A temperature-controlled cooling system for accurate quantitative post-mortem MRI

PurposeTo develop a temperature-controlled cooling system to facilitate accurate quantitative post-mortem MRI and enable scanning of unfixed tissue. MethodsA water cooling system was built and integrated with a 7T scanner to minimize temperature drift during MRI scans. The system was optimized for operational convenience and rapid deployment to ensure efficient workflow, which is critical for scanning unfixed post-mortem samples. The performance of the system was evaluated using a 7-h diffusion MRI protocol at 7T with a porcine tissue sample. Quantitative T-1, T-2, and ADC maps were interspersed with the diffusion scans at seven different time points to investigate the temperature dependence of MRI tissue parameters. The impact of temperature changes on biophysical model fitting of diffusion MRI data was investigated using simulation. ResultsTissue T-1, T-2, and ADC values remained stable throughout the diffusion MRI scan using the developed cooling system, but varied substantially using a conventional scan setup without temperature control. The cooling system enabled accurate estimation of biophysical model parameters by stabilizing the tissue temperature throughout the diffusion scan, while the conventional setup showed evidence of significantly biased estimation. ConclusionA temperature-controlled cooling system was developed to tackle the challenge of heating in post-mortem imaging, which shows potential to improve the accuracy and reliability of quantitative post-mortem imaging and enables long scans of unfixed tissue.

Automated summary

A temperature-controlled cooling system was developed to facilitate accurate quantitative post-mortem MRI and enable scanning of unfixed tissue. The system was optimized for operational convenience and rapid deployment to ensure efficient workflow. The results showed that the developed cooling system stabilized tissue temperature and enabled accurate estimation of biophysical model parameters, while the conventional setup showed biased estimation.