A prospective study on the added value of pulsed arterial spin-labeling and apparent diffusion coefficients in the grading of gliomas

Abstract

Background and purpose: The purpose of this study was to determine whether qualitative and quantitative measures obtained with pulsed arterial spin-labeling (PASL) and apparent diffusion coefficients (ADC) improve glioma grading compared with conventional MR images.

Materials and methods: We prospectively performed 2 qualitative consensus reviews in 33 suspected gliomas: 1) conventional MR images alone and 2) conventional MR images with PASL and ADC. To calculate the diagnostic performance parameters of PASL and ADC, we used a qualitative scoring system on the basis of the tumor perfusion signal intensity (sTP) and visual ADC scoring (sADC). We then analyzed quantitative regions of interest and calculated the ratio of the maximum tumor perfusion signal intensity (rTPmax) and the minimum ADC value (mADC).

Results: Two observers diagnosed accurate tumor grades in 23 of 33 (70%) lesions in the first review and in 29 of 33 (88%) lesions in the second review. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for determining a glioma grading by using combined sTP and sADC scoring were 90.9, 90.9, 95.2, and 83.3%, respectively. Statistical analysis gave a threshold value of 1.24 for rTPmax and 0.98 x 10(-3) mm/s(2) for mADC to provide a sensitivity, specificity, PPV, and NPV of 95.5, 81.8, 91.3, and 90.1% and 90.9, 81.8, 90.9, and 81.8%, respectively. The receiver operator characteristic curve analyses showed no significant difference between the quantitative and combined qualitative parameters.

Conclusion: PASL and ADC significantly improve the diagnostic accuracy of glioma grading compared with conventional imaging.

Fig 1.

A 41-year-old man with a suspected recurred tumor after surgery. There is a contrast-enhancing lesion around the previous tumor resection site on contrast-enhanced T1-weighted axial image (A). Tumor perfusion score (sTP) on pulsed arterial spin-labeling map is 4 (B, arrow). Visual apparent diffusion coefficient score (sADC) is 3 (C, arrow); therefore, the combined sTP and sADC is 7. Initial imaging diagnosis was glioblastoma multiforme (grade 4); however, on pathologic examination this tumor was confirmed as an anaplastic astrocytoma (grade 3).

Fig 2.

A 45-year-old woman with a suspected brain tumor in the right basal ganglia. Contrast-enhanced T1-weighted axial image shows enhancing mass in the right basal ganglia (A). Tumoral perfusion score (sTP) is 3 (B, arrow). The visual apparent diffusion coefficient score (sADC) is 2 (C); therefore, the combined sTP and sADC is 5. On pathologic examination, this tumor was confirmed as an anaplastic astrocytoma (grade 3).

Fig 3.

Interactive dot diagram. The sensitivity and specificity for the determination of a glioma grade with the ratio of maximum tumoral perfusion signal intensity (rTPmax) were 95.1 and 81.8, with a threshold value of 1.24.

Fig 4.

Interactive dot diagram. The sensitivity and specificity for the determination of a glioma grade with a minimum apparent diffusion coefficient value (mADC) were 90.9 and 81.8, with a threshold value of 0.98 × 10^–3 mm/s².

Fig 5.

The ROC curve of 2 quantitative (rTPmax, mADC) and 3 qualitative (sTP, sADC, sTP and sADC) parameters for glioma grading. The areas under the ROC curve for the 5 parameters are as follows: rTPmax, 0.97; sTP and sADC, 0.96; mADC, 0.93; sTP, 0.90; and sADC, 0.80.