Automated Brain Tumor Detection and Segmentation for Treatment Response Assessment Using Amino Acid PET

Evaluation of metabolic tumor volume (MTV) changes using amino acid PET has become an important tool for response assessment in brain tumor patients. MTV is usually determined by manual or semiautomatic delineation, which is laborious and may be prone to intra- and interobserver variability. The goal of our study was to develop a method for automated MTV segmentation and to evaluate its performance for response assessment in patients with gliomas. Methods: In total, 699 amino acid PET scans using the tracer O-(2-[18F]fluoroethyl)-l-tyrosine (18F-FET) from 555 brain tumor patients at initial diagnosis or during follow-up were retrospectively evaluated (mainly glioma patients, 76%). 18F-FET PET MTVs were segmented semiautomatically by experienced readers. An artificial neural network (no new U-Net) was configured on 476 scans from 399 patients, and the network performance was evaluated on a test dataset including 223 scans from 156 patients. Surface and volumetric Dice similarity coefficients (DSCs) were used to evaluate segmentation quality. Finally, the network was applied to a recently published 18F-FET PET study on response assessment in glioblastoma patients treated with adjuvant temozolomide chemotherapy for a fully automated response assessment in comparison to an experienced physician. Results: In the test dataset, 92% of lesions with increased uptake (n = 189) and 85% of lesions with iso- or hypometabolic uptake (n = 33) were correctly identified (F1 score, 92%). Single lesions with a contiguous uptake had the highest DSC, followed by lesions with heterogeneous, noncontiguous uptake and multifocal lesions (surface DSC: 0.96, 0.93, and 0.81 respectively; volume DSC: 0.83, 0.77, and 0.67, respectively). Change in MTV, as detected by the automated segmentation, was a significant determinant of disease-free and overall survival, in agreement with the physician's assessment. Conclusion: Our deep learning-based 18F-FET PET segmentation allows reliable, robust, and fully automated evaluation of MTV in brain tumor patients and demonstrates clinical value for automated response assessment.

Image Reconstruction Analysis for Positron Emission Tomography With Heterostructured Scintillators

The concept of structure engineering has been proposed for exploring the next generation of radiation detectors with improved performance. A TOF-PET geometry with heterostructured scintillators with a pixel size of 3.0 × 3.1 × 15 mm3 was simulated using Monte Carlo. The heterostructures consisted of alternating layers of BGO as a dense material with high stopping power and plastic (EJ232) as a fast light emitter. The detector time resolution was calculated as a function of the deposited and shared energy in both materials on an event-by-event basis. While sensitivity was reduced to 32% for 100-μm thick plastic layers and 52% for 50 μm, the coincidence time resolution (CTR) distribution improved to 204 ± 49 and 220 ± 41 ps, respectively, compared to 276 ps that we considered for bulk BGO. The complex distribution of timing resolutions was accounted for in the reconstruction. We divided the events into three groups based on their CTR and modeled them with different Gaussian TOF kernels. On an NEMA IQ phantom, the heterostructures had better contrast recovery in early iterations. On the other hand, BGO achieved a better contrast-to-noise ratio (CNR) after the 15th iteration due to the higher sensitivity. The developed simulation and reconstruction methods constitute new tools for evaluating different detector designs with complex time responses.

Advanced Monte Carlo simulations of emission tomography imaging systems with GATE

Built on top of the Geant4 toolkit, GATE is collaboratively developed for more than 15 years to design Monte Carlo simulations of nuclear-based imaging systems. It is, in particular, used by researchers and industrials to design, optimize, understand and create innovative emission tomography systems. In this paper, we reviewed the recent developments that have been proposed to simulate modern detectors and provide a comprehensive report on imaging systems that have been simulated and evaluated in GATE. Additionally, some methodological developments that are not specific for imaging but that can improve detector modeling and provide computation time gains, such as Variance Reduction Techniques and Artificial Intelligence integration, are described and discussed.

Design, evaluation and comparison of endorectal coils for hybrid MR-PET imaging of the prostate

Prostate cancer is one of the most common cancers among men and its early detection is critical for its successful treatment. The use of multimodal imaging, such as MR-PET, is most advantageous as it is able to provide detailed information about the prostate. However, as the human prostate is flexible and can move into different positions under external conditions, it is important to localise the focused region-of-interest using both MRI and PET under identical circumstances. In this work, we designed five commonly used linear and quadrature radiofrequency surface coils suitable for hybrid MR-PET use in endorectal applications. Due to the endorectal design and the shielded PET insert, the outer face of the coils investigated was curved and the region to be imaged was outside the volume of the coil. The tilting angles of the coils were varied with respect to the main magnetic field direction. This was done to approximate the various positions from which the prostate could be imaged. The transmit efficiencies and safety excitation efficiencies from simulations, together with the signal-to-noise ratios from the MR images were calculated and analysed. Overall, it was found that the overlapped loops driven in quadrature were superior to the other types of coils we tested. In order to determine the effect of the different coil designs on PET, transmission scans were carried out, and it was observed that the differences between attenuation maps with and without the coils were negligible. The findings of this work can provide useful guidance for the integration of such coil designs into MR-PET hybrid systems in the future.

Design and initial performance of PlanTIS: a high-resolution positron emission tomograph for plants

Positron emitters such as (11)C, (13)N and (18)F and their labelled compounds are widely used in clinical diagnosis and animal studies, but can also be used to study metabolic and physiological functions in plants dynamically and in vivo. A very particular tracer molecule is (11)CO(2) since it can be applied to a leaf as a gas. We have developed a Plant Tomographic Imaging System (PlanTIS), a high-resolution PET scanner for plant studies. Detectors, front-end electronics and data acquisition architecture of the scanner are based on the ClearPET system. The detectors consist of LSO and LuYAP crystals in phoswich configuration which are coupled to position-sensitive photomultiplier tubes. Signals are continuously sampled by free running ADCs, and data are stored in a list mode format. The detectors are arranged in a horizontal plane to allow the plants to be measured in the natural upright position. Two groups of four detector modules stand face-to-face and rotate around the field-of-view. This special system geometry requires dedicated image reconstruction and normalization procedures. We present the initial performance of the detector system and first phantom and plant measurements.

Combined MRI-PET dissects dynamic changes in plant structures and functions

Unravelling the factors determining the allocation of carbon to various plant organs is one of the great challenges of modern plant biology. Studying allocation under close to natural conditions requires non-invasive methods, which are now becoming available for measuring plants on a par with those developed for humans. By combining magnetic resonance imaging (MRI) and positron emission tomography (PET), we investigated three contrasting root/shoot systems growing in sand or soil, with respect to their structures, transport routes and the translocation dynamics of recently fixed photoassimilates labelled with the short-lived radioactive carbon isotope (11)C. Storage organs of sugar beet (Beta vulgaris) and radish plants (Raphanus sativus) were assessed using MRI, providing images of the internal structures of the organs with high spatial resolution, and while species-specific transport sectoralities, properties of assimilate allocation and unloading characteristics were measured using PET. Growth and carbon allocation within complex root systems were monitored in maize plants (Zea mays), and the results may be used to identify factors affecting root growth in natural substrates or in competition with roots of other plants. MRI-PET co-registration opens the door for non-invasive analysis of plant structures and transport processes that may change in response to genomic, developmental or environmental challenges. It is our aim to make the methods applicable for quantitative analyses of plant traits in phenotyping as well as in understanding the dynamics of key processes that are essential to plant performance.

The motion aftereffect: more than area V5/MT? Evidence from 15O-butanol PET studies

The motion aftereffect is a perceptual phenomenon which has been extensively investigated both psychologically and physiologically. Neuroimaging techniques have recently demonstrated that area V5/MT is activated during the perception of this illusion. The aim of this study was to test the hypothesis if a more broadly distributed network of brain regions subserves the motion aftereffect. To identify the neuronal structures involved in the perception of the motion aftereffect, regional cerebral blood flow (rCBF) measurements with positron emission tomography were performed in six normal volunteers. Data were analysed using SPM96. The motion-sensitive visual areas including area V5/MT were activated in both hemispheres. Additionally, the lateral parietal cortex bilaterally, the right dorsolateral prefrontal cortex, the anterior cingulate cortex and the left cerebellum showed significant increases in rCBF values during the experience of the waterfall illusion. In a further reference condition with identical attentional demand but no perception of a motion aftereffect elevated rCBF were found in these regions as well. In conclusion, our findings support the notion that the perceptual illusion of motion arises exclusively in the motion-sensitive visual area V5/MT. In addition, a more widespread network of brain regions including the prefrontal and parietal cortex is activated during the waterfall illusion which represents a non-motion aftereffect-specific subset of brain areas but is involved in more basic attentional processing and cognition.

The network of brain areas involved in the motion aftereffect

A network of brain areas is expected to be involved in supporting the motion aftereffect. The most active components of this network were determined by means of an fMRI study of nine subjects exposed to a visual stimulus of moving bars producing the effect. Across the subjects, common areas were identified during various stages of the effect, as well as networks of areas specific to a single stage. In addition to the well-known motion-sensitive area MT the prefrontal brain areas BA44 and 47 and the cingulate gyrus, as well as posterior sites such as BA37 and BA40, were important components during the period of the motion aftereffect experience. They appear to be involved in control circuitry for selecting which of a number of processing styles is appropriate. The experimental fMRI results of the activation levels and their time courses for the various areas are explored. Correlation analysis shows that there are effectively two separate and weakly coupled networks involved in the total process. Implications of the results for awareness of the effect itself are briefly considered in the final discussion.

Line bisection judgments implicate right parietal cortex and cerebellum as assessed by fMRI

OBJECTIVE

To use functional MRI (fMRI) to determine which brain regions are implicated when normal volunteers judge whether pretransected horizontal lines are correctly bisected (the Landmark test).

BACKGROUND

Manual line bisection and a variant thereof involving perceptual judgments of pretransected lines (the Landmark test) are widely used to assess unilateral visuospatial neglect in patients with neurologic disease. Although unilateral (left) neglect most often results from lesions to right temporoparietal cortex, the normal functional anatomy of the Landmark test has not been convincingly demonstrated.

METHODS

fMRI was carried out in 12 healthy right-handed male volunteers who judged whether horizontal lines were correctly prebisected. In the control task, subjects detected whether the horizontal lines contained a transection mark irrespective of the position of that mark. Response was by two-choice key press: on half the trials, subjects used the right, and on half, the left hand. Statistical analysis of evoked blood oxygenation level-dependent responses, measured with echoplanar imaging, employed statistical parametric mapping.

RESULTS

Performing the Landmark task showed neural activity (p < 0.05, corrected) in the right superior posterior and right inferior parietal lobe, early visual processing areas bilaterally, the cerebellar vermis, and the left cerebellar hemisphere. Only the latter area showed a significant interaction with hand used.

CONCLUSIONS

The right hemispheric dominance observed in inferior parietal cortex is consistent with the results of lesion studies. Right superior parietal cortex, vermis, and left cerebellar hemisphere have not been implicated in neglect, but all appear to play a cognitive role in the Landmark task.

Iodine-123 alpha-methyl tyrosine single-photon emission tomography of cerebral gliomas: standardised evaluation of tumour uptake and extent

Single-photon emission tomography (SPET) with the amino acid analogue l-3-[123I]iodo-alpha-methyl tyrosine (IMT) is helpful in the diagnosis and monitoring of cerebral gliomas. Radiolabelled amino acids seem to reflect tumour infiltration more specifically than conventional methods like magnetic resonance imaging and computed tomography. Automatic tumour delineation based on maximal tumour uptake may cause an overestimation of mean tumour uptake and an underestimation of tumour extension in tumours with circumscribed peaks. The aim of this study was to develop a program for tumour delineation and calculation of mean tumour uptake which takes into account the mean background activity and is thus optimised to the problem of tumour definition in IMT SPET. Using the frequency distribution of pixel intensities of the tomograms a program was developed which automatically detects a reference brain region and draws an isocontour region around the tumour taking into account mean brain radioactivity. Tumour area and tumour/brain ratios were calculated. A three-compartment phantom was simulated to test the program. The program was applied to IMT SPET studies of 20 patients with cerebral gliomas and was compared to the results of manual analysis by three different investigators. Activity ratios and chamber extension of the phantom were correctly calculated by the automatic analysis. A method based on image maxima alone failed to determine chamber extension correctly. Manual region of interest analysis in patient studies resulted in a mean inter-observer standard deviation of 8.7% +/ -6.1% (range 2.7% -25.0%). The mean value of the results of the manual analysis showed a significant correlation to the results of the automatic analysis (r = 0.91, P<0. 0001 for the uptake ratio; r = 0.87, P<0.0001 for the tumour area). We conclude that the algorithm proposed simplifies the calculation of uptake ratios and may be used for observer-independent evaluation of IMT SPET studies. Three-dimensional tumour recognition and transfer to co-registered morphological images based on this program may be useful for the planning of surgical and radiation treatment.

Whole-body kinetics and dosimetry of L-3--123I-iodo-alpha-methyltyrosine

The synthetic amino acid L-3--123I-iodo-alpha-methyltyrosine (IMT) is currently under clinical evaluation as a single-photon emission tomography (SPET) tracer of amino acid uptake in brain tumours. So far, dosimetric data in respect of IMT are not available. Therefore we investigated the whole-body distribution of IMT in six patients with cerebral gliomas and the radiation doses were estimated. Whole-body scans were acquired at 1.5, 3 and 5 h after i.v. injection of 370-550 MBq IMT. The bladder was voided prior to each scan and the radioactivity excreted in the urine was measured. Based on the MIRD-11 method and the updated MIRDOSE3, the mean absorbed doses for various organs and the effective dose were calculated from geometric means of the anterior and posterior whole-body scans using seven source organs and the residence time. IMT was predominantly excreted by the kidneys (52.8%+/-11.5% at 1.5 h p.i., 63.0%+/-15.7% at 3 h p.i. and 74.6%+/-9.8% at 5 h p.i.). No organ system other than the urinary tract showed significant retention of the tracer. Early whole-body scans revealed slightly increased tracer uptake in the liver and in the bowel. Highest absorbed doses were found for the urinary bladder wall (0.047 mGy/MBq), the kidneys (0.010 mGy/MBq), the lower large intestinal wall (0.011 mGy/MBq) and the upper large intestinal wall (0.008 mGy/MBq). The effective dose according to ICRP 60 was estimated to be 0.0073 mSv/MBq for adults. This leads to an effective dose of 3.65 mSv in a typical brain SPET study using 500 MBq IMT. The MIRDOSE3 scheme yielded similar results. Thus, in spite of the relatively high tracer dose required for optimal brain scanning, radiation exposure in SPET studies with IMT is in the normal range of routine nuclear medicine investigations.

Effect of partial volume correction on muscarinic cholinergic receptor imaging with single-photon emission tomography in patients with temporal lobe epilepsy

Animal experiments and preliminary results in humans have indicated alterations of hippocampal muscarinic acetylcholine receptors (mAChR) in temporal lobe epilepsy. Patients with temporal lobe epilepsy often present with a reduction in hippocampal volume. The aim of this study was to investigate the influence of hippocampal atrophy on the quantification of mAChR with single photon emission tomography (SPET) in patients with temporal lobe epilepsy. Cerebral uptake of the muscarinic cholinergic antagonist [123I]4-iododexetimide (IDex) was investigated by SPET in patients suffering from temporal lobe epilepsy of unilateral (n=6) or predominantly unilateral (n=1) onset. Regions of interest were drawn on co-registered magnetic resonance images. Hippocampal volume was determined in these regions and was used to correct the SPET results for partial volume effects. A ratio of hippocampal IDex binding on the affected side to that on the unaffected side was used to detect changes in muscarinic cholinergic receptor density. Before partial volume correction a decrease in hippocampal IDex binding on the focus side was found in each patient. After partial volume no convincing differences remained. Our results indicate that the reduction in hippocampal IDex binding in patients with epilepsy is due to a decrease in hippocampal volume rather than to a decrease in receptor concentration.

3-[123I]iodo-alpha-methyltyrosine and [methyl-11C]-L-methionine uptake in cerebral gliomas: a comparative study using SPECT and PET

This study compares the uptake of the nonmetabolizable amino acid analog 3-[123I]iodo-alpha-methyltyrosine (IMT) and of [methyl-11C]-L-methionine (MET) in cerebral gliomas.

METHODS

In 14 patients with cerebral gliomas, IMT uptake was measured using SPECT (10 dynamic, 4 static SPECT acquisitions) and, on the same day, MET uptake by dynamic PET. The IMT and MET data were compared with respect to tracer kinetics, tumor to brain ratios (T/B) and tumor size after converting the resolution of the PET scans to that of the SPECT scans (14 mm FWHM).

RESULTS

All gliomas showed increased uptake of both tracers in relation to normal brain tissue. Visual comparison of the scans yielded no differences in tumor size and shape with both methods. IMT showed a maximal tracer uptake in brain and in tumors at about 15 min postinjection which was followed by a washout of 45.0% +/- 13.5% in gliomas (mean +/- s.d., p < 0.001, n = 10) and 35.3% +/- 5.4% in normal brain (p < 0.001, n = 10) at 60 min postinjection. MET concentration in tumor tissue or brain tissue between 15 and 60 min remained constant. T/B ratios of IMT SPECT and MET PET showed a significant correlation at 15 min postinjection (r = 0.69, n = 10, p = 0.03), a low correlation for the mean values of the scans from 15-60 min postinjection (r = 0.54, n = 14, p = 0.05) and no correlation at 60 min postinjection (r = 0.09, n = 10, n.s.).

CONCLUSION

IMT and MET uptake in gliomas is similar in the early, transport dominated phase. There are some differences in tumor to brain ratios between both tracers within the first hour postinjection that are mainly caused by variable washout of IMT. Imaging of tumor extent with IMT SPECT is comparable to MET PET. Thus, amino acid SPECT using IMT is a promising tool to evaluate the biological activity and intracerebral infiltration of gliomas.

SPECT with HMPAO compared to PET with FDG in Huntington disease

OBJECTIVE

It is the aim of this study to compare the performance of 99mTc-d,l-hexamethylpropyleneamine oxime (HMPAO) SPECT with that of [18F]fluorodeoxyglucose (FDG) PET in detecting striatal dysfunction as it occurs in Huntington disease (HD).

MATERIALS AND METHODS

For the determination of regional cerebral glucose consumption, the PET camera PC-4096 was used; the cerebral uptake of HMPAO was measured using the three-head SPECT camera TRIAD. Eight patients with manifest HD, seven subjects at risk for HD, and nine normal individuals were included in the study. In both modalities data evaluation was performed using caudate-to-whole-slice (C/S) ratios. The patients' data were compared to 95% confidence intervals determined in the nine controls.

RESULTS

The PET and SPECT C/S values correlated significantly (n = 24; r = 0.87; p < 0.0001). The C/S values were significantly reduced in PET in all eight and in SPECT in seven of the eight HD patients studied. Five of the seven at-risk subjects had normal C/S values in PET and SPECT, one showed reduced C/S values in both diagnostic methods, and the remaining at-risk individual showed a reduced C/S value in PET only. Thus, concordant results between PET and SPECT were obtained in seven of eight patients and six of seven at-risk subjects studied, corresponding to an 87% accuracy of SPECT in the detection of striatal dysfunction as compared to the "gold standard" PET.

CONCLUSION

With use of a multidetector camera, HMPAO-SPECT comes near the performance of FDG-PET in the diagnosis of striatal dysfunction as it occurs in HD.