|
Activity Number:
|
473
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305809 |
|
Title:
|
Predicting Malignant Renal Lesions by Using Preoperative Color Doppler Ultrasonography: Building a Nomogram
|
|
Author(s):
|
Alexia Iasonos*+ and Ganesh V. Raj and Paul Russo
|
|
Companies:
|
Memorial Sloan-Kettering Cancer Center and Memorial Sloan-Kettering Cancer Center and Memorial Sloan-Kettering Cancer Center
|
|
Address:
|
307 E. 63rd street, 3rd Floor, New York, NY, 10021,
|
|
Keywords:
|
multivariate logistic regression ; nomogram ; ROC ; cross-validation
|
|
Abstract:
|
It is important to accurately identify patients with benign renal masses who may not need surgery for their lesion. Current methods like CT and ultrasonography (US) cannot adequately differentiate between malignant and benign lesions. We recently found an association between vascular flow on CDUS and clear cell carcinomas. However, the lack of specificity limits the ability of CDUS alone to reliably predict histology.In conjunction with patient characteristics and radiographic parameters, we developed a nomogram to predict histology, using a retrospective cohort of 299 lesions. Regression techniques for multivariate models, including statistically versus clinically significant variables, will be presented. The predictive accuracy of the model was assessed by the area under the Recite model was validated internally (10-fold cross-validation), and externally for overfit adjustment.
|
