Detection and localization of hyperfunctioning parathyroid glands on [18F]fluorocholine PET/ CT using deep learning - model performance and comparison to human experts.

BACKGROUND: In the setting of primary hyperparathyroidism (PHPT), [18F]fluorocholine PET/CT (FCH-PET) has excellent diagnostic performance, with experienced practitioners achieving 97.7% accuracy in localising hyperfunctioning parathyroid tissue (HPTT). Due to the relative triviality of the task for human readers, we explored the performance of deep learning (DL) methods for HPTT detection and localisation on FCH-PET images in the setting of PHPT. PATIENTS AND METHODS: We used a dataset of 93 subjects with PHPT imaged using FCH-PET, of which 74 subjects had visible HPTT while 19 controls had no visible HPTT on FCH-PET. A conventional Resnet10 as well as a novel mPETResnet10 DL model were trained and tested to detect (present, not present) and localise (upper left, lower left, upper right or lower right) HPTT. Our mPETResnet10 architecture also contained a region-of-interest masking algorithm that we evaluated qualitatively in order to try to explain the model's decision process. RESULTS: The models detected the presence of HPTT with an accuracy of 83% and determined the quadrant of HPTT with an accuracy of 74%. The DL methods performed statistically worse (p < 0.001) in both tasks compared to human readers, who localise HPTT with the accuracy of 97.7%. The produced region-of-interest mask, while not showing a consistent added value in the qualitative evaluation of model's decision process, had correctly identified the foreground PET signal. CONCLUSIONS: Our experiment is the first reported use of DL analysis of FCH-PET in PHPT. We have shown that it is possible to utilize DL methods with FCH-PET to detect and localize HPTT. Given our small dataset of 93 subjects, results are nevertheless promising for further research.

Image reconstruction using small-voxel size improves small lesion detection for positron emission tomography.

BACKGROUND: PET/CT imaging is widely used in oncology and provides both metabolic and anatomic information. Because of the relatively poor spatial resolution of PET, the detection of small lesions is limited. The low spatial resolution introduces the partial-volume effect (PVE) which negatively affects images both qualitatively and quantitatively. The aim of the study was to investigate the effect of small-voxel (2 mm in-line pixel size) vs. standard-voxel (4 mm in-line pixel size) reconstruction on lesion detection and image quality in a range of activity ratios. MATERIALS AND METHODS: The National Electrical Manufacturers Association (NEMA) body phantom and the Micro Hollow-Sphere phantom spheres were filled with a solution of [18F]fluorodeoxyglucose ([18F]FDG) in sphere-to-background ratios of 2:1, 3:1, 4:1 and 8:1. In all images reconstructed with 2 mm and 4 mm in-line pixel size the visual lesion delineation, contrast recovery coefficient (CRC) and contrast-to-noise ratio (CNR) were evaluated. RESULTS: For smaller (≤ 13 mm) phantom spheres, significantly higher CRC and CNR using small-voxel reconstructions were found, also improving visual lesion delineation. CRC did not differ significantly for larger (≥ 17 mm) spheres using 2 mm and 4 mm in-line pixel size, but CNR was significantly lower; however, lower CNR did not affect visual lesion delineation. CONCLUSIONS: Small-voxel reconstruction consistently improves precise small lesion delineation, lesion contrast and image quality.