Bayesian modeling of pretransplant variables accurately predicts kidney graft survival.

INTRODUCTION: Machine learning can enable the development of predictive models that incorporate multiple variables for a systems approach to organ allocation. We explored the principle of Bayesian Belief Network (BBN) to determine whether a predictive model of graft survival can be derived using pretransplant variables. Our hypothesis was that pretransplant donor and recipient variables, when considered together as a network, add incremental value to the classification of graft survival. METHODS: We performed a retrospective analysis of 5,144 randomly selected patients (age ≥18, deceased donor kidney only, first-time recipients) from the United States Renal Data System database between 2000 and 2001. Using this dataset, we developed a machine-learned BBN that functions as a pretransplant organ-matching tool. RESULTS: A network of 48 clinical variables was constructed and externally validated using an additional 2,204 patients of matching demographic characteristics. This model was able to predict graft failure within the first year or within 3 years (sensitivity 40%; specificity 80%; area under the curve, AUC, 0.63). Recipient BMI, gender, race, and donor age were amongst the pretransplant variables with strongest association to outcome. A 10-fold internal cross-validation showed similar results for 1-year (sensitivity 24%; specificity 80%; AUC 0.59) and 3-year (sensitivity 31%; specificity 80%; AUC 0.60) graft failure. CONCLUSION: We found recipient BMI, gender, race, and donor age to be influential predictors of outcome, while wait time and human leukocyte antigen matching were much less associated with outcome. BBN enabled us to examine variables from a large database to develop a robust predictive model.

Sixty-four slice multidetector computed tomographic angiography in the evaluation of vascular trauma.

BACKGROUND: Artifacts produced by metallic fragments and orthopedic hardware limit the usefulness of conventional computed tomography in many military trauma patients. Contemporary literature suggests that multidetector computed tomographic angiography (MDCTA) by resolving these limitations may provide a useful noninvasive alternative to invasive arteriography. The objective of this study is to review the utility of MDCTA in the evaluation of recent combat casualties with vascular injuries. METHODS: Data on all vascular trauma patients seen by our service has been collected prospectively and entered into a database. A retrospective review was conducted of patients seen from August through December 2006 who underwent MDCTA. Patient demographics, mechanism of injury, modality of evaluation, and findings were recorded. RESULTS: Twenty patients underwent MDCTA. Thirteen patients were injured by blast fragments. Seven patients were injured by gunshot wounds. Nineteen of 20 studies were diagnostic and one was judged to be indeterminate. Studies in nine patients identified arterial injuries. Multiple extremities were evaluated with a single study in 16 patients. Fifteen studies assessed the lower extremities, four the upper extremities and two the neck. Fourteen patients in this series had retained fragments, 10 had external fixators or intramedullary rods, and only 4 had neither retained fragments nor orthopedic hardware. MDCTA allowed for assessment of the arterial runoff despite hardware or fragments in 15 of 16 (94%) patients. Comparative studies were available in four patients in addition to MDCTA. There were no missed injuries in these four patients. CONCLUSIONS: MDCTA yielded high resolution images that were very useful for the delayed evaluation of combat casualties. The presence of metallic fragments or orthopedic hardware did not significantly interfere with MDCTA. It is a reliable and promising alternative to traditional arteriography for evaluating clinically occult vascular trauma.