Search

The purpose was to evaluate ex-vivo a prototype of a novel biopsy canula under CT fluoroscopy-guidance in ventilated porcine lung explants in respiratory motion simulations. Using an established chest phantom for porcine lung explants, n = 24 artificial lesions consisting of a fat-wax-Lipiodol mixture (approx. 70HU) were placed adjacent to sensible structures such as aorta, pericardium, diaphragm, bronchus and pulmonary artery. A piston pump connected to a reservoir beneath a flexible silicone reconstruction of a diaphragm simulated respiratory motion by rhythmic inflation and deflation of 1.5 L water. As biopsy device an 18-gauge prototype biopsy canula with a lancet-like, helically bended cutting edge was used. The artificial lesions were punctured under CT fluoroscopy-guidance (SOMATOM Sensation 64, Siemens, Erlangen, Germany; 30mAs/120 kV/5 mm slice thickness) implementing a dedicated protocol for CT fluoroscopy-guided lung biopsy. The mean-diameter of the artificial lesions was 8.3 +/- 2.6 mm, and the mean-distance of the phantom wall to the lesions was 54.1 +/- 13.5 mm. The mean-displacement of the lesions by respiratory motion was 14.1 +/- 4.0 mm. The mean-duration of CT fluoroscopy was 9.6 +/- 5.1 s. On a 4-point scale (1 = central; 2 = peripheral; 3 = marginal; 4 = off target), the mean-targeted precision was 1.9 +/- 0.9. No misplacement of the biopsy canula affecting adjacent structures could be detected. The novel steerable biopsy canula proved to be efficient in the ex-vivo set-up. The chest phantom enabling respiratory motion and the steerable biopsy canula offer a feasible ex-vivo system for evaluating and training CT fluoroscopy-guided lung biopsy adapted to respiratory motion.

Parallel imaging and echo sharing techniques have markedly reduced the acquisition times for MRI of large volumes. Dynamic 2 and 3-dimensional data sets of the chest with high temporal resolution (up to 10 images/s with single slice and 2 volume/s) allow an analysis of respiratory motion of the lungs and tumors. Time-resolved 2D series in preselected planes can be used to observe respiratory motion during free breathing or after respiratory commands, e.g. to exclude chest wall invasion by a tumor or for diagnosing impairment of respiratory mechanics. Time-resolved 3D-series (4D-MRI) allow monitoring of the spatial displacement of the lungs and tumors as a whole volume. Present limitations such as an overestimation of tumor size and an underestimation of displacement due to a limited temporal resolution are expected to be overcome with further technical developments. However, 4D-MRI already appears to be the appropriate tool to select patients for motion-adapted radiotherapy. In addition 4D-MRI is available for a broad spectrum of scientific applications, as it allows repeated and prolonged series of measurements without radiation exposure.

Magnetic resonance imaging of inhaled fluorinated inert gases ((19)F-MRI) such as sulfur hexafluoride (SF(6)) allows for analysis of ventilated air spaces. In this study, the possibility of using this technique to image lung function was assessed. For this, (19)F-MRI of inhaled SF(6) was compared with respiratory gas analysis, which is a global but reliable measure of alveolar gas fraction. Five anesthetized pigs underwent multiple-breath wash-in procedures with a gas mixture of 70% SF(6) and 30% oxygen. Two-dimensional (19)F-MRI and end-expiratory gas fraction analysis were performed after 4 to 24 inhaled breaths. Signal intensity of (19)F-MRI and end-expiratory SF(6) fraction were evaluated with respect to linear correlation and reproducibility. Time constants were estimated by both MRI and respiratory gas analysis data and compared for agreement. A good linear correlation between signal intensity and end-expiratory gas fraction was found (correlation coefficient 0.99+/-0.01). The data were reproducible (standard error of signal intensity 8% vs. that of gas fraction 5%) and the comparison of time constants yielded a sufficient agreement. According to the good linear correlation and the acceptable reproducibility, we suggest the (19)F-MRI to be a valuable tool for quantification of intrapulmonary SF(6) and hence lung function.

Purpose: Four-dimensional (4D) imaging is a key to motion-adapted radiotherapy of lung tumors. We evaluated in a ventilated ex vivo system how size and displacement of artificial pulmonary nodules are reproduced with helical 4D-CT, 4D-MRI, and linac-integrated cone beam CT (CBCT)., Methods and Materials: Four porcine lungs with 18 agarose nodules (mean diameters 1.3-1.9 cm), were ventilated inside a chest phantom at 8/min and subject to 4D-CT (collimation 24 x 1.2 mm, pitch 0.1, slice/increment 24 x 10(2)/1.5/0.8 mm, pitch 0.1, temporal resolution 0.5 s), 4D-MRI (echo-shared dynamic three-dimensional-flash; repetition/echo time 2.13/0.72 ms, voxel size 2.7 x 2.7 x 4.0 mm, temporal resolution 1.4 s) and linac-integrated 4D-CBCT (720 projections, 3-min rotation, temporal resolution approximately 1 s). Static CT without respiration served as control. Three observers recorded lesion size (RECIST-diameters x/y/z) and axial displacement. Interobserver- and interphase-variation coefficients (IO/IP VC) of measurements indicated reproducibility., Results: Mean x/y/z lesion diameters in cm were equal on static and dynamic CT (1.88/1.87; 1.30/1.39; 1.71/1.73; p > 0.05), but appeared larger on MRI and CBCT (2.06/1.95 [p < 0.05 vs. CT]; 1.47/1.28 [MRI vs. CT/CBCT p < 0.05]; 1.86/1.83 [CT vs. CBCT p < 0.05]). Interobserver-VC for lesion sizes were 2.54-4.47% (CT), 2.29-4.48% (4D-CT); 5.44-6.22% (MRI) and 4.86-6.97% (CBCT). Interphase-VC for lesion sizes ranged from 2.28% (4D-CT) to 10.0% (CBCT). Mean displacement in cm decreased from static CT (1.65) to 4D-CT (1.40), CBCT (1.23) and MRI (1.16)., Conclusions: Lesion sizes are exactly reproduced with 4D-CT but overestimated on 4D-MRI and CBCT with a larger variability due to limited temporal and spatial resolution. All 4D-modalities underestimate lesion displacement.

To compare the image quality of dynamic lung MRI with variations of steady-state free-precession (SSFP) and gradient echo (GRE) cine techniques at 1.5 T and 3 T. Ventilated porcine lungs with simulated lesions inside a chest phantom and four healthy human subjects were assessed with SSFP (TR/TE=2.9/1.22 ms; 3 ima/s) and GRE sequences (TR/TE=2.34/0.96 ms; 8 ima/s) as baseline at 1.5 and 3 T. Modified SSFPs were performed with nine to ten images/s (parallel imaging factors 2 and 3). Image quality for representative structures and artifacts was ranked by three observers independently. At 1.5 T, standard SSFP achieved the best image quality with superior spatial resolution and signal, but equal temporal resolution to GRE. SSFP with improved temporal resolution was ranked second best. Further acceleration (PI factor 3) was of no benefit, but increased artifacts. At 3 T, GRE outranged SSFP imaging with high lesion signal intensity, while artifacts on SSFP images increased visibly. At 1.5 T, a modified SSFP with moderate parallel imaging (PI factor 2) was considered the best compromise of temporal and spatial resolution. At 3 T, GRE sequences remain the best choice for dynamic lung MRI.

Chest wall invasion by a tumour and mediastinal masses are known to benefit from the superior soft tissue contrast of magnetic resonance imaging (MRI). However, helical computed tomography (CT) (i.e. with multiple row detector systems) remains the modality of choice to detect and follow lesions of the lung parenchyma. Since minimizing radiation exposure plays a minor role in oncologic patients, there are only few routine indications for which MRI of lung parenchyma is preferred to CT. This includes whole body MR imaging for staging or scientific studies with frequent follow-up examinations. MR-based lung imaging in this context was always considered as a weak point. Depending on the sequence technique and imaging conditions (i.e. ability to hold breath) the threshold for lung nodule detection with MRI using 1.5 T systems was estimated to be above 3-4 mm. The feasibility of lung MRI at 0.3-0.5 T and 3.0 T systems has been demonstrated. The clinical value of time-resolved lung nodule perfusion analysis cannot yet be determined, although the combination of perfusion characteristics with morphologic criteria contributes to estimate the integrity of a solitary lesion.

A method of four-dimensional (4D) magnetic resonance imaging (MRI) has been implemented and evaluated. It consists of retrospective sorting and slice stacking of two-dimensional (2D) images using an external signal for motion monitoring of the object to be imaged. The presented method aims to determine the tumour trajectories based on a signal that is appropriate for monitoring the movement of the target volume during radiotherapy such that the radiation delivery can be adapted to the movement. For evaluation of the 4D-MRI method, it has been applied to a dynamic lung phantom, which exhibits periodic respiratory movement of a porcine heart-lung explant with artificial pulmonary nodules. Anatomic changes of the lung phantom caused by respiratory motion have been quantified, revealing hysteresis. The results demonstrate the feasibility of the presented method of 4D-MRI. In particular, it enables the determination of trajectories of periodically moving objects with an uncertainty in the order of 1 mm.

Objectives: To compare the image quality and lesion contrast of lung MRI using 5 different pulse sequences at 1.5 T and 3 T., Materials and Methods: Lung MRI was performed at 1.5 T and 3 T using 5 pulse sequences which have been previously proposed for lung MRI: 3D volumetric interpolated breath-hold examination (VIBE), true fast imaging with steady-state precession (TrueFISP), half-Fourier single-shot turbo spin-echo (HASTE), short tau inversion recovery (STIR), T2-weighted turbo spin-echo (TSE). In addition to 4 healthy volunteers, 5 porcine lungs were examined in a dedicated chest phantom. Lung pathology (nodules and infiltrates) was simulated in the phantom by intrapulmonary and intrabronchial injections of agarose. CT was performed in the phantom for correlation. Image quality of the sequences was ranked in a side-by-side comparison by 3 blinded radiologists regarding the delineation of pulmonary and mediastinal anatomy, conspicuity of pulmonary nodules and infiltrates, and presence of artifacts. The contrast of nodules and infiltrates (CNODULES and CINFILTRATES) defined by the ratio of the signal intensities of the lesion and adjacent normal lung parenchyma was determined., Results: There were no relevant differences regarding the preference for the individual sequences between both field strengths. TSE was the preferred sequence for the visualization of the mediastinum at both field strengths. For the visualization of lung parenchyma the observers preferred TrueFISP in volunteers and TSE in the phantom studies. At both field strengths VIBE achieved the best rating for the depiction of nodules, whereas HASTE was rated best for the delineation of infiltrates. TrueFISP had the fewest artifacts in volunteers, whereas STIR showed the fewest artifacts in the phantom. For all but the TrueFISP sequence the lesion contrast increased from 1.5 T to 3 T. At both field strengths VIBE showed the highest CNODULES (6.6 and 7.1) and HASTE the highest CINFILTRATES (6.1 and 6.3)., Conclusion: The imaging characteristics of different pulse sequences used for lung MRI do not substantially differ between 1.5 T and 3 T. A higher lesion contrast can be expected at 3 T.