Assessment of 18F-DCFPyL PSMA PET/CT and PET/MR quantitative parameters for reference standard organs: Inter-reader, inter-modality, and inter-patient variability.

Prostate specific membrane antigen (PSMA)-based radiotracers have shown promise for prostate cancer assessment. Evaluation of quantitative variability and establishment of reference standards are important for optimal clinical and research utility. This work evaluates the variability of PSMA-based [18F]DCFPyL (PyL) PET quantitative reference standards. Consecutive eligible patients with biochemically recurrent prostate cancer were recruited for study participation from August 2016-October 2017. After PyL tracer injection, whole body PET/CT (wbPET/CT) was obtained with subsequent whole body PET/MR (wbPET/MR). Two readers independently created regions of interest (ROIs) including a 40% standardized uptake value (SUV) threshold ROI of the whole right parotid gland and separate spherical ROIs in the superior, mid, and inferior gland. Additional liver (right lobe) and blood pool spherical ROIs were defined. Bland-Altman analysis, including limits of agreement (LOA), as well as interquartile range (IQR) and coefficient of variance (CoV) was used. Twelve patients with prostate cancer were recruited (mean age, 61.8 yrs; range 54-72 years). One patient did not have wbPET/MR and was excluded. There was minimal inter-reader SUVmean variability (bias±LOA) for blood pool (-0.13±0.42; 0.01±0.41), liver (-0.55±0.82; -0.22±1.3), or whole parotid gland (-0.05±0.31; 0.08±0.24) for wbPET/CT and wbPET/MR, respectively. Greater inter-reader variability for the 1-cm parotid gland ROIs was present, for both wbPET/CT and wbPET/MR. Comparing wbPET/CT to the subsequently acquired wbPET/MR, blood pool had a slight decrease in SUVmean. The liver as well as parotid gland showed a slight increase in activity although the absolute bias only ranged from 0.45-1.28. The magnitude of inter-subject variability was higher for the parotid gland regardless of modality or reader. In conclusion, liver, blood pool, and whole parotid gland quantitation show promise as reliable reference normal organs for clinical/research PET applications. Variability with 1-cm parotid ROIs may limit its use.

Comparison of 11 automated PET segmentation methods in lymphoma.

Segmentation of lymphoma lesions in FDG PET/CT images is critical in both assessing individual lesions and quantifying patient disease burden. Simple thresholding methods remain common despite the large heterogeneity in lymphoma lesion location, size, and contrast. Here, we assess 11 automated PET segmentation methods for their use in two scenarios: individual lesion segmentation and patient-level disease quantification in lymphoma. Lesions on 18F-FDG PET/CT scans of 90 lymphoma patients were contoured by a nuclear medicine physician. Thresholding, active contours, clustering, adaptive region-growing, and convolutional neural network (CNN) methods were implemented on all physician-identified lesions. Lesion-level segmentation was evaluated using multiple segmentation performance metrics (Dice, Hausdorff Distance). Patient-level quantification of total disease burden (SUVtotal) and metabolic tumor volume (MTV) was assessed using Spearman's correlation coefficients between the segmentation output and physician contours. Lesion segmentation and patient quantification performance was compared to inter-physician agreement in a subset of 20 patients segmented by a second nuclear medicine physician. In total, 1223 lesions with median tumor-to-background ratio of 4.0 and volume of 1.8 cm3, were evaluated. When assessed for lesion segmentation, a 3D CNN, DeepMedic, achieved the highest performance across all evaluation metrics. DeepMedic, clustering methods, and an iterative threshold method had lesion-level segmentation performance comparable to the degree of inter-physician agreement. For patient-level SUVtotal and MTV quantification, all methods except 40% and 50% SUVmax and adaptive region-growing achieved a performance that was similar the agreement of the two physicians. Multiple methods, including a 3D CNN, clustering, and an iterative threshold method, achieved both good lesion-level segmentation and patient-level quantification performance in a population of 90 lymphoma patients. These methods are thus recommended over thresholding methods such as 40% and 50% SUVmax, which were consistently found to be significantly outside the limits defined by inter-physician agreement.

Convolutional Neural Networks for Automated PET/CT Detection of Diseased Lymph Node Burden in Patients with Lymphoma.

PURPOSE: To automatically detect lymph nodes involved in lymphoma on fluorine 18 (18F) fluorodeoxyglucose (FDG) PET/CT images using convolutional neural networks (CNNs). MATERIALS AND METHODS: In this retrospective study, baseline disease of 90 patients with lymphoma was segmented on 18F-FDG PET/CT images (acquired between 2005 and 2011) by a nuclear medicine physician. An ensemble of three-dimensional patch-based, multiresolution pathway CNNs was trained using fivefold cross-validation. Performance was assessed using the true-positive rate (TPR) and number of false-positive (FP) findings. CNN performance was compared with agreement between physicians by comparing the annotations of a second nuclear medicine physician to the first reader in 20 of the patients. Patient TPR was compared using Wilcoxon signed rank tests. RESULTS: Across all 90 patients, a range of 0-61 nodes per patient was detected. At an average of four FP findings per patient, the method achieved a TPR of 85% (923 of 1087 nodes). Performance varied widely across patients (TPR range, 33%-100%; FP range, 0-21 findings). In the 20 patients labeled by both physicians, a range of 1-49 nodes per patient was detected and labeled. The second reader identified 96% (210 of 219) of nodes with an additional 3.7 per patient compared with the first reader. In the same 20 patients, the CNN achieved a 90% (197 of 219) TPR at 3.7 FP findings per patient. CONCLUSION: An ensemble of three-dimensional CNNs detected lymph nodes at a performance nearly comparable to differences between two physicians' annotations. This preliminary study is a first step toward automated PET/CT assessment for lymphoma.© RSNA, 2020.