Clinical use of stereoscopic X-ray positioning of patients treated with conformal radiotherapy for prostate cancer

Patient-specific three-dimensional image reconstruction from a single X-ray projection using a convolutional neural network for on-line radiotherapy applications

Background and purpose: Radiotherapy is commonly chosen to treat thoracic and abdominal cancers. However, irradiating mobile tumors accurately is extremely complex due to the organs' breathing-related movements. Different methods have been studied and developed to treat mobile tumors properly. The combination of X-ray projection acquisition and implanted markers is used to locate the tumor in two dimensions (2D) but does not provide three-dimensional (3D) information. The aim of this work is to reconstruct a high-quality 3D computed tomography (3D-CT) image based on a single X-ray projection to locate the tumor in 3D without the need for implanted markers. Materials and Methods: Nine patients treated for a lung or liver cancer in radiotherapy were studied. For each patient, a data augmentation tool was used to create 500 new 3D-CT images from the planning four-dimensional computed tomography (4D-CT). For each 3D-CT, the corresponding digitally reconstructed radiograph was generated, and the 500 2D images were input into a convolutional neural network that then learned to reconstruct the 3D-CT. The dice score coefficient, normalized root mean squared error and difference between the ground-truth and the predicted 3D-CT images were computed and used as metrics. Results: Metrics' averages across all patients were 85.5% and 96.2% for the gross target volume, 0.04 and 0.45 Hounsfield unit (HU), respectively. Conclusions: The proposed method allows reconstruction of a 3D-CT image from a single digitally reconstructed radiograph that could be used in real-time for better tumor localization and improved treatment of mobile tumors without the need for implanted markers.

Artificial intelligence in image-guided radiotherapy: a review of treatment target localization

Modern conformal beam delivery techniques require image-guidance to ensure the prescribed dose to be delivered as planned. Recent advances in artificial intelligence (AI) have greatly augmented our ability to accurately localize the treatment target while sparing the normal tissues. In this paper, we review the applications of AI-based algorithms in image-guided radiotherapy (IGRT), and discuss the indications of these applications to the future of clinical practice of radiotherapy. The benefits, limitations and some important trends in research and development of the AI-based IGRT techniques are also discussed. AI-based IGRT techniques have the potential to monitor tumor motion, reduce treatment uncertainty and improve treatment precision. Particularly, these techniques also allow more healthy tissue to be spared while keeping tumor coverage the same or even better.

Ct Image Fusion as a Tool for Measuring in 3D the Setup Errors during Conformal Radiotherapy for Prostate Cancer

Aims and Background

The importance of optimal daily patient positioning has been stressed in order to ensure treatment reproducibility and gain in accuracy and precision. We report our data on the 3D setup uncertainty during radiation therapy for prostate cancer using the CT image fusion technique.

Methods

Ten consecutive patients scheduled for radiation therapy for prostate cancer underwent 5 prone position CT scans using an individualized immobilization cast. These different setups were analyzed using the image fusion module of the ERGO 3D-Line Medical System (Milan, Italy) treatment planning system. The isocenter and the body marker displacements were measured.

Results

The 3D isocenter dislocations were quantified: systematic error was Σ3D = 3.9 mm, whereas random error was σ3D = 1 mm. The mean of the minimum displacements was 0.2 ± 1 mm showing that the immobilization device used allows an accurate setup to be obtained. Single direction errors were also measured showing systematic errors, ΣAP = 2.6 mm, ΣLL = 0.6 mm, ΣSI = 3 mm in the anterior-posterior, latero-lateral, superior-inferior direction, respectively. Related random errors were σAP = 1 mm, σLL = 0.6 mm, σSI = 1.2 mm. In terms of accuracy, our uncertainties are similar to those reported in the literature.

Conclusions

By applying the CT image fusion technique, a 3D study on setup accuracy was performed. We demonstrated that the use of an individualized immobilization system for prostate treatment is adequate to obtain good setup accuracy, as long as a high-quality positioning control method, such as the stereoscopic X-ray-based positioning system, is used.

Incorporating imaging information from deep neural network layers into image guided radiation therapy (IGRT)

BACKGROUND AND PURPOSE

To investigate a novel markerless prostate localization strategy using a pre-trained deep learning model to interpret routine projection kilovoltage (kV) X-ray images in image-guided radiation therapy (IGRT).

MATERIALS AND METHODS

We developed a personalized region-based convolutional neural network to localize the prostate treatment target without implanted fiducials. To train the deep neural network (DNN), we used the patient's planning computed tomography (pCT) images with pre-delineated prostate target to generate a large amount of synthetic kV projection X-ray images in the geometry of onboard imager (OBI) system. The DNN model was evaluated by retrospectively studying 10 patients who underwent prostate IGRT. Three out of the ten patients who had implanted fiducials and the fiducials' positions in the OBI images acquired for treatment setup were examined to show the potential of the proposed method for prostate IGRT. Statistical analysis using Lin's concordance correlation coefficient was calculated to assess the results along with the difference between the digitally reconstructed radiographs (DRR) derived and DNN predicted locations of the prostate.

RESULTS

Differences between the predicted target positions using DNN and their actual positions are (mean ± standard deviation) 1.58 ± 0.43 mm, 1.64 ± 0.43 mm, and 1.67 ± 0.36 mm in anterior-posterior, lateral, and oblique directions, respectively. Prostate position identified on the OBI kV images is also found to be consistent with that derived from the implanted fiducials.

CONCLUSIONS

Highly accurate, markerless prostate localization based on deep learning is achievable. The proposed method is useful for daily patient positioning and real-time target tracking during prostate radiotherapy.

Inherent uncertainty involved in six-dimensional shift determination in ExacTrac imaging system

Objective

This study was conducted for the assessment of in-built systematic and random errors in the ExacTrac imaging system due to the software of Brainlab, on that basis; recommending a new quality control programme for ExacTrac imaging system.

Methods

A program was developed to compare the image dataset of real time anthropomorphic pelvic phantom using ExacTrac with the reference image dataset from computed tomography. Images were acquired 20 times in a day, on single sitting for 20 conjugative days. On the basic of these translational and rotational shifts, systematic and random errors were calculated that had arisen due to multiple time image acquisition and image registration between acquired and reference image dataset of the phantom.

Results

Random errors were found as 0·006 cm in right-left (Rt-Lt) direction, 0·008 cm in superior-inferior (Sup-Inf) direction and 0·012 cm in anterior-posterior (Ant-Post) direction. On this basic, margins were calculated using Van Herk formula; it was found that there were 0·02 cm inherent shift in Rt-Lt direction, 0·03 cm in Sup-Inf direction and 0·03 cm in Ant-Post direction.

Conclusion

This study concluded that there was inherent error in ExacTrac system which can be quantified and used as a quality assurance tool for the ExacTrac system.

A novel compound 6D-offset simulating phantom and quality assurance program for stereotactic image-guided radiation therapy system

A comprehensive quality assurance (QA) device cum program was developed for the commissioning and routine testing of the 6D IGRT systems. In this article, both the new QA system and the BrainLAB IGRT system which was added onto a Varian Clinac were evaluated. A novel compound 6D‐offset simulating phantom was designed and fabricated in the Prince of Wales Hospital (PWH), Hong Kong. The QA program generated random compound 6D‐offset values. The 6D phantom was simply set up and shifted accordingly. The BrainLAB ExacTrac X‐ray IGRT system detected the offsets and then corrected the phantom position automatically through the robotic couch. Routine QA works facilitated data analyses of the detection errors, the correction errors, and the correlations. Fifty sets of data acquired in 2011 in PWH were thoroughly analyzed. The 6D component detection errors and correction errors of the IGRT system were all within ±1mm and ±1° individually. Translational and rotational scalar resultant errors were found to be 0.50±0.27mmand0.54±0.23°, respectively. Most individual component errors were shown to be independent of their original offset values. The system characteristics were locally established. The BrainLAB 6D IGRT system added onto a regular linac is sufficiently precise for stereotactic RT This new QA methodology is competent to assure the IGRT system overall integrity. Annual grand analyses are recommended to check local system consistency and for external cross‐comparison. The target expansion policy of 1.5 mm 3D margin from CTV to PTV is confirmed for this IGRT system currently in PWH.

PACS numbers: 87.53.Ly, 87.55.Gh, 87.55.Qr, 87.56.Fc

Is the in vivo dosimetry with the OneDosePlusTM system able to detect intra-fraction motion? A retrospective analysis of in vivo data from breast and prostate patients

Background

The OneDosePlus^TM system, based on MOSFET solid-state radiation detectors and a handheld dosimetry reader, has been used to evaluate intra-fraction movements of patients with breast and prostate cancer.

Methods

An Action Threshold (AT), defined as the maximum acceptable discrepancy between measured dose and dose calculated with the Treatment Planning System (TPS) (for each field) has been determined from phantom data. To investigate the sensitivity of the system to direction of the patient movements, fixed displacements have been simulated in phantom. The AT has been used as an indicator to establish if patients move during a treatment session, after having verified the set-up with 2D and/or 3D images. Phantom tests have been performed matching different linear accelerators and two TPSs (TPS1 and TPS2).

Results

The ATs have been found to be very similar (5.0% for TPS1 and 4.5% for TPS2). From statistical data analysis, the system has been found not sensitive enough to reveal displacements smaller than 1 cm (within two standard deviations). The ATs applied to in vivo treatments showed that among the twenty five patients treated for breast cancer, only four of them moved during each measurement session. Splitting data into medial and lateral field, two patients have been found to move during all these sessions; the others, instead, moved only in the second part of the treatment. Patients with prostate cancer have behaved better than patients with breast cancer. Only two out of twenty five moved in each measurement session.

Conclusions

The method described in the paper, easily implemented in the clinical practice, combines all the advantages of in vivo procedures using the OneDosePlus^TM system with the possibility of detecting intra-fraction patient movements.

Interfraction rotation of the prostate as evaluated by kilovoltage X-ray fiducial marker imaging in intensity-modulated radiotherapy of localized prostate cancer

To quantify the daily rotation of the prostate during a radiotherapy course using stereoscopic kilovoltage (kV) x-ray imaging and intraprostatic fiducials for localization and positioning correction. From 2005 to 2009, radio-opaque fiducial markers were inserted into 38 patients via perineum into the prostate. The ExacTrac/Novalis Body X-ray 6-day image acquisition system (ET/NB; BrainLab AG, Feldkirchen, Germany) was used to determine and correct the target position. During the first period in 10 patients we recorded all rotation errors but used only Y (table) for correction. For the next 28 patients we used for correction all rotational coordinates, i.e., in addition Z (superior-inferior [SI] or roll) and X (left-right [LR] or tilt/pitch) according to the fiducial marker position by use of the Robotic Tilt Module and Varian Exact Couch. Rotation correction was applied above a threshold of 1° displacement. The systematic and random errors were specified. Overall, 993 software-assisted rotational corrections were performed. The interfraction rotation errors of the prostate as assessed from the radiodense surrogate markers around the three axes Y, Z, and X were on average 0.09, -0.52, and -0.01° with standard deviations of 2.01, 2.30, and 3.95°, respectively. The systematic uncertainty per patient for prostate rotation was estimated with 2.30, 1.56, and 4.13° and the mean random components with 1.81, 2.02, and 3.09°. The largest rotational errors occurred around the X-axis (pitch), but without preferring a certain orientation. Although the error around Z (roll) can be compensated on average by a transformation with 4 coordinates, a significant error around X remains and advocates the full correction with 6 coordinates. Rotational errors as assessed via daily stereoscopic online imaging are significant and dominate around X. Rotation possibly degrades the dosimetric coverage of the target volume and may require suitable strategies for correction.

Geometric Performance and Efficiency of an Optical Tracking System for Daily Pre-treatment Positioning in Pelvic Radiotherapy Patients

The purpose of this study was to characterize the accuracy of a novel in-house optical tracking system (OTS), and to determine its efficiency for daily pre-treatment positioning of pelvic radiotherapy patients compared to conventional optical distance indicator (ODI) methodology. The OTS is comprised of a passive infrared stereoscopic camera, and custom control software for use in assisting radiotherapy patient setup. Initially, the system was calibrated and tested for stability inside a radiation therapy treatment room. Subsequently, under an ethics approved protocol, the clinical efficiency of the OTS was compared to conventional ODI setup methodology through 17 pelvic radiotherapy patients. Differences between orthogonal source-to-skin distance (SSD) readings and overall set-up time resultant from both systems were compared. The precision of the OTS was 0.01 ± 0.01 mm, 0.02 ± 0.02 mm, and −0.01 ± 0.06 mm in the left/right (L/R), anterior/posterior (A/P), and cranial/caudal (C/C) directions, respectively. Discrepancies measured between the linac radiographic center in the treatment room and the calibrated origin of the camera (OTS) by two independent observers was submillimeter. Analysis of 146 fractions from 17 patients showed a high correlation between the SSD readings of the OTS and ODI setup methodologies (r = 0.99). The average time for pre-treatment positioning using the OTS couch shift calculation was 2.60 ± 0.69 minutes, and for conventional ODI setup, 3.62 ± 0.82 minutes; the difference of 1.02 minutes was statistically significant (p < 0.001). In conclusion, the OTS is a precise and robust tool for use as an independent check of treatment room patient positioning. The system is indicated as geometrically equivalent to current methods of daily pre-treatment patient positioning with potential for gains in efficiency by decreasing setup times in the treatment room.

Hypofractionated extracranial stereotactic radiotherapy boost for gynecologic tumors: a promising alternative to high-dose rate brachytherapy

The purpose of this study is to report toxicity and outcome results in patients with gynaecological tumours treated with a final boost using extra-cranial stereotactic radiotherapy (SRT) with a linac-based micro-multileaf collimator technique as an alternative to high-dose rate brachytherapy (HDR-BT). Since January 2002, 26 patients with either endometrial (n = 17) or cervical (n = 9) cancer were treated according to this protocol: 45-50.4 Gy external radiotherapy (RT) to the pelvic +/- para-aortic regions followed by a final SRT boost of 2 x 7 Gy to the vaginal vault (4-7 day interval between fractions). Median age was 62 years (37-74 range). Fifteen patients were diagnosed with adenocarcinoma, 7 with squamous-cell carcinoma, and 4 with sarcoma. FIGO stage I (n = 17), stage II (n = 7), and stage III (n = 2). Toxicity was scored according to RTOG/EORTC criteria. No severe (> grade-3) acute urinary or low-gastrointestinal (GI) toxicity was observed during treatment and up to 3 months after treatment completion. Moderate (grade < or = 3) acute urinary or low-GI toxicity was observed in 23% and 35% of patients, respectively. After a median follow-up of 47 months (4-77, range), late urinary, low-GI, and sexual > or = grade-2 (worst score) has been reported in 4%, 12% and 29.4% of patients, respectively. The 3-year loco-regional failure-free and overall survival rates were 96% and 95%, respectively. Preliminary results on feasibility, tolerance, and outcome with SRT are encouraging and may be considered a sound alternative to HDR-BT for gynecologic tumors.

Residual translational and rotational errors after kV X-ray image-guided correction of prostate location using implanted fiducials

PURPOSE

To evaluate the residual errors and required safety margins after stereoscopic kilovoltage (kV) X-ray target localization of the prostate in image-guided radiotherapy (IGRT) using internal fiducials.

PATIENTS AND METHODS

Radiopaque fiducial markers (FMs) have been inserted into the prostate in a cohort of 33 patients. The ExacTrac/Novalis Body™ X-ray 6d image acquisition system (BrainLAB AG, Feldkirchen, Germany) was used. Corrections were performed in left-right (LR), anterior-posterior (AP), and superior-inferior (SI) direction. Rotational errors around LR (x-axis), AP (y) and SI (z) have been recorded for the first series of nine patients, and since 2007 for the subsequent 24 patients in addition corrected in each fraction by using the Robotic Tilt Module™ and Varian Exact Couch™. After positioning, a second set of X-ray images was acquired for verification purposes. Residual errors were registered and again corrected.

RESULTS

Standard deviations (SD) of residual translational random errors in LR, AP, and SI coordinates were 1.3, 1.7, and 2.2 mm. Residual random rotation errors were found for lateral (around x, tilt), vertical (around y, table), and longitudinal (around z, roll) and of 3.2°, 1.8°, and 1.5°. Planning target volume (PTV)-clinical target volume (CTV) margins were calculated in LR, AP, and SI direction to 2.3, 3.0, and 3.7 mm. After a second repositioning, the margins could be reduced to 1.8, 2.1, and 1.8 mm.

CONCLUSION

On the basis of the residual setup error measurements, the margin required after one to two online X-ray corrections for the patients enrolled in this study would be at minimum 2 mm. The contribution of intrafractional motion to residual random errors has to be evaluated.

Schedule for CT image guidance in treating prostate cancer with helical tomotherapy

The aim of this study was to determine the effect of reducing the number of image guidance sessions and patient-specific target margins on the dose distribution in the treatment of prostate cancer with helical tomotherapy. 20 patients with prostate cancer who were treated with helical tomotherapy using daily megavoltage CT (MVCT) imaging before treatment served as the study population. The average geometric shifts applied for set-up corrections, as a result of co-registration of MVCT and planning kilovoltage CT studies over an increasing number of image guidance sessions, were determined. Simulation of the consequences of various imaging scenarios on the dose distribution was performed for two patients with different patterns of interfraction changes in anatomy. Our analysis of the daily set-up correction shifts for 20 prostate cancer patients suggests that the use of four fractions would result in a population average shift that was within 1 mm of the average obtained from the data accumulated over all daily MVCT sessions. Simulation of a scenario in which imaging sessions are performed at a reduced frequency and the planning target volume margin is adapted provided significantly better sparing of organs at risk, with acceptable reproducibility of dose delivery to the clinical target volume. Our results indicate that four MVCT sessions on helical tomotherapy are sufficient to provide information for the creation of personalised target margins and the establishment of the new reference position that accounts for the systematic error. This simplified approach reduces overall treatment session time and decreases the imaging dose to the patient.

Benefits of Six Degrees of Freedom for Optically Driven Patient Set-up Correction in SBRT

To quantify the advantages of a 6 degrees of freedom (dof) versus the conventional 3- or 4-dof correction modality for stereotactic body radiation therapy (SBRT) treatments. Eighty-five patients were fitted with 5–7 infra-red passive markers for optical localization. Data, acquired during the treatment, were analyzed retrospectively to simulate and evaluate the best approach for correcting patient misalignments. After the implementation of each correction, the new position of the target (tumor's center of mass) was estimated by means of a dedicated stereotactic algorithm. The Euclidean distance between the corrected and the planned location of target point was calculated and compared to the initial mismatching. Initial and after correction median±quartile displacements affecting external control points were 3.74±2.55 mm (initial), 2.45±0.91 mm (3-dof), 2.37±0.95 mm (4-dof), and 2.03±1.47 mm (6-dof). The benefit of a six-parameter adjustment was particularly evident when evaluating the results relative to the target position before and after the re-alignment. In this context, the Euclidean distance between the planned and the current target point turned to 0.82±1.12mm (median±quartile values) after the roto-translation versus the initial displacement of 2.98±2.32mm. No statistical improvements were found after 3- and 4-dof correction (2.73±1.22 mm and 2.60±1.31 mm, respectively). Angular errors were 0.09±0.93° (mean±std). Pitch rotation in abdomen site showed the most relevant deviation, being − 0.46±1.27° with a peak value of 5.46°. Translational misalignments were −0.68±2.60 mm (mean±std) with the maximum value of 12 mm along the cranio-caudal direction. We conclude that positioning system platforms featuring 6-dof are preferred for high precision radiation therapy. Data are in line with previous results relative to other sites and represent a relevant record in the framework of SBRT.

Residual set-up errors and margins in on-line image-guided prostate localization in radiotherapy

BACKGROUND AND PURPOSE

Image-guided on-line correction of the target position allows radiotherapy of prostate cancer with narrow set-up margins. The present study investigated the residual set-up error after on-line prostate localization and its impact on margins.

MATERIALS AND METHODS

Prostate localization based on two orthogonal X-ray images of gold markers implanted in the prostate was performed with an on-board imager at four treatment sessions for 90 patients. The set-up error in the sagittal plane residual after couch adjustment was evaluated on lateral verification portal images.

RESULTS

The set-up error was less than 3.0mm in 92% of the cases in the anterior-posterior (AP) direction and in 95% of the cases in the cranio-caudal (CC) direction. The set-up error was dominated by internal prostate motion taking place during the set-up procedure. Set-up margins were calculated using two formalisms: margins designed to ensure a minimum CTV dose of 95% for 90% of the patient population were 3.6mm (AP) and 3.5mm (CC). Patient-independent normal distributed set-up errors would result in margins of 4.3mm (AP) and 4.0mm (CC) to ensure complete CTV inclusion in the PTV with 90% probability.

CONCLUSION

Internal prostate motion during the set-up procedure was the main contributor to residual set-up errors.

Genetic evolutionary taboo search for optimal marker placement in infrared patient setup

In infrared patient setup adequate selection of the external fiducial configuration is required for compensating inner target displacements (target registration error, TRE). Genetic algorithms (GA) and taboo search (TS) were applied in a newly designed approach to optimal marker placement: the genetic evolutionary taboo search (GETS) algorithm. In the GETS paradigm, multiple solutions are simultaneously tested in a stochastic evolutionary scheme, where taboo-based decision making and adaptive memory guide the optimization process. The GETS algorithm was tested on a group of ten prostate patients, to be compared to standard optimization and to randomly selected configurations. The changes in the optimal marker configuration, when TRE is minimized for OARs, were specifically examined. Optimal GETS configurations ensured a 26.5% mean decrease in the TRE value, versus 19.4% for conventional quasi-Newton optimization. Common features in GETS marker configurations were highlighted in the dataset of ten patients, even when multiple runs of the stochastic algorithm were performed. Including OARs in TRE minimization did not considerably affect the spatial distribution of GETS marker configurations. In conclusion, the GETS algorithm proved to be highly effective in solving the optimal marker placement problem. Further work is needed to embed site-specific deformation models in the optimization process.

A feasibility study of image-guided hypofractionated conformal arc therapy for inoperable patients with localized non-small cell lung cancer

We treated 36 cases of stage I/II non-small cell lung cancer in inoperable patients. Treatments were planned to a total isocenter dose of 60Gy (8x7.5Gy) based on a dynamic field shaping arc, employing one arc to span as much area as possible and if needed additional weighted segments. The 2 year infield progression free probability is 65%. Disease-specific survival is 75% at 2 years. No patients experienced grade 3-4 toxicity.

Optical Tracking Technology in Stereotactic Radiation Therapy

The last decade has seen the introduction of advanced technologies that have enabled much more precise application of therapeutic radiation. These relatively new technologies include multileaf collimators, 3-dimensional conformal radiotherapy planning, and intensity modulated radiotherapy in radiotherapy. Therapeutic dose distributions have become more conformal to volumes of disease, sometimes utilizing sharp dose gradients to deliver high doses to target volumes while sparing nearby radiosensitive structures. Thus, accurate patient positioning has become even more important, so that the treatment delivered to the patient matches the virtual treatment plan in the computer treatment planning system. Optical and image-guided radiation therapy systems offer the potential to improve the precision of patient treatment by providing a more robust fiducial system than is typically used in conventional radiotherapy. The ability to accurately position internal targets relative to the linac isocenter and to provide real-time patient tracking theoretically enables significant reductions in the amount of normal tissue irradiated. This report reviews the concepts, technology, and clinical applications of optical tracking systems currently in use for stereotactic radiation therapy. Applications of radiotherapy optical tracking technology to respiratory gating and the monitoring of implanted fiducial markers are also discussed.

Assessment of secondary patient motion induced by automated couch movement during on-line 6 dimensional repositioning in prostate cancer treatment

BACKGROUND AND PURPOSE

The purpose of this study is to assess retrospectively secondary patient motion induced by 6D patient setup correction.

MATERIALS AND METHODS

For 104 patients, treated with Novalis, 6D setup correction prior to treatment was performed by ExacTrac5.0/NovalisBody in combination with the Robotic Tilt Module mounted underneath the Exact Couch top. This 6D correction might induce additional setup errors due to patient reaction against the rotations. To evaluate induced secondary motion, the 6D setup correction is verified and evaluated with respect to the tolerance limits.

RESULTS

The majority of measured secondary motions are found within the tolerance limits. Detected secondary motions are mostly found in longitudinal shifts and lateral rotations, and mainly found in only 1 dimension during the same verification. The verifications indicate that the patient population can be divided into a group that hardly moves and a group that moves throughout all 6D setup corrections. The patient's behavior can be predicted by the evaluation of the first five fractions as none of the patients demonstrate a learning curve during the treatment.

CONCLUSIONS

6D setup correction does not induce secondary motion for the majority of the patients and can therefore be applied for all treatment indications.

X-ray–assisted positioning of patients treated by conformal arc radiotherapy for prostate cancer: Comparison of setup accuracy using implanted markers versus bony structures

PURPOSE

The aim of this study was to compare setup accuracy of NovalisBody stereoscopic X-ray positioning using implanted markers in the prostate vs. bony structures in patients treated with dynamic conformal arc radiotherapy for prostate cancer.

METHODS AND MATERIALS

Random and systematic setup errors (RE and SE) of the isocenter with regard to the center of gravity of three fiducial markers were measured by means of orthogonal verification films in 120 treatment sessions in 12 patients. Positioning was performed using NovalisBody semiautomated marker fusion. The results were compared with a control group of 261 measurements in 15 patients who were positioned with NovalisBody automated bone fusion. In addition, interfraction and intrafraction prostate motion was registered in the patients with implanted markers.

RESULTS

Marker-based X-ray positioning resulted in a reduction of RE as well as SE in the anteroposterior, craniocaudal, and left-right directions compared with those in the control group. The interfraction prostate displacements with regard to the bony pelvis that could be avoided by marker positioning ranged between 1.6 and 2.8 mm for RE and between 1.3 and 4.3 mm for SE. Intrafraction random and systematic prostate movements ranged between 1.4 and 2.4 mm and between 0.8 and 1.3 mm, respectively.

CONCLUSION

The problem of interfraction prostate motion can be solved by using implanted markers. In addition, the NovalisBody X-ray system performs more accurately with markers compared with bone fusion. Intrafraction organ motion has become the limiting factor for margin reduction around the clinical target volume.

Optimal port placement and enhanced guidance in robotically assisted cardiac surgery

BACKGROUND

Optimal port placement and enhanced guidance in robotically assisted cardiac surgery is required to improve preoperative planning and intraoperative navigation.

METHODS

Offline optimal port placement is planned on a three-dimensional virtual reconstruction of the patient's computed tomography scan. Using this data, an accurate in vivo port placement can be performed, which is achieved by augmented reality techniques superimposing virtual models of the thorax and the teleoperator arms on top of the real worldview.

RESULTS

A new system incorporating both port placement planning and intraoperative navigation in robotically assisted minimally invasive heart surgery was established to aid the operative workflow. A significant reduction of operation time by improved planning and intraoperative support is anticipated.

CONCLUSIONS

The enhanced intraoperative orientation possibilities may lead to further decrease in operation time and have the continuing ability to improve quality.

IGRT of prostate cancer; is the margin reduction gained from daily IG time-dependent?

The aim of this study was to assess the set-up uncertainties and the possible CTV-PTV margin reduction when adopting daily IGRT. Further, to identify any intrafraction time trends in the prostate movements to ensure the margin reduction gained from IGRT. Fifteen prostate cancer patients treated with IMRT using daily IG of three implanted fiducial markers were included. The interfraction uncertainties were assessed by statistically evaluating the daily prostate marker displacement. The intrafraction uncertainties were represented by the difference in prostate marker displacement before and after beam delivery. To evaluate any intrafraction time trends, the data points were divided into two groups with respect to time duration and statistically analysed. This study confirmed that daily IG considerably reduces the set-up uncertainties. Our results implied that if IGRT is performed on a daily basis, both systematic and random set-up errors will be reduced to a minimum, leading to a required set-up margin of only 1.5 mm. Results from measurements of intrafraction motions in time durations ranging from 2 to 27 min, indicated that a margin enlargement of 1 mm was required to account for the intrafraction uncertainties. The results did not suggest any significant time trends in the intrafraction uncertainties.

Optimal control of set-up margins and internal margins for intra- and extracranial radiotherapy using stereoscopic kilovoltage imaging

In this paper the clinical introduction of stereoscopic kV-imaging in combination with a 6 degrees-of-freedom (6 DOF) robotics system and breathing synchronized irradiation will be discussed in view of optimally reducing interfractional as well as intrafractional geometric uncertainties in conformal radiation therapy. Extracranial cases represent approximately 70% of the patient population on the NOVALIS treatment machine (BrainLAB A.G., Germany) at the AZ-VUB, which is largely due to the efficiency of the real-time positioning features of the kV-imaging system. The prostate case will be used as an example of those target volumes showing considerable changes in position from day-to-day, yet with negligible motion during the actual course of the treatment. As such it will be used to illustrate the on-line target localization using kV-imaging and 6 DOF patient adjustment with and without implanted radio-opaque markers prior to treatment. Small lung lesion will be used to illustrate the system's potential to synchronize the irradiation with breathing in coping with intrafractional organ motion.

Setup accuracy of stereoscopic X-ray positioning with automated correction for rotational errors in patients treated with conformal arc radiotherapy for prostate cancer

We evaluated setup accuracy of NovalisBody stereoscopic X-ray positioning with automated correction for rotational errors with the Robotics Tilt Module in patients treated with conformal arc radiotherapy for prostate cancer. The correction of rotational errors was shown to reduce random and systematic errors in all directions. (NovalisBody and Robotics Tilt Module are products of BrainLAB A.G., Heimstetten, Germany).

Phase II study of a four-week hypofractionated external beam radiotherapy regimen for prostate cancer: Report on acute toxicity

PURPOSE

To evaluate the early side effects of a short course hypofractionated radiotherapy regimen in prostate cancer.

MATERIALS AND METHODS

Three institutions (IRE, AZ VUB, GUH) included 36 patients with T1-T3N0M0 prostate cancer in a phase II study. Patients were treated with 56 Gy in 16 fractions over 4 weeks. Early side effects were scored using the RTOG/EORTC criteria and the international prostate symptom index (IPSI) weekly during treatment and 1 and 2 months afterwards. The results were compared with two control groups of patients previously treated with conventional fractionation at AZ VUB (238 patients) and GUH (114 patients).

RESULTS

None of the patients experienced grade 3-4 toxicity. Grade 1-2 Gastro-intestinal (GI), grade 2 GI, grade 1-2 Genito-urinary (GU) and grade 2 GU toxicity occurred in 75%, 36%, 75% and 44% for the hypofractionation schedule. The corresponding figures were 25-44%, 6-29%, 47-53% and 16-44% for the control groups (p<0.01 for grade 1-2 GI and GU toxicity). Two months after treatment all GU and the majority of GI symptoms had resolved. The IPSI increased from (average +/-1 SD) 5.6+/-4 pre-treatment to 10.0+/-6 during week 2-4 and had normalized (5.2+/-4) two months after treatment.

CONCLUSIONS

Though no grade 3-4 side effects were observed, the investigated schedule results in a marked increase of grade 1-2 early side effects as compared to a conventional regimen. Side effects resolved within two months post-treatment.

Robust frameless stereotactic localization in extra-cranial radiotherapy

In the field of extra-cranial radiotherapy, several inaccuracies can make the application of frameless stereotactic localization techniques error-prone. When optical tracking systems based on surface fiducials are used, inter- and intra-fractional uncertainties in marker three-dimensional (3D) detection may lead to inexact tumor position estimation, resulting in erroneous patient setup. This is due to the fact that external fiducials misdetection results in deformation effects that are poorly handled in a rigid-body approach. In this work, the performance of two frameless stereotactic localization algorithms for 3D tumor position reconstruction in extra-cranial radiotherapy has been specifically tested. Two strategies, unweighted versus weighted, for stereotactic tumor localization were examined by exploiting data coming from 46 patients treated for extra-cranial lesions. Measured isocenter displacements and rotations were combined to define isocentric procedures, featuring 6 degrees of freedom, for correcting patient alignment (isocentric positioning correction). The sensitivity of the algorithms to uncertainties in the 3D localization of fiducials was investigated by means of 184 numerical simulations. The performance of the implemented isocentric positioning correction was compared to conventional point-based registration. The isocentric positioning correction algorithm was tested on a clinical dataset of inter-fractional and intra-fractional setup errors, which was collected by means of an optical tracker on the same group of patients. The weighted strategy exhibited a lower sensitivity to fiducial localization errors in simulated misalignments than those of the unweighted strategy. Isocenter 3D displacements provided by the weighted strategy were consistently smaller than those featured by the unweighted strategy. The peak decrease in median and quartile values of isocenter 3D displacements were 1.4 and 2.7 mm, respectively. Concerning clinical data, the weighted strategy isocentric positioning correction provided the reduction of fiducial registration errors, featuring up to 61.7% decrease in median values (versus 46.8% for the unweighted strategy) of initial displacements. The weighted strategy proved high performance in minimizing the effects of fiducial localization errors, showing a great potential in improving patient setup. The clinical data analysis revealed that the application of a robust reconstruction algorithm may provide high-quality results in patient setup verification, by properly managing external fiducials localization errors.

Performance and characteristics of an IR localizing system for radiation therapy

We report the development of a new system for interactive patient posture, position and respiratory control during radiation therapy treatment. The system consists of an infrared camera, retro-reflective markers and dedicated software that makes it practical to use in the clinic. The system is designed to be used with multiple retro-reflective markers to monitor not only position, but also the posture of the patient in real time. Specific features of the system include: 1. The system reports an absolute misalignment at several points on a patients, and also provides feedbacks on any necessary adjustments in terms of site specific set-up parameters, such as focus to surface distance (PIN), superior and inferior alignment, chest-wall angle, etc. 2. The system is based on the set of predefined templates containing number and position of control points and feedback parameters developed for different treatment sites. 3. A "virtual portal vision" procedure is developed to project organ contours in the beams-eye-view (BEV) based on the marker locations obtained in real time and compare them with digitally reconstructed radiographs (DRRs) from CT simulation. Assuming good correlation between external markers and internal anatomy, the system offers the possibility of mimicking a verification procedure without taking port-films, which can potentially reduce the setup time. In this paper, we concentrate on system properties and performance, while initial applications on a number of clinical sites is ongoing. Accuracy and precision of this system are evaluated in the context of breast/chest treatments using rigid phantoms. The system has an intrinsic uncertainty of +/- 1 mm; and when two systems in different rooms (CT and treatment room) are used for correlating positional information, the uncertainty is less than 2 mm.

Six dimensional analysis with daily stereoscopic x-ray imaging of intrafraction patient motion in head and neck treatments using five points fixation masks

The safety margins used to define the Planning Target Volume (PTV) should reflect the accuracy of the target localization during treatment that comprises both the reproducibility of the patient positioning and the positional uncertainty of the target, so both the inter- and intrafraction motion of the target. Our first aim in this study was to determine the intrafraction motion of patients immobilized with a five-point thermoplastic mask for head and neck treatments. The five-point masks have the advantage that the patient's shoulders as well as the cranial part of the patient's head is covered with the thermoplastic material that improves the overall immobilization of the head and neck region of the patient. Thirteen patients were consecutively assigned to use a five-point thermoplastic mask. The patients were positioned by tracking of infrared markers (IR) fixed to the immobilization device and stereoscopic x-ray images were used for daily on-line setup verification. Repositioning was carried out prior to treatment as needed; rotations were not corrected. Movements during treatment were monitored by real-time IR tracking. Intrafraction motion and rotation was supplementary assessed by a six-degree-of-freedom (6-D) fusion of x-ray images, taken before and after all 385 treatments, with DRR images generated from the planning CT data. The latter evaluates the movement of the patient within the thermoplastic mask independent from the mask movement, where IR tracking evaluates the movement of the mask caused by patient movement in the mask. These two movements are not necessarily equal to each other. The maximum intrafraction movement detected by IR tracking showed a shift [mean (SD; range)] of -0.1(0.7; 6.0), 0.1(0.6; 3.6), -0.2(0.8;5.5) mm in the vertical, longitudinal, and lateral direction, respectively, and rotations of 0.0(0.2; 1.6), 0.0(0.2; 1.7) and 0.2(0.2; 2.4) degrees about the vertical, longitudinal, and lateral axis, respectively. The standard deviations and ranges found with the 6-D fusion demonstrate intrafraction patient displacements of -0.5(1.2; 7.4), 0.3(0.7; 5.3), 0.0(0.7; 5.7) mm in the vertical, longitudinal, and lateral direction, respectively, and rotations of -0.1(0.6; 4.1), 0.1(0.7; 8.3) and -0.2(0.8; 8.2) degrees about the vertical, longitudinal, and lateral axis, respectively. The 6-D fusions are considerably larger (p < 0.05) than detected by IR tracking. This indicates that the external marker tracking underestimates the magnitude of the actual intrafraction motion and rotation of the patient. The intrafraction motion detected for the patients immobilized with a conventional thermoplastic mask was relatively large. The feasibility to reduce this intrafraction movement by the application of alternative five-point thermoplastic mask types was evaluated as a second aim of this study. The preliminary results showed a clear reduction in the range, being an indication for the random movements, of both the intrafraction shift and rotation for both alternative mask types. The 6-D fusion is found a useful tool for a fast evaluation of the actual patient's intrafraction shift and rotation and shows the latter is not negligible and needs to be taken into account additional to the initial setup accuracy when determining the PTV margin.

Clinical Implementation and Efficiency of Kilovoltage Image-Guided Radiation Therapy

This paper describes measurements of clinical efficiency and time requirements associated with image-guided radiation therapy (IGRT). In June 2004, the authors' institution installed an integrated kilovoltage (kV) imaging system attached to a medical linear accelerator for radiographic target localization. Over the past year, 242 patients have been localized with the kV radiographic imaging system for a total of 2,700 fractions. Data were analyzed by reviewing the time required for each patient's IGRT session, broken into both image acquisition and image analysis time. Average IGRT procedure time was reviewed pertaining to months, treatment sessions, disease sites, and radiation therapists. Results showed that the average IGRT procedure time was reduced from 450 to 237 seconds from June 2004 to June 2005. Further analysis revealed that each therapist showed improvement in reducing the IGRT procedure time from the first month of use to the month of June 2005. The routine use of IGRT may ultimately be performed within 3 to 4 minutes, with minimal disruption to the clinical treatment process.

Optically Guided Patient Positioning Techniques

Optical tracking determines an object's position by measuring light either emitted or reflected from the object. The hallmark of optical tracking systems is their high spatial resolution and measurement in real time; such systems can resolve the position of a point source within a fraction of a millimeter and report at a rate of 10 Hz or faster. Several systems have been developed for radiation therapy, all of which track infrared markers attached to the patient's external surface. The positions of the optical markers relative to the target volume, together with the desired marker positions relative to treatment isocenter, are determined during computed tomography simulation. In the treatment room, the real marker positions are measured relative to isocenter; rigid-body mathematics then determine marker displacements from their desired positions and hence target displacement from isocenter. Real-time feedback allows one to correct the patient's position. The first systems were used for intracranial stereotaxis radiotherapy; rigid arrays of optical markers were attached to the patient via a biteplate linkage. Subsequent systems for extracranial radiotherapy tracked external markers to determine patient position and/or gate the radiation beam based on patient motion. Lastly, optical tracking has been integrated with ultrasound or stereoscopic x-ray imaging to determine the position of internal anatomy targets relative to isocenter.

Fractionated stereotactic radiotherapy boost for gynecologic tumors: an alternative to brachytherapy?

PURPOSE

A brachytherapy (BT) boost to the vaginal vault is considered standard treatment for many endometrial or cervical cancers. We aimed to challenge this treatment standard by using stereotactic radiotherapy (SRT) with a linac-based micromultileaf collimator technique.

METHODS AND MATERIALS

Since January 2002, 16 patients with either endometrial (9) or cervical (7) cancer have been treated with a final boost to the areas at higher risk for relapse. In 14 patients, the target volume included the vaginal vault, the upper vagina, the parametria, or (if not operated) the uterus (clinical target volume [CTV]). In 2 patients with local relapse, the CTV was the tumor in the vaginal stump. Margins of 6-10 mm were added to the CTV to define the planning target volume (PTV). Hypofractionated dynamic-arc or intensity-modulated radiotherapy techniques were used. Postoperative treatment was delivered in 12 patients (2 x 7 Gy to the PTV with a 4-7-day interval between fractions). In the 4 nonoperated patients, a dose of 4 Gy/fraction in 5 fractions with 2 to 3 days' interval was delivered. Patients were immobilized in a customized vacuum body cast and optimally repositioned with an infrared-guided system developed for extracranial SRT. To further optimize daily repositioning and target immobilization, an inflated rectal balloon was used during each treatment fraction. In 10 patients, CT resimulation was performed before the last boost fraction to assess for repositioning reproducibility via CT-to-CT registration and to estimate PTV safety margins around the CTV. Finally, a comparative treatment planning study between BT and SRT was performed in 2 patients with an operated endometrial Stage I cancer.

RESULTS

No patient developed severe acute urinary or low-intestinal toxicity. No patient developed urinary late effects (>6 months). One patient with a vaginal relapse previously irradiated to the pelvic region presented with Grade 3 rectal bleeding 18 months after retreatment. A second patient known to suffer from irritable bowel syndrome presented with Grade 1 abdominal pain after treatment. The estimated PTV margins around the CTV were 9-10 mm with infrared marker registration. External SRT succeeded in improving dose homogeneity to the PTV and in reducing the maximum dose to the rectum, when compared to BT.

CONCLUSION

These results suggest that the use of external SRT to deliver a final boost to the areas at higher risk for relapse in endometrial or cervical cancer is feasible, well tolerated, and may well be considered an acceptable alternative to BT.

The patient positioning concept for the planned MedAustron centre

Especially for ion therapy, efficiency in form of patient throughput is becoming increasingly important, and here, patient positioning in treatment room isocenter is a key aspect. In order to ascertain high quality nonetheless, we suggest an alternative to the rigidly installed couch paradigm in form of real-time patient positioning onhand a ceiling mounted infrared photogrammetric system giving positioning information to a novel treatment couch with 6 degrees of freedom integrated on a rolling platform. All MedAustron treatment planning rooms and even the planning CT are not forseen to have a rigidly installed treatment couch.

Target repositioning optimization in prostate cancer: is intensity-modulated radiotherapy under stereotactic conditions feasible?

PURPOSE

To assess repositioning reproducibility of the prostate when treatment setup conditions before radiotherapy (RT) are optimized and internal organ motion is reduced with an endorectal inflatable balloon.

METHODS AND MATERIALS

Thirty-two patients were treated with 64 Gy to the prostate and seminal vesicles using a three-dimensional conformal radiotherapy technique, followed by a boost (two fractions of 5-8 Gy, 3-5 days apart) delivered to a reduced prostate volume (the peripheral tumor bearing zone with 3-mm margins) using intensity-modulated RT. A commercially available infrared-guided stereotactic repositioning system and a rectal balloon were used. Further improvement in repositioning could be obtained with a stereoscopic X-ray registration device matching the pelvic bones during treatment with the corresponding bones in the planning computed tomography (CT). To simulate repositioning reproducibility, CT resimulation was performed before the last boost fraction. Prostate repositioning was reassessed, first after CT-to-CT fusion with the stereotactic metallic body markers of the infrared-guided system, and second after CT-to-CT registration of the pelvic bony structures.

RESULTS

Standard deviations of the prostate (CTV) center of mass shifts in the three axes ranged from 2.2 to 3.6 mm with body marker registration and from 0.9 to 2.5 mm with pelvic bone registration. The latter improvement was significant, particularly in the right-to-left axis (3.5-fold improvement). In 10 patients, systematic rectal probe repositioning errors (i.e., >20-mL probe volume variations or >8-mm probe shifts in the perpendicular axes) were detected. Target repositioning was reassessed excluding these 10 patients. An additional improvement was observed in the anteroposterior axis with 1.7 times and 1.5 times reduction of the standard deviation with body markers and pelvic bone registrations, respectively.

CONCLUSIONS

Infrared-guided target repositioning for prostate cancer can be optimized with a stereoscopic X-ray positioning device mostly in the right-to-left axis. An optimally positioned inflatable rectal probe further optimizes target repositioning mostly along the anteroposterior axis. Thus a planning target volume with a margin of 2 (right-to-left), 4 (anteroposteriorly), and 6 (craniocaudally) mm around the CTV can be recommended under optimal setup conditions with pelvic bone registration and optimal repositioning of an inflated rectal balloon.

A phantom study on the positioning accuracy of the Novalis Body system

A phantom study was conducted to investigate inherent positioning accuracy of an image-guided patient positioning system-the Novalis Body system for three-dimensional (3-D) conformal radiotherapy. This positioning system consists of two infrared (IR) cameras and one video camera and two kV x-ray imaging devices. The initial patient setup was guided by the IR camera system and the target localization was accomplished using the kV x-ray imaging system. In this study, the IR marker shift and phantom rotation were simulated and their effects on the positioning accuracy were examined by a Rando phantom. The effects of CT slice thickness and treatment sites on the positioning accuracy were tested. In addition, the internal target shift was simulated and its effect on the positioning accuracy was examined by a water tank. With the application of the Novalis Body system, the positioning error of the planned isocenter was significantly reduced. The experimental results with the simulated IR marker shifts indicated that the positioning errors of the planned isocenter were 0.6 +/- 0.3, 0.5 +/- 0.2, and 0.7 +/- 0.2 mm along the lateral, longitudinal, and vertical axes, respectively. The experimental results with the simulated phantom rotations indicated that the positioning errors of the planned isocenter were 0.6 +/- 0.3, 0.7 +/- 0.2, and 0.5 +/- 0.2 mm along the three axes, respectively. The experimental results with the simulated target shifts indicated that the positioning errors of the planned isocenter were 0.6 +/- 0.3, 0.7 +/- 0.2, and 0.5 +/- 0.2 mm along the three axes, respectively. On average, the positioning accuracy of 1 mm for the planned isocenter was achieved using the Novalis Body system.

Application of video imaging for improvement of patient set-up

BACKGROUND AND PURPOSE

For radiotherapy of prostate cancer, the patient is usually positioned in the left-right (LR) direction by aligning a single marker on the skin with the projection of a room laser. The aim of this study is to investigate the feasibility of a room-mounted video camera in combination with previously acquired CT data to improve patient set-up along the LR axis.

MATERIAL AND METHODS

The camera was mounted in the treatment room at the caudal side of the patient. For 22 patients with prostate cancer 127 video and portal images were acquired. The set-up error determined by video imaging was found by matching video images with rendered CT images using various techniques. This set-up error was retrospectively compared with the set-up error derived from portal images. It was investigated whether the number of corrections based on portal imaging would decrease if the information obtained from the video images had been used prior to irradiation. Movement of the skin with respect to bone was quantified using an analysis of variance method.

RESULTS

The measurement of the set-up error was most accurate for a technique where outlines and groins on the left and right side of the patient were delineated and aligned individually to the corresponding features extracted from the rendered CT image. The standard deviations (SD) of the systematic and random components of the set-up errors derived from the portal images in the LR direction were 1.5 and 2.1 mm, respectively. When the set-up of the patients was retrospectively adjusted based on the video images, the SD of the systematic and random errors decreased to 1.1 and 1.3 mm, respectively. From retrospective analysis, a reduction of the number of set-up corrections (from nine to six corrections) is expected when the set-up would have been adjusted using the video images. The SD of the magnitude of motion of the skin of the patient with respect to the bony anatomy was estimated to be 1.1 mm.

CONCLUSION

Video imaging is an accurate technique for measuring the set-up of prostate cancer patients in the LR direction. The outline of the patient is a more accurate estimate of the set-up of the bony anatomy than the marker on the patient's abdomen.

Quality assurance of a system for improved target localization and patient set-up that combines real-time infrared tracking and stereoscopic X-ray imaging

BACKGROUND AND PURPOSE

The aim of this study is to investigate the positional accuracy of a prototype X-ray imaging tool in combination with a real-time infrared tracking device allowing automated patient set-up in three dimensions.

MATERIAL AND METHODS

A prototype X-ray imaging tool has been integrated with a commercially released real-time infrared tracking device. The system, consisting of two X-ray tubes mounted to the ceiling and a centrally located amorphous silicon detector has been developed for automated patient positioning from outside the treatment room prior to treatment. Two major functions are supported: (a) automated fusion of the actual treatment images with digitally reconstructed radiographs (DRRs) representing the desired position; (b) matching of implanted radio opaque markers. Measurements of known translational (up to 30.0mm) and rotational (up to 4.0 degrees ) set-up errors in three dimensions as well as hidden target tests have been performed on anthropomorphic phantoms.

RESULTS

The system's accuracy can be represented with the mean three-dimensional displacement vector, which yielded 0.6mm (with an overall SD of 0.9mm) for the fusion of DRRs and X-ray images. Average deviations between known translational errors and calculations varied from -0.3 to 0.6mm with a standard deviation in the range of 0.6-1.2mm. The marker matching algorithm yielded a three-dimensional uncertainty of 0.3mm (overall SD: 0.4mm), with averages ranging from 0.0 to 0.3mm and a standard deviation in the range between 0.3 and 0.4mm.

CONCLUSIONS

The stereoscopic X-ray imaging device integrated with the real-time infrared tracking device represents a positioning tool allowing for the geometrical accuracy that is required for conformal radiation therapy of abdominal and pelvic lesions, within an acceptable time-frame.