Liang-Chih Wu

Fluorine-18 fluoromisonidazole tumour to muscle retention ratio for the detection of hypoxia in nasopharyngeal carcinoma

In vivo demonstration of hypoxia is of significance for tumour patient management. Fluorine-18 fluoromisonidazole ([18F]FMISO) is a proven hypoxie imaging agent. We developed an [18F]FMISO tumour to muscle retention ratio (TMRR) for the detection of tumour hypoxia in nasopharyngeal carcinoma (NPC). Data were acquired by positron emission tomography (PET) of the nasopharynx and neck after intravenous injection of 370 MBq of [18F]FMISO. Two imaging protocols were adopted: a long protocol for comprehensive dynamic information and a short protocol for a simple, clinically convenient imaging procedure. Tomograms were reconstructed and evaluated visually. ROI analysis on the basis of time-activity curve evaluation was performed to calculate the TMRR of NPC or cervical nodal metastases (CNMs) in relation to the suboccipital muscles at 2 h. The calculation of the TMRR was exactly the same for both the long and the short protocol as two 30-min composite frames had been created immediately after intravenous injection and 2 h after injection of [18F]FMISO in the long protocol. The normal tissue to muscle retention ratio (NTMRR) was derived similarly from the normal nasopharynx. The data of 12 controls and 24 patients with NPC were analysed. The long protocol was used in 15 patients, and the short protocol in nine. In controls, the mean NTMRR±1 SD was 0.96±0.14. The mean TMRRs for NPC and CNMs were 2.56±1.50 and 1.35±0.51, respectively; these values were significantly higher than the mean NTMRR for normal controls (P<0.005 in each case). At the retention threshold value of 1.24, tumour hypoxia occurred in 100% of the primary lesions of NPC and 58% of CNMs. The TMRR for undifferentiated carcinoma was significantly lower than that for non-keratinized carcinoma (P<0.05). The [18F]FMISO TMRR is a simple and clinically useful index for detecting tumour hypoxia in NPC.

Persistent Left Superior Vena Cava Demonstrated by First-Pass Radionuclide Angiography.

Persistent left superior vena cava (PLSVC) is a development variation of the embryonic thoracic venous system. It can be isolated or associated with congenital heart disease combined with shunting problems. Many image findings of PLSVC have been reported, but few mentioned findings in a first-pass radionuclide angiography. We report a case of PLSVC found incidentally in a first-pass radionuclide angiography with tracer injection through the left jugular vein. The right ventricular ejection fraction was underestimated. Injection via the right jugular or right cubital vein is recommended to obtain accurate ejection fractions in cases of PLSVC without shunting.

A practical background correction method for an immediately repeated first-pass radionuclide angiography

Background: A satisfactory bolus injection is essential for a successful first‐pass radionuclide angiography (FPRNA). Rescheduling the FPRNA study is usually needed due to high background interference caused by an unsatisfactory bolus injection. We developed a protocol to correct the pre‐existing background activity subsequent to immediately repeating the study. Methods: Seventy‐four consecutive patients who had their bone scan and FPRNA scheduled on the same day were included for analysis. The initial 51 cases constituted the “validation‐only” group. In the other 23 cases, the “validation plus clearance constants” group, a 5‐min dynamic acquisition was performed during the 5‐min equilibrium to obtain the background clearance curve and the clearance constants. For all included 74 cases ejection fraction (EF) analysis was proceeded using the images from the first injection, second injection, and second injection with the corrected background to yield EF1, EF2, and EF2′, respectively. EF2 and EF2′ were then compared to the ejection fraction without background interference, the EF1. Results: For the LV, the mean difference between the EF1 and the uncorrected EF2 (|LVEF1‐LVEF2| in mean ± SD) was 3.1 ± 2.0% and the difference between the EF1 and the corrected EF2′ (|LVEF1‐LVEF2′|) was 1.6 ± 2.1%, while the mean differences for RV are 2.2 ± 1.9% and 1.8 ± 1.8%, respectively. A significant difference (p < 0.05) was observed between the uncorrected and the corrected data for both the LV and RV. Conclusion: In FPRNA, when a bolus injection is immediately readministered, both LVEF and RVEF can be underestimated. With our correction method, the results are superior to those without correction.

Effects of the Acute and Chronic Ethanol Intoxication on Acetate Metabolism and Kinetics in the Rat Brain

BACKGROUND Ethanol (EtOH) intoxication inhibits glucose transport and decreases overall brain glucose metabolism; however, humans with long-term EtOH consumption were found to have a significant increase in [1-11 C]-acetate uptake in the brain. The relationship between the cause and effect of [1-11 C]-acetate kinetics and acute/chronic EtOH intoxication, however, is still unclear. METHODS [1-11 C]-acetate positron emission tomography (PET) with dynamic measurement of K1 and k2 rate constants was used to investigate the changes in acetate metabolism in different brain regions of rats with acute or chronic EtOH intoxication. RESULTS PET imaging demonstrated decreased [1-11 C]-acetate uptake in rat brain with acute EtOH intoxication, but this increased with chronic EtOH intoxication. Tracer uptake rate constant K1 and clearance rate constant k2 were decreased in acutely intoxicated rats. No significant change was noted in K1 and k2 in chronic EtOH intoxication, although 6 of 7 brain regions showed slightly higher k2 than baseline. These results indicate that acute EtOH intoxication accelerated acetate transport and metabolism in the rat brain, whereas chronic EtOH intoxication status showed no significant effect. CONCLUSIONS In vivo PET study confirmed the modulatory role of EtOH, administered acutely or chronically, in [1-11 C]-acetate kinetics and metabolism in the rat brain. Acute EtOH intoxication may inhibit the transport and metabolism of acetate in the brain, whereas chronic EtOH exposure may lead to the adaptation of the rat brain to EtOH in acetate utilization. [1-11 C]-acetate PET imaging is a feasible approach to study the effect of EtOH on acetate metabolism in rat brain.

Using 3-D OFEM for movement correction and quantitative evaluation in dynamic cardiac NH3 PET images

Various forms of cardiac pathology, such as myocardial ischemia and infarction, can be characterized with 13NH3-PET images. In clinical situation, polar map (bullseye image), which derived by combining images from multiple planes (designated by the circle around the myocardium in the above images), so that information of the entire myocardium can be displayed in a single image for diagnosis. However, image artifact problem always arises from body movement or breathing motion in image acquisition period and results in indefinite myocardium disorder region shown in bullseye image. In this study, a 3-D motion and movement correction method is developed to solve the image artifact problem to improve the accuracy of diagnostic bullseye image. The proposed method is based on 3-D optical flow estimation method (OFEM) and cooperates with the particular dynamic imaging protocol, which snaps serial PET images (5 frames) in later half imaging period. The 3-D OFEM assigns to each image point in the visual 3-D flow velocity field, which associates with the non-rigid motion of the time-varying brightness of a sequence of images. It presents vectors of corresponding images position between frames for motion correction. To validate the performance of proposed method, 10 normal and 20 abnormal whole-body dynamic PET imaging studies were applied, and the results show that the bullseye images, which generated by corrected images, present clear and definite tissue region for clinical diagnosis.

Inter-subject MR-PET image registration and integration

A MR-PET inter-subject image integration technique is developed to provide more precise anatomical location based on a template MR image, and to examine the anatomical variation in sensory-motor stimulation or to obtain cross-subject signal averaging to enhance the delectability of focal brain activity detected by different subject PET images. In this study, a multimodality intra-subject image registration procedure is firstly applied to align MR and PET images of the same subject. The second procedure is to estimate an elastic image transformation that can nonlinearly deform each 3D brain MR image and map them to the template MR image. The estimation procedure of the elastic image transformation is based on a strategy that searches the best local image match to achieve an optimal global image match, iteratively. The final elastic image transformation estimated for each subject will then be used to deform the MR-PET registered PET image. After the nonlinear PET image deformation, MR-PET inter-subject mapping, averaging, and fusing are simultaneously accomplished. The developed technique has been implemented to an UNIX based workstation with Motif window system. The software named Elastic-IRIS has few requirements of user interaction. The registered anatomical location of 10 different subjects has a standard deviation of /spl sim/2 mm in the x, y, and z directions. The processing time for one MR-PET inter-subject registration ranged from 20 to 30 minutes on a SUN SPARC-20.