[en] The Pareto front reflects the optimal trade-offs between conflicting objectives and can be used to quantify the effect of different beam configurations on plan robustness and dose-volume histogram parameters. Therefore, our aim was to develop and implement a method to automatically approach the Pareto front in robust intensity-modulated proton therapy (IMPT) planning. Additionally, clinically relevant Pareto fronts based on different beam configurations will be derived and compared to enable beam configuration selection in cervical cancer proton therapy. A method to iteratively approach the Pareto front by automatically generating robustly optimized IMPT plans was developed. To verify plan quality, IMPT plans were evaluated on robustness by simulating range and position errors and recalculating the dose. For five retrospectively selected cervical cancer patients, this method was applied for IMPT plans with three different beam configurations using two, three and four beams. 3D Pareto fronts were optimized on target coverage (CTV D_9_9_%) and OAR doses (rectum V_3_0_G_y; bladder V_4_0_G_y). Per patient, proportions of non-approved IMPT plans were determined and differences between patient-specific Pareto fronts were quantified in terms of CTV D_9_9_%, rectum V_3_0_G_y and bladder V_4_0_G_y to perform beam configuration selection. Per patient and beam configuration, Pareto fronts were successfully sampled based on 200 IMPT plans of which on average 29% were non-approved plans. In all patients, IMPT plans based on the 2-beam set-up were completely dominated by plans with the 3-beam and 4-beam configuration. Compared to the 3-beam set-up, the 4-beam set-up increased the median CTV D_9_9_% on average by 0.2 Gy and decreased the median rectum V_3_0_G_y and median bladder V_4_0_G_y on average by 3.6% and 1.3%, respectively. This study demonstrates a method to automatically derive Pareto fronts in robust IMPT planning. For all patients, the defined four-beam configuration was found optimal in terms of plan robustness, target coverage and OAR sparing. (paper)

[en] Setup and range uncertainties compromise radiotherapy plan robustness. We introduce a method to evaluate the clinical effect of these uncertainties on the population using tumor control probability (TCP) and normal tissue complication probability (NTCP) models.

Eighteen oropharyngeal cancer patients treated with curative intent were retrospectively included. Both photon (VMAT) and proton (IMPT) plans were created using a planning target volume as planning objective. Plans were recalculated for uncertainty scenarios: two for range over/undershoot (IMPT) or CT-density scaling (VMAT), six for shifts. An average shift scenario () was calculated to assess random errors. Dose differences between nominal and scenarios were translated to TCP (2 models) and NTCP (15 models). A weighted average (W_Avg) of the TCP/NTCP based on Gaussian distribution over the variance scenarios was calculated to assess the clinical effect of systematic errors on the population.

TCP/NTCP uncertainties were larger in IMPT compared to VMAT. Although individual perturbations showed risks of plan deterioration, the scenario did not show a substantial decrease in any of the TCP endpoints suggesting evaluated plans in this cohort were robust for random errors. Evaluation of the W_Avg scenario to assess systematic errors showed in VMAT no substantial decrease in TCP endpoints and in IMPT a limited decrease. In IMPT, the W_Avg scenario had a mean TCP loss of 0%–2% depending on plan type and primary or nodal control. The W_Avg for NTCP endpoints was around 0%, except for mandible necrosis in IMPT (W_Avg: 3%).

The estimated population impact of setup and range uncertainties on TCP/NTCP following VMAT or IMPT of oropharyngeal cancer patients was small for both treatment modalities. The use of TCP/NTCP models allows for clinical interpretation of the population effect and could be considered for incorporation in robust evaluation methods.

Highlights:

– TCP/NTCP models allow for a clinical evaluation of uncertainty scenarios.

– For this cohort, in silico-PTV based IMPT plans and VMAT plans were robust for random setup errors.

– Effect of systematic errors on the population was limited: mean TCP loss was 0%–2%. (paper)

[en] Purpose: The aim of this study is to develop and validate a generic method for automatic bladder segmentation on cone beam computed tomography (CBCT), independent of gender and treatment position (prone or supine), using only pretreatment imaging data. Methods: Data of 20 patients, treated for tumors in the pelvic region with the entire bladder visible on CT and CBCT, were divided into four equally sized groups based on gender and treatment position. The full and empty bladder contour, that can be acquired with pretreatment CT imaging, were used to generate a patient-specific bladder shape model. This model was used to guide the segmentation process on CBCT. To obtain the bladder segmentation, the reference bladder contour was deformed iteratively by maximizing the cross-correlation between directional grey value gradients over the reference and CBCT bladder edge. To overcome incorrect segmentations caused by CBCT image artifacts, automatic adaptations were implemented. Moreover, locally incorrect segmentations could be adapted manually. After each adapted segmentation, the bladder shape model was expanded and new shape patterns were calculated for following segmentations. All available CBCTs were used to validate the segmentation algorithm. The bladder segmentations were validated by comparison with the manual delineations and the segmentation performance was quantified using the Dice similarity coefficient (DSC), surface distance error (SDE) and SD of contour-to-contour distances. Also, bladder volumes obtained by manual delineations and segmentations were compared using a Bland-Altman error analysis. Results: The mean DSC, mean SDE, and mean SD of contour-to-contour distances between segmentations and manual delineations were 0.87, 0.27 cm and 0.22 cm (female, prone), 0.85, 0.28 cm and 0.22 cm (female, supine), 0.89, 0.21 cm and 0.17 cm (male, supine) and 0.88, 0.23 cm and 0.17 cm (male, prone), respectively. Manual local adaptations improved the segmentation results significantly (p < 0.01) based on DSC (6.72%) and SD of contour-to-contour distances (0.08 cm) and decreased the 95% confidence intervals of the bladder volume differences. Moreover, expanding the shape model improved the segmentation results significantly (p < 0.01) based on DSC and SD of contour-to-contour distances. Conclusions: This patient-specific shape model based automatic bladder segmentation method on CBCT is accurate and generic. Our segmentation method only needs two pretreatment imaging data sets as prior knowledge, is independent of patient gender and patient treatment position and has the possibility to manually adapt the segmentation locally

[en] To study radiotherapy-related adverse effects, detailed dose information (3D distribution) is needed for accurate dose-effect modeling. For childhood cancer survivors who underwent radiotherapy in the pre-CT era, only 2D radiographs were acquired, thus 3D dose distributions must be reconstructed from limited information. State-of-the-art methods achieve this by using 3D surrogate anatomies. These can however lack personalization and lead to coarse reconstructions. We present and validate a surrogate-free dose reconstruction method based on Machine Learning (ML). Abdominal planning CTs (n = 142) of recently-treated childhood cancer patients were gathered, their organs at risk were segmented, and 300 artificial Wilms’ tumor plans were sampled automatically. Each artificial plan was automatically emulated on the 142 CTs, resulting in 42,600 3D dose distributions from which dose-volume metrics were derived. Anatomical features were extracted from digitally reconstructed radiographs simulated from the CTs to resemble historical radiographs. Further, patient and radiotherapy plan features typically available from historical treatment records were collected. An evolutionary ML algorithm was then used to link features to dose-volume metrics. Besides 5-fold cross validation, a further evaluation was done on an independent dataset of five CTs each associated with two clinical plans. Cross-validation resulted in mean absolute errors ≤ 0.6 Gy for organs completely inside or outside the field. For organs positioned at the edge of the field, mean absolute errors ≤ 1.7 Gy for $D_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}}$, ≤ 2.9 Gy for $D_{2\mathrm{c}\mathrm{c}}$, and ≤ 13% for $V_{5\mathrm{G}\mathrm{y}}$ and $V_{10\mathrm{G}\mathrm{y}}$, were obtained, without systematic bias. Similar results were found for the independent dataset. To conclude, we proposed a novel organ dose reconstruction method that uses ML models to predict dose-volume metric values given patient and plan features. Our approach is not only accurate, but also efficient, as the setup of a surrogate is no longer needed. (paper)

[en] Purpose: To develop a delineation tool that refines physician-drawn contours of the gross tumor volume (GTV) in nasopharynx cancer, using combined pixel value information from x-ray computed tomography (CT) and magnetic resonance imaging (MRI) during delineation. Methods: Operator-guided delineation assisted by a so-called ''snake'' algorithm was applied on weighted CT-MRI registered images. The physician delineates a rough tumor contour that is continuously adjusted by the snake algorithm using the underlying image characteristics. The algorithm was evaluated on five nasopharyngeal cancer patients. Different linear weightings CT and MRI were tested as input for the snake algorithm and compared according to contrast and tumor to noise ratio (TNR). The semi-automatic delineation was compared with manual contouring by seven experienced radiation oncologists. Results: A good compromise for TNR and contrast was obtained by weighing CT twice as strong as MRI. The new algorithm did not notably reduce interobserver variability, it did however, reduce the average delineation time by 6 min per case. Conclusions: The authors developed a user-driven tool for delineation and correction based a snake algorithm and registered weighted CT image and MRI. The algorithm adds morphological information from CT during the delineation on MRI and accelerates the delineation task.