Technology in Cancer Research and Treatment ISSN 1533-0346 Volume 13, Number 4, August 2014 © Adenine Press (2014)

Commissioning and Quality Assurance of an Integrated System for Patient Positioning and Setup Verification in Particle Therapy www.tcrt.org DOI: 10.7785/tcrt.2012.500386 In an increasing number of clinical indications, radiotherapy with accelerated particles shows relevant advantages when compared with high energy X-ray irradiation. However, due to the finite range of ions, particle therapy can be severely compromised by setup errors and geometric uncertainties. The purpose of this work is to describe the commissioning and the design of the quality assurance procedures for patient positioning and setup verification systems at the Italian National Center for Oncological Hadrontherapy (CNAO). The accuracy of systems installed in CNAO and devoted to patient positioning and setup verification have been assessed using a laser tracking device. The accuracy in calibration and image based setup verification relying on in room X-ray imaging system was also quantified. Quality assurance tests to check the integration among all patient setup systems were designed, and records of daily QA tests since the start of clinical operation (2011) are presented. The overall accuracy of the patient positioning system and the patient verification system motion was proved to be below 0.5 mm under all the examined conditions, with median values below the 0.3 mm threshold. Image based registration in phantom studies exhibited sub-millimetric accuracy in setup verification at both cranial and extra-cranial sites. The calibration residuals of the OTS were found consistent with the expectations, with peak values below 0.3 mm. Quality assurance tests, daily performed before clinical operation, confirm adequate integration and sub-millimetric setup accuracy. Robotic patient positioning was successfully integrated with optical tracking and stereoscopic X-ray verification for patient setup in particle therapy. Sub-millimetric setup accuracy was achieved and consistently verified in daily ­clinical operation. Key words: Particle therapy; Patient positioning system; Patient verification system; Setup validation.

A. Pella, Ph.D.1* M. Riboldi, Ph.D.1,2 B. Tagaste, B.Sc.2 D. Bianculli, M.Sc.3 M. Desplanques, M.Sc.1 G. Fontana, M.Sc.2 P. Cerveri, Ph.D.1 M. Seregni, M.Sc.1 G. Fattori, M.Sc.1 R. Orecchia, M.D.2,4,5,6 G. Baroni, Ph.D.1,2 Politecnico di Milano, ­Dipartimento

1

di Elettronica, Informazione e ­Bioingegneria, Milano, Italy CNAO Foundation, Clinical Division,

2

Pavia, Italy CNAO Foundation, Accelerator

3

­Division, Pavia, Italy CNAO Foundation, Scientific Director,

4

Pavia, Italy European Institute of Oncology,

5

­Division of Radiotherapy, Milano, Italy University of Milan, Milano, Italy

6

Introduction Particle therapy delivering protons and carbon ions is emerging as a valuable technique in cancer treatment (1-10). Particles have an inverse dose profile, that allows to concentrate a higher target dose, with simultaneous lower dosage to healthy tissues. As a consequence, they are characterized by a high geometrical selectivity (especially when active scanning is applied) granting millimeter – precision dose delivery. Abbreviations: CNAO: Centro Nazionale di Adroterapia Oncologica (Italian National Center for Oncological Hadrontherapy); PPS: Patient Positioning System; PVS: Patient Verification System; OTS: Optical Tracking System; CT: Computerized Tomography; QA: Quality Assurance.

*Corresponding author: Andrea Pella, Ph.D. E-mail: [email protected]

303 Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

304

Pella et al.

Particles also feature enhanced radio-biological effectiveness in the target with relatively low biological interaction in the entrance region (skin sparing). In particular, particle therapy decreases the integral dose, thus reducing the amount of irradiated healthy tissue. This makes particle therapy an excellent tool for high dose radiotherapy, resulting in high target dose coverage and reduction of low/moderate dose volumes. As a consequence, even if not already supported by randomized clinical trials, dose escalation in the target can ideally be achieved with a reduction of acute and late toxicity and risk of inducing secondary malignancies, if compared to conventional X-ray radiotherapy (11-13). Conversely, particle therapy is more sensitive to uncertainties than conventional radiotherapy (14). Inaccurate patient setup may result in severe under-dosage in the target and unacceptable dose levels in healthy tissues. It is evident that the potential advantages of particle therapy can be fully exploited only in combination with modern patient positioning and image guidance systems. These technologies, already entered in high energy X-ray radiotherapy (15-16) for the implementation of image guided treatment protocols Image-guided ­Radiation Therapy (IGRT), require a specific optimization for an application in particle therapy. Without excluding the use of conventional patient immobilization devices such as vacuum bags, custom molds, rigid stereotactic head frames, rectal balloons etc., a modern center for particle therapy is usually equipped with integrated systems for high precision patient positioning and setup verification. Because IGRT improves precision (14, 17-21), dose hypofractionation becomes more feasible, leading to a reduction of the number of treatment sessions. Particle therapy, and especially carbon ion therapy, exhibits a consistent trend toward hypofractionation (9, 22). According to current clinical practice, high mechanical precision of the patient positioning system is required to ensure a proper patient setup with respect to the isocenter. Multiple X-ray projections with diagnostic image quality and/or volumetric in-room imaging devices are applied to verify the concordance of patient’s anatomy with the treatment planning CT and/or with digitally reconstructed radiographs (DRRs), and correct inter-fractional setup errors. When multiple (typically two) X-ray projections and 2D-3D image registration methods are applied for patient setup verification, evident limitations apply due to the lack of soft-tissue visualization. In these cases, radiopaque seeds may be implanted near the tumor mass to help the identification of the target. Further image-guidance technologies (ultrasonography, markerless fluoroscopy) have been explored (16, 23-29).

Regardless the specific set of technologies that are put in operation for patient positioning and setup verification, a crucial issue is the need to define calibration procedures and daily/periodic quality assurance tests need to be carefully designed to verify systems performance. The number of facilities able to offer a treatment with accelerated particles is nowadays worldwide increasing (30). Related technologies are growing due to improvements in control and tools (31) for setup verification before and during irradiation. High costs are the principal reason why this type of treatment is not largely distributed (12, 32). Actually, in Europe only two facilities are equipped with a synchrotron able to accelerate both protons and carbon ions: the first one is in Germany Heidelberg Ion Therapy Center (HIT) and the second in Italy Centro Nazionale di Adroterapia Oncologica (CNAO) (33-35). The CNAO started clinical treatments with protons in September 2011 and with carbon ions in November 2012. In this work we describe the patient setup devices installed at CNAO and elucidate the commissioning activities, the design and initial results of quality assurance tests (36) performed during patient treatments in the two rooms that are currently ­performing clinical operation. Materials and Methods Patient Setup Procedures At CNAO, three rooms are available for treatment d­ elivery; two of them are equipped with a fixed horizontal beamline and one with a dual horizontal and vertical beamlines. Patient setup procedures are described for the two rooms equipped with the horizontal beamline, that are currently operational for clinical patient treatments. The basic concept of patient setup at CNAO is that patients are positioned and immobilized on dedicated treatment supports (couch or chair) in dedicated preparation rooms outside the treatment bunker and then brought into the treatment room on a transport system directly before irradiation. The concept envisages that the availability of multiple treatment supports and transport systems allows therapists to setup a patient outside the bunker, while the previous one is being treated, thus resulting into effective workflow optimization across the three available rooms. An analysis shows that in this way, the facility could reach the potential to deliver more than 18,000 treatment sessions per year under full clinical operation (33, 35). Once within the treatment bunker, the couch or chair is docked to a 6 degrees of freedom patient positioning device CNAO Patient Positioning System (CNAOPPS). The PPS drives the patient into the nominal position defined at treatment planning. At this stage, two

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

Design of a Patient Positioning and Setup Verification System for Particle Therapy 305 i­ndependent v­ erification systems are applied for patient setup refinement: – A first patient setup check is performed by means of an infrared optical tracking system (CNAO-OTS), that computes a first correction vector (CV) based on markers placed on patient immobilization device. – Orthogonal X-ray projections are acquired by means of a stereoscopic image based verification device CNAO Patient Verification System (CNAO-PVS) and a second correction vector, based on bony anatomy 2D-3D registration is estimated with radiation oncologists supervision.

The estimated patient setup correction vectors are sent to the CNAO-PPS that is used to improve patient setup accordingly. During the irradiation, the CNAO-OTS provides real time monitoring of the external surrogates as a safety measure against undesired patient motion and is able to evoke an interlock for beam delivery interruption, in case patient motion is detected. Figure 1 shows PPS, OTS and PVS, respectively. Details regarding the technical specifications of all systems are reported in Table I and described in the ­following paragraphs. Safety Measures CNAO-PPS and PVS are equipped with safety components to minimize the probability of patient injury even in presence of a mechanical and/or steering failure. Anticollision panels are installed on any moving components of both systems to guarantee the safety of operators and patients. There are two levels of alarm: the first is managed by software inhibiting the motion of the PPS-PVS to potentially dangerous configurations, the second is based on the anticollision panels and is mastered by a redundant safety programmable safety controller, which bypasses the prime system controller.

Figure 1:  Overview of the systems installed in room 1. Left panel: (A) PPS; (B) PVS; (C) OTS; (D) transport system. Right panel: (E-F) X-ray tubes; (G-H) flat panels.

As an additional safety feature, the PPS weight sensors provides a feedback to the control system, interrupting any operation if an over threshold change in weight load is detected during treatment operations.

Table I Technical specification of patient setup sub-systems (37).

PPS

OTS

Translation range of motion

X: 1000 mm Y: 1000 mm Z: from 700 mm to 1300 mm above ground

Rotation range of motion

Yaw: 185° (with respect to room coordinates) Pitch: (depending by the height, up to) 30° Roll: (depending by the height, up to) 15°

Motion accuracy

0.3 mm in translations 0.1° in rotations

Field of view (FOV) Peak error in 3-D fiducial localization Real time point\surface – based registration

500 3 500 3 300 mm3 around isocenter 0.5 mm Guarantee at 15 Hz s.r.

Image receptor

Model: Varian 4030E digital flat panels Resolution: 3200 3 2300 pixel ∼30 3 40 cm pixel area; 0.127 mm pixel size

X-ray generator

Model: Varian Frequency: up to 3 fps 150 kVp; 640 (optional 800) mA

Accuracy

Translations of flat panels: 0.15 mm Rotations of X-ray tube: 0.15 mm Rotation of the entire system: 0.1°

PVS

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

306

Pella et al.

As a whole, the system complies with the standards for human safety and with the directives for European Certification of medical devices and has obtained CE certification (number V-11-048) in 2012. System Commissioning The commissioning of the integrated system for patient positioning and setup verification (optical and X-ray) installed in the facility was focused on the following issues: 1. PPS motion accuracy (translational/rotational) with a peak acceptable error of 0.3 mm according to specifications (see Table I) 2. PVS structure rotation and deployment accuracy of tubes and panels with peak acceptable errors of less than 0.1° for system rotation and 0.15 mm for imaging deployment (see Table I) 3. OTS accuracy in marker spatial localization with threedimensional errors within 0.5 mm 4. Overall accuracy of image-based 2D-3D registration based on the commercial Verisuite software application with acceptable uncertainties in recovering optimal setup of 1 mm and 1°. Maximum time consumption 60?. For motion accuracy measurements (in particular of the PPS) we used a Leica laser tracker (LTD 500, Leica Geosystems Spa), (Figure 2). This laser measurement system has a built-in interferometer, featuring the following range of measurement: – horizontal 235° – vertical 45° – distance 0–35 m (0–115′) The instrument resolution, as declared by ­manufacturer, is 0.14 and 1.26 μm for angular and linear measurements,

respectively (absolute accuracy for a static target is  10 µm/m). The ­nominal uncertainty (RMS) of the treatment room network to determine 3D coordinates of the isocenter measured 13  μm. Although motion accuracy was tested experimentally with both treatment couch and chair docked to the PPS, the results relative to the treatment couch are reported in this paper, since the treatment chair is still currently under commissioning for clinical use. CNAO-PPS motion accuracy tests were performed with and without an additional weight fixed on the surface of the treatment couch, as a way to investigate the efficacy of the weight compensation featured by the PPS control system. Motion commissioning procedures were carried out using a corner cube reflector (Figure 2, right box) and a set of magnetic sockets, which were glued directly on the PPS/PVS at the beginning of each measurement session. Data were stored into the laser tracker control computer, and then exported as text files for further analysis. Patient Positioning System (PPS) The PPS is a robotic arm with a pantographic architecture featuring 6 degrees of freedom motion, manufactured for CNAO by Schaer Engineering (Flaach, Switzerland). The extremity of this device is fixed on the floor, while the basement dedicated at docking procedures of treatment tools is floating (see Figure 1A). In particular, the basement of the PPS moves on a granite floor, where friction is minimized by a flow of compressed air (133.3 l/min at 10 bar). The flatness of the granite surface is a basic requirement to achieve effective motion of the basement: flatness is specified within 0.02 mm for a surface with 10 mm radius, 0.1 mm for a surface with 500 mm radius and 0.3 mm for a surface with 3000 mm radius. The PPS was primarily designed for fixed-beam applications, and it can be coupled with a treatment couch or chair, providing an adequate range of patient setup possibilities. Payload of up to 200 kg can be supported by the PPS and aligned to the nominal treatment position. Additional weight on the treatment tool is measured by four sensors and compensated via software. This allows PPS to feature a high accuracy in rotational and translational motion within a fairly large range of motion (see Table I). PPS Commissioning

Figure 2:  Setup during commissioning (left) and detail of the corner cuve aligned at the isocenter (right).

The protocol for PPS commissioning envisaged to drive the system in 20 random 6 degrees of

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

Design of a Patient Positioning and Setup Verification System for Particle Therapy 307 freedom configurations (10 without and 10 with an additional 40 kg weight) in the range (delta peak to peak values) 441 mm lateral, 850 mm longitudinal, 262 mm vertical, 9° pitch, 6.5° roll, 220° yaw with respect to beam direction.

components, like laser scanners, and appropriate software modules. CNAO-OTS accuracy in markers 3D localization was assessed experimentally to be around 0.5 mm (see Table I).

At each commissioning session, a PPS reference position bringing one of the sockets fixed on the couch surface in correspondence of the room isocenter was recorded by means of the laser tracker. During this operation a micro-controlled moving platform was used to adjust the control point at room isocenter precisely (Figure 2, right box). Different operators then moved the PPS according to the above reported sets of rotations and translations. After each motion, all sockets were recorded by means of the laser tracker. Off-line data analysis was performed in Matlab (The MathWorks, Inc.), according to the following steps:

OTS Commissioning

1. Sockets reference configuration, as acquired by the laser tracker, was roto-translated according to the set of parameters sent to the PPS control system for each motion. 2. Measured and reference roto-translated socket configurations were compared with expected differences below the rotational and translational thresholds, according to specifications. Differences between these two corresponding dataset were assessed statistically. Optical Tracking System (OTS) The OTS is a non-invasive system to monitor passive markers placed on patient’s skin or immobilization devices, allowing continuous monitoring of their position and movement during the first phases of the treatment session (first, rough alignment) and during dose delivery (control, motion compensation). The CNAO-OTS is based on the SMART-D optical tracking (BTS Spa, Milan, Italy) equipped with three near infrared (890 nm) digital cameras for stereo-photogrammetric real-time 3-D localization of passive markers or laser spots (projected onto patient’s skin surface). The system provides also a real-time visual feedback for operators and ­on-line detection of the breathing phase. The CNAO-OTS layout was optimized to allow the integration of its functionalities under a unique control and manmachine interface, minimizing the impact on other hardware installed in the treatment room. Cameras are mounted on a custom suspension placed above the nozzle to ensure visibility (see Figure 1). At present, the system features point-based localization with passive markers which are fitted on patient immobilization device. Further developments are envisaged to ensure patient surface detection, by integrating additional hardware

The OTS accuracy in tracking passive markers depends upon the residuals errors of the stereometric calibration. The CNAO-OTS calibration is performed through a static acquisition of a calibration tern and a dynamic acquisition of a wand sequence (38). After nominal calibration, an additional procedure is required to map the reference system of the OTS (defined by the static acquisition of the marked tern) to the room isocentric reference system, centered in the room isocenter. The mapping operator (4 3 4 rototranslation matrix in homogenous co-ordinates) is calculated by acquiring a marker configuration fixed on the QA phantom provided for X-ray QA and geometric calibration (Brandis ­Medizintechnik Vertriebs GmbH, Weinheim, Germany) and rigidly registering the acquired marker positions with those measured by means of the laser tracker which were used as reference. The reference configuration was computed relying on the geometrical properties of the QA Brandis phantom and double checked by means of repeated laser tracker measurements, as described in paragraph system commissioning (3.2.4). CNAO-OTS marker localization accuracy was assessed and is reported by quantifying the quality of the calibration and of the mapping residuals. In order to evaluate the system accuracy in mapping a set of passive markers, we acquired a calibration wand in the working volume, immediately after the calibration of the OTS. Distances between markers were elaborated offline and results will be presented as the average of their standard deviation. Patient Verification System (PVS) The CNAO-PVS is a stereoscopic X-ray imaging device, contained in a cylindrical rotating structure hanging from the bunker ceiling. The PVS structure is supported by vertical bars and stabilized by horizontal bars fixed to the concrete bunker walls. Both X-ray tubes and flat panels are deployed only during imaging, by pivoting and sliding tubes and flat panels, respectively. The time needed for deployment is less than 30 seconds. The PVS is installed in rooms 1 and 3, where a horizontal beam line is available. No PVS is installed in room 2 (the one with two beam lines) because of the presence of the vertical nozzle. The PVS provides kilovoltage high quality X-ray images in double projection, relying on high-resolution flat panel detectors, which are used for patient setup verification

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

308

Pella et al.

by means of 2D-3D registration. Digitally Reconstructed Radiographs (DRRs), extracted from the treatment planning CT are optimized to fit the in-room geometry of PVS and thus compared with the acquired projections resulting in angle corrections. Patient setup verification by means of PVS can be repeated any time during treatment (typically between different irradiation fields) to ensure the consistency with treatment plan geometry. PVS X-ray components consist of two image receptors, two X-ray generators and tubes (boxes Table I for technical specifications). A commercial image registration software (VeriSuite®, MedCom GmbH, Darmstadt), provides several different software functionalities for PVS operations along with different modes for calculation of the correction vector (manual alignment, manual and automatic or fully automatic). PVS Commissioning Measurements of PVS accuracy in linear and rotational movements were performed by an external company (GDV Systems GmbH, Bad Schwartau, Germany). Several sockets were applied to the moving parts of the cylindrical structure and registered by the GDV proprietary laser tracker. Two sets of acquisitions were performed: (i) the first one was focused on the fine tuning of PVS installation and centering and related quantification of residual inaccuracies of PVS rotations, (ii) the second one on the mechanical accuracy of flat panels and X-ray tubes deployment. The mechanical center of rotation of the PVS was also measured and compared with the corresponding nominal position with expected negligible differences.

– Acquisition of stereoscopic X-ray images – 2D-3D registration and storage of correction parameters (CVCT) to refine the alignment to the treatment planning CT – Application of the calculated CVCT corrections to the PPS configuration Next steps describe how roto-translations were applied to the phantom to estimate deltas between these values and the correction vector suggested by the PVS software. – Application of known roto-translations (known shifts, KS) to the PPS configuration – Acquisition of stereoscopic X-ray images, automatic 2D-3D registration and storage of correction parameters (CVKS) – Comparison of CVKS vs. KS, with expected differences below the thresholds defined in system specifications (1 mm, 1°) Further analysis was carried out to analyze the performance of image registration procedures in extra-cranial sites. Due to the oblique projection angles of the stereoscopic imaging system, an advanced imaging procedure was designed to maximize the quality of the two projections used for registration, by combining PPS configurations with each projection acquisition. The reason behind this approach is that in particular cases (e.g. imaging of thorax or pelvic district), oblique projections lead X-rays to pass through a considerable length of patient tissues, which entails significant X-ray attenuation

Clinical Dry-run The accuracy of PPS, PVS and image registration software application was tested by means of an image-based registration procedure. We used an anthropomorphic tissue-equivalent phantom (Alderson Rando®, The Phantom Laboratory. P.O. Box 511, Salem, NY 12865-0511 USA) and simulated a clinical-like workflow consisting of the following steps. First steps are devoted to minimize the inaccuracies in phantom repositioning: – Drive of the PPS in a reference position, as determined by the treatment planning CT

Figure 3:  Registration accuracy measured with corresponding landmarks placed in the Rando phantom.

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

Design of a Patient Positioning and Setup Verification System for Particle Therapy 309 (i.e. poor image contrast). This advanced workflow includes the following steps: – PPS table rotation around vertical axis at 65° – Imaging with X-ray tube and panel A (quasi RL ­projection) – PPS table rotation at 110° – Imaging with X-ray tube and panel B (quasi LR ­projection) – Automatic registration with the acquired A/B images The accuracy of the advanced workflow was verified by applying random PPS displacements from the nominal position, considering both thorax and pelvis anatomy. The accuracy of the registration procedure was checked by computing the residual RMS displacement of selected landmarks, that were previously implanted in the radio-equivalent phantom (Figure 3). Statistical Analysis Concerning the commissioning of the PPS, statistical analysis was applied to investigate two different aspects: Case 1: Investigate whether the presence of weight (added to simulate the presence of a patient) may influence the accuracy in couch motion. Case 2: Investigate whether significant difference in PPS motion accuracy exist among the three rooms. The Lilliefors test was applied to investigate whether the distribution of experimental data was normal. In those cases in which data were not normally distributed, non parametric statistical analysis was applied (Kolmogorov-Smirnov, Friedman tests). When post-hoc comparison was applied, non parametric test (Tukey-Kramer) was used in the case of not normally distributed data, whilst analogous parametric tests were used when appropriate.

Quality Assurance Tests In order to evaluate the consistency of all patient positioning systems and allow performing an efficient daily accuracy check, a set of quality assurance (QA) tests have been designed. It is important to point out that the following procedure is currently performed every day before the clinical operation. The procedure entails different steps, as detailed below: – Mechanical fixation of the calibration phantom on the treatment couch in a predetermined position indicated by the couch indexing system. The phantom features internal radiopaque seeds and surface reflective markers in known, previously measured positions (see ­Figure 4). – Automatic setup of the phantom at three different positions at variable couch angles around the vertical axis (Pos 1: 0° couch rotation, Pos 2: 190°, Pos 3: 290°). These positions were defined, measured and stored in order to bring the center of the calibration phantom in correspondence to the room isocenter. – Acquisition of the external markers coordinates (Pos 1-3; reflective markers, as seen by the OTS) and of double projections with X-ray images (internal radiopaque seeds as seen by the PVS). – Evaluation of displacements of reflective markers with respect to the nominal position and displacement of the radiopaque seed aligned to the room isocenter. The first measure is provided by the OTS and describes the 3D displacements of the passive reflective marker configuration with respect to a reference models. The second measure is the distance between the projection of the reference central seed in the phantom and the center of the flat panels. The measurement performed at different couch rotations allows to span over a 180° motion range.

Figure 4:  Brandis phantom in position during QA procedures in room 3 (left box), and 3D rendering (right box).

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

310

Pella et al.

Results

reported for rooms 1 and 3. Mean and standard deviation of 3D reconstruction errors were found to be 0.23  0.24 mm in room 1 (n 5 6) and 0.22  0.24 mm in room 3 (n 5 3).

PPS In this paragraph, results of data collected during the final commission tests are reported. Table II reports median and max values of 3D displacements between measured and estimated positions of 7 fiducial sockets fixed on the treatment couch, after the application of 20 random motion with and without additional load, in the three CNAO treatment rooms. Table II Median and max values of 3D displacements between measured and estimated positions. Room 1

Room 2

Room 3

No weight Weight

No weight Weight

No weight Weight

Median [mm]

0.15

0.17

0.21

0.15

0.11

0.09

Max [mm]

0.33

0.27

0.35

0.44

0.24

0.24

Single 3D deviations for each fiducial in each room under both experimental conditions (weight-no weight) are reported in Figure 5. OTS Results of the CNAO-OTS commissioning are reported in terms of calibration plus mapping residuals. Results are

Two-dimensional (2D) residuals of each camera are reported in Table III. Table III Residual calibration errors for each-camera.

Camera 1 Camera 2 Camera 3

Room 1

Room 3

Mean [pixel] Stddev [pixel]

Mean [pixel] Stddev [pixel]

0.13 0.13 0.12

0.11 0.11 0.10

0.13 0.12 0.11

0.11 0.11 0.09

Standard deviation of 3D distances between markers were found to be 0.26 mm, 0.39 mm, 0.62 mm for rooms 1, 2 and 3 respectively. PVS and Registration Performance PVS overall centering accuracy (measured as the difference between the measured and theoretical center of rotation of the cylindrical structure) was 0.20, 0.20 and 0.10 mm (X, Y, Z) with respect to the room coordinate system in room 1. Similar result was obtained in room 3 (0.28, 0.46, 0.01 mm). Rotational inaccuracies affecting PVS installation (tilt of surfaces supposed to be perfectly horizontal) were found to be 0.01° in rooms 1 and 3, respectively.

Figure 5:  PPS and couch without additional load (left panel); PPS and couch with additional load (right panel). In both cases marker 0 correspond to the one aligned at the isocenter.

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

Design of a Patient Positioning and Setup Verification System for Particle Therapy 311

Figure 6:  Rotational accuracy of the PVS in room 1 (left) and room 3 (right), reporting angular errors for flat panels (FP) and X-rays tubes (RR).

The rotational motion accuracy of the PVS is reported in ­Figure 6, where angular displacements between measured and theoretical positions of laser tracker sockets placed on PVS flat panels and tubes after a set of 12 PVS structure rotations (range: 293°-90°) in rooms 1 and 3 are reported. Table IV reports the results of flat panels and tubes deployment repeatability in terms of interquartile range, which were obtained from 12 consecutive deployments. Table IV Interquartile range [mm] of flat panel and tubes deployment repeatability in rooms 1 and 3.

3°-1° quartile 3°-1° quartile

FP1

FP2

RR1

RR2

0.11

Room 1 0.11 0.225

0.16

0.11

Room 3 0.19 0.19

0.39

positions were used (min-max translation values: 25 mm: 15 mm; min-max rotation values: 21°:11°). Maximum deviations between applied known roto-translations and corrective motion estimated by the image-based procedure did not exceeded 0.3 mm and 0.2° for translations and rotations, respectively. Similar results for the Alderson Rando thorax and pelvis phantom are reported in Table VI. Results reported in Table VI concern the advanced workflow procedure, that was tested in room 3. In this case six representative positions were used (min-max translation values: 220 mm:120 mm; min-max rotation values: 23°:13°). Table VI Mean and max values of 2D-3D registration performance with advance workflow, Rando Thorax, and Pelvis in room 3. Delta

Results of overall accuracy in image-based 2D-3D registration assessed during the last commissioning experimental session are reported in Table V (Rooms 1 and 3, respectively) for the Alderson Rando head phantom. Five representative Table V Mean and max values of 2D-3D registration performance, Rando head, rooms 1 and 3. Delta Traslations [mm] Room 1 Room 3

Rotations [°]

Mean

Max

Mean

Max

20.05 0.06

0.3 0.3

0.01 0.03

0.2 0.1

Traslations [mm] Thorax Pelvis

Rotations [°]

Mean

Max

Mean

Max

0.38 0.41

0.8 0.8

0.19 0.29

0.3 0.6

Statistical Analysis Under a general scenario of specifications fulfillment, with PPS motion 3-D deviations below 0.3 mm (except for few outliers in room 2), non parametric statistical analysis was applied to test the influence of presence of weight on the treatment couch and possible different performance among different rooms. Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

312

Pella et al.

Case 1. Influence of weight: Two-sample KolmogorovSmirnov test shows a significant difference among the two investigated conditions (absence or presence of an additional weight, p-value 5 0.0085), with slightly better results in the first case. Case 2. Difference in between rooms: The ANOVA equivalent Friedman test highlights significant differences among the three rooms (p-value  2·10215). The post-hoc comparison (Tukey-Kramer test) indicates that the best performing system was the one in room 3, whereas PPSs installed in room 2 and room 1 were not significantly different at 95% confidence level. Concordance of PVS and OTS during QA QA tests results are reported for room 1 (n 5 546) and room 3 (n 5 59). Mean  standard deviation of the 2D distance between the center of the flat panels and the center of the reference seed of the QA/calibration phantom in the three QA PPS configurations (rotations around vertical axis: 0°, 190°, 290°) were 0.25  0.14 mm (max value: 0.83) and 0.21  0.10 mm (max value: 0.54). QA tests results involving the CNAO-OTS system are also reported (Figure 7) in terms of difference between measured position of the marker configuration on the QA/calibration phantom and the corresponding reference.

scanning. The activities for design, installation and commissioning lasted around 3 years in the framework of a tight and fruitful collaboration between CNAO and the system suppliers. High precision measurement devices, widely applied in industry and also used to align the components of the CNAO synchrotron, have been used to assess the accuracy of PPS and PVS motion. The PPS exhibited an overall average setup accuracy within 0.3 mm in all the treatment rooms, thus fulfilling particularly stringent specifications. Median errors for the three rooms ranged between 0.09-0.21 mm (minimum value in room 3, additional load condition; maximum value in room 2, without additional load). The maximum error (0.44 mm) was observed in room 2 (additional load). Even if this value is greater than 0.3 mm, it is important to point out that it refers to a landmark placed on the opposite side of the treatment area, hence with a negligible influence in terms of accuracy within the treatment field. The set of measurements with the corner cube placed at the isocenter (marker 0 in Figure 6) showed values always below 0.3 in all the rooms and conditions. Only one slightly overthreshold value (0.32 mm) was recorded in room 2. Statistical differences in PPS motion accuracy were observed as a function of the presence of the 40 kg weight on the

Figure 7:  Mean and standard deviation of absolute errors (delta X, Y, Z) between measured position of the marker configuration on the QA/calibration phantom by the OTS and the corresponding reference in room 1 (left box) and room 3 (right box).

Discussion In this paper, we report the performance of patient positioning and setup verification devices installed, commissioned and clinically operating at the CNAO facility, the first Italian and second European center for particle therapy with protons and carbon-ions delivered with active

treatment couch, with slightly better results observed under the no-weight condition. The PPS in room 3 showed best performance with respect to the PPSs in rooms 1 and 2. Although differences were not clinically relevant, with all the systems performing within specifications, PPS in room 3 was the last one being commissioned, thus inheriting the fine tuning experience matured in rooms 1 and 2.

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

Design of a Patient Positioning and Setup Verification System for Particle Therapy 313 The application of an infra-red optical tracking system finds its rationale in time reduction for patient setup by providing a first, fast patient alignment before image-based refinement. Moreover, the CNAO-OTS system provides beam interlock generation in case of undesired patient motion, and is ready for on-line breathing phase detection for time-resolved (gated) irradiation. The system features high intrinsic accuracy, which allowed obtaining an overall high performance in markers detection, with sub-millimeter 3-D peak errors within the 0.5 3 0.5 3  0.5 m3 working volume. Current experience of daily QA demonstrates fairly stable performance of CNAOOTS, thus limiting the need of system recalibration in occasion of the preventive maintenance activities planned every six months. The rotational and linear motion accuracy of the PVS revealed mechanical accuracy of the system full in specification. Also, the overall accuracy of the image-based patient verification and correction resulted well within the limits defined in the design phase. According to the results reported in Tables V-VI, maximum deviations in correcting known shifts applied to the PPS were consistently within specifications on the radio-equivalent anthropomorphic Rando phantom. The agreement of PVS and OTS at daily QAs was assessed and revealed a fairly high match between the two systems. As a whole, the clinical workflow for patient treatment is well established, with more than 90 patients treated at CNAO for cranial and extra-cranial pathologies since the start of clinical operation. According to our initial clinical experience, rare are the cases in which the radiation oncologists require more than one image-based registration procedure following the OTS preliminary correction. This also testifies the efficacy of the casting devices applied for patient immobilization. The agreement among OTS and PVS indications in terms of correction vectors is fairly high with more pronounced deviations in the pelvic area, as intuitively expected. The preparation of setup depends by multiple factors, including patient clinical condition and medical doctors indications. They are also responsible for the clinical evaluation of the in-room imaging vs. DRRs overlay. The time required for this check is strictly dependent on the clinical scenario and is therefore hardly predictable. However, according to our experience, an average setup-time of 15 minutes for cranial and 25 minutes for extra-cranial districts is required. In this second case extra time is necessary to compensate for larger displacements, especially in the pelvic region. The experience matured in the commissioning and start of the clinical operation in room 1 and 3 represents the technological and methodological basis for the in-room imaging project in CNAO room 2, featuring multiple X-ray projections and 2D-3D registration, as well as volumetric (CBCT) reconstruction capabilities for soft tissue visualization and 3D-3D registration.

Conflict of Interest We certify that regarding this paper, no actual or potential conflicts of interests exist; the work is original, has not been accepted for publication nor is concurrently under consideration elsewhere. All the authors have contributed directly to the planning, execution or analysis of the work reported or to the writing of the paper. References 1. Kraft, G. Radiobiological effects of very heavy ions: inactivation, inductions of chromosome aberrations and strand breaks. Nuclear Science Applications 3, 1-28 (1987) 2. Kraft, G. Radiobiological Effects of Highly Charged Ions: Their Relevance for Tumor Therapy and Radioprotection in Space, The Physics of Highly and Multiply Charged Ions Ch. 10 Ed. F. J. Currell Kluwer Academic Publisher (2002). 3. Kramer, M., Scholtz, M. Treatment planning for heavy-ion radiotherapy: calculation and optimization of biologically effective dose. Phys Med Biol 45, 3319-3330 (2000). DOI: 10.1088/0031-9155/45/11/314 4. Orecchia, R., Zurlo, A., Loasses, A., Krengli, M., Tosi, G., Zurrida, S., Zucali, P., Veronesi, U. Particle beam therapy (Hadrontherapy): basis for interest and clinical experience. European Journal of Cancer 34, 459-468 (1998). DOI: 10.1016/ S0959-8049(97)10044-2 5. Schulz-Ertner, D., Tsujii, H. Particle radiation therapy using proton and heavier ion beams. J Clin Oncol 25, 953-964 (2007). DOI: 10.1200/JCO.2006.09.7816 6. Schulz-Ertner, D., Karger, C. P., Feuerhake, A., Nikoghosyan, A., Combs, S. E., Jäkel, O., Edler, L., Scholz, M., Debus, J. Effectiveness of carbon ion radiotherapy in the treatment of skull-base chordomas. Int J Radiat Oncol Biol Phys 68, 449-457 (2007). DOI: 10.1016/ j.ijrobp.2006.12.059 7. Sisterson, J. Worldwide charged particle patient totals. Particles 23, 14 (1999). 8. Staab, A., Rutz, H. P., Ares, C., Timmermann, B., Schneider, R., Bolsi, A., Albertini, F., Lomax, A., Goitein, G., Hug, E. Spotscanning-based proton therapy for extracranial chordoma. Int J Radiation Oncology Biol Phys 81, 489-496 (2011). DOI: 10.1016/ j.ijrobp.2011.02.018 9. Tsujii, H., Mizoe, J., Kamada, T., Baba, M., Tsuji, H., Kato, H., Kato, S., Yamada, S., Yasuda, S., Ohno, T., Yanagi, T., Imai, R., Kagei, K., Hara, R., Hasegawa, A., Nakajima, M., Sugane, N., Tamaki, N., ­Takagi, R., Kandatsu, S., Yoshikawa, K., Kishimoto, R., Miyamoto, T. Clinical results of carbon ion radiotherapy at NIRS. J Radiat Res 48, A1-A13 (2007). DOI: 10.1269/jrr.48.A1 10. Blakely, E. A., Ngo, F. Q. H., Curtis, S. B., Tobias, C. A. Heavy-ion radiobiology: cellular studies. Adv Rad Biol 11, 295-389 (1984). 11. Suit, H., Goldberg, S., Niemierko, A., Trofimov, A., Adams, J., Paganetti, H., Chen, G. T., Bortfeld, T., Rosenthal, S., L ­ oeffler, J., Delaney, T. Proton beams to replace photon beams in radical dose treatments. Acta Oncol 42(8), 800-8 (2003). DOI: ­ 10.1080/02841860310017676 12. Jones, B. The case for particle therapy. Br J Radiology 79, 24-31 (2006). DOI: 10.1259/bjr/81790390 13. Tsujii, H., Kamada, T. A review of update clinical results of carbon ion radiotherapy. Jpn J Clin Oncol 42(8), 670-685 (2012). DOI:10.1093/jjco/hys104 14. Riboldi, M., Orecchia, R., Baroni, G. Real-time tumour tracking in particle therapy: technological developments and future perspectives. The Lancet Oncology 13, 383-391 (2012). DOI: 10.1016/ S1470-2045(12)70243-7

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015

314

Pella et al.

15. Dawson, L. A., Jaffray, D. A. Advances in image-guided radiation therapy. J Clin Oncol 25, 938-946 (2007). DOI: 10.1200/ JCO.2006.09.9515 16. Kessler, M. L. Image registration and data fusion in radiation therapy. British Journal of Radiology 79, S99-S108 (2006). DOI: 10.1259/ bjr/70617164 17. Balter, J. M., Kessler, M. L. Imaging and alignment for image-guided radiation therapy. J Clin Oncol 25, 931-937 (2007). DOI: 10.1200/ JCO.2006.09.7998 18. Baroni, G., Riboldi, M., Spadea, M. F., Tagaste, B., Garibaldi, C., Orecchia, R., Pedotti, A. Integration of enhanced optical tracking techniques and imaging in IGRT. J Radiat Res 48(Suppl.), A61-A74 (2007). DOI: 10.1269/jrr.48.A61 19. Sharpe, C., Moseley, D. J. Image guidance: treatment target localization systems. Front Radiat Ther Oncol 40, 72-93 (2007). DOI: 10.1159/000106029 20. Warlick, B. W. Image-guided radiation therapy: techniques and strategies. Community Oncology 5, 86-92 (2008). DOI: 10.1016/ S1548-5315(11)70437-5 21. Xing, L., Thorndyke, B., Schreibmann, E., Yang, Y., Li, T. F., Kim, G. Y., Luxton, G., Koong, A. Overview of image-guided radiation therapy. Medical Dosimetry 31, 91-112 (2006). DOI: 10.1016/ j.meddos.2005.12.004 22. Durante, M., Loeffler, J. S. Charged particles in radiation oncology. Nat Rev Clin Oncol 7, 37-43 (2010). DOI: 10.1038/nrclinonc.2009.183 23. Xu, Q., Hamilton, R. A novel respiratory gating method based on automated analysis of ultrasonic diaphragm motion. Med Phys 32, 2124 (2005). DOI: http://dx.doi.org/10.1118/1.2178451 24. Li, R., Lewis, J. H., Cerviño, L. I., Jiang, S. B. A feasibility study of markerless fluoroscopic gating for lung cancer radiotherapy using 4DCT templates. Phys Med Biol 54, N489-N500 (2009). DOI: 10.1088/0031-9155/54/20/N03 25. Lin, T., Cerviño, L. I., Tang, X., Vasconcelos, N., Jiang, S. B. Fluoroscopic tumor tracking for image-guided lung cancer radiotherapy. Phys Med Biol 54, 981-992 (2009). DOI: 10.1088/0031-9155/54/4/011 26. Cui, Y., Dy, J. G., Sharp, G. C., Alexander, B. M., Jiang, S. B. Learning Methods for Lung Tumor Markerless Gating in Image-guided Radiotherapy. Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2008). DOI: 10.1145/1401890.1401998 27. Lewis, J. H., Li, R., Watkins, W. T., Lawson, J. D., Segars, W. P., Cerviño, L. I., Song, W. Y., Jiang, S. B. Markerless lung tumor

tracking and trajectory reconstruction using rotational cone-beam projections: a feasibility study. Phys Med Biol 55, 2505-2522 (2010). DOI: 10.1088/0031-9155/55/9/006 28. Hoisak, J. D., Sixel, K. E., Tirona, R., Cheung, P. C., Pignol, J. P. Prediction of lung tumour position based on spirometry and on abdominal displacement: accuracy and reproducibility. Radiotherapy and Oncology 78, 339-346 (2006). DOI: 10.1016/j. radonc.2006.01.008 29. Amies, C., Bani-Hashemi, A., Celi, J. C., Grousset, G., Ghelmansarai, F., Hristov, D., Lane, D., Mitschke, M., Singh, A., Shukla, H., Stein, J., Wofford, M. A multi-platform approach to image guided radiation therapy (IGRT). Medical Dosimetry 31, 12-19 (2006). DOI: 10.1016/j.meddos.2005.12.013 30. Particle Therapy Cooperative Group 2011 [online], http://ptcog.web. psi.ch 31. Liauw, S. L., Connell, P. P., Weichselbaum, R. R. New paradigms and future challenges in radiation oncology: an update of biological targets and technology. Sci Transl Med 5(173), 173sr2 (2013). DOI: 10.1126/scitranslmed.3005148 32. Lodge, M., Pijls-Johannesma, M., Stirk, L., Munro, A. J., De Ruysscher, D., Jefferson, T. A systematic literature review of the clinical and cost-effectiveness of hadron therapy in cancer. Radiotherapy and Oncology 83, 110-122 (2007). DOI: 10.1016/ j.radonc.2007.04.007 33. Rossi, S. The status of CNAO. Eur Phys J Plus 126, 78 (2011). DOI: 10.1140/epjp/i2011-11078-8 34. Krengli, M., Orecchia, R. Medical aspects of the National Centre for Oncological Hadronterapy, Centro Nazionale Adroterapia Oncologica (CNAO) in Italy. Radiotherapy and Oncology 73, S21-S23 (2004). DOI: 10.1016/S0167-8140(04)80007-0 35. Orecchia, R., Fossati, P., Rossi, S. The national center for oncological hadron therapy: status of the project and future clinical use of the facility. Tumori 95, 169-176 (2009). DOI: 10.1700/422.5005 36. Riboldi, M., Pella, A., Tagaste, B., Baroni, G., Ciocca, M., Rossi, S., Orecchia, R. Installation and preliminary testing of a newly designed patient setup and monitoring system for particle beam therapy. Proceedings of PTCOG 48 Heidelberg 91 (2009). DOI: 10.3205/09ptcog161 37. Martin, C., Bachtold, D., Schar, H. Specification 6DPPS & Diagnostic Device. Technical Report (2007). 38. Borghese, N. A., Cerveri, P. Calibrating a video camera pair with a rigid bar. Pattern Recognition 33, 81-95 (2000) Received: June 10, 2013; Revised: July 25, 2013; Accepted: September 3, 2013

Technology in Cancer Research & Treatment, Volume 13, Number 4, August 2014

Downloaded from tct.sagepub.com at GEORGIAN COURT UNIV on May 5, 2015