No abstract

No abstract

Purpose:To determine whether a Bayesian network trained on a large database of patient demographic risk factors and radiologist-observed findings from consecutive clinical mammography examinations can exceed radiologist performance in the classification of mammographic findings as benign or malignant. Materials and Methods:The institutional review board exempted this HIPAA-compliant retrospective study from requiring informed consent. Structured reports from 48 744 consecutive pooled screening and diagnostic mammography examinations in 18 269 patients from April 5, 1999 to February 9, 2004 were collected. Mammographic findings were matched with a state cancer registry, which served as the reference standard. By using 10-fold cross validation, the Bayesian network was tested and trained to estimate breast cancer risk by using demographic risk factors (age, family and personal history of breast cancer, and use of hormone replacement therapy) and mammographic findings recorded in the Breast Imaging Reporting and Data System lexicon. The performance of radiologists compared with the Bayesian network was evaluated by using area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Results:The Bayesian network significantly exceeded the performance of interpreting radiologists in terms of AUC (0.960 vs 0.939, P ϭ .002), sensitivity (90.0% vs 85.3%, P Ͻ .001), and specificity (93.0% vs 88.1%, P Ͻ .001). Conclusion:On the basis of prospectively collected variables, the evaluated Bayesian network can predict the probability of breast cancer and exceed interpreting radiologist performance. Bayesian networks may help radiologists improve mammographic interpretation.

Purpose-To create a breast cancer risk estimation model based on the descriptors of National Mammography Database (NMD) format using logistic regression that can aid in decision-making for early detection of breast cancer. Material and Methods-InstitutionalReview Board waived this HIPAA-compliant retrospective study from requiring informed consent. We created two logistic regression models based on the mammography features and demographic data for 62,219 consecutive cases of mammography records from 48,744 studies in 18,270 patients reported using the Breast Imaging-Reporting and Data System (BI-RADS) lexicon and NMD format between 4/5/1999 and 2/9/2004. State cancer registry outcomes matched with our data served as the reference standard. The probability of cancer was the outcome in both models. Model-2 was built using all variables in Model-1 plus radiologists' BI-RADS assessment codes. We used 10-fold cross-validation to train and test the model and calculate the area under the receiver operating characteristic (ROC) curves (A z ) to measure the performance. Both models were compared to the radiologists' BI-RADS assessments.Results-Radiologists achieved an A z value of 0.939 ± 0.011. The A z was 0.927 ± 0.015 for Model-1 and 0.963 ± 0.009 for Model-2. At 90% specificity, the sensitivity of Model-2 (90%) was significantly better (P<0.001) than that of radiologists (82%) and Model-1 (83%). At 85% sensitivity, the specificity of Model 2 (96%) was significantly better (P<0.001) than that of radiologists (88%) and Model-1 (87%).Conclusions-Our logistic regression model can effectively discriminate between benign and malignant breast disease and identify the most important features associated with breast cancer.

We have used mechanically generated capillary wave and ellipsometric techniques to investigate interfacial viscoelastic properties of adsorbed monolayers of polystyrene-b-poly(methacrylic acid) diblock copolymer at an air−water interface, as a function of both the overall molecular weight, M w, and the nominal interfacial number density of the copolymer. This experiment is a follow-up of our earlier experiment, in which we studied adsorbed monolayers of the same diblock copolymer at the toluene−water interface (Macromolecules 1993, 26, 6595). Now we have changed the environment of the polystyrene block from toluene to air and have studied the effect of such a change. The most prominent effect of this change is that it is more difficult to attain equilibrium at the air−water interface. Unlike the toluene−water case, no clear saturation of surface pressure is observed at the air−water interface. The maximum surface pressure values measured at the air−water interface are smaller than the saturation surface pressure values in the toluene−water case for all the three molecular weights we have investigated. Ellipsometric study shows that only a very small fraction of the copolymer molecules added to the system is adsorbed at the air−water interface. However, substantial changes in the longitudinal elasticity and longitudinal viscosity are observed.