Semi-automated digital measurement as the method of choice for beta cell mass analysis

Abstract

Pancreas injury by partial duct ligation (PDL) activates beta cell differentiation and proliferation in adult mouse pancreas but remains controversial regarding the anticipated increase in beta cell volume. Several reports unable to show beta cell volume augmentation in PDL pancreas used automated digital image analysis software. We hypothesized that fully automatic beta cell morphometry without manual micrograph artifact remediation introduces bias and therefore might be responsible for reported discrepancies and controversy. However, our present results prove that standard digital image processing with automatic thresholding is sufficiently robust albeit less sensitive and less adequate to demonstrate a significant increase in beta cell volume in PDL versus Sham-operated pancreas. We therefore conclude that other confounding factors such as quality of surgery, selection of samples based on relative abundance of the transcription factor Neurogenin 3 (Ngn3) and tissue processing give rise to inter-laboratory inconsistencies in beta cell volume quantification in PDL pancreas.

Fig 1. Fully automated analysis without manual data verification of insulin⁺ area results in both inclusion of noise and exclusion of weak positivity.

(a) Inclusion of false positive signal (“noise”, white arrowheads). (b) Exclusion of weak positivity (white arrows). Left: raw image; middle: ROI (red) with automated threshold (green); right: ROI (red) on raw image.

Fig 2. Depiction of manual correction of false-positive and false-negative signal.

(a) Manual correction of false-positive noise (white arrowheads). (b) Manual correction of false-negative weaker signal (compare red encircled areas on both panels). Left: automatic threshold ROI, right: manually corrected ROI. ROI Manager: the blue highlight indicates the depicted ROI: highest ROI value: automatic insulin area (ins auto), lowest ROI value: corrected insulin area (ins corr). The yellow boxes demonstrate the difference in plane area before (ins auto, 1) and after (ins, corr, 2) manual ROI correction: a ins auto: 0.056965 mm², ins corr: 0.065549 mm²; b ins auto: 0.037369 mm², ins corr: 0.106420 mm².

Fig 3. Beta cell volume quantification on non-corrected (Auto, circles) and corrected (Corrected, squares) insulin+ ROI’s.

Investigator remediation of micrograph artifacts is not required to demonstrate a significant increase in BCV in PDL (right, n = 12) vs. Sham (left, n = 10) pancreas (2X2 ANOVA shows significant effect of condition but not of method nor of the interaction condition x method—F ratio of model = 12.45, p < 0.0001; t ratio of "Condition" = -6.07, p < 0.0001).