0796

Absolute Cerebral Blood Flow Derived from Low GBCA Dose DCE-MRI in Patients with  Type 2 Neurofibromatosis
Ka-Loh Li1, Daniel Lewis1,2, Xiaoping Zhu1, and Alan Jackson1

1Division of Informatics, Imaging and Data Science, University of Manchester, Manchester, United Kingdom, 2Department of Neurosurgery, Salford Royal NHS Foundation Trust, Manchester, United Kingdom

Synopsis

A newly developed low dose T1W-DCE-MRI method, ACcomb, was used to estimate the absolute CBF of vestibular schwannomas (VS) and normal appearing brain tissue in a group of 12 consecutive type 2 neurofibromatosis (NF2) patients, undergoing anti-angiogenic bevacizumab treatment. This new method consistently displayed excellent gray-white matter flow contrast and produced mean GM and WM CBF estimates consistent with previous literature values. Use of this new method showed that at 90 days post bevacizumab treatment, there was increased positive correlation between CBF and estimated plasma volume within the VS, and a significant increase in CBF within normal appearing white matter.

Introduction

We have recently developed a new methodology for absolute CBF estimation using low gadolinium based contrast agent (GBCA) dose T1-DCE-MRI. The new method combined two reciprocal ways for searching optimal points located in early transit time window of the tissue concentration curves for GBCA concentration averaging. The averaged GBCA concentration (AC) was then applied in a microsphere model for CBF calculation (1). In this study we applied this method, named as ACcomb, to a group of patients with type 2 neurofibromatosis (NF2), who underwent bevacizumab treatment, for estimation of perfusion in the vestibular schwannoma (VS) and normal appearing brain tissues.

Methods

Twelve consecutive NF2 patients with a total of 21 VS were recruited. Each tumor was classified according to response to anti-angiogenic therapy over the 90 day period. Response was defined as a volume reduction exceeding 0.125 cm3 or a relative volume decrease exceeding 5%. A low GBCA dose high temporal resolution (LDHT-) DCE-MRI with ∆t = 1.0 s was acquired, following intravenous injection of 0.02 mmol/kg of a macrocyclic type of GBCA. Tissue concentration curve, Ct(t), was derived using the LDHT-DCE and 3D native T1N maps from a series of 3D variable flip angle images, α = [2°,8°,15°,20°].

Absolute CBF was obtained by incorporating the ACcomb with individually measured vascular input function, VIF(1). To investigate possible variations in the shape and amplitude of VIFs, we retrospectively inspected VIFs from an extended cohort. Patients in the extended cohort were scanned with the same LDHT-DCE protocol (Fig. 1).

A concurrent kinetic analysis was also performed with the same DCE-MRI using extended-Tofts model, to produce the plasma volume (vp) and transfer constant (Ktrans).


Results

Figure 2 shows maps of Ktrans and CBFACcomb from a NF2 patient with a large left sided VS, scanned with LDHT-T1W-DCE-MRI pre-(day0) and 90 days post-treatment(day90).

Figure 3 shows CBFACcomb maps of a NF2 patient with a large left sided VS and four meningiomas, on day0 and day90. The vp images are also displayed for comparison. Slices from two levels of the 3D parametric images are shown. The appearances of the T1W-CBFACcomb are similar to vp maps. As can be seen there is some reduction of CBFACcomb and vp within the VS on 90 days after treatment.

Figure 4 shows an inter-subject, inter-tumor comparison between the mean values of T1W-CBFACcomb and vp from the 21 tumors (VS) of the 12 NF2 patients, undergoing bevacizumab treatment. Linear regression analysis shows that T1W-CBFACcomb and vp are significantly correlated, on both pre- and post-treatment (day0 and day90). The correlation is much stronger on day90 (R2 = 0.827, p < 0.0000) than day0 (R2 = 0.505, p < 0.001).

Figure 5a shows mean CBFACcom values in GM: 55.9 ± 14 (day0), 61.0 ± 8.3 (day90) ml/min/100ml, in WM: 25.8 ± 3.5 (day0), 28.4 ± 3.4 (day90) ml/min/100ml and mean gray/white ratio: 2.15 ± 0.3 (day0), 2.15 ± 0.2 (day90). All values were consistent to those reported in literature. There is a significant increase of CBF in normal appearing WM after treatment. Figure 5b shows differences between mean CBF in VS before and after treatment for the responders and non-responders.

Discussion

This study provides new information regarding cerebral blood flow in NF2 patients undergoing anti-angiogenic therapy. Firstly, our AC data provide evidence that the mean values of absolute CBF of normal appearing WM and GM are consistent with literature values in healthy volunteers measured using the model-based deconvolution approaches (2,3). Secondly, whilst the mean pre-treatment CBF values of the VS are within the range of normal appearing GM, the concurrent measured Ktrans is much higher than GM suggesting the tumor vessels are more leaky. Post bevacizumab treatment the correlation between the CBF and plasma volume of all VS was much stronger than pre-treatment, and there was heterogeneity in VS response, with responders showing a trend to decreasing tumor CBF whereas non-responders showed increasing CBF. Furthermore, within the normal appearing brain, there was an interesting finding of significantly increased WM CBF 90 days treatment.

The clinical application of the new ACcomb method consistently displayed excellent gray-white matter flow contrast compared with previous reported DCE-MRI study (4), while using a much lower GBCA dose than those of DSC-MRI (5). Reduced GBCA dose is a pertinent clinical concern, as patients with benign CNS tumors may receive many contrast-enhanced MR scans throughout their lifetime.

Conclusion

Combining the merits of the low GBCA dose, fast acquisition and large volume coverage, the T1W-CBFACcomb could be a potential valuable marker for monitoring the therapy responses of CNS tumors to anti-angiogenic therapies.

Acknowledgements

This work was supported by CRUK [C8742/A18097]. This is a contribution from the Cancer Imaging Centre in Cambridge & Manchester, which is funded by the EPSRC and Cancer Research UK.

References

1. Whole Brain Absolute CBF Measurements Using Ultra-Low GBCA DCE-MRI and Microsphere Model. Submitted to ISMRM 2018.

2. Larsson HB, Hansen AE, Berg HK, Rostrup E, Haraldseth O. Dynamic contrast-enhanced quantitative perfusion measurement of the brain using T1-weighted MRI at 3T. J Magn Reson Imaging 2008;27(4):754-762.

3. Sourbron S, Ingrisch M, Siefert A, Reiser M, Herrmann K. Quantification of cerebral blood flow, cerebral blood volume, and blood-brain-barrier leakage with DCE-MRI. Magn Reson Med 2009;62(1):205-217.

4. Moody AR, Martel A, Kenton A, et al. Contrast-reduced imaging of tissue concentration and arterial level (CRITICAL) for assessment of cerebral hemodynamics in acute stroke by magnetic resonance. Invest Radiol 2000;35(7):401-411.

5. Kwong KK, Chesler DA. Early time points perfusion imaging: theoretical analysis of correction factors for relative cerebral blood flow estimation given local arterial input function. Neuroimage 2011;57(1):182-189. 3.

Figures

Figure 1 shows variation in the shape of blood concentration, Cb(t), curves measured from the superior sagittal sinus (SSS) of different individuals. a). Distribution of number of exams with various number of data points on upslope of the first pass of the Cb(t) curves from 55 cases, who were injected GBCA bolus of 0.02 mM/kg at 3ml/s rate. Bottom row shows Cb(t) curves with N upslope data points. b) N = 4, from a 16-year old male (weight=85 kg), c) N = 6, 16-year old female (weight=60kg), d) N = 8, 25-year old male of (weight=107 kg).

Figure 2. The central slices of 3D Ktrans map (top row), derived from fitting the LDHT-T1W-DCE data to the extended Tofts model. and 3D T1W-CBFACcomb (bottom row), calculated using the microsphere model, from an NF2 patient with a large left sided vestibular schwannoma pre- (day0, left) and 90 days post-treatment (right). The VS shows the Ktrans values much higher than that of normal appearing GM (NAGM), but the CBF values close to that of NAGM. There are marked decreases of Ktrans and CBFACcomb accompanied by reduction of tumor volume of the VS after 90 days anti-angiogenic treatment.

Figure 3. Maps of T1W-CBFACcomb (top row) and vp by extended Tofts model (bottom row) from an NF2 patient with a large left sided vestibular schwannoma and four meningiomas. These maps were derived from the LDHT-T1W-DCE-MRI acquired on a 1.5T scanner, pre- and 90 days post- bevacizumab therapy. Slices from two levels of the 3D maps of CBF and vp are shown, the vestibular schwannomas within the left cerebellopontine angle (left two colomn), one of the four meningiomas (right two columns).

Figure 4. Scatter-plots of the tumor mean T1W-CBFACcomb and vp of the 21 vestibular schwannomas from 12 patients with NF2, for day0 (green) and day90 (red). The green line and red line represent the theoretical vp values calculated with the regression formulas for day0 (in green) and day90 (in red, on the right-down corner of the graph). Linear regression analysis shows that T1W-CBFACcomb and vp are significantly correlated, both pre- and post-treatment (day0 and day90). Nevertheless, the correlation is much stronger on day90 (R2 = 0.827, p < 0.0000) than day0 (R2 = 0.505, p < 0.001).

Figure 5. Comparison of T1W-CBFACcomb of WM, GM and VS from 12 NF2 patients before and after bevacizumab treatment. P-values are from paired two-tailed t-tests. (a) Differences of group mean CBF between day0 and day90 in normal appearing GM and WM; There is a significant increase of CBF in normal appearing WM. (b) Differences between group mean CBF of VS on day0 and day90 from the responders and non-responders to angiogenic therapy. The responders show a trend to decreasing CBF after bevacizumab treatment; the non-responders show a trend to increasing CBF after treatment.

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)
0796