Here, we aimed to develop and validate a fully automated artificial intelligence (AI)-based method for the detection and quantification of suspected prostate tumour/local recurrence, lymph node metastases, and bone metastases from [¹⁸F]PSMA-1007 positron emission tomography-computed tomography (PET-CT) images. Images from 660 patients were included. Segmentations by one expert reader were ground truth. A convolutional neural network (CNN) was developed and trained on a training set, and the performance was tested on a separate test set of 120 patients. The AI method was compared with manual segmentations performed by several nuclear medicine physicians. Assessment of tumour burden (total lesion volume (TLV) and total lesion uptake (TLU)) was performed. The sensitivity of the AI method was, on average, 79% for detecting prostate tumour/recurrence, 79% for lymph node metastases, and 62% for bone metastases. On average, nuclear medicine physicians’ corresponding sensitivities were 78%, 78%, and 59%, respectively. The correlations of TLV and TLU between AI and nuclear medicine physicians were all statistically significant and ranged from R = 0.53 to R = 0.83. In conclusion, the development of an AI-based method for prostate cancer detection with sensitivity on par with nuclear medicine physicians was possible. The developed AI tool is freely available for researchers.

Background: The aim was to evaluate whether an artificial intelligence (AI)-based tool for the automated quantification of the total metabolic tumour volume (tMTV) in patients with Hodgkin lymphoma (HL) could support nuclear medicine specialists in lesion segmentation and thereby enhance inter-observer agreement. Methods: Forty-eight consecutive patients who underwent staging with [18F]FDG PET/CT were included. Eight invited specialists from different hospitals were asked to manually segment lesions for tMTV calculations in 12 cases without AI advice, and to use automated AI segmentation in a further 12 cases, with editing as required, i.e., segmenting/adjusting 24 cases each. Each case was segmented by two specialists manually and by two different specialists using the AI tool, allowing for the pairwise comparison of inter-observer variability. Results: The median difference between two specialists performing manual tMTV segmentations was 26 cm³ (IQR 10–86 cm³) corresponding to 23% (IQR 7–50%) of the median tMTV in the dataset, while the median difference between two specialists tMTV adjustments using AI segmentations was 12 cm³ (IQR 4–39 cm³) corresponding to 9% (IQR 2–21%) (p = 0.023). The median difference in tMTV between measurements with and without AI was 3.3 cm³, corresponding to 2.3% of the median tMTV. Conclusions: An automated AI-based tool can significantly increase agreement among specialists quantifying tMTV in HL patients staged with [18F]FDG PET/CT, without markedly changing the measurements.