The application of Efron's bootstrap methods in validating feature classification using artificial neural networks for the analysis of mammographic masses

Efron's bootstrap resampling method is used to analyze the performance of artificial neural networks (ANNs) in the area of feature classification for the analysis of mammographic masses. The purpose of feature classification in mammography is to discover the salient information that can be used to discriminate benign from malignant masses. The performance of ANNs is typically measured in terms of the area under the receiver operating characteristics (ROC) curve (A/sub z/). Performance uncertainty problems and the generalization problems of ANNs are still the critical issues that impede the further application of ANNs in clinical medicine. It is unreasonable and impractical to justify the performance of one ANN being better than another just by its best A/sub z/ value. Efron's bootstrap methods make it possible to quantitatively analyze the performance of ANNs and anticipate its change tendency with relatively high accuracy. Our experimental results show that the probability model of A/sub z/ is close to a normal distribution. The performance of ANNs is more sensitive to the change of topology than that of the size and the composition of the training set. Bootstrap methods can be used to find the optimal epochs and avoid overfitting.

Partial volume effect in quantitative magnetic resonance perfusion imaging

In dynamic-susceptibility contrast (DSC) magnetic resonance (MR) perfusion imaging, the cerebral blood flow (CBF) is estimated from the tissue residue function obtained through deconvolution of the contrast concentration functions. However, the reliability of CBF estimates obtained by deconvolution is sensitive to various distortions. Among the most prominent experimental limitations is the image spatial resolution, leading to partial volume effect (PVE), which arises when the size of the voxel exceeds the volume containing the arterial input signal. PVE results in distortion of the arterial input function (AIF), and directly leads to miscalculation of the CBF. This work demonstrates the degree of the CBF estimation bias that could develop as a result of PVE.

Applying the transient error reconstruction algorithm in the assessment of the cerebral blood flow

The brain perfusion level, characterized by the cerebral blood flow (CBF) parameter, is a known indicator of blood supply in cerebral ischemic stroke. In magnetic resonance dynamic susceptibility contrast (DSC) perfusion studies the CBF parameter is estimated from the residue function obtained from deconvolving the tissue concentration curve by the arterial concentration curve. Deconvolution is a noise sensitive process and ensuring algorithmic stability leads to CBF biases. Distortions are introduced by noise reducing techniques in both the time-domain singular value decomposition (SVD) and frequency-domain based Fourier transform (FT) deconvolution approaches. We provide preliminary results of using the transient error reconstruction algorithm (TERA), an auto regressive moving average based technique, to compensate for these distortions. TERA is applied to determine the characteristics of the low-noise low frequency components of the residue function and then used to reconstruct the time-domain residue function. Results using noise-free signals indicate that the CBF estimates determined using TERA were less sensitive to the tissue mean transit time (MTT) than the time-domain SVD techniques. The difficulties encountered when applying TERA approach to signals with noise levels commonly found in MR perfusion studies are also discussed.

Regulation of Osteoblast Gene Expression and Phenotype by Polylactide-fatty Acid Surfaces

Cell function is influenced by surface structure and molecules. Molecules that enhance cellular differentiation can be applied to tissue scaffold surfaces to stimulate endogenous tissue regeneration. The application of this approach to bone implants yields surfaces coated with factors (proteins, peptides, etc...) that promote the differentiation of osteoblasts, the cells that make bone. Increased bone formation leads to increased healing and union of the implant with endogenous bone. To obtain better control over surface coating we developed PLLA copolymers with allyl (PLLA-co-DAG) and 3-hydroxypropyl (PLLA-co-HP) side chains to which we can attach functional groups. Given the potential of fatty acids being able to incorporate into lipid bilayers and/or influence gene expression, we grafted different fatty acid side chains to PLLA-co-HP by esterifying the corresponding fatty acids with the PLLA-co-HP 3-hydroxypropyl side chains. The effects of the polymer modifications on osteoblasts were then evaluated. While cellular morphology differed between surface coatings, they did not reflect changes in cellular phenotype. Changes in gene expression were most evident with arachidonate and 3-hydroxypropyl side-chains which exhibited osteoblast differentiating capabilities. Linoleate, myristate, oleate, and stearate ester side-chains did not have a significant influence on osteoblast phenotype. Growth characteristics of osteoblasts did not differ between the fatty acid copolymer films, although cells grown on PLLA-co-HP exhibited a trend toward increased growth. Taken together our findings demonstrate that surface fatty acid composition can impact osteoblast phenotype.

Is correction necessary when clinically determining quantitative cerebral perfusion parameters from multi-slice dynamic susceptibility contrast MR studies?

Several groups have modified the standard singular value decomposition (SVD) algorithm to produce delay-insensitive cerebral blood flow (CBF) estimates from dynamic susceptibility contrast (DSC) perfusion studies. However, new dependences of CBF estimates on bolus arrival times and slice position in multi-slice studies have been recently recognized. These conflicting findings can be reconciled by accounting for several experimental and algorithmic factors. Using simulation and clinical studies, the non-simultaneous measurement of arterial and tissue concentration curves (relative slice position) in a multi-slice study is shown to affect time-related perfusion parameters, e.g. arterial-tissue-delay measurements. However, the current clinical impact of relative slice position on amplitude-related perfusion parameters, e.g. CBF, can be expected to be small unless any of the following conditions are present individually or in combination: (a) high concentration curve signal-to-noise ratios, (b) small tissue mean transit times, (c) narrow arterial input functions or (d) low temporal resolution of the DSC image sequence. Recent improvements in magnetic resonance (MR) technology can easily be expected to lead to scenarios where these effects become increasingly important sources of inaccuracy for all perfusion parameter estimates. We show that using Fourier interpolated (high temporal resolution) residue functions reduces the systematic error of the perfusion parameters obtained from multi-slice studies.

Upper limits on a stochastic background of gravitational waves

The Laser Interferometer Gravitational-Wave Observatory has performed a third science run with much improved sensitivities of all three interferometers. We present an analysis of approximately 200 hours of data acquired during this run, used to search for a stochastic background of gravitational radiation. We place upper bounds on the energy density stored as gravitational radiation for three different spectral power laws. For the flat spectrum, our limit of omega0 < 8.4 x 10(-4) in the 69-156 Hz band is approximately 10(5) times lower than the previous result in this frequency range.

Limits on gravitational-wave emission from selected pulsars using LIGO data

We place direct upper limits on the amplitude of gravitational waves from 28 isolated radio pulsars by a coherent multidetector analysis of the data collected during the second science run of the LIGO interferometric detectors. These are the first direct upper limits for 26 of the 28 pulsars. We use coordinated radio observations for the first time to build radio-guided phase templates for the expected gravitational-wave signals. The unprecedented sensitivity of the detectors allows us to set strain upper limits as low as a few times 10(-24). These strain limits translate into limits on the equatorial ellipticities of the pulsars, which are smaller than 10(-5) for the four closest pulsars.

The status, quality, and expansion of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC)

The National Institutes of Health's Mammalian Gene Collection (MGC) project was designed to generate and sequence a publicly accessible cDNA resource containing a complete open reading frame (ORF) for every human and mouse gene. The project initially used a random strategy to select clones from a large number of cDNA libraries from diverse tissues. Candidate clones were chosen based on 5'-EST sequences, and then fully sequenced to high accuracy and analyzed by algorithms developed for this project. Currently, more than 11,000 human and 10,000 mouse genes are represented in MGC by at least one clone with a full ORF. The random selection approach is now reaching a saturation point, and a transition to protocols targeted at the missing transcripts is now required to complete the mouse and human collections. Comparison of the sequence of the MGC clones to reference genome sequences reveals that most cDNA clones are of very high sequence quality, although it is likely that some cDNAs may carry missense variants as a consequence of experimental artifact, such as PCR, cloning, or reverse transcriptase errors. Recently, a rat cDNA component was added to the project, and ongoing frog (Xenopus) and zebrafish (Danio) cDNA projects were expanded to take advantage of the high-throughput MGC pipeline.

Finasteride and bicalutamide as primary hormonal therapy in patients with advanced adenocarcinoma of the prostate

BACKGROUND

Medical or surgical castration is effective in advanced prostate cancer but with profound side-effects, particularly on sexual function. Effective, less toxic therapies are needed. This study examined whether the addition of finasteride to high-dose bicalutamide enhanced disease control, as measured by additional decreases in serum prostate-specific antigen (PSA).

PATIENTS AND METHODS

Forty-one patients with advanced prostate cancer received bicalutamide (150 mg/day). Finasteride (5 mg/day) was added at first PSA nadir. Serum PSA was measured every 2 weeks until disease progression. Questionnaires were administered to assess sexual function.

RESULTS

Median follow-up is 3.9 years. At the first PSA nadir, median decrease in PSA from baseline was 96.5%. Thirty of 41 patients (73%) achieved a second PSA nadir and median decrease of 98.5% from baseline. Median time to each nadir was 3.7 and 5.8 weeks, respectively. Median time to treatment failure was 21.3 months. Toxicities were minor, including gynecomastia. Seventeen of 29 (59%) and 12 of 24 (50%) men had normal sex drive at baseline and at second PSA nadir, respectively. One-third of men had spontaneous erection at both time points.

CONCLUSION

Finasteride provides additional intracellular androgen blockade when added to bicalutamide. Duration of control is comparable to castration, with preserved sexual function in some patients.

Removing the effect of SVD algorithmic artifacts present in quantitative MR perfusion studies

Quantitative cerebral blood flow (CBF) values can be obtained from dynamic susceptibility contrast (DSC) MR perfusion studies using the standard singular value decomposition (sSVD) deconvolution algorithm. Reports in the literature from simulation and in vivo studies suggest that CBF estimates obtained using sSVD deconvolution depend on the arterial-tissue delay (ATD). By contrast, Fourier transform (FT) deconvolution produces CBF estimates that are independent of ATD. The diagnostic reliability of quantitative CBF measurements to define areas of normal tissue flow and tissue at risk is brought into doubt by such gross sensitivity to the specifics of the deconvolution approach. This variation of CBF values with ATD is shown to be an artifact associated with the current implementation of the sSVD deconvolution algorithm. A reformulated version of the SVD deconvolution algorithm (rSVD) is presented and compared to the standard SVD algorithm through simulation and patient case studies.

Barking up the wrong tree? A survey of dog bite wound management

OBJECTIVES

Several trials have been published examining the role of antibiotics in dog bite wound management. A meta-analysis of these suggests that there is very little benefit to routine antibiotic prescription in these patients. All papers however incorporated rigorous wound care regimens involving large volume irrigation.

METHODS

The authors undertook a telephone survey to investigate wound care and prescribing practice in bite wound management in accident and emergency and minor injury units in the Yorkshire region.

RESULTS

Twenty one departments were contacted. Only 10% of departments routinely irrigate these wounds. Antibiotics are prescribed routinely in 71%.

CONCLUSIONS

Management of dog bite wounds would not seem to be evidence based in most departments in this sample.

Hazard identification by methods of animal-based toxicology

This paper is one of several prepared under the project "Food Safety In Europe: Risk Assessment of Chemicals in Food and Diet" (FOSIE), a European Commission Concerted Action Programme, organised by the International Life Sciences Institute, Europe (ILSI). The aim of the FOSIE project is to review the current state of the science of risk assessment of chemicals in food and diet, by consideration of the four stages of risk assessment, that is, hazard identification, hazard characterisation, exposure assessment and risk characterisation. The contribution of animal-based methods in toxicology to hazard identification of chemicals in food and diet is discussed. The importance of first applying existing technical and chemical knowledge to the design of safety testing programs for food chemicals is emphasised. There is consideration of the presently available and commonly used toxicity testing approaches and methodologies, including acute and repeated dose toxicity, reproductive and developmental toxicity, neurotoxicity, genotoxicity, carcinogenicity, immunotoxicity and food allergy. They are considered from the perspective of whether they are appropriate for assessing food chemicals and whether they are adequate to detect currently known or anticipated hazards from food. Gaps in knowledge and future research needs are identified; research on these could lead to improvements in the methods of hazard identification for food chemicals. The potential impact of some emerging techniques and toxicological issues on hazard identification for food chemicals, such as new measurement techniques, the use of transgenic animals, assessment of hormone balance and the possibilities for conducting studies in which common human diseases have been modelled, is also considered.

Application of transmission near-infrared spectroscopy to uniformity of content testing of intact steroid tablets

Transmission near-infrared (NIR) spectroscopy was used for the rapid and non-destructive determination of the content of a hormone steroid in single intact tablets. Tablets produced for clinical trial purposes containing 5, 10, 15, 20 and 30 mg (2.94, 5.88, 8.82, 11.76 and 17.64% m/m, respectively) were used to develop calibration models without the need to specially prepare any out of specification tablets. Reference values for the individual tablets used in the NIR calibration models and test set were measured by reversed-phase high performance liquid chromatography (HPLC). Partial least squares regression using standard normal variate transformed second-derivative spectra over the range 800 to 1040 nm gave the optimum calibration model with a standard error of calibration of 0.52 mg per tablet. Measurements of an independent test set gave comparable results (standard error of prediction 0.31 mg per tablet). Measurement errors for a single tablet (RSD < 2.5% for a given active level) were sufficiently small to allow the procedure to be applied to pharmacopoeial uniformity of content testing of batches of these tablets and permitted the non-destructive testing of 30 tablets in under 20 min as compared to 6 h by HPLC.

Lead poisoning of waterfowl by contaminated sediment in the Coeur d'Alene River

The Coeur d'Alene River basin in Idaho has been contaminated by mine tailings that have impaired the health of wildlife since the early 1900s. In other parts of the world, virtually all lead poisoning of waterfowl is caused by the ingestion of manmade lead artifacts, primarily spent lead shotshell pellets or, occasionally, fishing sinkers. However, in the Coeur d'Alene River basin in Idaho, nonartifactual lead poisoning was the ultimate cause of death of most of 219 (77%) of 285 waterfowl carcasses that had been found sick or dead from 1992 through 1997. The majority of these 219 waterfowl (172 tundra swans [Cygnus columbianus], 33 Canada geese [Branta canadensis], and 14 other species) were poisoned by ingesting river sediment that was contaminated with lead. The next most common cause of death (20 instances, 7%) was lead poisoning accompanied by ingested shotshell pellets. The remaining 46 waterfowl succumbed to trauma, infectious diseases (aspergillosis, avian cholera, tuberculosis), or miscellaneous problems, or the cause of death was not determined.

Pamidronate to prevent bone loss during androgen-deprivation therapy for prostate cancer

BACKGROUND

Treatment with a gonadotropin-releasing hormone agonist decreases bone mineral density and increases the risk of fracture in men with prostate cancer. We conducted a controlled study of the prevention of osteoporosis in men undergoing treatment with a gonadotropin-releasing hormone agonist.

METHODS

In a 48-week, open-label study, we randomly assigned 47 men with advanced or recurrent prostate cancer and no bone metastases to receive either leuprolide alone or leuprolide and pamidronate (60 mg intravenously every 12 weeks). Bone mineral density of the lumbar spine and the proximal femur was measured by dual-energy x-ray absorptiometry. Trabecular bone mineral density of the lumbar spine was measured by quantitative computed tomography. Forty-one men completed the study.

RESULTS

In men treated with leuprolide alone, the mean (+/-SE) bone mineral density decreased by 3.3+/-0.7 percent in the lumbar spine, 2.1+/-0.6 percent in the trochanter, and 1.8+/-0.4 percent in the total hip, and the mean trabecular bone mineral density of the lumbar spine decreased by 8.5+/-1.8 percent (P<0.001 for each comparison with the base-line value). In contrast, the mean bone mineral density did not change significantly at any skeletal site in men treated with both leuprolide and pamidronate. There were significant differences between the two groups in the mean changes in bone mineral density at 48 weeks in the lumbar spine (P<0.001), trochanter (P = 0.003), total hip (P=0.005), and trabecular bone of the lumbar spine (P=0.02).

CONCLUSIONS

Pamidronate prevents bone loss in the hip and lumbar spine in men receiving treatment for prostate cancer with a gonadotropin-releasing hormone agonist.

Regioselective aromatic borylation in an inert solvent

[reaction: see text]. A protocol for performing Rh catalyzed aromatic borylations in cyclohexane has been devised. Borylation at the 5-position of several 1,3-substituted aromatic species ranging from electron-rich (1,3-(NMe(2))(2)C(6)H(4)) to electron-deficient (1,3-(CF(3))(2)C(6)H(4)) yields the corresponding aryl boronate esters. Veratrole was selectively borylated at the 4-position, thus extending regioselectivity to 1,2-substituted benzenes. Selective borylation at the 3-position of an N-protected pyrrole has also been demonstrated, providing a valuable reagent for cross-coupling reactions in a single step.

Monocular optical constraints on collision control

A simulated ball-hitting task was used to explore the optical basis for collision control. Ball speed and size were manipulated in Experiments 1 and 2. Results showed a tendency for participants to respond earlier to slower and larger balls. Early in practice, participants would consistently miss the slowest and largest balls. Experiments 3 and 4 examined performance as a function of the range of speeds. Performance for identical speeds differed depending on whether the speeds were fastest or slowest within a range. Asymmetric transfer between the 2 ranges of speeds showed that those trained with slow speeds were very successful when tested with a faster range of speeds. Those trained with fast speeds did not do as well when tested on slower speeds. The pattern of results across 4 experiments suggests that participants were using optical angle and expansion rate as separate degrees of freedom for solving the collision task.

Neoadjuvant docetaxel followed by radical prostatectomy in patients with high-risk localized prostate cancer: a preliminary report

Effective treatment options for high-risk localized prostate cancer are limited. Patients at high risk for recurrence include those with biopsy Gleason scores of 8 to 10, prostate specific antigen (PSA) levels > 20 ng/mL, and clinical stage T3 disease. Docetaxel chemotherapy is active in hormone-refractory prostate cancer, either combined with estramustine or used as a single agent. To determine if systemic therapy can improve the outcome of radical prostatectomy in men with high-risk localized prostate cancer, we are undertaking a pilot phase II clinical trial of weekly docetaxel at 36 mg/m(2) for up to 6 months, followed by surgery. Patients are monitored with weekly visits, monthly digital rectal examinations, PSA measurement, and testosterone tests, and endorectal magnetic resonance imaging done at baseline, after two cycles, and again after six cycles. To date, 15 patients have been enrolled, and 70 cycles of chemotherapy have been administered. Toxicity has been mostly grade 1 in intensity, and fatigue has been the most common grade 2 toxicity reported. The primary endpoint of the trial is measurement of pathologic complete response rate, for which data are not yet available. Recruitment to the trial is ongoing.

The structure of dentine in the apical region of human teeth

AIM

To study the structure of the apical region of human teeth with emphasis on dentinal tubules and their branches.

METHODOLOGY

This descriptive histological study employed demineralized stained sections for light microscopy, demineralized unstained sections for scanning electron microscopy, and undemineralized, acid-etched specimens for confocal tandem scanning microscopy.

RESULTS

The apical portion of human teeth showed marked variations in structure, including accessory root canals, areas of resorption and repaired resorptions, occasional attached, embedded and free pulp stones, varied amounts of irregular secondary dentine, and even cementum-like tissue lining the apical root canal wall. The apex often deviated from the long axis of the root canal. Primary dentinal tubules were irregular in direction and density. Some areas were devoid of tubules.

CONCLUSIONS

The irregular and variable structure of the apical region of human teeth represent special challenges during endodontic therapy. Obturation techniques based on the penetration of adhesives into dentinal tubules are unlikely to be successful and adhesive techniques must depend on impregnation of a hybrid layer.

Low bone mineral density in hormone-naïve men with prostate carcinoma

BACKGROUND

The objective of this study was to determine the prevalence of low bone mineral density in men with prostate carcinoma and no history of androgen-deprivation therapy.

METHODS

The authors conducted a cross-sectional study in 41 hormone-naïve men with locally advanced, lymph node positive, or recurrent prostate carcinoma and no radiographic evidence of bone metastases. Bone mineral density of the total hip, posterior-anterior (PA) lumbar spine, and lateral lumbar spine was determined by dual-energy X-ray absorptiometry (DXA) using a densitometer. Trabecular bone mineral density of the lumbar spine was determined by quantitative computed tomography (QCT). Bone mineral density results were expressed in standard deviation units relative to young adult men (T score) and relative to age-matched men (Z score).

RESULTS

Fourteen of 41 men (34%; 95% confidence interval [95% CI], 20-51%) had T scores < -1.0 at one or more skeletal sites by DXA, 12 of 41 men (29%; 95% CI, 16-42%) had T scores between -1.0 and -2.5, and 2 of 41 men (5%; 95% CI, 1-17%) had T scores < -2.5. Thirty-nine of 41 men (95%; 95% CI, 83-99%) had T scores < -1.0 by QCT, 13 of 41 men (31%; 95% CI 18-48%) had T scores between -1.0 and -2.5, and 26 of 41 men (63%; 95% CI, 47-78%) had T scores < -2.5. T scores for trabecular bone mineral density of the lumbar spine were significantly lower than T scores for either the total hip (P < 0.001) or the PA lumbar spine (P < 0.001). The mean Z score for trabecular bone mineral density of the lumbar spine was -0.7 +/- 0.9. Hypogonadism, hypovitaminosis D, and dietary calcium intakes below the Recommended Daily Allowance were observed in 20%, and 17%, and 59% of study participants, respectively.

CONCLUSIONS

Many hormone-naïve men with prostate carcinoma have low bone mineral density. QCT is a more sensitive method than DXA for diagnosing low bone mineral density in this patient population. Trabecular bone mineral density is lower than expected for age and risk factors for osteoporosis are common.