Keywords:
Neuroradiology brain, Oncology, CNS, MR-Diffusion/Perfusion, MR, Diagnostic procedure, Imaging sequences, Observer performance, Cancer, Multidisciplinary cancer care
Authors:
S. Türk1, J. Kim2, R. Lobo3, J. Bapuraj3, T. Ma3, T. Johnson3, S. Camelo Piragua3, L. R. Junck3, A. Srinivasan3; 1Izmir/TR, 2Ann Arbor/US, 3Ann Arbor, MI/US
DOI:
10.26044/ecr2019/C-0958
Methods and materials
In this IRB approved,
HIPAA compliant study,
50 high-grade glioma patients with 150 MRI studies who underwent total tumor resection and adjuvant chemoradiation,
and had multiple follow up MRI studies between 2010-2018,
were evaluated.
The 94 image acquisitions (62%) were done on a 1.5 Tesla magnet,
there were 53 studies (35%) performed on a 3Tesla magnet and 3 studies (-2%) were done on a 1 Tesla magnet.
Philips and GE MRI machines were used.
12 case of pseudoprogression were included in this study together with 38 cases of true progression.
All cases were proven by histopathological evaluations following open biopsy or repeat surgery.
True progression cases included high-grade gliomas (n=7),
anaplastic gliomas (anaplastic oligoastrocytoma,
anaplastic oligocytoma,
anaplastic astrocytoma) (n=8) and glioblastoma multiforme (n=23).
50 % of the pathology reports contained at least a part of tissue with treatment effects,
however,
the dominant feature of either progression or pseudoprogression was given as the final verdict and the patients were then managed accordingly.
The last study before revision surgery with a known pathology result was chosen as a final time point.
This study and 2 prior follow up studies comprising of pre and -post-contrast axial axialT1,
axialT2 TSE,
axial T2-FLAIR,
axial ADC; and axial CBV,
CBF,
K2,
MTT maps were reviewed by 2 board-certified neuroradiologists.
Linear or nodular enhancement,
one or multiple enhancing lesions,
high or low ADC (compared to the contralateral white matter),
CBV,
CBF,
MTT,
K2 values,
FLAIR signal intensity changes within time were noted,
including interval changes in these parameters.
A scoring sheet was created depending on the RECIST,
MacDonald and RANO criteria.
(Table 1) Each statement was evaluated by the radiologist and marked as yes/no/nonapplicable.
Date 3 represented the last MRI study before surgery.
If a study did not contain appropriate images,
this statement marked as nonapplicable.
Radiologists had an independent final impression for each study which was described as true tumor progression,
pseudoprogression, or non-applicable.
The last criteria was employed for the first assessment since logically no conclusion could be drawn on a single assessment.
Odds ratios (OR) for each statement,
raters’ success of predicting true and pseudoprogression and interobserver reliability were calculated with R statistics software.
p<0.05 was considered significant,
OR<1 implied that the statement more strongly associated with pseudoprogression than true progression.