Improved diagnosis of breast implant rupture with sonographic findings and artificial neural networks.

RATIONALE AND OBJECTIVES: The authors evaluated the use of sonographic findings combined with artificial neural networks as an aid to the diagnosis of breast implant rupture. MATERIALS AND METHODS: From a database of 78 breast implants that were evaluated prospectively with sonography and then surgically removed, sonographic findings and surgical results were used to train and test backpropagation and radial basis function artificial neural networks by using the leave-one-out method. Receiver operating characteristic (ROC) curve analysis was used to compare the performance of the different neural networks with that of the radiologists involved. RESULTS: By using the ROC area index as a measure of performance, the artificial neural network (Az = 0.8744) outperformed the radiologists (Az = 0.8057), although not by a statistically significant difference (P = .09). The best-performing network used, in addition to the sonographic findings, the diagnosis of the radiologist as an input. This network (Az = 0.9245) outperformed both the radiologists and the "unaided" networks by a statistically significant margin (P = .02 for radiologists, P = .04 for the unaided network). The network performed remarkably well in those cases in which the radiologists classified the implant as indeterminate, predicting the correct diagnosis in 23 of 25 cases (92%). CONCLUSION: The results suggest that artificial neural networks in tandem with the unaided radiologic diagnosis can improve the accuracy rate in the detection of implant rupture based on sonographic findings. This "team" approach provided the best results.

Sonographic signs of breast implant rupture.

OBJECTIVE: This study evaluated sonography as a screening test for breast implant rupture and developed diagnostic criteria for implant rupture. SUBJECTS AND METHODS: Women contemplating implant removal were evaluated prospectively with sonography. Implants were classified as normal, indeterminate, or ruptured. Individual sonographic signs were analyzed for their statistic association with implant rupture. To measure the degree of interobserver variation, static images were evaluated by two different observers. RESULTS: Of 236 implants evaluated, surgical confirmation was available in 78, 22 of which were ruptured and 56, intact. Echogenic noise, multiple discontinuous parallel linear echoes, and echodense aggregates in the implant lumen were statistically associated with rupture. Based on sonographic findings, we classified 34 implants as intact (at surgery: 31 intact, 3 ruptured). We classified 19 as ruptured (at surgery: 11 ruptured, 8 intact). Of the 25 implants we classified as indeterminate, 17 were intact and 8 were ruptured at surgery. Therefore, sonography had a positive predictive value of 58%, a negative predictive value of 91%, a sensitivity of 50%, and a specificity of 55%. Receiver operating characteristic analysis suggests a learning curve effect and no significant interobserver variation. CONCLUSIONS: A normal sonographic result is highly predictive of an intact implant. Thus, sonography is useful in evaluating symptomatic women or women concerned about implant rupture. An indeterminate sonographic result suggests the need for further testing.

Introducing patient values into the decision making process for breast cancer screening.

Breast cancer is a serious and feared disease, and its management is a significant public health issue. Mammographic screening is a control strategy for this disease but its application in the United States is controversial. This article provides a brief review of the literature of physician/patient interaction styles, then proposes a comprehensive model that integrates the values and factors relevant to the decision over the spectrum of possible interaction styles. The issue of how screening program decisions for an individual woman should be made is considered in light of the current U.S. practice of offering population-wide screening guidelines. This approach is examined and contrasted with an approach in which individual patient values and preferences are used. The article offers some insights into how these values might be obtained so that they may be included in the decision-making process.

Tracheal and bronchial cartilaginous rings: warfarin sodium-induced calcification.

Progressive calcification of the cartilaginous rings (CCR) of the trachea and bronchi has been observed in patients undergoing prolonged prophylactic anticoagulant therapy with warfarin sodium. The purpose of this study was to validate the relationship of warfarin sodium and CCR, as well as to present the appearance and sex and age distribution of the normal degenerative CCR seen in the elderly. Chest radiographs were scrutinized for evidence of CCR in 92 patients who underwent warfarin sodium therapy and in 105 patients used as a control group. CCR was classified as not present (scored as 0), subtle (scored as 1), and extensive (scored as 2). In the warfarin sodium study group, 47% of patients (43 of 92) exhibited level 1 or level 2 CCR. This proportion was 19% (20 of 105) in the control group. The difference was statistically significant (P less than .001). A significant positive correlation (P less than .001) was also present between the duration of warfarin sodium therapy and increased levels of CCR.

Value of diagnostic tests for deep venous thrombosis: a decision analysis model.

A decision analysis model and data pooled from more than 12,000 patients were used to help elucidate conflicting results about the preferred method for diagnosing deep venous thrombosis (DVT). The prevalence of DVT in symptomatic legs is 40%, a value at which venography followed by Doppler ultrasound (in cases of nondiagnostic or unsuccessful venography) is preferred over the sequence of Doppler-plethysmography, provided that the sensitivity of the latter is less than 94%. If the prevalence of DVT decreases to 25% or the sensitivity of Doppler is 95% or higher, then the sequences of Doppler-plethysmography and venography-Doppler are equivalent in helping minimize resulting morbidity and mortality. Because of the greater clinical significance of proximal DVT, the analysis was modified to reflect changes in prevalence, sensitivity of noninvasive tests, and rate of pulmonary embolus due to DVT extending above the knee. In this case, performing plethysmography is slightly better than performing venography followed by Doppler or plethysmography.