Refine
Has Fulltext
- yes (2)
Is part of the Bibliography
- yes (2)
Year of publication
- 2021 (2) (remove)
Document Type
- Doctoral Thesis (2)
Keywords
- Lunge (2) (remove)
Hintergrund: Die pulmonale elektrische Impedanztomographie (EIT) stellt die regionale Ventilation der Lunge dar. Ziel dieser Studie war es, die postoperative Umverteilung der regionalen Ventilation nach lungenchirurgischen Eingriffen in Abhängigkeit von der Operationsseite und deren Zusammenhang mit der forcierten Vitalkapazität zu untersuchen.
Methoden: In dieser Beobachtungsstudie wurden Patienten untersucht, die sich rechts- und linksseitig offenen und videothorakoskopischen Eingriffen unterzogen. Es wurden Messungen mit EIT und Spirometrie durchgeführt. Jeweils 13 links- und 13 rechtsseitig operierte erwachsene Patienten, wurden präoperativ, sowie am dritten postoperativen Tag untersucht. Der Center of Ventilation (COV) innerhalb einer 32 x 32 Matrix wurde aus den EIT-Daten berechnet. COVx (links/rechts) wurde zu COVx' (ipsilateral/kontralateral) mit invertiertem Vektor der perioperativen Veränderung bei rechtsseitiger Operation modifiziert, um den Effekt der Zeit und der Seite der Operation in beiden Gruppen mit Two-way ANOVA für wiederholte Messungen zu testen.
Ergebnisse: Die perioperative Verschiebung von COVx' war abhängig von der Operationsseite (p=0,007) und war um -1,97 Matrixpunkte (p<0,001) bei rechtsseitiger und -0,61 Matrixpunkte (p=0,425) bei linksseitiger Operation von der Operationsseite weg gerichtet. Die forcierte Vitalkapazität (%vom Soll) verringerte sich von 94 (83-109) % [Median (Perzentile)] (linksseitig) und 89 (80-97) % (rechtsseitig) auf 61 (59-66) % bzw. 62 (40-72) % (p<0,05) und korrelierte mit COVx' für beide Gruppen.
Schlussfolgerung: Nur nach rechtsseitiger Lungenoperation konnte die EIT eine reduzierte Ventilation auf der Seite der Operation zeigen, während die Vitalkapazität in beiden Gruppen deutlich reduziert war.
In summary, the wave-CAIPI k-space trajectory presents an efficient sampling strategy for accelerated MR acquisitions. Using wave-CAIPI in parallel imaging reconstructions leads to a reduced noise level in the reconstructed images, compared to the Cartesian standard trajectory. This effect could be quantified by means of noise and SNR calculations. An SNR gain can be traded for a reduced scan time, i.e., additional undersampling, or for an enhanced image quality, keeping scan time constant.
Acceleration of MR imaging is especially important in dynamic applications, since these examinations are inherently time-consuming. The impact of wave-CAIPI sampling on image quality and its potential for scan time reduction was investigated for two dynamic applications: self-gated dynamic 3D lung MRI during free breathing and cardiac 4D flow MRI.
Dynamic 3D Lung MRI
By employing wave-CAIPI sampling in self-gated, free-breathing dynamic 3D lung MRI for the purpose of radiotherapy treatment planning, the image quality of accelerated scans could be enhanced. Volunteer examinations were used to quantify image quality by means of similarity between accelerated and reference images. To this end, the normalized mutual information and the root-mean-square error were chosen as quantitative image similarity measures.
The wave-CAIPI sampling was shown to exhibit superior quality, especially for short scan times. The values of the normalized mutual information were (10.2 +- 7.3)% higher in the wave-CAIPI case -- the root-mean-square error was (18.9 +- 13.2)% lower on average. SNR calculations suggest an average SNR benefit of around 14% for the wave-CAIPI, compared to Cartesian sampling.
Resolution of the lung in 8 breathing states can be achieved in only 2 minutes. By using the wave-CAIPI k-space trajectory, precise tumor delineation and assessment of respiration-induced displacement is facilitated.
Cardiac 4D Flow MRI
In 4D flow MRI, acceleration of the image acquisition is essential to incorporate the corresponding scan protocols into clinical routine. In this work, a retrospective 6-fold acceleration of the image acquisition was realized. Cartesian and wave-CAIPI 4D flow examinations of healthy volunteers were used to quantify uncertainties in flow parameters for the respective sampling schemes.
By employing wave-CAIPI sampling, the estimated errors in flow parameters in 6-fold accelerated scans could be reduced by up to 55%. Noise calculations showed that the noise level in 6-fold accelerated 4D flow acquisitions with wave-CAIPI is 43% lower, compared to Cartesian sampling. Comparisons between Cartesian and wave-CAIPI 4D flow examinations with a prospective acceleration factor R=2 revealed small, but partly statistically significant discrepancies. Differences between 2-fold and 6-fold accelerated wave-CAIPI scans are comparable to the differences between Cartesian and wave-CAIPI examinations at R=2.
Wave-CAIPI 4D flow acquisitions of the aorta could be performed with an average, simulated scan time of under 4 minutes, with reduced uncertainties in flow parameters. Important visualizations of hemodynamic flow patterns in the aorta were only slightly affected by undersampling in the wave-CAIPI case, whereas for Cartesian sampling, considerable discrepancies were observed.