All patients provided written, informed consent prior to enrollment, and the study protocol was approved by the Ethics Committee of Juntendo University. A total of 32 patients (female, 15; male, 17) 24 - 82 years old (mean age, 57.7 years, standard deviation [SD] ± 17.8) with single or recurrent episodes of low back pain and leg numbness and tingling, including pain, were examined using an Achieva whole-body 3.0 T MR scanner (Philips Medical Systems, Best, The Netherlands) with a phased array spine coil (Table 1). Subjects were excluded if

they had other intraspinal diseases, such as tumors, a history of lumbar surgery for any disease, or if image quality was unsatisfactory for diffusion metrics calculation [14] [15].

Images were acquired on an Achieva 3.0 T MR scanner (Philips Medical Systems) with the following imaging parameters for DKI: repetition time/echo time, 10,758/88 ms; number of excitations, two; slice thickness/gap, 4/0 mm; number of slices, 32; field of view, 64 × 64 mm; matrix, 128 × 128 reconstructed; imaging time, approximately 13 min; and four b-values (0, 700, 1400, and 2100 s/mm²) with diffusion encoding in six directions for each b-value. The gradient length (δ) was 9.8 ms and the time between the two leading edges of the diffusion gradient (Δ) was 44.1 ms. Image quality was improved by using a reduced field-of-view technique [14] [16] [17]. Prior to DKI, conventional turbo spin echo T1- and T2-weighted sagittal and axial images were obtained. For sagittal images, the imaging parameters were: repetition time/echo time, 400/10 ms for T1WI and 3246/128 ms for T2WI; echo train length, 4 for T1WI and 36 for T2WI; number of excitations, two; slice thickness/gap, 3/0.3 mm; number of slices, 11; field of view, 250 × 250 mm; and matrix, 512 × 512. For axial images, the following imaging parameters were used: repetition time/echo time, 726/10 ms for T1WI and 6196/93 ms for T2WI; echo train length, 5 for T1WI and 36 for T2WI; number of excitations, two; slice thickness/gap, 4/0.4 mm; number of slices, 24; field of view, 160 × 160 mm; and matrix, 512 × 512.

The free software dTV II FZRx and Volume-One 1.72 (Image Computing and Analysis Laboratory, Department of Radiology, The University of Tokyo Hospital, Tokyo, Japan) [18] were used for analyses of DKI on a PC running Windows (Microsoft, Redmond, WA). Disc degeneration was classified as grade I -V on the Pfirrmann classification in the midsagittal section on T2WI in a blinded manner (Figure 1 and Table 2). The MK map was generated using the MK values in the midsagittal section from sagittal sections centered on the lumbar midline region. The mean values in the regions of interest (ROI) were measured in the middle of five equal areas on three slices of the midsagittal section (Figure 2). Analyses were performed for each Pfirrmann grade to determine the MK values in the nucleus pulposus region [19].

The Mann-Whitney U test was used for statistical analyses with IBM SPSS Statistics version 23.0. In all analyses, P < 0.05 was taken to indicate statistical significance.

A highly organized framework of collagen fibrils, representing the main macromolecular component of the intervertebral disc [20], anchors the disc to the bone and provides tensile strength. The highest levels of collagen are seen in the outer annulus of the disc (as much as 70% by dry weight), which decreases to approximately 20% - 30% by dry weight in the nucleus of the adult human lumbar disc [21].

Proteolysis of matrix macromolecules plays a role in intervertebral disc degeneration, with the proteolytic products slowly diffusing out of the disc leading to a loss of matrix integrity, failure of biomechanical load response, and ultimately to the morphological features associated with degeneration. The sequential changes that occur in this process are still not well understood, but the earliest and most marked degenerative change in disc composition is the loss of glycosaminoglycan (GAG) [22], which decreases in parallel with increasing grade of disc degeneration. The loss of GAG results in a drop in swelling pressure [23], loss of hydration, and loss of disc height, with resulting adverse effects on the ability of the disc to respond appropriately to applied biomechanical loads. Disc degeneration also results in disorganization and destruction of the collagen network [24], and is generally classified based on T2WI findings using the system described by Pfirrmann et al. [10] [11] [12]. However, this classification is based on visual evaluation, and therefore it cannot be used to quantify the degree of degeneration. The water and proteoglycan contents within tissues can be quantified by magnetic resonance T2 mapping, and this method can be used to detect early cartilage abnormalities as well as to track responses to therapy [19]. The T2 value tended to decrease with increasing Pfirrmann classification grade, and T2 values differed significantly from grades I to IV, but not grades IV and V [5].

Diffusion-weighted imaging (DWI) [25], including diffusion tensor imaging (DTI) [26], is based on the Gaussian distribution and random movement of water molecules, while DKI [27] visualizes non-Gaussian water diffusion, and has been applied clinically for the characterization of normal and abnormal tissues. In contrast to the mature metrics of DTI, such as fractional anisotropy (FA), DKI and its derived parameters are currently still in the developmental stage [28] [29]. However, a number of reports have indicated important features of DKI for clinical applications, such as its potential usefulness in diagnostic imaging for the evaluation of neurological issues associated with aging [30] as well as various disorders, such as stroke, Alzheimer’s disease, schizophrenia, glioma [31] [32], Parkinson’s disease [33], and attention deficit hyperactivity disorder. There have also been trials to examine the applicability of DKI in the liver [34] and spine [18], and DKI may facilitate estimation of conventional DTI parameters with improved accuracy [35].

Intervertebral disc degeneration is commonly classified according to the method of Pfirrmann et al. [11], which involves determination of the amount of water in the disc based on the T2WI signal intensity, as the water content decreases with both age and disc degeneration [36]. Although T2WI is used clinically for evaluation of disc degeneration [37] [38] [39], it is not possible to determine the absolute signal intensity [40]. In the present study, we performed MK mapping to determine the MK value of the water content in intervertebral discs according to the Pfirrmann classification. Our results indicated that the MK values tended to decrease with increasing Pfirrmann classification grade in the nucleus pulposus, which may reflect reductions of both proteoglycan and water contents [8] [9]. However, there was no significant difference in MK value between grade IV and V of the Pfirrmann classification, and therefore it may be difficult to distinguish between these two grades based on the MK values. In most previous studies, the ROI was selected manually, which has low reproducibility due to the poor differentiation of the central nuclear region from the peripheral annulus. In the present study, the ROI was selected by measuring small equally divided areas. This technique has improved reproducibility compared to the manual approach as well as minimal measurement bias. In the present study, the MK values were shown to decrease with increasing Pfirrmann classification grade in the nucleus pulposus, which probably reflected decreases in both proteoglycan and water contents. Grade IV and V of the Pfirrmann classification increased MK values, since the spatial resolution was not sufficient, it would be difficult to draw an ROI on grade IV and V. Here, we proposed boundaries of the classification based on quantitative evaluation, and therefore our findings suggested that determination of intervertebral disc water content based on the MK value may be useful in clinical research regarding degenerative disc diseases.