A novel MRI diffusion metric 'slow diffusion coefficient' (SDC) for diagnosing isocitrate dehydrogenase (IDH) genotype in diffuse gliomas: initial promising results

Abstract

Background: Determining isocitrate dehydrogenase (IDH) mutation is crucial for glioma management. Slow diffusion coefficient (SDC) is a novel metric being proposed to measure in vivo tissue slow diffusion. In its basic form, SDC is derived from a high b-value diffusion-weighted image and a higher b-value diffusion-weighted image. The study attempts to distinguish IDH genotypes of diffuse gliomas using SDC alone and in combination with other two diffusion metrics of diffusion-derived vessel density (DDVD) and apparent diffusion coefficient (ADC).

Methods: This study enrolled 63 patients with diffuse gliomas (30 IDH-mutant and 33 IDH-wildtype) who underwent diffusion-weighted imaging at 3T. SDC was calculated with b = 500 and 750 mm²/s images. DDVD was calculated with b = 0 and 10 mm²/s images. ADC was calculated with b = 0 and 1000 mm²/s images. Correlations between the diffusion metrics and IDH genotypes were studied, as well as the correlation between SDC and Ki-67 expression.

Results: There was a significant difference among three histological grading of glioma (median value: 0.472 au/s for grade-2, 0.441 au/s for grade-3, 0.364 au/s for grade-4, p < 0.0001). Based on IDH gene testing, IDH mutant negative tumors had SDC value of 0.339 ± 0.055 au/s, IDH mutant positive tumors had SDC value of 0.437 ± 0.097 au/s, with AUROC of 0.828 for separation. SDC was negatively and weakly correlated with DDVD, with Pearson r of -0.212 (p = 0.096). A combination of SDC and DDVD separated IDH mutant -/+ tumors with an AUROC of 0.886. SDC was positively and moderately correlated with ADC, with Pearson r of 0.705 (p < 0.0001). AUROC analysis shows a combination of SDC, DDVD, and ADC separated IDH mutant -/+ tumors with an AUC of 0.897. If immunohistochemistry IDH partially positive tumors were not included, a combination of SDC, DDVD and ADC separated immunohistochemistry IDH mutant -/+ tumors with an AUC of 0.911. SDC was negatively and moderately correlated with Ki 67 LI, with Pearson r of -0.382 (p = 0.002).

Conclusion: SDC can help to distinguish IDH genotypes in diffuse gliomas. A combination of SDC and DDVD, or a combination of SDC, DDVD, and ADC, can further improve disease classification.

Clinical trial number: Not applicable.

Keywords: Diffusion magnetic resonance imaging; Glioma; Isocitrate dehydrogenase; Ki-67.

Fig. 1

Glioma patient enrolment flow diagram. IDH = isocitrate dehydrogenase, GBM = glioblastoma; AC = astrocytoma; OD = oligodendroglioma; NEC = not elsewhere classified (these gliomas are not included in histological grading)

Fig. 2

Examples of ROI delineation and representative cases of gliomas. (A) A patient with glioblastoma IDH-wildtype in the left frontal lobe (A1), showing relatively low SDC value (A2), high DDVD value (A3), and low ADC value (A4). (B) A patient with astrocytoma IDH-mutant (grade 2) in the left insular lobe (B1), showing high SDC value (B2), low DDVD value (B3), and high ADC value (B4). (C) A patient with astrocytoma IDH-mutant (grade 3) in the right temporal lobe (C1), showing high SDC value (C2), low DDVD value (C3), and relative high ADC value (C4). ROI = region of interest, IDH = isocitrate dehydrogenase, DDVD = diffusion-derived vessel density, SDC = slow diffusion coefficient

Fig. 3

Glioma SDC value Box and Whiskers plots. Tumor histological grade was significantly correlated with SDC, with higher-grade tumors exhibiting lower SDC values. As the tumor grade progressively increased from grade 2 to grade 4, SDC demonstrated a stepwise downward trend

Fig. 4

IDH mutant negative tumors had lower SDC values and IDH mutant positive gliomas had higher SDC values. Data A and B are based on IDH gene testing results. Data C and D are based on IDH immunohistochemistry results. In D, partially positive tumors were excluded. IDHm (-) or IDHm (0): IDH mutant negative; IDHm (+) or IDHm (1): IDH mutant positive; IDHm (0.5): IDH mutant partially positive

Fig. 5

Separating IDH mutant negative (IDHm-) and positive (IDH+) gliomas by a combination of SDC and DDVD. A and B include all 63 cases, and IDH results were based on gene testing. In C and D IDH, results were based on immunohistochemistry, and partially positive cases were removed. A comparison of A and C shows, IDH mutant partially positive cases were more likely to have diffusion metrics deviated from the clustering of IDH mutant positive cases and the clustering of IDH mutant negative cases (for example, the two cases labeled with arrow in A). ROC: Receiver operating characteristic curve

Fig. 6

Separating IDH mutant negative (IDHm-) and positive (IDHm+) gliomas by a combination of SDC, DDVD, and ADC. A and B include all 63 cases, and IDH results were based on gene testing. In C and D, IDH results were based on immunohistochemistry, and partially positive cases were removed. A comparison of A and C shows, IDH mutant partially positive cases were more likely to have diffusion metrics deviated from the clustering of IDH mutant positive cases and the clustering of IDH mutant negative cases (for example, the two cases labeled with arrow in A). SDC in au/s, DDVD in au/pixel, ADC in 10^− 4 mm²/s

Fig. 7

Probability determination of each glioma being IDH mutant positive or negative by a combination of SDC, DDVD, and ADC. This figure is the same as Fig. 6A. According to Eq. (4), the dot of A, B, C has the probability of 0.842, 0.053, and 0.466, respectively, for being IDH mutant glioma [m(+)]