Can We Reduce Neuroimaging in Ophthalmology?

The use of computed tomographic (CT) and magnetic resonance imaging (MRI) scans is becoming increasingly commonplace in ophthalmic practice.¹ The relative ease of ordering and the valuable anatomical information that scans provide make them a very useful tool.² However, this comes at a cost both in terms of radiation dose to patients (in the case of CT) and financial cost to the National Health Service. In ophthalmology, little work has been done in assessing the appropriateness of ophthalmic neuroimaging: the most recent study we found was conducted nearly 8 years ago.^1,3

We aimed to look at a sample of the current practice in diagnostic neuroimaging in a District General Hospital (DGH) with the aim of assessing appropriateness of requests and where possible suggesting changes to improve the accuracy of those requests.

We performed a retrospective case-note analysis of diagnostic neuroimaging at the Royal Preston Hospital between June 2008 and December 2008. Details of ophthalmology requests were obtained from the Radiology Department and 62 consecutive CT/MRI scans were examined. After identifying patients who underwent neuroimaging, the case notes were retrieved and the clinical indications and outcomes of the imaging were recorded. Appropriateness was determined, at the time of the case note review, by a combination of the reason for request and the criteria suggested by Lee et al.⁴

In total 23,773 patients attended the eye unit during the study period. Two hundred fifty-one CT or MRI scans were ordered, of which 62 were evaluated in our study. The main reason for requesting neuroimaging was suspected visual pathway lesion due to visual field defects (28/62; 45%). Other indications included suspected papilloedema (11/62; 18%), orbital pathology (8/62; 13%), suspected optic neuritis (4/62; 6%), ocular motility disorders (7/62; 11%) and miscellaneous (4/62; 6%). In total 8/62 (13%) of scans were deemed inappropriate, of which 6/62 (10%) were due to suspected visual pathway lesion.

CT and MRI scans in ophthalmology are becoming increasingly commonplace and 1.05% of our patients underwent neuroimaging during the study period. This is higher than the most recent similar study, which reported that 0.30% of patients underwent neuroimaging.¹ In our study, most of the scans were deemed appropriate (87%), which reflects similar previous studies (91% and 88%).^1,3

Suspected visual pathway lesion remains the most common reason for requesting scans, which confirms previous reports (45%).¹ Most of these were deemed appropriate but three scans were ordered for a single non-neurological visual field deficit without any other indication. Such situations do not require imaging. Also in this group, three scans were ordered for patients with advanced normal tension glaucoma (NTG). From the notes NTG was the most likely diagnosis and there was no indication that these patients required neuroimaging.

We feel that our sample is representative of other DGHs in England and we propose that the majority of our inappropriate scans could have been avoided had we developed more specific criteria when ordering scans for suspected visual pathway lesion. Such criteria that we suggest include

1. Suspected compressive lesions of the visual pathway—Request neuroimaging if there is reduced vision, a reproducible neurological visual field defect, pupillary defect, or suspicious optic disc.

2. Bilateral advanced disc cupping with normal IOP—These patients do not require neuroimaging unless there is an additional clinical feature suggestive of a compressive lesion.

Unfortunately, there is a “grey area” in the decision-making process to scan patients with suspected visual pathway lesions. Consequently an excess number of scans are ordered inappropriately. Our department has adopted these criteria and we hope they are useful to other units who face similar problems. This would have an impact in terms of reducing inappropriate radiation as well as financial benefits to the department.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.