Type of study:
This single institute retrospective study was carried out at a dedicated tertiary care centre providing neurosurgical and neuro-imaging services for patients with brain tumors. Informed consent was obtained from all subjects prior to imaging.
Subjects:
All cases of histopathologically proven glial tumors who had undergone surgical resection followed by radiotherapy with or without chemotherapy and underwent simultaneous amino acid PET-MRI for suspected recurrence between January’2019 and March’2020 were included in the study. Exclusion criteria included non-glial primary brain tumors, metastatic lesions and standard contraindications for PET and /or MRI like pregnancy, end-stage renal disease and presence of a cardiac pacemaker or MRI incompatible metallic implants.
Imaging technique:
All patients underwent simultaneous amino-acid PET-MR imaging on a 3 Tesla SIEMENS, Biograph mMR scanner (Erlangen, Germany). All patients were fasted 4-6 hours prior to scanning for baseline stable metabolic conditions. On the day of imaging, all patients were injected 360-378 MBq of C11 methionine on the table through IV cannula. Simultaneous acquisition of PET images was performed along with UTE MR attenuation correction sequence (MRAC) along with other standard and advanced MRI sequences for 40 min in LIST MODE. PET images were acquired using the following parameters: 500mm FOV, 400mm anterior-posterior FOV, 1.0 zoom, 3 interactions, 21 subsets, HD PET reconstruction method, and 2.0mm Gaussian filter.
The following MR sequences were obtained during the PET acquisition:3D Fluid Attenuated Inversion Recovery (FLAIR)- TR/TE= 5000/385msec, TI=1800msec, voxel size=1x1x1mm, FOV=256x256; Axial T1 spin echo- TR/TE= 550/15msec, slice thickness-4mm, FOV=230x230; Axial T2 spin echo- TR/TE=5500/92msec, slice thickness-4mm, FOV=230x230; Axial Susceptibility Weighted Imaging (SWI)- TR/TE=27/20msec, flip angle=15degree, slice thickness=2mm, FOV=230x230; Axial Diffusion Weighted Imaging (DWI)- TR/TE=3900/81msec, slice thickness-4mm, FOV=230x230 at b values of 50 and 1000.
DSC perfusion was performed after the administration of gadolinium-based contrast agent in a dose of 0.1-0.15 mmol/kg body weight at a rate of 5-6ml/sec using a dual chamber injector connected to a 16-gauge cannula placed in the antecubital vein followed by 25ml saline chase at the same rate. Echo planar sequence was acquired with parameters as follows TR/TE= 1900/30msec, flip angle=90 degrees, slice thickness=4mm, FOV= 230x230, no. of slices=25. This was followed by acquisition of post-contrast T1 Magnetization Prepared Rapid Gradient-Echo (MPRAGE) sequence with TR/TE= 2200/2.33msec, TI=900msec, flip angle=8 degrees, FOV=256x256, voxel size=1x1x1mm.
Image Analysis:
The PET and MRI scans were analyzed by a nuclear medicine specialist and neuroradiologist respectively.
PET Analysis:
Quantitative ROI analysis: C11 methionine PET images were loaded into SIEMENS SYNGO Via (VB30) workstation. The ROI was drawn semi-automatically using an individually adapted isocontour of the tumour maximum using a standard ROI with a fixed diameter of 1.6 cm centred on the tumour maximum yielding a volume of 2 mL. Similar mirror ROI was placed in the contralateral brain parenchyma to calculate the background /normal brain parenchymal uptake. The values SUV max, and SUV mean were obtained for both tumour and normal brain parenchyma and tabulated. Ratio TBR max and TBR mean (tumour to normal brain/background) were calculated for statistical analysis.
MRI Analysis-
Quantitative ROI analysis: DSC perfusion, diffusion trace images with ADC maps and post-contrast images were loaded into Philips Intellispace Portal version 6.0. Perfusion images were processed using the leakage correction algorithm. The colored CBV maps were co-registered with the post-contrast image and ROI drawn in the area showing maximum contrast enhancement. A mirror ROI was placed in the contralateral normal white matter and the relative CBV ratio obtained. The ADC maps were also co-registered with the post-contrast images and ROI drawn in the same region of maximum contrast enhancement to obtain the mean ADC value. Another ROI was drawn in the contralateral normal white matter and the ADC ratio calculated.
Qualitative grading of diffusion restriction: A qualitative ordinal scale was used to grade the degree of diffusion restriction on ADC maps with grade 1 assigned when the region of interest appeared brighter than the normal white matter, grade 2 when the signal intensity is same as white matter, grade 3- less than white matter and grade 4- avid, unequivocal diffusion restriction.
Qualitative visual assessment for detecting recurrence: A visual analysis of the rCBV maps, ADC maps and post-contrast images was independently performed to assess for the presence of recurrence without quantification.
Spatial concordance between PET, perfusion and diffusion: The colored rCBV map, ADC map and post-contrast image each were independently compared to the PET image and the spatial concordance between the area of uptake, elevated rCBV on perfusion, restricted diffusion on ADC map and enhancement on post-contrast image graded as follows: grade 0- discordance, 1- fair (less than 50%) and 2- moderate (more than 50 %) concordance between area of PET uptake and elevated perfusion on rCBV map/restriction on ADC map/enhancement on post-contrast T1 MPRAGE.
Final diagnosis: The final diagnosis of recurrence versus radiation necrosis was based on histopathology when available and on clinical and /or imaging follow up for cases where it was not. Disease progression clinically or on imaging was classified as recurrence while stable disease was considered as necrosis.
Statistical Analysis:
Quantitative variables were expressed as median or mean with standard deviation. Qualitative variables were expressed as percentages. The PET, perfusion and diffusion parameters between the two groups were compared using the Mann-Whitney U test. Receiver operating characteristic (ROC) curve analysis was used to assess the diagnostic performance of each parameter in detecting recurrence. ROC curve analysis in combination with logistic regression analysis was used to measure the area under curve of various combination of parameters. The degree of agreement between perfusion, diffusion, contrast enhancement, PET and final diagnosis was estimated using the Cohen kappa statistic with values of .01-.20, .21-.40, .41-.60, .61-.80 and .81-1.00 indicating slight, fair, moderate, substantial and perfect agreement. Spatial concordance between rCBV maps, ADC maps, post-contrast MRI and PET was expressed as percentages. All analysis was performed on IBM SPSS version 26. A p value of less than 0.05 was regarded as significant.