TY - JOUR
T1 - Posttreatment recurrence of malignant brain neoplasm
T2 - Accuracy of relative cerebral blood volume fraction in discriminating low from high malignant histologic volume fraction
AU - Gasparetto, Emerson L.
AU - Pawlak, Mikolaj A.
AU - Patel, Sohil H.
AU - Huse, Jason
AU - Woo, John H.
AU - Krejza, Jaroslaw
AU - Rosenfeld, Myrna R.
AU - O'Rourke, Donald M.
AU - Lustig, Robert
AU - Melhem, Elias R.
AU - Wolf, Ronald L.
N1 - Copyright:
Copyright 2012 Elsevier B.V., All rights reserved.
PY - 2009/3
Y1 - 2009/3
N2 - Purpose: To determine the accuracy of relative cerebral blood volume (rCBV) fraction for distinguishing high-grade recurrent neoplasm from treatment-related necrosis (TRN) in enhancing masses identified on surveillance magnetic resonance (MR) images following treatment for primary or secondary brain neoplasm. Materials and Methods: This institutional review board approved and HIPAA-com-pliant retrospective study included 30 patients undergoing resection of recurrent enhancing mass appearing after treatment with surgery and radiation, with or without chemotherapy. The enhancing mass volume was manually segmented on three-dimensional T1-weighted images. The rCBV maps were created by using T2-weighted dynamic susceptibility contrast perfusion MR imaging and registered to T1-weighted images, and the fraction of enhancing mass with rCBV above a range of thresholds was calculated. A receiver operating characteristic (ROC) curve was created by calculating sensitivity-specificity pairs at each threshold for rCBV fraction (≤20% or >20%) by using percentage of malignant features at histo-logic evaluation as the reference criterion. Relationships between rCBV and probability of recurrence were estimated by using logistic regression analysis. Results: ROC analysis showed excellent discriminating accuracy of rCBV fraction (area under the ROC curve, 0.97 ± 0.03 [standard error]) and high efficiency (93%) with an rCBV threshold of 1.8 times that of normal-appearing white matter. Logistic regression analysis showed that a unit increase of rCBV is associated with a 254-fold increase (95% confidence interval: 43, 1504, P < .001) of the odds that enhanced tissue is recurrence, adjusting for age, treatment, volume of enhancing tissue, and time to suspected recurrence. Conclusion: The fraction of malignant histologic features in enhancing masses recurring after treatment for brain neoplasms can be predicted by using the rCBV fraction, with improved differentiation between recurrent neoplasm and TRN.
AB - Purpose: To determine the accuracy of relative cerebral blood volume (rCBV) fraction for distinguishing high-grade recurrent neoplasm from treatment-related necrosis (TRN) in enhancing masses identified on surveillance magnetic resonance (MR) images following treatment for primary or secondary brain neoplasm. Materials and Methods: This institutional review board approved and HIPAA-com-pliant retrospective study included 30 patients undergoing resection of recurrent enhancing mass appearing after treatment with surgery and radiation, with or without chemotherapy. The enhancing mass volume was manually segmented on three-dimensional T1-weighted images. The rCBV maps were created by using T2-weighted dynamic susceptibility contrast perfusion MR imaging and registered to T1-weighted images, and the fraction of enhancing mass with rCBV above a range of thresholds was calculated. A receiver operating characteristic (ROC) curve was created by calculating sensitivity-specificity pairs at each threshold for rCBV fraction (≤20% or >20%) by using percentage of malignant features at histo-logic evaluation as the reference criterion. Relationships between rCBV and probability of recurrence were estimated by using logistic regression analysis. Results: ROC analysis showed excellent discriminating accuracy of rCBV fraction (area under the ROC curve, 0.97 ± 0.03 [standard error]) and high efficiency (93%) with an rCBV threshold of 1.8 times that of normal-appearing white matter. Logistic regression analysis showed that a unit increase of rCBV is associated with a 254-fold increase (95% confidence interval: 43, 1504, P < .001) of the odds that enhanced tissue is recurrence, adjusting for age, treatment, volume of enhancing tissue, and time to suspected recurrence. Conclusion: The fraction of malignant histologic features in enhancing masses recurring after treatment for brain neoplasms can be predicted by using the rCBV fraction, with improved differentiation between recurrent neoplasm and TRN.
UR - http://www.scopus.com/inward/record.url?scp=62649103363&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=62649103363&partnerID=8YFLogxK
U2 - 10.1148/radiol.2502071444
DO - 10.1148/radiol.2502071444
M3 - Article
C2 - 19244052
AN - SCOPUS:62649103363
SN - 0033-8419
VL - 250
SP - 887
EP - 896
JO - RADIOLOGY
JF - RADIOLOGY
IS - 3
ER -