- Email: [email protected]
Image Fusion in 2020
e2
Abstracts / Physica Medica 30 (2014) e1ee15
Recently, 2 scientific papers have confirmed good performance of our screening program. These studies are also a confirmation for the group of physicists who have worked at a strict physic-technical quality assurance system. The networking approach could be copied in other countries of course. Session number of session in which the abstract is presented: Scientific session: QA in radiology; Thursday afternoon, 14.30 e 15.30 Session title of session in which the abstract is presented FULLY AUTOMATED TREATMENT PLAN GENERATION IN DAILY ROUTINE B. Heijmen, P. Voet, M. Dirkx, A. Sharfo, L. Rossi, D. Fransen, J. Penninkhof, ndez Romero, A. Al-Mamgani, L. M. Hoogeman, S. Petit, J.-W. Mens, A. Me Incrocci, S. Breedveld. Erasmus MC Cancer Institute, Rotterdam, The Netherlands Background: Currently, treatment plans are generated by dosimetrists using a trial-and-error procedure. The process may take several hours and plan quality is dependent on the skills and experience of the dosimetrist, and on allotted time. We have developed and clinically introduced a system for fully automatic plan generation, using lexicographic multi-criterial optimization to replace the labour-intensive and operator-dependent trialand-error approach. Materials and Methods: For each patient, the treatment plan is fully automatically generated by the clinical treatment planning system (Monaco, Elekta AB), based on a patient-specific template that is automatically pre-generated with our in-house lexicographic multi-criterial optimizer (“Erasmus-iCycle”, Med Phys. 2012; 39(2): 951). Automatic plan generation in Erasmus-iCycle is based on a ‘wishlist’ with hard constraints and treatment objectives with assigned priorities. For each treatment site (e.g. H&N cancer), a single fixed wish-list is used for all patients. In case of IMRT, Erasmus-iCycle can be used for integrated beam profile optimization and (non-coplanar) beam angle selection. Results: In a prospective clinical H&N cancer study, radiation oncologists selected the AUTO-plan in 97% of cases rather than the MANUAL-plan generated by trial-and-error (IJROBP 2013; 85(3): 866-72). For a group of 44 cervical cancer patients, dual-arc VMAT AUTO-plans were superior to MANUAL-plans generated by an expert cervical cancer planner, spending many hours; reduced small bowel V15Gy, V45Gy, and Dmean, bladder Dmean, and rectum Dmean, p<0.001. For 30 prostate cancer patients, differences between VMAT AUTO- and MANUAL-plans, the latter generated by an expert planner with up to 4 hours planning hands-on time, were statistically insignificant (IJROBP 2014; 88(5): 1175-9). Discussion: Automatic plan generation with consistent high plan quality and vast reductions in planning workload is feasible and has been clinically introduced for major treatment sites.
imaging systems, as well as the ability to adopt technical advances more easily in combined imaging hardware, SPECT/CT, PET/CT and PET/MRI will likely be explored to their full potential. This includes a trend towards acquiring listmode data routinely and employing different type of models (kinetic, motion, etc) for subsequent corrections. With PET/MRI, in particular, we will likely see a trend towards assessing tumor heterogeneity through the adoption of dedicated MR sequences applied locoregionally. This information can be used further for improved imageguided interventions. We will likely see a revival of ideas on anatomyguided corrections (e.g., PVC) and reconstructions with the wider adoption of combined imaging. Image fusion in 2020, and beyond, will strongly depend on our ability to accept combined imaging as a collaborative affair and our intent and endurance to cooperate with neighbouring disciplines. This pertains to primarily to medical doctors, but also to technologists and medical physicists. THE ANTIKYTHERA MECHANISM: DECODING ANCIENT GREEK ASTRONOMICAL COMPUTER
AN
ASTONISHING
John H. Seiradakis. Aristotle University, Department of Physics, Section of Astrophysics, Astronomy & Mechanics, GR-541 24 Thessaloniki, Greece The Antikythera Mechanism was found by chance, in a shipwreck, close to the small Greek island of Antikythera, in April 1900, by sponge divers. The shipwreck was dated between 86 and 67 B.C. (coins from Pergamon). Later the Mechanism was stylistically dated, around the second half of the 2nd century B.C. (100 e 150 B.C.). About this time the great Greek astronomer Hipparchos (190 e 120 B.C.) lived in Rhodes. It was a portable (laptop-size), geared mechanism which calculated and displayed, with high precision, the movement of the Sun and the Moon on the sky and the phase of the Moon for a given epoch. It could also calculate the dates of the four-year cycle of the Olympic Games. It had one dial on the front and two on the back. Its 30, precisely cut, gears were driven by a manifold, with which the user could select, with the help of a pointer, any particular epoch (at the front dial). While doing so, several pointers were synchronously driven by the gears, to show the above mentioned celestial phenomena on several accurately marked annuli. It contained an extensive user’s manual. The exact function of the gears has finally been decoded and a large portion of the manual has been read after 2000 years by a major new investigation, using state of the art equipment. Based on new surface photography and high resolution tomography data, a new model has been built at the Aristotle University, revealing the technological abilities of ancient greeks. No complicated geared instruments are known before the Antikythera Mechanism and for several centuries after. Therefore, this astronomical device stands out as an extraordinary proof of high tech in ancient times.
IMAGE FUSION IN 2020 Thomas Beyer. Medical University of Vienna, Austria
CURRENT APPROACH IN CLINICAL ELECTRON BEAM DOSIMETRY
Image fusion describes the process of spatial-temporal alignment of complementary image information. Images can be acquired retrospectively or prospectively. Most commonly image sets originate from different image modalities. However, image sets from the same modality can be aligned and fused also for the purpose of longitudinal assessments. Image fusion is commonly motivated by the need to gain additional information from the aligned image sets, such as by localizing small lesions on functional, low-resolution images through matched, high-resolution anatomical images. This presentation reviews briefly the origin of clinical image fusion, which traces back to the 1960’s when the outline of the neck of patients was transposed manually on scintigrams for better localization of thyroid uptake and nodular disease. Image fusion advanced quickly by means of adopting computer-based automated processing. With the introduction of prototype PET/CT and SPECT/CT systems by the late 1990’s image fusion became hardware based, thus, offering a number of advantages over retrospective image fusion. Since 2006 combined PET/MR has become available for use in humans. This sets the stage for rapid advances in image fusion in the years to come. Given the availability of high-end imaging components in the combined
Dimitris Mihailidis PhD. Charleston Radiation Therapy and West Virginia University, USA The absolute dosimetry calibration of clinical electron beams is increasingly based on the AAPM Task Group #51 (TG-51) protocol. In addition, recently published dosimetry data on electrons beams bring up the question of: how would one need to modify the widely used TG-25 that originally was based on the older AAPM Task Group #21 (TG-21) calibration protocol? The 2009 Task Group #70 (TG-70) by the AAPM, is trying to address the issue above. TG-70 operates as supplement and update to TG-25 on issues that needed to be modified because of TG-51 approach to electron dosimetry and because of the more recent data on clinical electron beams. It describes in detail the method of converting measured depth-ionization curves with ion chambers into depth-dose curves, making use of recently published stopping-power ratios and other conversion factors. It also describes the use of water equivalent phantoms to perform relative electron dosimetry based on recently published conversions factors. The report discusses small and irregularly shaped electron field dosimetry using the
Abstracts / Physica Medica 30 (2014) e1ee15
Recently, 2 scientific papers have confirmed good performance of our screening program. These studies are also a confirmation for the group of physicists who have worked at a strict physic-technical quality assurance system. The networking approach could be copied in other countries of course. Session number of session in which the abstract is presented: Scientific session: QA in radiology; Thursday afternoon, 14.30 e 15.30 Session title of session in which the abstract is presented FULLY AUTOMATED TREATMENT PLAN GENERATION IN DAILY ROUTINE B. Heijmen, P. Voet, M. Dirkx, A. Sharfo, L. Rossi, D. Fransen, J. Penninkhof, ndez Romero, A. Al-Mamgani, L. M. Hoogeman, S. Petit, J.-W. Mens, A. Me Incrocci, S. Breedveld. Erasmus MC Cancer Institute, Rotterdam, The Netherlands Background: Currently, treatment plans are generated by dosimetrists using a trial-and-error procedure. The process may take several hours and plan quality is dependent on the skills and experience of the dosimetrist, and on allotted time. We have developed and clinically introduced a system for fully automatic plan generation, using lexicographic multi-criterial optimization to replace the labour-intensive and operator-dependent trialand-error approach. Materials and Methods: For each patient, the treatment plan is fully automatically generated by the clinical treatment planning system (Monaco, Elekta AB), based on a patient-specific template that is automatically pre-generated with our in-house lexicographic multi-criterial optimizer (“Erasmus-iCycle”, Med Phys. 2012; 39(2): 951). Automatic plan generation in Erasmus-iCycle is based on a ‘wishlist’ with hard constraints and treatment objectives with assigned priorities. For each treatment site (e.g. H&N cancer), a single fixed wish-list is used for all patients. In case of IMRT, Erasmus-iCycle can be used for integrated beam profile optimization and (non-coplanar) beam angle selection. Results: In a prospective clinical H&N cancer study, radiation oncologists selected the AUTO-plan in 97% of cases rather than the MANUAL-plan generated by trial-and-error (IJROBP 2013; 85(3): 866-72). For a group of 44 cervical cancer patients, dual-arc VMAT AUTO-plans were superior to MANUAL-plans generated by an expert cervical cancer planner, spending many hours; reduced small bowel V15Gy, V45Gy, and Dmean, bladder Dmean, and rectum Dmean, p<0.001. For 30 prostate cancer patients, differences between VMAT AUTO- and MANUAL-plans, the latter generated by an expert planner with up to 4 hours planning hands-on time, were statistically insignificant (IJROBP 2014; 88(5): 1175-9). Discussion: Automatic plan generation with consistent high plan quality and vast reductions in planning workload is feasible and has been clinically introduced for major treatment sites.
imaging systems, as well as the ability to adopt technical advances more easily in combined imaging hardware, SPECT/CT, PET/CT and PET/MRI will likely be explored to their full potential. This includes a trend towards acquiring listmode data routinely and employing different type of models (kinetic, motion, etc) for subsequent corrections. With PET/MRI, in particular, we will likely see a trend towards assessing tumor heterogeneity through the adoption of dedicated MR sequences applied locoregionally. This information can be used further for improved imageguided interventions. We will likely see a revival of ideas on anatomyguided corrections (e.g., PVC) and reconstructions with the wider adoption of combined imaging. Image fusion in 2020, and beyond, will strongly depend on our ability to accept combined imaging as a collaborative affair and our intent and endurance to cooperate with neighbouring disciplines. This pertains to primarily to medical doctors, but also to technologists and medical physicists. THE ANTIKYTHERA MECHANISM: DECODING ANCIENT GREEK ASTRONOMICAL COMPUTER
AN
ASTONISHING
John H. Seiradakis. Aristotle University, Department of Physics, Section of Astrophysics, Astronomy & Mechanics, GR-541 24 Thessaloniki, Greece The Antikythera Mechanism was found by chance, in a shipwreck, close to the small Greek island of Antikythera, in April 1900, by sponge divers. The shipwreck was dated between 86 and 67 B.C. (coins from Pergamon). Later the Mechanism was stylistically dated, around the second half of the 2nd century B.C. (100 e 150 B.C.). About this time the great Greek astronomer Hipparchos (190 e 120 B.C.) lived in Rhodes. It was a portable (laptop-size), geared mechanism which calculated and displayed, with high precision, the movement of the Sun and the Moon on the sky and the phase of the Moon for a given epoch. It could also calculate the dates of the four-year cycle of the Olympic Games. It had one dial on the front and two on the back. Its 30, precisely cut, gears were driven by a manifold, with which the user could select, with the help of a pointer, any particular epoch (at the front dial). While doing so, several pointers were synchronously driven by the gears, to show the above mentioned celestial phenomena on several accurately marked annuli. It contained an extensive user’s manual. The exact function of the gears has finally been decoded and a large portion of the manual has been read after 2000 years by a major new investigation, using state of the art equipment. Based on new surface photography and high resolution tomography data, a new model has been built at the Aristotle University, revealing the technological abilities of ancient greeks. No complicated geared instruments are known before the Antikythera Mechanism and for several centuries after. Therefore, this astronomical device stands out as an extraordinary proof of high tech in ancient times.
IMAGE FUSION IN 2020 Thomas Beyer. Medical University of Vienna, Austria
CURRENT APPROACH IN CLINICAL ELECTRON BEAM DOSIMETRY
Image fusion describes the process of spatial-temporal alignment of complementary image information. Images can be acquired retrospectively or prospectively. Most commonly image sets originate from different image modalities. However, image sets from the same modality can be aligned and fused also for the purpose of longitudinal assessments. Image fusion is commonly motivated by the need to gain additional information from the aligned image sets, such as by localizing small lesions on functional, low-resolution images through matched, high-resolution anatomical images. This presentation reviews briefly the origin of clinical image fusion, which traces back to the 1960’s when the outline of the neck of patients was transposed manually on scintigrams for better localization of thyroid uptake and nodular disease. Image fusion advanced quickly by means of adopting computer-based automated processing. With the introduction of prototype PET/CT and SPECT/CT systems by the late 1990’s image fusion became hardware based, thus, offering a number of advantages over retrospective image fusion. Since 2006 combined PET/MR has become available for use in humans. This sets the stage for rapid advances in image fusion in the years to come. Given the availability of high-end imaging components in the combined
Dimitris Mihailidis PhD. Charleston Radiation Therapy and West Virginia University, USA The absolute dosimetry calibration of clinical electron beams is increasingly based on the AAPM Task Group #51 (TG-51) protocol. In addition, recently published dosimetry data on electrons beams bring up the question of: how would one need to modify the widely used TG-25 that originally was based on the older AAPM Task Group #21 (TG-21) calibration protocol? The 2009 Task Group #70 (TG-70) by the AAPM, is trying to address the issue above. TG-70 operates as supplement and update to TG-25 on issues that needed to be modified because of TG-51 approach to electron dosimetry and because of the more recent data on clinical electron beams. It describes in detail the method of converting measured depth-ionization curves with ion chambers into depth-dose curves, making use of recently published stopping-power ratios and other conversion factors. It also describes the use of water equivalent phantoms to perform relative electron dosimetry based on recently published conversions factors. The report discusses small and irregularly shaped electron field dosimetry using the