In radiotherapy, automatic organ-at-risk segmentation algorithms allow faster delineation times, but clinically relevant contour evaluation remains challenging. Commonly used measures to assess automatic contours, such as volumetric Dice Similarity Coefficient (DSC) or Hausdorff distance, have shown to be good measures for geometric similarity, but do not always correlate with clinical applicability of the contours, or time needed to adjust them. This study aimed to evaluate the correlation of new and commonly used evaluation measures with time-saving during contouring. Twenty lung cancer patients were used to compare user-adjustments after atlas-based and deep-learning contouring with manual contouring. The absolute time needed (s) of adjusting the auto-contour compared to manual contouring was recorded, from this relative time-saving (%) was calculated. New evaluation measures (surface DSC and added path length, APL) and conventional evaluation measures (volumetric DSC and Hausdorff distance) were correlated with time-recordings and time-savings, quantified with the Pearson correlation coefficient, R. The highest correlation (R = 0.87) was found between APL and absolute adaption time. Lower correlations were found for APL with relative time-saving (R = −0.38), for surface DSC with absolute adaption time (R = −0.69) and relative time-saving (R = 0.57). Volumetric DSC and Hausdorff distance also showed lower correlation coefficients for absolute adaptation time (R = −0.32 and 0.64, respectively) and relative time-saving (R = 0.44 and −0.64, respectively). Surface DSC and APL are better indicators for contour adaptation time and time-saving when using auto-segmentation and provide more clinically relevant and better quantitative measures for automatically-generated contour quality, compared to commonly-used geometry-based measures.
Quality of automatic contouring is generally assessed by comparison with manual delineations, but the effect of contour differences on the resulting dose distribution remains unknown. This study evaluated dosimetric differences between treatment plans optimized using various organ-at-risk (OAR) contouring methods. OARs of twenty lung cancer patients were manually and automatically contoured, after which user-adjustments were made. For each contour set, an automated treatment plan was generated. The dosimetric effect of intra-observer contour variation and the influence of contour variations on treatment plan evaluation and generation were studied using dose-volume histogram (DVH)-parameters for thoracic OARs. Dosimetric effect of intra-observer contour variability was highest for Heart Dmax (3.4 ± 6.8 Gy) and lowest for Lungs-GTV Dmean (0.3 ± 0.4 Gy). The effect of contour variation on treatment plan evaluation was highest for Heart Dmax (6.0 ± 13.4 Gy) and Esophagus Dmax (8.7 ± 17.2 Gy). Dose differences for the various treatment plans, evaluated on the reference (manual) contour, were on average below 1 Gy/1%. For Heart Dmean, higher dose differences were found for overlap with PTV (median 0.2 Gy, 95% 1.7 Gy) vs. no PTV overlap (median 0 Gy, 95% 0.5 Gy). For Dmax-parameters, largest dose difference was found between 0-1 cm distance to PTV (median 1.5 Gy, 95% 4.7 Gy). Dose differences arising from automatic contour variations were of the same magnitude or lower than intra-observer contour variability. For Heart Dmean, we recommend delineation errors to be corrected when the heart overlaps with the PTV. For Dmax-parameters, we recommend checking contours if the distance is close to PTV (<5 cm). For the lungs, only obvious large errors need to be adjusted.
14 May 2022 in Physics and Imaging in Radiation Oncology
User-adjustments after deep-learning (DL) contouring in radiotherapy were evaluated to get insight in real-world editing during clinical practice. This study assessed the amount, type and spatial regions of editing of auto-contouring for organs-at-risk (OARs) in routine clinical workflow for patients in the thorax region. A total of 350 lung cancer and 362 breast cancer patients, contoured between March 2020 and March 2021 using a commercial DL-contouring method followed by manual adjustments were retrospectively analyzed. Subsampling was performed for some OARs, using an inter-slice gap of 1–3 slices. Commonly-used whole-organ contouring assessment measures were calculated, and all cases were registered to a common reference shape per OAR to identify regions of manual adjustment. Results were expressed as the median, 10th-90th percentile of adjustment and visualized using 3D renderings. Per OAR, the median amount of editing was below 1 mm. However, large adjustments were found in some locations for most OARs. In general, enlarging of the auto-contours was needed. Subsampling DL-contours showed less adjustments were made in the interpolated slices compared to simulated no-subsampling for these OARs. The real-world performance of automatic DL-contouring software was evaluated and proven useful in clinical practice. Specific regions-of-adjustment were identified per OAR in the thorax region, and separate models were found to be necessary for specific clinical indications different from training data. This analysis showed the need to perform routine clinical analysis especially when procedures or acquisition protocols change to have the best configuration of the workflow.
Atlas-based and deep-learning contouring (DLC) are methods for automatic segmentation of organs-at-risk (OARs). The European Particle Therapy Network (EPTN) published a consensus-based atlas for delineation of OARs in neuro-oncology. In this study, geometric and dosimetric evaluation of automatically-segmented neuro-oncological OARs was performed using CT- and MR-models following the EPTN-contouring atlas. Image and contouring data from 76 neuro-oncological patients were included. Two atlas-based models (CT-atlas and MR-atlas) and one DLC-model (MR-DLC) were created. Manual contours on registered CT-MR-images were used as ground-truth. Results were analyzed in terms of geometrical (volumetric Dice similarity coefficient (vDSC), surface DSC (sDSC), added path length (APL), and mean slice-wise Hausdorff distance (MSHD)) and dosimetrical accuracy. Distance-to-tumor analysis was performed to analyze to which extent the location of the OAR relative to planning target volume (PTV) has dosimetric impact, using Wilcoxon rank-sum tests. CT-atlas outperformed MR-atlas for 22/26 OARs. MR-DLC outperformed MR-atlas for all OARs. Highest median (95 %CI) vDSC and sDSC were found for the brainstem in MR-DLC: 0.92 (0.88-0.95) and 0.84 (0.77-0.89) respectively, as well as lowest MSHD: 0.27 (0.22-0.39)cm. Median dose differences (ΔD) were within ± 1 Gy for 24/26(92 %) OARs for all three models. Distance-to-tumor showed a significant correlation for ΔDmax,0.03cc-parameters when splitting the data in ≤ 4 cm and > 4 cm OAR-distance (p < 0.001). MR-based DLC and CT-based atlas-contouring enable high-quality segmentation. It was shown that a combination of both CT- and MR-autocontouring models results in the best quality.
Purpose: This study aims to create a deep learning (DL) model capable of accurately delineating the ventricles, and by extension, the periventricular space (PVS), following the 2021 EPTN Neuro-Oncology Atlas guidelines on T1-weighted contrast-enhanced MRI scans (T1CE). The performance of this DL model was quantitatively and qualitatively compared with an off-the-shelf model. Materials and Methods: An nnU-Net was trained for ventricle segmentation using both CT and T1CE MRI images from 78 patients. Its performance was compared to that of a publicly available pretrained segmentation model, SynthSeg. The evaluation was conducted on both internal (N = 18) and external (n = 18) test sets, with each consisting of paired CT and T1CE MRI images and expert-delineated ground truths (GTs). Segmentation accuracy was assessed using the volumetric Dice Similarity Coefficient (DSC), 95th percentile Hausdorff distance (HD95), surface DSC, and added path length (APL). Additionally, a local evaluation of ventricle segmentations quantified differences between manual and automatic segmentations across both test sets. All segmentations were scored by radiotherapy technicians for clinical acceptability using a 4-point Likert scale. Results: The nnU-Net significantly outperformed the SynthSeg model on the internal test dataset in terms of median [range] DSC, 0.93 [0.86–0.95] vs. 0.85 [0.67–0.91], HD95, 0.9 [0.7–2.5] mm vs. 2.2 [1.7–4.8] mm, surface DSC, 0.97 [0.90–0.98] vs. 0.84 [0.70–0.89], and APL, 876 [407–1298] mm vs. 2809 [2311–3622] mm, all with p < 0.001. No significant differences in these metrics were found in the external test set. However clinical ratings favored nnU-Net segmentations on the internal and external test sets. In addition, the nnU-Net had higher clinical ratings than the GT delineation on the internal and external test set. Conclusions: The nnU-Net model outperformed the SynthSeg model on the internal dataset in both segmentation metrics and clinician ratings. While segmentation metrics showed no significant differences between the models on the external set, clinician ratings favored nnU-Net, suggesting enhanced clinical acceptability. This suggests that nnU-Net could contribute to more time-efficient and streamlined radiotherapy planning workflows.
8 February 2025 in Clinical and Translational Radiation Oncology
Assess cognitive changes after radiotherapy (RT) in brain and head-and-neck (HN) cancer patients using patient-reported outcome measures (PROMs) and evaluate a dose–effect relationship for brain structures. Primary brain and HN cancer patients treated with RT between 2012–2021 were included. Patient characteristics, clinical parameters, and PROMs at baseline and 1-year follow-up were collected. Cognitive functioning (CF) from the EORTC QLQ-C30, communication deficit (CD) from the QLQ-BN20, and one cognition-related questions from the EQ6D questionnaire were used, the latter two only for brain patients. Missing data were imputed and the four-point scale scores were transformed to a 100-point scale. Change in scores from baseline to 1-year were categorized into improvement/constant or deterioration. Organs-at-risk (OARs) were contoured either clinically or retrospectively using autocontouring and dose to the OARs were calculated. A total of 110 brain and 356 HN cancer patients were included. Median age was 56 (brain) and 67.5 (HN) years. Baseline and 1-year CF was significantly lower for brain patients (p < 0.001). Univariate analysis for ΔCF showed that age at start RT ≤ 65 years, receiving chemotherapy, higher CF Baseline score, brain mean dose > 3 Gy, and multiple dose levels to left and right hippocampus were statistically associated with cognitive deterioration. Multivariate analysis for ΔCF identified age at RT ≤ 65 years, higher CF Baseline score, and brain mean dose > 3 Gy as significant predictors. This study identified risk factors for subjective cognitive decline and suggests that patients’ self-perceived cognitive deterioration may be related to age, CF baseline score and brain radiation dose above 3 Gy.