign, in situ, and invasive). We introduce a model to utilize the 2-class classification probabilities and classify the images into a 4-class classification. The proposed method yields promising results and achieves a classification accuracy of 97.50% for 4-class image classification and 98.6% for 2-class image classification on the ICIAR BACH dataset.

Evidence and evolution of Criegee intermediates, hydroperoxides and secondary organic aerosols formed via ozonolysis of α-pinene.

α-Pinene, the most abundant monoterpene in the atmosphere, accounts for more than 50% of global monoterpene emission. Though its reaction with ozone has been generally perceived as a major source of secondary organic aerosols (SOAs), direct evidence of its reaction intermediates (RI) and their evolution remain lacking. Here we study the ozonolysis of α-pinene between 180 and 298 K using a long-path, temperature-variable aerosol cooling chamber coupled to a rapid-scan time-resolved Fourier transform infrared spectrometer. The spectroscopic signatures of large Criegee intermediates (CIs) and hydroperoxides (HPs) were found for the first time. The aerosol size evolution during the reaction was also measured. In contrast to a previous perception, we show that temperature plays a determinant role in the ozonolysis kinetics. Finally, we show that the formation of HPs is an energetically favorable pathway to dissipate CIs. This study provides new insights into the ozonolysis of α-pinene and its contribution to SOA formation.

Does decarboxylation make 2,5-dihydroxybenzoic acid special in matrix-assisted laser desorption/ionization?

RATIONALE: Among the six positional isomers of dihydroxybenzoic acid (DHB), 2,5-DHB is a more favorable matrix for use in matrix-assisted laser desorption/ionization (MALDI) than the other isomers because of its high ion-generation efficiency at 337 and 355 nm. The generation of hydroquinone or p-benzoquinone through the decarboxylation of 2,5-DHB has been suggested to play a crucial role in the ion-generation efficiency of 2,5-DHB. METHODS: The mass spectra of desorbed neutrals generated from MALDI were measured using electron impact ionization (70 eV) and a quadrupole mass spectrometer and vacuum ultraviolet (118 nm) photoionization and a time-of-flight mass spectrometer. The mass spectra of desorbed ions generated from MALDI were investigated using a time-of-flight mass spectrometer. The dissociation barrier height and dissociation rate of decarboxylation were calculated by an ab initio method and RRKM theory. RESULTS: Decarboxylation of neutral 2,5-DHB and 2,5-DHB cations was not observed. Theoretical calculations indicated that decarboxylation of neutral 2,5-DHB and 2,5-DHB cations is too slow to occur. CONCLUSIONS: The high ion-generation efficiency of the 2,5-DHB matrix at 337 and 355 nm is not related to decarboxylation.

Photodissociation dynamics of benzaldehyde (C6H5CHO) at 266, 248, and 193 nm.

The photodissociation of gaseous benzaldehyde (C(6)H(5)CHO) at 193, 248, and 266 nm using multimass ion imaging and step-scan time-resolved Fourier-transform infrared emission techniques is investigated. We also characterize the potential energies with the CCSD(T)/6-311+G(3df,2p) method and predict the branching ratios for various channels of dissociation. Upon photolysis at 248 and 266 nm, two major channels for formation of HCO and CO, with relative branching of 0.37:0.63 and 0.20:0.80, respectively, are observed. The C(6)H(5)+HCO channel has two components with large and small recoil velocities; the rapid component with average translational energy of approximately 25 kJ mol(-1) dominates. The C(6)H(6)+CO channel has a similar distribution of translational energy for these two components. IR emission from internally excited C(6)H(5)CHO, ν(3) (v=1) of HCO, and levels v≤2, J≤43 of CO are observed; the latter has an average rotational energy of approximately 13 kJ mol(-1) and vibrational energy of approximately 6 kJ mol(-1). Upon photolysis at 193 nm, similar distributions of energy are observed, except that the C(6)H(5)+HCO channel becomes the only major channel with a branching ratio of 0.82±0.10 and an increased proportion of the slow component; IR emission from levels ν(1) (v=1) and ν(3) (v=1 and 2) of HCO and v≤2, J≤43 of CO are observed; the latter has an average energy similar to that observed in photolysis at 248 nm. The observed product yields at different dissociation energies are compared to statistical-theory predicted results based on the computed singlet and triplet potential-energy surfaces.

Photodissociation dynamics of benzoic acid.

The photodissociation of benzoic acid at 193 and 248 nm was investigated using multimass ion imaging techniques. Three dissociation channels were observed at 193 nm: (1) C(6)H(5)COOH-->C(6)H(5)+COOH, (2) C(6)H(5)COOH-->C(6)H(5)CO+OH, and (3) C(6)H(5)COOH-->C(6)H(6)+CO(2). Only channels, (2) and (3), were observed at 248 nm. Comparisons of the ion intensities and photofragment translational energy distributions with the potential energies obtained from ab initio calculations and the branching ratios obtained from the Rice-Ramsperger-Kassel-Marcus theory suggest that the dissociation occurs on many electronic states.

Photodissociation and photoionization of 2,5-dihydroxybenzoic acid at 193 and 355 nm.

Photodissociation and photoionization of 2,5-dihydroxybenzoic acid (25DHBA), at 193 and 355 nm were investigated separately in a molecular beam using multimass ion imaging techniques. Two channels competed after excitation by one 193 nm photon. One channel is dissociation from the repulsive excited state along O-H bond distance, resulting in H atom elimination from meta-OH functional group. The other channel is internal conversion to the ground state, followed by H(2)O elimination. Some of the fragments further proceeded to secondary dissociation. On the other hand, absorption of one 355 nm photon gave rise to H(2)O elimination channel on the ground state. Absorption of more than one 355 nm photon resulted in the three-body dissociation which also occurs on the ground state. Dissociation on the excited state does not play a role at 355 nm. The large concentration ratio (2×10(5)), between neutral fragments and cations produced from 355 nm multiphoton excitation indicates that internal conversion followed by dissociation, is the major channel for 355 nm multiphoton excitation. Multiphoton ionization is a minor channel. Multiphoton ionization of 25DHBA clusters only produces 25DHBA cations. Neither anion nor protonated 25DHBA cation were observed. It is very different from the ions produced from solid matrix-assisted laser desorption/ionization (MALDI), experiments. This suggests that protonated 25DHBA and negatively charged 25DHBA generated in MALDI experiments does not simply result from the ionization following proton transfer reactions or charge transfer reactions of the clusters in the gas phase.

355 nm multiphoton dissociation and ionization of 2, 5-dihydroxyacetophenone.

Multiphoton dissociation and ionization of 2,5-dihydroxyacetophenone (DHAP), an important matrix compound in UV matrix-assisted laser desorption/ionization (MALDI), is studied in a molecular beam at 355 nm using multimass ion imaging mass spectrometer and time-of-flight mass spectrometry. For laser fluence larger than 130 mJ/cm(2), nearly all of the irradiated molecules absorb at least one photon. The absorption cross section was found to be sigma = 1.3(+/-0.2) x 10(-17)cm(2). Molecules excited by two photons quickly dissociate into fragments. The major channels are (1) C(6)H(3)(OH)(2)COCH(3) --> C(6)H(3)(OH)(2)CO + CH(3) and (2) C(6)H(3)(OH)(2)COCH(3) --> C(6)H(3)(OH)(2) + COCH(3). Molecules absorbing three or more photons become parent ions or crack into smaller ionic fragments. The concentration ratio of ions (parent ions and ionic fragments) to neutral fragments is about 10(-6):1. Changing the molecular beam carrier gas from He at 250 Torr to Ar at 300 Torr results in molecular beam clustering (dimers and trimers). Multiphoton ionization of clusters by a 355 nm laser beam produces only dimer cations, (C(6)H(3)(OH)(2)COCH(3))(2)(+). Protonated clusters or negatively charged ions, observed from a solid sample of DHAP using 355 nm multiphoton ionization, were not found in the molecular beam. The experimental results indicate that the photoionization occurs in the gas phase after DHAP vaporizes from the solid phase may not play an important role in the MALDI process.

Interaction of human serum albumin with charge transfer probe ethyl ester of N,N-dimethylamino naphthyl acrylic acid: an extrinsic fluorescence probe for studying protein micro-environment.

The charge transfer (CT) probe ethyl ester of N,N-dimethylamino naphthyl acrylic acid (EDMANA) bound to Human Serum Albumin (HSA) serves as an efficient reporter of the polarity and conformational changes of protein in aqueous buffer (Tris-HCl buffer, pH=7.03) and in presence of denaturant, quencher and reverse micelles. The change in fluorescence intensity and the position of emission maxima of EDMANA in presence of HSA well reflect the nature of binding and location of the probe inside the proteinous environment. The increase in steady state anisotropy values with increase of protein concentration indicate restriction imposed on the mobility of the probe molecules in the proteinous medium. The results of fluorescence quenching of EDMANA by acrylamide, Fluorescence Resonance Energy Transfer (FRET) and Red Edge Excitation Shift (REES) studies throw light on the accessibility to the probe bound to HSA and hence indicate the probable location of the probe within the hydrophobic cavity of HSA. The complicated nature of protein unfolding in presence of urea is well studied by change in the fluorescence properties of EDMANA bound to HSA protein.