# Allometry correction in semi-quantitative analysis of dopamine transporter SPECT

## TL;DR

This study shows that correcting for brain size differences improves the accuracy of dopamine transporter SPECT scans in diagnosing striatal signal changes.

## Contribution

A novel allometry correction method is proposed to enhance semi-quantitative SBR analysis in dopamine transporter SPECT.

## Key findings

- Allometry correction significantly reduced the correlation between putamen SBR and brain volume scaling.
- Allometry correction increased the effect size between reduced and normal striatal signal groups.
- Diagnostic accuracy improved from 94.4% to 96.2% with allometry correction.

## Abstract

Striatal specific binding ratios (SBR) are widely used to support the interpretation of dopamine transporter SPECT scans. Automatic SBR computation often involves using affine transformations to map the individual SPECT images to an anatomical reference space for ROI analysis using predefined standard masks. This does not account for differences in volumetric scaling between brain structures since, by definition, affine transformations preserve volume ratios. However, striatal volume has been reported to scale proportional to (intracranial volume)0.4, indicating particularly pronounced “negative” allometric scaling. This study aimed to investigate the impact of disregarding allometric scaling on putamen SBR, and to propose an easy-to-implement method to avoid this issue. 656 [123I]FP-CIT SPECT (67.2 ± 11.4y, 44.2% females, 52.1/47.9% with reduced/normal striatal signal according to visual interpretation by an expert reader) were spatially normalized by affine transformation to the anatomical reference space of the Montreal Neurological Institute. Unilateral putamen SBR were estimated using hottest voxels analysis in large putamen masks predefined in the reference space. For conventional hottest voxels analysis, the number of hottest voxels was fixed at 1,250 (equivalent to 10 ml, the mean putamen volume in healthy adults). For allometry correction, the number of hottest voxels was determined separately for each subject as 1,250*DET(1−0.4), where DET is the Jacobian determinant of the affine transformation. To characterize the impact of the allometry correction on diagnostic performance, a data-driven Gaussian mixture model was employed.

Linear regression in the visually normal SPECT revealed a positive correlation between uncorrected putamen SBR and DET (Pearson’s R = 0.509, 95%-CI 0.422–0.587). The correlation was significantly (p < 0.00005) weaker when allometry-corrected ROI analysis was used (R = 0.285, 95%-CI 0.180–0.383). The effect size of the distance between reduced and normal putamen SBR as determined by the Gaussian mixture model was significantly larger with than without allometry correction (3.979 versus 3.335, one-sided p < 0.0001). When using the visual expert reading as diagnostic reference standard, the overall accuracy of the dichotomized SBR was significantly improved by allometry correction (from 94.4% to 96.2%, one-sided p = 0.026).

The diagnostic performance of semi-quantitative SBR analyses involving affine transformations to an anatomical reference space can be enhanced through the application of the proposed allometry correction.