Computed tomography of the human pineal gland for study of the sleep–wake rhythm: reproducibility of a semi-quantitative approach

Abstract

To propose a semi-quantitative computed tomography (CT) protocol for determining uncalcified pineal tissue (UCPT), and to evaluate its reproducibility in modification of studies showing that the degree of calcification is a potential marker of deficient melatonin production and may prove an instability marker of circadian rhythm. Twenty-two pineal gland autopsy specimens were scanned in a skull phantom with different slice thickness twice and the uncalcified tissue visually assessed using a four-point scale. The maximum gland density was measured and its inverse graded on a non-linear four-point scale. The sum of both scores was multiplied by the gland volume to yield the UCPT. The within-subject variance of UCPT was determined and compared between scans of different slice thickness. The UCPT of the first measurement, in arbitrary units, was 39+/-52.5 for 1 mm slice thickness, 44+/-51.1 for 2 mm, 45+/-34.8 for 4 mm, and 84+/-58.0 for 8 mm. Significant differences of within-subject variance of UCPT were found between 1 and 4 mm, 1 and 8 mm, and 2 and 8 mm slice thicknesses (P<0.05). A superior reproducibility of the semi-quantitative CT determination of UCPT was found using 1 and 2 mm slice thicknesses. These data support the use of thin slices of 1 and 2 mm. The benefit in reproducibility from thin slices has to be carefully weighted against their considerably higher radiation exposure.