Precision of liver and pancreas apparent diffusion coefficients using motion-compensated gradient waveforms in DWI

Abstract

Background: Diffusion weighted imaging (DWI) with optimized motion-compensated gradient waveforms reduces signal dropouts in the liver and pancreas caused by cardiovascular-associated motion, however its precision is unknown. We hypothesized that DWI with motion-compensated DW gradient waveforms would improve apparent diffusion coefficient (ADC)-repeatability and inter-reader reproducibility compared to conventional DWI in these organs.

Methods: In this IRB-approved, prospective, single center study, subjects recruited between October 2019 and March 2020 were scanned twice on a 3 T scanner, with repositioning between test and retest. Each scan included two respiratory-triggered DWI series with comparable acquisition time: 1) conventional (monopolar) 2) motion- compensated diffusion gradients. Three readers measured ADC values. One-way ANOVA, Bland-Altman analysis were used for statistical analysis.

Results: Eight healthy participants (4 male/4 female), with a mean age of 29 ± 4 years, underwent the liver and pancreas MRI protocol. Four patients with liver metastases (2 male/2 female) with a mean age of 58 ± 5 years underwent the liver MRI protocol. In healthy participants, motion-compensated DWI outperformed conventional DWI with mean repeatability coefficient of 0.14 × 10^-3 (CI:0.12-0.17) vs. 0.31 × 10^-3 (CI:0.27-0.37) mm²/s for liver, and 0.11 × 10^-3 (CI:0.08-0.15) vs. 0.34 × 10^-3 (CI:0.27-0.49) mm²/s for pancreas; and with mean reproducibility coefficient of 0.20 × 10^-3 (CI:0.18-0.23) vs. 0.51 × 10^-3 (CI:0.46-0.58) mm²/s for liver, and 0.16 × 10^-3 (CI:0.13-0.20) vs. 0.42 × 10^-3 (CI:0.34-0.52) mm²/s for pancreas. In patients, improved repeatability was observed for motion-compensated DWI in comparison to conventional with repeatability coefficient of 0.51 × 10^{- 3} mm²/s (CI:0.35-0.89) vs. 0.70 × 10^-3 mm²/s (CI:0.49-1.20).

Conclusion: Motion-compensated DWI enhances the precision of ADC measurements in the liver and pancreas compared to conventional DWI.

Keywords: ADC; DWI; Liver; Pancreas; Repeatability.

Figure 1.. Study design.

For each MRI scan, two respiratory-triggered DWI series were acquired: 1) conventional (monopolar) 2) motion-compensated diffusion gradients. Eight healthy participants received a pancreas and liver protocol in two scanning sessions. Patients with liver metastases underwent a liver scan.

Figure 2.. Example DW images and ADC maps of the liver in a healthy participant.

Motion-compensated DWI compared to standard monopolar DWI, was affected less by motion artifacts and inhomogeneities in the liver (arrows).

Figure 3.. Motion-compensated DWI provides improved liver ADC repeatability.

Improved repeatability in the liver was observed for motion-compensated DWI in comparison to conventional monopolar-gradient DWI, by all 3 readers. RC, repeatability coefficient (×10⁻³ mm²/s).

Figure 4.. Motion-compensated DWI provides improved liver ADC reproducibility.

Improved inter-reader reproducibility in the liver was observed for motion-compensated DWI in comparison to conventional monopolar-gradient DWI. The data given represents mean and difference of the two radiologist readers. RDC, reproducibility coefficient (×10⁻³ mm²/s).

Figure 5.. Example DW images and ADC maps of the pancreas in a healthy participant.

Improvement in image quality of pancreas DWI using motion-compensated DW gradient waveform compared to conventional monopolar-gradient waveform DW, with reduction of inhomogeneities (arrows) in the pancreas body.

Figure 6.. Motion-compensated DWI provides improved pancreas ADC repeatability.

Improved test-retest repeatability in the pancreas was observed for motion-compensated DWI in comparison to conventional monopolar-gradient DWI, by all 3 readers. RC, repeatability coefficient (×10⁻³ mm²/s).

Figure 7.. Motion-compensated DWI provides improved pancreas ADC reproducibility.

Improved inter-reader reproducibility in the pancreas was observed for motion-compensated DWI in comparison to conventional monopolar-gradient DWI. The data given represents mean and difference of the two radiologist readers. RDC, reproducibility coefficient (×10⁻³ mm²/s).

Figure 8.. Motion-compensated DWI reduces artifacts and improves repeatability in patients with liver metastases.

Patients with liver metastases. Reduction in artifacts using motion-compensated DWI led to more homogeneous ADC signal of left liver lobe metastatic lesion (marked by arrows) of 65-year-old male patient with gastric adenocarcinoma (a). In addition, an improved repeatability was observed for motion-compensated DWI in patients with metastases in comparison to conventional DWI (b).