Stefan Hindel1, Anika Söhner, Marc Maa, Wolfgang Sauerwein, Hideo Andreas Baba, Martin Kramer, Lutz Lüdemann. 1. From the *Department of Radiotherapy, Medical Physics, University Hospital Essen, Essen, North Rhine-Westphalia; †Department of General and Visceral Surgery at Wesel Protestant Hospital, Wesel; ‡Department of Pathology, University Hospital Essen, Essen, North Rhine-Westphalia; §Hospital of Veterinary Medicine, Department of Small Animal Surgery, Justus Liebig University Giessen, Giessen, Hesse, Germany.
Abstract
OBJECTIVES: The aim of our study was to assess the accuracy of fractional interstitial volume determination in low perfused and low vascularized tissue by using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). MATERIALS AND METHODS: The fractional interstitial volume (ve) was determined in the medial thigh muscle of 12 female pigs by using a 3-dimensional gradient echo sequence with k-space sharing and administering gadolinium-based contrast agent (gadoterate meglumine). Analysis was performed using 3 pharmacokinetic models: the simple Tofts model (TM), the extended TM (ETM), and the 2-compartment exchange model (2CXM). We investigated the effect of varying acquisition durations (ADs) on the model parameter estimates of the 3 models and compared the ve values with the results of histological examinations of muscle sections of the medial thigh muscle. RESULTS: Histological measurements yielded a median value (25%-75% quartile) of 4.8% (3.7%-6.2%) for ve. The interstitial fractional volume determined by DCE-MRI was comparable to the histological results but varied strongly with AD for the TM and ETM. For the TM and the ETM, the results were virtually the same. Choosing arterial hematocrit to Hcta = 0.4, the lowest median ve value determined by DCE-MRI was 5.2% (3.3%-6.1%) for the ETM at a 6-minute AD. The maximum ve value determined with the ETM at a 15-minute AD was 7.7% (4.5%-9.0%). The variation with AD of median ve values obtained with the 2CXM was much smaller: 6.2% (3.1%-9.2%) for the 6-minute AD and 6.3% (4.3%-9.8%) for the 15-minute AD. The best fit for the 2CXM was found at the 10-minute AD with ve values of 6.6% (3.7%-8.2%). No significant correlation between the histological and any DCE-MRI modeling results was found. Considering the expected accuracy of histological measurements, the medians of the MR modeling results were in good agreement with the histological prediction. A parameter determination uncertainty was identified with the use of the TMs. This is due to underfitting and has a major effect even on the analysis of tissues with low vascularization and low perfusion, where the estimated ve values depend on the AD. For the TM and ETM, the results best matched the histological measurements for an AD of 6 minutes. CONCLUSIONS: Owing to more fitting parameters, the 2CXM yielded better fits and the median interstitium-to-plasma rate constant kep was less depending on the AD; however, the uncertainty expressed by the 25% to 75% quartile range was found to be larger. An AD of 10 minutes was needed for the 2CXM to achieve accuracy comparable to those of the TMs with shorter ADs.
OBJECTIVES: The aim of our study was to assess the accuracy of fractional interstitial volume determination in low perfused and low vascularized tissue by using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). MATERIALS AND METHODS: The fractional interstitial volume (ve) was determined in the medial thigh muscle of 12 female pigs by using a 3-dimensional gradient echo sequence with k-space sharing and administering gadolinium-based contrast agent (gadoterate meglumine). Analysis was performed using 3 pharmacokinetic models: the simple Tofts model (TM), the extended TM (ETM), and the 2-compartment exchange model (2CXM). We investigated the effect of varying acquisition durations (ADs) on the model parameter estimates of the 3 models and compared the ve values with the results of histological examinations of muscle sections of the medial thigh muscle. RESULTS: Histological measurements yielded a median value (25%-75% quartile) of 4.8% (3.7%-6.2%) for ve. The interstitial fractional volume determined by DCE-MRI was comparable to the histological results but varied strongly with AD for the TM and ETM. For the TM and the ETM, the results were virtually the same. Choosing arterial hematocrit to Hcta = 0.4, the lowest median ve value determined by DCE-MRI was 5.2% (3.3%-6.1%) for the ETM at a 6-minute AD. The maximum ve value determined with the ETM at a 15-minute AD was 7.7% (4.5%-9.0%). The variation with AD of median ve values obtained with the 2CXM was much smaller: 6.2% (3.1%-9.2%) for the 6-minute AD and 6.3% (4.3%-9.8%) for the 15-minute AD. The best fit for the 2CXM was found at the 10-minute AD with ve values of 6.6% (3.7%-8.2%). No significant correlation between the histological and any DCE-MRI modeling results was found. Considering the expected accuracy of histological measurements, the medians of the MR modeling results were in good agreement with the histological prediction. A parameter determination uncertainty was identified with the use of the TMs. This is due to underfitting and has a major effect even on the analysis of tissues with low vascularization and low perfusion, where the estimated ve values depend on the AD. For the TM and ETM, the results best matched the histological measurements for an AD of 6 minutes. CONCLUSIONS: Owing to more fitting parameters, the 2CXM yielded better fits and the median interstitium-to-plasma rate constant kep was less depending on the AD; however, the uncertainty expressed by the 25% to 75% quartile range was found to be larger. An AD of 10 minutes was needed for the 2CXM to achieve accuracy comparable to those of the TMs with shorter ADs.
Authors: Stefan Hindel; Anika Söhner; Marc Maaß; Wolfgang Sauerwein; Dorothe Möllmann; Hideo Andreas Baba; Martin Kramer; Lutz Lüdemann Journal: PLoS One Date: 2017-01-31 Impact factor: 3.240