The field of diffuse optical imaging has made significant advances since the two reviews of “optical imaging in medicine” were published in Physics in Medicine and Biology in 1997. Imaging systems have become faster, more robust, less susceptible to error, and able to acquire data with more source-detector combinations and at more wavelengths. Images are now routinely reconstructed in 3D, using more sophisticated techniques which can be adapted to the clinical situation by incorporating prior information and by compensating for some of the unavoidable sources of measurement error. The greatest recent progress has probably been in the demonstration of clinical applications: many hundreds of optical breast and brain examinations have now been carried out worldwide.
However, diffuse optical imaging is still primarily a laboratory-based technique and has yet to progress to routine clinical use. Improvements are still required in qualitative and quantitative accuracy, both of which are limited by poor spatial resolution. Recent developments, in particular the use of anatomical prior information, are already indicating that improved image quality is achievable. It seems increasingly probable that optical techniques could play a major clinical role when used as adjuncts to other imaging systems. For example, the addition of an integrated optical imaging system to a new MR imaging facility may provide enhanced clinical information about haemodynamics and metabolism at minimal additional cost and complexity compared to the MRI system itself.
A further role which optical imaging could fill is as a low-cost, portable imaging system for use in primary care situations or at the bedside. Similarly, optical imaging has already led to increased knowledge of brain physiology when used by psychology researchers who may not have access to an expensive fMRI system, and it can provide unique information about neuro-vascular coupling and brain development.
The transfer of new techniques and ideas for diffuse optical imaging into clinical tools remains a demanding problem, requiring close collaboration between engineers, clinicians, scientists and mathematicians. We will only know what clinical utility diffuse optical imaging can ultimately have when the ideal instrumentation is combined with the best modelling and reconstruction methods, and applied to the most appropriate clinical question.