Automatic segmentation of prostate cancer metastases in PSMA PET/CT images using deep neural networks with weighted batch-wise dice loss

• Published in: Computers in biology and medicine 2023-05, Vol.158, p.106882-106882, Article 106882
• Main Authors: Xu, Yixi, Klyuzhin, Ivan, Harsini, Sara, Ortiz, Anthony, Zhang, Shun, Bénard, François, Dodhia, Rahul, Uribe, Carlos F., Rahmim, Arman, Lavista Ferres, Juan
• Format: Article
• Language: English
• Subjects: Antigens; Artificial neural networks; Castration; Computed tomography; Data augmentation; Deep learning; Dosimetry; Humans; Image processing; Image segmentation; Lesion detection; Lesions; Male; Medical imaging; Metastases; Metastasis; Model testing; Neural networks; Neural Networks, Computer; Patients; Performance evaluation; Pharmaceuticals; Positron emission; Positron Emission Tomography Computed Tomography - methods; Prostate cancer; Prostatic Neoplasms - diagnostic imaging; PSMA PET/CT imaging; Radioactive tracers; Radiochemistry; Radioisotopes; Radiopharmaceuticals; Three dimensional models; Tumors
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2023.106882

Automatic and accurate segmentation of lesions in images of metastatic castration-resistant prostate cancer has the potential to enable personalized radiopharmaceutical therapy and advanced treatment response monitoring. The aim of this study is to develop a convolutional neural networks-based framework for fully-automated detection and segmentation of metastatic prostate cancer lesions in whole-body PET/CT images. 525 whole-body PET/CT images of patients with metastatic prostate cancer were available for the study, acquired with the [18F]DCFPyL radiotracer that targets prostate-specific membrane antigen (PSMA). U-Net (1)-based convolutional neural networks (CNNs) were trained to identify lesions on paired axial PET/CT slices. Baseline models were trained using batch-wise dice loss, as well as the proposed weighted batch-wise dice loss (wDice), and the lesion detection performance was quantified, with a particular emphasis on lesion size, intensity, and location. We used 418 images for model training, 30 for model validation, and 77 for model testing. In addition, we allowed our model to take n = 0,2, …, 12 neighboring axial slices to examine how incorporating greater amounts of 3D context influences model performance. We selected the optimal number of neighboring axial slices that maximized the detection rate on the 30 validation images, and trained five neural networks with different architectures. Model performance was evaluated using the detection rate, Dice similarity coefficient (DSC) and sensitivity. We found that the proposed wDice loss significantly improved the lesion detection rate, lesion-wise DSC and lesion-wise sensitivity compared to the baseline, with corresponding average increases of 0.07 (p-value = 0.01), 0.03 (p-value = 0.01) and 0.04 (p-value = 0.01), respectively. The inclusion of the first two neighboring axial slices in the input likewise increased the detection rate by 0.17, lesion-wise DSC by 0.05, and lesion-wise mean sensitivity by 0.16. However, there was a minimal effect from including more distant neighboring slices. We ultimately chose to use a number of neighboring slices equal to 2 and the wDice loss function to train our final model. To evaluate the model's performance, we trained three models using identical hyperparameters on three different data splits. The results showed that, on average, the model was able to detect 80% of all testing lesions, with a detection rate of 93% for lesions with maximum standardized uptake