- Email: [email protected]
Image-guided total-body irradiation with a movable electronic portal imaging device for bone marrow transplant conditioning
ZEMEDI-10808; No. of Pages 7
ARTICLE IN PRESS
Image-guided total-body irradiation with a movable electronic portal imaging device for bone marrow transplant conditioning Giovanna Dipasquale a,∗ , Raymond Miralbell a , Giorgio Lamanna a , Philippe Nouet a , Manuel Montero b , Michel Rouzaud a , Thomas Zilli a a b
Radiation Oncology Department, Geneva University Hospital, Geneva, Switzerland Radiology Department, Geneva University Hospital, Geneva, Switzerland
Received 5 July 2019; accepted 11 November 2019
Abstract Introduction: To prevent radiation pneumonitis following total body irradiation (TBI) clinicians usually use lung shield blocks. The correct position of these shields relative to the patient’s lungs is usually verified via mega-voltage imaging and computed radiographic (CR) films. In order to improve this time-consuming procedure, we developed in our department a dedicated, movable, real-time imaging system for image-guided TBI. Material & Methods: The system consists of an electronic portal imaging device (EPID) mounted on a dedicated support whose motion along a rail can be controlled from the linac console outside the bunker room. Images are acquired online using a stand-alone console. To test the system efficacy we retrospectively analyzed data of lung blocks positioning from two groups of 10 patients imaged with EPID or CR-films, respectively. Results: The median number of portal images per fraction was 2 (range 1-5) and 1 (range 1-2) for the EPID and the CR-film system, respectively. The minimum time required for an EPID image acquisition, without interpretation and no need of patient position correction in the bunker, was 20 seconds against 214 seconds for the CR-film. Lung shielding positioning in the right-left and superior-inferior directions was improved using the EPID system (p< 0.01). Conclusions: Compared to CR-films, our movable real-time imaging EPID system is a simple technical solution able to reduce the minimum imaging time for lung shielding by a factor of 10. With the increased possibility to acquire more images as compared to CR-film system the EPID system has the potential to improve patient alignment, as well as patient’s comfort and overall setup time. Keywords: Total body irradiation, IGRT, EPID, Radiation pneumonitis
Introduction Total body irradiation (TBI) followed or preceeded by intensive chemotherapy has become one of the standard conditioning regimen before bone marrow transplantation (BMT) in
Abbreviations: BMT, Bone Marrow Transplantation; CR, Computed radiographic; EPID, electronic portal imaging device; TBI, Total Body Irradiation. ∗ Corresponding author. Giovanna Dipasquale, Radiation Oncology Department, Geneva University Hospital. CH-1211 Geneva 14, SWITZERLAND. Tel.: +41 22 37 23 273. fax: +41 22 37 27 117 E-mail: [email protected] (G. Dipasquale).
Z Med Phys xxx (2019) xxx–xxx https://doi.org/10.1016/j.zemedi.2019.11.003 www.elsevier.com/locate/zemedi
treating patients with malignant hematologic disorders. Several randomized trials have proved superior outcomes with TBI compared with non TBI-containing regimens [1–3]. At present, there is no TBI technique that is considered standard, but rather a large number of procedures differing in the total dose, fractionation, dose-rate, patient positioning, beam modifiers, and other technical factors [1]. Ideally, an optimal TBI regimen in allografted patients would result in optimal immunosuppression and bone marrow ablation, a low toxicity rate (particularly pneumonitis) and, in association with chemotherapy, a greater cytotoxic effect [4–6]. To reduce the risk of pneumonitis, many centers utilize lung block shields during TBI to decrease lung dose [5,6]. The correct alignment of the blocks relative to the patient’s lungs
ARTICLE IN PRESS 2
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
is verified via mega voltage imaging, using computed radiographic (CR) film systems, a procedure that is strongly advised for the quality of the delivery but very time consuming. In order to improve and optimize this procedure, we developed in our department a dedicated movable real-time imaging system for image-guided TBI. The aim of this report is to describe this technical solution developed to improve both imaging time and treatment quality for patients undergoing TBI treatments and analyze data of 20 patients simulated with CR film system or EPID.
Materials and Methods In our department and for more than 25 years TBI treatments have been performed using a set of opposed antero-posterior fields [7,8]. Patients are immobilized in a semi-standing position on a special stand with arm bars and hand grips and lung shielding blocks [7]. The stand is a modular structure built of aluminum tubes of 4 cm diameter and wood with a plastified surface (Figure 1 - right). The floor of the stand is elevated 20 cm from the ground and is supported by 4 wheels. The structure is fixed to the treatment room wall with a pair of hooks. A 170 cm high, 95 cm wide, and 1 cm thick removable lucite screen hangs from the front top of the structure at an average distance of 40-50 cm from the patient. This screen has a double purpose: as a “beam spoiler” aiming to reach a skin dose of ≥ 95%, and as a lung shield holder. Lung blocks, manufactured with cerrobend material, are used allowing for a ≈25% dose reduction at the center of both lungs. Personalized blocks are created based on the patient’s simulation images with an outer margins of 2.5 cm from the diaphragm and the lung apex and of 2 cm from the lateral rib cage limits and the vertebral body edge. The blocks are positioned and taped on a 0.2 cm thin lucite tray bearing drawn points corresponding to the top and bottom of the lead marker used for simulation purposes. The tray with the blocks is fixed to the lucite screen with adjustable screws, allowing for fast up-and-down displacements and alignment using skin marks. Patient’s torso rotation and flexion are corrected by moving the patient while controlling for the penumbra of the shielding of the lungs projected from the gantry field light and using the pre-defined skin marks. The treatment is performed at extended (>400 cm) source-skin distance, using a 6MV beam field measuring 40x40 cm2 at the isocenter. Standard verification procedure of proper lung shielding is routinely performed at the beginning of every TBI session by acquiring portal images using completely open mega-voltage beams. A KODAK 2000RT CR system [9] has been used for several years replacing “traditional films”. This system is composed of a CR-film cassette that is positioned behind the patient’s TBI chair, with the CR-film reader set in a separate room not far from the linear accelerator (linac) console area at 20 steps distance. Imaging using a CR system requires approximately 5 minutes per portal image and consists of: radiographic
acquisition with the CR screen-film cassette directly in the treatment bunker, recovery of the CR-film, CR-film “development” followed by a proper read-out and interpretation and patient re-positioning if necessary. Considering that this procedure may have to be repeated several times before achieving a correct positioning of the patient, it could extend considerably the treatment session. Lung shielding accuracy must be balanced against patient discomfort. The longer the overall TBI treatment time, the greater is the discomfort for the patient thus challenging the positioning quality and correct lung shielding. Therefore in clinical reality not more than 2 CR-films are acquired and the final lungs shielding position (after eventual modification) is not always verified. In order to optimize the above procedure, we developed in our department in 2015 a dedicated, movable, real-time imaging system for image-guided TBI treatments. The system consists of an electronic portal imaging device (EPID) mounted on a dedicated support (Figure 1). This support can be moved automatically along a rail fixed inside the bunker, with the controls located in the linac console room. The imaging unit can switch between 2 positions: out-position and home-position, protected from the primary radiation fields. The home-position allows storing the EPID in an area of the treatment room that does not interfere with the normal routine work and also preserves the electronics core of the imaging device. To optimize the position of the EPID in the patient’s craniocaudal direction, the dedicated EPID support can be adjusted using up-down manual buttons. The rail and button systems are composed of standard elements of an automatic sliding door. The EPID is made of amorphous silicon (512x1024 pixels, size 30 x 40 cm2 ) salvaged from a decommissioned linac (2100C series, Varian Medical Systems, CA). The system is connected via a switch-box to the linac-IX (Varian Medical Systems, CA) and uses the Varian maintenance software (AM module) to command the EPID and acquire images. The acquired images can be exported in DICOM (Digital imaging and communications in medicine) format via a USB (Universal Serial Bus) key into the ARIA system (Varian medical systems, CA) to be stored in the patient’s treatment record; Figure 2 shows an example of a typical EPID lung shielding verification image acquired during a TBI session. To assess accuracy improvements of the use of our system, datasets of 10 consecutive patients treated with standard TBI doses (10 Gy or 12 Gy in 5 or 6 fractions, twice daily over three days) and the movable EPID using lung shielding were compared to datasets of 10 patients imaged with CR-films. To verify the accuracy of the lung shield location, distances between the borders of the lung blocks and the diaphragm, the lung apex, the lateral rib cage limits and the vertebral body edge were measured for each image and compared to expected distances from treatment planning to assess accuracy. For the EPID group, information on time between acquisitions was extracted from the time stamps present on repeated
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
3
Figure 1. Movable EPID device on the rails.
Figure 2. EPID image acquisition during a TBI treatment showing lung blocks imported on the ARIA system (Varian Medical systems, CA).
portal images. Individual steps used in the imaging process were also evaluated. Since CR-films images were often acquired only once per session, the time needed to image blocks using a CR-film system was derived by simulating a patient treatment and a CR-film development. Lung block positioning accuracy was tested using the Mann-Whitney test with a threshold for statistical significance of p 0.05.
Results As can be observed from Table 1, for the EPID data, TBI sessions are characterized by a large individual variability in the number of images required to verify the correct patient lungs-shielding positioning. For example, patient #10 required a minimum of 4 consecutive portal images to match
the block shields to the lungs, while patient #8 generally required one single acquisition due to the optimal repositioning. Overall, the median number of portal images per fraction required to ensure a correct lung-shield repositioning was 2, ranging from 1 to 5. On the contrary, the lower number of images acquired per fraction in the CR-film system group (median of 1, ranging between 1 and 2) is explained by the longer acquisition time required to image the lung shield with this technique, (Table 2) and the constrains to keep the session treatment time short. As illustrated on Table 1, in half of the patients treated with the CR-film system, no additional images were acquired to verify the correct repositioning after adjustment of the patient (right-left) or blocks (cranio-caudal) positions. These data confirm that, depending on the patient’s clinical conditions (fatigue, nausea, etc.) and patient’s
ARTICLE IN PRESS 4
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
Table 1 Clinical and portal imaging characteristics for 20 patients treated with total body irradiation using the EPID-system or the CR-film system for lung shielding positioning. Patient ID
Patient Sex
Patientage (years)
Fraction (fx) number: 1 fx
2 fx
3 fx
4 fx
Median # portal images/ session 5 fx
6 fx
3 1 2 2 1 1 4 1 1 1 1 1
3 1 2 2 1 1* 1* 1
# of portal images EPID
FILM
1 2 3 4 5 6 7§ 8 9 10 1 2 3 4 5 6 7 8 9 10
F M M M F F F M F M F M M F F F M M M M
51 39 40 27 24 35 41 41 56 41 54 47 46 14 30 57 58 51 18 23
3 1 1 3 1 2 2 1 2 4 1 1 1 1 1 1 2* 1 1 1
4 2 1 3 1 2 2 1 1 5 1 1* 1 1 1 1 2* 1 1 2
2 3 3 3 2 2 1 1 2 4 1 1 1 1 1* 1 2* 1 1 2
4 3 4 3 4 1 1 3 4 1 1 1* 1 1 1 1 2 1 1
3.5 3 1 3 2 2 1.5 1 2 4 1 1 1 1 1 1 2 1 1 1
Abbreviations EPID: electronic portal imaging device; F: female; M: male; § patient with no left lung shield from 3rd session; * patient positioning corrected but not re-imaged afterwards.
Table 2 Total body irradiation imaging process. TBI imaging procedure at Geneva University Hospital
Time (sec)
Individual steps in each process
Film system reader in a separate room (20 steps distance from linac console)
EPID-system with patient’s repositioning required
EPID-system with no patient’s repositioning
Open the bunker door and move in, while the door is opening Move in the bunker toward the patient Collect the screen film system Move out of the bunker Walk to the film reader Create the patient (identity, field direction, etc) in the imaging database Open the cassette to recover the film and deposit it in the scanner of the film reader Film scanning Film regeneration (to allow reusing the film) Film image manipulation on the computer screen Film image interpretation Walk back from the film reader to the linac console Correct patient position if needed Let the staff next to the patient leave the bunker Close the door of the bunker End of treatment export images to ARIA system Total time (without image interpretation and patient repositioning) in seconds (minutes)
5 12 20 12 10 50
5 12 12 15
0 0 0 15
20
-
-
25 30 5 x 10 x 5 10 x 214 (3.6)
5 x x 0 10 x 59 (1)
5 x 0 0 0 x 20 (0.33)
Abbreviations: EPID = Electronic portal imaging device; X represents a variable time, which depends on the patient imaged but not on the imaging system or procedure used. Therefore it is not accounted for in the total time calculation.
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
5
Table 3 Examples of estimated total body irradiation imaging process, without interpretation time, as a function of the number of images acquired and the imaging system.
# of images acquired CR-film EPID CR-film EPID CR-film EPID
Total time (min) per 1 image acquisition requiring patient repositioning (without image interpretation and patient repositioning, see Table 2)
Possible time (min) spent aligning the patient to the blocks.
Time (min.) as a function of the number of images acquired and time to reposition the patient between 2 images.
3.6 1 3.6 1 3.6 1
0.5 0.5 1 1 1.5 1.5
2 8.2 3 9.2 4 10.2 5
cooperation, lung shielding can be a time consuming and patient dependent issue in TBI treatments. Using the EPID-based movable real-time imaging system, the median time required for a new portal image acquisition, interpretation and correction of the patient position in the bunker was 2 minutes (range, 1-5 minutes). Table 2 shows that, excluding time for image interpretation and patient/block re-alignment, the whole procedure requires approximately 3.6 minutes using the CR-film system against 1 minute with the EPID movable device. If no patient re-alignment is necessary, i.e. no bunker entrance necessary, the EPID imaging time is further reduced to 20 seconds. Table 3 shows how imaging time in a session can grow depending on the imaging device used to acquire the images but also the time to align the patient and the number of images acquired per session. There is the potential to improve overall set-up time and increasing patient comfort when using the movable EPID system compared to CR-based film system due to the fact that acquisition time for a single image is reduced. Accuracy in lung shielding positioning was statistically improved by using our EPID system compared to CR-films (Figure 3). Median deviations in the right-left and superoinferior directions were 0.0 cm (range, -3.3 - 2.0) vs. 0.2 cm (range, -1.5 – 2.3), and -0.3 cm (range, -5.5 – 3.5) vs. 0.6 cm (range, -1.5 – 2.3) for the EPID and the CR-film systems, respectively (p< 0.01). Focusing on the distance of the lung shielding border to the apex (reference value of 2.5 cm), the use of the EPID system resulted in a superior positioning at this level compared to the CR-films: the median distance was 2.7 cm (range, 0.5- 5.5) vs. 2.0 cm (range, 0.5-3.1) for the EPID and CR-film systems, respectively (p< 0.0001).
Discussion Compared to the traditional use of CR-films our movable imaging system is able to significantly reduce imaging acquisition time thus accelerating quality control procedures. Preliminary imaging performed to ensure the correct
3 12.3 4.5 13.8 6 15.3 7.5
4 16.4 6 18.4 8 20.4 10
5 20.5 7.5 23 10 25.5 12.5
positioning of the pulmonary shielding blocks can be performed automatically real-time, without the operator entering the treatment room and with no waiting time for CR-film “development”. This user friendly simplicity allows repeating the images as many times as needed to ensure inter-fractional reproducibility in positioning. Four monitor units are used to acquire a CR-film (between 0.26 and 0.32 cGy for a typical 2500-3000 MU TBI session) compared to 1 (between 0.06 and 0.08 cGy for a typical 25003000 MU TBI session) using the EPID-system, but this is not a clinical issue considering the typical number of monitor units of a treatment session, i.e., 2500-3000 and the equivalent image quality of the two techniques. Furthermore using an EPID system, intra-fractional motion could also be monitored and assessed. In fact, while the beam is stopped the EPID can be moved out of the home position to acquire an image with one monitor unit and immediately “homed” to continue the treatment if no intervention is necessary. The mural supportsystem is compatible with any EPID. In addition, with any TBI technique treating patients with horizontal X-ray beams, it would allow online in-vivo monitoring because the electronic of the EPID could be positioned out of the irradiation beam. Though the use of EPID for TBI is not novel (reported for the first time by Gladstone et al. [10] in 1993 and later by Salk et al. [11]), compared to previous prototypes, our device is the first EPID system developed for image-guided TBI treatments completely movable and automatically controlled. A stand-alone mobile TBI imager on wheels, trolley system, is commercially available (Cablon Medical, The Netherlands). To avoid the irradiation of the electronics, this system is adapted only for TBI treatments using horizontal fields with patients’ laying on a table and fields smaller than 40x40 cm2 . This imager would be incompatible with our TBI technique [7]. Furthermore the trolley system needs to be moved in place for the TBI imaging sessions and stored afterwards requiring extra preparation time at the linac to position the imager. It also requires a large treatment room allowing moving the trolley around the patient lying-down for TBI.
ARTICLE IN PRESS 6
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
Figure 3. Histograms of the difference between the measured distances of the lung shielding border to the reference structures at treatment and the planned distance at simulation. Data is plotted per imaging system, CR-film versus electronic portal imaging device (EPID) for the medio-lateral direction, i.e right-left (RL), and for the caudal-cranial direction (CC). Positioning accuracy measurements were taken using the last image acquired for the final setup image for the EPID-system and for the CR-film. When the CR-film images were not re-acquired, after patient’s adjustment, a perfect alignment in the direction of the correction was considered (8 out of 49 sessions), favoring the CR-film over the EPID system.
Our prototype system instead could be used for any TBI technique using fields directed to a wall regardless of the patient position. Nevertheless, several limitations have to be addressed in this study: the retrospective and mono-institutional nature of the study, with inherent heterogeneities induced by the use of two different sets of patients to compare the two systems and bias related to equipment used; the impossibility to know what the patient-specific lung dose is, because it is not possible to CT scan these patients in treatment position, making it difficult to evaluate patient outcome especially as related to lung toxicity. Imaging more frequently and allowing recovering more information on the treatments could potentially improve outcomes like for external beam radiation therapy.
as compared to CR-film systems, as well as to improve patient comfort and overall setup time. In the future a shielding wall constructed between the electronic core of the movable EPID and the treatment beam may also allow continuous intra-fractional monitoring of patient’s motion for virtually any TBI configuration.
Conflicts of Interest Notification The author & co-authors have no conflicts of interest to declare
Acknowledgments Conclusions In conclusion, the movable, real-time, imaging EPID system presented in this study represents a simple technical solution able to reduce the minimum imaging time for lung shielding by a factor of 10, while using 25% less of dose. This presents the potential to improve patient alignment, because of the possibility to acquire more images to best align the patient
The implementation of this project has been funded by a STARTUP 2014 DISIM and aPRD 6-II-2012 (Projet de Recherche et Développement) grants of the University Hospital of Geneva and a grant of the Fondation Dr Henri Dubois-Ferrière Dinu Lipatti. We would like to thank M. Antonio Rebelo for his assistance with the collection of the data.
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
References [1] Blaise D, Maraninchi D, Michallet M, Reiffers J, Jouet JP, Milpied N, et al. Long-term follow-up of a randomized trial comparing the combination of cyclophosphamide with total body irradiation or busulfan as conditioning regimen for patients receiving HLA-identical marrow grafts for acute myeloblastic leukemia in first complete remission. Blood 2001;97:3669–71. [2] Bunin N, Aplenc R, Kamani N, Shaw K, Cnaan A, Simms S. Randomized trial of busulfan vs total body irradiation containing conditioning regimens for children with acute lymphoblastic leukemia: a Pediatric Blood and Marrow Transplant Consortium study. Bone Marrow Transplant 2003;32:543–8. [3] Dusenbery KE, Daniels KA, McClure JS, McGlave PB, Ramsay NK, Blazar BR, et al. Randomized comparison of cyclophosphamide-total body irradiation versus busulfan-cyclophosphamide conditioning in autologous bone marrow transplantation for acute myeloid leukemia. Int J Radiat Oncol Biol Phys 1995;31:119–28. [4] Van Dyk J, Galvin JM, Glasgow GP, Podgorsak EB. The physical aspects of total and half body photon irradiation: Report of AAPM Radiation Therapy Committee Task Group No 1986;29. [5] Giebel S, Miszczyk L, Slosarek K, Moukhtari L, Ciceri F, Esteve J, et al. Extreme heterogeneity of myeloablative total body irradiation techniques in clinical practice: a survey of the Acute Leukemia
[6]
[7]
[8]
[9]
[10]
[11]
7
Working Party of the European Group for Blood and Marrow Transplantation. Cancer 2014;120:2760–5. Wong JYC, Filippi AR, Dabaja BS, Yahalom J, Specht L. Total Body Irradiation: Guidelines from the International Lymphoma Radiation Oncology Group (ILROG). Int J Radiat Oncol Biol Phys 2018;101:521–9. Miralbell R, Rouzaud M, Grob E, Nouet P, Bieri S, Majno SB, et al. Can a total body irradiation technique be fast and reproducible? Int J Radiat Oncol Biol Phys 1994;29:1167–73. Miralbell R, Sancho G, Bieri S, Carrio I, Helg C, Brunet S, et al. Renal insufficiency in patients with hematologic malignancies undergoing total body irradiation and bone marrow transplantation: a prospective assessment. Int J Radiat Oncol Biol Phys 2004;58:809–16. Conte L, Bianchi C, Cassani E, Monciardini M, Mordacchini C, Novario R, et al. The use of photostimulable phosphor systems for periodic quality assurance in radiotherapy. Phys Med 2008;24:34–7. Gladstone DJ, van Herk M, Chin LM. Verification of lung attenuator positioning before total body irradiation using an electronic portal imaging device. Int J Radiat Oncol Biol Phys 1993;27: 449–54. Salk J, Blank P, Machold U, Rau E, Schneider E, Röttinger EM. Verification of Lung and Kidney Shielding in Total Body Irradiation using an EPID with Extended Image Detection Unit. 5th International Workshop on Electronic Portal Imaging. Phoenix. Arizona 1998.
Available online at www.sciencedirect.com
ScienceDirect
ARTICLE IN PRESS
Image-guided total-body irradiation with a movable electronic portal imaging device for bone marrow transplant conditioning Giovanna Dipasquale a,∗ , Raymond Miralbell a , Giorgio Lamanna a , Philippe Nouet a , Manuel Montero b , Michel Rouzaud a , Thomas Zilli a a b
Radiation Oncology Department, Geneva University Hospital, Geneva, Switzerland Radiology Department, Geneva University Hospital, Geneva, Switzerland
Received 5 July 2019; accepted 11 November 2019
Abstract Introduction: To prevent radiation pneumonitis following total body irradiation (TBI) clinicians usually use lung shield blocks. The correct position of these shields relative to the patient’s lungs is usually verified via mega-voltage imaging and computed radiographic (CR) films. In order to improve this time-consuming procedure, we developed in our department a dedicated, movable, real-time imaging system for image-guided TBI. Material & Methods: The system consists of an electronic portal imaging device (EPID) mounted on a dedicated support whose motion along a rail can be controlled from the linac console outside the bunker room. Images are acquired online using a stand-alone console. To test the system efficacy we retrospectively analyzed data of lung blocks positioning from two groups of 10 patients imaged with EPID or CR-films, respectively. Results: The median number of portal images per fraction was 2 (range 1-5) and 1 (range 1-2) for the EPID and the CR-film system, respectively. The minimum time required for an EPID image acquisition, without interpretation and no need of patient position correction in the bunker, was 20 seconds against 214 seconds for the CR-film. Lung shielding positioning in the right-left and superior-inferior directions was improved using the EPID system (p< 0.01). Conclusions: Compared to CR-films, our movable real-time imaging EPID system is a simple technical solution able to reduce the minimum imaging time for lung shielding by a factor of 10. With the increased possibility to acquire more images as compared to CR-film system the EPID system has the potential to improve patient alignment, as well as patient’s comfort and overall setup time. Keywords: Total body irradiation, IGRT, EPID, Radiation pneumonitis
Introduction Total body irradiation (TBI) followed or preceeded by intensive chemotherapy has become one of the standard conditioning regimen before bone marrow transplantation (BMT) in
Abbreviations: BMT, Bone Marrow Transplantation; CR, Computed radiographic; EPID, electronic portal imaging device; TBI, Total Body Irradiation. ∗ Corresponding author. Giovanna Dipasquale, Radiation Oncology Department, Geneva University Hospital. CH-1211 Geneva 14, SWITZERLAND. Tel.: +41 22 37 23 273. fax: +41 22 37 27 117 E-mail: [email protected] (G. Dipasquale).
Z Med Phys xxx (2019) xxx–xxx https://doi.org/10.1016/j.zemedi.2019.11.003 www.elsevier.com/locate/zemedi
treating patients with malignant hematologic disorders. Several randomized trials have proved superior outcomes with TBI compared with non TBI-containing regimens [1–3]. At present, there is no TBI technique that is considered standard, but rather a large number of procedures differing in the total dose, fractionation, dose-rate, patient positioning, beam modifiers, and other technical factors [1]. Ideally, an optimal TBI regimen in allografted patients would result in optimal immunosuppression and bone marrow ablation, a low toxicity rate (particularly pneumonitis) and, in association with chemotherapy, a greater cytotoxic effect [4–6]. To reduce the risk of pneumonitis, many centers utilize lung block shields during TBI to decrease lung dose [5,6]. The correct alignment of the blocks relative to the patient’s lungs
ARTICLE IN PRESS 2
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
is verified via mega voltage imaging, using computed radiographic (CR) film systems, a procedure that is strongly advised for the quality of the delivery but very time consuming. In order to improve and optimize this procedure, we developed in our department a dedicated movable real-time imaging system for image-guided TBI. The aim of this report is to describe this technical solution developed to improve both imaging time and treatment quality for patients undergoing TBI treatments and analyze data of 20 patients simulated with CR film system or EPID.
Materials and Methods In our department and for more than 25 years TBI treatments have been performed using a set of opposed antero-posterior fields [7,8]. Patients are immobilized in a semi-standing position on a special stand with arm bars and hand grips and lung shielding blocks [7]. The stand is a modular structure built of aluminum tubes of 4 cm diameter and wood with a plastified surface (Figure 1 - right). The floor of the stand is elevated 20 cm from the ground and is supported by 4 wheels. The structure is fixed to the treatment room wall with a pair of hooks. A 170 cm high, 95 cm wide, and 1 cm thick removable lucite screen hangs from the front top of the structure at an average distance of 40-50 cm from the patient. This screen has a double purpose: as a “beam spoiler” aiming to reach a skin dose of ≥ 95%, and as a lung shield holder. Lung blocks, manufactured with cerrobend material, are used allowing for a ≈25% dose reduction at the center of both lungs. Personalized blocks are created based on the patient’s simulation images with an outer margins of 2.5 cm from the diaphragm and the lung apex and of 2 cm from the lateral rib cage limits and the vertebral body edge. The blocks are positioned and taped on a 0.2 cm thin lucite tray bearing drawn points corresponding to the top and bottom of the lead marker used for simulation purposes. The tray with the blocks is fixed to the lucite screen with adjustable screws, allowing for fast up-and-down displacements and alignment using skin marks. Patient’s torso rotation and flexion are corrected by moving the patient while controlling for the penumbra of the shielding of the lungs projected from the gantry field light and using the pre-defined skin marks. The treatment is performed at extended (>400 cm) source-skin distance, using a 6MV beam field measuring 40x40 cm2 at the isocenter. Standard verification procedure of proper lung shielding is routinely performed at the beginning of every TBI session by acquiring portal images using completely open mega-voltage beams. A KODAK 2000RT CR system [9] has been used for several years replacing “traditional films”. This system is composed of a CR-film cassette that is positioned behind the patient’s TBI chair, with the CR-film reader set in a separate room not far from the linear accelerator (linac) console area at 20 steps distance. Imaging using a CR system requires approximately 5 minutes per portal image and consists of: radiographic
acquisition with the CR screen-film cassette directly in the treatment bunker, recovery of the CR-film, CR-film “development” followed by a proper read-out and interpretation and patient re-positioning if necessary. Considering that this procedure may have to be repeated several times before achieving a correct positioning of the patient, it could extend considerably the treatment session. Lung shielding accuracy must be balanced against patient discomfort. The longer the overall TBI treatment time, the greater is the discomfort for the patient thus challenging the positioning quality and correct lung shielding. Therefore in clinical reality not more than 2 CR-films are acquired and the final lungs shielding position (after eventual modification) is not always verified. In order to optimize the above procedure, we developed in our department in 2015 a dedicated, movable, real-time imaging system for image-guided TBI treatments. The system consists of an electronic portal imaging device (EPID) mounted on a dedicated support (Figure 1). This support can be moved automatically along a rail fixed inside the bunker, with the controls located in the linac console room. The imaging unit can switch between 2 positions: out-position and home-position, protected from the primary radiation fields. The home-position allows storing the EPID in an area of the treatment room that does not interfere with the normal routine work and also preserves the electronics core of the imaging device. To optimize the position of the EPID in the patient’s craniocaudal direction, the dedicated EPID support can be adjusted using up-down manual buttons. The rail and button systems are composed of standard elements of an automatic sliding door. The EPID is made of amorphous silicon (512x1024 pixels, size 30 x 40 cm2 ) salvaged from a decommissioned linac (2100C series, Varian Medical Systems, CA). The system is connected via a switch-box to the linac-IX (Varian Medical Systems, CA) and uses the Varian maintenance software (AM module) to command the EPID and acquire images. The acquired images can be exported in DICOM (Digital imaging and communications in medicine) format via a USB (Universal Serial Bus) key into the ARIA system (Varian medical systems, CA) to be stored in the patient’s treatment record; Figure 2 shows an example of a typical EPID lung shielding verification image acquired during a TBI session. To assess accuracy improvements of the use of our system, datasets of 10 consecutive patients treated with standard TBI doses (10 Gy or 12 Gy in 5 or 6 fractions, twice daily over three days) and the movable EPID using lung shielding were compared to datasets of 10 patients imaged with CR-films. To verify the accuracy of the lung shield location, distances between the borders of the lung blocks and the diaphragm, the lung apex, the lateral rib cage limits and the vertebral body edge were measured for each image and compared to expected distances from treatment planning to assess accuracy. For the EPID group, information on time between acquisitions was extracted from the time stamps present on repeated
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
3
Figure 1. Movable EPID device on the rails.
Figure 2. EPID image acquisition during a TBI treatment showing lung blocks imported on the ARIA system (Varian Medical systems, CA).
portal images. Individual steps used in the imaging process were also evaluated. Since CR-films images were often acquired only once per session, the time needed to image blocks using a CR-film system was derived by simulating a patient treatment and a CR-film development. Lung block positioning accuracy was tested using the Mann-Whitney test with a threshold for statistical significance of p 0.05.
Results As can be observed from Table 1, for the EPID data, TBI sessions are characterized by a large individual variability in the number of images required to verify the correct patient lungs-shielding positioning. For example, patient #10 required a minimum of 4 consecutive portal images to match
the block shields to the lungs, while patient #8 generally required one single acquisition due to the optimal repositioning. Overall, the median number of portal images per fraction required to ensure a correct lung-shield repositioning was 2, ranging from 1 to 5. On the contrary, the lower number of images acquired per fraction in the CR-film system group (median of 1, ranging between 1 and 2) is explained by the longer acquisition time required to image the lung shield with this technique, (Table 2) and the constrains to keep the session treatment time short. As illustrated on Table 1, in half of the patients treated with the CR-film system, no additional images were acquired to verify the correct repositioning after adjustment of the patient (right-left) or blocks (cranio-caudal) positions. These data confirm that, depending on the patient’s clinical conditions (fatigue, nausea, etc.) and patient’s
ARTICLE IN PRESS 4
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
Table 1 Clinical and portal imaging characteristics for 20 patients treated with total body irradiation using the EPID-system or the CR-film system for lung shielding positioning. Patient ID
Patient Sex
Patientage (years)
Fraction (fx) number: 1 fx
2 fx
3 fx
4 fx
Median # portal images/ session 5 fx
6 fx
3 1 2 2 1 1 4 1 1 1 1 1
3 1 2 2 1 1* 1* 1
# of portal images EPID
FILM
1 2 3 4 5 6 7§ 8 9 10 1 2 3 4 5 6 7 8 9 10
F M M M F F F M F M F M M F F F M M M M
51 39 40 27 24 35 41 41 56 41 54 47 46 14 30 57 58 51 18 23
3 1 1 3 1 2 2 1 2 4 1 1 1 1 1 1 2* 1 1 1
4 2 1 3 1 2 2 1 1 5 1 1* 1 1 1 1 2* 1 1 2
2 3 3 3 2 2 1 1 2 4 1 1 1 1 1* 1 2* 1 1 2
4 3 4 3 4 1 1 3 4 1 1 1* 1 1 1 1 2 1 1
3.5 3 1 3 2 2 1.5 1 2 4 1 1 1 1 1 1 2 1 1 1
Abbreviations EPID: electronic portal imaging device; F: female; M: male; § patient with no left lung shield from 3rd session; * patient positioning corrected but not re-imaged afterwards.
Table 2 Total body irradiation imaging process. TBI imaging procedure at Geneva University Hospital
Time (sec)
Individual steps in each process
Film system reader in a separate room (20 steps distance from linac console)
EPID-system with patient’s repositioning required
EPID-system with no patient’s repositioning
Open the bunker door and move in, while the door is opening Move in the bunker toward the patient Collect the screen film system Move out of the bunker Walk to the film reader Create the patient (identity, field direction, etc) in the imaging database Open the cassette to recover the film and deposit it in the scanner of the film reader Film scanning Film regeneration (to allow reusing the film) Film image manipulation on the computer screen Film image interpretation Walk back from the film reader to the linac console Correct patient position if needed Let the staff next to the patient leave the bunker Close the door of the bunker End of treatment export images to ARIA system Total time (without image interpretation and patient repositioning) in seconds (minutes)
5 12 20 12 10 50
5 12 12 15
0 0 0 15
20
-
-
25 30 5 x 10 x 5 10 x 214 (3.6)
5 x x 0 10 x 59 (1)
5 x 0 0 0 x 20 (0.33)
Abbreviations: EPID = Electronic portal imaging device; X represents a variable time, which depends on the patient imaged but not on the imaging system or procedure used. Therefore it is not accounted for in the total time calculation.
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
5
Table 3 Examples of estimated total body irradiation imaging process, without interpretation time, as a function of the number of images acquired and the imaging system.
# of images acquired CR-film EPID CR-film EPID CR-film EPID
Total time (min) per 1 image acquisition requiring patient repositioning (without image interpretation and patient repositioning, see Table 2)
Possible time (min) spent aligning the patient to the blocks.
Time (min.) as a function of the number of images acquired and time to reposition the patient between 2 images.
3.6 1 3.6 1 3.6 1
0.5 0.5 1 1 1.5 1.5
2 8.2 3 9.2 4 10.2 5
cooperation, lung shielding can be a time consuming and patient dependent issue in TBI treatments. Using the EPID-based movable real-time imaging system, the median time required for a new portal image acquisition, interpretation and correction of the patient position in the bunker was 2 minutes (range, 1-5 minutes). Table 2 shows that, excluding time for image interpretation and patient/block re-alignment, the whole procedure requires approximately 3.6 minutes using the CR-film system against 1 minute with the EPID movable device. If no patient re-alignment is necessary, i.e. no bunker entrance necessary, the EPID imaging time is further reduced to 20 seconds. Table 3 shows how imaging time in a session can grow depending on the imaging device used to acquire the images but also the time to align the patient and the number of images acquired per session. There is the potential to improve overall set-up time and increasing patient comfort when using the movable EPID system compared to CR-based film system due to the fact that acquisition time for a single image is reduced. Accuracy in lung shielding positioning was statistically improved by using our EPID system compared to CR-films (Figure 3). Median deviations in the right-left and superoinferior directions were 0.0 cm (range, -3.3 - 2.0) vs. 0.2 cm (range, -1.5 – 2.3), and -0.3 cm (range, -5.5 – 3.5) vs. 0.6 cm (range, -1.5 – 2.3) for the EPID and the CR-film systems, respectively (p< 0.01). Focusing on the distance of the lung shielding border to the apex (reference value of 2.5 cm), the use of the EPID system resulted in a superior positioning at this level compared to the CR-films: the median distance was 2.7 cm (range, 0.5- 5.5) vs. 2.0 cm (range, 0.5-3.1) for the EPID and CR-film systems, respectively (p< 0.0001).
Discussion Compared to the traditional use of CR-films our movable imaging system is able to significantly reduce imaging acquisition time thus accelerating quality control procedures. Preliminary imaging performed to ensure the correct
3 12.3 4.5 13.8 6 15.3 7.5
4 16.4 6 18.4 8 20.4 10
5 20.5 7.5 23 10 25.5 12.5
positioning of the pulmonary shielding blocks can be performed automatically real-time, without the operator entering the treatment room and with no waiting time for CR-film “development”. This user friendly simplicity allows repeating the images as many times as needed to ensure inter-fractional reproducibility in positioning. Four monitor units are used to acquire a CR-film (between 0.26 and 0.32 cGy for a typical 2500-3000 MU TBI session) compared to 1 (between 0.06 and 0.08 cGy for a typical 25003000 MU TBI session) using the EPID-system, but this is not a clinical issue considering the typical number of monitor units of a treatment session, i.e., 2500-3000 and the equivalent image quality of the two techniques. Furthermore using an EPID system, intra-fractional motion could also be monitored and assessed. In fact, while the beam is stopped the EPID can be moved out of the home position to acquire an image with one monitor unit and immediately “homed” to continue the treatment if no intervention is necessary. The mural supportsystem is compatible with any EPID. In addition, with any TBI technique treating patients with horizontal X-ray beams, it would allow online in-vivo monitoring because the electronic of the EPID could be positioned out of the irradiation beam. Though the use of EPID for TBI is not novel (reported for the first time by Gladstone et al. [10] in 1993 and later by Salk et al. [11]), compared to previous prototypes, our device is the first EPID system developed for image-guided TBI treatments completely movable and automatically controlled. A stand-alone mobile TBI imager on wheels, trolley system, is commercially available (Cablon Medical, The Netherlands). To avoid the irradiation of the electronics, this system is adapted only for TBI treatments using horizontal fields with patients’ laying on a table and fields smaller than 40x40 cm2 . This imager would be incompatible with our TBI technique [7]. Furthermore the trolley system needs to be moved in place for the TBI imaging sessions and stored afterwards requiring extra preparation time at the linac to position the imager. It also requires a large treatment room allowing moving the trolley around the patient lying-down for TBI.
ARTICLE IN PRESS 6
G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
Figure 3. Histograms of the difference between the measured distances of the lung shielding border to the reference structures at treatment and the planned distance at simulation. Data is plotted per imaging system, CR-film versus electronic portal imaging device (EPID) for the medio-lateral direction, i.e right-left (RL), and for the caudal-cranial direction (CC). Positioning accuracy measurements were taken using the last image acquired for the final setup image for the EPID-system and for the CR-film. When the CR-film images were not re-acquired, after patient’s adjustment, a perfect alignment in the direction of the correction was considered (8 out of 49 sessions), favoring the CR-film over the EPID system.
Our prototype system instead could be used for any TBI technique using fields directed to a wall regardless of the patient position. Nevertheless, several limitations have to be addressed in this study: the retrospective and mono-institutional nature of the study, with inherent heterogeneities induced by the use of two different sets of patients to compare the two systems and bias related to equipment used; the impossibility to know what the patient-specific lung dose is, because it is not possible to CT scan these patients in treatment position, making it difficult to evaluate patient outcome especially as related to lung toxicity. Imaging more frequently and allowing recovering more information on the treatments could potentially improve outcomes like for external beam radiation therapy.
as compared to CR-film systems, as well as to improve patient comfort and overall setup time. In the future a shielding wall constructed between the electronic core of the movable EPID and the treatment beam may also allow continuous intra-fractional monitoring of patient’s motion for virtually any TBI configuration.
Conflicts of Interest Notification The author & co-authors have no conflicts of interest to declare
Acknowledgments Conclusions In conclusion, the movable, real-time, imaging EPID system presented in this study represents a simple technical solution able to reduce the minimum imaging time for lung shielding by a factor of 10, while using 25% less of dose. This presents the potential to improve patient alignment, because of the possibility to acquire more images to best align the patient
The implementation of this project has been funded by a STARTUP 2014 DISIM and aPRD 6-II-2012 (Projet de Recherche et Développement) grants of the University Hospital of Geneva and a grant of the Fondation Dr Henri Dubois-Ferrière Dinu Lipatti. We would like to thank M. Antonio Rebelo for his assistance with the collection of the data.
ARTICLE IN PRESS G. Dipasquale et al. / Z Med Phys xxx (2019) xxx–xxx
References [1] Blaise D, Maraninchi D, Michallet M, Reiffers J, Jouet JP, Milpied N, et al. Long-term follow-up of a randomized trial comparing the combination of cyclophosphamide with total body irradiation or busulfan as conditioning regimen for patients receiving HLA-identical marrow grafts for acute myeloblastic leukemia in first complete remission. Blood 2001;97:3669–71. [2] Bunin N, Aplenc R, Kamani N, Shaw K, Cnaan A, Simms S. Randomized trial of busulfan vs total body irradiation containing conditioning regimens for children with acute lymphoblastic leukemia: a Pediatric Blood and Marrow Transplant Consortium study. Bone Marrow Transplant 2003;32:543–8. [3] Dusenbery KE, Daniels KA, McClure JS, McGlave PB, Ramsay NK, Blazar BR, et al. Randomized comparison of cyclophosphamide-total body irradiation versus busulfan-cyclophosphamide conditioning in autologous bone marrow transplantation for acute myeloid leukemia. Int J Radiat Oncol Biol Phys 1995;31:119–28. [4] Van Dyk J, Galvin JM, Glasgow GP, Podgorsak EB. The physical aspects of total and half body photon irradiation: Report of AAPM Radiation Therapy Committee Task Group No 1986;29. [5] Giebel S, Miszczyk L, Slosarek K, Moukhtari L, Ciceri F, Esteve J, et al. Extreme heterogeneity of myeloablative total body irradiation techniques in clinical practice: a survey of the Acute Leukemia
[6]
[7]
[8]
[9]
[10]
[11]
7
Working Party of the European Group for Blood and Marrow Transplantation. Cancer 2014;120:2760–5. Wong JYC, Filippi AR, Dabaja BS, Yahalom J, Specht L. Total Body Irradiation: Guidelines from the International Lymphoma Radiation Oncology Group (ILROG). Int J Radiat Oncol Biol Phys 2018;101:521–9. Miralbell R, Rouzaud M, Grob E, Nouet P, Bieri S, Majno SB, et al. Can a total body irradiation technique be fast and reproducible? Int J Radiat Oncol Biol Phys 1994;29:1167–73. Miralbell R, Sancho G, Bieri S, Carrio I, Helg C, Brunet S, et al. Renal insufficiency in patients with hematologic malignancies undergoing total body irradiation and bone marrow transplantation: a prospective assessment. Int J Radiat Oncol Biol Phys 2004;58:809–16. Conte L, Bianchi C, Cassani E, Monciardini M, Mordacchini C, Novario R, et al. The use of photostimulable phosphor systems for periodic quality assurance in radiotherapy. Phys Med 2008;24:34–7. Gladstone DJ, van Herk M, Chin LM. Verification of lung attenuator positioning before total body irradiation using an electronic portal imaging device. Int J Radiat Oncol Biol Phys 1993;27: 449–54. Salk J, Blank P, Machold U, Rau E, Schneider E, Röttinger EM. Verification of Lung and Kidney Shielding in Total Body Irradiation using an EPID with Extended Image Detection Unit. 5th International Workshop on Electronic Portal Imaging. Phoenix. Arizona 1998.
Available online at www.sciencedirect.com
ScienceDirect