Synopsis

In this work we demonstrate that careful T₁ mapping of the vitreous humour can identify disease in the retina. We studied a cohort of patients with central vein retinal occlusion (a type of retinal ischaemia) and show that significant decreases in T₁ of the vitreous humour are observed compared to healthy control eyes. We speculate that the decreases may be the result of increased pO₂ that could arise when the oxygen demand of the retina is reduced as a consequence of damage. We show preliminary data from patients with proliferative diabetic retinopathy and ocular ischaemic syndrome that suggest that it may be possible to discriminate different forms of retinal ischaemia completely non-invasively with MRI.

Aim

Introduction

The vitreous humour is the clear gel that fills the eye ball between the retina and the lens. Recent research has demonstrated that it has a role in the pathogenesis of a number of conditions such as central retinal vein occlusion (CRVO). Reduced retinal blood supply in conditions such as CRVO is concomitant with a decrease in oxygen supply that triggers a cascade of damage to the retina that may result in severe visual loss. Direct MRI of the retina itself is particularly difficult: it is <0.5 mm thick and suffers from significant movement artefact. However, retinal health can be inferred by probing the vitreous humour that fills the bulk of the eye. The vitreous is in contact with the retina and has the advantage of being largely homogeneous and MR measurements are insensitive to small-scale movement. The partial pressure of oxygen (pO₂) can be estimated with MRI using T₁ mapping since molecular oxygen is paramagnetic and the longitudinal relaxation rate (1/T₁) of the vitreous humour is proportional to pO₂^1-3. In this work we use MR pO₂ mapping in the eye to assess the pO₂ in a cohort of patients with CRVO to demonstrate that MRI offers a non-invasive approach to detecting abnormalities in the vitreous. We also took preliminary measures of pO₂ in patients with proliferative diabetic retinopathy (PDR) and ocular ischaemic syndrome (OIS) to determine whether it may be possible to discriminate between sub-types of retinal ischaemia.

Materials and methods

Scanning protocol: Imaging was performed on a Siemens Avanto 1.5 T scanner (Erlangen, Germany). T₁ mapping was performed using an inversion recovery (IR)-trueFISP sequence with 17 inversion times in the range TI=0.7s–30s. A single slice was positioned through the centre of both eyes in the axial oblique plane. Other trueFISP parameters were: TR=(20+TI) s, TE = 1.52 ms, FA = 80°, matrix = 256x256, voxel dimensions = 0.9x0.9x4 mm³. The total scan time for T₁ measurement was 15 mins. Eye movement was controlled by instructing the participant to fixate using a visual aid¹. Fixation was only required for the duration of k-space acquisition for each TI (approx. 1s only).

Data Analysis: T₁ mapping involved a pixel-by-pixel three-parameter fit of the signal intensity S (at each TI) to the equation S(TI) = A + Bexp{-TI/T₁}; A and B are parameters that account for inversion pulse flip angle, equilibrium signal intensity and TR. Since flip angle is included, the technique is resilient to B₁ errors.

Participants: 8 patients (6 with CVRO, 1 with PDR and 1 with OIS). Age (years) mean=65; range 53-85. Exclusion criteria included bilateral eye disease. In this way, we could use the contralateral eye as a healthy control.

Results and discussion

There is a highly significant (p=0.001) decrease in mean T₁ value in the vitreous humour of patients with CRVO compared to T₁ measured in healthy control eyes (Figure 1; mean T₁ (CRVO)=4.25s, (control)=4.52s). Of note, all CRVO patients exhibited reduced T₁ in the vitreous of the diseased eye compared to their healthy eye; none opposed this trend (Figure 2). This finding could be suggestive of increased pO₂ as a result of retinal damage. This is particularly intriguing since hypoxic conditions have previously been associated with acute pathology in the retina⁴. However, it is important to note that these decreased oxygen levels were measured at surgery while hypoxia was suspected to be present as a result of active disease. A possible explanation for increased pO₂ in the present study is that when the retina has already been damaged, and it no longer forms an effective oxygen sink. In healthy eyes, the vitreous has a pO₂ of approximately 10 mmHg, much less than atmospheric pO₂ (130 mmHg) and pO₂ in the retinal arterioles (80 mmHg). When the retina is compromised, oxygen diffusing into the eye through the sclera and/or from plasma would lead to an increase in vitreous pO₂, as seen here. Exploratory measures of T₁ in OIS did not show a reduction compared to measures in control vitreous, while in PDR, a T₁ reduction was seen. This may be indicative of the differing levels of retinal ischaemia that characterize these conditions, although we are cautious about drawing conclusions from single-subject measures.

Conclusions

Acknowledgements

References

1. Dowell NG, Hughes EH, Simpson ARH, Tofts P. Determination of T1-dependence on oxygenation in the eye using a simple phantom model. Abstract presented at: Proceedings of the International Society of Magnetic Resonance in Medicine(ISMRM); May 5–11, 2012; Melbourne, Australia.

2. Simpson ARH, Dowell NG, Jackson TL, Tofts PS, Hughes EHH (2013). Measuring the Effect of Pars Plana Vitrectomy on Vitreous Oxygenation Using Magnetic Resonance Imaging IOVS, 54:2028-2034.

3. Muir ER, Zhang Y, Nateras OSE, Peng Q, Duong TQ (2013). Human Vitreous: MR Imaging of Oxygen Partial Pressure. Radiology, 266: 905–911.

4. Williamson TH, Grewal J, Gupta B, Mokete B, Lim M, Fry CH (2009). Measurement of pO2 during vitrectomy for central retinal veinocclusion, a pilot study. Graefes Arch Clin Exp Ophthalmol. 247:1019–1023.