Post mortem temperature and its effect on quantitative magnetic resonance imaging

Post mortem magnetic resonance imaging (PMMRI) has the ability to enrich au- topsy results in forensic death investigations and to reveal otherwise undetectable findings due to its high soft tissue contrast. Additionally, PMMRI enables the validation and development of in vivo magnetic resonance imaging (MRI) by the possibility of subsequent histopathological examinations. Nevertheless, due to the cooling of the deceased, PMMRI of the intact body (so called in situ PMMRI) is limited by the temperature sensitivity of the MRI parameters. Therefore, in order to exploit the benefits of PMMRI, temperature correction is inevitable. Prior studies proposed temperature correction methods for PMMRI of the brain using the rectal temperature. However, it is known that the cooling of the body is inhomogenous after death. Thus, the goal of this thesis was to develop an accurate temperature correction model for PMMRI of the brain.
The relations between brain temperature and the in situ PMMRI relaxation param- eters T1, T2 and T∗2, as well as the diffusion parameters mean diffusivity (MD), and fractional anisotropy (FA) were investigated in the first study of this thesis (Publi- cation 1; status accepted). Significant linear relations have been found between the brain temperature and T1, T∗2, MD, and FA in gray matter, as well as T2, T∗2, and MD in white matter. The findings allow the correction of these MRI parameters for the brain temperature and, thus, the analysis of PMMRI independently of the deceased’s temperature. The second study (Publication 2; status accepted) exam- ined to which extent white matter fiber orientation dependent R∗2 depends on the brain temperature and differs between in vivo and post mortem in situ examinations. Decreased R∗2 fiber orientation dependency has been observed post mortem in situ compared to in vivo subjects, which may be attributed primarily to the deceased’s lower brain temperature. The third study (Publication 3; status accepted) revealed the relation between the brain temperature and null point inversion time (TInull) (time point at which the longitudinal magnetization of cerebral spinal fluid (CSF) is zero). A significant linear relation has been found between CSF’s TInull and the temperature. This allows the adaption of TInull for the temperature, leading to an optimal suppression of the CSF signal and, hence, enabling the application of the fluid attenuated inversion recovery (FLAIR) sequence post mortem. Nevertheless, brain temperature measurement is invasive and cannot be acquired in real-time dur- ing the MRI scan due to its required MRI incompatible temperature measurement method. Therefore, in the fourth study (Publication 4; status submitted) the post mortem temperature cooling of different body sites was investigated, in order to find a non invasive and real-time brain temperature model during the MRI scan. It has been found that the forehead temperature correlates linearly with the brain temperature. Thus, a temperature correction model for the MRI parameters using the forehead temperature has been examined (Publication 5; status submitted). A significant temperature sensitivity was found for T2 and MD in white matter, for T1 in cerebral cortex, as well as for T1 and MD in deep gray matter. This enables a real-time and non invasive temperature correction of the parameters taking into account temperature variations during the MRI scan.
As a conclusion, the findings allow the temperature correction in PMMRI, either based on the brain temperature or in real-time based on the forehead temperature.