High-dose busulfan, melphalan, thiotepa and peripheral blood stem cell infusion for the treatment of metastatic breast cancer

The purpose of this study was to determine the outcome of patients with metastatic breast cancer treated with high-dose busulfan (Bu), melphalan (Mel) and thiotepa (TT) followed by peripheral blood stem cell (PBSC) infusion. Fifty-one patients with chemotherapy refractory (n = 32) or responsive (n = 19) metastatic breast cancer received Bu (12 mg/kg), Mel (100 mg/m2) and TT (500 mg/m2) followed by PBSC collected after chemotherapy and growth factor (n = 43) or growth factor alone (n = 8). The 100 day treatment-related mortality was 8% including one death from cytomegalovirus pneumonia, one from aspiration pneumonia and two from regimen-related toxicity (RRT). Seven of 28 refractory (25%) and 5/7 (71%) responsive patients with evaluable disease achieved a complete response of all measurable disease or all soft tissue disease with at least improvement in bone lesions (PR*). Fifteen of 51 patients (29%) are alive and progression-free a median of 423 days (range 353-934) after treatment, 5/32 (16%) with refractory disease and 10/19 (53%) with responsive disease. The probabilities of progression-free survival (PFS) at 1.5 years for the patients with refractory (n = 32) and responsive (n = 19) disease were 0.24 and 0.53, respectively. These preliminary data suggest that high-dose Bu/Mel/TT has significant activity in patients with advanced breast cancer and may be superior to some previously published regimens.     

Epitope specificity of CD44 for monoclonal antibody-dependent facilitation of marrow engraftment in a canine model

Primary graft rejection after marrow transplantation occurs more frequently in patients receiving HLA-haploidentical compared with HLA-identical sibling transplants. Both human and experimental animal data suggest that the cells responsible for this phenomenon are either host natural killer (NK) cells, T cells, or both. To investigate the mechanisms of graft rejection, we have developed a canine model of marrow transplantation, which uses DLA-nonidentical unrelated donors in the absence of postgrafting immunosuppression. In this model most animals rejected their marrow grafts after a preparative regimen of 9.2 Gy total body irradiation (TBI). However, engraftment of DLA-nonidentical marrow can be facilitated when the recipients are pretreated with monoclonal antibody (MoAb) S5, which recognizes CD44. In this report, we extended these observations by first cloning the canine CD44 and, next, mapping the epitope recognized by S5, which was located in a region conserved among human and canine CD44 and was distinct from the hyaluronan binding domain. However, in vitro binding of S5 caused a conformational change in CD44, which allowed increased hyaluronan binding. Then, we reexamined the in vivo model of marrow transplantation and compared results with MoAb S5 to those with two other anti-CD44 MoAbs, IM7 and S3. Only MoAb S5 significantly increased the engraftment rate of DLA-nonidentical unrelated marrow, whereas the two other anti-CD44 MoAbs were ineffective. The enhanced in vivo effect was not related to differences in the MoAbs' avidities, since both S5 and IM7 had equivalent binding to CD44, but most likely related to the specific epitope that S5 recognizes. Thus, this study shows that the effect of the anti-CD44 MoAb S5 in facilitating engraftment is epitope specific and if one is to use an anti-CD44 to facilitate engraftment of marrow in humans, one cannot assume that any anti-CD44 would work.     

Two studies of the long-term follow-up of minimal therapist contact treatments of vascular and tension headache.

Tension and vascular headache patients, initially treated with biofeedback and/or relaxation training in either a minimal therapist contact protocol (3 visits) or an intensive individual protocol (10 or 16 visits) were followed-up prospectively for 2 years. In the first study, for the first 6 months of follow-up, half of all patients continued to keep headache diaries and were seen monthly and the other half had only minimal contact. The results at 1-year follow-up, based on 4 weeks of daily headache diaries, revealed equally good maintenance from both treatment protocols and from both follow-up conditions. In Study 2, we found that patients remained improved over pretreatment baseline levels at the 2-year follow-up regardless of initial treatment intensity. Approximately three quarters of vascular patients who were initially improved at posttreatment remained improved at 2 years. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Antipneumococcal activities of levofloxacin and clarithromycin as determined by agar dilution, microdilution, E-test, and disk diffusion methodologies

The activities of levofloxacin and clarithromycin against 199 penicillin- and macrolide-susceptible and -resistant pneumococci were tested by agar and microdilution methods in air and by disk diffusion and E-test methods in air and CO2. For levofloxacin, >/=99. 0% of strains were susceptible at /=17 mm, regardless of incubation in air or CO2. Although zone sizes were smaller and E-test MICs were higher for clarithromycin in CO2 than those in air, category differences were minor, and susceptibility rates for clarithromycin were similar to those obtained by agar and microdilution in air (range, 76.9 to 80.9% by all methods). For clarithromycin, adjustment of breakpoints based upon distribution of results resulted in susceptibility rates which were similar by all methods (75.8 to 76.9% susceptible, 0 to 1.5% intermediate, 22.6 to 23.1% resistant). Minor discrepancies were obtained with levofloxacin for one strain (0.5%) by microdilution and two strains (1.0%) by disk diffusion in CO2. For clarithromycin, minor discrepancies were found in three strains (1.5%) by microdilution, seven strains (3.5%) by agar dilution, four strains (2.0%) by E-test in air, six strains (3.0%) by disk diffusion in air, and five strains (2.5%) by disk diffusion in CO2. Major discrepancies occurred with levofloxacin in one strain (0.5%) by microdilution but were not found with clarithromycin. Very major discrepancies were not seen with levofloxacin, but occurred with clarithromycin in five strains (2.5%) by microdilution, three strains (1.5%) by agar dilution, two strains (1.0%) by E-test in air, eight strains (4.0%) by disk diffusion in air, and one strain (0.5%) by disk diffusion in CO2.     

Evaluation of Face Datasets as Tools for Assessing the Performance of Face Recognition Methods

Face datasets are considered a primary tool for evaluating the efficacy of face recognition methods. Here we show that in many of the commonly used face datasets, face images can be recognized accurately at a rate significantly higher than random even when no face, hair or clothes features appear in the image. The experiments were done by cutting a small background area from each face image, so that each face dataset provided a new image dataset which included only seemingly blank images. Then, an image classification method was used in order to check the classification accuracy. Experimental results show that the classification accuracy ranged between 13.5% (color FERET) to 99% (YaleB). These results indicate that the performance of face recognition methods measured using face image datasets may be biased. Compilable source code used for this experiment is freely available for download via the Internet.     

Learning dataset representation for automatic machine learning algorithm selection

The algorithm selection problem is defined as identifying the best-performing machine learning (ML) algorithm for a given combination of dataset, task, and evaluation measure. The human expertise required to evaluate the increasing number of ML algorithms available has resulted in the need to automate the algorithm selection task. Various approaches have emerged to handle the automatic algorithm selection challenge, including meta-learning. Meta-learning is a popular approach that leverages accumulated experience for future learning and typically involves dataset characterization. Existing meta-learning methods often represent a dataset using predefined features and thus cannot be generalized across different ML tasks, or alternatively, learn a dataset’s representation in a supervised manner and therefore are unable to deal with unsupervised tasks. In this study, we propose a novel learning-based task-agnostic method for producing dataset representations. Then, we introduce TRIO, a meta-learning approach, that utilizes the proposed dataset representations to accurately recommend top-performing algorithms for previously unseen datasets. TRIO first learns graphical representations for the datasets, using four tools to learn the latent interactions among dataset instances and then utilizes a graph convolutional neural network technique to extract embedding representations from the graphs obtained. We extensively evaluate the effectiveness of our approach on 337 datasets and 195 ML algorithms, demonstrating that TRIO significantly outperforms state-of-the-art methods for algorithm selection for both supervised (classification and regression) and unsupervised (clustering) tasks.

     

Wikiometrics: a Wikipedia based ranking system

We present a new concept—Wikiometrics—the derivation of metrics and indicators from Wikipedia. Wikipedia provides an accurate representation of the real world due to its size, structure, editing policy and popularity. We demonstrate an innovative “mining” methodology, where different elements of Wikipedia – content, structure, editorial actions and reader reviews – are used to rank items in a manner which is by no means inferior to rankings produced by experts or other methods. We test our proposed method by applying it to two real-world ranking problems: top world universities and academic journals. Our proposed ranking methods were compared to leading and widely accepted benchmarks, and were found to be extremely correlative but with the advantage of the data being publically available.     

Parameter estimation and screening of solar cells

The aggregation (sorting) of the individual solar cells into an array is commonly based on a single operating point on the current-voltage (I-V) characteristic curve. an alternative approach for cell performance prediction and cell screening is provided by modelling the cell using an equivalent electrical circuit, in which the parameters involved are related to the physical phenomena in the device. These analytical models may be represented by a double exponential I-V characteristic with seven parameters, by a double exponential model with five parameters or by a single exponential equation with four or five parameters. In this article we address issues concerning methodologies for the determination of solar cell parameters based on measured data points of the I-V characteristic, and introduce a procedure for screening solar cells for arrays. We show that common curve-fitting techniques, e.g. least-squares, may produce many combinations of parameter values while maintaining a good fit between the fitted and measured I-V characteristics of the cell. Therefore, techniques relying on curve-fitting criteria alone cannot be used directly for cell parameterization. We propose a consistent procedure that takes into account the entire set of parameter values for a batch of cells. This procedure is based on a definition of a mean cell representing the batch, and takes into account the relative contribution of each parameter to the overall goodness of fit. the procedure is demonstrated on a batch of 50 silicon cells for Space Station Freedom.     

Collective-agreement-based pruning of ensembles

Ensemble methods combine several individual pattern classifiers in order to achieve better classification. The challenge is to choose the minimal number of classifiers that achieve the best performance. An ensemble that contains too many members might incur large storage requirements and even reduce the classification performance. The goal of ensemble pruning is to identify a subset of ensemble members that performs at least as good as the original ensemble and discard any other members.In this paper, we introduce the Collective-Agreement-based Pruning (CAP) method. Rather than ranking individual members, CAP ranks subsets by considering the individual predictive ability of each member along with the degree of redundancy among them. Subsets whose members highly agree with the class while having low inter-agreement are preferred.