Evaluating Variability in Knee CartiGram MRI – A Quantitative Study

(RSNA 2018, Thu Nov 29 2018 11:50AM – 12:00PM ROOM Z15)

PURPOSE

MRI cartilage assessment using CartiGram is a widely used T2 mapping sequence to non-invasively detect changes in cartilage. We tested the robustness of the technique by quantitatively measuring intra- and inter-scanner variability of T2 values.

METHOD AND MATERIALS

Our study had two parts. First, a phantom containing NaCl, GdCl3 and Agarose was created to mimic human meniscus, muscles, cartilage and synovium. A T2 CartiGram (test) and T1w IR-TSE (gold standard) were performed on a 3.0T (750W, GE Healthcare) and 1.5T (HDxt, GE Healthcare) MRI scanners. A phantom integrity test was performed at the end. Obtained data was evaluated by creating T2 maps and calculation of T2 mean±SD. Second, in two healthy volunteers, a T2 CartiGram was performed twice, at an interval of 10 minutes with subject lying still in the scanner, each on both the scanners on the same day. T2 maps were created and mean±SD and Relative Percentage Difference (RPD) calculated. Additional 2-D wear maps were created to check for anatomical variability.

RESULTS

Phantom: 3.0T MRI showed T2 values of 26.5±1.4 in meniscus, 61±3.5 in muscles and 56±2.9, 71.4±3.8 and 78±5.5 in three cartilage samples. The 1.5T MRI showed T2 values of 28±1.2 in meniscus, 62.3±3.4 in muscles and 52.9±2.9, 77.5±5.4 and 89±7.3 in the three cartilage samples.
Healthy Volunteers: The RPDs on the same scanner for subject 1 were 3.5% (on 3.0T) and 3.8% (on 1.5T) on the medial femoral cartilage and -1.7% (on 3.0T) and -1% (on 1.5T) on the lateral, and for subject 2, they were -0.4% (on 3.0T) and 4.8% (on 1.5T) on the medial and 1.2% (on 3.0T) and 5.5% (on 1.5T) on the lateral. The 1.5T scanner reported a lower overall T2 value than the 3.0T, in contrast to the phantom results. Visual inspection of the 2D wear maps by a musculoskeletal radiologist revealed variability of T2 signal with no observable pattern.

CONCLUSION

There exists variability in T2 values of CartiGram when performed in healthy volunteers across both, different time points and different field-strengths. Further studies are needed to re-evaluate the threshold of 40ms for cartilage pathology and define MRI machine-specific guidelines.

CLINICAL RELEVANCE/APPLICATION

Significant differences in T2 values on CartiGram can lead to difficulty in diagnoses of borderline cases in clinical practice.