A Preliminary Study to Use SUVmax of FDG PET-CT as an Identifier of Lesion for Artificial Intelligence

Abstract

Background: Diagnostic reports contribute not only to the particular patient, but also to constructing massive training dataset in the era of artificial intelligence (AI). The maximum standardized uptake value (SUVmax) is often described in daily diagnostic reports of [¹⁸F] fluorodeoxyglucose (FDG) positron emission tomography (PET) - computed tomography (CT). If SUVmax can be used as an identifier of lesion, that would greatly help AI interpret diagnostic reports. We aimed to clarify whether the lesion can be localized using SUVmax strings. Methods: The institutional review board approved this retrospective study. We investigated a total of 112 lesions from 30 FDG PET-CT images acquired with 3 different scanners. SUVmax was calculated from DICOM files based on the latest Quantitative Imaging Biomarkers Alliance (QIBA) publication. The voxels showing the given SUVmax were exhaustively searched in the whole-body images and counted. SUVmax was provided with 5 different degrees of precision: integer (e.g., 3), 1st decimal places (DP) (3.1), 2nd DP (3.14), 3rd DP (3.142), and 4th DP (3.1416). For instance, when SUVmax = 3.14 was given, the voxels with 3.135 ≤ SUVmax < 3.145 were extracted. We also evaluated whether local maximum restriction could improve the identifying performance, where only the voxels showing the highest intensity within some neighborhood were considered. We defined that "identical detection" was achieved when only single voxel satisfied the criterion. Results: A total of 112 lesions from 30 FDG PET-CT images were investigated. SUVmax ranged from 1.3 to 49.1 (median = 5.6). Generally, when larger and more precise SUVmax values were given, fewer voxels satisfied the criterion. The local maximum restriction was very effective. When SUVmax was determined to 4 decimal places (e.g., 3.1416) and the local maximum restriction was applied, identical detection was achieved in 33.3% (lesions with SUVmax < 2), 79.5% (2 ≤ SUVmax < 5), and 97.8% (5 ≤ SUVmax) of lesions. Conclusion: In this preliminary study, SUVmax of FDG PET-CT could be used as an identifier to localize the lesion if precise SUVmax is provided and local maximum restriction was applied, although the lesions showing SUVmax < 2 were difficult to identify. The proposed method may have potential to make use of diagnostic reports retrospectively for constructing training datasets for AI.

Keywords: FDG PET; SUVmax; artificial intelligence; diagnostic report; identifier; maximum of standardized uptake value.

Figure 1

A conceptual image of AI generating the visual summary of the report of FDG PET. SUVmax in the sentence appearing in the report text is used for localization. In this case, the primary lesion (right palatine tonsil) and metastatic nodes show high FDG uptakes. Note that SUVmax should be round before attending physicians read the report.

Figure 2

The findings for a patient who underwent FDG PET-CT for lung nodules. The true SUVmax of the nodule in the left upper lobe was 2.97177 (arrow). When local maximum restriction was not applied, 21031, 2176, 210, 33, and 33 voxels were extracted for 3, 3.0, 2.97, 2.972, and 2.9718, respectively. When 3 × 3 × 3 local maximum restriction was applied, 254, 32, 4, 2, and 2 voxels were extracted. When 5 × 5 × 5 local maximum restriction was applied, 126, 14, 1, 1, and 1 voxel(s) were extracted, achieving identical detection.

Figure 3

The same case as depicted in Figure 2. The true SUVmax of the nodule in the right upper lobe was 1.53924 (arrow). When local maximum restriction was not applied, 74952, 13442, 1427, 198, and 198 voxels were extracted for 2, 1.5, 1.54, 1.539, and 1.5392, respectively. When 3 × 3 × 3 local maximum restriction was applied, 782, 104, 6, 2, and 2 voxels were extracted. When 5 × 5 × 5 local maximum restriction was applied, 410, 60, 4, 2, and 2 voxels were extracted. Thus, identical detection was not achieved for this lesion.

Figure 4

The findings for a patient who underwent FDG PET-CT for a spinal code lesion. The true SUVmax of the nodule in the spinal code lesion was 5.56218 (arrow). When local maximum restriction was not applied, 21116, 1953, 186, 25, and 5 voxels were extracted for 6, 5.6, 5.56, 5.562, and 5.5622, respectively. When 3 × 3 × 3 local maximum restriction was applied, 12, 3, 1, 1, and 1 voxel(s) were extracted. When 5 × 5 × 5 local maximum restriction was applied, 8, 2, 1, 1, and 1 voxel(s) were extracted, achieving identical detection.

Figure 5

The number of voxels extracted by a given SUVmax with various levels of precision. Top row, local maximum restriction was not applied; middle row, 3 × 3 × 3 local maximum restriction was applied; bottom row, 5 × 5 × 5 local maximum restriction was applied.

Figure 6

The overall rate of identical detection of the lesion. DP, decimal places. Free, 3 × 3 × 3, and 5 × 5 × 5 express no restriction and each local maximum restriction.

Figure 7

The results of sub-group analysis of the rate of identical detection of the lesions with SUVmax < 2 (A, N = 6), 2 < SUVmax < 5 (B, N = 44), and 5 < SUVmax (C, N = 62). DP, decimal places. Free, 3 × 3 × 3, and 5 × 5 × 5 express no restriction and each local maximum restriction.

Figure 8

A histogram of SUV over the whole-body image of a patient (semi-log plot).