Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
SUPPLEMENTARY MATERIAL Quantitative DCE-MRI data acquisition and reconstruction. The imaging protocol included coronal fast imaging with steady progression (TR/TE, 6.3/3.0; flip angle, 70°; matrix, 256 x 256; slice thickness/interslice gap, 6/2.4 mm; field of view, 350 x 350 mm), axial fat-suppressed turbo spin-echo T2-weighted (TR/effective TE, 5,000/182; echo-train length, 29; matrix, 97 x 256; field of view, 247 x 330 mm; slice thickness/interslice gap, 6/1.8), and axial fat-suppressed fast low-angle shot (FLASH) sequence (TR/TE, 261/4.8; flip angle, 70°; matrix, 96 x 256; field of view, 247 x 330 mm; slice thickness/interslice gap, 6/1.8) or similar. The DCE-MRI was aimed at a fixed slice where the largest metastasis is located in each patient, with 2D T1-weighted turbo FLASH sequences (301/1.2; flip angle, 8°; matrix, 125 x 256; field of view, 285 x 380 mm for coronal sections or 247 x 330 mm for axial sections; slice thickness, 6.5 mm). The acquisition time for each slice was 0.6 sec (or similar). Dynamic scanning started 5 sec after the initiation of contrast bolus injection. A 0.1 mmol/kg bolus of Gd-DTPA (Magnevist®, Bayer Schering Pharma, Berlin, Germany) was rapidly administered (at a rate of 2.5 mL/sec) by an injector via a previously placed 22-gauge IV cannula in an antecubital vein. Immediately afterwards, a 20-mL saline flush was administered at the same injection rate. One set of dynamic turbo FLASH sequences was performed. 500 dynamic images were obtained, always in the same anatomical section. It took a total of +/- 10 min from the initiation of contrast injection to the completion of the whole dynamic MRI study. A delayed contrast-enhanced axial fat-suppressed FLASH sequence was performed after the completion of the dynamic study. FDG-PET/CT image acquisition and reconstruction. 1 The patients fasted for at least 6 h prior to intravenous injection of [18F]-FDG and received 3.7 – 5.5 MBq/kg of body weight (0.10 – 0.15 mCi/kg) of FDG intravenously followed by 250 ml of sodium chloride and 20 mg of furosemide. Image acquisition started 50-70 minutes after injection of FDG in a relaxed supine position with the use of an integrated PET/CT scanner (Philips Gemini PET-CT, Philips Medical Systems) which consists of a gadolinium oxyorthosilicate full-ring PET scanner with 5.0-mm spatial resolution and a 16-slice helical CT scanner. First a total body CT with an iodinated intravenous contrast agent was performed. PET scanning was performed immediately after acquisition of the CT images, without changing the patient position. Patients were instructed to breathe normally during the acquisition of CT and PET images. The PET images were reconstructed using a three-dimensional row action maximum likelihood algorithm (3D RAMLA) as provided by the manufacturer. Images were corrected for decay, scatter, random events and attenuation. For the latter, CT data were used. FDG-PET/CT image analysis The metabolic response was defined by changes in FDG uptake in two liver target lesions, if available, as expressed by the standardized uptake value (SUV). This value is defined as the measured activity concentration (Bq/ml) divided by the injected dose (Bq) and normalized for total body weight (g). On the baseline image, a volume-ofinterest (VOI) was created around the two most metabolically active lesions (as determined by SUVmax) and around the lesion evaluated with DCE-MRI using an isocontour threshold method with the threshold set at 2 standard deviations above the mean SUV in healthy liver parenchyma. In 13 patients, the lesion assessed by DCE-MRI also showed the highest FDG-uptake. The VOIs were copied to the same lesions on the PET images 2 acquired at day 15 of cycle 5. If the lesions had disappeared at this time point, the VOIs were placed carefully in the same location using both PET and CT images and if necessary, the VOIs were slightly adapted to avoid extrahepatic expansion. The lesion SUVs were determined using the maximum value (SUVmax) measured in these VOIs. Pathological examination. Immunohistochemical staining was performed on formalin-fixed, paraffin embedded material using a standardized and automated immunostainer (Benchmark XT; Ventana Medical Systems, Tucson, AZ). Four-micrometer-thick sections were deparaffinized, pretreated with EDTA buffer (Ventana Medical Systems) and incubated with anti-Ki67 antibody (Neomarkers, Fremont, CA) or anti-CD31 antibody (Thermo Fisher Scientific, Fremont, CA). The ultraView Universal DAB Detection Kit (Ventana Medical Systems) was used as detection method to visualize the expression. Five separate high power fields (HPF) with the highest vascularization were selected and in these 5 HPF the number of blood vessels were counted. MVD is expressed as the total number of vessels divided by the number of x400 HPF. Ki-67 expression was quantified by counting the number of Ki-67 positive nuclei among 50 tumor cells in selected areas with the highest expression.. Ki-67 is expressed as the mean percentage of five positive areas. 3