Clinical paradigms and challenges in surface guided radiation therapy: Where do we go from here?

Open AccessPublished:September 25, 2020DOI:https://doi.org/10.1016/j.radonc.2020.09.041

      Highlights

      • SGRT is a 3D positioning tool that can improve treatment efficiency and safety in radiotherapy.
      • Impediments to clinical adoption include training and integration with other IGRT workflows.
      • Contradictory perspectives about SGRT with clinical replications are debated.
      • SGRT offers vast amounts of real-time data as it is a non-invasive imaging modality.
      • The promising future outlook is discussed for SGRT to become fully integrated into radiotherapy.

      Abstract

      Surface guided radiotherapy (SGRT) is becoming a routine tool for patient positioning for specific clinical sites in many clinics. However, it has not yet gained its full potential in terms of widespread adoption. This vision paper first examines some of the difficulties in transitioning to SGRT before exploring the current and future role of SGRT alongside and in concert with other imaging techniques. Finally, future horizons and innovative ideas that may shape and impact the direction of SGRT going forward are reviewed.

      Keywords

      Surface imaging is an emerging modality that generates three-dimensional (3D) surfaces in real-time enabling image guidance for radiotherapy, a process referred to as surface-guided radiotherapy (SGRT). The main advantage of SGRT is its non-ionizing nature, thus enabling its use throughout the entire radiotherapy workflow. While the initial motivation for implementing SGRT was to replace lasers and skin marks for patient positioning [
      • Johnson L.
      • Milliken B.
      • Hadley S.
      • Pelizzari C.
      • Haraf D.
      • Chen G.
      Initial clinical experience with a video-based patient positioning system.
      ], its adoption has been largely limited to a few clinical applications (e.g., breast and stereotactic radiosurgery) even in centers with multiple years of experience [
      • Padilla L.
      • Havnen-Smith A.
      • Cerviño L.
      • Al-Hallaq H.A.
      A survey of surface imaging use in radiation oncology in the United States.
      ]. Given that the capabilities afforded by SGRT are much broader than replacing 3-point positioning, it is somewhat surprising that SGRT has not yet found widespread adoption in radiotherapy.
      In this review, the role of SGRT as a clinical tool will be discussed as well as the issues that may have hindered its more widespread adoption. To gain some insight into this, we will begin by discussing how learners process knowledge and implement new technologies, particularly in a multi-disciplinary setting such as radiation oncology. Next, we will compare and contrast the seeming dualities of SGRT. SGRT was intended to replace the simple 3-point positioning workflow, but perhaps the additional information and sheer amount of data output by a 3D imaging system has been too overwhelming to be easily incorporated into a fast-paced clinical setting. On the other hand, the data collected by SGRT could potentially be harnessed with appropriate tools to aid in making high-level clinical decisions. Other contrasting dualities will be discussed including whether SGRT can simultaneously enable individualization of patient positioning/tracking while also standardizing workflows. Finally, while it is expected that SGRT could enhance patient safety and treatment quality, might it also introduce unexpected new sources of error?
      SGRT has the potential to improve clinical workflows beyond its current capabilities and reconfigure clinical practice. As an example, the vast amounts of data being collected about the position, surface, and respiratory state could be used to make predictions about the patient that could guide radiotherapy outside of simply positioning the patient. An example of this would be the identification of an outlier respiratory pattern for a single patient or even across the patient population. Additionally, SGRT could be used to augment information in the treatment room to further aid radiation therapy technologists (RTTs) in efficiently and accurately delivering a treatment. Acceleration of such innovations into clinical practice would necessitate a close collaboration between researchers and vendors.
      It is our intention to provide a perspective on the current vision for future horizons of SGRT in radiotherapy and offer conceptual ideas and predictions in terms of innovations that could transform SGRT from merely being a patient positioning tool utilized in some clinics for a few applications to becoming an indispensable and integral component of future radiotherapy workflows. It assumes that the reader has a basic understanding of SGRT and associated clinical workflows, which have been previously summarized elsewhere [

      H. Al-Hallaq, A. Gutierrez, and L. Padilla, ‘Surface Image-guided radiotherapy: clinical applications and motion management’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      ,

      A. Gutierrez, H. Al-Hallaq, and D. Stanley, ‘Surface Image-guided radiotherapy: overview and quality assurance’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      ,
      ,

      H. A. Al-Hallaq, A. N. Gutierrez, and L. I. Cerviño, ‘Surface guidance in radiation therapy’, In The modern technology of radiation therapy, vol. 4, 4 vols, J. Van Dyk, Ed. Madison, WI: Medical Physics Publishing.

      ]. While guidelines for clinical implementation and QA are being developed within AAPM (TG-302) and ESTRO, these have yet to be finalized and will not be reviewed in detail here. Instead, we aim to provide an overview of SGRT as a clinical translational tool and demonstrate that it has the potential to morph into a tool that could potentially benefit every radiotherapy patient by guiding multiple aspects of the clinical decision chain besides positioning.

      Evolving multidisciplinary roles

      The clinical implementation of SGRT requires a multidisciplinary team, including clinicians, medical physicists and RTTs. Because the qualified medical physicist (QMP) is responsible for commissioning the system, they become the de facto teacher for other users (i.e., RTTs). The QMP is typically responsible for providing guidance to the clinical team regarding the SGRT system performance, and since the technology is quickly evolving, for also communicating changes that affect the clinical workflow [

      IAEA, ‘Roles and responsibilities, and education and training requirements for clinically qualified medical physicists’, Internat. Atomic Energy Agency, 25, 2013.

      ]. One of the most difficult parts of implementing SGRT is the training and transition from using traditional data (i.e., 3-point setup) to being able to process much more complex but information-rich data. It should also be noted that RTT educational requirements and competencies have been evolving [

      M. Coffey, M. Leech, and ESTRO, the RTT committee, ‘Introduction to the ESTRO European Qualifications Framework (EQF) 7 and 8: Benchmarking Radiation Therapist (RTT) advanced education’, Tech Innov Patient Support Radiat Oncol, vol. 8, pp. 19–21, Dec. 2018, doi: 10.1016/j.tipsro.2018.09.008.

      ], with new topics such as image guidance and adaptive radiotherapy recently added to their curricula. New technologies such as SGRT should be included in the scope of RTT training, including system troubleshooting, identification of specific patient-related issues, and collaborative development of clinical workflows.
      When a new technology is implemented, the clinical team must acquire specific skills for each component of the system, practice integrating them, and know when to apply what they have learned [
      ]. However, it is not until the various skills are combined and practiced in complex situations that the learner fully develops a deeper understanding of the entire system. Thus, to fully integrate a new technology into clinical practice, it is necessary to ensure that the organization allocates time in the schedule to accommodate the educational process. Not only does this mean that enough time to practice is allocated, but also that other distracting events are minimized. Otherwise, the cognitive load can easily be exceeded, and the clinical team members are unable to pay attention and assimilate the new information. Another way to reduce the cognitive load is to allow the team to focus on one clinical application at a time (i.e., SGRT for breast cancer patients) rather than implement it for all patients and clinical sites at once. Finally, the cognitive load may also be exceeded when a learner is presented with too many choices or is required to make multiple novel decisions in sequence within a short timeframe [
      • Miller G.A.
      The magical number seven, plus or minus two: Some limits on our capacity for processing information.
      ]. Underestimating the cognitive load when implementing a new technology, as well as the existing pressures of the clinical environment, can lead to incorrect estimation of the time required for the novice learner to develop the necessary skills. This, in turn, could lead to frustration on part of the learner and a dislike for the technology. Thus, it is important that the QMP adequately estimates the time required for the clinical team to learn and assimilate the new technology. Several institutions have had success implementing SGRT by using it initially for positioning breast cancer patients, where its advantages are readily evident [

      H. Al-Hallaq, A. Gutierrez, and L. Padilla, ‘Surface Image-guided radiotherapy: clinical applications and motion management’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      ,
      ,
      • Hoisak J.D.P.
      • Pawlicki T.
      The role of optical surface imaging systems in radiation therapy.
      ]. After the clinical team develops a consistency and fluency with SGRT for this task, they may apply what has been learned to another patient cohort.
      One example of how SGRT alters clinical workflows is when it is used for patient positioning in place of traditional 3-point positioning using skin marks (see 'Patient positioning (3-point vs. SGRT)'). For this change to occur, day-to-day routine habits must be unlearned and replaced by new knowledge and routines. This is challenging since routine habits require a lower cognitive load and provide a sense of safety [

      R. Rushmer and H. T. O. Davies, ‘Unlearning in health care’, Qual Saf Health Care, vol. 13 Suppl 2, pp. ii10-15, Dec. 2004, doi: 10.1136/qhc.13.suppl_2.ii10.

      ]. In response to the resulting stress and higher cognitive load, the clinical team may apply SGRT workflows in an unsystematic way in an effort to confirm their prior knowledge (i.e., their belief that skin marks are more accurate than SGRT) [
      • Tucker A.L.
      • Edmondson A.C.
      Why hospitals don’t learn from failures: organizational and psychological dynamics that inhibit system change.
      ]. Also, “quick fixes” have been observed, where the clinical team sometimes chooses to ignore the information from the new SGRT system, because it does not agree with their prior beliefs. An example of this would be, neglecting to correct the arm/shoulder position for a breast treatment because this information was not available prior to SGRT and therefore is not thought to be relevant. Ultimately, some clinical users may experience an “aha”-moment, which provides insight that they do in fact need to unlearn their day-to-day routine to improve the treatment for the patient and to adopt SGRT in place of simple 3-point marks. Among the clinical users who have experienced this “aha”-moment, many describe it as a turning point where they realize that they cannot go back because they cannot deny observing a situation that was uncovered by SGRT that would have otherwise led to an incorrect treatment. An example would be when SGRT prevents a “near-miss” by identifying an incorrect isocenter location or incorrect slant board angle (see 'Patient safety vs. a new source of risk'). Ultimately, patience and communication among team members, including vendors (see 'Requirements to move forward'), is important in order to sustain and further refine a new SGRT workflow and to develop new tools that can augment the decision-making process (see 'Decision support').

      Dualities

      For clinics that have adopted SGRT, its utility and versatility to solve various clinical challenges is evident. But the tug between existing clinical workflows and changes to these workflows caused by SGRT is strong. Here, we compare and contrast the potential dualities that are instigated by SGRT.

       Patient positioning (3-point vs. SGRT)

      The current standard of care for marking patients during CT simulation for reference and alignment on the treatment machine is by means of 3-point localization. While tattoos are most commonly used, they are not popular amongst patients for various reasons [
      • Jimenez R.B.
      • et al.
      Tattoo free setup for partial breast irradiation: A feasibility study.
      ]. SGRT is well suited as a replacement for tattoos, but many clinics continue to use tattoos even after implementing SGRT, either as a backup option in case of malfunction of the SGRT system or to provide an additional safety check. Compared to 3-point localization, SGRT provides 3D visualization of the patient’s posture and topographical information [

      G. Carl et al., ‘Optical Surface Scanning for Patient Positioning in Radiation Therapy: A Prospective Analysis of 1902 Fractions’, Technol. Cancer Res. Treat., vol. 17, p. 1533033818806002, 01 2018, doi: 10.1177/1533033818806002.

      ], including rotations, which allows the patients’ position to be corrected prior to verification imaging or treatment [
      • Padilla L.
      • Kang H.
      • Washington M.
      • Hasan Y.
      • Chmura S.J.
      • Al-Hallaq H.
      Assessment of interfractional variation of the breast surface following conventional patient positioning for whole-breast radiotherapy.
      ]. On the other hand, with markers and 2D x-ray imaging, corrections are carried out predominantly along planes perpendicular to the imaging direction without intuitive indicators of any potential rotations and/or deformations in the patient setup. This duality highlights the paradigm shift with SGRT, namely visualizing and guiding setup in a proactive manner prior to treatment start versus acting after imaging has been completed. It should be noted that the adoption of SGRT requires that some immobilization devices be altered, such as converting from a closed to an open mask system or by adjusting vacuum bag heights to expose more of the patient surface.
      Several studies have shown that SGRT for patient positioning provides similar or improved accuracy as 3-point localization [
      • Walter F.
      • Freislederer P.
      • Belka C.
      • Heinz C.
      • Söhn M.
      • Roeder F.
      Evaluation of daily patient positioning for radiotherapy with a commercial 3D surface-imaging system (CatalystTM).
      ,
      • Stanley D.N.
      • McConnell K.A.
      • Kirby N.
      • Gutiérrez A.N.
      • Papanikolaou N.
      • Rasmussen K.
      Comparison of initial patient setup accuracy between surface imaging and three point localization: A retrospective analysis.
      ,
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ]. The localization accuracy varies with anatomical site. In head and neck, small shifts compared to cone-beam CT (CBCT) have been observed, which has been attributed to the use of extended immobilization and minimal effects of respiratory motion [

      G. Carl et al., ‘Optical Surface Scanning for Patient Positioning in Radiation Therapy: A Prospective Analysis of 1902 Fractions’, Technol. Cancer Res. Treat., vol. 17, p. 1533033818806002, 01 2018, doi: 10.1177/1533033818806002.

      ,
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ]. SGRT has proven useful in positioning breast cancer patients compared to 3-point localization. This is because the surface is an adequate surrogate for the target volume in the case of whole breast [
      • Padilla L.
      • Kang H.
      • Washington M.
      • Hasan Y.
      • Chmura S.J.
      • Al-Hallaq H.
      Assessment of interfractional variation of the breast surface following conventional patient positioning for whole-breast radiotherapy.
      ,
      • Shah A.P.
      • Dvorak T.
      • Curry M.S.
      • Buchholz D.J.
      • Meeks S.L.
      Clinical evaluation of interfractional variations for whole breast radiotherapy using 3-dimensional surface imaging.
      ,
      • Crop F.
      • et al.
      Surface imaging, laser positioning or volumetric imaging for breast cancer with nodal involvement treated by helical TomoTherapy.
      ,
      • Hattel S.H.
      • Andersen P.A.
      • Wahlstedt I.H.
      • Damkjaer S.
      • Saini A.
      • Thomsen J.B.
      Evaluation of setup and intrafraction motion for surface guided whole-breast cancer radiotherapy.
      ,
      • Kügele M.
      • et al.
      Surface guided radiotherapy (SGRT) improves breast cancer patient setup accuracy.
      ,
      • Laaksomaa M.
      • et al.
      AlignRT® and CatalystTM in whole-breast radiotherapy with DIBH: Is IGRT still needed?.
      ], and partial breast irradiation [
      • Jimenez R.B.
      • et al.
      Tattoo free setup for partial breast irradiation: A feasibility study.
      ,
      • Chang A.J.
      • et al.
      Video surface image guidance for external beam partial breast irradiation.
      ] and because the extended field-of-view (FOV) of SGRT enables correction of the patient's arm position, which leads to improved treatment posture [
      • Bert C.
      • Metheany K.G.
      • Doppke K.P.
      • Taghian A.G.
      • Powell S.N.
      • Chen G.T.Y.
      Clinical experience with a 3D surface patient setup system for alignment of partial-breast irradiation patients.
      ]. For treatment volumes in the abdomen and pelvis, a reduced accuracy in positioning for both 3-point localization and SGRT is generally observed [

      G. Carl et al., ‘Optical Surface Scanning for Patient Positioning in Radiation Therapy: A Prospective Analysis of 1902 Fractions’, Technol. Cancer Res. Treat., vol. 17, p. 1533033818806002, 01 2018, doi: 10.1177/1533033818806002.

      ,
      • Walter F.
      • Freislederer P.
      • Belka C.
      • Heinz C.
      • Söhn M.
      • Roeder F.
      Evaluation of daily patient positioning for radiotherapy with a commercial 3D surface-imaging system (CatalystTM).
      ,
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ]. Since the surface does not always correlate well with the internal treatment volume in the abdomen and pelvis, it is advisable to combine SGRT with internal verification images (see IGRT vs. SGRT/IGRT). Also, the registration between surfaces can be challenging and result in a larger uncertainty in certain patients (i.e., obese patients) [
      • Heinzerling J.H.
      • et al.
      Use of surface-guided radiation therapy in combination with IGRT for setup and intrafraction motion monitoring during stereotactic body radiation therapy treatments of the lung and abdomen.
      ].
      The use of SGRT for patient positioning not only contributes to positioning accuracy, but after successful implementation also improves positioning efficiency compared with laser- or x-ray-based positioning [
      • Batin E.
      • Depauw N.
      • MacDonald S.
      • Lu H.-M.
      Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
      ]. A few studies have shown SGRT saves time when positioning patients for treatment using traditional LINACs [
      • Jimenez R.B.
      • et al.
      Tattoo free setup for partial breast irradiation: A feasibility study.
      ,

      A. Mannerberg et al., ‘Increased accuracy in reduced time - surface guided RT for hypofractionated prostate cancer patients.’, in ESTRO 2020, pp. 204–205, [Online]. Available: https://cld.bz/8huStZo/214/.

      ], closed-bore LINACs [
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ], or proton therapy [
      • Batin E.
      • Depauw N.
      • MacDonald S.
      • Lu H.-M.
      Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
      ]. The mean setup time was reduced by approximately 1 min per treatment fraction using SGRT compared to 3-point localization for patients undergoing partial breast irradiation [
      • Jimenez R.B.
      • et al.
      Tattoo free setup for partial breast irradiation: A feasibility study.
      ] or prostate stereotactic body radiotherapy (SBRT) [

      A. Mannerberg et al., ‘Increased accuracy in reduced time - surface guided RT for hypofractionated prostate cancer patients.’, in ESTRO 2020, pp. 204–205, [Online]. Available: https://cld.bz/8huStZo/214/.

      ]. With SGRT, larger reductions in setup time, from on average of 11 minutes to 6 minutes, were observed for patients treated with proton post-mastectomy chest wall irradiation [
      • Batin E.
      • Depauw N.
      • MacDonald S.
      • Lu H.-M.
      Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
      ]. Similar time reductions of 3.8 – 4.8 minutes per treatment fraction were observed for patients with brain, head and neck, thoracic and abdominal tumors treated with a closed-bore LINAC [
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ].

       IGRT vs. SGRT/IGRT

      Image-guided radiotherapy (IGRT) encompasses any imaging technique that is used for anatomical or functional localization as well as response assessment in the treatment room [
      ]. By far, the most widely used IGRT technique is on-board CBCT, which has become standard-of-care for patient positioning of deep-seated tumors due to its availability on most treatment units. Non-ionizing imaging modalities such as transabdominal 3D ultrasound imaging and magnetic resonance imaging (MRI) are also used for IGRT when available. These provide additional benefits over CBCT due to their real-time imaging capabilities and superior soft tissue contrast particularly in the case of MRI [
      ]. Despite the undisputed benefits of 3D volumetric imaging, SGRT can play an important role as a complementary IGRT tool. This is because SGRT provides real-time imaging and can be easily added to most treatment units to improve the accuracy, safety, and efficacy for all radiotherapy patients.
      SGRT accuracy requires a reliable correlation between the target and the patient surface. For deep seated organs, especially those that move independently of bony anatomy, such as the prostate, other imaging techniques are generally more accurate for localization [
      • Heinzerling J.H.
      • et al.
      Use of surface-guided radiation therapy in combination with IGRT for setup and intrafraction motion monitoring during stereotactic body radiation therapy treatments of the lung and abdomen.
      ,
      • Krengli M.
      • et al.
      Three-dimensional surface and ultrasound imaging for daily IGRT of prostate cancer.
      ]. Several publications [
      • Laaksomaa M.
      • et al.
      AlignRT® and CatalystTM in whole-breast radiotherapy with DIBH: Is IGRT still needed?.
      ,
      • Gierga D.P.
      • et al.
      Comparison of target registration errors for multiple image-guided techniques in accelerated partial breast irradiation.
      ] have evaluated the synergy between both IGRT and SGRT for patient positioning, demonstrating congruence between both methods. However, SGRT should be recognized as an independent IGRT modality, to be integrated into an imaging strategy and complemented by 3D volumetric imaging when appropriate. This is because SGRT is the only modality that combines the following desirable characteristics: imaging without dose, real-time feedback, sub-millimeter spatial resolution in 3D, and the largest FOV available in radiotherapy. These characteristics enable SGRT to verify immobilization accuracy, correct the patient’s posture, track the respiratory state, and provide intra-fraction monitoring (i.e. surveillance of the patient from initial positioning to the end of treatment). Such applications of SGRT could benefit most radiotherapy patients regardless of whether another complementary IGRT modality will be used for anatomical localization.
      Intra-fraction monitoring with SGRT can be used to augment 3D volumetric imaging workflows. First, SGRT can monitor intra-fraction motion during the time interval required for the physician to evaluate 3D IGRT, which can often be considerable. Also, SGRT can be used to verify the accuracy of any resulting shifts. This is particularly important when large shifts are necessary, a robotic table is used since its mechanical center of rotation differs from the machine isocenter, and/or additional translational shifts are required to correct for rotations [
      • Meyer J.
      • et al.
      Positioning accuracy of cone-beam computed tomography in combination with a HexaPOD robot treatment table.
      ]. Second, SGRT can be used to monitor the correlation between chest and abdominal breathing during lung SBRT to detect baseline drifts [
      • Hughes S.
      • et al.
      Assessment of two novel ventilatory surrogates for use in the delivery of gated/tracked radiotherapy for non-small cell lung cancer.
      ], which may alter the tumor position even if the patient position has not changed.
      Because SGRT is used in tandem with IGRT, its independent role is often not acknowledged, particularly when medical billing regulations limit the use of IGRT codes to a single modality. However, hybrid imaging protocols that combine SGRT and radiographic imaging techniques should be defined and tailored to the specific clinical situation, i.e. tumor location, treatment delivery technique, and correlation between surface and tumor.

       Standardization vs Individualization of Workflows

      Optimization and standardization of workflows are a constant focus in clinics in an effort to efficiently utilize resources. In the era of personalized radiotherapy, these two points of view may appear to collide. On the one hand, SGRT promotes the reduction of inter-operator variability, mitigates users’ subjectivity, and compels a rigid/specific workflow [
      • Shah A.P.
      • Dvorak T.
      • Curry M.S.
      • Buchholz D.J.
      • Meeks S.L.
      Clinical evaluation of interfractional variations for whole breast radiotherapy using 3-dimensional surface imaging.
      ], which might reduce both treatment time [
      • Haraldsson A.
      • Ceberg S.
      • Ceberg C.
      • Bäck S.
      • Engelholm S.
      • Engström P.E.
      Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
      ,
      • Batin E.
      • Depauw N.
      • MacDonald S.
      • Lu H.-M.
      Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
      ,

      A. Mannerberg et al., ‘Increased accuracy in reduced time - surface guided RT for hypofractionated prostate cancer patients.’, in ESTRO 2020, pp. 204–205, [Online]. Available: https://cld.bz/8huStZo/214/.

      ] and repeat imaging [
      • Kügele M.
      • et al.
      Surface guided radiotherapy (SGRT) improves breast cancer patient setup accuracy.
      ]. On the other hand, SGRT quantifies variations in individual patients due to natural differences (i.e., body size and shape) or those due to various physiologic processes (i.e., breast seroma shrinkage, weight loss). Finding the balance between standardization and individualization of workflows on a per-patient basis can be challenging and overwhelming even for expert users.
      SGRT provides the opportunity to reproduce the patient’s external surface over the entire body habitus from the CT simulation at each treatment session in the most reproducible way. This is facilitated by the large FOV of SGRT, which can be as large as 110 × 140 × 240 cm3, enabling the characterization of overall posture and identification of any misalignment of the limbs or the head position that are not captured by 3D volumetric imaging. Despite the use of standardized regions-of-interest [
      • Gopan O.
      • Wu Q.
      Evaluation of the accuracy of a 3D surface imaging system for patient setup in head and neck cancer radiotherapy.
      ] and site-specific thresholds [
      • Stanley D.N.
      • McConnell K.A.
      • Kirby N.
      • Gutiérrez A.N.
      • Papanikolaou N.
      • Rasmussen K.
      Comparison of initial patient setup accuracy between surface imaging and three point localization: A retrospective analysis.
      ], SGRT response will vary for individual patients. Heinzerling et al. [
      • Heinzerling J.H.
      • et al.
      Use of surface-guided radiation therapy in combination with IGRT for setup and intrafraction motion monitoring during stereotactic body radiation therapy treatments of the lung and abdomen.
      ] reported differences when monitoring female and male SBRT patients, as well as those with different body-mass indices. Batin et al. [
      • Batin E.
      • Depauw N.
      • Jimenez R.B.
      • MacDonald S.
      • Lu H.-M.
      Reducing X-ray imaging for proton postmastectomy chest wall patients.
      ] also identified variations in the setup accuracy in patients with mastectomy and with different breast implant sizes. Zhao et al. [
      • Zhao B.
      • Maquilan G.
      • Jiang S.
      • Schwartz D.L.
      Minimal mask immobilization with optical surface guidance for head and neck radiotherapy.
      ] described how SGRT helped to detect weight loss in head and neck patients. Likewise, breast cancer patients can undergo changes in breast shape or volume, as reported by Padilla et al. [
      • Padilla L.
      • Kang H.
      • Washington M.
      • Hasan Y.
      • Chmura S.J.
      • Al-Hallaq H.
      Assessment of interfractional variation of the breast surface following conventional patient positioning for whole-breast radiotherapy.
      ]. Because it is difficult to tease out the patient-specific components from the overall positioning recommendations, it can make interpretation of SGRT results challenging. While prior experience with these clinical scenarios is beneficial for deciding when to forgo the standard workflow in favor of an individualized one, the development of decision support tools (see Decision support) would aid the goal to alert and guide the clinical team when SGRT results highlight a clinical outlier (i.e., veered away from the population-based treatment margins). While such decision support tools currently do not exist for SGRT, they have been successfully developed for similar scenarios in radiotherapy [
      • Luk S.M.H.
      • et al.
      Characterization of a Bayesian network-based radiotherapy plan verification model.
      ,
      • Lam K.L.
      • Ten Haken R.K.
      • Litzenberg D.
      • Balter J.M.
      • Pollock S.M.
      An application of Bayesian statistical methods to adaptive radiotherapy.
      ]. Although SGRT enables standardized workflows that are important for patient safety, it also provides the ability to individualize workflows on a per-patient basis.

       Patient safety vs. a new source of risk

      The complexity of workflows and treatments in radiotherapy demand quality management (QM) strategies, in particular risk management strategies [
      • Malicki J.
      • Przybylska K.
      General guidelines on risk management in external beam radiotherapy.
      ,
      • Cherry P.
      • Duxbury A.M.
      Practical radiotherapy: physics and equipment.
      ]. International guidelines strongly advise that the introduction of new techniques or workflow changes be subject to a risk assessment analysis [

      J. Shafiq, Barton, Michaell, L. Douglas Claire Noble, and L. J. Donaldson, ‘An international review of patient safety measures in radiotherapy practice’, doi: 10.1016/j.radonc.2009.03.007.

      ], where potential failure events and their causes are identified. Similar to other IGRT techniques, a QM strategy for SGRT should include imaging protocols, the definition of technical parameters, action thresholds, documentation, and the decision-making process [
      • Jaffray D.A.
      • et al.
      Safety considerations for IGRT: Executive summary.
      ]. The results of this assessment should promote the mitigation of risk events and allow safe clinical use of SGRT.
      The clinical implementation of SGRT might add new risk sources, as a result of system failures, inadequate workflows, or a lack of process-specific training [
      • Manger R.P.
      • Paxton A.B.
      • Pawlicki T.
      • Kim G.-Y.
      Failure mode and effects analysis and fault tree analysis of surface image guided cranial radiosurgery.
      ,
      • Huq M.S.
      • et al.
      The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management.
      ]. Examples of SGRT failure events include: inconsistency between the patient- or plan-name between the Oncology Information System (OIS) and the SGRT-database, a mismatch between the SGRT or kV/MV/radiation isocenters, an improper reference surface, an inappropriate baseline, an incorrect system configuration (i.e., region-of-interest for tracking, frame-rate, thresholds, etc.), or even faulty reporting (e.g., beam-gating window is not correctly documented). The level of risk may differ when SGRT is used as a decision-support tool (see Decision support) since this will impact the patient’s entire treatment course compared to when it is used for patient positioning and verified by kV imaging. An example of a risk analysis in the context of SGRT for stereotactic radiosurgery was performed by Manger et al. [
      • Manger R.P.
      • Paxton A.B.
      • Pawlicki T.
      • Kim G.-Y.
      Failure mode and effects analysis and fault tree analysis of surface image guided cranial radiosurgery.
      ]. Although risk analysis will always be workflow- and department-specific, this study provides insight into potential issues related to SGRT. Finally, it should be acknowledged that some sources of risk may be more difficult to identify and thus mitigate. One example stems from the resistance to embrace SGRT, which may lead the user to apply recommended workflows unsystematically rather than in a standard fashion (see Evolving multidisciplinary roles).
      Nonetheless, SGRT provides an additional safety measure in the clinical environment [
      ,
      • Al-Hallaq H.
      • Salter B.J.
      Safety and quality improvements with SGRT.
      ]. When integrated with the LINAC and OIS, it has the potential to provide real-time feedback of patient positioning and monitoring during the entire treatment. Furthermore, additional information that improves workflows and provides safety benefits, such as patient identification [
      • Silverstein E.
      • Snyder M.
      Implementation of facial recognition with Microsoft Kinect v2 sensor for patient verification.
      ,

      J. Svoboda M.M. Bronstein M. Drahansky Contactless biometric hand geometry recognition using a low-cost 3D camera 2015 Phuket, Thailand, May 452 457 10.1109/ICB.2015.7139109.

      ], collision avoidance [
      • Padilla L.
      • Pearson E.A.
      • Pelizzari C.A.
      Collision prediction software for radiotherapy treatments.
      ], cross-check of immobilization devices and bolus position (Fig. 1) [

      H. Zhao, ‘Verification of Patient Treatment Accessories and Posture’, presented at the AAPM Annual Meeting, Nashville, TN, 2018, Accessed: Dec. 16, 2018. [Online]. Available: https://aapm.onlinelibrary.wiley.com/doi/abs/10.1002/mp.12938.

      ], will be discussed in Potential new SGRT applications). SGRT promises to be a powerful risk mitigation strategy in the radiotherapy chain, when fully integrated, through the following measures: i) by reducing human errors and subjective decisions when automating the patient positioning/immobilization process, ii) by increasing the detection rate of failure events since it can serve as an “independent observer” in the treatment room [
      • Al-Hallaq H.
      • Salter B.J.
      Safety and quality improvements with SGRT.
      ], and iii) by compelling workflow standardization, which is one of the main recommendations to reduce failure-events [
      • Huq M.S.
      • et al.
      The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management.
      ,
      ].
      Figure thumbnail gr1
      Fig. 1SGRT positioning using video of a phantom in various immobilization devices including a) a breast board, b) a head and neck board, and c) an abdominal compression plate as well as placement of d) bolus material respectively. The pink outline resulted from a conversion of a DICOM-/SGRT-surface into a video-layer.

       Big data vs. too much data

      Despite acquiring images to reconstruct 3D surfaces in real-time, the only quantitative output of most clinical systems is a 6 degree-of-freedom translational/rotational shift. This is intended to simplify the interpretation for the end user (i.e., RTT) in that it mimics the output of x-ray IGRT systems while providing a single recommendation for re-positioning the patient. However, valuable information is discarded by collapsing 3D data into a single set of numbers. Because RTTs intuitively understand that a patient cannot be summarized by a single set of registration parameters and because they routinely encounter discrepancies between x-ray IGRT and SGRT [

      H. Al-Hallaq, A. Gutierrez, and L. Padilla, ‘Surface Image-guided radiotherapy: clinical applications and motion management’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      ,

      H. A. Al-Hallaq, A. N. Gutierrez, and L. I. Cerviño, ‘Surface guidance in radiation therapy’, In The modern technology of radiation therapy, vol. 4, 4 vols, J. Van Dyk, Ed. Madison, WI: Medical Physics Publishing.

      ], they may lose confidence in the output of the system. The impetus to limit information overflow to the end user must be balanced against utilizing the “big data” afforded by SGRT.
      To address this tug between too little and too much data, more sophisticated analysis methods and decision support tools could be incorporated into clinical systems. For example, statistical process control has been used to analyze the results of IGRT systems, which demonstrated that population control charts could identify patients whose metrics were outliers [
      • Shiraishi S.
      • Grams M.P.
      • Fong de Los Santos L.E.
      Image-guided radiotherapy quality control: Statistical process control using image similarity metrics.
      ]. For SGRT, it might be useful to identify outliers both across the patient population within an institution or across institutions (provided that they have agreed to pool their data) or even within the same patient (see Decision support). Additionally, a tool could be developed to incorporate x-ray IGRT information into the process. Because SGRT focuses on the patient’s surface while x-ray imaging focuses on internal anatomy, combining these complementary modes of information could reduce the discrepancy between the two modalities. This could for instance be accomplished by calculating translations and rotations in 6 degrees even from orthogonal x-ray images and comparing to shifts recommended by SGRT in order to suggest a unified set of translations/rotations rather than overwhelming the treatment team with two disparate shift recommendations. In terms of CBCT, an automatic registration to the external surface could be performed to provide a direct comparison to the surface registration calculated by SGRT systems. Known factors, such as e.g. weight gain/loss, that would lead to a discrepancy between internal and external anatomy could thus be evaluated. By discerning patterns that could be used to guide and trouble-shoot the treatment process, SGRT systems could provide an additional value not otherwise available clinically, which in turn may enhance adoption of the technology.

      Harnessing the synergy between research and clinical practice

      While SGRT is becoming relatively well established as a patient positioning and motion/respiratory monitoring tool, researchers and vendors are now exploring new avenues to further enhance the capabilities of these optical systems as described here.

       Potential new SGRT applications

      Markerless tracking for 4D image reconstruction
      One useful existing application, which is not new per se, is the replacement of physical tracking markers with markerless surface tracking to obtain a respiratory signal for 4D image reconstruction. This capability has been demonstrated for CT, PET and CBCT imaging [
      • Kauweloa K.I.
      • et al.
      GateCTTM surface tracking system for respiratory signal reconstruction in 4DCT imaging.
      ,
      • Jönsson M.
      • Ceberg S.
      • Nordström F.
      • Thornberg C.
      • Bäck S.Å.J.
      Technical evaluation of a laser-based optical surface scanning system for prospective and retrospective breathing adapted computed tomography.
      ,
      • Heß M.
      • Büther F.
      • Gigengack F.
      • Dawood M.
      • Schäfers K.P.
      A dual-Kinect approach to determine torso surface motion for respiratory motion correction in PET.
      ] and eliminates the need to place a physical surrogate on the patient, which has implications for patient comfort and infection control, and obviates the need for a third-party system. This method could be further developed to generate respiratory motion models at the time of CT simulation that can be verified at treatment, particularly for SBRT either at quiet respiration or during breath-hold.
      Biometric patient identification
      Acquiring biometric information for patient identification [
      • Wiant D.B.
      • et al.
      A novel method for radiotherapy patient identification using surface imaging.
      ] is another area that is being pursued by vendors and beginning to be clinically implemented. This is useful in the realm of patient safety and optimization of workflows. Either the patient’s body surface or face can be used for identification [
      • Silverstein E.
      • Snyder M.
      Implementation of facial recognition with Microsoft Kinect v2 sensor for patient verification.
      ,
      • Wiant D.B.
      • et al.
      A novel method for radiotherapy patient identification using surface imaging.
      ]. Also, surface fluctuations in color due to de-oxygenated blood can be used to quantify the heart and respiratory rates [
      • Silverstein E.
      • Snyder M.
      Comparative analysis of respiratory motion tracking using Microsoft Kinect v2 sensor.
      ,

      E. Gambi et al., ‘Heart Rate Detection Using Microsoft Kinect: Validation and Comparison to Wearable Devices’, Sensors (Basel), vol. 17, no. 8, Aug. 2017, doi: 10.3390/s17081776.

      ]. Unifying these numerous functions into a single tool is advantageous in terms of both cost and efficiency.
      Immobilization device identification
      Similarly, SGRT can be used to verify the presence and the correct position of immobilization devices. Solutions to detect immobilization devices with embedded RFID (radio-frequency identification) chips [
      • Harry T.
      • Taylor M.
      • Fletcher R.L.
      • Mundt A.J.
      • Pawlicki T.
      Passive tracking of linac clinical flow using radiofrequency identification technology.
      ] already exist in some clinical systems. However, imaging the devices with SGRT would allow direct verification of adjustable devices, such as breast board angle, head and neck board and their relative position to each other (e.g., peg positions of upper and lower devices, or abdominal compression plate compared with the support-arc) as shown in Fig. 1.
      In-room scene mapping
      Because SGRT is non-ionizing and has a FOV that is larger than other IGRT systems, this opens up the possibility to generate additional 3D information by scene mapping [
      • Newcombe R.A.
      • et al.
      KinectFusion: Real-time dense surface mapping and tracking.
      ], essentially acquiring a 3D surface image of the treatment room that can be utilized for various purposes. For instance, the idea of using optical cameras for collision avoidance was mentioned in the early visionary paper by Brahme et al [
      • Brahme A.
      • Nyman P.
      • Skatt B.
      4D laser camera for accurate patient positioning, collision avoidance, image fusion and adaptive approaches during diagnostic and therapeutic procedures.
      ]. In this context, surface information can be used in two different ways. First, SGRT could be used detect potential collisions among the various components of the treatment unit or with the patient in real-time with the option to automatically stop the gantry motion and shut off the beam prior to a collision [

      N. Islam, J. Kilian‐Meneghin, S. deBoer, and M. Podgorsak, ‘A collision prediction framework for noncoplanar radiotherapy planning and delivery’, J Appl Clin Med Phys, vol. n/a, no. n/a, doi: 10.1002/acm2.12920.

      ]. Alternatively, SGRT could be used to preventatively check for potential collisions during the CT-simulation and/or treatment planning process provided that a scan of the patient and a 3D model of the treatment unit are available [
      • Padilla L.
      • Pearson E.A.
      • Pelizzari C.A.
      Collision prediction software for radiotherapy treatments.
      ].
      Another useful application that takes advantage of the large FOV of SGRT systems is its utility for providing missing data when CT scanning large patients. CT scanners have a limited FOV and parts of the external anatomy might be truncated or incorrectly reconstructed. SGRT could be conveniently used to complement missing information (e.g. clipped elbows or truncated anatomy) [
      • Maltz J.S.
      • Bose S.
      • Shukla H.P.
      • Bani-Hashemi A.R.
      CT Truncation artifact removal using water-equivalent thicknesses derived from truncated projection data.
      ] if an interface to the treatment planning system (TPS) exists for importation of the surface scan. Confidently delineating the external contour could make the difference between a treatment plan limited to a simple 2 or 4-field arrangement versus a more complex intensity-modulated radiation treatment (IMRT), although uncertainties in tissue and immobilization heterogeneities would need to be taken into account.
      In the same vein, but for total body irradiation (TBI), SGRT provides a unique method to scan the entire patient without CT as shown in Fig. 2. Typically, a treatment plan is generated based on manually measured patient dimensions at a few locations. Continuous contour measurements acquired with optical cameras and converted into a DICOM RT structure would allow for more accurate and reproducible dose calculations when imported into a TPS. In addition, the overall position can be monitored during the course of treatment, which can last for an hour or longer, particularly when higher doses are prescribed or if lung block placement requires multiple iterations. As TBI treatments are usually implemented at extended distance, this would require additional optical cameras suitably placed to avoid occlusion during treatment by the scatter shield. Furthermore, reproducible placement of lung blocks and in-vivo diodes could be guided by superimposing their placement into the scene by means of augmented reality [

      J. Talbot, J. Meyer, R. Watts, and R. Grasset, ‘An Augmented Reality Application for Patient Positioning and Monitoring in Radiotherapy’, in World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany, Berlin, Heidelberg, 2009, pp. 620–623, doi: 10.1007/978-3-642-03474-9_174.

      ], which will be addressed next. Finally, there is a potential role for SGRT even for isocentric TBI treatments, such as intensity-modulated total marrow irradiation (TMI), which typically requires multiple isocenters to implement on a LINAC [
      • Aydogan B.
      • Mundt A.J.
      • Roeske J.C.
      Linac-based intensity modulated total marrow irradiation (IM-TMI).
      ]. Because there is currently no imaging method to aid in positioning the entire patient, the isocenters are verified piecemeal in an iterative and often time-consuming fashion.
      Figure thumbnail gr2
      Fig. 2Proof-of-concept in a volunteer of the use of SGRT to position and monitor TBI patients.
      Augmented reality
      The idea of using augmented reality (AR) for patient positioning was proposed and demonstrated as far back as 2009 [

      J. Talbot, J. Meyer, R. Watts, and R. Grasset, ‘An Augmented Reality Application for Patient Positioning and Monitoring in Radiotherapy’, in World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany, Berlin, Heidelberg, 2009, pp. 620–623, doi: 10.1007/978-3-642-03474-9_174.

      ,

      JWT Talbot, ‘A patient position guidance system in radiotherapy using augmented reality’, 2009.

      ,
      • Talbot J.
      • Meyer J.
      • Watts R.
      • Grasset R.
      A method for patient set-up guidance in radiotherapy using augmented reality.
      ]. AR is a computer visualization concept that embeds virtual objects into a real-life scenario. AR supplements the real world by superimposing virtual objects that appear to coexist with the real environment. These virtual objects are usually added with the intention of providing additional perceptual information that would not be apparent without augmentation. This can have implications for workflow efficiency as well as treatment accuracy and patient safety. Specifically, virtually superimposing the intended immobilization devices including their indexed positions could assist with setting up the treatment room and prevent errors that are human caused (i.e., selection of an incorrect breast board angle) (see Fig. 1). As a starting point, some clinical SGRT systems are already capable of projecting colors onto the patient surface to guide the RTTs in terms of how to reposition the patient as shown in Fig. 3.
      Figure thumbnail gr3
      Fig. 3Illustration of interactive visual guidance via color map projected onto a torso phantom positioned with a pitch rotation of 3°. The red and yellow color wash (shown in arrows of the same color) indicate deviations from a pre-determined tolerance.
      It has also been proposed to augment the scene with relevant beam or structure information that could be extracted from the DICOM objects [
      • Mueller A.
      Visualisation of radiation treatment parameters in an Augmented Reality environment.
      ]. For instance, the intersection between the treatment beam and the reference patient surface could be calculated and virtually projected onto the patient. Such a tool could serve to verify the radiation light field of the multi-leaf collimator (MLC) or the shape and position of an electron block to ensure that the correct treatment field will enter at the intended location [
      • Mueller A.
      Visualisation of radiation treatment parameters in an Augmented Reality environment.
      ] as illustrated in Fig. 4. Similarly, internal information such as the contours of the target volume or critical structures could be projected onto the surface to provide additional visual information during setup. Both scenarios are shown for a prostate case in Fig. 5. Other examples of the potential use of AR to enhance workflows and safety include the augmentation of bolus placement, in-vivo dose points for diode placement, positioning of physical lung blocks, and verification of field junctions (e.g., cranio-spinal irradiation, multiple isocenters for breast and supraclavicular fields, gaps between adjacent fields).
      Figure thumbnail gr4
      Fig. 4Illustration of the beam information projected onto the patient surface as well as a projection of an internal object (from
      [
      • Mueller A.
      Visualisation of radiation treatment parameters in an Augmented Reality environment.
      ]
      with permission).
      Figure thumbnail gr5
      Fig. 5Example of objects projected from both inside and outside of a male pelvis onto the surface to augment the scene with additional information. (a) Planned MLC position projected onto the external surface and (b) prostate target projected onto the external surface (modified from
      [
      • Mueller A.
      Visualisation of radiation treatment parameters in an Augmented Reality environment.
      ]
      with permission). In a clinical setting the information would be projected directly onto the patient or augmented into the scene as a virtual object on a live image (AR) to provide additional information for patient setup.
      Adaptive radiotherapy
      As dedicated commercial treatment units for adaptive radiotherapy (ART) are being developed, such as MR-LINACs or the Varian Ethos system, there is an opportunity for SGRT to address a weakness in the ART workflows: adaptive planning takes time [
      • Green O.L.
      • Henke L.E.
      • Hugo G.D.
      Practical clinical workflows for online and offline adaptive radiation therapy.
      ]. The premise of ART is to generate a plan of the day if adaptation is deemed necessary. Generating a new plan on-the-fly requires re-contouring, re-optimization, reevaluation and quality checks during which the patient remains in the treatment position for a considerable amount of time. SGRT is ideally suited to track the patient’s position throughout this process without additional imaging dose. Some of these adaptive treatment units have a limited degree of freedom in terms of couch movement. As a result, corrections cannot be applied in all directions. Initial positioning of patients accurately using SGRT would reduce the frequency of re-positioning and limit the need for complex ART to situations in which there are true changes in internal anatomy.
      QA Tool
      Finally, SGRT may prove useful as a QA tool. Due to the submillimeter accuracy of optical systems they lend themselves easily to a wide range of QA tasks. Examples include the validation of the isocentricity of a robotic couch [

      O. El‐Sherif, N. B. Remmes, and J. J. Kruse, ‘Validating robotic couch isocentricity with 3D surface imaging’, J Appl Clin Med Phys, doi: 10.1002/acm2.12939.

      ], and as a QA tool for proton range compensators [
      • Kim M.
      • et al.
      Development of a 3D optical scanning-based automatic quality assurance system for proton range compensators.
      ].

       Decision support

      In radiotherapy, decision support tools could be implemented to ensure that the planned treatment is delivered despite the variations that might occur over multiple days of treatment (i.e., plan robustness). The crudest way to achieve robustness is through the use of institution-specific population-based planning margins. SGRT could potentially be used to calculate population-based margins for anatomical sites in which the surface is an adequate surrogate for the target (i.e., breast or brain) and to continually update these margins with the most recent clinical data (e.g., past 6–12 months). Additionally, SGRT could be used to calculate margins necessary to account for respiratory-related motion (i.e., internal target volume), by monitoring the correlation between the surface and tumor over the entire treatment course. To ensure robustness, outliers must be identified prior to treatment and the plan must be adapted accordingly. If a single tool such as SGRT could be used to model expected (i.e., respiratory) and unexpected (i.e., tumor shrinkage or weight loss) changes throughout the course of treatment and if the dosimetric impact of such changes was well characterized, SGRT could serve as the trigger in an adaptive radiotherapy feedback loop.
      First, daily SGRT could be used to identify when plan adaptation is necessary following anatomic changes [

      H. Al-Hallaq, A. Gutierrez, and L. Padilla, ‘Surface Image-guided radiotherapy: clinical applications and motion management’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      ]. Since this would be more likely to succeed for anatomic changes that would affect the surface, it may be more promising for tumors that are close to the surface (e.g., breast cancer [
      • Padilla L.
      • Kang H.
      • Washington M.
      • Hasan Y.
      • Chmura S.J.
      • Al-Hallaq H.
      Assessment of interfractional variation of the breast surface following conventional patient positioning for whole-breast radiotherapy.
      ,
      • Shah A.P.
      • Dvorak T.
      • Curry M.S.
      • Buchholz D.J.
      • Meeks S.L.
      Clinical evaluation of interfractional variations for whole breast radiotherapy using 3-dimensional surface imaging.
      ], lymphoma) or for physiologic changes that affect the surface (e.g., weight loss in head and neck [
      • Gopan O.
      • Wu Q.
      Evaluation of the accuracy of a 3D surface imaging system for patient setup in head and neck cancer radiotherapy.
      ]). Providing a quantitative action level to guide clinicians regarding potential plan adaptions is expected to improve the quality of treatments.
      SGRT data could also be mined to identify outliers in the patient’s respiration pattern and to guide the 4DCT process, as some vendors have begun to do. Because it is non-ionizing, SGRT could be used to observe a patient over a much longer timeframe to accurately characterize the respiratory pattern. For breath-hold treatments, current SGRT systems quantify the amplitude and 3D surface position. However, the pattern of the breath-hold cycle may be a more sensitive measure of whether the patient is reproducing their position at breath-hold during treatment. Such a patient-specific “breath-hold signature” could be used to augment patient coaching during treatment. While the simplest way to accomplish this would be to display the breath-hold pattern to the patient during treatment, vector maps could be analyzed and grouped into several categories that the therapists could use to provide additional coaching (e.g., increase breath-hold amplitude, expand chest, reduce belly movement). In this context, Bayesian networks may lend themselves well as a decision support tool that has been applied to similar problems in radiotherapy [
      • Luk S.M.H.
      • et al.
      Characterization of a Bayesian network-based radiotherapy plan verification model.
      ].

       Requirements to move forward

      Data repositories and vendors’ involvement
      Beyond their responsibility for clinical implementation, physicists are continuously critically analyzing the tools and data at their disposal to learn how to advance clinical care in the future. Although SGRT systems generate real-time 3D surfaces, these are typically not archived nor are they exportable in DICOM format. While this is primarily due to the data storage requirements, users should be offered the option of archiving the surface data as it could be retrospectively mined. Such analysis has identified concrete steps of the clinical workflow that require improvement such as fabrication of immobilization devices [
      • Batin E.
      • Depauw N.
      • MacDonald S.
      • Lu H.-M.
      Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
      ]. More sophisticated analysis techniques (see Decision support) could also be developed using SGRT data, particularly if vendors would provide built-in tools for data anonymization and sharing across institutions to facilitate creation of multi-institutional data repositories.
      On the other hand, vendors’ roles should go beyond the development of hardware and software by actively participating in the application of new technologies in a safe and accurate manner [
      • Pawlicki T.
      • Dunscombe P.
      • Mundt A.J.
      • Scalliet P.
      Quality and safety in radiotherapy.
      ]. The vendors should provide users with adequate multidisciplinary training and support during the commissioning and clinical implementation phases. Due to the complexity when transitioning from a simple 3-point setup to a more complex and data-intensive positioning workflow provided by SGRT systems, more emphasis should be placed on providing tools that help the user to understand and interpret the multidimensional information (i.e., decision support). As an example, the behavior of deformable surface matching algorithms can be non-intuitive (i.e., “a black box”) [
      • Meyer J.
      • et al.
      Characterizing a deformable registration algorithm for surface-guided breast radiotherapy.
      ], which might hinder fast adoption despite the fact that they have been shown to be accurate. Establishing a network of users and facilitating communication among them when rolling out a new technology should become standard practice by vendors. Vendors should also bear the responsibility for communicating system- and workflow-pitfalls upfront to avoid any breach in trust between the users and vendors. Additionally, the vendors should make an effort to establish collaboration protocols with companies whose products interface with SGRT systems, in order to maximize the potential of the technology and to facilitate developmental projects (see Potential new SGRT applications) and clinical integration while optimizing workflows and reducing the risk profile.
      Hence, a strong collaboration between vendors, researchers, and clinical users is a key-factor to enable development, testing and clinical implementation of these sophisticated tools. This requires that SGRT is recognized as an important clinical tool that is paradigm changing.
      Costs and clinical trials
      Another barrier to more widespread adoption of SGRT is its cost. It has been demonstrated that SGRT has many benefits in terms of workflow efficiencies, patient safety and improved positioning accuracy in both initial setup and intra-fraction monitoring. However, the value proposition can be challenging to make to hospital administrators and healthcare insurance payers alike. The central notion of health technology assessment is the cost per quality-adjusted life-years (QALY), which requires outcome data from clinical trials, of which very few are underway [

      University of Zurich, ‘Randomized Controlled Trial Comparing Closed vs. Open Face Masks for Cranial Radiotherapy’, clinicaltrials.gov, Clinical trial registration NCT04079595, Sep. 2019. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT04079595.

      ,

      D. I. Karam, ‘Assessment of Left-sided Cardiac Sparing Through the Use of 3-dimensional Surface Matching-based Deep Inspiration Breath Hold and Active Breathing Control’, clinicaltrials.gov, Clinical trial registration NCT03459898, Jun. 2020. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT03459898.

      ,

      The University of Texas Health Science Center at San Antonio, ‘Demonstration of a Novel Approach Using Surface-Image Guidance to Improve Delivery of Breast Radiotherapy’, clinicaltrials.gov, Clinical trial registration NCT03799523, Jul. 2020. Accessed: Sep. 02, 2020. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT03799523.

      ,

      ClinicalTrials.gov, ‘Radiotherapy Delivery in Deep Inspiration for Pediatric Patients’. https://clinicaltrials.gov/ct2/show/NCT03315546 (accessed Sep. 03, 2020).

      ].

      Summary

      Every so often, technical innovations cause a paradigm shift in radiotherapy (e.g., IMRT, IGRT). Occasionally the shift can be anticipated, but other times it cannot. SGRT was developed initially as a tool to replace lasers and skin marks for patient positioning. However, the volume of information rich data provided by SGRT has been harnessed by researchers and vendors to develop it into a powerful tool that could affect patient safety, treatment quality, motion management, risk surveillance and management, as well as to potentially provide decision support in radiotherapy. Those who routinely use SGRT cannot deny the fact that it is much more than a positioning tool. However, this has created an implementation challenge for SGRT because users can become overwhelmed by the sheer amount of information. As these tools are refined to provide the proper decision support and offer unique capabilities not otherwise available with other IGRT modalities, such as augmented reality, SGRT is poised to become an integral component of all radiotherapy treatments causing an unanticipated but welcome paradigm shift.

      Funding sources

      There are no funding sources to report.

      Disclosures

      HA reports travel funds from AAPM and ESTRO to co-chair the 3rd Physics Workshop on SGRT. VB represents her Hospital in a cooperation agreement and in a reference site agreement with Vision RT. MK reports research founding to her Hospital from C-rad Positioning AB. JM reports no disclosures/conflicts related to this manuscript.

      Declaration of Competing Interest

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgements

      We would like to thank the ESTRO Physics Committee for sponsoring a discussion on SGRT at the 3rd Physics Workshop in October, 2019 in Budapest, Hungary. We would like to thank the participants of this Workshop for useful discussions regarding the future horizons of SGRT. We would also like to thank AAPM (The American Association of Physicists in Medicine), United States of America for co-sponsoring the SGRT workshop with ESTRO (European Society for Radiotherapy and Oncology), Belgium.

      References

        • Johnson L.
        • Milliken B.
        • Hadley S.
        • Pelizzari C.
        • Haraf D.
        • Chen G.
        Initial clinical experience with a video-based patient positioning system.
        Int J Radiat Oncol Biol Phys. 1999; 45: 205-213
        • Padilla L.
        • Havnen-Smith A.
        • Cerviño L.
        • Al-Hallaq H.A.
        A survey of surface imaging use in radiation oncology in the United States.
        J Appl Clin Med Phys. 2019; 20: 70-77https://doi.org/10.1002/acm2.12762
      1. H. Al-Hallaq, A. Gutierrez, and L. Padilla, ‘Surface Image-guided radiotherapy: clinical applications and motion management’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      2. A. Gutierrez, H. Al-Hallaq, and D. Stanley, ‘Surface Image-guided radiotherapy: overview and quality assurance’, in Image guidance in radiation therapy: techniques, accuracy, and limitations: American Association of Physicists in Medicine 2018 Summer School Proceedings, Vanderbilt University, Nashville, Tennessee, July 26-28, 2018, P. Alaei and G. Ding, Eds. Madison, WI: Medical Physics Publishing, 2018.

      3. Hoisak J.D.P. Paxton A.B. Waghorn B. Pawlicki T.A. Surface guided radiation therapy. 1st ed. Taylor and Francis, Boca Raton, FL2020
      4. H. A. Al-Hallaq, A. N. Gutierrez, and L. I. Cerviño, ‘Surface guidance in radiation therapy’, In The modern technology of radiation therapy, vol. 4, 4 vols, J. Van Dyk, Ed. Madison, WI: Medical Physics Publishing.

      5. IAEA, ‘Roles and responsibilities, and education and training requirements for clinically qualified medical physicists’, Internat. Atomic Energy Agency, 25, 2013.

      6. M. Coffey, M. Leech, and ESTRO, the RTT committee, ‘Introduction to the ESTRO European Qualifications Framework (EQF) 7 and 8: Benchmarking Radiation Therapist (RTT) advanced education’, Tech Innov Patient Support Radiat Oncol, vol. 8, pp. 19–21, Dec. 2018, doi: 10.1016/j.tipsro.2018.09.008.

        • Miller G.A.
        The magical number seven, plus or minus two: Some limits on our capacity for processing information.
        Psychol Rev. 1994; 101: 343-352https://doi.org/10.1037/0033-295X.101.2.343
        • Hoisak J.D.P.
        • Pawlicki T.
        The role of optical surface imaging systems in radiation therapy.
        Semin Radiat Oncol. 2018; 28: 185-193https://doi.org/10.1016/j.semradonc.2018.02.003
      7. R. Rushmer and H. T. O. Davies, ‘Unlearning in health care’, Qual Saf Health Care, vol. 13 Suppl 2, pp. ii10-15, Dec. 2004, doi: 10.1136/qhc.13.suppl_2.ii10.

        • Tucker A.L.
        • Edmondson A.C.
        Why hospitals don’t learn from failures: organizational and psychological dynamics that inhibit system change.
        California Manage Rev. 2003; 45: 55-72https://doi.org/10.2307/41166165
        • Jimenez R.B.
        • et al.
        Tattoo free setup for partial breast irradiation: A feasibility study.
        J Appl Clin Med Phys. 2019; 20: 45-50https://doi.org/10.1002/acm2.12557
      8. G. Carl et al., ‘Optical Surface Scanning for Patient Positioning in Radiation Therapy: A Prospective Analysis of 1902 Fractions’, Technol. Cancer Res. Treat., vol. 17, p. 1533033818806002, 01 2018, doi: 10.1177/1533033818806002.

        • Padilla L.
        • Kang H.
        • Washington M.
        • Hasan Y.
        • Chmura S.J.
        • Al-Hallaq H.
        Assessment of interfractional variation of the breast surface following conventional patient positioning for whole-breast radiotherapy.
        J Appl Clin Med Phys. 2014; 15: 177-189https://doi.org/10.1120/jacmp.v15i5.4921
        • Walter F.
        • Freislederer P.
        • Belka C.
        • Heinz C.
        • Söhn M.
        • Roeder F.
        Evaluation of daily patient positioning for radiotherapy with a commercial 3D surface-imaging system (CatalystTM).
        Radiat Oncol. 2016; 11https://doi.org/10.1186/s13014-016-0728-1
        • Stanley D.N.
        • McConnell K.A.
        • Kirby N.
        • Gutiérrez A.N.
        • Papanikolaou N.
        • Rasmussen K.
        Comparison of initial patient setup accuracy between surface imaging and three point localization: A retrospective analysis.
        J Appl Clin Med Phys. 2017; 18: 58-61https://doi.org/10.1002/acm2.12183
        • Haraldsson A.
        • Ceberg S.
        • Ceberg C.
        • Bäck S.
        • Engelholm S.
        • Engström P.E.
        Surface-guided tomotherapy improves positioning and reduces treatment time: A retrospective analysis of 16 835 treatment fractions.
        J Appl Clin Med Phys. 2020; https://doi.org/10.1002/acm2.12936
        • Shah A.P.
        • Dvorak T.
        • Curry M.S.
        • Buchholz D.J.
        • Meeks S.L.
        Clinical evaluation of interfractional variations for whole breast radiotherapy using 3-dimensional surface imaging.
        Pract Radiat Oncol. 2013; 3: 16-25https://doi.org/10.1016/j.prro.2012.03.002
        • Crop F.
        • et al.
        Surface imaging, laser positioning or volumetric imaging for breast cancer with nodal involvement treated by helical TomoTherapy.
        J Appl Clin Med Phys. 2016; 17: 200-211https://doi.org/10.1120/jacmp.v17i5.6041
        • Hattel S.H.
        • Andersen P.A.
        • Wahlstedt I.H.
        • Damkjaer S.
        • Saini A.
        • Thomsen J.B.
        Evaluation of setup and intrafraction motion for surface guided whole-breast cancer radiotherapy.
        J Appl Clin Med Phys. 2019; 20: 39-44https://doi.org/10.1002/acm2.12599
        • Kügele M.
        • et al.
        Surface guided radiotherapy (SGRT) improves breast cancer patient setup accuracy.
        J Appl Clin Med Phys. 2019; 20: 61-68
        • Laaksomaa M.
        • et al.
        AlignRT® and CatalystTM in whole-breast radiotherapy with DIBH: Is IGRT still needed?.
        J Appl Clin Med Phys. 2019; 20: 97-104https://doi.org/10.1002/acm2.12553
        • Chang A.J.
        • et al.
        Video surface image guidance for external beam partial breast irradiation.
        Pract Radiat Oncol. 2012; 2: 97-105https://doi.org/10.1016/j.prro.2011.06.013
        • Bert C.
        • Metheany K.G.
        • Doppke K.P.
        • Taghian A.G.
        • Powell S.N.
        • Chen G.T.Y.
        Clinical experience with a 3D surface patient setup system for alignment of partial-breast irradiation patients.
        Int J Radiat Oncol Biol Phys. 2006; 64: 1265-1274https://doi.org/10.1016/j.ijrobp.2005.11.008
        • Heinzerling J.H.
        • et al.
        Use of surface-guided radiation therapy in combination with IGRT for setup and intrafraction motion monitoring during stereotactic body radiation therapy treatments of the lung and abdomen.
        J Appl Clin Med Phys. 2020; https://doi.org/10.1002/acm2.12852
        • Batin E.
        • Depauw N.
        • MacDonald S.
        • Lu H.-M.
        Can surface imaging improve the patient setup for proton postmastectomy chest wall irradiation?.
        Pract Radiat Oncol. 2016; 6: e235-e241https://doi.org/10.1016/j.prro.2016.02.001
      9. A. Mannerberg et al., ‘Increased accuracy in reduced time - surface guided RT for hypofractionated prostate cancer patients.’, in ESTRO 2020, pp. 204–205, [Online]. Available: https://cld.bz/8huStZo/214/.

      10. Alaei P. Ding G.X. Image guidance in radiation therapy: techniques, accuracy, and limitations. first ed. Medical Physics Pub, Madison, WI2018
        • Krengli M.
        • et al.
        Three-dimensional surface and ultrasound imaging for daily IGRT of prostate cancer.
        Radiat Oncol. 2016; 11https://doi.org/10.1186/s13014-016-0734-3
        • Gierga D.P.
        • et al.
        Comparison of target registration errors for multiple image-guided techniques in accelerated partial breast irradiation.
        Int J Radiat Oncol Biol Phys. 2008; 70: 1239-1246https://doi.org/10.1016/j.ijrobp.2007.11.020
        • Meyer J.
        • et al.
        Positioning accuracy of cone-beam computed tomography in combination with a HexaPOD robot treatment table.
        Int J Radiat Oncol Biol Phys. 2007; 67: 1220-1228https://doi.org/10.1016/j.ijrobp.2006.11.010
        • Hughes S.
        • et al.
        Assessment of two novel ventilatory surrogates for use in the delivery of gated/tracked radiotherapy for non-small cell lung cancer.
        Radiother Oncol. 2009; 91: 336-341https://doi.org/10.1016/j.radonc.2009.03.016
        • Gopan O.
        • Wu Q.
        Evaluation of the accuracy of a 3D surface imaging system for patient setup in head and neck cancer radiotherapy.
        Int J Radiat Oncol Biol Phys. 2012; 84: 547-552https://doi.org/10.1016/j.ijrobp.2011.12.004
        • Batin E.
        • Depauw N.
        • Jimenez R.B.
        • MacDonald S.
        • Lu H.-M.
        Reducing X-ray imaging for proton postmastectomy chest wall patients.
        Practical Radiation Oncology. 2018; 8: e266-e274https://doi.org/10.1016/j.prro.2018.03.002
        • Zhao B.
        • Maquilan G.
        • Jiang S.
        • Schwartz D.L.
        Minimal mask immobilization with optical surface guidance for head and neck radiotherapy.
        J Appl Clin Med Phys. 2018; 19: 17-24https://doi.org/10.1002/acm2.12211
        • Luk S.M.H.
        • et al.
        Characterization of a Bayesian network-based radiotherapy plan verification model.
        Med Phys. 2019; 46: 2006-2014https://doi.org/10.1002/mp.13515
        • Lam K.L.
        • Ten Haken R.K.
        • Litzenberg D.
        • Balter J.M.
        • Pollock S.M.
        An application of Bayesian statistical methods to adaptive radiotherapy.
        Phys Med Biol. 2005; 50: 3849
        • Malicki J.
        • Przybylska K.
        General guidelines on risk management in external beam radiotherapy.
        Radiat Protect. 2015;
        • Cherry P.
        • Duxbury A.M.
        Practical radiotherapy: physics and equipment.
        John Wiley & Sons, 2019
      11. J. Shafiq, Barton, Michaell, L. Douglas Claire Noble, and L. J. Donaldson, ‘An international review of patient safety measures in radiotherapy practice’, doi: 10.1016/j.radonc.2009.03.007.

        • Jaffray D.A.
        • et al.
        Safety considerations for IGRT: Executive summary.
        Pract Radiat Oncol. 2013; 3: 167-170https://doi.org/10.1016/j.prro.2013.01.004
        • Manger R.P.
        • Paxton A.B.
        • Pawlicki T.
        • Kim G.-Y.
        Failure mode and effects analysis and fault tree analysis of surface image guided cranial radiosurgery.
        Med Phys. 2015; 42: 2449-2461https://doi.org/10.1118/1.4918319
        • Huq M.S.
        • et al.
        The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management.
        Med Phys. 2016; 43: 4209-4262https://doi.org/10.1118/1.4947547
        • Silverstein E.
        • Snyder M.
        Implementation of facial recognition with Microsoft Kinect v2 sensor for patient verification.
        Med Phys. 2017; 44: 2391-2399https://doi.org/10.1002/mp.12241
      12. J. Svoboda M.M. Bronstein M. Drahansky Contactless biometric hand geometry recognition using a low-cost 3D camera 2015 Phuket, Thailand, May 452 457 10.1109/ICB.2015.7139109.

        • Padilla L.
        • Pearson E.A.
        • Pelizzari C.A.
        Collision prediction software for radiotherapy treatments.
        Med Phys. 2015; 42: 6448-6456https://doi.org/10.1118/1.4932628
      13. H. Zhao, ‘Verification of Patient Treatment Accessories and Posture’, presented at the AAPM Annual Meeting, Nashville, TN, 2018, Accessed: Dec. 16, 2018. [Online]. Available: https://aapm.onlinelibrary.wiley.com/doi/abs/10.1002/mp.12938.

        • Al-Hallaq H.
        • Salter B.J.
        Safety and quality improvements with SGRT.
        in: Hoisak J.D.P. Paxton A.B. Waghorn B. Pawlicki T.A. Surface guided radiation therapy. 1st ed. Taylor and Francis, Boca Raton, FL2020: 25-50
      14. Zietman A. Palta J. Steinberg M. Safety is no accident: a framework for quality radiation oncology and care. Arlington, VA, ASTRO2012
        • Shiraishi S.
        • Grams M.P.
        • Fong de Los Santos L.E.
        Image-guided radiotherapy quality control: Statistical process control using image similarity metrics.
        Med Phys. 2018; 45: 1811-1821https://doi.org/10.1002/mp.12859
        • Kauweloa K.I.
        • et al.
        GateCTTM surface tracking system for respiratory signal reconstruction in 4DCT imaging.
        Med Phys. 2012; 39: 492-502https://doi.org/10.1118/1.3671941
        • Jönsson M.
        • Ceberg S.
        • Nordström F.
        • Thornberg C.
        • Bäck S.Å.J.
        Technical evaluation of a laser-based optical surface scanning system for prospective and retrospective breathing adapted computed tomography.
        Acta Oncol. 2015; 54: 261-265https://doi.org/10.3109/0284186X.2014.948059
        • Heß M.
        • Büther F.
        • Gigengack F.
        • Dawood M.
        • Schäfers K.P.
        A dual-Kinect approach to determine torso surface motion for respiratory motion correction in PET.
        Med Phys. 2015; 42: 2276-2286https://doi.org/10.1118/1.4917163
        • Wiant D.B.
        • et al.
        A novel method for radiotherapy patient identification using surface imaging.
        J Appl Clin Med Phys. 2016; 17: 271-278https://doi.org/10.1120/jacmp.v17i2.6066
        • Silverstein E.
        • Snyder M.
        Comparative analysis of respiratory motion tracking using Microsoft Kinect v2 sensor.
        J Appl Clin Med Phys. 2018; 19: 193-204https://doi.org/10.1002/acm2.12318
      15. E. Gambi et al., ‘Heart Rate Detection Using Microsoft Kinect: Validation and Comparison to Wearable Devices’, Sensors (Basel), vol. 17, no. 8, Aug. 2017, doi: 10.3390/s17081776.

        • Harry T.
        • Taylor M.
        • Fletcher R.L.
        • Mundt A.J.
        • Pawlicki T.
        Passive tracking of linac clinical flow using radiofrequency identification technology.
        Practical Radiat Oncol. 2014; 4: e85-e90https://doi.org/10.1016/j.prro.2013.03.005
        • Newcombe R.A.
        • et al.
        KinectFusion: Real-time dense surface mapping and tracking.
        in: 2011 10th IEEE international symposium on mixed and augmented reality. 2011: 127-136 (doi: 10.1109/ISMAR.2011.6092378)
        • Brahme A.
        • Nyman P.
        • Skatt B.
        4D laser camera for accurate patient positioning, collision avoidance, image fusion and adaptive approaches during diagnostic and therapeutic procedures.
        Med Phys. 2008; 35: 1670-1681https://doi.org/10.1118/1.2889720
      16. N. Islam, J. Kilian‐Meneghin, S. deBoer, and M. Podgorsak, ‘A collision prediction framework for noncoplanar radiotherapy planning and delivery’, J Appl Clin Med Phys, vol. n/a, no. n/a, doi: 10.1002/acm2.12920.

        • Maltz J.S.
        • Bose S.
        • Shukla H.P.
        • Bani-Hashemi A.R.
        CT Truncation artifact removal using water-equivalent thicknesses derived from truncated projection data.
        in: 2007 29th annual international conference of the IEEE engineering in medicine and biology society. 2007: 2907-2911 (doi: 10.1109/IEMBS.2007.4352937)
      17. J. Talbot, J. Meyer, R. Watts, and R. Grasset, ‘An Augmented Reality Application for Patient Positioning and Monitoring in Radiotherapy’, in World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany, Berlin, Heidelberg, 2009, pp. 620–623, doi: 10.1007/978-3-642-03474-9_174.

        • Aydogan B.
        • Mundt A.J.
        • Roeske J.C.
        Linac-based intensity modulated total marrow irradiation (IM-TMI).
        Technol Cancer Res Treat. 2006; 5: 513-519https://doi.org/10.1177/153303460600500508
      18. JWT Talbot, ‘A patient position guidance system in radiotherapy using augmented reality’, 2009.

        • Talbot J.
        • Meyer J.
        • Watts R.
        • Grasset R.
        A method for patient set-up guidance in radiotherapy using augmented reality.
        Australas Phys Eng Sci Med. 2009; 32: 203-211https://doi.org/10.1007/BF03179240
        • Mueller A.
        Visualisation of radiation treatment parameters in an Augmented Reality environment.
        (BSc(Hons) Thesis) University of Canterbury, Christchurch, New Zealand2009
        • Green O.L.
        • Henke L.E.
        • Hugo G.D.
        Practical clinical workflows for online and offline adaptive radiation therapy.
        Semin Radiat Oncol. 2019; 29: 219-227https://doi.org/10.1016/j.semradonc.2019.02.004
      19. O. El‐Sherif, N. B. Remmes, and J. J. Kruse, ‘Validating robotic couch isocentricity with 3D surface imaging’, J Appl Clin Med Phys, doi: 10.1002/acm2.12939.

        • Kim M.
        • et al.
        Development of a 3D optical scanning-based automatic quality assurance system for proton range compensators.
        Med Phys. 2015; 42: 1071-1079https://doi.org/10.1118/1.4906131
        • Pawlicki T.
        • Dunscombe P.
        • Mundt A.J.
        • Scalliet P.
        Quality and safety in radiotherapy.
        CRC Press, 2010
        • Meyer J.
        • et al.
        Characterizing a deformable registration algorithm for surface-guided breast radiotherapy.
        Med Phys. 2019; https://doi.org/10.1002/mp.13921
      20. University of Zurich, ‘Randomized Controlled Trial Comparing Closed vs. Open Face Masks for Cranial Radiotherapy’, clinicaltrials.gov, Clinical trial registration NCT04079595, Sep. 2019. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT04079595.

      21. D. I. Karam, ‘Assessment of Left-sided Cardiac Sparing Through the Use of 3-dimensional Surface Matching-based Deep Inspiration Breath Hold and Active Breathing Control’, clinicaltrials.gov, Clinical trial registration NCT03459898, Jun. 2020. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT03459898.

      22. The University of Texas Health Science Center at San Antonio, ‘Demonstration of a Novel Approach Using Surface-Image Guidance to Improve Delivery of Breast Radiotherapy’, clinicaltrials.gov, Clinical trial registration NCT03799523, Jul. 2020. Accessed: Sep. 02, 2020. [Online]. Available: https://clinicaltrials.gov/ct2/show/NCT03799523.

      23. ClinicalTrials.gov, ‘Radiotherapy Delivery in Deep Inspiration for Pediatric Patients’. https://clinicaltrials.gov/ct2/show/NCT03315546 (accessed Sep. 03, 2020).