The impact of multicentric datasets for the automated tumor delineation in primary prostate cancer using convolutional neural networks on 18F-PSMA-1007 PET.

PURPOSE: Convolutional Neural Networks (CNNs) have emerged as transformative tools in the field of radiation oncology, significantly advancing the precision of contouring practices. However, the adaptability of these algorithms across diverse scanners, institutions, and imaging protocols remains a considerable obstacle. This study aims to investigate the effects of incorporating institution-specific datasets into the training regimen of CNNs to assess their generalization ability in real-world clinical environments. Focusing on a data-centric analysis, the influence of varying multi- and single center training approaches on algorithm performance is conducted. METHODS: nnU-Net is trained using a dataset comprising 161 18F-PSMA-1007 PET images collected from four distinct institutions (Freiburg: n = 96, Munich: n = 19, Cyprus: n = 32, Dresden: n = 14). The dataset is partitioned such that data from each center are systematically excluded from training and used solely for testing to assess the model's generalizability and adaptability to data from unfamiliar sources. Performance is compared through a 5-Fold Cross-Validation, providing a detailed comparison between models trained on datasets from single centers to those trained on aggregated multi-center datasets. Dice Similarity Score, Hausdorff distance and volumetric analysis are used as primary evaluation metrics. RESULTS: The mixed training approach yielded a median DSC of 0.76 (IQR: 0.64-0.84) in a five-fold cross-validation, showing no significant differences (p = 0.18) compared to models trained with data exclusion from each center, which performed with a median DSC of 0.74 (IQR: 0.56-0.86). Significant performance improvements regarding multi-center training were observed for the Dresden cohort (multi-center median DSC 0.71, IQR: 0.58-0.80 vs. single-center 0.68, IQR: 0.50-0.80, p < 0.001) and Cyprus cohort (multi-center 0.74, IQR: 0.62-0.83 vs. single-center 0.72, IQR: 0.54-0.82, p < 0.01). While Munich and Freiburg also showed performance improvements with multi-center training, results showed no statistical significance (Munich: multi-center DSC 0.74, IQR: 0.60-0.80 vs. single-center 0.72, IQR: 0.59-0.82, p > 0.05; Freiburg: multi-center 0.78, IQR: 0.53-0.87 vs. single-center 0.71, IQR: 0.53-0.83, p = 0.23). CONCLUSION: CNNs trained for auto contouring intraprostatic GTV in 18F-PSMA-1007 PET on a diverse dataset from multiple centers mostly generalize well to unseen data from other centers. Training on a multicentric dataset can improve performance compared to training exclusively with a single-center dataset regarding intraprostatic 18F-PSMA-1007 PET GTV segmentation. The segmentation performance of the same CNN can vary depending on the dataset employed for training and testing.

Deep learning based automated delineation of the intraprostatic gross tumour volume in PSMA-PET for patients with primary prostate cancer.

PURPOSE: With the increased use of focal radiation dose escalation for primary prostate cancer (PCa), accurate delineation of gross tumor volume (GTV) in prostate-specific membrane antigen PET (PSMA-PET) becomes crucial. Manual approaches are time-consuming and observer dependent. The purpose of this study was to create a deep learning model for the accurate delineation of the intraprostatic GTV in PSMA-PET. METHODS: A 3D U-Net was trained on 128 different 18F-PSMA-1007 PET images from three different institutions. Testing was done on 52 patients including one independent internal cohort (Freiburg: n = 19) and three independent external cohorts (Dresden: n = 14 18F-PSMA-1007, Boston: Massachusetts General Hospital (MGH): n = 9 18F-DCFPyL-PSMA and Dana-Farber Cancer Institute (DFCI): n = 10 68Ga-PSMA-11). Expert contours were generated in consensus using a validated technique. CNN predictions were compared to expert contours using Dice similarity coefficient (DSC). Co-registered whole-mount histology was used for the internal testing cohort to assess sensitivity/specificity. RESULTS: Median DSCs were Freiburg: 0.82 (IQR: 0.73-0.88), Dresden: 0.71 (IQR: 0.53-0.75), MGH: 0.80 (IQR: 0.64-0.83) and DFCI: 0.80 (IQR: 0.67-0.84), respectively. Median sensitivity for CNN and expert contours were 0.88 (IQR: 0.68-0.97) and 0.85 (IQR: 0.75-0.88) (p = 0.40), respectively. GTV volumes did not differ significantly (p > 0.1 for all comparisons). Median specificity of 0.83 (IQR: 0.57-0.97) and 0.88 (IQR: 0.69-0.98) were observed for CNN and expert contours (p = 0.014), respectively. CNN prediction took 3.81 seconds on average per patient. CONCLUSION: The CNN was trained and tested on internal and external datasets as well as histopathology reference, achieving a fast GTV segmentation for three PSMA-PET tracers with high diagnostic accuracy comparable to manual experts.

Intraindividual Comparison Between [18F] PSMA-1007 PET/CT and Multiparametric MRI for Radiotherapy Planning in Primary Prostate Cancer Patients.

Introduction: Accurate detection and segmentation of the intraprostatic gross tumor volume (GTV) is pivotal for radiotherapy (RT) in primary prostate cancer (PCa) since it influences focal therapy target volumes and the patients' cT stage. The study aimed to compare the performance of multiparametric resonance imaging (mpMRI) with [18F] PSMA-1007 positron emission tomography (PET) for intraprostatic GTV detection as well as delineation and to evaluate their respective influence on RT concepts. Materials and Methods: In total, 93 patients from two German University Hospitals with [18F] PSMA-1007-PET/CT and MRI (Freiburg) or [18F] PSMA-1007-PET/MRI (Dresden) were retrospectively enrolled. Validated contouring techniques were applied for GTV-PET and -MRI segmentation. Absolute tumor volume and cT status were determined for each imaging method. The PCa distribution from histopathological reports based on biopsy cores and surgery specimen was used as reference in terms of laterality (unilateral vs. bilateral). Results: In the Freiburg cohort (n = 84), mpMRI and PET detected in median 2 (range: 1-5) and 3 (range: 1-8) GTVs, respectively (p < 0.01). The median GTV-MRI was significantly smaller than the GTV-PET, measuring 2.05 vs. 3.65 ml (p = 0.0005). PET had a statistically significant higher concordance in laterality with surgery specimen compared to mpMRI (p = 0.04) and biopsy (p < 0.01), respectively. PSMA PET led to more cT2c and cT3b stages, whereas cT3a stage was more pronounced in mpMRI. Based on the cT stage derived from mpMRI and PET information, 21 and 23 as well as 59 and 60 patients, respectively, were intermediate- and high-risk according to the National Comprehensive Cancer Network (NCCN) v1.2022 criteria. In the Dresden cohort (n = 9), similar results were observed. Conclusion: Intraprostatic GTV segmentation based on [18F] PSMA-1007 PET results in more and larger GTVs compared to mpMRI. This influences focal RT target volumes and cT stage definition, but not the NCCN risk group.