Introduction Fuji Intelligent Chromo Endoscopy (FICE) is a post-processing image enhancement technique designed by Fuji for their video endoscopy systems. FICE is a method that takes the original white light image and separates the image into discrete light wavelengths, each setting has a different combination of three selected wavelengths (from 400 to 700 nm), representing RGB, that are independently weighted, combined and superimposed on the white light image to produce the FICE image. Fuji currently provides 10 different settings. The aims of this project are: to evaluate the current FICE settings on images captured during upper GI endoscopy, and the creation of new oesophageal settings to simplify the selection and improve the application of FICE in the diagnosis of oesophageal pathology.
Methods We used a PC based FICE simulator, provided by Fuji, to process images offline, using the FICE settings. A series of images were captured during diagnostic endoscopies of various conditions (eg, varicies, Barrett's oesophagus) at various points in the oesophagus using the capture facilities on the Fuji endoscopy EPX-4400 processor. Using the FICE simulator, two new FICE settings were created to maximise the enhancement of the pathology while maintaining the anatomical colouring. For each RGB component, the selected wavelength was altered independently in 5 nm steps. Once the wavelength was fixed, the gain was altered to minimise artefacts and maximise enhancement. Forty images were selected from the series that covered various conditions, and were processed via the FICE simulator using the 10 standard and two new settings. The images were randomised for evaluation by five blinded endoscopists. Each endoscopist compared the original and FICE image and scored the degree of enhancement over the original image from 0 (no enhancement) to 5 (maximum). The highest score FICE settings were translated to the EPX-4400 processor for further clinical evaluation.
Results The oesophageal mucosa presented with two distinct shades of pink (eg, Light and Dark). Two settings were created, one for each mucosal shade (L1—Light and L2—Dark). From the 40 images 65% would be characterised as Light and 35% Dark mucosa. Out of a possible enhancement score of 1000, the standard FICE Settings (0–9) scored between 202 and 319, with settings 2 scoring the highest (319). The Lothian FICE settings L1 and L2 scored 463 and 387 respectively.
Conclusion In conclusion the L1/2 FICE settings were found to provide further enhancement compared with current FICE settings by improving the colour discrimination between normal and abnormal mucosa. FICE is a useful and flexible system with a lot of potential but still requires optimisation.
Competing interests S Inglis Grant/Research Support from: Fuji provided software for Trial, S Alexandridis Grant/Research Support from: Fuji/Imotech Medical Provided funding for a Fellowship, J Plevris Grant/Research Support from: Fuji Provided Loan Equipment for trial.