Parkinson’s: MRI differences in the locus coeruleus, without individual diagnostic validation

A review focused on an imaging measure

Neuromelanin-sensitive MRI can assess the integrity of the locus coeruleus in vivo. The researchers wanted to determine whether the differences observed between people with Parkinson’s and healthy controls were consistent across different cohorts and methods.

The review includes 65 reports, of which 17 contributed to at least one of the main pooled outcomes. The primary signal analysis is based on 14 independent cohorts, comprising 1,078 people with Parkinson’s and 543 healthy controls. Only the PubMed abstract is available in the source dossier, not the full text.

A lower average signal, but substantial variability

The locus coeruleus signal is lower on average in the Parkinson’s group. The standardized difference, expressed as Hedges’ g, is −0.95, with a 95% confidence interval from −1.32 to −0.57. This figure describes a difference between groups: it is neither a percentage of signal lost nor a probability of diagnosis.

A major limitation accompanies this result: between-study heterogeneity is substantial, with an I² of 88.4%. The 95% prediction interval, from −2.28 to 0.38, includes zero. In other words, the average difference does not guarantee that a difference in the same direction will be found in another study setting.

Sensitivity analyses retain a negative average difference in signal. However, all prediction intervals reported for these analyses also include zero.

Volume also differs, without resolving the diagnostic question

The locus coeruleus volume analysis pools five independent cohorts, with 215 people with Parkinson’s and 134 controls. It also reports lower values in Parkinson’s, with a Hedges’ g of −0.87 and a 95% confidence interval from −1.13 to −0.62. Estimated heterogeneity is zero in this analysis.

However, this result is not enough to establish individual diagnostic accuracy. The authors call for harmonized acquisition, analysis and reporting, as well as independent, multicenter, longitudinal validation before clinical translation.

What we can take away

This review supports the existence of group-level differences in locus coeruleus imaging. It does not validate this measure as a stand-alone test for diagnosing Parkinson’s in an individual.