WO2018091265A1

WO2018091265A1 - Phantom, simulation installation and method for preparing a detection system for detecting a lesion under influence of a respiratory movement

Info

Publication number: WO2018091265A1
Application number: PCT/EP2017/077898
Authority: WO
Inventors: Laura DEVOS; Pascal Deleporte; Maximilien Vermandel
Original assignee: INSERM (Institut National de la Santé et de la Recherche Médicale); Universite De Lille 2 Droit Et Sante; Centre Hospitalier Regional Et Universitaire De Lille (Chru)
Priority date: 2016-11-17
Filing date: 2017-10-31
Publication date: 2018-05-24

Abstract

Phantom (10) comprising: - a body (11) having an inner space (12), - a driving system (50) simulating a respiratory movement, - at least one insert (35) configured to simulate a lesion, - at least one membrane (30) configured to simulate tissues at an interface between two adjacent anatomical structures, the membrane (30) being flexible and arranged within the body (11) to divide the inner space (12) in two separate compartments (12a, 12b) intended to be filled with respective mediums each simulating one of the anatomical structures, at least one of the compartments (12a, 12b) receiving said at least one insert (35).

Description

Phantom, simulation installation and method for preparing a detection system for detecting a lesion under influence of a respiratory movement

The invention relates to a phantom, a simulation installation and a method for preparing a detection system for detecting a lesion under influence of a respiratory movement.

Although not limited thereto, the invention specifically applies to the study of tumorous lesion of the lung and/or the liver.

During the extension checkups, the liver is a common site of metastases not only in colorectal, pancreatic or neuroendocrine cancer but also in an increasing number of other neoplasias (Gandy RC, Bergamin PA, Haghighi KS, Hepatic resection of non-colorectal non-endocrine liver metastases, ANZ J Surg. March 31, 2016). The detection of l iver metastatic sites is crucial in caring the patient because their presence in a limited number and l ittle scattered is accessible to surgical treatment ( Khatri VP. Petrel l i NJ, Belghiti J, Extending the frontiers of surgical therapy for hepatic colorectal metastases: is there a limit?, J Clin Oncol Off J Am Soc Clin Oncol., November 20, 2005; 23(33):8490 ^"9) potentially curative and increasing overall survival of the patient (de Ridder JAM, van der Stok EP, Mekenkamp LJ, Wiering B, Koopman M, Punt CJA, et al., Management of liver metastases in colorectal cancer patients: A retrospective case-control study of systemic therapy versus liver resection, Eur J Cancer Oxf Engl 1990, May 2016; 59: 13 -21).

Positron emission tomography ( PET ) imaging with 1 8 - fl u orodeo x y g 1 uc o se ( 1 8- FDG) coupled to computed tomography (CT) imaging has become a key examination in cl inical routine practice for many years. Thanks to the information combining metabolic activity of the tissues obtained by PET imaging and morphological precision of data obtained by the CT imaging, it is increasingly used in oncology as diagnostic aid for characterizing suspicious lesions with morphologic imaging, for detecting neoplastic recurrence and increasingly for evaluating therapeutic response (d^'Amico A. Review of clinical practice utility of positron emission tomography with 18F-fluorodeoxy glucose in assessing tumour response to therapy, Radiol Med (Torino), 2015; 120(4):345 -51).

1 8-FDG PET also has an important place in the pre-surgical assessment of liver metastasis. The sensitivity and specificity of 1 8-FDG PET for the detection of hepatic lesions are 80% and 92% respectively. In addition, it changes the support in a little more than 25 % of cases (Wiering B, Krabbe PFM, Jager GJ, Oyen WJG, Ruers TJ M. The impact of fluor-18-deoxy glucose-positron emission tomography in the management of colorectal liver metastases, Cancer. December 15, 2005; 104( 12 ):2658 - 70 ). Currently, the index of quantification that is commonly used at most is the standard uptake value (SUV) (Hunter GJ, Hamberg LM, Alpert NM, Choi NC, Fischman AJ, Simplified measurement of deoxyglucose utilization rate, J Nucl Med Off Publ Soc Nucl Med., June 1996; 37(6):950 "5) and that is known at most by cl inicians is the SUV max. corresponding to the ma imum concentration within a voxel belonging to the analyzed lesion. Nowadays, other quantification parameters are increasingly used, particularly metabolic volumes. However these parameters are subject to numerous variations as biological data, tomographic reconstruction data, correction of the PSF, attenuation correction, distribution, partial volume effect and respiratory movement ( Buvat I. Les limites du SUV, Http://www.em- Premiumcomdoc-Distantuniv-Lille2frdatarevues092812580031000407000459 [Internet], 6 aout 2007).

In particular, respiratory movement is a source of artifact and therefore of false negative in PET imaging. Indeed, hepatic dome and pulmonary bases are very mobile zones during the respiratory movement driven by the diaphragm and displacement may reach several centimeters (Callahan J. Kron T, Schneider-Kolsky M. Hicks RJ. The clinical significance and management of lesion motion due to respiration during PET/CT scanning. Cancer Imaging, December 28. 201 1 ; l l(l):224-36), then causing more errors in PET images (Nehmeh SA. Erdi YE. Ling CC, Rosenzweig KE, Squire OD. Braban LE, et al.. Effect of respiratory gating on reducing lung motion artifacts in PET imaging of lung cancer, Med Phys. March 2002; 29(3):366 "71). First, a bad fusion betw een CT and PET images results in a distorted attenuation correction and thus in errors in the quantification toward a decrease of SUVs. This decrease causes false negative in PET imaging for the detection of secondary l iver lesions. Second, a spreading of hypermetabolic foci occurs, causing an increase in measured metabolic volumes which can result, for example, in an increase of doses accumulated during radiotherapy treatment (Wijsman R. Grootjans W. Troost EG, van der Heijden EH, Visser EP, de Geus-Oei L-F, et al., Evaluating the use of optimally respiratory gated 18F-FDG-PET in target volume delineation and its influence on radiation doses to the organs at risk in non-small-cell lung cancer patients, Nucl Med Commun. January 2016; 37(l):66-73).

To counterbalance these artifacts due to respiratory movement, there are now several PET and/or CT data acquisition techniques coupled to a measuring device for monitoring the respiratory cycle. Nowadays, two methods of compensation are routinely available, where the respiratory cycle is divided into n intervals or bins either by frequency analysis or by amplitude analysis (Pepin A. Daouk J, Bailly P. Hapdey S, Meyer M-E. Management of respiratory motion in PET/computed tomography: the state of the art, Nucl Med Commun. February 2014; 35(2): 113 -22).

In order to evaluate and to set detection systems combining an imaging device according to a modality such as PET or CT and a measuring device, phantoms are implemented. Known phantoms, such as that exploited by the company Modus Medical Devices Inc. under the designations QUASAR ™ Respiratory Motion Phantom and QUASAR™ Cylindrical Respiratory Motion Phantom, are of the type comprising:

- a body having an inner space extending along a longitudinal axis,

- a driving system simulating a respiratory movement within the inner space, and

- at least one insert configured to simulate a lesion, the insert being movably mounted within the inner space of the body.

A phantom of the aforementioned type further comprising at least one membrane configured to simulate tissues at an interface between two adjacent anatomical structures is disclosed in US 2008/298540.

However, there is a need for a phantom that enables images to be acquired in conditions that match more closely that of the clinical reality.

The invention aims to fulfil the above mentioned need.

To this end, according to a first aspect, the invention provides a phantom comprising:

- a body having an inner space extending along a longitudinal axis,

- a driving system simulating a respiratory movement within the inner space, and - at least one insert configured to simulate a lesion, the insert being movably mounted within the inner space of the body,

- at least one membrane configured to simulate tissues at an interface between two adjacent anatomical structures, the membrane being flexible and arranged within the body to divide the inner space in two separate compartments intended to be filled with respective mediums each simulating one of the anatomical structures, at least one of the compartments receiving said at least one insert,

wherein the insert is attached to the driving system so as to be moved by the driving system. Hence, the phantom of the invention enables a lesion subjected to respiratory movement and located at the vicinity of an interface between two anatomical structures, such as the lungs and the liver, to be simulated. Such localization, which is often met in clinical reality, renders the lesion difficult to detect. With the phantom of the invention, the detection system combining the imaging device with the measuring device can be appropriately prepared and set, especially by evaluating different image acquisition parameters of the imaging device and/or by optimizing image acquisition and/or reconstruction protocols. Moreover, thanks to the attachment of the insert to the driving system, the movement of the insert can be accurately controlled and the configuration of the phantom can be modulated in a simple manner.

The body may include a lateral wall around the longitudinal axis, the membrane being secured to the lateral wall of the body, the membrane extending transversally with respect to the longitudinal axis and the compartments being adjacent along the longitudinal axis. These provisions make it possible to simulate a displacement of the lesion in a cranio- caudal direction matching an orientation of a patient laying on an examination table.

The lateral wall may have opposite end edges and the body further includes at least one removable cover removably mounted to the one of the end edges of the lateral wall, the driving system being mounted to the removable cover.

The driving system may comprise an arm driven to reciprocate back and forth in translation along the longitudinal axis, the insert being attached to the arm so as to be movable in translation along the longitudinal axis in a back and forth movement along a stroke about a median position with respect to the membrane.

The median position of the insert within the compartment may be adjustable.

The insert may be mounted on a support connected to the driving system.

The insert may be removably mounted within the body.

Said at least one insert may comprise a plurality of inserts of different sizes.

The insert may include a hollow envelop intended to be filled with a medium that differs from the medium of which the compartment receiving said insert is filled.

The body may further include a support surface on which the phantom rests in use, the support surface extending parallel to the longitudinal axis.

The body may be provided with at least one aperture opening in each compartment, the aperture presenting an opened state in which a passage for the medium is allowed, and a closed state in which a passage for the medium is prevented. According a second aspect, the invention provides a simulation installation comprising a phantom as previously defined and a detection system for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the detection system comprising a measuring device configured to monitor a respiratory movement of the patient and an imaging device configured to acquire an image of the lesion, the imaging device being synchronised with the measuring device.

The imaging device may be a positron emission tomography (PET) imaging device.

According a third aspect, the invention provides a method for preparing a detection system for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the method comprising the steps of:

- providing a measuring device configured to monitor a respiratory movement of the patient and an imaging device configured to acquire an image of the lesion, synchronising the imaging device with the measuring device,

- providing a phantom according to any of claims 1 to 11, the phantom comprising: a body having an inner space extending along a longitudinal axis,

at least one membrane configured to simulate tissues at an interface between two adjacent anatomical structures, the membrane being flexible and arranged within the body to divide the inner space in two separate compartments filled with respective mediums each simulating one of the anatomical structures,

a driving system, and

at least one insert configured to simulate a lesion, the insert being attached to the driving system so as to be moved within at least one of the compartments of the inner space of the body by the driving system,

- actuating the driving system of the phantom to simulate a respiratory movement within the inner space of the body of the phantom,

- monitoring the respiratory movement simulated by the driving system with the measuring device and acquiring an image of the insert with the imaging device.

The method may further comprise the step of defining respiratory phases of the respiratory movement and acquiring an image of the insert during each respiratory phase.

Other objects and advantages of the invention will emerge from the following disclosure of a particular embodiment of the invention given as a non-limitative example, the disclosure being made in reference to the enclosed drawings in which:

- Figure 1 is a perspective view from a first direction of a phantom according to an embodiment of the invention, the phantom is implemented for preparing a detection system for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the phantom comprises a body having an inner space divided in two separate compartments by a membrane, the compartments being intended to be filled with respective mediums each simulating an anatomical structure and the membrane simulating tissues at an interface between the anatomical structures, the phantom further comprising several inserts simulating lesions mounted in one of the compartments and connected to a driving system simulating a respiratory movement,

- Figure 2 is a perspective view from a second direction of the phantom of Figure 1,

- Figure 3 is a side view of the phantom of Figure 1, illustrating the driving system and the inserts moved to a forward position of a stroke of the inserts in a back and forth movement along a longitudinal axis of the body, the inserts contacting the membrane and deforming it,

- Figure 4 is a side view of the phantom of Figure 1, illustrating the driving system and the inserts moved to a backward position of the stroke of the inserts, the inserts being spaced apart from the membrane,

- Figure 5 is a side view of the phantom of Figure 1, illustrating a following system coupled to the driving system to enable the respiratory motion simulated by the driving system to be monitored,

- Figure 6 is a partial perspective view of a simulation installation comprising the phantom of Figure 1 and a detection system for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the detection system comprising a measuring device configured to monitor a respiratory movement of the patient and an imaging device configured to acquire an image of the lesion,

- Figures 7 to 12 are graphs related to an experimentation implementing the phantom of Figure 1 to compare performances of three acquisition protocols with PET and

CT 4D imaging, to compare reconstructed data from PET 4D imaging with either data from CT 413 imaging or CT 3D imaging, as well as to analyze variations of three quantitative parameters depending on a position of insert relative to the membrane, Figure 7 being a graph illustrating a variation of covering coefficients with respect to a volume of the inserts for a contrast 1/5, Figure 8 being a graph illustrating a variation of covering coefficients with respect to a volume of the inserts for a contrast 1/10, Figure 9 being a graph illustrating measurement errors of quantitative parameters for each 4D protocol at a contrast 1/10, Figure 10 being a graph illustrating measurement errors of quantitative parameters for each 4D protocol at a contrast 1/5, Figure 11 being a graph illustrating a mean of measurement errors regarding a volume of 50% as a function of the volume of the inserts, Figure 12 being a graph comparing three quantitative parameters as a function of the position of the insert with respect to the membrane.

On the Figures, the same reference numbers refer to the same or similar elements. Figures 1 to 5 represent an embodiment of a phantom 10 for preparing a detection system 1 for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement. In particular, the phantom 10 may be used to choose and evaluate components and settings of the detection system 1 so that accurate and reliable detection on the patient himself can be performed subsequently.

The phantom 10 represented on the drawings is notably but not exclusively implemented in the study of one or several tumorous lesions at the vicinity of tissues at an interface between two anatomical structures subjected to the respiratory movement. In particular, the phantom 10 may be implemented to simulate one or several hepatic and/or pulmonary tumorous lesions at the vicinity of the diaphragm between the lungs and the liver.

On Figures 1 and 2, the phantom 10 comprises a body 11 having an inner space 12 extending along a longitudinal axis L. As it will become apparent from the following, the body 11 is made of one or more materials that are compatible with the detection system and especially with at least one modality of a medical imaging device, such as a Positron emission tomography (PET) imaging device 2 and/or a computed tomography (CT) imaging device.

The body 11 includes a lateral wall 13 around the longitudinal axis L. The lateral wall 13 has an internal surface which delimits the inner space 12, an external surface opposite the internal surface and opposite end edges spaced apart from each other along the longitudinal axis L. On the Figures, the lateral wall 13. made of a transparent or translucent plastic material, is cylindrical of circular cross-section. In particular, the lateral wall 13 is made of two halves 13a, 13b. Each half 13a, 13b of the lateral wall 13 is provided at respective ends with first 14 and second 15 end flanges extending radially with respect to the longitudinal axis L. The first end flange 14 of one of the halves 13b is secured, for example bolted, to the second end flange 15 of the other half 13a. Each half 13a, 13b is further provided with an aperture 16 arranged substantially centrally with respect to the ends and opening in the inner space 12. Each aperture 16 comprises a collar 17, annular around an axis extending radially with respect to the longitudinal axis L, and a cap 18 removably mounted on the collar 17 so that the aperture 16 presents a closed state when the cap 18 is mounted on the collar 17 so as to prevent a passage through it, and an opened state when the cap 18 is removed from the collar 17 so as to allow a passage through it.

The body 11 further includes two removable covers 21 removably mounted respectively on the end edges of the lateral wall 13 to close the inner space 12 along the longitudinal axis L. In particular, each removable cover 21 comprises a transverse wall with respect to the longitudinal axis L and is secured, for example bolted, to one of the free end flanges 14, 15 of the halves 13a, 13b.

The body 11 further includes a support in the form of two hoops 22 mounted respectively on the external surface of the halves 13a, 13b of the lateral wall 13, at a distance from each other. The hoops 22 have feet 23 defining a support surface extending parallel to the longitudinal axis L. Therefore, when the phantom 10 rests on its support surface in use, the longitudinal axis L is horizontal.

The phantom 10 further comprises a membrane 30 arranged within the inner space 12 of the body 11. The membrane 30 is flexible and, as for the body 11, preferably made of one or more materials that are compatible with the modality of the medical imaging device 2 of the detection system 1 . For example, the membrane 30 is made of silicone. In the represented embodiment, the membrane 30 is secured to the lateral wall 13 of the body 1 1 so as to extend transversally with respect to the longitudinal axis L. In particular, a peripheral edge of the membrane 30 is tightened between the end flanges 14, 15 secured to each other of the halves 13a, 13b of the lateral wall 13.

With such arrangement, the membrane 30 divides the inner space in two separate compartments 12a, 12b adjacent along the longitudinal axis L. In the represented embodiment, the compartments 12a, 12b are each delimited by the internal surface of one of the halves 13a, 13b of the lateral wall 13. The compartments 12a, 12b can be filled though their apertures 16 in the opened state with respective mediums each simulating an anatomical structure, the membrane 30 simulating the tissues at the interface between the two adjacent anatomical structures. Each of the mediums can be of any kind to simulate an anatomical structure. In particular, each medium may be include any kind of fluid, possibly with solid particles distributed therein, having suitable characteristics especially in terms of density or viscosity. The medium of one of the compartments 12a, 12b may also include any suitable agent and, in particular, a contrast agent compatible with the modality of the medical imaging device 2 of the detection system 1 to provide a contrast that is different from that of the other compartment 12a, 12b. For example, the medium of one of the compartments 12a, 12b may be air to simulate lungs and the medium of the other compartment 12a, 12b may be water with 18-fluorodeoxyglucose ( 18-FDG ) as contrast agent to simulate liver, the membrane 30 simulating the diaphragm. The mediums can be subsequently removed from their respective compartments 12a, 12b and replaced by other mediums.

Four inserts 35, of different sizes and configured to simulate a lesion, are mounted in a removable manner within the inner space 12 of the body 11 and more specifically within a first 12a of the compartments. Each insert 35 includes a spherical hollow envelop that can be filled with a medium that differs from the medium of which the first compartment 12a receiving the insert 35 is filled. As for the body 11 and the membrane 30, the inserts 35 are made of a material compatible with the modality of the medical imaging device 2 of the detection system 1.

As shown on Figures 3 and 4, the inserts 35 are mounted on a support 40 connected to a driving system 50 which simulates a respiratory movement within the inner space 12. In the represented embodiment, the driving system 50 comprises an actuator 51 and a transmission mechanism 53 mounted to the removable cover 21 attached to the first compartment 12a. The actuator 51 is arranged in a case 54 outside the inner space 12 of the body 11 and comprises an electrical motor connected to an autonomous power supply source, such as a battery. The actuator 51 has a shaft 52 driven in rotation about a rotation axis perpendicular to the longitudinal axis L. The transmission mechanism 53 comprises a driving cam 55 fitted on the shaft 52 of the actuator 51, and an arm 56 slidably mounted through the removable cover 21 along the longitudinal direction L so that an outer portion is arranged outside the inner space 12 of the body 11 and an inner portion is arranged inside the inner space 12 of the body 11. The outer portion of the arm 56 has two rollers 57 arranged within a slot 58 extending parallel to the longitudinal axis L, on either sides of the driving cam 55.

The support 40 comprises a plate 41 attached to a free end of the inner portion of the arm 56 and four rods 42 extending each from the plate 41 parallel to the longitudinal axis L towards a free end on which one of the insert 35 is attached.

Therefore, as the shaft 52 rotates upon actuation of the actuator 51, the driving cam 55 moves the rollers 57 along the slot 58. The arm 56 reciprocates back and forth along the longitudinal axis L thereby moving the inserts 35 in translation along a stroke between a forward position shown on Figure 3 and a backward position shown on Figure 4 equally spaced from a median position with respect to the membrane 30. On Figure 3, in the forward position, the inserts 35 contact the membrane 30 and deform it towards the second compartment 12b. On Figure 4, in the backward position, the inserts 35 are spaced apart from the membrane 30. The median position of each insert 35 within the first compartment 12a may be adjustable, for example by adjusting a length of the rods 42.

As shown on Figure 5, the phantom 10 further comprises a following system 60 coupled to the driving system 50 to enable the respiratory motion simulated by the driving system 50 to be monitored. The following system 60 comprises a following cam 61 fitted on the shaft 52 of the actuator 51 and arranged within an opening of a plate 62 slidably mounted on the case 54 along a direction perpendicular to the rotation axis and the longitudinal axis L. The plate 62 may reciprocate up and down in accordance with the back and forth movement of the inserts 35. The plate 62 may comprise a platform 63 on which a locating element may be mounted to be detected by a measuring device 3 of the detection system 1 configured to monitor a respiratory movement of the patient.

An example of a detection system 1 for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement is shown on Figure 6. The measuring device 3 of the detection system 1 may be chosen among:

- a Real Time Management (RTM) system of the type exploited by the company Varian based on an infrared camera able to detect and follow displacements of a locating element first placed on the platform 63 of the plate 62 of the following system 60 and second placed on the chest of the patient,

- a system of the type exploited by the company Anzai or of the type exploited by the company Siemens Medical Solution under the denomination HD-Chest® comprising a belt measuring variations in the circumference of the chest of the patient by impedance measurement or pressure measurement.

The imaging device configured to acquire an image of the lesion of the detection system 1 may be a positron emission tomography (PET) imaging device 2. The imaging device 2 is synchronised with the measuring device 3.

A method for preparing the detection system 1 disclosed previously is now disclosed.

The PET imaging device 2 and the measuring device 3 are chosen and synchronised. Before mounting the removable cover 21 together with the inserts 35 and the driving system 50 onto the lateral wall 13 of the body 11, the inserts 35 are filled with the medium. The removable cover 21 is then mounted to the lateral wall 13 of the body 11 and the compartments 12a, 12b can be filled through the apertures 16 in the opened state with the appropriate mediums. The mediums preferably include different contrast agents or a contrast agent at different concentration to provide different contrasts on the resulting image. The apertures 16 are placed in their closed state to prevent a passage of the medium and the phantom 10 is placed with the following system 60 facing the measuring device 3 and the inserts 35 arranged within the PET imaging device 2.

The driving system 35 may then be actuated to simulate the respiratory movement within the inner space 12 of the body 11 of the phantom 10 and to move the lesions in accordance with this respiratory movement.

The respiratory movement simulated by the driving system 35 is monitored by the measuring device 3 and one or several images of the inserts 35 are acquired by the PET imaging device 2.

Different experimentation protocols may be studied. For example, respiratory phases of the respiratory movement may be defined so as acquire an image of each insert 35 during each respiratory phase. Also, different settings of the detection system 1 may be tested, in particular, to evaluate different image acquisition parameters of the imaging device 2 to optimize image acquisition and/or reconstruction protocols.

The invention has been disclosed in relation to a phantom 10 dedicated to a simulation of one or several hepatic and/or pulmonary tumorous lesions at the vicinity of the diaphragm between the lungs and the liver. The invention is, however, not limited thereto and could be implemented to simulate any other kind of lesion at the vicinity of tissues at an interface between two anatomical structures subjected to the respiratory movement. Any other arrangement of one or several inserts in one or several compartments, separated by pairs by one or several membrane, could be appropriately provided. Also, the respiratory movement could be imposed in any other appropriate manner by a driving system acting on the insert or any other suitable component of the phantom and especially the membrane or the medium of which the compartment receiving the insert is filled.

In relation with Figures 7 to 12, an experimentation implementing the phantom 10 is now disclosed as a non-limitative example of implementation of the phantom 10. The experimentation is performed to compare performances of three acquisition protocols with 4D PET and CT imaging, to compare reconstructed data from 4D PET imaging with either data from 4D CT imaging or 3D CT imaging, as well as to analyze variations of three quantitative parameters depending on a position of the insert 35 relative to the membrane 30. Material and methods

Description of the phantom

In the experimentation, the phantom 10 of the type previously disclosed has a total length of 561 mm and a maximum diameter of 188 mm.

The second compartment 12b has a volume V2 of 2 L, a length of 210 mm and a diameter of I 10 mm. It simulates the lung parenchyma. The first compartment 1 2a attached to the driv ing system 50 has a volume V 1 of 1 .5 L. a length of 158 mm and a diameter of 1 1 0 mm. It simulates the l iver parenchyma. The inserts 35 each simulate a lesion of liver dome, located near the membrane 30 separating the two compartments 1 2a. 12b and constituting an outl ine of the diaphragm. The driving system 50 located outside inner space of the body 1 1 containing a radioactive tracer allows to transcribe a repetitive back and forth movement of translation of the inserts 35 mimicking a regular respiratory movement over a stroke of approximately 15 mm. This movement creates a deformation of the membrane 30 towards the second compartment 12b, simulating l iver dome movements and thus those of possible hypermetabol ic lesions thereof. Attached to the top portion of the driving system 50 is the following system 60 performing an up and down movement in synergy with the inserts 35. simulating the variation of the amplitude of the abdominal and chest cavity. This following system 60 is compatible with a casing comprising a sensor for an infrared camera, if the R I^'M system is used, but also with a respiratory motion detection belt as used by Siemens Medical .Solution ( HD-Chest ®).

The actuator of the driving system is supplied by a 9 V battery and is connected to a control unit configured to actuate it.

PET Acquisition

For the first acquisitions, four inserts 35 of different volumes were used: S 1 = 8 mL. S2 = 3.05 ml. S3 = 1 . 17 ml and S4 = 17. 1 5 ml..

Two different contrasts between background (BDF) and the activity of the spheres were made. In a contrast 1/10, 105 MBq were placed in the volume V 1 of the first compartment 12a. 6.56 MBq were placed into the volume S I of the insert. 2.50 ! MBq were placed into the volume S2 of the insert and 0.9594 MBq were placed into the volume S3 of the insert. For the volume S4 of the insert requiring another series of acquisition. 145 MBq were placed into the volume V I of the first compartment 12a and 7.6 MBq were placed into the volume S4 of the insert. For a contrast 1/5, 1 8 1 .7 MBq were placed in the volume V 1 of the first compartment 12a, and respectively 4.8 MBq, 1 .83 MBq and 0.702 MBq were placed in the volumes S I, S2 and S3 of the inserts and 130 Bq and 7.35 MBq in the volume S4 of the insert. The volume V2 of the second compartment 12b remained empty only filled with air. Alternatively, it could be filled polystyrene micro- balls impregnated with 1 8-FDG. By knowing each activity, a theoretical maximum SUV of each compartment 12a, 12b and of each insert 35 is known.

The acquisitions were made on the PET imaging device of the type Discovery 1 1 D RX 16 exploited by General Electric Medical System and which consists in four rings of 70 LYSO detector blocks ( Lutetium. Orthosi l icate yttrium ). For each contrast, a standard acquisition of two pitches of 2 minutes followed by a traditional CT acquisition was performed. Then for acquisitions synchronized with respiratory movements performed on the same day as standard acquisitions, the Real Time Management (RTM) system of Varian was used. The experimentation began by acquiring 4D CT followed immediately by acquiring 4D PET constituted by a pitch centered on the junction of the compartments 1 2a. 12b and performed simultaneously with the measurement of respiratory movements.

For each of the contrasts, the CT data and 4D PET were cut into 4, 5 and 6 phases each lasting 2 minutes. The inserts of volumes S I, S2 and S3 were placed together and the insert of volume S4 was alone during acquisitions. Two methods of reconstruction were used. For 4D synchronized acquisitions, each phase was reconstructed w ith the corresponding CT phase. For 3D acquisitions, each bin of each synchronized PET protocol was reconstructed a posteriori with a conventional scanner scan before the respiratory gating. 4D PET / 4D CT data and 4D PET / 3D CT data were then obtained.

For the reconstructions of data, parameters of clinical routine acquisitions were used, namely an iterative reconstruction algorithm of the type OS EM 3D with 2 iterations and 2 1 subassembly and a post filtering of Gaussian typ of 6 mm . Attenuation corrections, scattered and random coincidences were applied.

Analysis of data

Once the 4D PET images have been reconstructed by acquiring corresponding 4D

CT. they were analyzed on a 4.6 Advantage Workstation console exploited by the company GE Medical Systems. Covering coefficients were first calculated for determining an SUV max. a mean SUV at 50 % isocontour and a mean SUV at A50 isocontour. The covering coefficients are defined as the ratio between the concentration of measured activity and the concentration of theoretical activity (calculated from the activity put into each insert taking into account for radioactive decrease ). These values were obtained with a manually taken region of interest (ROI) of size suitable for each insert.

Then for each insert, at each bin of each phase and for the two contrasts, the SUV max were determined. Two segmentation methods were carried out for measuring their metabolic volume at 50 % isocontour of the SUV max and by adaptive thresholding A50.

It is calculated as follows:

A50 = [(SUV max + BDF) 12] I SUV max.

For each parameter evaluated on 4D PET data, an error between the theoretical value and the measured value was calculated. For example, for the SUV max, the error obtained by the following formula:

Error SUV max = (measured SUV - theoretical SUV) / theoretical SUV max.

Each 4D PET protocol having a different phase number the respiratory cycle (4 phases. 5 phases or 6 phases ), the mean and standard deviation of their errors between theory and reality in each bin of the respiratory cycle were calculated to compare each series of acquisition.

During 4D PET CT acquisitions, inserts had a regular translational movement over a stroke of about 1 5 mm transmitted by the driving system. For each bin of each acquisition protocol, the different inserts were at a distance from the membrane that was measured on the CT information in coronal section.

Acquisitions were thus separated into two groups:

- those in which the insert was at a distance less than its radius from the membrane,

- those in which the insert was remote from the membrane of a distance greater than its radius.

Statistical analysis

To compare the three acquisition 4D PET CT protocols according to their performances to the extent of SUV max and metabolic volumes according both thresholding methods previously described. Kruskal Wallis test was achieved.

To compare the two methods of reconstruction of PET images according to 3D or

4D TDM, a variance analysis ( ANOVA ) was performed to compare the measurement errors of the three quantitative parameters in the two contrasts.

Finally to compare the two groups identified above depending on the position of the insert relative to the membrane, the data were compared with a Mann Whitney test. These tests were performed on the XLSTAT software and the results are given with 95% confidence interval.

Results

Validation of the phantom

At first, an evolution of the covering coefficients as a function of the volume of each insert for each contrast (1/5 and 1/10) was studied for the max SUV, the SUV mean at 50 % isocontour and the SUV mean at A50 isocontour. Only data related to bin 2 of 4D PET CT protocol in 6 phases are presented, since the other acquisitions follow this trend and lead to similar concl usions.

Figure 7 is a graph illustrating a variation of covering coefficients with respect to a volume of the inserts for a contrast 1/5 and Figure 8 is a graph illustrating a variation of covering coefficients with respect to a volume of the inserts for a contrast 1/10.

It can be observed that covering coefficient increase with the volume of the inserts, whatever the quantitative parameter is. There is, however, an overestimate of the values of the SUV ma of about 15 % to 20 % compared to other SUVs. Moreover, it can be noted that the covering coefficients are higher when the contrast between the lesions and the background is better. In any case, it is that al l covering coefficients are in the range between about 70 % to 120 %, which is within the scope of the legislation NEMA NU 2 2012 regarding phantoms used for control quality of PET machines.

Comparison of the 3 protocoles of 4D TEP CT

For each protocol 4D PET CT. means of measurement errors for each insert, in both contrasts, are compared and no statistically significant difference has been found, would it be in the measurement of SUV max : p = 0. 144. the measurement of the metabol ic volume 50 % isocontour p = 0.2 1 8 or in the measurement of metabolic volume A 50 threshold p =

0.335.

Errors are shown for each contrast in Figures 9 and 10, Figure 9 illustrating measurement errors of quantitative parameters for each 4D protocol at a contrast 1/10 and Figure 10 illustrating measurement errors of quantitative parameters for each 4D protocol at a contrast 1/5. Comparison of data 4D PET reconstruction method according to 4D CT vs 3D CT All means of measurement errors per insert, per protocol and per contrast are summarized in Tables 1 , 2 and 3 below respectively for the max SUV and two metabolic volumes. Multivariate analysis between the two PET reconstruction methods, synchronized with respiratory movement, shows a significant difference for the measurement of SUV ma p = 0.001, in favor of reconstruction with the 4D CT. As regards metabolic volumes, there is no significant difference between the two reconstruction methods, p = 0. 210 for metabolic volume at 50 % isocontour and p = 0.700 for metabolic volume at A50 threshold.

Table 1 - Mean of measurement errors for SUV max

Mean Mean Mean

Mean Mean Mean Mean

error error error error error error error

Volume contrast 5 contrast contrast 5 contrast 5 contrast 5 contrast 5 contrast 5

5 Bins 5 4 Bins Standard 6 Bins 4D 6 Bins 3D 5 Bins 4D 4 Bins 4D

3D 3D PET

1,17 0,19 0,15 0,23 0,32 0,27 0,35 0,15

3,05 0,11 0,19 0,09 0,14 0,1 0,07 0,23

8 0,15 0,19 0,13 0,17 0,12 0,18 0,18

17,15 0,06 0,29 0,05 0,22 0,06 0,23 0,23

Table 2 - Mean of measurement errors for metabolic volume at 50 % isocontour

Mean

Mean Mean

Mean Mean Mean Mean error error error

error error error error contrast

Volume contrast 5 contrast

contrast 5 contrast 5 contrast 5 contrast 5 5

4 Bins 5 4 Bins

6 Bins 4D 6 Bins 3D 5 Bins 4D 5 Bins 3D Standard

4D 3D

PET

1,17 0,17 0,09 0,10 0,75 0,31 0,75 0,33

3,05 0,44 0,36 0,43 0,46 0,41 0,46 0,46

8 0,05 0,03 0,04 0,02 0,03 0,1 0,07

17,15 0,11 0,16 0,11 0,11 0,12 0,14 0,13

Table 3 - Mean of measurement errors for metabolic volume at A50 threshold

Figure 11 further illustrates a mean of measurement errors regarding a volume of 50% as a function of the volume of the inserts. Measurement errors of metabolic volumes are statistically higher for the insert of 3.0 ml. (p <0.0001) regardless of the thresholding methods, regardless of the method of reconstruction of the PET data or the contrast in the image.

Study of the position of the sphere relative to the diaphragm on performance 4D PET CT

Figure 12 is a graph comparing three quantitative parameters as a function of the position of the insert with respect to the membrane.

The comparison between the two groups showed no significant difference between measurements of SUV max and metabolic volume at isocontour 50%, but there is a trend to means of errors of less important when the insert is remote from the diaphragm: p = 0. 16 and p = 0.088 respectively.

However, we found a significant difference in measurement of the metabolic volume at threshold 50: p = 0.043 with measurements closer to the theory in the group where the insert was farther from the diaphragm.

Discussion

First, a great similarity in the evolution of covering coefficients mainly of SUV max is found between two different acquisitions of different contrasts. The phantom has all the prerequisites for monitoring different targets of which several parameters can be varied.

No significant difference between the th ee protocols 4D TEP / 4D CT can be found, suggesting a minimal loss of information would it be for the measurement of the max SUV or for both methods of metabolic volume measurement. Then it could be considered using the fastest protocol (knowing that a bin corresponds to 2 minutes of examination) including cases of algetic patient, or for example claustrophobic, without altering the performance of the examination.

Acquisitions reconstructed w ith the 4D CT decreased significantly the measurement errors of the SUV max. This result is consistent with several studies carried out on phantoms (Vines DC. Keller H. Hoisak JDP, Breen SL. Quantitative PET Comparing Gated with Nongated Acquisitions Using a NEMA Phantom with Respiratory- Simulated Motion, J Nucl Med Technol. December 1, 2007; 35(4):246-51 ; Miwa K, Wagatsuma K, Umeda T, Miyaji N, Murata T, Osawa A et al., Improvement of quantitative accuracy using phase -based respiratory-gated PET/CT in phantom and clinical studies, Nihon Hoshasen Gijutsu Gakkai Zasshi. Novembre 2014; 70(11): 1235 -42) and patients (Farid K, Poullias X, Alifano M, Regnard J-F, Servois V, Caillat-Vigneron N et al., Respiratory-gated imaging in metabolic evaluation of small solitary pulmonary nodules: 18F-FDG PET/CT and correlation with histology, Nucl Med Commun. July 2015; 36(7):722-7) that show an increase in detectabiiity of lung hypermetabol ic lesions with a difference of 38 % between 3D and 4D SUV max. The same appl ies in l iver lesions (Revheim M-E, Haugvik S-P, Johnsrud K, Mathisen 0, Fjeld JG, Skretting A, Respiratory gated and prolonged acquisition I8F-FDG PET improve preoperative assessment of colorectal liver metastases, Acta Radiol Stockh Swed 1987, April 2015; 56(4):397 -403) where the sensitivity of the 4D PET is increased, but without difference w ith a late centered pitch.

These conclusions remain the same regardless of the technique of synchronization to the respiratory cycle: by frequency analysis, as used in this experimentation, but also by amplitude analysis. This difference is explained by better attenuation correction with the 4D CT. By contrast, a sub correction of attenuation can be observed when the insert is close to the diaphragm, and thus partially surrounded by least attenuating medium: the lung parenchyma that was represented by air only. In this case, the SUV max was always higher and especially overestimated compared with the theory (124% of theory vs. 113% in the acquisition phase 6 of the insert of 8L).

The analysis by type of correction does not show any significant difference in measurement errors of metabol ic volumes.

Regarding the position of the insert relative to the diaphragm, only the metabolic volume at A50 thresholding emerges between the two groups (p = 0.043) but. as a global tendency, a decrease of errors in measurements of quantitative parameters when targets are at a distance from the diaphragm higher than their radius is observed.

Claims

1. Phantom (10) for preparing a detection system (1) for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the detection system (1) comprising a measuring device (3) configured to monitor a respiratory movement of the patient and an imaging device (2) configured to acquire an image of the lesion, the imaging device (2) being synchronised with the measuring device (3), the phantom (10) comprising:

- a body (11) having an inner space (12) extending along a longitudinal axis (L), - a driving system (50) simulating a respiratory movement within the inner space

(12) , and

- at least one insert (35) configured to simulate a lesion, the insert (35) being movably mounted within the inner space (12) of the body (11),

- at least one membrane (30) configured to simulate tissues at an interface between two adjacent anatomical structures, the membrane (30) being flexible and arranged within the body (11) to divide the inner space (12) in two separate compartments (12a, 12b) intended to be filled with respective mediums each simulating one of the anatomical structures, at least one of the compartments (12a, 12b) receiving said at least one insert (35),

the phantom (10) being characterized in that the insert (35) is attached to the driving system (50) so as to be moved by the driving system (50).

2. Phantom (10) according to claim 1, wherein the body (11) includes a lateral wall

(13) around the longitudinal axis (L), the membrane (30) being secured to the lateral wall (13) of the body (11), the membrane (30) extending transversally with respect to the longitudinal axis (L) and the compartments (12a, 12b) being adjacent along the longitudinal axis (L).

3. Phantom (10) according to claim 2, wherein the lateral wall (13) has opposite end edges and the body (11) further includes at least one removable cover (21) removably mounted to the one of the end edges of the lateral wall (13), the driving system (50) being mounted to the removable cover (21).

4. Phantom (10) according to any of claims 1 to 3, wherein the driving system (50) comprises an arm (56) driven to reciprocate back and forth in translation along the longitudinal axis (L), the insert (35) being attached to the arm so as to be movable in translation along the longitudinal axis (L) in a back and forth movement along a stroke about a median position with respect to the membrane (30).

5. Phantom (10) according to claim 4, wherein the median position of the insert (35) within the compartment (12a, 12b) is adjustable.

6. Phantom (10) according to any of claims 1 to 5, wherein the insert (35) is removably mounted within the body (11).

7. Phantom (10) according to any of claims 1 to 6, wherein said at least one insert (35) comprises a plurality of inserts (35) of different sizes.

8. Phantom (10) according to any of claims 1 to 7, wherein the insert (35) includes a hollow envelop intended to be filled with a medium that differs from the medium of which the compartment (12a, 12b) receiving said insert (35) is filled.

9. Phantom (10) according to any of claims 1 to 8, wherein the body (11) further includes a support surface on which the phantom (10) rests in use, the support surface extending parallel to the longitudinal axis (L).

10. Phantom (10) according to any of claims 1 to 9, wherein the body (11) is provided with at least one aperture (16) opening in each compartment (12a, 12b), the aperture (16) presenting an opened state in which a passage for the medium is allowed, and a closed state in which a passage for the medium is prevented.

11. Simulation installation comprising a phantom (10) according to any of claims 1 to 10 and a detection system (1) for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the detection system (1) comprising a measuring device (3) configured to monitor a respiratory movement of the patient and an imaging device (2) configured to acquire an image of the lesion, the imaging device (2) being synchronised with the measuring device (3).

12. Simulation installation according to claim 11, wherein the imaging device (2) is a positron emission tomography (PET) imaging device.

13. Method for preparing a detection system (1) for detecting a lesion in an anatomical structure of a patient under influence of a respiratory movement, the method comprising the steps of:

- providing a measuring device (3) configured to monitor a respiratory movement of the patient and an imaging device (2) configured to acquire an image of the lesion, synchronising the imaging device (2) with the measuring device (3),

- providing a phantom (10) according to any of claims 1 to 10, the phantom (10) comprising: a body (11) having an inner space (12) extending along a longitudinal axis (L), at least one membrane (30) configured to simulate tissues at an interface between two adjacent anatomical structures, the membrane (30) being flexible and arranged within the body (11) to divide the inner space (12) in two separate compartments (12a, 12b) filled with respective mediums each simulating one of the anatomical structures,

a driving system (50), and

at least one insert (35) configured to simulate a lesion, the insert (35) being attached to the driving system (50) so as to be moved within at least one of the compartments (12a, 12b) of the inner space (12) of the body (11) by the driving system (50),

- actuating the driving system (50) of the phantom (10) to simulate a respiratory movement within the inner space (12) of the body (11) of the phantom (10),

- monitoring the respiratory movement simulated by the driving system (50) with the measuring device (3) and acquiring an image of the insert (35) with the imaging device (2).

14. Method according to claim 13, further comprising the step of defining respiratory phases of the respiratory movement and acquiring an image of the insert (35) during each respiratory phase.