Fig 1.
Relationship between erythema index and haemoglobin for palpebral and forniceal conjunctiva.
(A) Representative calibrated image from a non-anaemic participant showing the palpebral (p) and forniceal (f) portions of the conjunctiva. The in-frame colour calibration target is shown. (B) Relationship between palpebral conjunctival EI derived from LX5 images and measured haemoglobin (n = 94). Solid line represents best fit by linear regression; slope = 0.092 ± 0.016; P < 0.0001 of zero slope. Dashed line represents 95% confidence interval. (C) Relationship between forniceal conjunctival EI derived from LX5 images and measured haemoglobin (n = 94). Solid line represents best fit by linear regression; slope = 0.040 ± 0.014; P = 0.004 of zero slope. Dashed line represents 95% confidence interval.
Fig 2.
Test characteristics of LX5 and iPhone images.
(A) Receiver-operator characteristic (ROC) curve (continuous line) of palpebral conjunctival EI calculated from LX5 camera images for detection of haemoglobin < 110 g/L in the training set (n = 47). Dashed line represents unity. (B) ROC curve (continuous line) of palpebral conjunctival EI calculated from iPhone 5S images for detection of haemoglobin < 110 g/L in the training set (n = 47). Dashed line represents unity. (C) Palpebral conjunctival EI values calculated from LX5 images (n = 94), grouped by haemoglobin concentration. Identical data represented in Fig 1B. Bars represent means. Error bars represent 95% confidence intervals. Dashed line represents EI cut-off for anaemia detection of < 18.14. **** P < 0.0001; *** P < 0.001; *P < 0.05; One way ANOVA with Bonferroni’s multiple comparison test. (D) Palpebral conjunctival EI values calculated from iPhone 5S images (n = 94), grouped by haemoglobin concentration. Bars represent means. Error bars represent 95% confidence intervals. Dashed line represents EI cut-off of < 30.49 for anaemia detection. **** P < 0.0001; ** P < 0.01; One way ANOVA with Bonferroni’s multiple comparison test.
Table 1.
Comparison of palpebral conjunctival EI with clinician assessment.