Exploring Plastic-Management Policy in China: Status, Challenges and Policy Insights

Plastic pollution is recognized as one of the most urgent global environmental concerns. China is the top producer and consumer of plastics and creates the most plastic waste globally. To evaluate policy options to control plastic pollution in China, we first reviewed the relevant policies and action plans in place. Then, we examined plastic-material flows and changes at the national level based on officially published data to evaluate the current situation and efficacy of policies at the macro-level. Results showed that 2016, the start of the 13th Five-Year Plan, was a pivotal year in the history of China's plastic policies tackling plastic issues nationally and internationally. Since 2016, the growth trend in the production and consumption of plastic products has slowed and the recycling rate has risen, surpassing 30% in 2021. To further tackle plastic pollution, key suggestions with important policy implications were provided, covering better integration of policies, the combined management of vertical–horizontal governance, tracking-system implementation, the introduction of a quality-certification system, the development of behaviour-based consumer-oriented solutions, the promotion of stakeholder collaboration, and the need for appropriate post-COVID-19 policies.

COVID-19 imaging, where do we go from here? Bibliometric analysis of medical imaging in COVID-19.

OBJECTIVES: We conducted a systematic and comprehensive bibliometric analysis of COVID-19-related medical imaging to determine the current status and indicate possible future directions. METHODS: This research provides an analysis of Web of Science Core Collection (WoSCC) indexed articles on COVID-19 and medical imaging published between 1 January 2020 and 30 June 2022, using the search terms "COVID-19" and medical imaging terms (such as "X-ray" or "CT"). Publications based solely on COVID-19 themes or medical image themes were excluded. CiteSpace was used to identify the predominant topics and generate a visual map of countries, institutions, authors, and keyword networks. RESULTS: The search included 4444 publications. The journal with the most publications was European Radiology, and the most co-cited journal was Radiology. China was the most frequently cited country in terms of co-authorship, with the Huazhong University of Science and Technology being the institution contributing with the highest number of relevant co-authorships. Research trends and leading topics included: assessment of initial COVID-19-related clinical imaging features, differential diagnosis using artificial intelligence (AI) technology and model interpretability, diagnosis systems construction, COVID-19 vaccination, complications, and predicting prognosis. CONCLUSIONS: This bibliometric analysis of COVID-19-related medical imaging helps clarify the current research situation and developmental trends. Subsequent trends in COVID-19 imaging are likely to shift from lung structure to function, from lung tissue to other related organs, and from COVID-19 to the impact of COVID-19 on the diagnosis and treatment of other diseases. Key Points • We conducted a systematic and comprehensive bibliometric analysis of COVID-19-related medical imaging from 1 January 2020 to 30 June 2022. • Research trends and leading topics included assessment of initial COVID-19-related clinical imaging features, differential diagnosis using AI technology and model interpretability, diagnosis systems construction, COVID-19 vaccination, complications, and predicting prognosis. • Future trends in COVID-19-related imaging are likely to involve a shift from lung structure to function, from lung tissue to other related organs, and from COVID-19 to the impact of COVID-19 on the diagnosis and treatment of other diseases.

Routine saliva testing for SARS-CoV-2 in children: Methods for partnering with community childcare centers.

While considerable attention was placed on SARS-CoV-2 testing and surveillance programs in the K-12 setting, younger age groups in childcare centers were largely overlooked. Childcare facilities are vital to communities, allowing parents/guardians to remain at work and providing safe environments for both children and staff. Therefore, early in the COVID-19 pandemic (October 2020), we established a PCR-based COVID-19 surveillance program in childcare facilities, testing children and staff with the goal of collecting actionable public health data and aiding communities in the progressive resumption of standard operations and ways of life. In this study we describe the development of a weekly saliva testing program and provide early results from our experience implementing this in childcare centers. We enrolled children (aged 6 months to 7 years) and staff at seven childcare facilities and trained participants in saliva collection using video chat technology. Weekly surveys were sent out to assess exposures, symptoms, and vaccination status changes. Participants submitted weekly saliva samples at school. Samples were transported to a partnering clinical laboratory or RT-PCR testing using SalivaDirect and results were uploaded to each participant's online patient portal within 24 h. SARS-CoV-2 screening and routine testing programs have focused less on the childcare population, resulting in knowledge gaps in this critical age group, especially as many are still ineligible for vaccination. SalivaDirect testing for SARS-CoV-2 provides a feasible method of asymptomatic screening and symptomatic testing for children and childcare center staff. Given the relative aversion to nasal swabs in younger age groups, an at-home saliva collection method provides an attractive alternative, especially as a routine surveillance tool. Results can be shared rapidly electronically through participants' private medical chart portals, and video chat technology allows for discussion and instruction between investigators and participants. This study fosters a cooperative partnership with participating childcare centers, parents/guardians, and staff with the goal of mitigating COVID-19 transmission in childcare centers. Age-related challenges in saliva collection can be overcome by working with parents/guardians to conceptualize new collection strategies and by offering parents/guardians continued virtual guidance and support.

Close relatives of MERS-CoV in bats use ACE2 as their functional receptors.

Middle East respiratory syndrome coronavirus (MERS-CoV) and several bat coronaviruses use dipeptidyl peptidase-4 (DPP4) as an entry receptor1-4. However, the receptor for NeoCoV-the closest known MERS-CoV relative found in bats-remains unclear5. Here, using a pseudotype virus entry assay, we found that NeoCoV and its close relative, PDF-2180, can efficiently bind to and use specific bat angiotensin-converting enzyme 2 (ACE2) orthologues and, less favourably, human ACE2 as entry receptors through their receptor-binding domains (RBDs) on the spike (S) proteins. Cryo-electron microscopy analysis revealed an RBD-ACE2 binding interface involving protein-glycan interactions, distinct from those of other known ACE2-using coronaviruses. We identified residues 337-342 of human ACE2 as a molecular determinant restricting NeoCoV entry, whereas a NeoCoV S pseudotyped virus containing a T510F RBD mutation efficiently entered cells expressing human ACE2. Although polyclonal SARS-CoV-2 antibodies or MERS-CoV RBD-specific nanobodies did not cross-neutralize NeoCoV or PDF-2180, an ACE2-specific antibody and two broadly neutralizing betacoronavirus antibodies efficiently inhibited these two pseudotyped viruses. We describe MERS-CoV-related viruses that use ACE2 as an entry receptor, underscoring a promiscuity of receptor use and a potential zoonotic threat.

The correlation between COVID-19 segmentation volume based on artificial intelligence technology and gastric wall edema: a multi-center study in Wuhan.

Purpose: This study aimed to investigate manifestations of the gastric wall and related risk factors in COVID-19 patients with gastrointestinal symptoms by CT. Materials and methods: Two hundred and forty patients diagnosed with COVID-19 by RT-PCR were enrolled from January 2020 to April 2020. Patients showed gastrointestinal symptoms, including nausea, vomiting, or diarrhea. Results of the initial laboratory examination were performed after admission. Chest CT was performed for all patients, with the lower bound including the gastric antrum. The volume of COVID-19 and lungs was segmented, and the ratio was calculated as follows: PV/LV = Volumepneumonia/Volumelungs. Results: Among the 240 patients, 109 presented with gastric wall edema (edema group), and 131 showed no gastric wall edema (non-edema group); the PV/LV values between the two groups were significantly different (P = 0.002). Univariate analysis revealed the following: fibrinogen (Fib), thrombin time (TT), activated partial thromboplastin time (APTT), and albumin (ALB) significantly differed between the two groups (P < 0.05). Binary logistic regression analysis showed that only APTT had a negative effect on gastric wall edema (P = 0.003). Conclusions: SARS-CoV-2 invades the gastrointestinal tract, gastric wall edema is the primary CT manifestation, and gastric wall edema is more likely to occur with a shorter APTT and severe pneumonia, with a slightly longer hospitalization time. Patients with gastric wall edema observed by CT should intervene early, which may improve digestive function, and further strengthen immune potency against COVID-19.

Whole genome sequence analysis of SARS-CoV-2 from 40 imported cases of confirmed COVID-19 in Sichuan

Objective To understand the genomic characteristics of SARS-CoV-2 from 40 imported cases with confirmed COVID-19 in Sichuan during January and March 2022. Methods Total viral RNA was extracted from respiratory samples of 182 confirmed COVID-19 cases who entered China through Chendu International Airport from January to March 2022.Mutation nucleic acid detection kit was used to identify the mutant strains and Illumina sequencing platform was applied for whole genome sequence（WGS） of virus.SARS-CoV-2 reference sequences were downloaded from NCBI database for genetic evolution and antigen variation analysis.The Nextclade and Pangolin online virus analysis platform were used to determine the virus family and type,and to analyze the mutation loci of the virus.The phylogenetic tree was constructed,along with the epidemiological data of cases to analyze the source and correlation of viruses. Results Among 182 imported COVID-19 cases,B.1.617.2 mutations were identified in 3 cases and B.1.1.529 mutations were detected in 57 cases.A total of 40 SARS-CoV-2 whole genome sequences with coverage>95% were obtained in this study.Nextclade typing analysis showed that 3 sequences belonged to 21J（Delta）,5 sequences belonged to 21K（Omicron）and the remaining 32 sequences belonged to 21L（Omicron）.Pangolin typing analysis showed that the 3 sequences of 21J（Delta）belonged to AY.4,AY.109and B.1.617.2,the 5sequences of 21K（Omicron）all belonged to BA.1.1,and the remaining 32 sequences of 21L（Omicron）belonged to BA.2.Our sequence results were99.7% consistency with the Omicron variants sequences in current GISAID database.Compared with the reference sequence strain Wuhan-Hu-1（NC＿045512.2）,45,47and 42nucleotide variation sites and 36,25 and 36amino acid variation sites were found in the 3 sequences of 21J（Delta）.There were average 59（26-64）nucleotide mutation sites and 48（10-53）amino acid mutation sites in the 5sequences of 21K（Omicron）.The median number of nucleotide mutation sites of 71（66-76）and amino acid mutation sites of 53（40-56）were identified in the 32sequences of 21L（Omicron）.Phylogenetic tree analysis showed that 40SARS-CoV-2WGSs were all related to the current variants of concern（VOC）. Conclusions Continuous sequencing of SARS-CoV-2whole genome from imported cases with confirmed COVID-19is of great significance for the prevention and control of the outbreak and prevalence of local epidemic caused by imported viruses in Sichuan.

An extracellular matrix biosensing mimetic for evaluating cathepsin as a host target for COVID-19.

To combat the new virus currently ravaging the whole world, every possible anti-virus strategy should be explored. As the main strategy of targeting the virus itself is being frustrated by the rapid mutation of the virus, people are seeking an alternative "host targeting" strategy: neutralizing proteins in the human body that cooperate with the virus. The cathepsin family is such a group of promising host targets, the main biological function of which is to digest the extracellular matrix (ECM) to clear a path for virus spreading. To evaluate the potential of cathepsin as a host target, we have constructed a biosensing interface mimicking the ECM, which can detect cathepsin from 3.3 pM to 33 nM with the limit of detection of 1 pM. Based on our quantitative analysis enabled by this biosensing interface, it is clear that patients with background diseases such as chronic inflammation and tumor, tend to have higher cathepsin activity, confirming the potential of cathepsin to serve as a host target for combating COVID-19 virus.

COVID-19 & differential effects in twins: Insights from Placenta Pathology.

INTRODUCTION: COVID-19 has been associated with several adverse pregnancy outcomes, including perinatal loss. Differential effects of COVID-19 in a twin pregnancy may provide unique insights into virus-placental interactions. We present a case of perinatal loss of a female fetus with survival of the male co-twin in a pregnancy complicated by COVID-19 and premature delivery. METHODS: Viral detection methods recommended by the NICHD task force were used to identify SARS-CoV-2 and its viral receptors in the placentas and fetal tissue (Antoun et al., 2020) [1] RESULTS: Compared with the surviving twin, we found a more severe intervillous necrosis and a relatively low detection of ACE2 membranous expression in the syncytiotrophoblasts of the female twin that succumbed. DISCUSSION: The interactions of SARS-CoV-2 and ACE2 at the maternal fetal interface within the placenta may play a significant role in perinatal loss, and the effects of fetal sex and gestational age at time of infection need to be explored further.

Research of the Changes in the Psychological Status of Chinese University Students and the Influencing Factors During the COVID-19 Pandemic.

Background: Psychological dynamics of college students have changed during the COVID-19 outbreak but little research has been done in this area. The purpose of this study is to investigate the dynamic changes in the mental health status of college students since the outbreak of the COVID-19 pandemic 1 year and the influencing factors. Methods: The research period was from February 2020 to August 2021. 384 college students were analyzed three times during this period on the recognition and psychological state of the pandemic. Results: During the period from February 2020 to August 2021, in general, the positive scores rose from 20.79 to 23.46, while the negative scores dropped from 17.41 to 14.00. The regression analysis results on the influencing factors showed the degree of recognition of the pandemic is all significant in the three phases (p < 0.05). Conclusion: With the effective control of the pandemic, the mental state of the students showed a slight improvement in the environment of sporadic cases. Behavior has a partial mediating effect between the source of fear and psychological changes. Correct behavior guidance can effectively reduce the psychological changes caused by college students' fear.

Detection of SARS-CoV-2 in tissue: the comparative roles of RT-qPCR, in situ RNA hybridization, and immunohistochemistry.

INTRODUCTION: The emergence of SARS-CoV-2, the causative agent the COVID-19 pandemic, has led to a rapidly expanding arsenal of molecular diagnostic assays for the detection of viral material in tissue specimens. AREAS COVERED: We review the value and shortcomings of available tissue-based assays for SARS-CoV-2 detection in formalin-fixed paraffin-embedded (FFPE) tissue, including immunohistochemistry, in situ hybridization, and quantitative reverse transcription PCR (RT-qPCR). The validation, accuracy, and comparative utility of each method is discussed. Subsequently, we identify commercially available antibodies which render the greatest specificity and reproducibility of staining in FFPE specimens. EXPERT OPINION: We offer expert opinion on the efficacy of such techniques and guidance for future implementation, both clinical and experimental.

Combining genomic and epidemiological data to compare the transmissibility of SARS-CoV-2 variants Alpha and Iota.

SARS-CoV-2 variants shaped the second year of the COVID-19 pandemic and the discourse around effective control measures. Evaluating the threat posed by a new variant is essential for adapting response efforts when community transmission is detected. In this study, we compare the dynamics of two variants, Alpha and Iota, by integrating genomic surveillance data to estimate the effective reproduction number (Rt) of the variants. We use Connecticut, United States, in which Alpha and Iota co-circulated in 2021. We find that the Rt of these variants were up to 50% larger than that of other variants. We then use phylogeography to show that while both variants were introduced into Connecticut at comparable frequencies, clades that resulted from introductions of Alpha were larger than those resulting from Iota introductions. By monitoring the dynamics of individual variants throughout our study period, we demonstrate the importance of routine surveillance in the response to COVID-19.

SaliVISION: a rapid saliva-based COVID-19 screening and diagnostic test with high sensitivity and specificity.

The Coronavirus disease 2019 (COVID-19) pandemic-caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)- has posed a global threat and presented with it a multitude of economic and public-health challenges. Establishing a reliable means of readily available, rapid diagnostic testing is of paramount importance in halting the spread of COVID-19, as governments continue to ease lockdown restrictions. The current standard for laboratory testing utilizes reverse transcription quantitative polymerase chain reaction (RT-qPCR); however, this method presents clear limitations in requiring a longer run-time as well as reduced on-site testing capability. Therefore, we investigated the feasibility of a reverse transcription looped-mediated isothermal amplification (RT-LAMP)-based model of rapid COVID-19 diagnostic testing which allows for less invasive sample collection, named SaliVISION. This novel, two-step, RT-LAMP assay utilizes a customized multiplex primer set specifically targeting SARS-CoV-2 and a visual report system that is ready to interpret within 40 min from the start of sample processing and does not require a BSL-2 level testing environment or special laboratory equipment. When compared to the SalivaDirect and Thermo Fisher Scientific TaqPath RT-qPCR testing platforms, the respective sensitivities of the SaliVISION assay are 94.29% and 98.28% while assay specificity was 100% when compared to either testing platform. Our data illustrate a robust, rapid diagnostic assay in our novel RT-LAMP test design, with potential for greater testing throughput than is currently available through laboratory testing and increased on-site testing capability.

Co-infecting pathogens can contribute to inflammatory responses and severe symptoms in COVID-19.

Background: The current COVID-19 pandemic is posing a major challenge to public health on a global scale. While it is generally believed that severe COVID-19 results from over-expression of inflammatory mediators (i.e., a "cytokine storm"), it is still unclear whether and how co-infecting pathogens contribute to disease pathogenesis. To address this, we followed the entire course of the disease in cases with severe or critical COVID-19 to determine the presence and abundance of all potential pathogens present-the total "infectome"-and how they interact with the host immune system in the context of severe COVID-19. Methods: We examined one severe and three critical cases of COVID-19, as well as a set of healthy controls, with longitudinal samples (throat swab, whole blood, and serum) collected from each case. Total RNA sequencing (meta-transcriptomics) was performed to simultaneously investigate pathogen diversity and abundance, as well as host immune responses, in each sample. A Bio-Plex method was used to measure serum cytokine and chemokine levels. Results: Eight pathogens, SARS-CoV-2, Aspergillus fumigatus (A. fumigatus), Mycoplasma orale (M. orale), Myroides odoratus (M. odoratus), Acinetobacter baumannii (A. baumannii), Candida tropicalis, herpes simplex virus (HSV) and human cytomegalovirus (CMV), identified in patients with COVID-19 appeared at different stages of the disease. The dynamics of inflammatory mediators in serum and the respiratory tract were more strongly associated with the dynamics of the infectome compared with SARS-CoV-2 alone. Correlation analysis revealed that pulmonary injury was directly associated with cytokine levels, which in turn were associated with the proliferation of SARS-CoV-2 and co-infecting pathogens. Conclusions: For each patient, the cytokine storm that resulted in acute lung injury and death involved a dynamic and highly complex infectome, of which SARS-CoV-2 was a component. These results indicate the need for a precision medicine approach to investigate both the infection and host response as a standard means of infectious disease characterization.

Drastic decline in sera neutralization against SARS-CoV-2 Omicron variant in Wuhan COVID-19 convalescents.

Global concern has been raised by the emergence and rapid transmission of the heavily mutated SARS-CoV-2 Omicron variant (B.1.1.529). So far, the infection features and immune escape ability of the Omicron variant have not been extensively studied. Here, we produced the Omicron pseudovirus and compared its entry, membrane fusion, and immune escape efficiency with the original strain and the dominating Delta variant. We found the Omicron variant showed slightly higher infectivity than the Delta variant and a similar ability to compete with the Delta variant in using Angiotensin-converting enzyme 2 (ACE2) in a BHK21-ACE2 cell line. However, the Omicron showed a significantly reduced fusogenicity than the original strain and the Delta variant in both BHK21-ACE2 and Vero-E6 cells. The neutralization assay testing the Wuhan convalescents' sera one-year post-infection showed a more dramatic reduction (10.15 fold) of neutralization against the Omicron variant than the Delta variant (1.79 fold) compared with the original strain with D614G. Notably, immune-boosting through three vaccine shots significantly improved the convalescents' immunity against the Omicron variants. Our results reveal a reduced fusogenicity and a striking immune escape ability of the Omicron variant, highlighting the importance of booster shots against the challenge of the SARS-CoV-2 antigenic drift.

Impact of chemotherapy and immunotherapy on the composition and function of immune cells in COVID-19 convalescent with gynecological tumors.

Ongoing pandemic and potential resurgence of Coronavirus disease 2019 (COVID-19) has prompted urgent efforts to investigate the immunological memory of convalescent patients, especially in patients with active cancers. Here we performed single-cell RNA sequencing in peripheral blood samples of 3 healthy donors (HDs), 4 COVID-19 patients (Covs) and 4 COVID-19 patients with active gynecological tumor (TCs) pre- and post- anti-tumor treatment. All Covs patients had recovered from their acute infection. Interestingly, the molecular features of PBMCs in TCs are similar to that in Covs, suggesting that convalescent COVID-19 with gynecologic tumors do not have major immunological changes and may be protected against reinfection similar to COVID-19 patients without tumors. Moreover, the chemotherapy given to these patients mainly caused neutropenia, while having little effect on the proportion and functional phenotype of T and B cells, and T cell clonal expansion. Notably, anti-PD-L1 treatment massively increased cytotoxic scores of NK cells, and T cells, and facilitated clonal expansion of T cells in these patients. It is likely that T cells could protect patients from SARS-CoV-2 virus reinfection and anti-PD-L1 treatment can enhance the anti-viral activity of the T cells.

Cross-Site Severity Assessment of COVID-19 From CT Images via Domain Adaptation.

Early and accurate severity assessment of Coronavirus disease 2019 (COVID-19) based on computed tomography (CT) images offers a great help to the estimation of intensive care unit event and the clinical decision of treatment planning. To augment the labeled data and improve the generalization ability of the classification model, it is necessary to aggregate data from multiple sites. This task faces several challenges including class imbalance between mild and severe infections, domain distribution discrepancy between sites, and presence of heterogeneous features. In this paper, we propose a novel domain adaptation (DA) method with two components to address these problems. The first component is a stochastic class-balanced boosting sampling strategy that overcomes the imbalanced learning problem and improves the classification performance on poorly-predicted classes. The second component is a representation learning that guarantees three properties: 1) domain-transferability by prototype triplet loss, 2) discriminant by conditional maximum mean discrepancy loss, and 3) completeness by multi-view reconstruction loss. Particularly, we propose a domain translator and align the heterogeneous data to the estimated class prototypes (i.e., class centers) in a hyper-sphere manifold. Experiments on cross-site severity assessment of COVID-19 from CT images show that the proposed method can effectively tackle the imbalanced learning problem and outperform recent DA approaches.

Concurrent Treatment with Vitamin K2 and D3 on Spine Fusion in Patients with Osteoporosis-Associated Lumbar Degenerative Disorders.

STUDY DESIGN: A prospective and nonrandomized concurrent controlled trial. OBJECTIVE: To address the early effects of concurrent treatment with vitamin K2 and vitamin D3 on fusion rates in patients who have undergone spinal surgery. SUMMARY OF BACKGROUND DATA: Intervertebral pseudarthrosis has been reported after transforaminal lumbar interbody fusion (TLIF) or posterior lumbar interbody fusion (PLIF), especially in patients with osteopenia or osteoporosis. No study has assessed the early effects of concurrent treatment with vitamin K2 and vitamin D3 on fusion rates. METHODS: Patients with osteopenia or osteoporosis who underwent TLIF or PLIF in our department were included. Patients in the VK2+VD3 group received vitamin K2, vitamin D3, and calcium treatment, whereas subjects in the control group only received calcium and vitamin D3. Spine fusion was evaluated by computed tomography. The Japanese Orthopedic Association Back Pain Evaluation Questionnaire (JOA-BPEQ) and visual analog scale (VAS) were used to assess the clinical and neurological symptoms. Bone mineral density (BMD) and bone metabolism markers were measured for osteoporotic evaluation. RESULTS: Seventy-eight patients were included, and nine patients subsequently discontinued because of 2019-nCoV. At six months postoperatively, complete fusion rates were significantly higher in the VK2+VD3 group than that in the control group (91.18% vs 71.43%, P = 0.036). At six months postoperatively, BMD was increased in the VK2+VD3 group and was higher than that in the control group, although there was no significant difference. At three months postoperatively, a significant increase in procollagen type I amino terminal propeptide (91.81%) and a slight decrease in C-terminal end peptide (8.06%) were observed in the VK2+VD3 group. In both groups, the JOA-BPEQ and VAS scores were significantly improved after spine surgery. CONCLUSION: Administration of vitamin K2 and vitamin D3 can increase lumbar interbody fusion rates, improve clinical symptoms, promote bone information, and avoid further decline in BMD within six months after TLIF or PLIF.Level of Evidence: 3.

C3D-UNET: A Comprehensive 3D Unet for Covid-19 Segmentation with Intact Encoding and Local Attention.

For COVID-19 prevention and treatment, it is essential to screen the pneumonia lesions in the lung region and analyze them in a qualitative and quantitative manner. Three-dimensional (3D) computed tomography (CT) volumes can provide sufficient information; however, extra boundaries of the lesions are also needed. The major challenge of automatic 3D segmentation of COVID-19 from CT volumes lies in the inadequacy of datasets and the wide variations of pneumonia lesions in their appearance, shape, and location. In this paper, we introduce a novel network called Comprehensive 3D UNet (C3D-UNet). Compared to 3D-UNet, an intact encoding (IE) strategy designed as residual dilated convolutional blocks with increased dilation rates is proposed to extract features from wider receptive fields. Moreover, a local attention (LA) mechanism is applied in skip connections for more robust and effective information fusion. We conduct five-fold cross-validation on a private dataset and independent offline evaluation on a public dataset. Experimental results demonstrate that our method outperforms other compared methods.

IMAL-Net: Interpretable multi-task attention learning network for invasive lung adenocarcinoma screening in CT images.

PURPOSE: Feature maps created from deep convolutional neural networks (DCNNs) have been widely used for visual explanation of DCNN-based classification tasks. However, many clinical applications such as benign-malignant classification of lung nodules normally require quantitative and objective interpretability, rather than just visualization. In this paper, we propose a novel interpretable multi-task attention learning network named IMAL-Net for early invasive adenocarcinoma screening in chest computed tomography images, which takes advantage of segmentation prior to assist interpretable classification. METHODS: Two sub-ResNets are firstly integrated together via a prior-attention mechanism for simultaneous nodule segmentation and invasiveness classification. Then, numerous radiomic features from the segmentation results are concatenated with high-level semantic features from the classification subnetwork by FC layers to achieve superior performance. Meanwhile, an end-to-end feature selection mechanism (named FSM) is designed to quantify crucial radiomic features greatly affecting the prediction of each sample, and thus it can provide clinically applicable interpretability to the prediction result. RESULTS: Nodule samples from a total of 1626 patients were collected from two grade-A hospitals for large-scale verification. Five-fold cross validation demonstrated that the proposed IMAL-Net can achieve an AUC score of 93.8% ± 1.1% and a recall score of 93.8% ± 2.8% for identification of invasive lung adenocarcinoma. CONCLUSIONS: It can be concluded that fusing semantic features and radiomic features can achieve obvious improvements in the invasiveness classification task. Moreover, by learning more fine-grained semantic features and highlighting the most important radiomics features, the proposed attention and FSM mechanisms not only can further improve the performance but also can be used for both visual explanations and objective analysis of the classification results.