Intelligent deep learning supports biomedical image detection and classification of oral cancer.

BACKGROUND: Oral cancer is a malignant tumor that usually occurs within the tissues of the mouth. This type of cancer mainly includes tumors in the lining of the mouth, tongue, lips, buccal mucosa and gums. Oral cancer is on the rise globally, especially in some specific risk groups. The early stage of oral cancer is usually asymptomatic, while the late stage may present with ulcers, lumps, bleeding, etc. OBJECTIVE: The objective of this paper is to propose an effective and accurate method for the identification and classification of oral cancer. METHODS: We applied two deep learning methods, CNN and Transformers. First, we propose a new CANet classification model for oral cancer, which uses attention mechanisms combined with neglected location information to explore the complex combination of attention mechanisms and deep networks, and fully tap the potential of attention mechanisms. Secondly, we design a classification model based on Swim transform. The image is segmented into a series of two-dimensional image blocks, which are then processed by multiple layers of conversion blocks. RESULTS: The proposed classification model was trained and predicted on Kaggle Oral Cancer Images Dataset, and satisfactory results were obtained. The average accuracy, sensitivity, specificity and F1-Socre of Swin transformer architecture are 94.95%, 95.37%, 95.52% and 94.66%, respectively. The average accuracy, sensitivity, specificity and F1-Score of CANet model were 97.00%, 97.82%, 97.82% and 96.61%, respectively. CONCLUSIONS: We studied different deep learning algorithms for oral cancer classification, including convolutional neural networks, converters, etc. Our Attention module in CANet leverages the benefits of channel attention to model the relationships between channels while encoding precise location information that captures the long-term dependencies of the network. The model achieves a high classification effect with an accuracy of 97.00%, which can be used in the automatic recognition and classification of oral cancer.

Indirect Emissions from Organophosphite Antioxidants Result in Significant Organophosphate Ester Contamination in China.

Organophosphite antioxidants (OPAs) have been seriously neglected as potential sources of organophosphate esters (OPEs) in environments. This study utilizes a modeling approach to quantify for the first time national emissions and multimedia distributions of triphenyl phosphate (TPHP)─a well-known flame retardant─and three novel OPEs: tris(2,4-ditert-butylphenyl) phosphate (AO168═O), bis(2,4-ditert-butylphenyl) pentaerythritol diphosphate (AO626═O2), and trisnonylphenol phosphate (TNPP). Emphasis is on the quantitative assessment of OPA source in China. TPHP has 1.1-9.7 times higher emission (300 Mg/year in 2019 with half from OPA sources) than AO168═O (278 Mg/year), AO626═O2 (53 Mg/year), and TNPP (32 Mg/year), but AO168═O is predominant in environments (63-79%) except freshwaters. About 72-99% of the studied OPEs are emitted via air, with 88-99% ultimately distributed into soils as the major sink. OPA-source emissions contribute 9.5-57% and 4.7-56% of TPHP masses and concentrations (except in sediments) in different media, respectively. Both AO168═O and AO626═O2 exhibit high overall persistence ranging between 2 and 11 years. Source emissions and environmental concentrations are elevated in economically developed areas, while persistence is higher in northern areas, where precipitation and temperature are lower. This study shows the significance of the sources of OPA to OPE contamination, which supports chemical management of these substances.

Selective and efficient removal of emerging contaminants by sponge-like manganese ferrite synthesized using a solvent-free method: Crucial role of the three-dimensional porous structure.

Ubiquitous macromolecular natural organic matter (NOM) in wastewater seriously influences the removal of emerging small-molecule contaminants via heterogeneous advanced oxidation processes because this material covers active sites and quenches reactive oxygen species. Here, sponge-like magnetic manganese ferrite (MnFe2O4-S) with a three-dimensional hierarchical porous structure was prepared via a facile solvent-free molten method. Compared with the particle-like structure of MnFe2O4-P, the sponge-like structure of MnFe2O4-S presents an enlarged specific surface area (112.14 m2·g-1 vs. 58.73 m2·g-1) and a smaller macropore diameter (68.2-77.2 nm vs. 946.5 nm). Enlarging the specific surface area increases the exposure of active sites, and adjusting the pore size helps sieve NOM and emerging contaminants. These changes are expected to effectively improve the degradation activity and overcome interference. To confirm the superiority of the sponge-like structure, MnFe2O4-S was used to activate peroxymonosulfate (PMS) for the degradation of multiple emerging contaminants, and its ability to degrade bisphenol A with and without humic acid (HA) was compared with that of MnFe2O4-P. The degradation activity of MnFe2O4-S was 1.6 times greater than that of MnFe2O4-P. Moreover, 20 mg·L-1 HA inhibited the degradation activity of MnFe2O4-S by only 7.1%, which was much lower than that obtained for MnFe2O4-P (53.4%). In addition, the excellent performance was maintained in multiple water matrices. Notably, under lake water matrices, the degradation activity of MnFe2O4-P was inhibited by 35.6% while that of MnFe2O4-S was hardly inhibited. More importantly, the MnFe2O4-S/PMS system was also applicable to the treatment of actual wastewater and 73.0% and 90.1% of total organic carbon and chemical oxygen demand was removed from bio-treated coking wastewater containing non-biodegradable contaminants and NOM. This study provides an alternative route for the green production of high-activity porous spinel ferrites with environmental anti-interference properties.