Likelihood-based inference under nonconvex boundary constraints

Likelihood-based inference under nonconvex constraints on model parameters has become increasingly common in biomedical research. In this paper, we establish large-sample properties of the maximum likelihood estimator when the true parameter value lies at the boundary of a nonconvex parameter space. We further derive the asymptotic distribution of the likelihood ratio test statistic under nonconvex constraints on model parameters. A general Monte Carlo procedure for generating the limiting distribution is provided. The theoretical results are demonstrated by five examples in Anderson's stereotype logistic regression model, genetic association studies, gene-environment interaction tests, cost-constrained linear regression and fairness-constrained linear regression.

CXCR6 defines therapeutic subtypes of CD4+ cytotoxic T cell lineage for adoptive cell transfer therapy in pediatric B cell acute lymphoblastic leukemia

The potential of cytotoxic CD4+ T cells and tissue resident memory T cells (Trm) in achieving adult leukemia remission have been highlighted [1,2]. We hypothesized that CXCR6 could serve as a marker for cytotoxic CD4+ Trm cells in the bone marrow (BM) of pediatric B-ALL patients. Flow cytometry (FCM) and published single cell RNA sequencing (scRNA-seq) datasets were employed to characterize CXCR6+CD4+ T cells in the BM and peripheral blood (PB) of pediatric B-ALL patients and healthy donors. FCM, scRNA-seq and co-culture were utilized to explore the cytotoxicity of CXCR6+CD4+ T cells in vitro based on in vitro induction of CXCR6+CD4+ T cells using tumor antigens and peripheral blood mononuclear cells (PBMCs). The ssGSEA based on the cell markers identified according to the in vivo scRNA-seq data, the TARGET-ALL-P2 datasets, and integrated machine learning algorithm were employed to figure out the key cells with prognostic values, followed by simulation of adoptive cell transfer therapy (ACT). Integrated machine learning identified the high-risk cells for disease free survival, and overall survival, while simulation of ACT therapy using CXCR6+CD4+T cells indicated that CXCR6+CD4+ T cells could remodel the bone marrow microenvironments towards anti-tumor. Based on the expression of genes involved in formation of resident memory T cells, CXCR6 is not a marker of resident memory CD4+T cells but defines therapeutic subtypes of CD4+ cytotoxic T cell lineage for pediatric B-ALL.

[Role of innate lymphoid cells in oral squamous cell carcinoma microenvironment]

Oral squamous cell carcinoma (OSCC) is the most common oral malignancy. It has a high incidence, strong invasion ability, easy metastasis, poor curative effect, and poor prognosis. Innate lymphoid cells (ILCs) are an important part of immune cells located in the mucosal barrier, which play an important role in the occurrence, development and outcome of tumors. ILCs are the key cells for decoding the regulatory mechanism of tumor microenvironment and the signatures for tumor progression. This paper reviewed the latest progress on ILCs, summarized the possible characteristics and functions of ILCs in the microenvironment of OSCC, and explored the relationship between ILCs and the occurrence, development and immunotherapy of OSCC.

[Preliminary outcomes of neoadjuvant chemoimmunotherapy combined with transoral robotic surgery for locally advanced oropharyngeal squamous cell carcinoma]

Objective: To evaluate the efficacy of neoadjuvant chemoimmunotherapy (NACI) combined with transoral robotic surgery (TORS) in the treatment of locally advanced oropharyngeal squamous cell carcinoma (OPSCC). Methods: This was a retrospective study of 15 patients with locally advanced OPSCC who underwent TORS after neoadjuvant therapy (NAT) at the Department of Otolaryngology-Head and Neck Surgery of Sun Yat-sen Memorial Hospital of Sun Yat-sen University from April 2019 to February 2023. There were 12 males and 3 females, aged 31 to 74 years. Twelve cases were tonsil cancer, and 3 cases were tongue base cancer. There were 11 cases in stage Ⅲ and 4 cases in stage Ⅳ. Two patients received neoadjuvant chemotherapy and 13 patients received NACI, with 2 to 3 cycles, and all patients underwent TORS after multidisciplinary team consultation. The clinicopathological characteristics, surgical outcomes, and oncological results were summarized. Results: All surgeries were successfully completed with negative surgical margins, and no case was required conversion surgery. All patients were fed via nasogastric tubes postoperatively, with a median gastric tube stay of 7 days (range: 2-60 days). No tracheotomy was applied. There were no major complications such as postoperative bleeding. Pathological complete response (pCR) was found in 10 cases (76.9%) among the 13 patients with NACI. The follow-up time was 21 months (range: 10-47 months), and there was no death or distant metastasis. One patient with rT0N3M0 tonsil cancer had local recurrence 5 months after surgery. The 2-year overall survival and 2-year disease-free survival were respectively 100.0% and 93.3% in the 15 patients. Conclusion: NACI combined with TORS provides a safe, effective and minimally invasive treatment for patients with locally advanced oropharyngeal squamous cell carcinoma.

Genome-wide association study of growth curve parameters reveals novel genomic regions and candidate genes associated with metatarsal bone traits in chickens

The growth and development of chicken bones have an enormous impact on the health and production performance of chickens. However, the development pattern and genetic regulation of the chicken skeleton are poorly understood. This study aimed to evaluate metatarsal bone growth and development patterns in chickens via non-linear models, and to identify the genetic determinants of metatarsal bone traits using a genome-wide association study (GWAS) based on growth curve parameters. Data on metatarsal length (MeL) and metatarsal circumference (MeC) were obtained from 471 F2 chickens (generated by crossing broiler sires, derived from a line selected for high abdominal fat, with Baier layer dams) at 4, 6, 8, 10, and 12 weeks of age. Four non-linear models (Gompertz, Logistic, von Bertalanffy, and Brody) were used to fit the MeL and MeC growth curves. Subsequently, the estimated growth curve parameters of the mature MeL or MeC (A), time-scale parameter (b), and maturity rate (K) from the non-linear models were utilized as substitutes for the original bone data in GWAS. The Logistic and Brody models displayed the best goodness-of-fit for MeL and MeC, respectively. Single-trait and multi-trait GWASs based on the growth curve parameters of the Logistic and Brody models revealed 4 618 significant single nucleotide polymorphisms (SNPs), annotated to 332 genes, associated with metatarsal bone traits. The majority of these significant SNPs were located on Gallus gallus chromosome (GGA) 1 (167.433-176.318 Mb), GGA2 (96.791-103.543 Mb), GGA4 (65.003-83.104 Mb) and GGA6 (64.685-95.285 Mb). Notably, we identified 12 novel GWAS loci associated with chicken metatarsal bone traits, encompassing 35 candidate genes. In summary, the combination of single-trait and multi-trait GWASs based on growth curve parameters uncovered numerous genomic regions and candidate genes associated with chicken bone traits. The findings benefit an in-depth understanding of the genetic architecture underlying metatarsal growth and development in chickens.

Method of smoldering combustion for the treatment of oil sludge-contaminated soil

There is an urgent need to globally remediate oil sludge-contaminated soil (OSS). Smoldering combustion is a new low-energy approach for the treatment of organic waste. Therefore, the feasibility of smoldering combustion for the treatment of OSS was investigated in this study using a series of laboratory-scale experiments. The effective remediation of OSS was found to be achievable when the mass ratio of oil sludge in the sample reached 1/12 and above. Experimental results showed that smoldering at peak temperatures above 500 °C was found to completely remove petroleum hydrocarbons from the samples. The mass ratio of oil sludge in the sample had little effect on the distribution of the major elements (Si, Al, and Ca) in the smoldering products, and most of the minerals in the oil sludge adhered to the surface of the soil particles after smoldering. The smoldering heating environment is detrimental to the reusability of the soil, increases soil pH and available phosphorus content, and decreases organic carbon and total nitrogen content. Moreover, the influence of the airflow rate and material height on smoldering characteristics was investigated. Matching the appropriate airflow rate can help maintain optimal smoldering conditions, and smoldering remains stable with increasing material height. The addition of recovered oil to a sample with a low mass ratio of oil sludge can help with smoldering ignition and improve the removal efficiency of petroleum hydrocarbons. This study has confirmed that smoldering can be used to treat OSS within a broad range of oil sludge concentrations without pretreatment.

Discovery of New Isotopes ^{160}Os and ^{156}W: Revealing Enhanced Stability of the N=82 Shell Closure on the Neutron-Deficient Side

Using the fusion-evaporation reaction ^{106}Cd(^{58}Ni,4n)^{160}Os and the gas-filled recoil separator SHANS, two new isotopes _{76}^{160}Os and _{74}^{156}W have been identified. The α decay of ^{160}Os, measured with an α-particle energy of 7080(26) keV and a half-life of 201_{-37}^{+58}  μs, is assigned to originate from the ground state. The daughter nucleus ^{156}W is a β^{+} emitter with a half-life of 291_{-61}^{+86}  ms. The newly measured α-decay data allow us to derive α-decay reduced widths (δ^{2}) for the N=84 isotones up to osmium (Z=76), which are found to decrease with increasing atomic number above Z=68. The reduction of δ^{2} is interpreted as evidence for the strengthening of the N=82 shell closure toward the proton drip line, supported by the increase of the neutron-shell gaps predicted in theoretical models.

[Annual review of bronchiectasis research in 2023]

This review focuses on the latest advances in bronchiectasis from October 1st, 2022 to September 30th, 2023, including the etiology, diagnosis, treatment, comorbidities, and management of bronchiectasis in order to provide a reference in clinical diagnosis and treatment, and future research of bronchiectasis for domestic peers.

[Efficacy of 0.05% cyclosporine A combined with vitamin A palmitate in the treatment of meibomian gland dysfunction-related dry eye]

Objective: To evaluate the efficacy of 0.05% cyclosporine A eye drops combined with vitamin A palmitate eye gel in the treatment of dry eye associated with meibomian gland dysfunction (MGD). Methods: A single-center, prospective, randomized, parallel controlled trial design was used to include patients diagnosed with MGD-associated dry eye. The patients were randomly divided into three groups and administered with medications binocularly for 12 weeks. The CsA+VA group was given 0.05% cyclosporine A eye drops twice a day and vitamin A palmitate eye gel three times a day. The CsA+HA group was given 0.05% cyclosporine A eye drops twice a day and 0.1% sodium hyaluronate eye drops three times a day. The HA group was given 0.1% sodium hyaluronate eye drops 3 times a day. The OSDI score, tear meniscus height, fluorescein tear break-up time, Schirmer Ⅰ test (without anesthesia), tear film lipid layer thickness, meibomian gland morphology and function examination, and corneal fluorescein sodium staining score were evaluated at baseline, 4, 8, and 12 weeks after the initiation of the treatment, respectively. Results: A total of 120 patients with MGD-related dry eye met the enrollment criteria, but 10 patients were lost to follow-up; 110 patients were finally included for observation, including 36 patients in the CsA+VA group, 38 in the CsA+HA group and 36 in the HA group. The OSDI score, tear meniscus height, fluorescein tear break-up time and meibomian gland secretion of the 3 groups were significantly improved. At the 12th week of the treatment, the differences of the CsA+VA group [25.45±15.11, (0.30±0.13) mm, (3.72±1.40) s, (5.03±2.52) points] and the CsA+HA group [26.98±16.89, (0.27±0.10) mm, (4.34±1.76) s, (5.11±2.39) points] from the HA group [24.57±11.26, (0.24±0.06) mm, (3.18±1.11) s, (9.11±3.34) points] were statistically significant (P<0.05). Compared with the CsA+HA group [(68.39±26.66) nm], the tear film lipid layer thickness in the CsA+VA group [(72.61±23.65) nm] was significantly increased (P<0.05). In the CsA+VA group, the meibomian gland secretion characters and discharge capacity among patients with severe abnormalities [(6.28±2.59) and (5.89±2.77) points at the 12th week of treatment], moderate abnormalities [(4.27±2.02) and (4.64±2.02) points at the 12th week of treatment] and mild abnormalities [(2.80±0.84) and (2.60±0.55) points at the 12th week of treatment] were significantly different (P<0.05). Conclusion: 0.05% cyclosporine A combined with vitamin A palmitate can significantly improve the symptoms and signs of patients with MGD-related dry eye, especially the tear film lipid layer thickness and the meibomian gland secretion characters and discharge capacity in severe cases.

Learning Interpretable Rules for Scalable Data Representation and Classification

Rule-based models, e.g., decision trees, are widely used in scenarios demanding high model interpretability for their transparent inner structures and good model expressivity. However, rule-based models are hard to optimize, especially on large data sets, due to their discrete parameters and structures. Ensemble methods and fuzzy/soft rules are commonly used to improve performance, but they sacrifice the model interpretability. To obtain both good scalability and interpretability, we propose a new classifier, named Rule-based Representation Learner (RRL), that automatically learns interpretable non-fuzzy rules for data representation and classification. To train the non-differentiable RRL effectively, we project it to a continuous space and propose a novel training method, called Gradient Grafting, that can directly optimize the discrete model using gradient descent. A novel design of logical activation functions is also devised to increase the scalability of RRL and enable it to discretize the continuous features end-to-end. Exhaustive experiments on ten small and four large data sets show that RRL outperforms the competitive interpretable approaches and can be easily adjusted to obtain a trade-off between classification accuracy and model complexity for different scenarios.

Livestock grazing modifies soil nematode body size structure in mosaic grassland habitats

Body size is closely related to the trophic level and abundance of soil fauna, particularly nematodes. Therefore, size-based analyses are increasingly prominent in unveiling soil food web structure and its responses to anthropogenic disturbances, such as livestock grazing. Yet, little is known about the effects of different livestock on the body size structure of soil nematodes, especially in grasslands characterized by local habitat heterogeneity. A four-year field grazing experiment from 2017 to 2020 was conducted in a meadow steppe characterized by typical mosaics of degraded hypersaline patches and undegraded hyposaline patches to assess the impacts of cattle and sheep grazing on the body size structure of soil nematodes within and across trophic groups. Without grazing, the hypersaline patches harbored higher abundance of large-bodied nematodes in the community compared to the hyposaline patches. Livestock grazing decreased large-bodied nematodes within and across trophic groups mainly by reducing soil microbial biomass in the hypersaline patches, with sheep grazing resulting in more substantial reductions compared to cattle grazing. The reduction in large-bodied nematode individuals correspondingly resulted in decreases in nematode community-weighted mean (CWM) body size, nematode biomass, and size spectra slopes. However, both cattle and sheep grazing had minimal impacts on the CWM body size and size spectra of total nematodes in the hyposaline patches. Our findings suggest that livestock grazing, especially sheep grazing, has the potential to simplify soil food webs by reducing large-bodied nematodes in degraded habitats, which may aggravate soil degradation by weakening the bioturbation activities of soil fauna. In light of the widespread land use of grasslands by herbivores of various species and the ongoing global grassland degradation of mosaic patches, the recognition of the trends revealed by our findings is critical for developing appropriate strategies for grassland grazing management.

Readily releasable β cells with tight Ca2+-exocytosis coupling dictate biphasic glucose-stimulated insulin secretion

Biphasic glucose-stimulated insulin secretion (GSIS) is essential for blood glucose regulation, but a mechanistic model incorporating the recently identified islet β cell heterogeneity remains elusive. Here, we show that insulin secretion is spatially and dynamically heterogeneous across the islet. Using a zinc-based fluorophore with spinning-disc confocal microscopy, we reveal that approximately 40% of islet cells, which we call readily releasable β cells (RRβs), are responsible for 80% of insulin exocytosis events. Although glucose up to 18.2 mM fully mobilized RRβs to release insulin synchronously (first phase), even higher glucose concentrations enhanced the sustained secretion from these cells (second phase). Release-incompetent β cells show similarities to RRβs in glucose-evoked Ca2+ transients but exhibit Ca2+-exocytosis coupling deficiency. A decreased number of RRβs and their altered secretory ability are associated with impaired GSIS progression in ob/ob mice. Our data reveal functional heterogeneity at the level of exocytosis among β cells and identify RRβs as a subpopulation of β cells that make a disproportionally large contribution to biphasic GSIS from mouse islets.

[Comparison of short-term safety of two anastomotic techniques when resecting Siewert type II adenocarcinoma of the esophagogastric junction: a multicenter retrospective cohort study]

Objective: In this study, we aimed to compare the short-term safety of two digestive tract reconstruction techniques, laparoscopic total abdominal overlap anastomosis and laparoscopic-assisted end-to-side anastomosis, following radical resection of Siewert Type II adenocarcinoma of the esophagogastric junction. Methods: In this retrospective cohort study, we analyzed relevant clinical data of 139 patients who had undergone radical surgery for Siewert Type II esophagogastric junction adenocarcinoma. These included 89 patients treated at the First Affiliated Hospital of Air Force Medical University from November 2021 to July 2023, 36 patients treated at the First Affiliated Hospital of Xi'an Jiaotong University from December 2020 to June 2021, and 14 patients treated at the Yuncheng Central Hospital in Shanxi Province from September 2021 to November 2022. The group consisted of 107 men (77.0%) and 32 women (23.0%) of mean age 62.5±9.3 years. Forty-eight patients underwent laparoscopic total abdominal overlap anastomosis (overlap group), and 91 laparoscopic-assisted end-to-side anastomosis (end-to-side group). Clinical data, surgical information, pathological findings, postoperative recovery, and related complications were compared between the two groups. Results: There were no significant differences in general clinical data between the overlap and end-to-side anastomosis groups (all P>0.05), indicating comparability. There was no significant difference in operation time (267.2±60.1 minutes vs. 262.8±70.6 minutes, t=0.370, P=0.712). However, the intraoperative blood loss in the overlap group (100 [50, 100] mL) was significantly lower compared to the end-to-side group (100[50, 175] mL, Z=2.776, P=0.005). Compared to the end-to-side group, longer distances between the tumor and distal resection margin proximal(1.7±1.0 cm vs. 1.3±0.9 cm, t=2.487, P=0.014) and the tumor and distal resection margin (9.5±2.9 cm vs. 7.9±3.5 cm, t=2.667, P=0.009) were achieved in the overlap group. Compared with the end-to-side group, the overlap group achieved significantly earlier postoperative ambulation (1.0 [1.0, 2.0] days vs. 2.0 [1.0, 3.0] days, Z=3.117, P=0.002), earlier time to first drink (4.7±2.6 days vs. 6.2±3.0 days, t=2.851, P=0.005), and earlier time to first meal (6.0±2.7 days vs. 7.1±3.0 days, t=2.170, P=0.032). However, the hospitalization costs were higher in the overlap group (113, 105.5±37, 766.3) yuan vs. (97, 250.2±27, 746.9) yuan; this difference is significant (t=2.818, P=0.006). There were no significant differences between the two groups in postoperative hospital stay, total number of lymph nodes cleared, or time to first postoperative flatus (all P>0.05). The incidence of surgery-related complications was 22.9%(11/48) in the overlap group and 19.8% (18/91) in the end-to-side group; this difference is not significant (χ²=0.187, P=0.831). Further comparison of complications using the Clavien-Dindo classification also showed no significant differences (Z=0.406, P=0.685). Conclusions: Both laparoscopic total abdominal overlap anastomosis and laparoscopic-assisted end-to-side anastomosis are feasible for radical surgery for Siewert Type II esophagogastric junction adenocarcinoma. Laparoscopic total abdominal overlap anastomosis achieves longer proximal and distal resection margins and better postoperative recovery; however, end-to-side anastomosis is more cost-effective.

[Application value of questionnaires in the screening obstructive sleep apnea syndrome in pregnancy across trimesters]

Objective: To evaluate the clinical utility value of questionnaires of Berlin, STOP, STOP-Bang (SBQ), Epworth Sleepiness Scale (ESS) in screening obstructive sleep apnea syndrome (OSAS) in pregnant women of different trimesters. Methods: Consecutive pregnant women at high risk for OSAS were enrolled from January, 2021 to April, 2022 at the obstetric clinic of Peking University People's Hospital. They completed questionnaires of Berlin, STOP, SBQ, ESS and also underwent an overnight polysomnography (PSG). To evaluate the accuracy of questionnaires of Berlin, STOP, SBQ, ESS, sensitivity, specificity, positive predictive values, negative predictive values and the area under the receiver operating characteristics (ROC) curve of these questionnaires in pregnancy across trimesters (Pregnancy 1-15 weeks was the first stage, pregnancy 16-27 weeks was the second stage, and pregnancy 28-40 weeks was the third stage) were calculated. Results: A total of 100 pregnant women [(34.5±4.3) years old (26-46 years old)] were included in this study, including 20, 35 and 45 pregnant women in the first, second and third trimester of pregnancy, respectively. Based on PSG results, 45 (45%) of 100 pregnant women were diagnosed with OSAS. The overall predictive values of the four questionnaires were not good, area under[AUC(95%CI)] the ROC curve ESS, Berlin questionnaire STOP and SBQ were 0.54(0.43, 0.66), 0.59 (0.47, 0.70), 0.62(0.51, 0.73) and 0.61 (0.49, 0.72), respectively, sensitivity was 35.6%, 65.9%, 48.9%, 28.9%, specificity was 71.7%, 52.5%, 73.6%, 92.5%. When categorized according to trimesters, the predicted values of the four questionnaires increased in the first trimester, the AUC (95%CI) of STOP questionnaire was 0.81 (0.61, 1.00), sensitivity was 75.0%, specificity was 87.5%. Conclusion: The overall predictive power of the four screening questionnaires is limited in pregnant women. But predictive value of STOP questionnaire is acceptable in the first trimester.

Translatability of In Vitro Stress for Predicting Deamidation and Oxidation Biotransformation on Biotherapeutics

Biotransformation leading to single residue modifications (e.g., deamidation, oxidation) can contribute to decreased efficacy/potency, poor pharmacokinetics, and/or toxicity/immunogenicity for protein therapeutics. Identifying and characterizing such liabilities in vivo are emerging needs for biologics drug discovery. In vitro stress assays involving PBS for deamidation or AAPH for oxidation are commonly used for predicting liabilities in manufacturing and storage and are sometimes considered a predictive tool for in vivo liabilities. However, reports discussing their in vivo translatability are limited. Herein, we introduce a mass spectrometry workflow that characterizes in vivo oxidation and deamidation in pharmacokinetically relevant compartments for diverse protein therapeutic modalities. The workflow has low bias of <10% in quantitating degradation in the relevant pharmacokinetic concentration range for monkey and rabbit serum/plasma (1-100 μg/mL) and allows for high sequence coverage (∼85%) for discovery/monitoring of amino acid modifications. For oxidation and deamidation, the assay was precise, with percent coefficient of variation of <8% at 1-100 μg/mL and ≤6% method-induced artifacts. A high degree of in vitro and in vivo correlation was observed for deamidation on the six diverse protein therapeutics (seven liability sites) tested. In vivo translatability for oxidation liabilities were not observed for the 11 molecules tested using in vitro AAPH stress. One of the molecules dosed in eyes resulted in a false positive and a false negative prediction for in vivo oxidation following AAPH stress. Finally, peroxide stress was also tested but resulted in limited success (1 out of 4 molecules) in predicting oxidation liabilities.

Structural basis of antibody inhibition and chemokine activation of the human CC chemokine receptor 8

The C-C motif chemokine receptor 8 (CCR8) is a class A G-protein coupled receptor that has emerged as a promising therapeutic target in cancer. Targeting CCR8 with an antibody has appeared to be an attractive therapeutic approach, but the molecular basis for chemokine-mediated activation and antibody-mediated inhibition of CCR8 are not fully elucidated. Here, we obtain an antagonist antibody against human CCR8 and determine structures of CCR8 in complex with either the antibody or the endogenous agonist ligand CCL1. Our studies reveal characteristic antibody features allowing recognition of the CCR8 extracellular loops and CCL1-CCR8 interaction modes that are distinct from other chemokine receptor - ligand pairs. Informed by these structural insights, we demonstrate that CCL1 follows a two-step, two-site binding sequence to CCR8 and that antibody-mediated inhibition of CCL1 signaling can occur by preventing the second binding event. Together, our results provide a detailed structural and mechanistic framework of CCR8 activation and inhibition that expands our molecular understanding of chemokine - receptor interactions and offers insight into the development of therapeutic antibodies targeting chemokine GPCRs.

First observation of fluid-like eddy-dominant bursty bulk flow turbulence in the Earth's tail plasma sheet

Turbulence is a ubiquitous phenomenon in neutral and conductive fluids. According to classical theory, turbulence is a rotating flow containing vortices of different scales. Eddies play a fundamental role in the nonlinear cascade of kinetic energy at different scales in turbulent flow. In conductive fluids, the Alfvénic/kinetic Alfvénic wave (AW/KAW) is the new "cell" of magnetohydrodynamic (MHD) turbulence (frozen-in condition). Wave energy, which has equal kinetic and magnetic energy, is redistributed among multiple-scale Fourier modes and transferred from the large MHD scale to the small kinetic scale through the collision of counter-propagating Alfvénic wave packages propagating along the magnetic field line. Fluid-like eddy-dominant plasma flow turbulence has never been found in space since the launch of the first satellite in 1957. In this paper, we report the first observation of eddy-dominant turbulence within magnetic reconnection-generated fast flow in the Earth's tail plasma sheet by the Magnetospheric Multiscale Spacecraft (MMS). In eddy-dominant turbulent reconnection jet, ions dominate the flow field while electrons dominate current and magnetic fluctuations. Our findings shed new light on the nonlinear kinetic and magnetic energy cascade in MHD turbulence.

[Effects of electric welding on hearing loss and respiratory damage]

Objective: To investigate the occupational health status of electric welding workers and explore the effects of electric welding on hearing loss and respiratory damage. Methods: From August to December 2021, the cluster sampling method was used to select workers from an automobile manufacturer in Guangzhou City as research subjects: 636 welding workers in the welding workshop as the welding group, 757 assembly workers in the engine workshop and the final assembly workshop exposed to pure noise as the assembly group. Occupational disease hazard factors were detected for welding positions and assembly positions, and occupational health examination was carried out for research subjects. The occupational health status, the trends of hearing loss and respiratory abnormalities with working age were compared and analyzed between the two groups. Binary logistic regression was used to analyze the association between hearing loss and respiratory abnormalities in welding workers. Results: The excess rates of welding fumes, manganese and its compounds in the welding position were both 9.68% (3/31). Its noise exposure intensity [ (85.36±2.68) dB (A) ] and excess rate [48.39% (15/31) ] were not significantly different from those in the assembly position [ (84.86±3.28) dB (A) and 43.24% (16/37) ] (P>0.05). The results of the occupational health examination showed that the detection rates of hearing loss, digital radiography (DR) chest X-ray abnormality, alanine aminotransferase abnormality, deazelaic aminotransferase abnormality and white blood cell count abnormality of workers in the welding group were higher than those in the assembly group (P<0.05). The detection rates of hearing loss, DR chest X-ray abnormality, pulmonary ventilation abnormality in the welding group and the detection rate of hearing loss in the assembly group increased with the working age of the workers (P<0.05). The hearing loss detection rate and DR chest X-ray abnormality detection rate of the workers with ≥9 years working age in the welding group were both higher than those in the assembly group workers with same working age (P<0.05). The binary logistic regression analysis showed that abnormal pulmonary ventilation and abnormal DR chest X-ray were the risk factors for hearing loss in welding workers (OR=10.83, 95%CI: 7.31-16.06; OR=16.59, 95%CI: 5.72-48.10; P<0.05) . Conclusion: Hearing loss and respiratory damage are prominent problems among welding workers, and the detection rates of abnormality increase with the working age of the workers. Hearing loss in welding workers is associated with abnormal pulmonary ventilation and abnormal DR chest X-ray.

A novel, label-free, pre-equilibrium assay to determine the association and dissociation rate constants of therapeutic antibodies on living cells

BACKGROUND AND PURPOSE

Monoclonal antibodies (Ab) represent the fastest growing drug class. Knowledge of the biophysical parameters (kon , koff and KD ) that dictate Ab:receptor interaction is critical during the drug discovery process. However, with the increasing complexity of Ab formats and their targets, it became apparent that existing technologies present limitations and are not always suitable to determine these parameters. Therefore, novel affinity determination methods represent an unmet assay need.

EXPERIMENTAL APPROACH

We developed a pre-equilibrium kinetic exclusion assay using recent mathematical advances to determine the kon , koff and KD of monoclonal Ab:receptor interactions on living cells. The assay is amenable to all human IgG1 and rabbit Abs.

KEY RESULTS

Using our novel assay, we demonstrated for several monoclonal Ab:receptor pairs that the calculated kinetic rate constants were comparable with orthogonal methods that were lower throughput or more resource consuming. We ran simulations to predict the critical conditions to improve the performance of the assays. We further showed that this method could successfully be applied to both suspension and adherent cells. Finally, we demonstrated that kon and koff , but not KD , correlate with in vitro potency for a panel of monoclonal Abs.

CONCLUSIONS AND IMPLICATIONS

Our novel assay has the potential to systematically probe binding kinetics of monoclonal Abs to cells and can be incorporated in a screening cascade to identify new therapeutic candidates. Wide-spread adoption of pre-equilibrium assays using physiologically relevant systems will lead to a more holistic understanding of how Ab binding kinetics influence their potency.

Fixed-Field IMRT for Cervix Carcinoma Patients on an MR-LINAC Platform: Dosimetric Feasibility and Challenges

PURPOSE/OBJECTIVE(S)

To investigate the impact of MR-LINAC performance characteristics and inverse planning implementation on the feasibility of fixed-field IMRT for cervix carcinoma patients by benchmarking MR-LINAC plans against clinically used VMAT plans in a single institution study and multi-institutional treatment planning challenge.

MATERIALS/METHODS

For 10 cervix cancer patients who had previously received Linac-based VMAT, new treatment plans were optimized for MR-LINAC IMRT using 6X FFF fixed fields with maximum available field size of 27.4 x 24.1 cm2. Dose optimization was performed on the clinically used planning CT and structure set. Prescribed dose was 48.6 Gy in 27 fractions for all patients with 6 patients receiving an additional integrated boost for a total of 58.05 Gy to involved nodes. Constraints were based on our institutional protocol as per Table 1. IMRT delivery time was limited to 20 min. Original clinically used VMAT plans were generated on Eclipse (Varian Medical System) using 3 to 4 arcs. For the multi-institutional planning challenge, the data set from a single patient was anonymized and shared to participants. Participants used a single MR-based Linac planning platform to generate a plan based on our institutional constraints, with maximum treatment time limited to 20 min. For all analyses, a paired samples t-test was used to compare the significance defined at p < 0.05.

RESULTS

For MR-LINAC plans, the mean number of fields used was 23, mean number of segments 229, and the average estimated treatment delivery time was 17.3 minutes. MR-LINAC plans showed a significantly higher heterogeneity and dose to organs at risk compared to VMAT plans (Table 1). For the planning challenge, a total of 30 participants registered interest. Of this, seven plans were submitted to the challenge. On average, participants generated a plan that would be acceptable based on our institutional constraints (Table 1). However, the volumetric dose to bowel and pelvic bones were higher on MR-LINAC plans compared to the reference VMAT plan.

CONCLUSION

MR-LINAC fixed-field IMRT for cervix cancer patients is feasible but system constraints and optimization implementation result in greater dose heterogeneity and worse organ-at-risk sparing compared to Linac based VMAT. Further research is needed to determine if potential reduction of treatment margins, allowed by better MRI soft-tissue visualization, will result in MR-LINAC IMRT superior to Linac VMAT.

Clinical Implementation of an Automated IMRT/VMAT Treatment Planning Tool

PURPOSE/OBJECTIVE(S)

To create an in-house automated treatment planning tool for IMRT/VMAT treatments and evaluate the dosimetric plan quality against manually generated plans.

MATERIALS/METHODS

A scripting application programming interface is employed to interact with a commercial treatment planning system (TPS) to implement automatic plan evaluation and update optimization parameters by mimicking the human planning process. The automated planning performs in an iterative fashion until reaching an acceptable tradeoff among target coverage/dose homogeneity and sparing of critical organs at risk. In each iteration, the dose constraints, priorities, and optimization structures for are automatically updated based on the results of the current iteration. Twenty previously treated plans (10 prostate and 10 head and neck), were preliminarily used to evaluate the performance of the automated planning tool. The differences in target and organ-at-risk metrics from the manually generated clinical plans were analyzed using paired t-test to evaluate clinical acceptability of tour automated planning tool. The current in-house-developed automated planning solution is able to create plans for different disease sites, including head & neck, prostate, pelvis, and lung. So far, the VMAT plans for more than 150 different cases have been generated with the tool. The results for these were also evaluated.

RESULTS

Compared to the manually generated clinical head and neck plans, all auto plans achieved PTV D95% coverage and critical organs at risk sparing without statistically significant change in average global Dmax (107.4% for manual vs 107.3% for automated plans). The auto-planning solution provided reduced maximum doses to brainstem and spinal cord (average reductions with standard deviations of 5.1 ± 2.6 Gy and 2.9 ± 1.4 Gy, respectively, all p <0.03), reduced average mean doses to contralateral parotid, ipsilateral parotid, contralateral submandibular gland, pharynx, esophagus, cochleae (reductions of 2.2 ± 2.9 Gy, 4.8 ± 4.7 Gy, 3.6 ± 5.2 Gy, 2.0 ± 7.1 Gy, 3.9 ± 2.6 Gy, 3.8 ± 5.0 Gy, respectively, all p < 0.045). Similar results were observed for the prostate plans. With the same PTV coverage and without statistically significant change in average global Dmax (106.5% for manual vs 106.8% for automated plans), the automated solution provided superior sparing for both bladder and rectum. Bladder V75, V70, V65 were reduced by 0.6% ± 2.1%, 0.8% ± 2.5%, and 0.9% ± 2.9% (all p <0.04), respectively. Rectum V75, V70, V65, V60 were reduced by 1.0% ± 2.3%, 1.2% ± 2.8%, 1.3% ± 3.2%, 1.6% ± 3.6% (all p < 0.01), respectively.

CONCLUSION

Our automated treatment planning solution is capable of efficiently generating VMAT plans for different disease sites with superior dosimetric indices compared to manually generated plans. Our tool is integrated within a commercial TPS platform, so it has the advantage of seamless adoption into the standard workflow to improve plan quality and treatment planning efficiency in our clinic.

The Lyman Normal Tissue Complication Probability Model and Risk Prediction for Temporal Lobe Injury after Re-Irradiation in Patients with Recurrent Nasopharyngeal Carcinoma

PURPOSE/OBJECTIVE(S)

The risk of temporal lobe injury (TLI) in recurrent nasopharyngeal carcinoma (rNPC) patients with intensity-modulated radiation therapy (IMRT) is high. We aimed to construct the normal tissue complication probability (NTCP) model for TLI of rNPC and establish a risk predictive model.

MATERIALS/METHODS

We retrospectively analyzed 103 patients with rNPC who had received two courses of IMRT in our institution. The 206 temporal lobes (TLs) of these patients were randomly divided into a training (n = 144) and validation group (n = 62). We determined the mean value of the following parameters to construct the Lyman NTCP model: TD50(1) (the dose with a 50% probability of complications to an organ when all volumes are irradiated), m [steepness of the dose-response at TD50(1)], and n (the parameter related to volume effect). The most predictive dosimetric parameter and clinical variables were integrated in Cox proportional hazards models. A nomogram was developed for predicting risk of TLs.

RESULTS

The parameters of the fitted NTCP model were TD50(1) = 107.84 Gy (95% confidence interval (CI), [97.15, 118.54]), m = 0.16 (95% CI, [0.14, 0.19]), and n = 0.04 (95% CI, [0.01, 0.06]). The cumulative dose delivered to 0.1 cm3 of temporal lobe volume (D0.1cc-c) was the most predictive dosimetric parameter for TLI. The Kaplan-Meier curves showed a significant difference in 2-year TLI-free survival among different risk groups according to the total score of nomograms.

CONCLUSION

The TD50(1) of TLI in patients with rNPC is 107.84 Gy in Lyman NTCP model. The nomogram model can accurately predict the risk of TLI for individual.

Prospective Clinical Adoption of Artificial Intelligence for Organ Contouring in Head and Neck Radiation Treatment Planning

PURPOSE/OBJECTIVE(S)

Patients that undergo head and neck (H&N) radiation therapy (RT) require laborious delineation of organs-at-risk (OARs) on computed tomography (CT) scans in a treatment planning system (TPS) to minimize radiation to normal tissue. This task can be completed rapidly and accurately with recently developed artificial intelligence-based semantic segmentation models. The current study aims to deploy and evaluate a strategy for improving clinical practice with this technology.

MATERIALS/METHODS

Deep learning models were trained and tested with CT scans and OAR contours from previous H&N RT cases at our clinic. Two medical physicists vetted the models and selected a 2.5D U-Net for further implementation. The model was embedded in a dedicated server at the hospital, programmed to read H&N CT scans staged for import into the TPS, generate auto-contours, and write them into a TPS-compatible format made available alongside the scan. In the pilot implementation, the auto-contouring service was utilized for more than 60 cases, prospectively. The auto-contours were quantitatively evaluated against the treatment-approved contours to determine how much modification was performed by the clinical team.

RESULTS

The 2.5D U-Net selected for clinical integration segments 21 OARs in less than 3 minutes per scan. Across all the prospective cases, the mean Dice score and mean 95th percentile Hausdorff distance (mm) between the auto-contour and treatment-approved contour for each of the 21 OARs were as follows, respectively: brainstem (0.93, 1.94), optic chiasm (0.70, 2.96), left cochlea (0.69, 2.37), right cochlea (0.68, 2.44), esophagus (0.88, 2.46), left globe (0.93, 1.50), right globe (0.93, 1.63), glottis (0.91, 2.13), larynx (0.93, 2.76), mandible (0.90, 4.86), left optic nerve (0.78, 1.64), right optic nerve (0.82, 1.65), oral cavity (0.86, 8.46), left parotid gland (0.91, 2.78), right parotid gland (0.91, 2.39), pharynx (0.85, 2.39), spinal cord (0.87, 2.27), left submandibular gland (0.85, 3.46), right submandibular gland (0.83, 3.69), left temporal lobe (0.94, 2.20), and right temporal lobe (0.95, 2.09). The auto-contours for the optic chiasm, optic nerves, cochleas, and submandibular glands differed substantially from the final contours, a finding corroborated by the clinical team; the rest were clinically acceptable with minor or no edits necessary.

CONCLUSION

The proposed strategy provides a sophisticated starting point for treatment planning that has garnered overall favorable feedback from the participating radiation oncologists and dosimetrists. Consequently, the technique is being extended to other treatment sites.

Parametric Response Mapping of Co-Registered Ivim MRI and PET to Identify Radioresistant Sub-Volumes in Locally Advanced Cervical Carcinoma Undergoing CCRT

PURPOSE/OBJECTIVE(S)

To investigate parametric response mapping (PRM) of co-registered positron emission tomography (PET) and intravoxel incoherent motion (IVIM) MRI in locally advanced cervical carcinoma patients to identify sub-volumes that may predict treatment response to adjuvant concurrent chemoradiation therapy (CCRT).

MATERIALS/METHODS

Pre- and on-treatment (∼after four weeks of CCRT) diffusion weighted images (DWIs) MRI and pre-treatment PET were performed on 20 cervical cancer patients (median [range] age: 63 [41 to 85]; FIGO staging: 3 IIB, 1 IIIB, 10 IIIC1, 4 IIIC2, 2 IVA; histology: 11 squamous cell carcinoma, 4 endocervical adenocarcinoma; LN status: 4 negative, 16 positive). IVIM MRI maps were generated by Bayesian fitting of a two-compartment IVIM model to the DWIs to produce three-dimensional diffusion coefficient (D) and perfusion fraction (f). Three-dimensional IVIM D and f maps were co-registered to PET standardized uptake value (SUV) maps. Population means of PET SUV, IVIM D and f from pre-treatment scans were calculated to generate thresholds to identify low versus high metabolic (μSUV) regions, low versus high diffused (μD) regions as well as low versus high perfused (μf) regions, respectively. PRM maps were generated using voxel-wise joint histogram analysis to classify voxels within the tumor as highly metabolic and with low cellular density (SUV↑D↑), highly metabolic and with high cellular density (SUV↑D↓), metabolically inactive and with high cellular density (SUV↓D↓), or metabolically inactive and with low cellular density (SUV↓D↑) tissue based on the population mean thresholds. Similar PRM maps were generated using the joint histogram analysis with SUV and f. Summary statistics for these and other imaging parameters were calculated pre- and on-treatment. Univariate analysis was performed to determine relationships between relative change in gross tumor volume (ΔGTV) and pre-treatment imaging measurements.

RESULTS

In tumors, on- versus pre-treatment tumor volume (p<.001) significantly decreased, while IVIM f (p = .002) and D (p = .03) significantly increased. Pre-treatment tumor volume (r = .45, p = .04) and PRM SUV↑D↓ (r = .65, p = .002) regions were positively related with ΔGTV, while pre-treatment IVIM D (r = -.64, p = .002), PRM SUV↓f↑ (r = -.52, p = .02) and PRM SUV↓D↑ (r = -.74, p<.001) regions were negatively related with ΔGTV. The latter PRM result suggests that larger regions of low cellular density (as represented by elevated D) + low metabolism result in a reduced change in tumor volume on-treatment, potentially representing a radioresistant sub-volume.

CONCLUSION

PRM, generated from PET and IVIM MRI, was applied to patients with locally advanced cervical carcinoma and radioresistant sub-volumes were identified which may predict treatment response. The complementary information provided from PET and IVIM, combined using PRM, may assist in decision-making to individualize therapies, such as with image guided brachytherapy, to improve patient outcomes.

Prediction of Progression After Cervix Cancer Radiotherapy Using a Machine-Learning Model on Pre-Treatment MRI

PURPOSE/OBJECTIVE(S)

MRI may be useful to identify women with cervical cancer at high risk of disease progression to test strategies of treatment intensification. The purpose of this study was to determine the value of a machine-learning model built on pre-treatment MRI for prediction of risk of progression after radiation therapy.

MATERIALS/METHODS

MagneticResonance Imaging (MRI) data for women with cervical cancer was collected from The Cancer Genome Atlas Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma Collection (TCGA-CESC) on the Cancer Imaging Archive (TCIA), which reported clinical, treatment, and imaging data from a single institution. 27 patients who had received radiation for cervical cancer were selected for input into a custom 3-D Residual Neural Network (ResNet) model with added custom layers specific to DICOM data in tensorflow python package. One T2 MRI per patient was used to predict recurrence free survival after radiation treatment, where patients were predicted to be "high risk" or "low risk" for disease recurrence as the output of the model. All slices of the T2 MRI were used. The model was validated using five-fold cross validation; 80% of the data was used to train each fold and 20% was used for testing. Final model statistical significance was confirmed through shuffle test at the p < 0.01 level. The clinical outcomes of patients and the model's "low-risk" and "high-risk" prediction were compared.

RESULTS

There were 27 patients in the study with mean age of 51 years (range 29-79). 20 patients had squamous cell carcinoma and 7 patients had adenocarcinoma. The stage breakdown consisted of 9 women IB, 2 IIA, 9 IIB, 2 IIIA, 2 IIIB, and 3 stage IV. 10 women were treated with radiation alone and 17 with chemo-radiation. 5 women received surgery in addition to radiation or chemoradiation. 21 patients received brachytherapy. Median follow-up of patients was 29 months (range 3-64). The model predicted 7 patients as "high risk" for recurrence; all 7 developed a recurrence during follow up. None of the 20 patients predicted to be "low risk" developed disease recurrence. Among all patients in the study, the two-year progression free survival (PFS) was 82.0%. Patients identified as "low risk" and "high risk" by model had two-year PFS of 100% and 43%, respectively. Among patients with recurrence, 3 developed local recurrence and 4 developed distant metastases. The ResNet model achieved cross-validated accuracy of 92% for prediction of progression-free survival (p<0.01).

CONCLUSION

A 3-D ResNet machine-learning model using pretreatment MRI image data can accurately predict clinical outcomes for cervical cancer following radiation therapy. Future work to confirm generalizability should focus on validation with a larger clinical dataset.

[Relationships between hypertensive disorders in pregnancy and obstructive sleep apnea syndrome]

Objective: To investigate the impact of obstructive sleep apnea syndrome (OSAS) on pregnancy outcomes, especially the relationship between OSAS and hypertensive disorders in pregnancy (HDP). Methods: A total of 228 pregnant women with high risk of OSAS who underwent sleep monitoring during pregnancy in Peking University People's Hospital from January 2021 to April 2022 were collected by reviewing their medical records for retrospective analysis. According to the diagnosis of OSAS, the pregnant women were divided into OSAS group (105 cases) and non-OSAS group (123 cases). The non-parametric Mann-Whitney U test, χ2 test or Fisher's exact test were used to compare the general data and maternal and fetal outcomes between the two groups, and the occurrence of each type of HDP was further compared. Results: (1) Compared with the non-OSAS group, the median pre-pregnancy body mass index (23.6 vs 27.6 kg/m2) and the proportion of snoring [28.9% (33/114) vs 59.2% (61/103)] in the OSAS group were higher, and the differences were both statistically significant (both P<0.001). (2) The incidence of HDP [67.6% (71/105) vs 39.0% (48/123)] and gestational diabetes mellitus [GDM; 40.0% (42/105) vs 26.8% (33/123)] of pregnant women in the OSAS group were higher than those in the non-OSAS group, and the median delivery week was shorter than that in the non-OSAS group (38.4 vs 39.0 weeks). The differences were all statistically significant (all P<0.05). Between-group differences for the delivery way, postpartum hemorrhage, the rate of intensive care unit admission, preterm birth, small for gestational age infants, neonatal asphyxia, the rate of neonatal intensive care unit admission, newborn birth weight and the proportion of umbilical artery blood pH<7.00 were not statistically significant (all P>0.05). (3) Compared with the non-OSAS group, the incidence of chronic hypertension [11.4% (14/123) vs 22.9% (24/105)] and chronic hypertension with superimposed pre-eclampsia [11.4% (14/123) vs 30.5% (32/105)] were higher in the OSAS group, and the differences were both statistically significant (both P<0.01). Conclusion: OSAS is related to HDP (especially chronic hypertension and chronic hypertension with superimposed pre-eclampsia) and GDM, which could provide a practical basis for the screening, diagnosis and treatment of OSAS in pregnant women at high risk.

[Dose-dense paclitaxel plus carboplatin in combination with trastuzumab neoadjuvant versus standard adjuvant therapy in human epidermal growth factor receptor-2 positive and hormone receptor negative breast cancer: a prospective cohort study]

Objective: To provide survival evidence of anthracycline-free neoadjuvant chemotherapy for patients with stages Ⅱ-Ⅲ human epidermal growth factor receptor-2 (HER-2) positive and hormone receptor (HR) negative breast cancer. Methods: The prospective cohort study was conducted at the Department of Medical Oncology of Cancer Hospital, Chinese Academy of Medical Sciences. Patients with HER-2 positive and HR negative breast cancer in stages Ⅱ-Ⅲ were enrolled to receive neoadjuvant therapy (NAT) of dose-dense paclitaxel (175 mg/m(2)) plus carboplatin (AUC=4.0) biweekly for 6 cycles in combination with trastuzumab (PCbH), and matched patients who received standard adjuvant therapy of physicians' choice were recruited for survival and safety comparison. Results: From July 2013 to November 2019, 166 patients were included (neoadjuvant 51, adjuvant 115). Compared with those who received adjuvant therapy, patients receiving NAT were younger (<35 years: 19.6% vs 5.2%, P=0.014), had larger tumors (T3: 62.7% vs 7.8%, P<0.001) and more advanced diseases (stage ⅡA: 2.0% vs 41.7%, P<0.001). Patients in the neoadjuvant group all received surgery, and 96 (83.5%) in the adjuvant group received anthracycline-and-taxane-containing regimens. A total of 98 patients (49 pairs) were matched, and the covariates between the two groups were acceptably balanced. Within a median follow-up of 46.5 (range, 14-87) months, the 4-year recurrence-free survival (RFS) rate among patients who received NAT was 73.3% (95% CI: 59.0%-87.6%), versus 80.6% (95% CI: 67.9%-93.3%) among those in the adjuvant group without statistical difference (P=0.418). A similar result was observed for the 4-year overall survival (OS) [neoadjuvant versus adjuvant: 91.5% (95% CI: 81.7%-100.0%) vs 97.8% (95% CI: 93.5%-100.0%), P=0.314]. Compared with standard adjuvant therapy, PCbH was related to less neutropenia and better cardiac safety. Conclusions: These results support the consideration of anthracycline-free neoadjuvant chemotherapy combined with anti-HER-2 therapy for patients with stages Ⅱ-Ⅲ HER-2-positive and HR-negative breast cancer. Optimized regimens with both efficacy and safety are needed and to be further investigated.

Toward a stable and low-resource PLM-based medical diagnostic system via prompt tuning and MoE structure

Machine learning (ML) has been extensively involved in assistant disease diagnosis and prediction systems to emancipate the serious dependence on medical resources and improve healthcare quality. Moreover, with the booming of pre-training language models (PLMs), the application prospect and promotion potential of machine learning methods in the relevant field have been further inspired. PLMs have recently achieved tremendous success in diverse text processing tasks, whereas limited by the significant semantic gap between the pre-training corpus and the structured electronic health records (EHRs), PLMs cannot converge to anticipated disease diagnosis and prediction results. Unfortunately, establishing connections between PLMs and EHRs typically requires the extraction of curated predictor variables from structured EHR resources, which is tedious and labor-intensive, and even discards vast implicit information.In this work, we propose an Input Prompting and Discriminative language model with the Mixture-of-experts framework (IPDM) by promoting the model's capabilities to learn knowledge from heterogeneous information and facilitating the feature-aware ability of the model. Furthermore, leveraging the prompt-tuning mechanism, IPDM can inherit the impacts of the pre-training in downstream tasks exclusively through minor modifications. IPDM remarkably outperforms existing models, proved by experiments on one disease diagnosis task and two disease prediction tasks. Finally, experiments with few-feature and few-sample demonstrate that IPDM achieves significant stability and impressive performance in predicting chronic diseases with unclear early-onset characteristics or sudden diseases with insufficient data, which verifies the superiority of IPDM over existing mainstream methods, and reveals the IPDM can powerfully address the aforementioned challenges via establishing a stable and low-resource medical diagnostic system for various clinical scenarios.

Integrated Transcriptomic and Metabolomic Analyses Uncover the Differential Mechanism in Saline-Alkaline Tolerance between Indica and Japonica Rice at the Seedling Stage

Saline-alkaline stress is one of the major damages that severely affects rice (Oryza sativa L.) growth and grain yield; however, the mechanism of the tolerance remains largely unknown in rice. Herein, we comparatively investigated the transcriptome and metabolome of two contrasting rice subspecies genotypes, Luohui 9 (abbreviation for Chao2R under study, O. sativa ssp. indica, saline-alkaline-sensitive) and RPY geng (O. sativa ssp. japonica, saline-alkaline-tolerant), to identify the main pathways and important factors related to saline-alkaline tolerance. Transcriptome analysis showed that 68 genes involved in fatty acid, amino acid (such as phenylalanine and tryptophan), phenylpropanoid biosynthesis, energy metabolism (such as Glycolysis and TCA cycle), as well as signal transduction (such as hormone and MAPK signaling) were identified to be specifically upregulated in RPY geng under saline-alkaline conditions, implying that a series of cascade changes from these genes promotes saline-alkaline stress tolerance. The transcriptome changes observed in RPY geng were in high accordance with the specifically accumulation of metabolites, consisting mainly of 14 phenolic acids, 8 alkaloids, and 19 lipids based on the combination analysis of transcriptome and metabolome. Moreover, some genes involved in signal transduction as hub genes, such as PR5, FLS2, BRI1, and NAC, may participate in the saline-alkaline stress response of RPY geng by modulating key genes involved in fatty acid, phenylpropanoid biosynthesis, amino acid metabolism, and glycolysis metabolic pathways based on the gene co-expression network analysis. The present research results not only provide important insights for understanding the mechanism underlying of rice saline-alkaline tolerance at the transcriptome and metabolome levels but also provide key candidate target genes for further enhancing rice saline-alkaline stress tolerance.

[Analysis of risk factors of radiation-induced toxicity in limited-stage small cell lung cancer treated with hypofractionated intensity-modulated radiotherapy]

Objective: To compare the incidence of radiation-related toxicities between conventional and hypofractionated intensity-modulated radiation therapy (IMRT) for limited-stage small cell lung cancer (SCLC), and to explore the risk factors of hypofractionated radiotherapy-induced toxicities. Methods: Data were retrospectively collected from consecutive limited-stage SCLC patients treated with definitive concurrent chemoradiotherapy in Cancer Hospital of Chinese Academy of Medical Sciences from March 2016 to April 2022. The enrolled patients were divided into two groups according to radiation fractionated regimens. Common Terminology Criteria for Adverse Events (CTCAE, version 5.0) was used to evaluate the grade of radiation esophagus injuries and lung injuries. Logistic regression analyses were used to identify factors associated with radiation-related toxicities in the hypofractionated radiotherapy group. Results: Among 211 enrolled patients, 108 cases underwent conventional IMRT and 103 patients received hypofractionated IMRT. The cumulative incidences of acute esophagitis grade ≥2 [38.9% (42/108) vs 35.0% (36/103), P=0.895] and grade ≥ 3 [1.9% (2/108) vs 5.8% (6/103), P=0.132] were similar between conventional and hypofractionated IMRT group. Late esophagus injuries grade ≥2 occurred in one patient in either group. No differences in the cumulative incidence of acute pneumonitis grade ≥2[12.0% (13/108) vs 5.8% (6/103), P=0.172] and late lung injuries grade ≥2[5.6% (6/108) vs 10.7% (11/103), P=0.277] were observed. There was no grade ≥3 lung injuries occurred in either group. Using multiple regression analysis, mean esophageal dose ≥13 Gy (OR=3.33, 95% CI: 1.23-9.01, P=0.018) and the overlapping volume between planning target volume (PTV) and esophageal ≥8 cm(3)(OR=3.99, 95% CI: 1.24-12.79, P=0.020) were identified as the independent risk factors associated with acute esophagitis grade ≥2 in the hypofractionated radiotherapy group. Acute pneumonitis grade ≥2 was correlated with presence of chronic obstructive pulmonary disease (COPD, P=0.025). Late lung injuries grade ≥2 was correlated with tumor location(P=0.036). Conclusions: Hypofractionated IMRT are tolerated with manageable toxicities for limited-stage SCLC patients treated with IMRT. Mean esophageal dose and the overlapping volume between PTV and esophageal are independently predictive factors of acute esophagitis grade ≥2, and COPD and tumor location are valuable factors of lung injuries for limited-stage SCLC patients receiving hyofractionated radiotherapy. Prospective studies are needed to confirm these results.

[The prospect and challenge of liquid biopsy in the diagnosis and treatment of chest malignancy]

In recent years, the incidence of chest malignant tumors in China has increased year by year, which has seriously threatened the health problems of people. Among them, early screening and intervention of patients with chest malignancies is the key to cancer prevention. Early detection, early diagnosis, and early treatment as the "three early prevention" of clinical practice are conducive to improve the survival rate of tumor patients. As a non-invasive and real-time reflection of tumor status, liquid biopsy has gradually received attention in clinical diagnosis and treatment. Circulating tumor cells (CTCs), circulating tumor DNA (ctDNA) and exosomes as liquid biopsy "Three carriages" are not only widely used in the diagnosis, monitoring and prognostic evaluation of chest malignancies, but also face many unknown challenges. In this article, the application of liquid biopsy in chest malignancies in recent years is elaborated in detail, which provides a reference for the formulation of clinical tumor prevention and diagnosis and treatment strategies.

[Spatial correlation between the prevalence of dental fluorosis and the chemical elemental composition of drinking water sources in a typical coal-fired pollution fluorosis area]

Objective: To investigate the spatial distribution characteristics and correlation between the prevalence of dental fluorosis and the chemical elemental composition of drinking water sources in coal-fired fluorosis areas. Methods: Based on the survey data on the prevalence of dental fluorosis at CDC in Guizhou Province in 2022, 274 original surface drinking water sources were collected in typical coal-fired fluorosis areas, and fluoride (F), calcium (Ca), magnesium (Mg), aluminum (Al), titanium (Ti), chromium (Cr), manganese (Mn), iron (Fe), nickel (Ni), copper (Cu), zinc (Zn), arsenic (As), selenium (Se), molybdenum (Mo), cadmium (Cd), barium (Ba), lead (Pb) 17 elements; apply Moran's I index, Getis-Ord Gi* hotspot analysis of the global spatial autocorrelation of chemical elements in drinking water and the degree of aggregation of each element on the local area, and correlation analysis with the prevalence of dental fluorosis in the region. Results: Except for Cu, Zn, and Cd, global spatial autocorrelation Moran's I was negative, and all other elements were positive. F, Ca, Al, Ti, As, Mo, Cd, and Cu elements showed high values of aggregation in the southeastern low-altitude area; Mg, Ba, Pb, Cr, Mn, and Fe elements were mainly aggregated in the central altitude terrain transition area, Zn and Se elements in water sources are significantly positively correlated with the prevalence of dental fluorosis (P<0.05). In contrast, F, Mg, Al, Ti, As, Mo, Cd, Ba, and Pb elements negatively correlate (P<0.05). Elements in the central region were high-high aggregation, as a hot spot aggregation area with high disease incidence, while F, Al, Mn, Mo, Cd, and Ba elements in the western region were low-low aggregation, as a cold spot aggregation area with a low incidence of fluorosis. Conclusions: The risk of population fluoride exposure in surface drinking water sources is shallow. However, the chemical element content of drinking water sources in coal-fired polluted endemic fluorosis areas has prominent spatial geographical distribution characteristics. There is a significant spatial aggregation effect with the prevalence of dental fluorosis, which may play a synergistic or antagonistic effect on the occurrence and prevalence of dental fluorosis.

[Preliminary experience of gasless transoral vestibular robotic thyroidectomy]

Objective: To explore the feasibility and safety of the gasless transoral vestibular robotic thyroidectomy using skin suspension. Methods: The clinical data of 20 patients underwent gasless transoral vestibular robotic thyroidectomy in the Department of Otorhinolaryngology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University from February 2022 to May 2022 were retrospectively analyzed. Among them, 18 were females and 2 were males, aged (38.7±8.0) years old. The intraoperative blood loss, operation time, postoperative hospital stay, postoperative drainage volume, postoperative pain visual analogue scale (VAS) score, postoperative swallowing function swallowing impairment score-6 (SIS-6), postoperative aesthetic VAS score, postoperative voice handicap index-10 (VHI-10) voice quality, postoperative pathology and complications were recorded. SPSS 25.0 was used for statistical analysis of the data. Results: The operations were successfully completed without conversion to open surgery in all patients. Pathological examination showed papillary thyroid carcinoma in 18 cases, retrosternal nodular goiter in 1 case, and cystic change in goiter in 1 case. The operative time for thyroid cancer was 161.50 (152.75, 182.50) min [M (P₂₅, P₇₅), the same below] and the average operative time for benign thyroid diseases was 166.50 minutes. The intraoperative blood loss 25.00 (21.25, 30.00) ml. In 18 cases of thyroid cancer, the mean diameter of the tumors was (7.22±2.02) mm, and lymph nodes (6.56±2.14) were dissected in the central region, with a lymph node metastasis rate of 61.11%. The postoperative pain VAS score was 3.00 (2.25, 4.00) points at 24 hours, the mean postoperative drainage volume was (118.35±24.32) ml, the postoperative hospital stay was 3.00 (3.00, 3.75) days, the postoperative SIS-6 score was (4.90±1.58) points at 3 months, and the postoperative VHI-10 score was 7.50 (2.00, 11.00) points at 3 months. Seven patients had mild mandibular numbness, 10 patients had mild cervical numbness, and 3 patients had temporary hypothyroidism three months after surgery and 1 patient had skin flap burn, but recovered one month after surgery. All patients were satisfied with the postoperative aesthetic effects, and the postoperative aesthetic VAS score was 10.00 (10.00, 10.00). Conclusion: Gasless transoral vestibular robotic thyroidectomy using skin suspension is a safe and feasible option with good postoperative aesthetic effect, which can provide a new treatment option for some selected patients with thyroid tumors.

[Endoscopic ultrasonographic features of submucosal lesions of upper digestive tract suspected gastrointestinal stromal tumors and their correlation with progression and pathological risk grade of the lesions]

Objective: To investigate the endoscopic ultrasonographic (EUS) characteristics of submucosal lesions of upper digestive tract suspected gastrointestinal stromal tumors (GIST) and their correlation with biological behaviors and pathological risk grade of the tumors. Methods: Retrospective cohort study. The EUS findings, follow-up review, surgical treatment and pathological data of patients with suspected GIST at the Gastrointestinal Endoscopy Center of Peking University People's Hospital from January 2013 to April 2021 were collected. All samples were divided into follow-up group and treatment group based on the pathological condition and the patient's treatment intention. According to whether or not the tumor was enlarged in EUS, the follow-up group was divided into non-enlarged group and enlarged group. Paired T-test was used to compare the lesion size before and after follow-up, and logistic regression was used to analyze the risk factors of tumor enlargement. According to the treatment methods, the treatment group was further divided into endoscopic treatment group and surgical treatment group. According to the pathological results and risk grade, the treatment group was further divided into the low-risk group and the medium-risk group. The risk factors of pathological malignant risk were analyzed by logistic regression, and the tumor diameter of patients with moderate or above pathological risk was predicted by receiver operation characteristic (ROC) curve. The relationship between the findings of EUS and the progression and pathological risk of GIST were also explored. Results: Seventy-three cases including 23 males and 50 females, with an age of 58 (30-88) years, were included in the follow-up group, with a mean lesion diameter of (1.21±0.49) cm before follow-up, median follow-up interval of 33.8 months, and a lesion diameter of (1.18±0.49) cm after follow-up. There was no significant difference (all P>0.05) in lesion diameter between before and after follow-up. There was no significant difference (all P>0.05) between tumor enlargement group (18 cases, 24.7%) and non-enlargement group (55 cases, 75.3%). One hundred and thirty-eight cases, including 52 males and 86 females, with an age of 60 (19-84) years, were enrolled in the treatment group, with a mean EUS estimated diameter of (2.55±1.35) cm and pathological diameters of (3.43±2.42) cm. Ninety-five (68.8%) of these cases were pathologically confirmed as GIST while 43 cases were diagnosed as other tumor types, including 37 benign tumors and 6 malignant tumors. In multifactorial logistic regression analysis, only the increase of tumor diameter [OR (95%CI): 1.800 (1.172-2.766), P=0.007] was a risk factor for pathological intermediate or higher risk. The optimal tumor diameter for predicting pathological intermediate or higher risk using ROC curve analysis was 2.75 cm, with a sensitivity 71.4%, specificity 79.0%, Youden index 0.5 and area under ROC curve 0.807 (95%CI: 0.703-0.909). Conclusions: EUS is essential for assessing the risk of progression and malignancy of submucosal lesions of upper digestive tract suspected GIST. For lesions of small diameter, the interval of follow-up shall be relatively extended while the indication of treatment could be partially waived.

The causal mutation in ARR3 gene for high myopia and progressive color vision defect

The ARR3 gene, also known as cone arrestin, belongs to the arrestin family and is expressed in cone cells, inactivating phosphorylated-opsins and preventing cone signals. Variants of ARR3 reportedly cause X-linked dominant female-limited early-onset (age < 7 years old) high myopia (< - 6D). Here, we reveal a new mutation (c.228T>A, p.Tyr76*) in ARR3 gene that can cause early-onset high myopia (eoHM) limited to female carriers. Protan/deutan color vision defects were also found in family members, affecting both genders. Using ten years of clinical follow-up data, we identified gradually worsening cone dysfunction/color vision as a key feature among affected individuals. We present a hypothesis that higher visual contrast due to the mosaic of mutated ARR3 expression in cones contributes to the development of myopia in female carriers.

[Transient elastography technology role assessment during the diagnosis and treatment of patients with chronic hepatitis B virus infection]

Objective: To explore the role of transient elastography technology in the assessment of disease staging and treatment in patients with chronic hepatitis B virus (HBV) infection. Methods: Patients who were clinically diagnosed with chronic HBV infection at Beijing Tsinghua Changgung Hospital from January 2018 to December 2021 was collected. Liver stiffness measurement (LSM) examination was performed more than once by transient elastography. The count data were expressed as cases (%) and the χ (2) test was made. Fisher's exact test was used with theoretical frequency less than 5. The measurement data between two groups was compared by t-test. Multiple groups were compared with an analysis of variance. Results: 1 055 patients were included in this study, including 669 (63.4%) males and 386 (36.6%) females. 757 (71.8%) patients were untreated. Among the untreated patients, the LSM value in the immune clearance (10.2 ± 3.8) kPa (187 cases, 40.4%), and the reactivation stages (9.1 ± 3.4) kPa (114 cases, 24.6%) was significantly higher than that in the immune tolerance (8.7 ± 3.6) kPa (78 cases, 16.8%) and immune control stages (8.4 ± 3.5) KPa (84 cases, 18.1%), and the difference between the four groups was statistically significant (F = 5.31 and P = 0.03). With ALT (male: 30 U/L, female: 19 U/L) as defined the normal value, the LSM value in the immune tolerance and the immune control stages were (5.8 ± 0.9) kPa and (7.1 ± 2.5) kPa, respectively, which were significantly lower than those of patients in the immune tolerance and immune control stages, and the difference was statistically significant (P < 0.01). There were 294 (38.8%) patients with uncertain period, excluding patients with fatty liver. Patients with uncertain periods were divided into four gray zone (GZ) groups: immune tolerance stage: LSM (5.1 ± 1.3) kPa was significantly lower than GZ-A (6.5 ± 2.4) kPa, t = 2.06, P = 0.03, and the difference was statistically significant; immune control stage: LSM was (5.6 ± 1.5) kPa, which was also lower than GZ-C (6.8 ± 1.3) kPa, t = 3.08, P = 0.02, and the difference was statistically significant; immune clearance stage: LSM > 8.0 kPa. LSM values showed a year-by-year reduction in patients with expanded indications who started antiviral treatment and were followed up for three years. Conclusion: The LSM value is significantly lower after the decrease of the defined high-normal ALT value in patients with the immune tolerance and immune control stages of chronic HBV infection. The LSM values of GZ-A and GZ-C in the uncertain periods of chronic HBV infection are higher than those of patients in the immune tolerance and immune control stages.

Insight on genetic features prevalent in five Ipomoea species using comparative codon pattern analysis reveals differences in major codons and reduced GC content at the 5’ end of CDS

Ipomoea plants possess important commercial, medicinal, and ornamental value. Molecular and morphological studies have confirmed that most species of this genus exhibit similar phenotypes but complex phylogenetic relationships. To date, limited information is available on these evolutionary relationships. In this study, systematic analysis of diverse species from Ipomoea was used to elucidate the relationships in this genus. To this end, we employed the concept of codon usage bias (CUB) to analyze the codon usage bias of five Ipomoea species such as effective number of codons (ENC) and GC content at the third synonym codon position (GC3s). Three types of plots including ENC-GC3s, parity rule 2 (PR2) and neutrality plots were employed to discover the factors determining CUB, and the frequency of hydrogen bonds and nucleotide were calculated to dissect changes in GC content at the 5'-end of the coding sequence. Our results showed little distinctness in CUB among the five species, with a reduction of hydrogen bonds content at the 5'-end (with similar changes in cytosines). In addition, optimal codons of Ipomoea aquatica ended with G or C, different from those of the other four species, which ended in A or T. These results may be useful for exploring the evolutionary relationships among this group, and for understanding the reasons for the variation among Ipomoea species.

A Real-Time Image-Based Efferocytosis Assay for the Discovery of Functionally Inhibitory Anti-MerTK Antibodies

Efferocytosis is a phagocytic process by which apoptotic cells are cleared by professional and nonprofessional phagocytic cells. In tumors, efferocytosis of apoptotic cancer cells by tumor-associated macrophages prevents Ag presentation and suppresses the host immune response against the tumor. Therefore, reactivating the immune response by blockade of tumor-associated macrophage-mediated efferocytosis is an attractive strategy for cancer immunotherapy. Even though several methods have been developed to monitor efferocytosis, an automated and high-throughput quantitative assay should offer highly desirable advantages for drug discovery. In this study, we describe a real-time efferocytosis assay with an imaging system for live-cell analysis. Using this assay, we successfully discovered potent anti-MerTK Abs that block tumor-associated macrophage-mediated efferocytosis in mice. Furthermore, we used primary human and cynomolgus monkey macrophages to identify and characterize anti-MerTK Abs for potential clinical development. By studying the phagocytic activities of different types of macrophages, we demonstrated that our efferocytosis assay is robust for screening and characterization of drug candidates that inhibit unwanted efferocytosis. Moreover, our assay is also applicable to investigating the kinetics and molecular mechanisms of efferocytosis/phagocytosis.

Social Link Inference via Multiview Matching Network From Spatiotemporal Trajectories

In this article, we investigate the problem of social link inference in a target location-aware social network (LSN), which aims at predicting the unobserved links between users within the network. This problem is critical for downstream applications, including network completion and friend recommendation. In addition to the network structures commonly used in general link prediction, the studies tailored for social link inference in an LSN leverage user trajectories from the spatial aspect. However, the temporal factor lying in user trajectories is largely overlooked by most of the prior studies, limiting the capabilities of capturing the temporal relevance between users. Moreover, effective user matching by fusing different views, i.e., social, spatial, and temporal factors, remains unresolved, which hinders the potential improvement of link inference. To this end, this article devises a novel multiview matching network (MVMN) by regarding each of the three factors as one view of any target user pair. MVMN enjoys the flexibility and completeness of modeling each factor by developing its suitable matching module: 1) location matching module; 2) time-series matching module; and 3) relation matching module. Each module learns a view-specific representation for matching, and MVMN fuses them for final link inference. Extensive experiments on two real-world data sets demonstrate the superiority of our approach against several competitive baselines for link prediction and sequence matching, validating the contribution of its key components.

Soil contamination in nearby natural areas mirrors that in urban greenspaces worldwide

Soil contamination is one of the main threats to ecosystem health and sustainability. Yet little is known about the extent to which soil contaminants differ between urban greenspaces and natural ecosystems. Here we show that urban greenspaces and adjacent natural areas (i.e., natural/semi-natural ecosystems) shared similar levels of multiple soil contaminants (metal(loid)s, pesticides, microplastics, and antibiotic resistance genes) across the globe. We reveal that human influence explained many forms of soil contamination worldwide. Socio-economic factors were integral to explaining the occurrence of soil contaminants worldwide. We further show that increased levels of multiple soil contaminants were linked with changes in microbial traits including genes associated with environmental stress resistance, nutrient cycling, and pathogenesis. Taken together, our work demonstrates that human-driven soil contamination in nearby natural areas mirrors that in urban greenspaces globally, and highlights that soil contaminants have the potential to cause dire consequences for ecosystem sustainability and human wellbeing.

[Study on protective effects and mechanisms of total glucosides of Paeony on Graves disease in BALB/c mice]

Objective: To investigate the protective effect and its immunoregulatory mechanism of Total Glucosides of Paeony (TGP) against Graves' Disease (GD) model on BALB/c mice. Methods: Fifty female (6 weeks old, weighing 16-18 g) BALB/c mice of specific pathogen free were divided into control group according to random number table method, model group, early low-dose TGP intervention group (250 mg·kg^-1·d^-1), early high-dose TGP intervention group (500 mg·kg^-1·d^-1), and late TGP intervention group, with 10 mice in each group. Except the control group, the other 4 groups were immunized 3 times (0, 3^rd, and 6^th week) with recombinant adenovirus expressing the thyroid stimulating hormone receptor (TSHR) A subunit to establish the GD model. The early low-dose and high-dose intervention group were given diets containing different doses of TGP throughout the whole process, and the late intervention group was given diets containing low doses of TGP from the 1^st week after the 2^nd immunization (week 4). The levels of thyrotropin receptor antibody (TRAb) and total thyroxine (TT4) were detected in the tail venous blood of mice at the 4^th week. At the 10^th week, the serum TRAb and TT4 levels and the ratio of regulatory T cells (Treg) in each group were detected, and the pathological changes of thyroid tissue were observed. Serum helper T cell 1(Th1) and Th2 cell-related factors interleukin-2 (IL-2), IL-4, IL-5, IL-10, IL-12p70, granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon-γ (IFN-γ) and tumor necrosis factors-α (TNF-α) were detected to investigate the protective effect of TGP on GD model in BALB/c mice and its mechanism. Results: At the 4^th week, The level of TT4 [(55.07±12.89) μg/L] in early high-dose intervention group was lower than that in model group [(74.33±8.63) μg/L] (all P<0.05). The level of TT4 in early low-dose intervention group and late intervention group and model group had no statistical significance (all P>0.05). TRAb level of mice between early low-dose, early high-dose, late intervention groups and model group was no significant difference (all P>0.05). At the 10^th week, TRAb [(90.00±26.89) U/L] and TT4[(32.66±8.11) μg/L] levels in the early high-dose intervention group were lower than those in the model group [(396.97±95.35) U/L, (73.70±16.33) μg/L] (all P<0.05). The TRAb and TT4 levels in the early low-dose intervention group and late intervention group were not significantly different from those in the model group (all P>0.05). The thyroid tissue of hyperthyroidism mice in the early high dose intervention group showed focal hypertrophic changes, while the thyroid tissue of other hyperthyroidism mice showed diffuse hypertrophic changes. The CD4⁺CD25⁺/CD4⁺Treg ratio in early high-dose intervention group was higher than that in model group at the 10^th week (4 weeks after three recombinant adenovirus immunization) (P<0.05). Compared with the model group at the 10^th week, the levels of IL-2, IL-12p70 and IFN-γ in the early high-dose intervention group were all decreased (all P<0.05), and the levels of IL-10 were increased (P<0.05). Conclusion: Early high-dose (500 mg·kg^-1·d^-1) TGP intervention group displays a protective effect against GD mice, the mechanism of which may be related to regulatory T cell function changes and Th1/Th2 cytokine balance restoration.

The Gap in the Thickness: Estimating Effectiveness of Pulmonary Nodule Detection in Thick- and Thin-Section CT Images with 3D Deep Neural Networks

BACKGROUND AND OBJECTIVES

There is a noticeable gap in diagnostic evidence strength between the thick and thin scans of Low-Dose CT (LDCT) for pulmonary nodule detection. When the thin scans are needed is unknown, especially when aided with an artificial intelligence nodule detection system.

METHODS

A case study is conducted with a set of 1,000 pulmonary nodule screening LDCT scans with both thick (5.0mm), and thin (1.0mm) section scans available. Pulmonary nodule detection is performed by human and artificial intelligence models for nodule detection developed using 3D convolutional neural networks (CNNs). The intra-sample consistency is evaluated with thick and thin scans, for both clinical doctor and NN (neural network) models. Free receiver operating characteristic (FROC) is used to measure the accuracy of humans and NNs.

RESULTS

Trained NNs outperform humans with small nodules < 6.0mm, which is a good complement to human ability. For nodules > 6.0mm, human and NNs perform similarly while human takes a fractional advantage. By allowing a few more FPs, a significant sensitivity improvement can be achieved with NNs.

CONCLUSIONS

There is a performance gap between the thick and thin scans for pulmonary nodule detection regarding both false negatives and false positives. NNs can help reduce false negatives when the nodules are small and trade off the false negatives for sensitivity. A combination of human and trained NNs is a promising way to achieve a fast and accurate diagnosis.

Fine-Grained Interaction Modeling with Multi-Relational Transformer for Knowledge Tracing

Knowledge tracing, the goal of which is predicting students’ future performance given their past question response sequences to trace their knowledge states, is pivotal for computer-aided education and intelligent tutoring systems. Although many technical efforts have been devoted to modeling students based on their question-response sequences, fine-grained interaction modeling between question-response pairs within each sequence is underexplored. This causes question-response representations less contextualized and further limits student modeling. To address this issue, we first conduct a data analysis and reveal the existence of complex cross effects between different question-response pairs within a sequence. Consequently, we propose MRT-KT, a multi-relational transformer for knowledge tracing, to enable fine-grained interaction modeling between question-response pairs. It introduces a novel relation encoding scheme based on knowledge concepts and student performance. Comprehensive experimental results show that MRT-KT outperforms state-of-the-art knowledge tracing methods on four widely-used datasets, validating the effectiveness of considering fine-grained interaction for knowledge tracing.

The Arabidopsis thaliana trehalose-6-phosphate phosphatase gene AtTPPI regulates primary root growth and lateral root elongation

Roots are the main organs through which plants absorb water and nutrients. As the key phytohormone involved in root growth, auxin functions in plant environmental responses by modulating auxin synthesis, distribution and polar transport. The Arabidopsis thaliana trehalose-6-phosphate phosphatase gene AtTPPI can improve root architecture, and tppi1 mutants have significantly shortened primary roots. However, the mechanism underlying the short roots of the tppi1 mutant and the upstream signaling pathway and downstream genes regulated by AtTPPI are unclear. Here, we demonstrated that the AtTPPI gene could promote auxin accumulation in AtTPPI-overexpressing plants. By comparing the transcriptomic data of tppi1 and wild-type roots, we found several upregulations of auxin-related genes, including GH3.3, GH3.9 and GH3.12, may play an important role in the AtTPPI gene-mediated auxin transport signaling pathway, ultimately leading to changes in auxin content and primary root length. Moreover, increased AtTPPI expression can regulate primary root growth and lateral root elongation under different concentration of nitrate conditions. Overall, constitutive expression of AtTPPI increased auxin contents and improved lateral root elongation, constituting a new method for improving the nitrogen utilization efficiency of plants.

Alternating Early Afterdepolarizations Underlying Bradycardia-Dependent Macroscopic T Wave and Discordant Mechanical Alternans

BACKGROUND

Macroscopic T wave alternans (macro-TWA) often heralds the onset of Torsades de Pointes in patients with QT prolongation. However, the mechanisms underlying macro-TWA remain unclear. We examined the cellular and ionic basis for macro-TWA in rabbits with left ventricular hypertrophy (LVH).

METHODS

The renovascular hypertension model was used to induce LVH in rabbits. Action potentials were simultaneously recorded from epicardium and endocardium together with a transmural ECG and isometric contractility in arterially perfused left ventricular wedges. Late sodium current (I_Na-L) was recorded in single-isolated left ventricular myocytes with the whole cell patch-clamp technique.

RESULTS

Macro-TWA and accompanied mechanical alternans occurred spontaneously in 8 of 33 LVH rabbits (P<0.05, versus 0/15 in controls) and were induced by an I_Na-L enhancer ATX-II at 1 to 3 nM in additional 7. Macro-TWA and mechanical alternans occurred discordantly, that is, that longer QT interval and larger T wave were associated with weaker isometric contvractility. Alternating early afterdepolarizations in the endocardium caused macro-TWA in 12 of 15 LVH rabbits and, therefore, early afterdepolarization-dependent R-from-T extrasystoles and Torsades de Pointes always originated from the beats with longer QT and larger T wave during macro-TWA. I_Na-L density was significantly larger in LVH myocytes than that of control myocytes. Macro-TWA, mechanical alternans, R-from-T extrasystoles, and Torsades de Pointes were all abolished by I_Na-L blocker ranolazine or mexiletine.

CONCLUSIONS

LVH enhances I_Na-L density and promotes alternating early afterdepolarizations in the left ventricular endocardium that manifest as macro-TWA with discordant mechanical alternans. I_Na-L blockade abolishes macro-TWA, mechanical alternans, early afterdepolarization-dependent R-from-T extrasystoles, and Torsades de Pointes.

Mixture 2D Convolutions for 3D Medical Image Segmentation

Three-dimensional (3D) medical image segmentation plays a crucial role in medical care applications. Although various two-dimensional (2D) and 3D neural network models have been applied to 3D medical image segmentation and achieved impressive results, a trade-off remains between efficiency and accuracy. To address this issue, a novel mixture convolutional network (MixConvNet) is proposed, in which traditional 2D/3D convolutional blocks are replaced with novel MixConv blocks. In the MixConv block, 3D convolution is decomposed into a mixture of 2D convolutions from different views. Therefore, the MixConv block fully utilizes the advantages of 2D convolution and maintains the learning ability of 3D convolution. It acts as 3D convolutions and thus can process volumetric input directly and learn intra-slice features, which are absent in the traditional 2D convolutional block. By contrast, the proposed MixConv block only contains 2D convolutions; hence, it has significantly fewer trainable parameters and less computation budget than a block containing 3D convolutions. Furthermore, the proposed MixConvNet is pre-trained with small input patches and fine-tuned with large input patches to improve segmentation performance further. In experiments on the Decathlon Heart dataset and Sliver07 dataset, the proposed MixConvNet outperformed the state-of-the-art methods such as UNet3D, VNet, and nnUnet.