With axial loading during MRI diurnal T2-value changes in lumbar discs are neglectable: a cross sectional study

In this cross sectional study six healthy volunteers were examined with alMRI, at three time points during 1 day, with 30 lumbar IVDs analyzed. None of the volunteers suffered from any known medical history of back pain or spine related disease. The inclusion of the volunteers aimed to largely reflect a LBP cohort in terms of age and IVD degeneration (four men, mean age: 38y, age range 27-63y and two women, mean age: 37y, age 28–46y), with all Pfirrmann grades represented.

Each volunteer was investigated at three different sessions during the same day. Every session consisted of one MRI scan with the spine in an unloaded position (referred to as uMRI) instantly followed by one scan with alMRI. The first session was at 07 am, the second at 11.30 am and the last at 04 pm. No restrictions regarding activity were imposed before or between sessions. The alMRI was carried out in supine position with a loading device (DynaWell compression device, Dynawell diagnostics AB, Las Vegas, NV, USA), set at half the body weight of the actual volunteer. During uMRI, the lower leg of the volunteer rested on a foam support of 10 cm height, resulting in slight bending of the hip and knee joints. During alMRI, a cushion of approximately 8 cm height was used between the bed and the lumbar back, creating a lumbar lordosis in order to simulate normal upright position [23].

The MR-images were acquired with a Siemens Magnetom Aera 1,5 T scanner (Erlangen, Germany). The scan protocol consisted of a sagittal turbo spin echo (TSE) T1 weighted (w) sequence, a sagittal TSE T2w sequence and an axial TSE T2w sequence. A sagittal multi echo spin echo (SE) T2-sequence was incorporated at the end of the protocol enabling sagittal T2-maps to be calculated. High quality standardized T2-maps (256 × 256 matrix, slice thickness 3.5 mm, FOV 220 × 220 mm ² , NEX: 1) were finally reconstructed.

All image analysis were made by a radiology resident, supervised by an experienced radiologist. Degeneration of the IVDs were graded according to the Pfirrmann classification [24]. To enable further image analysis, multiplanar reconstructions were made from the T1w-sequences. 10 mm thick (in order to enable volumetric IVD analysis) midsagittal slices of sagittal TSE T1w-sequences were fused with sagittal SE T2-mapping sequences using software from syngovia (Siemens, Erlangen Germany).

The entire IVD was delineated in the sagittal plane using the polygonal region of interest (ROI) tool. Craniocaudally, the IVD was delineated by the endplates and anterioposteriorly by the anterior longitudinal ligament and the posterior longitudinal ligament. The entire IVD was further divided into five volumetric sub-regions with equal length in the sagittal plane, referred to as ROI1 to ROI5. Thus, ROI1 corresponded to the anterior parts of the IVD, ROI5 the posterior IVD parts and ROI2–4 the regions in between (Fig. 1).

The mean T2-time and standard deviation of the mean T2-time were assessed for the entire IVDs and for each sub-region, on both alMRI and uMRI. To ensure reliable measurements, blinded inter-rater agreement was performed before initiating the study. This was done on 5 clinical LBP patients, in which both uMRI and alMRI had been performed. 36 IVDs were measured with approximately half evaluated at alMRI. Inter-rater agreement was substantial for ROI1 and almost perfect regarding ROI2–5 (Table 1) [25, 26]. To test the robustness of the scanning, three repeated scans with alMRI and three repeated scans with uMRI were made on one volunteer at the same session.

The alMRI T2-value slightly decreased over the day in ROI2 to ROI4, from 102 ms (ROI2), 128 ms (ROI3) and 118 ms (ROI4) in the morning to 94 ms, 121 ms and 111 ms, respectively in the afternoon (Fig. 2/Table 2). Thus, small alterations in absolute numbers of alMRI T2-values over the day were discerned. However, no significant changes were found (Table 2), neither for the entire IVDs nor for the separate sub-regions. For all comparative analyzes, the p-values were larger than 0.2.

Neither for the entire IVDs nor for the separate sub-regions were any significant diurnal alterations found with uMRI. P-values larger than 0.2 were found for all comparative analyzes (Fig. 3).

When IVDs were compared between alMRI and uMRI, alterations in T2-values were seen for every sub-region both in the morning, at lunchtime and in the afternoon session. For example, mean T2-value for uMRI in ROI1 was 30 ms (SD 11 ms) compared to 35 ms (SD 12 ms) with alMRI, corresponding to an increase of 17%. Likewise, a decrease of 20% in the mean T2-value was seen in ROI5, where the mean T2-value was 59 ms (SD 19 ms) for uMRI and 47 ms (SD 14 ms) for alMRI.

Among the volunteers investigated IVDs with all Pfirrmann grades were represented (Fig. 4), including one IVD with Pfirrmann grade 5. This IVD, localized at L5-S1, was so severely degenerated that only the most 3 ventral ROIs were enabled to be outlined. Most of the IVDs were Pfirrmann grade 2 (Fig. 4). Large diversity of T2-values were detected between all the investigated IVDs. At the morning session, the alMRI T2-values of the entire IVD varied between 38 and 138 ms, at lunch between 33 and 143 ms, whereas the afternoon T2-values varied between 31 to 147 ms. Thus, the difference between the highest and lowest T2-values of the various IVDs were over 300% (Fig. 4). Degenerated IVDs had, in general, lower T2-values. For example, the highest T2-value assessed in an IVD with Pfirrmann grade 5 was 39 ms while the highest value for the entire IVD with Pfirrmann grade 1 was 147 ms. Large regional variations were further seen in non-degenerated IVDs. In a Pfirrmann grade 1 disc for example, the mean alMRI T2-value in the morning was 35 ms in ROI1, 102 ms in ROI2, 128 ms in ROI3, 118 ms in ROI4 and 47 ms in ROI5. In degenerated IVDs, the regional variations were much smaller. In the IVD with Pfirrmann grade 5, the highest T2 value found was 52 ms (ROI3) while the lowest T2 value found was 32 ms (ROI2). When the alMRI T2-values were correlated to Pfirrmann grade, however, no significant changes over the day could be revealed neither on entire IVD level nor within the various sub-regions.

Repetitive scans of the same volunteer showed robustness of the scanning with only minor T2-value variations, however, a 10% alteration was seen at one occasion with alMRI at level L5-S1 between scan 1 (36 ms, SD 14 ms) and scan 2 (40 ms, SD 15 ms). Intra-rater measurements are displayed in Table 1.

Only a small number of volunteers were included in the present study. Since there is a lack of similar studies, in which quantitative T2-mapping measurements are performed with alMRI, estimation of the sample size was based on expected changes in regional IVD T2-values over the day for unloaded MRI [19, 25]. With a diurnal effect of 10%, a sample size of 5 individuals and 25 IVDs should be sufficient at 80% power with 95% significance and two sided test. One extra healthy volunteer was included in the study to ensure power. With alMRI, however, the effect size was found to be much smaller and, hence, no statistical significance could be verified. Significant or not, the diurnal changes during alMRI were small and thus likely not clinically relevant. Hence, the alMRI method can be considered as robust over the day. Moreover, the included volunteers were well representative for a LBP cohort, with all degenerations grades represented, and thus these data likely are more clinically relevant compared to results from study cohorts with younger mean age and less degenerated IVDs.

With alMRI, only minor diurnal T2-value changes were found in the lumbar discs. Nonsignificant and neglectable diurnal changes are advantageous both for research purposes, as well as in the clinical setting, giving comparable and robust data regardless of at what time-point the alMRI is performed.