Sequential Strategy Research Articles

Bayesian optimization for personalized dose-finding trials with combination therapies

Abstract Identification of optimal dose combinations in early-phase dose-finding trials is challenging, due to the trade-off between precisely estimating the many parameters required to flexibly model the possibly nonmonotonic dose-response surface, and the small sample sizes in early-phase trials. This difficulty is even more pertinent in the context of personalized dose-finding, where patient characteristics are used to identify tailored optimal dose combinations. To overcome these challenges, we propose the use of Bayesian optimization for finding optimal dose combinations in standard (one size fits all) and personalized multi-agent dose-finding trials. Bayesian optimization is a method for estimating the global optima of expensive-to-evaluate objective functions. The objective function is approximated by a surrogate model, commonly a Gaussian process, paired with a sequential design strategy to select the next point via an acquisition function. This work is motivated by an industry-sponsored problem, where the focus is on optimizing a dual-agent therapy in a setting featuring minimal toxicity. To compare the performance of the standard and personalized methods under this setting, simulation studies are performed for a variety of scenarios. Our study concludes that taking a personalized approach is highly beneficial in the presence of heterogeneity.

Time-lapse 4D full-waveform inversion for ocean-bottom cable seismic data with seawater velocity changes

Time-lapse (4D) full-waveform inversion (FWI) is a valuable technology for high-resolution imaging of reservoir changes caused by hydrocarbon production and CO$_2$ storage. However, it still faces challenges in dealing with non-repeatability issues due to changes in seawater or near-surface velocity between baseline and monitor surveys. Despite recent advances, the ability of 4D FWI to address this problem has rarely been demonstrated. {\color{red}We} investigate the effectiveness of current 4D FWI strategies, including parallel, double-difference, sequential, and common-model strategies, in resolving non-repeatability issues for 4D OBC (ocean-bottom cable) seismic data. Additionally, a three-stage 4D FWI strategy is developed for 4D OBC seismic data, involving the estimation of seawater velocities in the baseline and monitor models, obtaining a good common starting model, and final convergence to obtain subsurface 4D changes. The synthetic data tests conducted with varying levels of seawater velocity changes indicate that among the current strategies tested, the common-model strategy performs the best, and that our new strategy can further enhance the 4D inversion results.

Paired Electro‐Synthesis of Remote Amino Alcohols with/in H2O

Amino alcohols, particularly remote amino alcohols and peptide alcohols, are valuable due to their functional diversity in biologically active compounds. However, traditional synthesis methods face significant challenges, making electrochemistry an attractive alternative. We have developed a mild and biocompatible sequential paired electrolysis strategy, leveraging copper‐electrocatalysis to synthesize diverse remote amino alcohols, including unnatural peptide alcohols. Both experimental results and density functional theory (DFT) calculations demonstrate that water serves as both the hydroxyl source and the solvent, facilitating the generation of CuH with Cu(I) at the cathode, which in turn reduces the aldehyde intermediates formed during the reaction.

PairedElectro-Synthesis of Remote Amino Alcohols with/in H2O.

Amino alcohols, particularly remote amino alcohols and peptide alcohols, are valuable due to their functional diversity in biologically active compounds. However, traditional synthesis methods face significant challenges, making electrochemistry an attractive alternative. We have developed a mild and biocompatible sequential paired electrolysis strategy, leveraging copper-electrocatalysis to synthesize diverse remote amino alcohols, including unnatural peptide alcohols. Both experimental results and density functional theory (DFT) calculations demonstrate that water serves as both the hydroxyl source and the solvent, facilitating the generation of CuH with Cu(I) at the cathode, which in turn reduces the aldehyde intermediates formed during the reaction.

OpenPrep: Sensor Event-driven Open Column Chromatographic Sample Preparation System.

We propose a sensor event-driven, open-column chromatographic sample preparation system─OpenPrep. This system replaces the problematic flow selector valve with a noncontact built-in sensor, eliminating carryover and clogging issues, which are commonly encountered in pump-driven chromatographic systems. The innovative valveless column-flow design, combined with a compact gantry dispenser and motion stage-based fraction collector, reduces the sample flow path to a disposable component with a postcolumn dead volume of only 0.5 mL. This configuration is superior to conventional dynamic flow systems owing to its ability to minimize sample consumption, eluent loss, and cleaning waste generation. Thus, OpenPrep can be ideal for preventing cross-contamination. Owing to the rapid flow-transfer and stop approach, OpenPrep achieves higher reagent delivery accuracy and precision compared to ISO 8655-2 piston-operated pipettes, by minimizing liquid residue on the dispensing tip. Sensor-driven flow control ensures precise eluent collection and automatic regulation of the mobile-phase flow rate through an online calibration algorithm, meeting all the requirements for accurate and reproducible column flow control. OpenPrep features an open architecture for both hardware and software, offering high flexibility for analyzing samples with considerable variations in radioactivity and matrix compositions, such as radioactive waste. The proof-of-concept implementation of a multistage sequential separation strategy, using five resin columns to separate Re, Sr, Fe, and Ni from a single sample, demonstrates that OpenPrep provides enhanced flexibility and reconfigurability, improved operational efficiency, and cost-effective customization. Excellent chemical yields were obtained for Re, Sr, and Ni in the range of 83-97%, with less than 2.6% relative standard deviation at a flow rate three times faster than the gravity method. Additionally, improved recovery performance of Fe was achieved by applying a slower flow rate compared to gravity flow.

Inhibiting Mitochondrial Damage for Efficient Treatment of Cerebral Ischemia-Reperfusion Injury Through Sequential Targeting Nanomedicine of Neuronal Mitochondria in Affected Brain Tissue.

Oxidative stress, predominantly from neuronal mitochondrial damage and the resultant cytokine storm, is central to cerebral ischemia-reperfusion injury (CIRI). However, delivering drugs to neuronal mitochondria remains challenging due to the blood-brain barrier (BBB), which impedes drug entry into affected brain tissues. This study introduces an innovative tannic acid (TA) and melanin-modified heteropolyacid nanomedicine (MHT), which highly specifically eliminates the neuronal mitochondrial reactive oxygen radicals burst to efficiently reduce neuronal mitochondrial damage through a strategically designed sequential targeting strategy from affected brain tissue to neuronal mitochondria. TA endows MHT with sequential targeting ability by binding to matrix proteins exposed to the damaged BBB and mitochondrial outer membrane proteins of neurons, while melanin significantly enhances the antioxidant capacity of MHT. Consequently, MHT effectively inhibits neuronal apoptosis by protecting mitochondria and reversing the inflammatory immune environment through the deactivation of the cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) pathway. MHT demonstrated a strong therapeutic effect on CIRI, with an ultralow dose (2 mgkg-1) proving effective in reversing the condition. This work not only introduces a new avenue to significantly reduce CIRI through sequential targeting therapy but also offers a new paradigm for treating other ischemia-reperfusion injury diseases.

A sequential treatment strategy by copper ion-doped nanofiber dressing for highly efficient biofilm combating and rapid wound healing

Copper ions are considered as a promising alternative to combat infected wound and accelerate wound healing. However, exhibiting good antibiofilm activity meanwhile minimizing cell toxicity remains a great challenge. To address this issue, we provide a sequential treatment strategy to combat biofilm infected wound with low copper ion (Cu2+) release (as low as 1.2 μg·mL−1). Polydopamine (PDA) adherent copper ion complex is constructed on thermoplastic polyurethane (TPU) to develop the Cu2+@PTPU nanofibers. With the application of Cu2+@PTPU wound dressing, nitric oxide (NO) gas is released by catalyzing endogenous donor, thereby rapidly eliminating more than 90 % biofilm biomass. Subsequent moderate photothermal treatment (∼47℃) significantly enhanced bactericidal effect by deconstructing bacteria membrane. As a result, the biofilm infection was eradicated in only ∼ 6h, significantly more effective than existing methods (∼24 h). Additionally, Cu2+@PTPU dressing not only exhibits excellent cell compatibility, but also promotes wound healing. In vivo study demonstrates that it reduces the level of inflammatory responses by 104 % and increases angiogenesis capacity by 115 %, compared with untreated group, thereby narrowing the wound site area by 72 % at day 14 after treatment. Overall, the Cu2+@PTPU dressing and sequential treatment strategy provide quick, safe, yet efficacious biofilm combating outcomes and facilitates rapid wound healing.

Romosozumab following denosumab improves lumbar spine bone mineral density and trabecular bone score greater than denosumab continuation in postmenopausal women.

Romosozumab following anti-resorptive can be an effective sequential treatment strategy to improve bone strength. However, whether the transition to romosozumab after denosumab is associated with greater improvement in bone mineral density (BMD) and trabecular bone score (TBS) compared to denosumab continuation remains unclear. In this propensity score-matched cohort study, we analyzed data from postmenopausal women who initiated denosumab between 2017 and 2020. Individuals who were transited to 12months of romosozumab after denosumab were 1:1 matched to those who continued an additional 12months of denosumab (n = 86 for each group; denosumab-romosozumab [DR] and denosumab-denosumab [DD]). Mean BMD gain by denosumab treatment in matched DR and DD groups from denosumab initiation to transition (median 4 times [range 2 to 8]) was +4.8% and + 2.0% in the lumbar spine and total hip. DR group showed greater LS BMD gain compared to the DD group (+6.8 vs. +3.3% point, P<.001) for 12months post-transition independent of the duration of prior denosumab treatment, yielding greater overall LS BMD gain in DR compared to DD (+11.6% vs. +8.0%, P<.001). DD group showed continued improvement of hip BMD, whereas hip BMD was maintained but not improved in the DR group. DR group was associated with greater TBS improvement than the DD group (2.9% vs 1.0%, P=.042). One month after the transition to romosozumab from denosumab, P1NP immediately increased above the level of denosumab initiation with relatively suppressed CTx, creating a transient anabolic window. For 12months follow-up, one incident morphometric vertebral fracture and one patella fracture were observed in DD, whereas one ankle fracture was observed in the DR group. Romosozumab following denosumab improved lumbar spine BMD and TBS greater than denosumab continuation in postmenopausal women.

From inbreeding to social behaviour: exploring female-biased sex allocation in the silk wasp Microstigmus rosae

In the Hymenoptera, female-biased sex ratios are often attributed to Local Mate Competition (LMC), which occurs when related males compete for access to the same mates, reducing fitness returns for male offspring relative to females. In social species, advantageous interactions between cooperating females can also promote female bias via Local Resource Enhancement (LRE). Despite recent theory suggesting potential interactions between these phenomena, studies of LMC in social species remain scarce. This study investigates female-biased sex ratio adjustment in the silk wasp Microstigmus rosae, a species exhibiting both incipient sociality and likely localised mating. Our analyses reveal several genetic indicators of local mating, including heterozygote deficiency and high inbreeding coefficients. Moreover, pairwise relatedness values between mothers and their inferred mates indicate that 35% of mating pairs were siblings or parent-offspring. In further support of the role of LMC, we find that our observed sex investment ratio of 0.35 is statistically equivalent to the theoretically expected evolutionary stable sex ratio, as predicted by partial LMC. We also find a clearly defined male-first sequential sex allocation strategy (M-F-F), a pattern typically attributed to extreme LMC. In contrast, we find no perceivable effect of cooperative behaviour (LRE) on the brood sex ratio. This rare demonstration of inbreeding and LMC in a social aculeate hymenopteran highlights the potential role of these factors in driving the evolution of social group formation in this species.

Sequential bioaugmentation of the dominant microorganisms to improve the fermentation and flavor of cereal vinegar

Traditional cereal vinegars are fermented by microorganisms that are spontaneously enriched, leading to uncertainty in regulating the fermentation process and flavor. The objective of this study was to elucidate the impact of the predominant microorganisms, provenly Lactobacillus helveticus and Acetobacter pasteurianus, on the solid-state fermentation (SSF) and flavor profile of cereal vinegar by several bioaugmentation strategies. The results indicated that the sequential bioaugmentation of predominant microorganisms improved the utilization of raw material and most key flavor compounds. Through sequential bioaugmentation strategy, bacterial diversity was regulated due to minimizing acetic acid inhibition in the early stages, and the non-volatile acid was targetedly improved by Lactobacillus. Furthermore, the important flavor of non-volatile acid, esters, acetoin, and tetramethyl-pyrazine content was enhanced by sequential bioaugmentation. Therefore, the sensory score on taste and odor were improved. These results provide a reference for the targeted regulation of the SSF and the flavor quality of cereal vinegar.

Enhanced biopolymer composites through micro-integrated lignin and anhydride modified nonwoven bast fiber

The growing need for sustainable materials in various industries has driven research towards eco-friendly biopolymer composites. In this study, lignin microfillers were integrated into PHBV/modified bast fiber matrices, offering a promising solution for sustainable material innovation with improved performance and functionality. Nonwoven bast fiber mats were initially treated with propionic, succinic, and maleic anhydrides, evaluated through weight gain analysis, chemical characterization, thermal stability, and wettability assessments. These modified bast fibers were then reinforced with poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) biopolymer at 30 % loading using specialized molding techniques. The resulting biocomposites were subjected to comprehensive characterization, including moisture related properties, mechanical strength, thermal behavior, and fracturographical properties. In a final stage, lignin microparticles (LMPs) as biofillers were integrated at 1 %, 3 %, 5 %, and 7 % concentrations to enhance biocomposite properties. Experimental findings highlighted succinylated bast fibers as offering the greatest improvements in mechanical strength, thermal stability, and morphological integrity among the modified fibers. At an optimal 5 % lignin microfiller content, the biocomposites exhibited superior wettability and improved stiffness. However, excessive lignin loading led to particle agglomeration, diminishing overall performance. Fracturographical analysis provided insights into the fracture behavior of the biocomposites, highlighting the cohesive and interfacial characteristics of the materials under stress conditions. This research highlights the efficacy of sequential modification strategies in developing high-performance sustainable materials tailored for diverse industrial applications.

Submacular hemorrhage: My personal journey to the goal.

Acute submacular hemorrhage (SMH) can be caused by various diseases including age-related macular degeneration (AMD), polypoidal choroidal vasculopathy (PCV), and retinal arterial microaneurysm (RAM). The natural course of submacular hemorrhage is generally poor. Animal studies have suggested that the removal of subretinal hemorrhage may effectively reduce retinal damage caused by hemorrhage in humans and removal of submacular hemorrhage have been performed with limited visual outcomes. Pneumatic displacement involving intravitreal expansile gas with or without adjunctive intravitreal injection of tissue plasminogen activator (tPA) has demonstrated effective displacement of SMH and improvement in visual acuity in the majority of cases. Although tPA may not be indispensable, its use may facilitate displacement. Combining pneumatic displacement with vitrectomy and subretinal injection of tPA may achieve superior displacement of SMH compared to pneumatic displacement of SMH, implying that pneumatic displacement of SMH with vitrectomy and subretinal injection may offer enhanced effectiveness in SMH displacement,while noobvious different was found in visual outcomes between the two treatments.Complications associated with these procedures encompass breakthrough hemorrhage, retinal detachment and macular hole formation. Breakthrough hemorrhage is more commonly observed following pneumatic displacement whereas retinal detachment appears to be more prevalent following vitrectomy. Macular hole formation subsequent to vitrectomy represents a significant complication, particularly in eyes with SMH attributed to ruptured retinal arterial microaneurysm. Both pneumatic displacement and vitrectomy present advantages and disadvantages, and the superiority between the two remains undetermined. Sequential strategy for the treatment of submacular hemorrhage is another option. As the initial step, pneumatic displacement of SMH should be attempted, and if displacement is insufficient, pneumatic displacement following vitrectomy with subretinal injection of tPA may be pursued. Further investigations are warranted to ascertain optimal management strategies for SMH leading to improved outcomes. KEY MESSAGES: What is known • Pneumatic displacement with/without intravitreal tPA injection, and vitrectomy with subretinal tPA injection and gas are the two major treatments for submacular hemorrhage. What is new • No obvious different was found in visual outcomes between vitrectomy, subretinal tPA injection and gas, and intravitreal tPA injection and gas while vitrectomy with subretinal tPA injection and gas may achieve better displacement of submacular hemorrhage. • Macular hole formation is a specific complication for submacular hemorrhage due to ruptured retinal arterial macroaneurysm.

Context-Dependent Criteria for Dirichlet Process in Sequential Decision-Making Problems

In models with insufficient initial information, parameter estimation can be subject to statistical uncertainty, potentially resulting in suboptimal decision-making; however, delaying implementation to gather more information can also incur costs. This paper examines an extension of information-theoretic approaches designed to address this classical dilemma, focusing on balancing the expected profits and the information needed to be obtained about all of the possible outcomes. Initially utilized in binary outcome scenarios, these methods leverage information measures to harmonize competing objectives efficiently. Building upon the foundations laid by existing research, this methodology is expanded to encompass experiments with multiple outcome categories using Dirichlet processes. The core of our approach is centered around weighted entropy measures, particularly in scenarios dictated by Dirichlet distributions, which have not been extensively explored previously. We innovatively adapt the technique initially applied to binary case to Dirichlet distributions/processes. The primary contribution of our work is the formulation of a sequential minimization strategy for the main term of an asymptotic expansion of differential entropy, which scales with sample size, for non-binary outcomes. This paper provides a theoretical grounding, extended empirical applications, and comprehensive proofs, setting a robust framework for further interdisciplinary applications of information-theoretic paradigms in sequential decision-making.

An Enabling Peptide Ligation Induced by Thiol-Salicylaldehyde Ester for Chemical Protein Synthesis.

Chemical protein synthesis by amide-forming ligation of two unprotected peptide segments offers an effective strategy for the preparation of protein derivatives that are not accessible through bioengineering approaches. Herein, an unprecedented chemical ligation between peptides with C-terminal 2-mercaptobenzaldehyde (thiol-salicylaldehyde, TSAL) esters and peptides bearing N-terminal cysteine/penicillamine is reported. Reactive peptide TSAL esters can be obtained from peptide hydrazides in an operationally simple and highly effective manner. This chemoselective peptide ligation enables the rapid production of N,S-benzylidene acetal intermediates, which can readily be converted into native amide bonds even at sterically hindered junctions. In addition, the current method can be applied compatibly in concert with other types of ligations and subsequent desulfurization chemistry, thereby facilitating convergent protein synthesis. The effectiveness of this new method is also showcased by the total synthesis of proteins ubiquitin and hyalomin-3 (Hyal-3), the efficient synthesis of protein ubiquitin-fold modifier 1 (UFM1) via a C-to-N sequential TSAL ester-induced ligation strategy, and the chemical synthesis of protein Mtb CM through a combined strategy of Ser/Thr ligation and TSAL ester-induced ligations.

Optimizing Treatment Strategies for Egfr-Mutated Non-Small-Cell Lung Cancer Treated with Osimertinib: Real-World Outcomes and Insights.

Background: Osimertinib, a third-generation EGFR tyrosine kinase inhibitor (TKI), demonstrated superior efficacy over first-generation TKIs in the FLAURA trial, resulting in its approval as first-line therapy for metastatic non-small-cell lung cancer (NSCLC). However, the real-world application of these trial results requires an evaluation of sequential therapeutic strategies. Methods: This retrospective, non-interventional study utilized data from the Epidemiological Strategy and Medical Economics (ESME) platform, which includes information on patients treated for lung cancer since 2015. Out of 39,974 patients in the database, 624 patients with EGFR-mutant advanced NSCLC treated with osimertinib as first-line (L1, n = 198) or second-line (L2, n = 426) treatment after first- or second-generation TKIs (n = 1262) were identified. Patient demographics, disease characteristics, treatment strategies, and disease progression were examined. Survival analyses were performed using Kaplan-Meier estimates and Cox proportional-hazards models. Results: In the study population (n = 624), 73.4% were female, with a median age of 70 years (range 28-93). Brain metastases at the start of osimertinib treatment were observed in 282 patients. ECOG PS-2 was reported in 29.4% of patients. The T790M mutation in exon 20 was identified in 257/426 patients (60.3%) receiving osimertinib in L2. Median progression-free survival (PFS) was 12.4 months (95% CI [10.7-14.7]) for L1 and 7.4 months (95% CI [6.2-8.7]) for L2. Median overall survival (OS) from advanced diagnosis was 28.5 months (95% CI [26.3-38.7]) for osimertinib L1 and 29.9 months (95% CI [28.6-31.8]) for osimertinib L2 (HR = 0.93; 95% CI [0.75-1.16]; p = 0.50). For L1, median OS was 27.1 months (95% CI [22.0-30.2]) for patients with cerebral metastases and 38.7 months (95% CI [26.3-52.8]) for those without (HR = 0.73; 95% CI [0.48-1.11]; p = 0.15). Discussion: Patients in the real-world ESME database exhibited a poorer prognosis compared to those in the FLAURA trial. The presence of cerebral metastases at diagnosis worsens the prognosis.

Selenium‐Rich Sb2(S,Se)3 Thin Films Deposited via Sequential Chemical Bath Deposition for High‐Efficiency Solar Cells

AbstractRecently, solution‐processed antimony sulfoselenide (Sb2(S,Se)3) thin films and solar cells have gained popularity and achieved impressive results. One of the most effective improvements is the chemical bath deposition (CBD) approach, which is gentle, highly adjustable, and scalable. However, development in alloyed Sb2(S,Se)3 using this process has been modest, with a power conversion efficiency (PCE) of roughly 8%. In this work, a sequential CBD deposition strategy is designed, introducing sulfur (S) or selenium (Se) sources at different stages of the CBD process to achieve high‐quality Sb2(S,Se)3 thin films and devices. Impressively, adding a moderate amount of sodium thiosulfate and thioacetamide as mixed sulfur sources during the CBD deposition of Sb2Se3 can produce Se‐rich Sb2(S,Se)3 films with high crystallinity, suitable bandgaps, low defect density, and favorable energy level alignment. With a short‐circuit current density (JSC) of 26.80 mA cm−2 and a fill factor (FF) of 65.89%, the solar cells with the FTO/CdS/Sb2(S,Se)3/Spiro‐OMeTAD/Au structure obtained an impressive PCE of 9.29%, the highest to date for Sb2(S,Se)3 solar cells manufactured with the CBD process. This study provides a feasible approach for enhancing the performance of CBD‐produced Sb2(S,Se)3 solar cells and offers new insights and techniques for the synthesis of other complex multicomponent thin films.

PrescDRL: deep reinforcement learning for herbal prescription planning in treatment of chronic diseases

Treatment planning for chronic diseases is a critical task in medical artificial intelligence, particularly in traditional Chinese medicine (TCM). However, generating optimized sequential treatment strategies for patients with chronic diseases in different clinical encounters remains a challenging issue that requires further exploration. In this study, we proposed a TCM herbal prescription planning framework based on deep reinforcement learning for chronic disease treatment (PrescDRL). PrescDRL is a sequential herbal prescription optimization model that focuses on long-term effectiveness rather than achieving maximum reward at every step, thereby ensuring better patient outcomes. We constructed a high-quality benchmark dataset for sequential diagnosis and treatment of diabetes and evaluated PrescDRL against this benchmark. Our results showed that PrescDRL achieved a higher curative effect, with the single-step reward improving by 117% and 153% compared to doctors. Furthermore, PrescDRL outperformed the benchmark in prescription prediction, with precision improving by 40.5% and recall improving by 63%. Overall, our study demonstrates the potential of using artificial intelligence to improve clinical intelligent diagnosis and treatment in TCM.

Sequentially Activated-Dumbbell DNA Nanodevices for Accurate Detection of Uracil-DNA Glycosylase via PER-Based Orthogonal Signal Outputs.

Accurate and reliable detection of uracil-DNA glycosylase (UDG) activity is crucial for clinical diagnosis and prognosis assessment. However, current techniques for accurately monitoring UDG activity still face significant challenges due to the single input or output signal modes. Here, we develop a sequentially activated-dumbbell DNA nanodevice (SEAD) that enables precise and reliable evaluation of UDG activity through primer exchange reactions (PER)-based orthogonal signal output. The SEAD incorporates a double-hairpin structure with a stem containing two deoxyuridine (dU) sites for target recognition and two preblocked primer binding regions for target amplification and signal output. Upon UDG recognition of dU, the SEAD can be cleaved by apurinic/apyrimidinic endonuclease 1 (APE1), generating two different hairpins with exposed primer binding regions. These hairpins serve as templates to initiate the parallel PER, enabling the extending of two different amplification products: a long single-stranded DNA (ssDNA) with repetitive sequences and a short ferrocene-labeled ssDNA with complementary sequences. These products further self-assemble into DNA nano-strings in an orthogonal manner that act as an electrochemiluminescence signal switch, enabling precise detection of low-abundance UDG. This work develops a sequential input and orthogonal output strategy for accurately monitoring UDG activity, highlighting the significant potential in cancer diagnosis and treatment.

Effectiveness of Risk-Adapted Upper Gastrointestinal Cancer Screening in China: Prospective Cohort Study.

Previous studies have proved the effectiveness of endoscopic screening in rural areas; however, long-term, high-quality evidence regarding the effectiveness of risk-adapted upper gastrointestinal cancer (UGC) sequential screening strategies in resource-rich regions is currently lacking. The objectives were to validate the effectiveness of risk-adapted sequential screening strategies in UGC prevention and control and assess the potential of sequential screening to lower mortality rates. Based on the Cancer Screening Program in Urban China, a prospective, large-scale cohort study based on population was conducted to recruit individuals from 4 cities in China from 2013-2019. Those identified as having a high risk of UGC according to a validated risk-score model were advised to undergo endoscopy tests. Follow-up outcomes were tracked until June 2021. Incidence of UGC, UGC-related mortality, and all-cause mortality were evaluated between the screened and nonscreened cohorts. The study included 153,079 participants at baseline. In total, 113,916 (74.42%) of the participants were designated as low risk of UGC. The remaining 39,163 (25.68%) participants were deemed to be at high risk of UGC and were offered gastroscopy tests. Among the high-risk participants, 9627 (compliance rate 24.6%) adhered to the gastroscopy tests. Over a median follow-up of 6.05 (IQR 3.06-7.06) years, 622 UGC cases, 180 UGC deaths, and 1958 all-cause death cases were traced. The screened cohort exhibited the highest cumulative incidence of UGC (119.2 per 100,000 person-years), followed by the nonscreened and low-risk cohorts. Obvious reductions in both all-cause mortality and UGC mortality were observed between those who undertook screening (153.7 and 4.7 per 100,000 person-years, respectively) and the nonscreened group (245.3 and 27 per 100,000 person-years, respectively). The screening population showed a significant 36% and 82% reduction in both all-cause mortality (hazard ratio [HR] 0.64, 95% CI 0.49-0.83, P<.001) and UGC mortality (HR 0.18, 95% CI 0.04-0.74, P=.02), respectively, compared to the nonscreened group. Reductions of 35% in all-cause mortality (HR 0.65, 95% CI 0.49-0.86, P=.003) and 81% in UGC mortality (HR 0.19, 95% CI 0.05-0.80, P=.02) were observed in participants aged older than 55 years in the screened group compared to the nonscreened group. The reductions in all-cause mortality and UGC mortality were statistically significant in males (all-cause mortality: HR 0.64, 95% CI 0.47-0.88, P=.005; UGC mortality: HR 0.10, 95% CI 0.01-0.72, P=.02), but significant reductions were not observed in females (all P values were >.05). Our study suggests the significance of one-off risk-adapted UGC screening in reducing both all-cause mortality and UGC mortality, particularly among high-risk individuals, indicating its effectiveness in UGC prevention and management.

A deep-learning-based surrogate modeling method with application to plasma processing

Rapid and accurate system evolution predictions are crucial in scientific and engineering research. However, the complexity of processing systems, involving multiple physical field couplings and slow convergence of iterative numerical algorithms, leads to low computational efficiency. Hence, this paper introduces a systematic deep-learning-based surrogate modeling methodology for multi-physics-coupled process systems with limited data and long-range time evolution, accurately predicting physics dynamics and considerably improving computational efficiency and generalization. The methodology comprises three main components: (1) generating datasets using a sequential sampling strategy, (2) modeling multi-physics spatio-temporal dynamics by designing a heterogeneous Convolutional Autoencoder and Recurrent Neural Network, and (3) training high-precision models with limited data and long-range time evolution via a dual-phase training strategy. A holistic evaluation using a 2D low-temperature plasma processing example demonstrates the methodology’s superior computational efficiency, accuracy, and generalization capabilities. It predicts plasma dynamics approximately 105 times faster than traditional numerical solvers, with a consistent 2% relative error across different generalization tasks. Furthermore, the potential for transferability across various geometries is explored, and the model’s transfer capability is demonstrated with two distinct geometric domain examples.