In this study,
we directly correlated postmortem NmMRI and neuropathological findings.
On the postmortem NmMRI of the midbrain,
the signal intensity of the SNc was higher than that of the SC in the normal control case,
whereas the signal intensity of the SNc was as same as that of the SC in the DLB and PD cases.
These findings are same as findings with antemortem NmMRI [6].
Also,
the correlation between postmortem NmMRI and neuropathological findings revealed that the diffuse hyperintense areas in the SNc reflected well-preserved neuromelanin-containing neurons in the normal control case,
whereas the iso-intense area in the SNc showed severe loss of neuromelanin-containing neurons in the DLB and PD cases.
In the normal control case,
hyperintensity with multiple hypointense spots was seen in the SNc.
Histologically,
these hypointense areas reflected dilated perivascular spaces and bundles of myelinated fibers in the SNc.
These structures were also seen in cases with DLB and PD.
Decreased signal intensity reflecting bundles of myelinated fibers seems to be influenced by formalin fixation in postmortem T1-weighted images,
a finding that is not seen in antemortem T1-weighted images.
As to hypointense spot reflecting dilated perivascular spaces,
the signal intensity of hypointense spots reflects water in the dilated perivascular spaces.
This finding can also be seen on antemortem T1-weighted images.
Therefore,
preserved neuromelanin-containing neurons contribute to high signal intensity and dilated perivascular spaces contribute to loss of signal intensity in the SNc.
We evaluated whether iron deposition contributes to high signal intensity in NmMRI.
In all cases,
NmMRI of the midbrain showed hypointensity in the red nucleus and white matter such as the medial lemniscus and cerebral peduncle.
Furthermore,
diffuse ferritin deposition in these regions was stronger than in the SNc.
Therefore,
increased signal intensity in the SNc seems to not be influenced by iron deposition.
The correlation between the contrast signal ratio of each part of the SNc and the number of neuromelanin-containing neurons in the SNc was also evaluated.
High contrast signal ratios were seen in the normal control case,
and relatively low contrast signal ratios were seen in the DLB and PD cases.
These findings seemed to reflect neuropathological changes in the SNc.
Furthermore,
a significant positive correlation between the signal intensity and the number of neuromelanin-containing neurons was seen in all parts of the SNc,
showing that signal intensity of the SNc is closely related to the quantity of neuromelanin-containing neurons.
Limitations
There are several limitations to this study.
First,
we did not evaluate signal changes with iron imaging such as spin echo T2-weighted imaging,
T2*-weighted imaging,
and susceptibility-weighted imaging.
Instead,
iron deposition in tissue was directly evaluated using ferritin histochemistry.
Iron deposition apparently did not contribute to the high signal intensity on NmMRI.
Second,
this study involved a small number of subjects.
Third,
the LC was not evaluated in this study,
although this structure is usually involved in PD and DLB.
Despite these limitations,
our results show that signal intensity in the SNc is closely related to the quantity of neuromelanin-containing neurons as shown by the direct correlation between neuromelanin imaging and neuropathological findings.
As far as we know,
this is the first study directly comparing postmortem NmMRI and neuropathological findings.
According to the recent study by Kashihara et al.
[17,
18],
reduced signal changes are seen on NmMRI in the SNc not only in cases with PD,
but also in cases with multiple system atrophy,
progressive supranuclear palsy,
and corticobasal degeneration.
These findings correspond to each type of neuropathological changes.
Our results provide supporting evidence for these previous results.
Based on the direct correlation between NmMRI and neuropathological findings,
signal intensity in the substantia nigra is closely related to the quantity of neuromelanin-containing neurons and is apparently not influenced by iron deposition.