| 1. |
|
|
| 2. |
- Edvardsson, Anneli, et al.
(författare)
-
Robustness and dosimetric verification of hippocampal-sparing craniospinal pencil beam scanning proton plans for pediatric medulloblastoma
- 2024
-
Ingår i: Physics and Imaging in Radiation Oncology. - : Elsevier. - 2405-6316. ; 29
-
Tidskriftsartikel (refereegranskat)abstract
- Background and PurposeHippocampal-sparing (HS) is a method that can potentially reduce late cognitive complications for pediatric medulloblastoma (MB) patients treated with craniospinal proton therapy (PT). The aim of this study was to investigate robustness and dosimetric plan verification of pencil beam scanning HS PT.Materials and MethodsHS and non-HS PT plans for the whole brain part of craniospinal treatment were created for 15 pediatric MB patients. A robust evaluation of the plans was performed. Plans were recalculated in a water phantom and measured field-by-field using an ion chamber detector at depths corresponding to the central part of hippocampi. All HS and non-HS fields were measured with the standard resolution of the detector and in addition 16 HS fields were measured with high resolution. Measured and planned dose distributions were compared using gamma evaluation.ResultsThe median mean hippocampus dose was reduced from 22.9 Gy (RBE) to 8.9 Gy (RBE), while keeping CTV V95% above 95 % for all nominal HS plans. HS plans were relatively robust regarding hippocampus mean dose, however, less robust regarding target coverage and maximum dose compared to non-HS plans. For standard resolution measurements, median pass rates were 99.7 % for HS and 99.5 % for non-HS plans (p < 0.001). For high-resolution measurements, median pass rates were 100 % in the hippocampus region and 98.2 % in the surrounding region.ConclusionsA substantial reduction of dose in the hippocampus region appeared feasible. Dosimetric accuracy of HS plans was comparable to non-HS plans and agreed well with planned dose distribution in the hippocampus region.
|
|
| 3. |
- Gorgisyan, Jenny, et al.
(författare)
-
Evalutation of two commercial deep learning OAR segmentation models for prostate cancer treatment
- 2022
-
Konferensbidrag (refereegranskat)abstract
- Purpose or ObjectiveTo evaluate two commercial, CE labeled deep learning-based models for automatic organs at risk segmentation on planning CT images for prostate cancer radiotherapy. Model evaluation was focused on assessing both geometrical metrics and evaluating a potential time saving.Material and MethodsThe evaluated models consisted of RayStation 10B Deep Learning Segmentation (RaySearch Laboratories AB, Stockholm, Sweden) and MVision AI Segmentation Service (MVision, Helsinki, Finland) and were applied to CT images for a dataset of 54 male pelvis patients. The RaySearch model was re-trained with 44 clinic specific patients (Skåne University Hospital, Lund, Sweden) for the femoral head structures to adjust the model to our specific delineation guidelines. The model was evaluated on 10 patients from the same clinic. Dice similarity coefficient (DSC) and Hausdorff distance (95th percentile) was computed for model evaluation, using an in-house developed Python script. The average time for manual and AI model delineations was recorded.ResultsAverage DSC scores and Hausdorff distances for all patients and both models are presented in Figure 1 and Table 1, respectively. The femoral head segmentations in the re-trained RaySearch model had increased overlap with our clinical data, with a DSC (mean±1 STD) for the right femoral head of 0.55±0.06 (n=53) increasing to 0.91±0.02 (n=10) and mean Hausdorff (mm) decreasing from 55±7 (n=53) to 4±1 (n=10) (similar results for the left femoral head). The deviation in femoral head compared to the RaySearch and MVision original models occurred due to a difference in the femoral head segmentation guideline in the clinic specific data, see Figure 2. Time recording of manual delineation was 13 minutes compared to 0.5 minutes (RaySearch) and 1.4 minutes (MVision) for the AI models, manual correction not included.ConclusionBoth AI models demonstrate good segmentation performance for bladder and rectum. Clinic specific training data (or data that complies to the clinic specific delineation guideline) might be necessary to achieve segmentation results in accordance to the clinical specific standard for some anatomical structures, such as the femoral heads in our case. The time saving was around 90%, not including manual correction.
|
|
| 4. |
- Nenoff, Lena, et al.
(författare)
-
Daily Adaptive Proton Therapy : Is it Appropriate to Use Analytical Dose Calculations for Plan Adaption?
- 2020
-
Ingår i: International Journal of Radiation Oncology Biology Physics. - : Elsevier BV. - 0360-3016. ; 107:4, s. 747-755
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: The accuracy of analytical dose calculations (ADC) and dose uncertainties resulting from anatomical changes are both limiting factors in proton therapy. For the latter, rapid plan adaption is necessary; for the former, Monte Carlo (MC) approaches are increasingly recommended. These, however, are inherently slower than analytical approaches, potentially limiting the ability to rapidly adapt plans. Here, we compare the clinical relevance of uncertainties resulting from both. Methods and Materials: Five patients with non-small cell lung cancer with up to 9 computed tomography (CT) scans acquired during treatment and five paranasal (head and neck) patients with 10 simulated anatomical changes (sinus filling) were analyzed. On the initial planning CT scans, treatment plans were optimized and calculated using an ADC and then recalculated with MC. Additionally, all plans were recalculated (non-adapted) and reoptimized (adapted) on each repeated CT using the same ADC as for the initial plan, and the resulting dose distributions were compared. Results: When comparing analytical and MC calculations in the initial treatment plan and averaged over all patients, 94.2% (non-small cell lung cancer) and 98.5% (head and neck) of voxels had differences <±5%, and only minor differences in clinical target volume (CTV) V95 (average <2%) were observed. In contrast, when recalculating nominal plans on the repeat (anatomically changed) CT scans, CTV V95 degraded by up to 34%. Plan adaption, however, restored CTV V95 differences between adapted and nominal plans to <0.5%. Adapted organ-at-risk doses remained the same or improved. Conclusions: Dose degradations caused by anatomic changes are substantially larger than uncertainties introduced by the use of analytical instead of MC dose calculations. Thus, if the use of analytical calculations can enable more rapid and efficient plan adaption than MC approaches, they can and should be used for plan adaption for these patient groups.
|
|
