Search

Background: This study aimed to investigate the superiority of nab -paclitaxel plus S-1 (AS) over oxaliplatin plus S-1 (SOX) in patients with advanced gastric cancer (AGC). Methods: In this multicenter, randomized, phase III superiority trial, eligible patients with unresectable, locally advanced gastric adenocarcinoma were recruited and randomly assigned (1:1) to receive AS ( nab -paclitaxel 260 mg/m 2 on day 1 or 130 mg/m 2 on days 1 and 8; oral S-1 40–60 mg twice daily for 14 days) or SOX (130 mg/m 2 oxaliplatin on day 1; oral S-1 40–60 mg twice daily for 14 days) every 3 weeks for up to six cycles. The primary endpoint was progression-free survival (PFS), and the secondary endpoints were overall survival, objective response rate, and safety. Results: Owing to slow enrolment, an unplanned interim analysis was performed, resulting in the early termination of the study on 31 December 2021 (data cutoff). Between March 2019 and March 2021, 97 patients (AS, n = 48; SOX, n = 49) were treated and evaluated for efficacy and safety of AS and SOX. As of the data cutoff, the median follow-up was 23.13 months [95% confidence interval (CI), 13.39–32.87]. The median PFS was 9.03 months (95% CI, 6.50–11.56) in the AS group and 5.07 months (95% CI, 4.33–5.81) in the SOX group, demonstrating a better PFS tendency following AS treatment than SOX treatment (hazard ratio = 0.59; 95% CI, 0.37–0.94; p = 0.03). The most common grade 3 or worse adverse events were anemia, neutropenia, and leukopenia in both groups, with a higher incidence of thrombocytopenia in the SOX group. Conclusion: Although this study was terminated early, the results demonstrated a better PFS tendency in patients with AGC who were treated with AS than in those treated with SOX, with controllable toxicities. Trial registration: Clinical Trials.gov identifiers: NCT03801668. Registered January 11, 2019.

Based on the idea of maximum determinant positive definite matrix completion, Yamashita proposed a sparse quasi-Newton update, called MCQN, for unconstrained optimization problems with sparse Hessian structures. Such an MCQN update keeps the sparsity structure of the Hessian while relaxing the secant condition. In this paper, we propose an alternative to the MCQN update, in which the quasi-Newton matrix satisfies the secant condition, but does not have the same sparsity structure as the Hessian in general. Our numerical results demonstrate the usefulness of the new MCQN update with the BFGS formula for a collection of test problems. A local and superlinear convergence analysis is also provided for the new MCQN update with the DFP formula.

Linear hyperspectral unmixing (LHU) aims to extract endmembers from mixed pixels in hyperspectral images, which is also a large-scale problem as the number of pixels in hyperspectral images is very large. In this paper, by virtue of the Moreau envelope, we formulate the LHU as an unconstrained optimization problem. Then we adopt the proximal gradient descent method to solve the model, and consider a stochastic version of the method which is for dealing with large-scale scenario. Numerical results are provided to demonstrate the simplicity and efficiency of our proposed model. [ABSTRACT FROM AUTHOR]