Two-Stage Orthogonal Least Squares Methods for Neural Network Construction

A number of neural networks can be formulated as the linear-in-the-parameters models. Training such networks can be transformed to a model selection problem where a compact model is selected from all the candidates using subset selection algorithms. Forward selection methods are popular fast subset selection approaches. However, they may only produce suboptimal models and can be trapped into a local minimum. More recently, a two-stage fast recursive algorithm (TSFRA) combining forward selection and backward model refinement has been proposed to improve the compactness and generalization performance of the model. This paper proposes unified two-stage orthogonal least squares methods instead of the fast recursive-based methods. In contrast to the TSFRA, this paper derives a new simplified relationship between the forward and the backward stages to avoid repetitive computations using the inherent orthogonal properties of the least squares methods. Furthermore, a new term exchanging scheme for backward model refinement is introduced to reduce computational demand. Finally, given the error reduction ratio criterion, effective and efficient forward and backward subset selection procedures are proposed. Extensive examples are presented to demonstrate the improved model compactness constructed by the proposed technique in comparison with some popular methods.

Solubility parameter based screening methods for early stage formulation development of itraconazole amorphous solid dispersions

Objectives: This article uses conventional and newly extended solubility parameter (δ) methods to identify polymeric materials capable of forming amorphous dispersions with itraconazole (itz). Methods: Combinations of itz and Soluplus, Eudragit E PO (EPO), Kollidon 17PF (17PF) or Kollidon VA64 (VA64) were prepared as amorphous solid dispersions using quench cooling and hot melt extrusion. Storage stability was evaluated under a range of conditions using differential scanning calorimetry and powder X-ray diffraction. Key findings: The rank order of itz miscibility with polymers using both conventional and novel δ-based approaches was 17PF > VA64 > Soluplus > EPO, and the application of the Flory–Huggins lattice model to itz–excipient binary systems corroborated the findings. The solid-state characterisation analyses of the formulations manufactured by melt extrusion correlated well with pre-formulation screening. Long-term storage studies showed that the physical stability of 17PF/vitamin E TPGS–itz was poor compared with Soluplus and VA64 formulations, and for EPO/itz systems variation in stability may be observed depending on the preparation method. Conclusion: Results have demonstrated that although δ-based screening may be useful in predicting the initial state of amorphous solid dispersions, assessment of the physical behaviour of the formulations at relevant temperatures may be more appropriate for the successful development of commercially acceptable amorphous drug products.

High-energy monoenergetic proton beams from two stage acceleration with a slow laser pulse

We present a new regime to generate high-energy quasimonoenergetic proton beams in a "slow-pulse" regime, where the laser group velocity vg<c is reduced by an extended near-critical density plasma. In this regime, for properly matched laser intensity and group velocity, ions initially accelerated by the light sail (LS) mode can be further trapped and reflected by the snowplough potential generated by the laser in the near-critical density plasma. These two acceleration stages are connected by the onset of Rayleigh-Taylor-like (RT) instability. The usual ion energy spectrum broadening by RT instability is controlled and high quality proton beams can be generated. It is shown by multidimensional particle-in-cell simulation that quasimonoenergetic proton beams with energy up to hundreds of MeV can be generated at laser intensities of 1021W/cm².

Laser-driven three-stage heavy-ion acceleration from relativistic laser-plasma interaction

A three-stage heavy ion acceleration scheme for generation of high-energy quasimonoenergetic heavy ion beams is investigated using two-dimensional particle-in-cell simulation and analytical modeling. The scheme is based on the interaction of an intense linearly polarized laser pulse with a compound two-layer target (a front heavy ion layer + a second light ion layer). We identify that, under appropriate conditions, the heavy ions preaccelerated by a two-stage acceleration process in the front layer can be injected into the light ion shock wave in the second layer for a further third-stage acceleration. These injected heavy ions are not influenced by the screening effect from the light ions, and an isolated high-energy heavy ion beam with relatively low-energy spread is thus formed. Two-dimensional particle-in-cell simulations show that ∼100MeV/u quasimonoenergetic Fe24+ beams can be obtained by linearly polarized laser pulses at intensities of 1.1×1021W/cm². 

Integrin genetic variants and stage-specific tumor recurrence in patients with stage II and III colon cancer

Integrins (ITGs) are key elements in cancer biology, regulating tumor growth, angiogenesis and lymphangiogenesis through interactions of the tumor cells with the microenvironment. Moving from the hypothesis that ITGs could have different effects in stage II and III colon cancer, we tested whether a comprehensive panel of germline single-nucleotide polymorphisms (SNPs) in ITG genes could predict stage-specific time to tumor recurrence (TTR). A total of 234 patients treated with 5-fluorouracil-based chemotherapy at the University of Southern California were included in this study. Whole-blood samples were analyzed for germline SNPs in ITG genes using PCR-restriction fragment length polymorphism or direct DNA sequencing. In the multivariable analysis, stage II colon cancer patients with at least one G allele for ITGB3 rs4642 had higher risk of recurrence (hazard ratio (HR)=4.027, 95% confidence interval (95% CI) 1.556-10.421, P=0.004). This association was also significant in the combined stage II-III cohort (HR=1.975, 95% CI 1.194-3.269, P=0.008). The predominant role of ITGB3 rs4642 in stage II diseases was confirmed using recursive partitioning, showing that ITGB3 rs4642 was the most important factor in stage II diseases. In contrast, in stage III diseases the combined analysis of ITGB1 rs2298141 and ITGA4 rs7562325 allowed to identify three distinct prognostic subgroups (P=0.009). The interaction between stage and the combined ITGB1 rs2298141 and ITGA4 rs7562325 on TTR was significant (P=0.025). This study identifies germline polymorphisms in ITG genes as independent stage-specific prognostic markers for stage II and III colon cancer. These data may help to select subgroups of patients who may benefit from ITG-targeted treatments.

Plastin Polymorphisms Predict Gender- and Stage-Specific Colon Cancer Recurrence after Adjuvant Chemotherapy

Tumor recurrence after curative resection remains a major problem in patients with locally advanced colorectal cancer treated with adjuvant chemotherapy. Genetic single-nucleotide polymorphisms (SNP) may serve as useful molecular markers to predict clinical outcomes in these patients and identify targets for future drug development. Recent in vitro and in vivo studies have demonstrated that the plastin genes PLS3 and LCP1 are overexpressed in colon cancer cells and play an important role in tumor cell invasion, adhesion, and migration. Hence, we hypothesized that functional genetic variations of plastin may have direct effects on the progression and prognosis of locally advanced colorectal cancer. We tested whether functional tagging polymorphisms of PLS3 and LCP1 predict time to tumor recurrence (TTR) in 732 patients (training set, 234; validation set, 498) with stage II/III colorectal cancer. The PLS3 rs11342 and LCP1 rs4941543 polymorphisms were associated with a significantly increased risk for recurrence in the training set. PLS3 rs6643869 showed a consistent association with TTR in the training and validation set, when stratified by gender and tumor location. Female patients with the PLS3 rs6643869 AA genotype had the shortest median TTR compared with those with any G allele in the training set [1.7 vs. 9.4 years; HR, 2.84; 95% confidence interval (CI), 1.32-6.1; P = 0.005] and validation set (3.3 vs. 13.7 years; HR, 2.07; 95% CI, 1.09-3.91; P = 0.021). Our findings suggest that several SNPs of the PLS3 and LCP1 genes could serve as gender- and/or stage-specific molecular predictors of tumor recurrence in stage II/III patients with colorectal cancer as well as potential therapeutic targets.

Gender-specific profiling in SCN1A polymorphisms and time-to-recurrence in patients with stage II/III colorectal cancer treated with adjuvant 5-fluoruracil chemotherapy

This study was designed to analyze the gender-related association between SCN1A polymorphisms (voltage-gated sodium channels; α-subunit) and time-to-recurrence (TTR) in patients with colorectal cancer (CRC) treated with 5-fluoruracil (5-FU)-based adjuvant chemotherapy. We enrolled from a prospective database patients with stage II and III CRC treated with adjuvant 5-FU-based chemotherapy. Genotypes for SCN1A rs3812718 and rs229877 were determined by direct DNA sequencing. One hundred twenty-seven males and 107 females were included in the study. In the univariate and multivariate analysis, the shortest TTR was associated with female patients carrying the rs3812718-TT genotype (hazard ratio (HR): 2.26 (95% confidence interval (CI): 0.89, 5.70), P=0.039) but with male patients carrying the rs3812718-CC genotype (HR: 0.49 (95% CI: 0.18, 1.38), P=0.048). For rs229877 the CT genotype was associated with a trend for shorter TTR in both gender populations. The study validated gender-dependent association between genomic SCN1A rs3812718 polymorphism and TTR in CRC patients treated with adjuvant 5-FU-based chemotherapy. This study confirms that voltage-gated Na+ channels may be a potential therapeutic target and a useful predictive biomarker before 5-FU infusion.