Thermal imaging of the inner canthi: a potential non-pharmaceutical intervention for pandemic preparedness

Lead supervisor: Dr Susannah Fleming

Co-supervisor: Professor Richard Stevens

Commercial partner: National Physical Laboratory

Background 

Epidemics and pandemics of infectious disease are a significant risk to human health and cause significant social disruption.  Experience from the COVID-19 pandemic shows that in the initial stages of the pandemic, non-pharmaceutical interventions (NPIs) are a key part of the response to limit spread of the infection whilst allowing essential services to continue and buy essential time for development of lateral flow tests and vaccines.  NPIs include universal interventions such as handwashing and limiting interpersonal contact, as well as targeted interventions, such as fever screening, to identify potentially infected individuals.

Fever is a common feature of infection and has been used as a key indicator to identify subjects in previous infection outbreaks such as SARS, pandemic influenza, and COVID-19.  However, reliable temperature measurement modalities require contact with a potential infected individual to obtain the required accuracy. Current non-contact thermometry methods (forehead/tympanic) do not have sufficient accuracy for reliable fever detection in public settings.

Thermal imaging of the inner canthi of the eye has the potential for accurate, non-contact and rapid screening of large numbers of subjects for fever, but this approach has not yet been well evaluated as a possible NPI for future pandemics.

Aims

The overall aim of the project is to evaluate whether thermal imaging of the inner canthi can reliably identify febrile individuals in public environments.  This will be achieved through:

1)           Identifying one or more candidate thermal imagers as well as automated methodology for segmenting the inner canthi from the thermal image, through a systematic review / horizon scan of current state of the art in thermal imaging capability. 

2)           Assessing the accuracy of thermal imaging of the inner canthi, compared to independently measured core body temperatures using a traceable accurate body temperature measurement.  This will be carried out in a controlled laboratory or clinical environment.

3)           Assessing real-world accuracy and reliability, through testing of thermal imaging of inner canthi in a public environment such as transport hub or hospital waiting room with automated segmentation and flagging of potentially febrile individuals.

This project is closely aligned to, and helps deliver, the MRC remit; particularly, we will develop high quality capacity in diagnostic and data analytics, demonstrating for the first time a definitive link between inner canthi derived temperatures and core body temperature (and so for the first time demonstrate reliable remote fever detection). We will do this through use of tools from data science working at the interface of human health (as indicated by reliable fever detection) and biology (including machine learning and AI). To achieve our objectives this project required blending of interdisciplinary skills bringing together leading capability in temperature metrology, medical statistics and physiological measurements for healthcare and we will hard-wire into the project  researcher mobility between different parts of the innovation ecosystem as the researcher will spend time at facilities in Oxford and NPL. Finally, the success of this project depends upon blending complementary skills in leading temperature metrology from NPL and clinical study design and medical statistics from Oxford.

Apply using course: DPhil in Primary Health Care