Is proton beam therapy always better than photon irradiation? Lessons from two cases.

Proton beam therapy (PBT) is increasingly used to treat cancers, especially in the paediatric and adolescent and young adult (AYA) population. As PBT becomes more accessible, determining when PBT should be used instead of photon irradiation can be difficult. There is a need to balance patient, tumour and treatment factors when making this decision. Comparing the dosimetry between these two modalities plays an important role in this process. PBT can reduce low to intermediate doses to organs at risk (OAR), but photon irradiation has its dosimetric advantages. We present two cases with brain tumours, one paediatric and one AYA, in which treatment plan comparison between photons and protons showed dosimetric advantages of photon irradiation. The first case was an 18-month-old child diagnosed with posterior fossa ependymoma requiring adjuvant radiotherapy. Photon irradiation using volumetric modulated arc therapy (VMAT) had lower doses to the hippocampi but higher doses to the pituitary gland. The second case was a 21-year-old with an optic pathway glioma. There was better sparing of the critical optic structures and pituitary gland using fractionated stereotactic radiation therapy over PBT. The dosimetric advantages of photon irradiation over PBT have been demonstrated in these cases. This highlights the role of proton-to-photon comparative treatment planning to better understand which patients might benefit from photon irradiation versus PBT.

Efficacious patient-specific QA for Vertebra SBRT using a high-resolution detector array SRS MapCHECK: AAPM TG-218 analysis.

PURPOSE: Patient-specific quality assurance (PSQA) for vertebra stereotactic body radiation therapy (SBRT) presents challenges due to highly modulated small fields with high-dose gradients between the target and spinal cord. This study aims to explore the use of the SRS MapCHECK® (SRSMC) for vertebra SBRT PSQA. METHODS: Twenty vertebra SBRT treatment plans including prescriptions 20 Gy/1 fraction and 24 Gy/2 fractions were selected for each of Millennium (M)-Multileaf Collimator (MLC), and high-definition (HD)-MLC. All 40 plans were measured using Gafchromic EBT3 film (film) and SRSMC, using the StereoPHAN phantom. Plan complexity was assessed using modulation complexity score (MCS), edge metric (EM) (mm-1), modulation factor (MU/cGy), and average leaf pair opening (ALPO) (mm) and its correlation with gamma-pass rate was investigated. The high dose gradient between the target and the spinal cord was analyzed for film and SRSMC and compared against the treatment planning system (TPS). Applying the methodology proposed by AAPM TG-218, action and tolerance values specific to the SRSMC for vertebra SBRT were determined for ß values ranging from 5 to 8. RESULTS: Film and SRSMC gamma-pass rates showed no correlation (p > 0.05). A moderate negative correlation (R = -0.57, p = 0.01) is present between EM and SRSMC 3%/1 mm gamma-pass rate for HD-MLC plans. Both film and SRSMC accurately measured high dose gradients between the target and the spinal cord (R2 > 0.86, p ≤ 0.05). Notably, dose-gradient of HD-MLC plans is 22% steeper and has a smaller standard deviation to M-MLC plans (p ≤ 0.05). Applying TG-218, the film tolerance limit was 96% with action limit 95% for 5%/1 mm (ß = 6) and for the SRSMC tolerance limit was 97% with an action limit of 96% for 4%/1 mm (ß = 6). CONCLUSION: Our findings suggest that universal TG-218 limits may not be suitable for vertebra SBRT PSQA. This study demonstrates that SRSMC is a viable tool for vertebra SBRT PSQA, supported by TG-218 implementation of process-based tolerance and action limits.

Developing a comparative photon-proton planning service in Victoria: the experience at Peter MacCallum Cancer Centre.

Proton-beam therapy (PBT) is a cutting-edge radiation therapy modality that is currently not available in Australia. Comparative photon-proton (CPP) planning is required for the medical treatment overseas programme (MTOP) and will be required for access to PBT in Australia in the future. Comparative planning brings professional development benefits to all members of the radiation therapy team. This service was also created to support future proposals for a PBT facility in Victoria. We report our experience developing an in-house CPP service at Peter MacCallum Cancer Centre. A set of resources to support CPP planning was established. Training of relevant staff was undertaken after which an in-house training programme was developed. A standard protocol for PBT planning parameters was established. All CPP plans were reviewed. Future goals for the CPP planning programme were described. In total, 62 cases were comparatively planned over 54 months. Of these, 60% were paediatric cases, 14% were adolescents and young adults (15-25 years) and 26% were adults. The vast majority (over 75%) of patients comparatively planned required irradiation to the central nervous system including brain and cranio-spinal irradiation. A variety of proton plans were reviewed by international PBT experts to confirm their deliverability. Our team at Peter MacCallum Cancer Centre has gained significant experience in CPP planning and will continue to develop this further. Local expertise will help support decentralisation of patient selection for proton treatments in the near future and the PBT business case in Victoria.

Partially Ablative Body Radiotherapy (PABR): A novel approach for palliative radiotherapy of locally advanced bulky unresectable sarcomas.

BACKGROUND: Locally advanced, bulky, unresectable sarcomas cause significant tumour mass effects, leading to burdensome symptoms. We have developed a novel Partially Ablative Body Radiotherapy (PABR) technique that delivers a high, ablative dose to the tumour core and a low, palliative dose to its periphery aiming to increase overall tumour response without significantly increasing treatment toxicity. AIM: This study aims to report the safety and oncologic outcomes of PABR in patients with bulky, unresectable sarcomas. METHODS AND MATERIALS: A total of 18 patients with histologically proven sarcoma treated with PABR from January 2020 to October 2023 were retrospectively reviewed. The primary endpoints were symptomatic and structural response rates. Secondary endpoints were overall survival, freedom from local progression, freedom from distant progression, and acute and late toxicity rates. RESULTS: All patients had tumours ≥5 cm with a median tumour volume of 985 cc, and the most common symptom was pain. The median age is 72.5 years and 44.5 % were ECOG 2-3. The most common regimen used was 20 Gy in 5 fractions with an intratumoral boost dose of 50 Gy (83.3 %). After a median follow-up of 11 months, 88.9 % of patients exhibited a partial response with a mean absolute tumour volume reduction of 49.5 %. All symptomatic patients experienced symptom improvement. One-year OS, FFLP and FFDP were 61 %, 83.3 % and 34.8 %, respectively. There were no grade 3 or higher toxicities. CONCLUSION: PABR for bulky, unresectable sarcomas appears to be safe and may provide good symptomatic response, tumour debulking, and local control. Further study is underway.

Deep Learning Auto-Segmentation Network for Pediatric Computed Tomography Data Sets: Can We Extrapolate From Adults?

PURPOSE: Artificial intelligence (AI)-based auto-segmentation models hold promise for enhanced efficiency and consistency in organ contouring for adaptive radiation therapy and radiation therapy planning. However, their performance on pediatric computed tomography (CT) data and cross-scanner compatibility remain unclear. This study aimed to evaluate the performance of AI-based auto-segmentation models trained on adult CT data when applied to pediatric data sets and explore the improvement in performance gained by including pediatric training data. It also examined their ability to accurately segment CT data acquired from different scanners. METHODS AND MATERIALS: Using the nnU-Net framework, segmentation models were trained on data sets of adult, pediatric, and combined CT scans for 7 pelvic/thoracic organs. Each model was trained on 290 to 300 cases per category and organ. Training data sets included a combination of clinical data and several open repositories. The study incorporated a database of 459 pediatric (0-16 years) CT scans and 950 adults (>18 years), ensuring all scans had human expert ground-truth contours of the selected organs. Performance was evaluated based on Dice similarity coefficients (DSC) of the model-generated contours. RESULTS: AI models trained exclusively on adult data underperformed on pediatric data, especially for the 0 to 2 age group: mean DSC was below 0.5 for the bladder and spleen. The addition of pediatric training data demonstrated significant improvement for all age groups, achieving a mean DSC of above 0.85 for all organs in every age group. Larger organs like the liver and kidneys maintained consistent performance for all models across age groups. No significant difference emerged in the cross-scanner performance evaluation, suggesting robust cross-scanner generalization. CONCLUSIONS: For optimal segmentation across age groups, it is important to include pediatric data in the training of segmentation models. The successful cross-scanner generalization also supports the real-world clinical applicability of these AI models. This study emphasizes the significance of data set diversity in training robust AI systems for medical image interpretation tasks.

Additive manufacturing of patient specific bolus for radiotherapy: large scale production and quality assurance.

Bolus is commonly used to improve dose distributions in radiotherapy in particular if dose to skin must be optimised such as in breast or head and neck cancer. We are documenting four years of experience with 3D printed bolus at a large cancer centre. In addition to this we review the quality assurance (QA) program developed to support it. More than 2000 boluses were produced between Nov 2018 and Feb 2023 using fused deposition modelling (FDM) printing with polylactic acid (PLA) on up to five Raise 3D printers. Bolus is designed in the radiotherapy treatment planning system (Varian Eclipse), exported to an STL file followed by pre-processing. After checking each bolus with CT scanning initially we now produce standard quality control (QC) wedges every month and whenever a major change in printing processes occurs. A database records every bolus printed and manufacturing details. It takes about 3 days from designing the bolus in the planning system to delivering it to treatment. A 'premium' PLA material (Spidermaker) was found to be best in terms of homogeneity and CT number consistency (80 HU +/- 8HU). Most boluses were produced for photon beams (93.6%) with the rest used for electrons. We process about 120 kg of PLA per year with a typical bolus weighing less than 500 g and the majority of boluses 5 mm thick. Print times are proportional to bolus weight with about 24 h required for 500 g material deposited. 3D printing using FDM produces smooth and reproducible boluses. Quality control is essential but can be streamlined.

Development of an automated treatment planning approach for lattice radiation therapy.

BACKGROUND: Lattice radiation therapy (LRT) alternates regions of high and low doses within the target. The heterogeneous dose distribution is delivered to a geometrical structure of vertices segmented inside the tumor. LRT is typically used to treat patients with large tumor volumes with cytoreduction intent. Due to the geometric complexity of the target volume and the required dose distribution, LRT treatment planning demands additional resources, which may limit clinical integration. PURPOSE: We introduce a fully automated method to (1) generate an ordered lattice of vertices with various sizes and center-to-center distances and (2) perform dose optimization and calculation. We aim to report the dosimetry associated with these lattices to help clinical decision-making. METHODS: Sarcoma cancer patients with tumor volume between 100 cm3 and 1500 cm3 who received radiotherapy treatment between 2010 and 2018 at our institution were considered for inclusion. Automated segmentation and dose optimization/calculation were performed by using the Eclipse Scripting Application Programming Interface (ESAPI, v16, Varian Medical Systems, Palo Alto, USA). Vertices were modeled by spheres segmented within the gross tumor volume (GTV) with 1 cm/1.5 cm/2 cm diameters (LRT-1 cm/1.5 cm/2 cm) and 2 to 5 cm center-to-center distance on square lattices alternating along the superior-inferior direction. Organs at risk were modeled by subtracting the GTV from the body structure (body-GTV). The prescription dose was that 50% of the vertice volume should receive at least 20 Gy in one fraction. The automated dose optimization included three stages. The vertices optimization objectives were refined during optimization according to their values at the end of the first and second stages. Lattices were classified according to a score based on the minimization of body-GTV max dose and the maximization of GTV dose uniformity (measured with the equivalent uniform dose [EUD]), GTV dose heterogeneity (measured with the GTV D90%/D10% ratio), and the number of patients with more than one vertex inserted in the GTV. Plan complexity was measured with the modulation complexity score (MCS). Correlations were assessed with the Spearman correlation coefficient (r) and its associated p-value. RESULTS: Thirty-three patients with GTV volumes between 150 and 1350 cm3 (median GTV volume = 494 cm3 , IQR = 272-779 cm3 were included. The median time required for segmentation/planning was 1 min/21 min. The number of vertices was strongly correlated with GTV volume in each LRT lattice for each center-to-center distance (r > 0.85, p-values < 0.001 in each case). Lattices with center-to-center distance = 2.5 cm/3 cm/3.5 cm in LRT-1.5 cm and center-to-center distance = 4 cm in LRT-1 cm had the best scores. These lattices were characterized by high heterogeneity (median GTV D90%/D10% between 0.06 and 0.19). The generated plans were moderately complex (median MCS ranged between 0.19 and 0.40). CONCLUSIONS: The automated LRT planning method allows for the efficacious generation of vertices arranged in an ordered lattice and the refinement of planning objectives during dose optimization, enabling the systematic evaluation of LRT dosimetry from various lattice geometries.

Evaluating synthetic computed tomography images for adaptive radiotherapy decision making in head and neck cancer.

Background and purpose: Adaptive radiotherapy (ART) decision-making benefits from dosimetric information to supplement image inspection when assessing the significance of anatomical changes. This study evaluated a dosimetry-based clinical decision workflow for ART utilizing deformable registration of the original planning computed tomography (CT) image to the daily Cone Beam CT (CBCT) to replace the need for a replan CT for dose estimation. Materials and methods: We used 12 retrospective Head & Neck patient cases having a ground truth - a replan CT (rCT) in response to anatomical changes apparent in the daily CBCT - to evaluate the accuracy of dosimetric assessment conducted on synthetic CTs (sCT) generated by deforming the original planning CT Hounsfield Units to the daily CBCT anatomy.The original plan was applied to the sCT and dosimetric accuracy of the sCT was assessed by analyzing plan objectives for targets and organs-at-risk compared to calculations on the ground-truth rCT. Three commercial DIR algorithms were compared. Results: For the best-performing algorithms, the majority of dose metrics calculated on the sCTs differed by less than 4 Gy (5.7% of 70 Gy prescription dose). An uncertainty of ±2.5 Gy (3.6% of 70 Gy prescription) is recommended as a conservative tolerance when evaluating dose metrics on sCTs for head and neck. Conclusions: Synthetic CTs present a valuable addition to the adaptive radiotherapy workflow, and synthetic CT dose estimates can be effectively used in addition to the current practice of visually inspecting the overlay of the planning CT and CBCT to assess the significance of anatomical change.

Proton-to-photon comparative treatment planning guidelines for the Australian context.

Proton-to-photon comparative treatment planning is a current requirement of Australian Government funding for patients to receive proton beam therapy (PBT) overseas, and a future requirement for Medicare funding of PBT in Australia. Because of the fundamental differences in treatment plan creation and evaluation between PBT and conventional radiation therapy with x-rays (XRT), there is the potential for a lack of consistency in the process of comparing PBT and XRT treatment plans. This may have an impact on patient eligibility assessment for PBT. The objective of these guidelines is to provide a practical reference document for centres performing proton-to-photon comparative planning and thereby facilitate national uniformity.

Selection of motion management in liver stereotactic body radiotherapy and its impact on treatment time.

Background and purpose: Reduction of respiratory tumour motion is important in liver stereotactic body radiation therapy (SBRT) to reduce side effects and improve tumour control probability. We have assessed the distribution of use of voluntary exhale breath hold (EBH), abdominal compression (AC), free breathing gating (gating) and free breathing (FB), and the impact of these on treatment time. Materials and Methods: We assessed all patients treated in a single institution with liver SBRT between September 2017 and September 2021. Data from pre-simulation motion management assessment using fluoroscopic assessment of liver dome position in repeat breath holds, and motion with and without AC, was reviewed to determine liver dome position consistency in EBH and the impact of AC on motion. Treatment time was assessed for all fractions as time from first image acquisition to last treatment beam off. Results: Of 136 patients treated with 145 courses of liver SBRT, 68 % were treated in EBH, 20 % with AC, 7 % in gating and 5 % in FB. AC resulted in motion reduction < 1 mm in 9/26 patients assessed. Median treatment time was higher using EBH (39 min) or gating (42 min) compared with AC (30 min) or FB (24 min) treatments. Conclusions: Motion management in liver SBRT needs to be assessed per-patient to ensure appropriate techniques are applied. Motion management significantly impacts treatment time therefore patient comfort must also be taken into account when selecting the technique for each patient.

Treatment Time Optimization in Single Fraction Stereotactic Ablative Radiation Therapy: A 10-Year Institutional Experience.

Purpose: Stereotactic ablative radiation therapy (SABR) delivered in a single fraction (SF) can be considered to have higher uncertainty given that the error probability is concentrated in a single session. This study aims to report the variation in technology and technique used and its effect on intrafraction motion based on a 10 years of experience in SF SABR. Methods and Materials: Records of patients receiving SF SABR delivered at our instruction between 2010 and 2019 were included. Treatment parameters were extracted from the patient management database by using an in-house script. Treatment time was defined as the time difference between the first image acquisition to the last beam off of a single session. The intrafraction variation was measured from the 3-dimensional couch displacement measured after the first cone beam computed tomography (CBCT) acquired during a treatment. Results: The number of SF SABR increased continuously from 2010 to 2019 and were mainly lung treatments. Treatment time was minimized by using volumetric modulated arc therapy, flattening filter-free dose rate, and coplanar field (24 ± 9 min). Treatment time increased as the number of CBCTs per session increased. The most common scenario involved both 2 and 3 CBCTs per session. On the average, a CBCT acquisition added 6 minutes to the treatment time. All treatments considered, the average intrafraction variation was 1.7 ± 1.6 mm. Conclusions: SF SABR usage increased with time in our institution. The intrafraction motion was acceptable and therefore a single fraction is an efficacious treatment option when considering SABR.

MIRSIG position paper: the use of image registration and fusion algorithms in radiotherapy.

The report of the American Association of Physicists in Medicine (AAPM) Task Group No. 132 published in 2017 reviewed rigid image registration and deformable image registration (DIR) approaches and solutions to provide recommendations for quality assurance and quality control of clinical image registration and fusion techniques in radiotherapy. However, that report did not include the use of DIR for advanced applications such as dose warping or warping of other matrices of interest. Considering that DIR warping tools are now readily available, discussions were hosted by the Medical Image Registration Special Interest Group (MIRSIG) of the Australasian College of Physical Scientists & Engineers in Medicine in 2018 to form a consensus on best practice guidelines. This position statement authored by MIRSIG endorses the recommendations of the report of AAPM task group 132 and expands on the best practice advice from the 'Deforming to Best Practice' MIRSIG publication to provide guidelines on the use of DIR for advanced applications.

Active small bowel sparing in intracavitary brachytherapy for cervical cancer.

OBJECTIVE: To propose and evaluate an active method for sparing the small bowel in the treatment field of cervical cancer brachytherapy by prone position procedure. METHODS: The prone position procedure consists of five steps: making bladder empty, prone-positioning a patient on belly board, making the small bowel move to abdomen, filling the bladder with Foley catheter and finally turning the patient into the supine position. The proposed method was applied for the treatment of seven cervical cancer patients. Its effectiveness was evaluated and a correlation between the patient characteristics and the volumetric dose reduction of small bowel was also investigated. Brachytherapy treatment plans were built before and after the proposed method, and their dose-volume histograms were compared for targets and organs-at-risk. In this comparison, all plans were normalized to satisfy the same D90% for high-risk clinical target volume. RESULTS: For the enrolled patients, the average dose of small bowel was significantly reduced from 75.2 ± 4.9 Gy before to 60.2 ± 4.0 Gy after the prone position procedure, while minor dosimetric changes were observed in rectum, sigmoid and bladder. The linear correlation to body mass index, thickness and width of abdominopelvic cavity and bladder volume were 76.2, 69.7, 28.8 and -36.3%, respectively. CONCLUSIONS: The application of prone position procedure could effectively lower the volumetric dose of the small bowel. The dose reduction in the small bowel had a strong correlation with the patient's obesity and abdominal thickness. This means the patients for whom the proposed method would be beneficial can be judiciously selected for safe brachytherapy.

A customizable anthropomorphic phantom for dosimetric verification of 3D-printed lung, tissue, and bone density materials.

PURPOSE: To design and manufacture a customized thoracic phantom slab utilizing the 3D printing process, also known as additive manufacturing, consisting of different tissue density materials. Here, we demonstrate the 3D-printed phantom's clinical feasibility for imaging and dosimetric verification of volumetric modulated arc radiotherapy (VMAT) plans for lung and spine stereotactic ablative body radiotherapy (SABR) through end-to-end dosimetric verification. METHODS: A customizable anthropomorphic phantom slab was designed using the CT dataset of a commercial phantom (adult female ATOM dosimetry phantom, CIRS Inc.). Material extrusion 3D printing was utilized to manufacture the phantom slab consisting of acrylonitrile butadiene styrene material for the lung and the associated lesion, polylactic acid (PLA) material for soft tissue and spinal cord, and both PLA and iron-reinforced PLA materials for bone. CT images were acquired for both the commercial phantom and 3D-printed phantom for HU comparison. VMAT plans were generated for spine and lung SABR scenarios and were delivered as per departmental SABR protocols using a Varian TrueBeam STx linear accelerator. End-to-end dosimetry was implemented with radiochromic films, analyzed with gamma criteria of 5% dose difference, and a distance-to-agreement of 1 mm, at a 10% low-dose threshold by comparing with calculated dose using the Acuros algorithm of the Eclipse treatment planning system (v15.6). RESULTS: 3D-printed phantom inserts were observed to produce HU ranging from -750 to 2100. The 3D-printed phantom slab was observed to achieve a similar range of HU from the commercial phantom including a mean HU of -760 for lung tissue, a mean HU of 50 for soft tissue, and a mean HU of 220 and 630 for low- and high-density bone, respectively. Film dosimetry results show 2D-gamma passing rates for lung SABR (internal and superior) and spine SABR (inferior and superior) over 98% and 90%, respectively. CONCLUSIONS: The end-to-end testing of VMAT plans for spine and lung SABR suggests the clinical feasibility of the 3D-printed phantom, consisting of different tissue density materials that emulate lung, soft tissue, and bone in kV imaging and megavoltage photon dosimetry. Further investigation of the proposed 3D printing techniques for manufacturability and reproducibility will enable the development of clinical 3D-printed phantoms in radiotherapy.

Out-of-field dose in stereotactic radiotherapy for paediatric patients.

BACKGROUND AND PURPOSE: Stereotactic radiotherapy combines image guidance and high precision delivery with small fields to deliver high doses per fraction in short treatment courses. In preparation for extension of these treatment techniques to paediatric patients we characterised and compared doses out-of-field in a paediatric anthropomorphic phantom for small flattened and flattening filter free (FFF) photon beams. METHOD AND MATERIALS: Dose measurements were taken in several organs and structures outside the primary field in an anthropomorphic phantom of a 5 year old child (CIRS) using thermoluminescence dosimetry (LiF:Mg,Cu,P). Out-of-field doses from a medical linear accelerator were assessed for 6 MV flattened and FFF beams of field sizes between 2 × 2 and 10 × 10 cm2. RESULTS: FFF beams resulted in reduced out-of-field doses for all field sizes when compared to flattened beams. Doses for FFF and flattened beams converged for all field sizes at larger distances (>40 cm) from the central axis as leakage becomes the primary source of out-of-field dose. Rotating the collimator to place the MLC bank in the longitudinal axis of the patient was shown to reduce the peripheral doses measured by up to 50% in Varian linear accelerators. CONCLUSION: Minimising out-of-field doses by using FFF beams and aligning the couch and collimator to provide tertiary shielding demonstrated advantages of small field, FFF treatments in a paediatric setting.

A retrospective analysis of setup and intrafraction positional variation in stereotactic radiotherapy treatments.

PURPOSE: The aim of this study was to provide a comprehensive assessment of patient intrafraction motion in linac-based frameless stereotactic radiosurgery (SRS) and radiotherapy (SRT). METHODS: A retrospective review was performed on 101 intracranial SRS/SRT patients immobilized with the Klarity stereotactic thermoplastic mask (compatible with the Brainlab frameless stereotactic system) and aligned on a 6 Degree of Freedom (DoF) couch with the Brainlab ExacTrac image guidance system. Both pretreatment and intrafraction correction data are provided as observed by the ExacTrac system. The effects of couch angle and treatment duration on positioning outcomes are also explored. RESULTS: Initial setup data for patients is shown to vary by up to ±4.18 mm, ±2.97°, but when corrected with a single x-ray image set with ExacTrac, patient positions are corrected to within ±2.11 mm, ±2.27°. Intrafraction patient motion is shown to be uniformly random and independent of both time and couch angle. Patient motion was also limited to within approximately 3 mm, 3° by the thermoplastic mask. CONCLUSIONS: Our results indicate that since patient intrafraction motion is unrelated to couch rotation and treatment duration, intrafraction patient monitoring in 6 DoF is required to minimize intracranial SRS/SRT margins.

Deforming to Best Practice: Key considerations for deformable image registration in radiotherapy.

Image registration is a process that underlies many new techniques in radiation oncology - from multimodal imaging and contour propagation in treatment planning to dose accumulation throughout treatment. Deformable image registration (DIR) is a subset of image registration subject to high levels of complexity in process and validation. A need for local guidance to assist in high-quality utilisation and best practice was identified within the Australian community, leading to collaborative activity and workshops. This report communicates the current limitations and best practice advice from early adopters to help guide those implementing DIR in the clinic at this early stage. They are based on the state of image registration applications in radiotherapy in Australia and New Zealand (ANZ), and consensus discussions made at the 'Deforming to Best Practice' workshops in 2018. The current status of clinical application use cases is presented, including multimodal imaging, automatic segmentation, adaptive radiotherapy, retreatment, dose accumulation and response assessment, along with uptake, accuracy and limitations. Key areas of concern and preliminary suggestions for commissioning, quality assurance, education and training, and the use of automation are also reported. Many questions remain, and the radiotherapy community will benefit from continued research in this area. However, DIR is available to clinics and this report is intended to aid departments using or about to use DIR tools now.

Independent review of 4DCT scans used for SABR treatment planning.

Four-dimensional computerized tomography (4DCT) is required for stereotactic ablative body radiotherapy (SABR) of mobile targets to account for tumor motion during treatment planning and delivery. In this study, we report on the impact of an image review quality assurance process performed prior to treatment planning by medical physicists for 4DCT scans used for SABR treatment. Reviews were performed of 211 4DCT scans (193 patients) over a 3-yr period (October 2014 to October 2017). Treatment sites included lung (n = 168), kidney/adrenal/adrenal gland (n = 12), rib (n = 4), mediastinum (n = 10), liver (n = 2), T-spine (n = 1), and other abdominal sites (n = 14). It was found that in 23% (n = 49) of cases patient management was altered due to the review process. The most frequent intervention involved patient-specific contouring advice (n = 35 cases, 17%) including adjustment of internal target volume (ITV) margins. In 13 cases (6%) a rescan was requested due to extensive motion artifact rendering the scan inadequate for SABR treatment planning. 4DCT review by medical physicists was found to be an effective method to improve plan quality for SABR.

Comparative clinical evaluation of atlas and deep-learning-based auto-segmentation of organ structures in liver cancer.

BACKGROUND: Accurate and standardized descriptions of organs at risk (OARs) are essential in radiation therapy for treatment planning and evaluation. Traditionally, physicians have contoured patient images manually, which, is time-consuming and subject to inter-observer variability. This study aims to a) investigate whether customized, deep-learning-based auto-segmentation could overcome the limitations of manual contouring and b) compare its performance against a typical, atlas-based auto-segmentation method organ structures in liver cancer. METHODS: On-contrast computer tomography image sets of 70 liver cancer patients were used, and four OARs (heart, liver, kidney, and stomach) were manually delineated by three experienced physicians as reference structures. Atlas and deep learning auto-segmentations were respectively performed with MIM Maestro 6.5 (MIM Software Inc., Cleveland, OH) and, with a deep convolution neural network (DCNN). The Hausdorff distance (HD) and, dice similarity coefficient (DSC), volume overlap error (VOE), and relative volume difference (RVD) were used to quantitatively evaluate the four different methods in the case of the reference set of the four OAR structures. RESULTS: The atlas-based method yielded the following average DSC and standard deviation values (SD) for the heart, liver, right kidney, left kidney, and stomach: 0.92 ± 0.04 (DSC ± SD), 0.93 ± 0.02, 0.86 ± 0.07, 0.85 ± 0.11, and 0.60 ± 0.13 respectively. The deep-learning-based method yielded corresponding values for the OARs of 0.94 ± 0.01, 0.93 ± 0.01, 0.88 ± 0.03, 0.86 ± 0.03, and 0.73 ± 0.09. The segmentation results show that the deep learning framework is superior to the atlas-based framwork except in the case of the liver. Specifically, in the case of the stomach, the DSC, VOE, and RVD showed a maximum difference of 21.67, 25.11, 28.80% respectively. CONCLUSIONS: In this study, we demonstrated that a deep learning framework could be used more effectively and efficiently compared to atlas-based auto-segmentation for most OARs in human liver cancer. Extended use of the deep-learning-based framework is anticipated for auto-segmentations of other body sites.