Fig 1.
CONSORT Flow diagram for Part 1 of clinical study.
Flow diagram of the progress through the phases of enrollment, allocation, follow-up, and data analysis for Part 1 of the clinical study.
Fig 2.
CONSORT Flow diagram for Part 2 of clinical study.
Flow diagram of the progress through the phases of enrollment, allocation, follow-up, and data analysis for Part 2 of the clinical study.
Fig 3.
Photographs of the handheld prototype PRODIGI imaging device and its use for real-time autofluorescence imaging of bacterial load invisible by white light examination.
A. Front view of PRODIGI showing wound fluorescence image displayed in real-time on the LCD screen in high definition. B. Back view of PRODIGI showing white light and 405 nm LED arrays providing illumination of the wound, while the fluorescence mission filter is placed in front of the CCD sensor. Inset shows side profile of the device. C-E. Photograph of PRODIGI device used to examine a diabetic foot ulcer with room lights on, in a hard shell carrying case in a typical wound clinic setting, and placed on typical wound care cart, respectively. Room lights are turned off for fluorescence imaging. F. PRODIGI white light image of type II diabetic foot ulcer in a 52 y old male patient. G. Corresponding AF image taken in < 1 sec showing bright red fluorescence of pathogenic bacteria in the wound periphery (yellow arrow) and in ‘off site’ areas (white arrow) away from the primary wound (confirmed by swab microbiology as mainly heavy growth S. aureus). Bacterial fluorescence appears red against a background of green fluorescence from connective tissues of the healthy skin, which provides anatomical context for localizing the bacteria within and around the wound. The bacterial regions were not seen under white-light visualization. H. A magnified view of G. showing S. aureus growing within the fissures of the wound periphery. Bright fluorescent ‘hot spots’ (yellow arrow) illustrate heterogeneity in the distribution of bioburden in the wound periphery. Fluorescence imaging allowed targeted swabbing of bacterial areas not possible by white light visualization. The heavy growth S. aureus growing in the off-site area was invisible by traditional clinical examination. Scale bars: A. 2 cm, B. 2 cm, C. 1 cm.
Fig 4.
Autofluorescence detection of clinically-significant bacterial load in wound periphery and off-site areas.
A. White light image of a type II diabetic foot ulcer in a 78 y old female. B. Corresponding AF image showing heavy growth of S. aureus in the wound periphery missed by white light imaging. C. White light shows unremarkable areas between toes, while in D. the corresponding AF imaging detected bacterial biofilm, confirmed by microbiology. E. Schematic illustrates different wound locations where fluorescence imaging detected clinically significant bacterial load. F. Comparison of accuracy for correctly detecting clinically-significant bioburden between standard WL and AF imaging in wound bed, wound periphery and off-site areas. Scale bars: A,B. 1 cm; C. 2 cm, D. 1 cm.
Fig 5.
Wound swabs taken under autofluorescence imaging guidance, indicating detected polymicrobial species prevalence.
The percentage of bacteria for each species are plotted for A. all swabs, B. swabs obtained from the wound bed, C. from the wound periphery and D. ‘off-site’ areas away from the primary wound.
Fig 6.
Quantitative longitudinal tracking of bacterial load in chronic wounds.
A. Sequential white light (top row) and AF images (middle row) of a non-healing diabetic foot ulcer in a 67 y old female performed over 5.5 months. AF images revealed bluish-green fluorescence within the wound bed and bright red bacteria around the wound periphery. A fluorescence intensity-based segmentation algorithm was used to quantify bacterial load changes over time (bottom row), with the bacteria false-colored and overlaid on original AF images of the middle row. B. Quantitative changes in bacterial load over time measured by relative bacterial fluorescence amount (total red AF area in cm2) indicate clinically-significant microbial load in the wound periphery, both of which are missed during conventional clinical examination. Scale bar: A. ~1 cm.
Fig 7.
Fluorescence image-guided treatment accelerates wound closure over time compared with non-guided conventional treatment.
A. Plot showing wound area measurements for twelve individual patients as a function of time from the onset of the study over the first control period (blue circles, the fluorescence image-guided period (red solid circles) and the second control period (green triangles). During the control periods, treatment was administered without real-time fluorescence image guidance. B. Plot showing the rate of change in the average wound area over the course of the study estimated from the regression model. The slopes and their corresponding p-values are shown for each study period. The data indicate that fluorescence image-guided treatment increased the rate of wound closure in a statistically significant manner compared with the control (non-guided) periods indicating that fluorescence image guided wound treatment could be beneficial addition to conventional wound therapy protocols. * p-value tests for change in the growth rate of average wound area in the guided period from the previous period. ** p-value tests for change in the growth rate of average wound area in the 2nd control period from the guided period.