Atlas-based automatic segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed as a way to improve the accuracy and execution time of segmentation, assuming that, the more similar the atlas is to the patient, the better the results will be. This paper presents an analysis of atlas selection methods in the context of radiotherapy treatment planning. For a range of commonly contoured OARs, a thorough comparison of a large class of typical atlas selection methods has been performed. For this evaluation, clinically contoured CT images of the head and neck (N = 316) and thorax (N = 280) were used. The state-of-the-art intensity and deformation similarity-based atlas selection methods were found to compare poorly to perfect atlas selection. Counter-intuitively, atlas selection methods based on a fixed set of representative atlases outperformed atlas selection methods based on the patient image. This study suggests that atlas-based segmentation with currently available selection methods compares poorly to the potential best performance, hampering the clinical utility of atlas-based segmentation. Effective atlas selection remains an open challenge in atlas-based segmentation for radiotherapy planning.

Segmentation can degrade a high-quality dose distribution obtained by fluence map optimisation (FMO). A novel algorithm is proposed for generation of MLC segments to deliver an FMO plan with step-and-shoot IMRT while minimising quality loss. All beams are considered simultaneously while generating MLC segments for reproducing the 3-dimensional FMO dose distribution. Segment generation is only steered by the 3D FMO dose distribution, i.e. underlying FMO fluence profiles are not considered. The algorithm features prioritised generation of segments, focusing on accurate reproduction of clinical objectives with the highest priorities. The performance of the segmentation algorithm was evaluated for 20 prostate patients, 15 head-and-neck patients, and 12 liver patients. FMO dose distributions were generated by automated multi-criteria treatment planning (Pareto-optimal plans) and subsequently segmented using the proposed method. Various segmentation strategies were investigated regarding prioritisation of objectives and limitation of the number of segments. Segmented plans were dosimetrically similar to FMO plans and for all patients a clinically acceptable segmented plan could be generated. Substantial differences between FMO and segmented fluence profiles were observed. Avoidance of the usual reconstruction of 2D FMO fluence profiles for segment generation, and instead simultaneously generating segments for all beams to directly reproduce the 3D FMO dose distribution is a likely explanation for the obtained results. For the strategies of limiting the number of segments large reductions in number of segments were observed with minimal impact on plan quality.

Atlas-based segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed to improve the performance of segmentation, assuming that the more similar the atlas is to the patient the better the result. It follows that the larger the database of atlases from which to select, the better the results should be. This work seeks to estimate a clinically achievable expected performance under this assumption. Assuming a perfect atlas selection, Extreme Value Theory has been applied to estimate accuracy of single-atlas and multi-atlas segmentation given a large database of atlases. For this purpose, clinical contours of most common OARs on CT of the head and neck (N=316) and thoracic (N=280) cases were used. This study found that while for most organs "perfect" segmentation cannot be reasonably expected, auto-contouring performance of a level corresponding to clinical quality could be consistently expected given a database of 5000 atlases under the assumption of perfect atlas selection.

Radiotherapy treatment planning requires accurate modeling of the delivered patient dose, including radiation scatter effects, multi-leaf collimator (MLC) leaf transmission, interleaf-leakage, etc. In fluence map optimization (FMO), a simple dose model is used to first generate an intermediate plan based on pencil-beams. In a second step (segmentation phase), this intermediate plan is then converted into a deliverable treatment plan with MLC segments. In this paper, we investigate novel approaches for the use of a clinical dose engine (CDE) for segmentation of FMO plans in robotic radiotherapy. Segments are sequentially added to the plan. Generation of each next segment is based on the total 3D dose distribution, resulting from already selected segments and the desired FMO dose, considering all treatment beams as candidates for delivery of the new segment. Three versions of the segmentation algorithm were investigated with differences in the integration of the CDE. The combined use of pencil-beams and segments in a segmentation method is non-trivial. Therefore, new methods were developed for the use of segment doses calculated with the CDE in combination with pencil-beams, used for the selection of new segments. For 20 patients with prostate cancer and 12 with liver cancer, segmented plans were compared with FMO plans. All three versions of the proposed segmentation algorithm could well mimic FMO dose distributions. Segmentation with a fully integrated CDE provided the best plan quality and lowest numbers of monitor units and segments at the cost of increased calculation time.

To propose and validate a fully automated multi-criterial treatment planning solution for a CyberKnife ® equipped with an InCise™ 2 multi-leaf collimator. Methods: The AUTO BAO plans are generated with fully-automated prioritized multi-criterial optimization (AUTO MCO) of pencil-beam fluence maps with integrated non-coplanar beam angle optimization (BAO), followed by MLC segment generation. Both the AUTO MCO and segmentation algorithms have been in-house developed. AUTO MCO generates for each patient a single, high-quality Pareto-optimal IMRT plan. The segmentation algorithm then accurately mimics the AUTO MCO 3D dose distribution, while considering all candidate beams simultaneously, rather than replicating the fluence maps. Pencil-beams, segment dose depositions, and final dose calculations are performed with a stand-alone version of the clinical dose calculation engine. For validation, AUTO BAO plans were generated for 33 prostate SBRT patients and compared to reference plans (REF) that were manually generated with the commercial treatment planning system (TPS), in absence of time pressure. REF plans were also compared to AUTO RB plans, for which fluence map optimization was performed for the beam angle configuration used in the REF plan, and the segmentation could use all these beams or only a subset, depending on the dosimetry. Results: AUTO BAO plans were clinically acceptable and dosimetrically similar to REF plans, but had on average reduced numbers of beams ((beams in AUTO BAO)/(beams in REF) (relative improvement): 24.7/48.3 (-49%)), segments (59.5/98.9 (-40%)), and delivery times (17.1/22.3 min. (-23%)). Dosimetry of AUTO RB and REF were also similar, but AUTO RB used on average fewer beams (38.0/48.3 (-21%)) and had on average shorter delivery times (18.6/22.3 min. (-17%)). Delivered Monitor Units (MU) were similar for all three planning approaches. Conclusions: A new, vendor-independent optimization workflow for fully automated generation of deliverable high-quality CyberKnife ® plans was proposed, including BAO. Compared to manual planning with the commercial TPS, fraction delivery times i Accepted ArticleThis article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as