Project Title:

Precisely Mapping Choroidal Tumour Margins

Chief Investigator:

Dr Roderick O’Day

Co-Investigators:

Dr Xavier Hadoux, Prof Peter van Wijngaarden

Aim

To assess the accuracy of choroidal tumour margin delineation by standard imaging techniques (colour fundus photography, scanning laser ophthalmoscopy, and optical coherence tomography) and examine the potential application of hyperspectral retinal imaging in this setting.

Methods

Participants with a choroidal naevus or melanoma were imaged standard imaging techniques and hyperspectral retinal imaging during a single visit. Images across multiple technologies were precisely aligned to enable comparison. Margins were delineated by two ocular oncologists. The accuracy of each of the techniques and hyperspectral retinal imaging was examined.

Key results

84 choroidal tumours were imaged. Multimodal imaging (in particular the addition of optical coherence tomography) was more accurate than unimodal imaging at assessing tumour margins. Tumours with low pigmentation were poorly assessed by colour fundus photography and better by scanning laser ophthalmoscopy.
Hyperspectral retinal imaging was as accurate at assessing tumour margins as the best available imaging technologies.

 

Implications for Clinical Practice/Science and Future Research

A follow-on study has commenced to confirm and expand on our findings. Future studies will assess whether hyperspectral retinal imaging can identify the imaging predictors of choroidal naevi growth.

Conclusion

Choroidal tumour margin assessment can be inaccurate when using conventional imaging technologies,
particularly in patients with non-pigmented choroidal tumours. A multimodal imaging approach reduces
these inaccuracies and hyperspectral retinal imaging shows promise in assessment of choroidal tumour margins.

 

Lay Summary of Outcomes

Ophthalmic imaging technology has transformed the management of patients with benign and malignant eye tumours over the past 20 years. This study used Australian-developed software to better image the hard to identify tumours and a novel technology (hyperspectral retinal imaging) shows excellent promise in being able to assess choroidal tumours.

 

This research project was funded by AVR in 2023