15th European Molecular Imaging Meeting
supported by:

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).
*.ics

Best of vEMIM Poster-Pitches | Imaging Technologies

   
Shortcut: BO-08
Date: Friday, 28 August, 2020, 3:30 p.m. - 5:00 p.m.
Session type: Spotlight Symposium

Contents

Abstract/Video opens by clicking at the talk title.

BO-08-01

Imaging breast malignancies with the second generation Twente photoacoustic mammoscope

Sjoukje M. Schoustra1, 2, Tim J. P. M. op 't Root3, Caroline A. H. Klazen4, Margreet van der Schaaf4, Jeroen Veltman5, Wiendelt Steenbergen2, Srirang Manohar1

1 University of Twente, Multi-Modality Medical Imaging, Enschede, Netherlands
2 University of Twente, Biomedical Photonic Imaging, Enschede, Netherlands
3 PA Imaging R&D B.V., Enschede, Netherlands
4 Medisch Spectrum Twente, Enschede, Netherlands
5 Ziekenhuisgroep Twente, Almelo/Hengelo, Netherlands

Introduction

Breast cancer is the most frequently diagnosed cancer among women worldwide and is responsible for most cancer deaths. [1] Imaging is crucial for diagnosis, treatment and follow-up of breast cancer. [2] Currently used clinical imaging techniques have limitations, motivating research into new imaging methods. Photoacoustics combines functional optical contrast with ultrasound resolution. [3] Tumor-induced angiogenesis leads to a higher blood vessel density, generating absorption contrast between tumor and healthy tissue for photoacoustics. [4]

Methods

Recently, we presented our second generation version mammoscope (PAM 2) [5] employing a tomographic imaging configuration, where we imaged vascularization in the breasts of healthy volunteers. This system has been used to image patients suspicious for malignant breast cancer (BI-RADS 4 or 5 after anamnesis, mammography, and/or ultrasound). Of the 24 patients included in the study, both breasts were imaged with PAM 2 at two excitation wavelengths (755 and 1064 nm). Two patients were measured with a PVC cup to support and position the breast. All measurements were approved by an institutional review board (METC Twente, Enschede). The obtained photoacoustic images are compared to conventional clinical images; the gold standard for diagnosis is histopathology.

Results/Discussion

Of 24 patients included with malignancies (proven by histopathology after our measurement), three patients could not be measured due to technical or subject’s mobility issues. The images of the 21 measured subjects vary greatly in quality. Current hypotheses to explain this include a sub-optimal light delivery and detection geometry for larger breasts as well as positioning problems and motion artifacts. These last mentioned problems have been solved using PVC cups to support and position the breast, which improves image quality. A figure is attached showing maximum intensity projections of a healthy volunteer (age 55) imaged with help of a breast supporting cup. These cups also ease the implementation of a two-layer speed-of-sound model (discerning water and tissue) in the reconstruction. Results of patient measurements are being analyzed in detail and will be presented at the conference.

Conclusions

Three-dimensional photoacoustic breast images have been obtained with our PAM 2 system of patients with malignancies. Results will be analyzed and compared to conventional images in order to extract and study features indicative of malignant lesions. Images of the breast of a healthy volunteer are promising, since they show detailed vascularization. We will report on the results of patient measurements during the conference.

Acknowledgment

Stichting Achmea Gezondheidszorg (SAG) funds are acknowledged for support. Authors would like to thank Rutger Pompe van Meerdervoort, Laurens Alink, and Wouter Muller Kobold, all from PA Imaging R&D B.V., for their contributions.

References
[1] Torre, L.A., et al., Global cancer statistics, 2012. CA Cancer Journal for Clinicians, 2015. 65(2): p. 87-108.
[2] Baltzer, P.A.T., et al., New diagnostic tools for breast cancer. Memo - Magazine of European Medical Oncology, 2017. 10(3): p. 175-180.
[3] Manohar, S. and Razansky, D., Photoacoustics: A historical review. Advances in Optics and Photonics, 2016. 8(4): p. 586-617.
[4] Beard, P., Biomedical photoacoustic imaging. Interface Focus, 2011. 1(4): p. 602-631.
[5] Schoustra, S.M., et al., Twente Photoacoustic Mammoscope 2: system overview and three-dimensional vascular network images in healthy breasts. Journal of biomedical optics, 2019. 24(12): p. 121909.
Reconstructed healthy volunteer measurement

breast illuminated with 755 nm; 3 maximum intensity projections (MIPs),

Keywords: optoacoustics, photoacoustics, tomography, breast cancer, breast imaging
BO-08-02

New statistical reconstruction method for quantitative imaging in preclinical studies: preliminary results

Cristobal Martinez Sanchez1, 2, Manuel Desco Menéndez1, 2, 4, Mónica Abella García1, 2, 3

1 Universidad Carlos III de Madrid, Dpto. Bioingeniería e Ingeniería Aeroespacial, Leganes, Spain
2 Instituto de Investigación Sanitaria Gregorio Marañón, Madrid, Spain
3 Centro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC), Madrid, Spain
4 Centro de investigación en red salud mental (CIBER- SAM), Madrid, Spain

Introduction

The polychromatic nature of the spectra in commercial CT scanners cause an effect  known as beam hardening (BH), which hinders the reconstruction of quantitative values. The BH correction in combination with the Hounsfield Units (HU) calibration allow to obtain similar values of only for soft tissue for different kVp but not for bone [1], introducing a possible confounding factor in bone studies. We propose a new reconstruction method that recovers the tissue density values that are independent of the energy or the acquisition scanner.

Methods

We propose a new statistical method that includes the modelling of the polychromatic nature of the spectrum through the BH function, which relates the total attenuation value with the density thickness traversed. This function is obtained in a calibration step with a phantom composed of PMMA and Al-6082, which have equivalent attenuation properties to soft tissue and bone respectively (Figure 1). The algorithm uses this function on the forward model and minimizes the log-likelihood with a Huber function as regularizer.
Preliminary evaluation was done in simulation, calculating the root mean square error (RMSE) with respect to the real density image, and in real data with two rodent studies acquired with the ARGUS-CT scanner.

Results/Discussion

We can see in top panel of Figure 2 a reduction of the dark bands between the bones and streaks due to the lack of projections with the proposed method. For the conventional reconstruction, the RMSE was 9% of the mean value in bone and 17% in soft tissue. With the proposed reconstruction method these values were reduced to 1.5% in both bone and soft tissue.
We can see a reduction of dark bands in the soft tissue and a reduction of the bone values also in real data. However, further evaluation with known density values of the acquired sample would be advisable to evaluate quantitatively the real cases.

Conclusions

We have presented a new statistical reconstruction algorithm that models the polyenergetic nature of the X-ray source with a simple calibration step. Results showed a complete restoration of the density values in simulation data and a complete elimination of the artificats due to beam-hardening effect in real data.

AcknowledgmentThis work has been supported by Ministerio de Ciencia, Innovación y Universidades, Agencia Estatal de Investigación, projects “DPI2016 79075 R AEI/FEDER, UE ”, Instituto de Salud Carlos III, project “DTS17/00122 ”, co funded by European Regional Development Fund (ERDF), “A way of making Europe” Europe”. The CNIC is supported by the Ministerio de Ciencia, Innovación y Universidades and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (SEV 2015 0505). 
References
[1] Cann, C. E. (1988). Quantitative CT for determination of bone mineral density: a review. Radiology, 166(2), 509-522.
Workflow of the calibration step
Axial slice of the reconstruction of simulated and rodent data
Keywords: X-ray tomography, Quantitave, Polychromatic, Image reconstruction, Beam hardening
BO-08-03

Cerenkov luminescence imaging in pulmonary and hepatic metastasectomy

Esther Ciarrocchi1, 2, Nicola Belcari1, 2, Francesco Bartoli3, Angela G. Cataldi3, Pinuccia Faviana4, Luca Morelli5, 6, Marco Lucchi7, 8, Claudio Traino9, Sara Vitali3, Paola A. Erba3, 10

1 University of Pisa, Department of Physics, Pisa, Italy
2 National Institute of Nuclear Physics, Section of Pisa, Pisa, Italy
3 University of Pisa, Department of Translational Research and of New Surgical and Medical Technologies, Pisa, Italy
4 University Hospital of Pisa (Azienda Ospedaliero-Universitaria Pisana - AOUP), Unit of Anatomy and histopathology 3, Pisa, Italy
5 University of Pisa, Department of General Surgery, Pisa, Italy
6 University Hospital of Pisa (Azienda Ospedaliero-Universitaria Pisana - AOUP), Unit of General Surgery, Pisa, Italy
7 University of Pisa, Department of Thoracic Surgery, Pisa, Italy
8 University Hospital of Pisa (Azienda Ospedaliero-Universitaria Pisana - AOUP), Unit of Thoracic Surgery, Pisa, Italy
9 University Hospital of Pisa (Azienda Ospedaliero-Universitaria Pisana - AOUP), Unit of Medical Physics, Pisa, Italy
10 University Hospital of Pisa (Azienda Ospedaliero-Universitaria Pisana - AOUP), Unit of Nuclear Medicine, Pisa, Italy

Introduction

Cerenkov luminescence imaging (CLI) is an optical imaging modality to detect distributions of radiopharmaceuticals. CLI can be used to visualize surgical margins immediately after resection and to refine surgery in a single procedure [1]. We are planning a clinical study to evaluate the impact of CLI during surgery of lung and liver metastasis from various primary tumors with respect to conventional post-operative histology, and we are performing in-vitro simulation measurements to optimize the clinical protocol in terms of patient inclusion criteria, activity to inject, radiation monitoring.

Methods

We analyzed PET/CT data of 15 patients performed with [18F]-FDG in pulmonary and hepatic metastases and with 68Ga-DOTATOC in neuroendocrine tumors (NETs) to determine typical injected activities, lesion volumes, uptakes and time delays between injection and imaging. We are collecting data for typical histological margins. Since the Cerenkov signal depends on the spectrum of the beta particles and on the optical properties of the tissue, we prepared phantoms to measure the minimum detectable activity as a function of the type of radiopharmaceutical, the type of tissue and the source depth in tissue. The phantoms were imaged with a LightPath system with acquisition time acceptable for clinical needs. We used two short-pass filters to discriminate the depth of origin of the detected light.

Results/Discussion

Patient data are summarized in Figure 1. The delay between injection and imaging was ~1 hour. For our clinical study, we expect delays up to 4-5 hours between injection and CLI. The decay-corrected mean uptake values to account for this delay (5-11 kBq/cc) are comparable to the minimum detectable activity level of 8 kBq/cc that we measured for [18F]-FDG [2]. For 68Ga-DOTATOC, the final uptake of 4-7 kBq/cc should be well detectable, because our first tests suggest a 13x signal increase with respect to 18F, but an enhancement up to 22x can be expected [3]. Figure 2a shows a representative phantom image. The attenuation of 68Ga signal in the various animal liver samples is shown in Fig. 2b. Independent data-sets for the same type of tissue suggest good reproducibility. We are finalizing the data analysis to determine the target-to-background ratio in both the patient and phantom data.

Conclusions

Patient data suggests that CLI can be performed with standard clinical activities and 5-minute exposure times. The typical lesion volumes are suitable for LightPath imaging. Phantom data for signal attenuation in biological tissue show good reproducibility. We are collecting additional data for lung phantoms, and we are studying the target-to-background ratio for 18F and 68Ga and a method to extract the source depth from the spectral images.

References
[1] M.R. Grootendorst, et al. "Intraoperative assessment of tumor resection margins in breast-conserving surgery using 18F-FDG Cerenkov luminescence imaging: a first-in-human feasibility study." Journal of Nuclear Medicine 58.6 (2017): 891-898.
[2] E. Ciarrocchi, et al. "Performance evaluation of the LightPath imaging system for intra-operative Cerenkov luminescence imaging." Physica Medica 52 (2018): 122-128.
[3] J. olde Heuvel, et al. "Performance evaluation of Cerenkov luminescence imaging: a comparison of 68 Ga with 18 F." EJNMMI physics 6.1 (2019): 17.
Figure 1. Summary of patient PET/CT data for three tumor types.
Total injected activity, total lesion glycolysis (TLG), total activity in the volume, and uptake for PET/CT data of 15 patients with lung or liver metastases or NETs, imaged with 18F-FDG or 68Ga-DOTATOC.
Figure 2. Results of the in-vitro simulation measurements.
a) Representative CLI image (false color) overlaid on the reference photo (black and white). 68GaCl3 was diluted in 4 ml of distilled water and covered with 4 mm of pork liver. The image was acquired with 300 s exposure, binning 8x8 and no optical filters. b) Attenuation of the CLI signal from 68Ga as a function of the source depth in various liver specimens, for phantoms as shown in a). Data were normalized for the source activity and the exposure time. No optical filters were used in this case.
Keywords: cerenkov luminescence imaging, cancer surgery, surgical margin assessment
BO-08-04

Fast 3D Hyperpolarised 13C Metabolic MRI at 7 T using Spectrally-Selective bSSFP

Geoffrey Topping1, Jason G. Skinner1, Irina Heid2, Maximilian Aigner1, Martin Grashei1, Christian Hundshammer1, Lukas Kritzner2, Frits van Heijster1, Rickmer Braren2, Franz Schilling1

1 Technical University of Munich, Department of Nuclear Medicine, Klinikum rechts der Isar, Munich, Germany
2 Technical University of Munich, Institute of Radiology, Klinikum rechts der Isar, Munich, Germany

Introduction

Metabolic magnetic resonance spectroscopic imaging (MRSI) using hyperpolarized (HP) compounds, such as 13C-labelled pyruvate to lactate conversion, requires pulse sequences that encode spatial and spectral information quickly and efficiently. Spectrally-selective excitations exploit the sparsity of the hyperpolarized 13C spectrum, allowing spatial and temporal resolution to be improved. This work establishes a 3D balanced steady-state free precession (bSSFP) sequence using this approach, which is applied to preclinical metabolic imaging at 7 T.

Methods

A 3D bSSFP sequence was modified by disabling slice selection and rephasing gradients, and changing the excitation frequency1,2 and power after each 3D volume is acquired. Single-side-lobed sinc-shaped excitation RF pulses with narrow 400 Hz (TR 11.27 ms) or 900 Hz (TR 6.29 ms) FWHM bandwidth were used, alternating between frequencies close to each resonance, which was also in a bSSFP pass band, to produce separate images of 13C-pyruvate and lactate.
A small animal 7 T MRI (Agilent/Bruker) was used for HP 13C-pyruvate-lactate MRSI and anatomical MRI of 6 pancreatic tumour bearing (PDAC) mice3 and 12 healthy mice. 13C pyruvate was hyperpolarized (Hypersense) and injected by tail vail (80 mM) immediately before bSSFP acquisition. T2-weighted anatomical MRI of these mice were also acquired.

Results/Discussion

With 400 Hz RF pulses, images were obtained with 3 mm isotropic resolution in 950 ms per (single metabolite) 3D image (Fig. 1). With 900 Hz RF pulses, images were obtained with 1.75 mm isotropic resolution in 1212 ms per 3D image (Fig. 2).
Evidence of heterogeneity was detected in PDAC mice tumours: AUC ratios were 1.26 for ROI1 and 1.41 for ROI2 (Fig. 1).
Shortening TR by relaxing the RF pulse FWHM from 400 Hz to 900 Hz lead to a substantial increase in resolution, from (3 mm)3 to (1.75 mm)3, with only a slight increase in scantime (950 ms vs. 1212 ms). Furthermore, the passbands became broader (88.74 Hz to 158.98 Hz), simplifying interpretation of quantification, because B0 inhomogeneities are less likely to lead to banding artifacts. Placement of the excitations on the far sides of the resonances means only the target resonance is excited, despite the broader RF excitation bandwidth.

Conclusions

Spectrally-selective 3D bSSFP can provide high spatiotemporal resolution, banding-artefact-free images of metabolic activity in mice at 7 T. We demonstrate this in healthy mice and mice with pancreatic tumours (PDAC). Potential tumour heterogeneity was detected in PDAC mice.

AcknowledgmentWe acknowledge support from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation – 391523415, SFB 824).
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 820374.
We thank Sybille Reder and Markus Mittelhäuser for performing the PET measurements, and Sandra Sühnel for assisting with the MRSI experiments.
References
[1] Milshteyn E, von Morze C, Gordon JW, Zhu Z, Larson PE, Vigneron DBJ 2018, 'High spatiotemporal resolution bSSFP imaging of hyperpolarized [1‐13C] pyruvate and [1‐13C] lactate with spectral suppression of alanine and pyruvate‐hydrate' MRM 80(3), 1048-1060
[2] Shang H, Sukumar S, von Morze C, Bok RA, Marco‐Rius I, Kerr A, Reed GD, Milshteyn E, Ohliger MA, Kurhanewicz JJ 2017, 'Spectrally selective three‐dimensional dynamic balanced steady‐state free precession for hyperpolarized C‐13 metabolic imaging with spectrally selective radiofrequency pulses', MRM 78(3), 963-975
[3] Heid I, Steiger K, Trajkovic-Arsic M, Settles M, Eßwein MR, Erkan M, Kleeff J, Jäger C, Friess H, Haller BJ 2017, 'Co-clinical assessment of tumor cellularity in pancreatic cancer', CCR 23(6), 1461-1470.
Spectrally selective (400 Hz FWHM) 3D bSSFP in a PDAC mouse
Hyperpolarized 13C pyruvate-lactate bSSFP images of a PDAC tumour mouse. Per-metabolite scantime = 950 ms, TR = 11.27 ms, resolution = (3 mm)3, αPy = 2° (bSSFP lobe 1), αLac = 10° (bSSFP lobe 1). All 13C images are overlaid on T2w 1H anatomical images. A: 6 select lactate image timepoints marked with two ROIs (green and blue) that correspond to the metabolite signal dynamics plotted in B&C respectively. AUC ratios for ROI1 and ROI2 = 1.26 and 1.41, potentially reflecting heterogeneity within the tumour.
Spectrally selective (900 Hz FWHM) 3D bSSFP in a healthy mouse
Hyperpolarized 13C pyruvate-lactate bSSFP images of a healthy mouse. Per-metabolite scantime = 1212 ms, TR = 6.29 ms, resolution = (1.75 mm)3. 13C images are overlaid on T2w 1H anatomical images. A&B show all 12 slices of the 13C images for 14 timepoints for pyruvate (αPy = 4°, bSSFP lobe 2) and lactate (αLac = 90°, bSSFP lobe 5). C&D are select images from A&B, indicated by the white boxes. Vena cava, heart, and kidneys are visible in both metabolites. The lactate phantom is visible in D&B slice 8 and absent in A slice 8. E shows metabolite signal dynamics in the marked kidney.
Keywords: Hyperpolarized 13C, bSSFP, PDAC, MRSI
BO-08-05

Performance of nanoScan PET/CT and PET/MRI and comparison of image derived quantifications with ex vivo tissue distribution in tumor bearing mice

Marion Chomet1, Maxime Schreurs1, Ricardo Vos1, Marc Huisman1, Mariska Verlaan1, Esther Kooijman1, Guus A. M. S. van Dongen1, Danielle J. Vugts1, Wissam Beaino1

1 Amsterdam UMC, VU University, Radiology and Nuclear medicine, Radionuclide Center, Amsterdam, Netherlands

Introduction

Ex vivo tissue distribution is the gold standard for the evaluation and quantification of radiotracers accumulation in tumor xenografts. This technique uses high number of animals especially when it is performed at multiple time points. Positron emission tomography (PET) imaging can be used as an alternative technique that allows longitudinal evaluation of tracer distribution in the same animal at different time points. Our aim was to evaluate the performance of a nanoPET/CT and PET/MRI scanners and determine their potential in accurate quantification of tumor uptake.

Methods

NEMA NU 4-2008 phantoms [1] were filled with 20 MBq of 11C, 18F, 89Zr or 68Ga and scanned until decay in a Mediso nanoScan PET/CT and PET/MRI. N87 xenograft nu/nu mice (n=20) were injected IV with [18F]FDG (≥10MBq), kept 50 min under anesthesia and further imaged for 20 min in the PET/CT or PET/MRI camera. At end of scan, animals were sacrificed immediately and organs of interest were collected and measured in a gamma counter to determine %ID/g. Data were reconstructed using TeraTomo reconstruction algorithm with attenuation and scatter correction. Amide analysis software was used for phantom analysis. Analysis on tumor bearing mice was performed using Vivoquant software. Uptake in tumors and other organs was compared with ex vivo tissue distribution.

Results/Discussion

With the PET/CT, the highest recovery coefficient, thus the lowest Partial Volume Effect (PVE) was obtained for 18F with a recovery coefficient of 80% in the 5 mm cylinder ROI (Fig.1). 11C and 89Zr had a lower but comparable recovery with respectively 76% and 77%. Finally, 68Ga had the lowest recovery with 54% in the 5 mm ROI. The comparison of [18F]-FDG uptake (%ID/g) in the tumor obtained from the PET/CT image analysis and the biodistribution showed a correlation of R2=0.61. When the total counts (Bq) in the tumors were compared, the correlation was higher with an R2 of 0.94 (Fig.2). For the PET/MRI a similar correlation between the image quantification of [18F]-FDG uptake in the tumor and the biodistribution data was also observed (R2=0.68 for the %ID/g and 0.92 for the total uptake in Bq). The comparison of [18F]-FDG brain uptake obtained from the tissue distribution and the PET/CT showed a very good correlation with an R2=0.99 (%ID/g) and R2=0.91 (total counts in Bq).

Conclusions

In conclusion, the Mediso nanoScan PET/CT and PET/MRI showed very similar performance with a slightly better recovery and lower PVE for the PET/CT when attenuation and scatter correction was applied.  PET imaging can potentially be used as an alternative for the classic ex vivo tissue distribution to determine the tracer uptake in tumor xenografts. However, in depth analysis needs to be performed and compared with other PET isotopes.

Acknowledgment

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska–Curie grant agreement No 675417

References
[1] Szanda I, Mackewn J, Patay G, Major P, Sunassee K, Mullen GE, Nemeth G, Haemisch Y, Blower PJ, Marsden PK, National Electrical Manufacturers Association NU-4 Performance Evaluation of the PET Component of the NanoPET/CT Preclinical PET/CT Scanner, J. Nucl. Med. 52, 1741–1747 (2011).
Fig.1
Determination of PVE for 11C, 18F, 89Zr and 68Ga  with the PET/CT (A) and the PET/MRI (B).
Fig.2
Correlation between PET/CT assessed tumor uptake and ex vivo tissue distribution.
Keywords: preclinical imaging, PET/CT, PET/MRI, quantification, tissue distribution
BO-08-06

A new method for the quantification of PET radiotracer Arterial Input Function in the Rat Carotid Artery

Mailyn Perez-Liva1, Thulaciga Yoganathan1, 2, Mariana Aramburo1, Jacobo Cal4, Mickael Tanter3, Jean Provost3, Thomas Viel1, Bertrand Tavitian1, 2

1 Paris-Cardiovascular Research Center at HEGP, INSERM U970, Paris, France
2 Université Paris Descartes, Sorbonne Paris Cité, Faculté de Médecine, Paris, France
3 Physics for Medicine Institute, Inserm U1273, ESPCI Paris, CNRS, PSL Research university, Paris, France, Paris, France
4 Ion Beam Applications (IBA) Spain, PT Center Madrid, Pozuelo De Alarcon, Madrid, Spain

Introduction

Assessment of the arterial input function (AIF) in the carotid artery would be ideal for radiotracer modeling of cerebral PET. This is highly challenging in rodents as the internal diameter of the carotid artery is smaller than the spatial resolution of preclinical PET scanners [1]. Here, we used co-registration of Ultrafast ultrasound Doppler of the rat carotid artery with PET dynamic acquisition to enable segmentation-based correction of motion and partial volume effect for improved image radiotracer quantification.

Methods

Animal experiments were performed under ethical approval N°18-146. Measurements were performed using our hybrid imaging system PETRUS that combines an Ultrafast Ultrasound Imaging (UUI) scanner with a small animal PET/CT [2,3]. PET and UUI volumes were acquired simultaneously and precisely registered in the same spatial coordinates [2]. Adult male rats (n=4) were injected IV with 55 MBq of [18F]FDG and imaged during 60 minutes. Radioactivity concentration counted in serial blood samples drawn from the left common carotid were compared with image-derived AIF of the right common carotid, segmented from the Doppler volume using isodata unsupervised classification [4]. A local projection method [5] with motion deconvolution was applied to the dynamic PET sequences.

Results/Discussion

The maximum longitudinal and axial displacements of the carotid artery estimated with ultrasound were 2.1 and 0.7 mm, respectively. The maximal difference between ground-truth (i.e. obtained from blood sampling) values of the AIF and uncorrected image-derived AIF was 72%. Visually, both the peak value, amplitude and shape of the two AIFs differed considerably. After correction for PVE, maximum difference was reduced to 16 %, and after motion and PVE correction, it was reduced to 7 %. The fit in shape and amplitude between the two curves were also considerably improved.

Conclusions

Ultrafast Doppler co-registration with dynamic PET can be used for motion and PVE correction of small moving structures in preclinical PET. This allows for increased accuracy of image-derived AIF quantification in the rat carotid artery. Better quantification of the AIF should benefit PET radiotracer quantification in rodent brain.

AcknowledgmentThis project was funded in part by Plan Cancer (ASC16026HSA-C16026HS) and by LABEX WIFI (Laboratory of Excellence ANR-10-LABX-24) within the French program “Investments for the Future” under reference ANR-10-IDEX-0001-02 PSL In vivo imaging was performed at the Life Imaging Facility of Paris Descartes University (Plateforme Imageries du Vivant - PIV), supported by France Life Imaging (grant ANR-11-INBS-0006) and Infrastructures Biologie-Santé (IBISA). It was also funded in part by a CARPEM Siric grant (to BT)
References
[1] Moses, W. W. (2011). Fundamental limits of spatial resolution in PET. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 648, S236-S240.
[2] Provost J, et al., (2018). Simultaneous Positron Emission Tomography and Ultrafast Ultrasound for Hybrid Molecular, Anatomical, and Functional Imaging. Nature Biomed Eng., 2(2):85-94.
[3] Pérez-Liva M, et al., (2018). Performance evaluation of the PET component of a hybrid PET/CT-ultrafast ultrasound imaging instrument. Phys Med Biol. 2018 doi: 10.1088/1361-6560/aad946. PubMed PMID: 30091723.
[4] Magid, A., Rotman, S. R., & Weiss, A. M. (1990). Comments on Picture thresholding using an iterative selection method. IEEE transactions on systems, man, and cybernetics, 20(5), 1238-1239.
[5] Cal-González, et al, (2017). Impact of motion compensation and partial volume correction for 18F-NaF PET/CT imaging of coronary plaque. Physics in Medicine & Biology, 63(1), 015005.
Keywords: simultaneous PET/CT-UUI system, Image Derive Arterial Input Function, carotid artery, partial volume effect correction, motion correction
BO-08-07

Anatomically Informed Time-of-Flight PET Image Reconstruction with STIR Toolkit

Palak Wadhwa1, 2, Daniel Deidda3, Kris Thielemans4, William Hallett2, Roger Gunn2, David Buckley1, Charalampos Tsoumpas1, 2, 5

1 University of Leeds, Biomedical Imaging Science Department, Leeds, United Kingdom
2 Invicro, London, United Kingdom
3 National Physical Laboratory, Teddington, United Kingdom
4 University College London, Institute of Nuclear Medicine, London, United Kingdom
5 Icahn School of Medicine at Mount Sinai, Biomedical Engineering & Imaging Institute, Mount Sinai, United States of America

Introduction

PET/MR imaging modality is capable of highly sensitive and specific molecular and anatomical imaging. The combination of these modalities allows to exploit MR anatomical information to reconstruct PET images and improve the quantitative accuracy and whilst reducing the noise. The kernelised expectation maximisation (KEM) algorithm [1] has been recently implemented in Software for Tomographic Image Reconstruction (STIR, http://stir.sf.net) library [2]. This investigation aims at studying the performance of KEM reconstruction for TOF PET data from GE SIGNA PET/MR scanner to reconstruct datasets.

Methods

Any TOF-PET listmode uncompressed file can be extracted from the GE SIGNA PET scanner and histogrammed into TOF histograms using existing classes and utilities in STIR [3]. Normalisation and attenuation correction histograms are calculated using also classes and utilities implemented within STIR for GE SIGNA data [4]. Background events (i.e. randoms and scatter) are extracted within STIR space using custom utilities implemented during this work. A clinical dataset using STIR the TOF-KEM algorithm is reconstructed and demonstrated without the manufacturer’s software. Gaussian post-filtering with FWHM of 4 mm is applied to the reconstructed images. TOF-KEM and standard TOF-OSEM reconstructions are compared using standardised uptake value ratio (SUVR) and coefficient of variation (CoV).

Results/Discussion

The kernel parameters are optimised using TOF-PET data from the GE SIGNA PET/MR scanner. The kernel parameters that produce images with lower noise without degrading image resolution are selected. These parameters are: σm = 0.5 and σdm = 3. The reconstructed images displayed visual improvement with TOF-KEM over TOF-OSEM algorithm. SUVR comparisons were conducted by choosing region of interests (ROIs) in the liver and the lungs. The SUVR comparisons conducted using TOF-OSEM produce the value of 18.5 and the same comparison conducted with TOF-KEM demonstrates a rise of 1.62\% in SUVR. The ground truth of SUVR is assumed to be the value extracted from the image reconstructed using vendor's reconstruction software. The value calculated using TOF-OSEM with GE toolbox is 25.5. CoV was also calculated for ROI drawn in the liver with respect to the lung over PET images reconstructed for the first 6 iterations. TOF-KEM demonstrates improvement in the uniformity over TOF-OSEM.

Conclusions

This work demonstrates that the incorporation of the MR kernel in the reconstruction algorithm improves the uniformity and quantitative accuracy of the images over the standard algorithm using STIR toolkit. This work further broadens the capabilities of STIR and allows TOF-KEM image reconstructions for data extracted from GE SIGNA PET/MR. Future work aims at demonstrating the improved performance of TOF-KEM over TOF-OSEM for low count datasets.

Acknowledgment

This work is funded by the Medical Research Council (MR/M01746X/1). Dr Charalampos Tsoumpas is supported by a Royal Society Industry Fellowship (IF170011) and EPSRC (EP/P022200/1).

We would like to thank Floris Jansen for GE support. We would also like to thank Nikos Efthimiou, Ottavia Bertolli and Elise Emond. Furthermore, we are thankful Kristen Wangerin, Timothy Deller and Michel Tohme for GE toolbox support and for the supplied information. Finally, we are grateful to the CCP PET-MR network (EPSRC grant EP/M022587/1) for providing necessary support and resources to implement this work.

Ethics number 17/WM/0084 with permission from a clinical study performed at Invicro.

References
[1] Wang, G. and Qi, J., 2014, 'PET image reconstruction using kernel method.', IEEE transactions on medical imaging, 34(1), pp.61-71.
[2] Deidda, D., Karakatsanis, N.A., Robson, P.M., Efthimiou, N., Fayad, Z.A., Aykroyd, R.G. and Tsoumpas, C., 2018, 'Effect of PET-MR inconsistency in the kernel image reconstruction method.', IEEE Transactions on Radiation and Plasma Medical Sciences, 3(4), pp.400-409.
[3] Efthimiou, N., Emond, E., Wadhwa, P., Cawthorne, C., Tsoumpas, C. and Thielemans, K., 2019, 'Implementation and validation of time-of-flight PET image reconstruction module for listmode and sinogram projection data in the STIR library.', Physics in Medicine & Biology, 64(3), p.035004.
[4] Wadhwa, P., Thielemans, K., Efthimiou, N., Bertolli, O., Emond, E., Thomas, B.A., Tohme, M., Wangerin, K.A., Delso, G., Hallett, W. and Gunn, R.N., Tsoumpas, C., 2018, 'Implementation of Image Reconstruction for GE SIGNA PET/MR PET Data in the STIR Library.', 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference Proceedings (NSS/MIC), pp. 1-3.
Demonstration of TOF-OSEM and TOF-KEM reconstructions with STIR

This figure demonstrates TOF-OSEM and TOF-KEM reconstructions for lung fibrosis patients injected with experimental 18-F radiotracer. TOF-OSEM reconstructions are Gaussian post-filtered with FWHM of 4 mm. TOF-KEM reconstructions are reconstructed using optimised kernel parameters: σm = 0.5 and σdm = 3. TOF-OSEM and TOF-KEM reconstructions are conducted using STIR toolkit with 28 subsets for 2 iterations.

Keywords: TOF-PET, PET/MR, KEM, TOF-KEM