Article Figures & Data
Figures
- FIGURE 1.
Schematic of proposed DCNN-based AC performed in image space (left), compared with conventional AC performed through system matrix during SPECT image reconstruction (right). BN = batch normalization; conv = convolution; recon = reconstruction; ReLU = rectified linear unit.
- FIGURE 2.
Joint histogram of voxels in the myocardium: SPECTNC vs. SPECTCTAC (A) and SPECTDL vs. SPECTCTAC (B). Counts were log-scaled (i.e., log10(counts)) to visualize small counts in joint histograms.
- FIGURE 3.
Bland–Altman plots for percentage segmental errors across all subjects, male and female subjects, and subjects with HLU and LLU: SPECTNC (A) and SPECTDL (B).
- FIGURE 4.
Polar map (right) and box plots (left) for percentage segmental errors across all subjects.
- FIGURE 5.
Examples of 3 subjects (A, male; B, male; C, male) with smallest absolute mean segmental error in short-axis (SA) and vertical long-axis (VLA) views and polar maps.
- FIGURE 6.
Examples of 3 subjects with 25th (A, male), 50th (B, male), and 75th (C, female) percentiles of absolute mean segmental errors in short-axis (SA) and vertical long-axis (VLA) views and polar maps.
- FIGURE 7.
Examples of 2 subjects with most overestimated (A, male) and most underestimated (B, female) mean segmental errors in short-axis (SA) and vertical long-axis (VLA) views and polar maps.
Tables
- TABLE 1
Average Error, Average Absolute Error, and Correlation Coefficients of All Segments Across All Subjects
Cohort SPECT type Error (%) |Error| (%) Pearson’s correlation coefficient (%) P All (n = 100) NC −6.11 ± 8.06 7.96 ± 6.23 88.05 <0.001 DL 0.49 ± 4.35 3.31 ± 2.87 96.17 <0.001 Male (n = 58) NC −6.96 ± 8.62 8.80 ± 6.72 86.66 <0.001 DL 0.22 ± 4.28 3.19 ± 2.85 96.22 0.10 Female (n = 42) NC −4.95 ± 7.05 6.80 ± 5.29 90.36 <0.001 DL 0.85 ± 4.44 3.46 ± 2.90 96.12 <0.001 HLU (n = 44) NC −7.30 ± 8.28 8.71 ± 6.79 86.72 <0.001 DL −0.21 ± 4.99 3.75 ± 3.30 94.75 0.25 LLU (n = 56) NC −5.18 ± 7.75 7.38 ± 5.70 89.36 <0.001 DL 1.03 ± 3.69 2.96 ± 2.43 97.33 <0.001 Data are mean ± SD.
- TABLE 2
Selected Subjects for Figures 5–7 and Their Average Errors Across 17 Segments for SPECTDL and SPECTNC
Subject no. Figure no. DL error (%) NC error (%) Criteria for DL selection 66 (male, HLU) 5A 0.10 ± 4.08 −12.61 ± 10.04 Smallest |error| 54 (male, HLU) 5B 0.10 ± 2.26 −8.73 ± 8.44 Smallest |error| 93 (male, LLU) 5C −0.10 ± 2.99 −6.32 ± 6.69 Smallest |error| 9 (male, LLU) 6A 1.10 ± 2.46 −9.04 ± 7.49 25th percentile |error| 89 (male, HLL) 6B 1.99 ± 3.05 −4.59 ± 6.85 50th percentile |error| 77 (female, HLL) 6C 3.15 ± 4.07 1.17 ± 5.25 75th percentile |error| 7 (male, LLU) 7A 9.20 ± 4.55 2.65 ± 2.63 Most overestimated error 57 (female, HLU) 7B −7.21 ± 4.85 −12.25 ± 7.25 Most underestimated error Supplemental Figure 1 provides average segmental errors of all subjects through box plots.
Supplemental Data
Files in this Data Supplement:
