Evidence-Based Guidelines for the Rational Use of Neuroimaging

Published 23 August 2017. Last updated 7 September 2017.

This project involved a retrospective audit of patients who received neuroimaging at Sydney Eye Hospital (SEH) and surrounding facilities, to identify elements of an appropriate neuroimaging request.


To develop evidence-based guidelines for the rational use of neuroimaging for all neuro-ophthalmic presentations at SEH.


  • Provides doctors with evidence-based guidelines.
  • Reduces healthcare costs associated with neuroimaging.
  • Reduces inconvenience for patients and their families.
  • Reduces discovery of irrelevant incidental findings.
  • Reduces exposure to unnecessary radiation and contrast agents.


Neuroimaging is often necessary when assessing patients who present to hospital with neuro-ophthalmic conditions. However, if it is performed when not clinically required, it can lead to increased healthcare costs, inconvenience for patients and their families, irrelevant incidental findings and unnecessary exposure to radiation and contrast agents. As such, it is important to only request neuroimaging for certain neuro-ophthalmic presentations.

In 2014-15, the radiology budget at SEH was $2 million, which did not include magnetic resonance imaging scans (MRI) performed away from the SEH campus. Other costs include ambulance transfers for inpatients, as well as time and transport for outpatients to attend the scan. As there are currently no neuroimaging guidelines for SEH, it is important that only those who require scans receive them, and the right scan is performed.

To make sure patients only receive neuroimaging when clinically required, a study was required to identify which patients require neuroimaging, what type of neuroimaging is most appropriate, the time at which neuroimaging should be conducted, and whether patients require neuroimaging as an inpatient or outpatient. These factors will allow evidence-based guidelines to be developed, to support appropriate neuroimaging requests.


A retrospective audit was conducted on all neuroimaging studies ordered for SEH neuro-ophthalmic and glaucoma presentations between June 2011 and June 2016. The audit included brain, orbit and spine computed tomography (CT) and MRI scans performed at 15 sites in total, including SEH, Prince of Wales Hospital, St Vincent’s Hospital Sydney and private radiology practices within a five kilometre radius of SEH. Data about the clinical presentation was collected from patient records reviewed at Sydney Hospital Medical Records Department between August 2016 and March 2017.

Project status

Implementation – The project is ready for implementation or is currently being implemented, piloted or tested.

Key dates

Project planning commenced in December 2015, with ethics approval from July 2016. Data collection was completed by March 2016, with processing and analysis by June 2016. After convening a committee of experts to develop the guidelines, the guidelines will be trialled at SEH during the 2018 clinical year. Data will then be collected at the end of 2018 and compared with previous year to evaluate the project’s success.

Implementation sites

Sydney Eye Hospital, South Eastern Sydney Local Health District


Save Sight Institute, University of Sydney


  • 839 patients had neuroimaging studies for neuro-ophthalmic or glaucoma presentations between June 2011 and June 2016.
  • 403 (48 per cent) of participants were women and the average age was 53.
  • 1042 neuroimaging studies were conducted during the audit period, with 596 (57 per cent) CT scans and 446 (43 per cent) MRI scans.
  • The most common clinical presentations that led to neuroimaging were blurred vision (20 per cent), diplopia (15 per cent) and disc swelling (14 per cent).
  • Field defects (39 per cent), disc pallor (36 per cent) and ptosis (33 per cent) had the highest rate of relevant findings in neuroimaging studies.
  • 33 per cent of scans led to an incidental finding, with only three per cent amenable to management.
  • The diagnosis of 23 per cent of presentations remained unknown.
  • The most common diagnosis for blurred vision was optic neuritis (n=62), with 50 per cent of patients receiving confirmation of this diagnosis through radiology tests.
  • Three per cent of patients with visual field defects (and no suspected optic neuritis or migraine) were diagnosed with acute haemorrhagic stroke with an urgent non-contrast brain CT scan.
  • Imaging for transient binocular vision loss revealed that six per cent of patients had an intracranial haemorrhage and nine per cent had a space-occupying lesion, while imaging transient monocular visual loss had no remarkable findings.
  • In patients with diplopia, 1.5 per cent had an aneurysm and 0.5 per cent had a carotid artery dissection, all of which were associated with a headache, ptosis or other cranial nerve palsy.
  • The 31 per cent of patients with relevant findings on imaging for ocular pain all had other clinical features that suggested the diagnosis was optic neuritis.
  • Three per cent of patients with bilateral disc swelling had underlying space-occupying lesions, while 40 per cent of glaucoma patients with sudden worsening of visual fields had a space-occupying lesion.

View a poster of this project, presented at the ACI Eyes on the future forum.

Lessons learnt

For eye-related presentations, a brain MRI is the most appropriate imaging in most clinical circumstances, with exceptions including suspicion of acute haemorrhage, a calcified pathology, proptosis or a rapidly-evolving emergency situation. Regardless of what type of imaging is ordered, contrast improves the sensitivity of the test and should be requested unless contraindicated or not required.

Clinical presentations that were found to have a low yield on neuroimaging included transient monocular visual loss, migraine, isolated ocular pain, isolated visual aura, flashes, floaters or headaches without red flag features. While the individual clinical context must be considered, routine neuroimaging for these conditions does not appear appropriate. These principles can be applied to other eye care centres and emergency departments.

Further reading

  • Mathews J, Mathews D, Walker S et al. Can ophthalmic requests for neuroimaging be improved? Eye 2004;18(3):290-2.
  • Mehta S, Loevner LA, Mikityansky I et al. The diagnostic and economic yield of neuroimaging in neuro-ophthalmology. Journal of Neuro-Ophthalmology 2012;32(2):139-44.


Shanil Dhanji
Senior Resident Medical Officer
Sydney Hospital and Sydney Eye Hospital
South Eastern Sydney Local Health District
Phone: 02 9382 7111

Mitchell Lawlor
Neuro-ophthalmologist and Glaucoma Surgeon
Sydney Hospital and Sydney Eye Hospital
South Eastern Sydney Local Health District
Phone: 02 9382 7111


Fill in our feedback form to find out more about this project or get in touch with the project manager.

Is this your project?

Fill in our feedback form to update your story or contact details.

Browse similar projects

Back to top