Optimized approach to decision fusion of heterogeneous data for breast cancer diagnosis.

As more diagnostic testing options become available to physicians, it becomes more difficult to combine various types of medical information together in order to optimize the overall diagnosis. To improve diagnostic performance, here we introduce an approach to optimize a decision-fusion technique to combine heterogeneous information, such as from different modalities, feature categories, or institutions. For classifier comparison we used two performance metrics: The receiving operator characteristic (ROC) area under the curve [area under the ROC curve (AUC)] and the normalized partial area under the curve (pAUC). This study used four classifiers: Linear discriminant analysis (LDA), artificial neural network (ANN), and two variants of our decision-fusion technique, AUC-optimized (DF-A) and pAUC-optimized (DF-P) decision fusion. We applied each of these classifiers with 100-fold cross-validation to two heterogeneous breast cancer data sets: One of mass lesion features and a much more challenging one of microcalcification lesion features. For the calcification data set, DF-A outperformed the other classifiers in terms of AUC (p < 0.02) and achieved AUC=0.85 +/- 0.01. The DF-P surpassed the other classifiers in terms of pAUC (p < 0.01) and reached pAUC=0.38 +/- 0.02. For the mass data set, DF-A outperformed both the ANN and the LDA (p < 0.04) and achieved AUC=0.94 +/- 0.01. Although for this data set there were no statistically significant differences among the classifiers' pAUC values (pAUC=0.57 +/- 0.07 to 0.67 +/- 0.05, p > 0.10), the DF-P did significantly improve specificity versus the LDA at both 98% and 100% sensitivity (p < 0.04). In conclusion, decision fusion directly optimized clinically significant performance measures, such as AUC and pAUC, and sometimes outperformed two well-known machine-learning techniques when applied to two different breast cancer data sets.

Deposition of silver nanoparticles in geochemically heterogeneous porous media: predicting affinity from surface composition analysis.

The transport of uncoated silver nanoparticles (AgNPs) in a porous medium composed of silica glass beads modified with a partial coverage of iron oxide (hematite) was studied and compared to that in a porous medium composed of unmodified glass beads (GB). At a pH lower than the point of zero charge (PZC) of hematite, the affinity of AgNPs for a hematite-coated glass bead (FeO-GB) surface was significantly higher than that for an uncoated surface. There was a linear correlation between the average nanoparticle affinity for media composed of mixtures of FeO-GB and GB collectors and the relative composition of those media as quantified by the attachment efficiency over a range of mixing mass ratios of the two types of collectors, so that the average AgNPs affinity for these media is readily predicted from the mass (or surface) weighted average of affinities for each of the surface types. X-ray photoelectron spectroscopy (XPS) was used to quantify the composition of the collector surface as a basis for predicting the affinity between the nanoparticles for a heterogeneous collector surface. A correlation was also observed between the local abundances of AgNPs and FeO on the collector surface.

Stable isotope ratios indicate diet and habitat use in New World monkeys.

This paper demonstrates the use of stable isotope ratios of carbon and nitrogen in animal tissue for indicating aspects of species behavioral strategy. We analyzed hair from individuals representing four species of New World monkeys (Alouatta palliata, the mantled howler; Ateles geoffroyi, the spider monkey; Cebus capucinus, the capuchin; and Brachyteles arachnoides, the woolly-spider monkey or muriqui) for delta 13C and delta 15N using previously developed methods. There are no significant differences in either carbon or nitrogen ratios between sexes, sampling year, or year of analysis. Seasonal differences in delta 13C reached a low level of significance but do not affect general patterns. Variation within species was similar to that recorded previously within single individuals. The omega 13C data show a bimodal distribution with significant difference between the means. The two monkey populations living in an evergreen forest were similar to each other and different from the other two monkey populations that inhabited dry, deciduous forests. This bimodal distribution is independent of any particular species' diet and reflects the level of leaf cover in the two types of forest. The delta 15N data display three significantly different modes. The omnivorous capuchins were most positive reflecting a trophic level offset. The spider monkeys and the muriquis were similar to one another and significantly more positive than the howlers. This distribution among totally herbivorous species correlates with the ingestion of legumes by the howler monkey population. In combination, these data indicate that museum-curated primate material can be analyzed to yield information on forest cover and diet in populations and species lacking behavioral data.