British Journal of Radiology (2005) 78, 836-840
© 2005 British Institute of Radiology
doi: 10.1259/bjr/29483804
Quantitative analysis of lung and tumour mobility: comparison of two time-resolved MRI sequences
C Plathow, MD, MSc1,2,
M Klopp, MD3,
C Fink, MD1,
A Sandner, MD4,
H Hof, MD5,
M Puderbach, MD1,
F Herth, MD6,
A Schmähl, MD7 and
H-U Kauczor, MD1
1 Department of Radiology, German Cancer Research Center, Heidelberg, Germany, 2 Department of Diagnostic Radiology, Eberhard Karls University, Tübingen, Germany, 3 Department of Thoracic Surgery, 4 Department of Oncology, 6 Department of Pneumology, 7 Department of Diagnostic Radiology, Clinic for Thoracic Disease, Heidelberg, Germany and 5 Department of Radiotherapy, University of Heidelberg, Heidelberg, Germany
Correspondence: Christian Plathow, Department of Radiology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany
 |
Abstract
|
|---|
The purpose of this study was to describe the use of parallel imaging technique (PAT) using dynamic MRI in lung and tumour mobility during the breathing cycle. 20 patients with stage I non-small cell lung carcinoma were investigated using two dynamic gradient echo sequences with PAT (TrueFISP (fast imaging with steady precession), and fast low angle shot (FLASH). Craniocaudal distance from the apex to the diaphragm of the thorax and tumour mobility during the breathing cycle were measured. Signal-to-noise ratio (SNR) of the tumour was determined. In spite of the different temporal resolutions both trueFISP and FLASH sequence proved to be adequate to continuously measure lung motion and tumour mobility. SNR of the tumour was significantly higher using the trueFISP sequence than FLASH sequence (20.7±3.6 vs 5.8±2.3, p<0.01). Mobility of the tumour bearing hemithorax was significantly lower compared with the non-tumour bearing hemithorax (p<0.05). Dynamic MRI using PAT allows for continuous quantitative documentation of tumour mobility and lung motion. Because of the higher SNR, trueFISP sequence provides a better delineation of intrapulmonary lesions with a sufficient temporal resolution.
|
Introduction
|
|---|
To date, reduced respiratory mechanics are indirectly evaluated by clinical lung function tests. These techniques are capable of detecting global changes of lung mechanics. Early or localized pathological changes are hardly detectable by these techniques. Furthermore knowledge of local changes of lung mechanics, e.g. caused by the tumour and the tumour mobility itself are of high interest for a sufficient safety margin concept in high-precision radiotherapy techniques [1, 2].
Dynamic MRI has been proposed as an additional tool for the evaluation of lung mechanics [3] using a temporal resolution of one image s–1. However, this temporal resolution was not sufficient to continuously document the breathing cycle. Recently, it has been shown that dynamic MRI with a high temporal resolution of 3 images s–1 allows for the instantaneous assessment of respiratory motion of the lung [4] and intrapulmonary lesions [5]. Parallel MRI aquisition techniques (PAT), such as sensitivity encoding (SENSE) or simultaneous acquisition of spatial harmonics (SMASH) allow for a significant reduction of the sampled k-space data, thus allowing for an increased temporal resolution without trade-offs in spatial resolution and anatomic coverage [6–8]. In applications where sufficient signal-to-noise ratio (SNR) is available, the employment of these techniques often results in higher protocol flexibility. A faster acquisition speed can be converted into increased temporal resolution. In this regard, PAT might also offer the possibility to measure tumour mobility and lung motion during the breathing cycle with a temporal resolution above 3 images s–1.
The aim of this study was to evaluate two sequences with PAT employing the generalized autocalibrating partially parallel acquisition (GRAPPA [8]) algorithm: one sequence with high spatial resolution (trueFISP) and one with high temporal resolution (FLASH). These techniques were investigated in patients with tumours to detect regional differences in lung motion and tumour mobility during the breathing cycle.
|
Materials and methods
|
|---|
Patients
20 patients with a histologically proven solitary stage I non-small cell lung cancer (NSCLC) (13 males; 7 females; mean age 51 years old; range 46–71 years; tumour diameter: mean 3.5 cm; range 3.0–3.9 cm) were included in this study. Patients were in good clinical conditions and had no subjective problems with breathing in supine position. Examinations were performed at the time point of the first therapy control after single dose high-precision radiotherapy [1] (12±2 weeks after therapy). After the nature of the procedure had been fully explained, informed consent was signed by all participants under an independent institutionally approved subjects research protocol.
MRI
All examinations were performed using a 1.5 T whole-body scanner (Magnetom Symphony; Siemens Medical Solutions, Erlangen, Germany). For signal reception a combination of two spine array coils and two four-element body phased array coils was used. The imaging protocol included two different gradient echo pulse sequences (no echo-sharing technique) with GRAPPA [8] (Table 1
). Focus of the first sequence was a high spatial resolution with a sufficient (
3 images s–1) temporal resolution. Focus of the second sequence was high temporal resolution, with sufficient spatial resolution.
The first sequence was a fast imaging with steady precession (trueFISP) pulse sequence using the following imaging parameters: echo time/repetition time (TE/TR) 1.7/37.3 ms; flip angle 65°; receiver bandwidth 977 Hz pixel–1; acceleration factor 2; reference k-space lines for calibration 31; field of view (FOV) 375 x 400; matrix 112 x 192; slice thickness 10 mm [4]). Three frames per second (scan time per frame 0.32 s) were acquired within an acquisition time of 30 s. Furthermore a fast low angle shot (FLASH) sequence with the following image parameters was used: TE/TR 1.03/2.49 ms; flip angle 5°; receiver bandwidth 980 Hz pixel–1; acceleration factor 2; reference k-space lines for calibration: 24; FOV 480 x 350–404 mm; matrix 78 x 128; slice thickness: 10 mm). 10 frames per second (scan time per frame: 0.10 s) were acquired within an acquisition time of 30 s.
Using both techniques lung motion and tumour mobility were measured and compared using time–distance curves. All patients were instructed to change from quiet tidal breathing to maximal inspiration followed by maximal expiration with as much effort as possible [4] (Figures 1 and 2). This procedure was rehearsed several times to ensure constant conditions.
View larger version (69K):
[in this window]
[in a new window]
|
Figure 1. Continuous measurement of craniocaudal (CC) displacement of the thorax from deep inspiration to deep expiration using a trueFISP (3 images s–1, images above the time line) and a FLASH-sequence (10 images s–1, images below the time line) in a patient with a solitary stage I non-small cell lung cancer (NSCLC) within the left hemithorax. A significant decrease of the motion of the left hemithorax in comparison with the right side is demonstrated. Using the trueFISP sequence the tumourous areas exhibit a higher signal to noise ratio (SNR).
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Figure 2. (a) Craniocaudal (CC) displacement of the non-tumour bearing hemithorax from deep inspiration to deep expiration, comparison between trueFISP and FLASH sequence. (b) Complete breathing cycle from quiet tidal breathing followed by maximum inspiration and expiration using a trueFISP sequence. Non-tumour bearing hemithorax (–tumour) is compared with the tumour bearing hemithorax (+tumour).
|
|
For an intraindividual quantitative comparison of both MR sequences – even though sequences with parallel imaging techniques have been applied – SNR was calculated based on signal intensity measurements in circular regions-of-interest (ROI, minimum diameter of 6 mm) positioned in the centre of lung tumours. Absolute signal intensity (SI) measurements were related to noise, which was defined as the standard deviation of SI measurements of a ROI placed outside the body. SNR were calculated with the following equation: SNR=SItumour/noise.
Measurement of tumour mobility and lung motion
For tumour mobility examination coronal planes through the centre of the tumour were acquired in each patient. Mobility of the tumour was measured in the coronal plane (craniocaudal (CC) displacement) from its external edge. CC displacement was measured from the T (thoracic) 6/T7 disc space to the proximal external tumour edge [9]. If tumour mobility proved to be >10 mm additional slices were acquired [5]. Time–distance curves were documented and measured (Figure 2
).
For lung motion examination coronal planes crossing the trachea were acquired in each patient [11]. Motion of the lung was measured from the apex of the lung to the diaphragmatic dome in the CC plane in the middle of the hemithorax. Measurement in the mediolateral and anteroposterior direction was not performed as major lung motion and tumour mobility is in the CC direction [4, 5] and main focus of the study was the comparison of two MRI sequence techniques in this field. Tumour bearing and non-tumour bearing hemithoraces were measured separately. Time–distance curves were documented and measured. Because of the interindividual differences of the duration of the inspiration and expiration plateaus bars were inserted in the figures (Figure 2). Maximum displacement was defined as maximum difference between deep inspiration and expiration using MRI (Figure 3).
View larger version (14K):
[in this window]
[in a new window]
|
Figure 3. Maximum craniocaudal (CC) displacement of the non-tumour bearing hemithorax (–tumour), tumour bearing hemithorax (+tumour) and the tumour. Comparison between trueFISP and FLASH sequence (* p<0.01).
|
|
Statistical analysis
The statistical analysis of the data was performed with SAS software (SAS Institute Inc., Cary, NC). The mean and standard deviation were calculated. In addition, a t-test for paired samples was performed. Correlation was investigated by Spearman's correlation. p-values less than 0.05 were considered statistically significant, p-values less than 0.01 were considered statistically highly significant.
|
Results
|
|---|
Parallel MRI techniques using a trueFISP and FLASH sequence showed regular synchronous diaphragm and chest wall motions with diagnostic quality.
With the defined sequence parameters quantitative analysis revealed significantly higher SNR values of the tumour using the trueFISP sequence (20.7±3.6) compared with the FLASH sequence (5.8±2.3). Using GRAPPA with an acceleration factor of 2 the temporal resolution could be increased by the factor 1.34 (trueFISP) and 1.6 (FLASH, Table 2). The acquisition of three images per second using a trueFISP sequence allowed for continuous recording during the breathing cycle and correlated highly significant with the acquisition of 10 images per second using a FLASH sequence (Table 2).
Tumour mobility and lung motion using trueFISP and FLASH sequence
Continuous measurement of CC displacement of the tumour position from deep inspiration followed by deep expiration is shown in Figure 1. Using the trueFISP sequence maximum tumour mobility was 1.5±0.5 cm, using the FLASH sequence measurements 1.6±0.7 cm (Figure 3). The difference between these two MRI techniques was not significant. Correlation of time distance curves was high (r=0.93, p<0.01).
Measuring CC distance from the apex to the diaphragmatic dome of the tumour and non-tumour bearing hemithoraces revealed a significant correlation between both techniques (r=0.89, p<0.01). Using the trueFISP sequence maximum CC distance of the tumour bearing hemithorax was 3.9±0.5 cm, using FLASH sequence 3.7±0.6 cm. In the non-tumour bearing hemithorax distance was 5.6±0.5 cm using the trueFISP sequence and 5.6±0.6 cm using the FLASH sequence. Difference between tumour bearing and non-tumour bearing hemithorax was highly significant (p<0.01).
Using trueFISP sequence there was a substantially lower standard deviation while measuring tumour mobility and CC distance compared with FLASH sequence (e.g. Figure 2a).
|
Discussion
|
|---|
The application of PAT enabled an increased temporal resolution of dynamic MRI of tumour mobility and lung motion. Motion can be quantified continuously in quiet respiration as well as in deep inspiration and expiration.
In the past, quantification of tumour mobility and lung motion from MRI data has been hampered by various factors. The main problem was the insufficient temporal resolution, leading to a non-continuous documentation of the organ motion during the breathing cycle.
Previous studies have demonstrated the feasibility of quantification of lung motion using a FLASH sequence with a temporal resolution of only one image per second. Suga et al [3] applied this sequence in six healthy volunteers and 28 patients with pulmonary emphysema to measure maximum inspiration and expiration. This technique was able to encompass the whole thorax and assess impaired respiratory mechanics directly with results similar to our study. But in order to be clinically accepted, the measurement of tumour mobility and lung motion by MRI will require high spatial and temporal resolution in order to continuously detect changes of the regional mobility for a precise integration, e.g. into radiotherapy planning [5].
The trueFISP technique became available in recent years because of significant hardware improvements. It can now be applied for a variety of clinical applications due to distinct improvements in SNR [10–12]. It has also been proposed for the detection of pericardial effusions and nodules of the lung at low field [13]. Recently this sequence was investigated to measure lung motion in correlation with spirometry [4]. Schäfer et al discouraged the use of trueFISP sequence for lung imaging at low field MRI because the SNR of an intrapulmonary tumour was lower than with a FLASH sequence [14]. Using a high-field MR scanner and different sequence parameters, however intrapulmonary tumours were precisely investigable using a dynamic trueFISP sequence [5, 15]. In our current study the significantly higher SNR using the trueFISP sequence compared with the FLASH sequence and the lower standard deviation confirms the latter observation.
The implementation of PAT enables a decrease of the data acquisition time. The acceleration in data acquisition can be employed to either reduce the overall acquisition time or to acquire more data to increase the temporal resolution. A major drawback of PAT imaging, however, is a relevant reduction of the achievable SNR [16]. Consequently, with the present hardware configuration the acceleration factor could not be increased to more than n=2 without compromising tumour visualization. In this study, PAT was used to reduce the scan time per acquisition by the factor of 1.6 and 1.34 for a FLASH and a trueFISP sequence, respectively. Thus, temporal resolution was substantially improved compared with previous studies of tumour mobility and lung motion MRI. Both techniques were significantly correlated and a temporal resolution of 3 images s–1 proved to be sufficient for a continuous and precise documentation of tumour mobility during the breathing cycle. Recently it has been shown that in healthy adults there is no difference between the right and left hemithorax [4]. In this study patients with a tumour showed a significant decrease of the motion of the tumour bearing hemithorax. Whether this decrease might be caused by the tumour itself, the therapy and/or a reduced compliance has to be shown in further research. But this technique proved to be sensitive enough even to detect small changes in lung mechanics. It might be argued that a slice thickness of 10 mm is insufficient for accurate assessment of tumour mobility. But, as recently published [5], such a slice thickness is adequate in the measurement of tumour mobility, at least in tumours with a substantial diameter (here >3 cm). Additional slices were also performed if tumour mobility proved to be more than 10 mm, thus the applied slice thickness was no problem in this study.
In a comprehensive investigation of small intrapulmonary lesions and lesions with surrounding lung pathologies during the breathing cycle the trueFISP sequence seems to be preferable. The high SNR also offers better conditions to further increase PAT acceleration factors with a still sufficient SNR. Thus in potential integrations of tumour mobility, e.g. in radiotherapy planning, this sequence seems to be more suitable than the FLASH sequence. Whether the use of a FLASH sequence with its better temporal resolution leads to significantly higher precision in forced breathing manoeuvres [17] has to be investigated further. Additionally, the better temporal resolution can also be used for three-dimensional (3D) measurements of 3D lung motion as recently has been shown [18].
A limitation of our study is the small number of included patients. But as patients were compared intraindividually using the two different techniques with similar results we propose this technique to be adequate to detect local tissue changes during the breathing cycle. Whether both sequences remain almost equivalent when it comes to subtle changes in different diseases (e.g. COPD) has to be investigated in further studies.
|
Conclusion
|
|---|
Parallel MRI allows for a quantitative assessment of tumour mobility and local lung motion. Future studies will have to assess the clinical value of this information for diagnosis, therapy planning and management of pulmonary diseases.
Received for publication October 10, 2004.
Revision received March 1, 2005.
Accepted for publication April 18, 2005.
|
References
|
|---|
- Hof H, Herfarth KK, Munter M, Hoess A, Motsch J, Wannenmacher M, et al. Stereotactic single-dose radiotherapy of stage I non-small-cell lung cancer (NSCLC). Int J Radiat Oncol Biol Phys 2003;56:335–41.[CrossRef][Medline]
- Gold W. Pulmonary function testing. In: Murray JF, Nadel JA (editors). Textbook of respiratory medicine (3rd edn). Philadelphia, PA: WB Saunders, 2000;783–804.
- Suga K, Tsukuda T, Awaya H, Takano K, Koike S, Matsunaga N, et al. Impaired respiratory mechanics in pulmonary emphysema: evaluation with dynamic breathing MRI. J Magn Reson Imaging 1999;10:510–20.[CrossRef][Medline]
- Plathow C, Ley S, Fink C, Puderbach M, Heilmann M, Zuna I, et al. Evaluation of chest motion and volumetry during the breathing cycle by dynamic MRI in healthy subjects: comparison with pulmonary function tests. Invest Radiol 2004;39:202–9.[CrossRef][Medline]
- Plathow C, Ley S, Fink C, Puderbach M, Hosch W, Schmähl A, et al. Analysis of intrathoracic tumor mobility during the whole breathing cycle by dynamic MRI. Int J Radiat Oncol Biol Phys 2004;59:943–51.[CrossRef][Medline]
- Sodicksen DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays. Magn Reson Med 1997;38:591–603.[Medline]
- Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: sensitivity encoding for fast MRI. Magn Reson Med 1999;42:952–62.[CrossRef][Medline]
- Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially parallel acquisition (GRAPPA). Magn Reson Med 2002;47:1202–10.[CrossRef][Medline]
- Giraud P, De Rycke Y, Dubray B, Helfre S, Voican D, Guo L, et al. Conformal radiotherapy (CRT) planning for lung cancer: analysis of intrathoracic organ motion during extreme phases of breathing. Int J Radiat Oncol Biol Phys 2001;51:1081–92.[CrossRef][Medline]
- Barkhausen J, Goyen M, Ruhm SG, Eggebrecht H, Debatin JF, Ladd ME. Assessment of ventricular function with single breath-hold real-time steady-state free precession cine MR imaging. AJR Am J Roentgenol 2002;178:731–5.[Abstract/Free Full Text]
- Barkhausen J, Goyen M, von Winterfeld F, Lauenstein T, Arweiler-Harbeck D, Debatin JF. Visualization of swallowing using real-time TrueFISP MR fluoroscopy. Eur Radiol 2002;12:129–33.[CrossRef][Medline]
- Quick HH, Ladd ME, Hoevel M, Bosk S, Debatin JF, Laub G, et al. Real-time MRI of joint movement with trueFISP. J Magn Reson Imaging 2002;15:710–5.[CrossRef][Medline]
- Heussel CP, Sandner A, Voigtlander T, Heike M, Deimling M, Kuth R, et al. Prospective feasibility study of chest X-ray vs. thoracic MRI in breath-hold technique at an open low-field scanner. Fortschr Rontgenstr 2002;174:854–61.[CrossRef]
- Schäfer JF, Vollmar J, Schick F, Seemann MD, Mehnert F, Vontheim R, et al. Imaging diagnosis of solitary pulmonary nodules on an open low-field MRI system—comparison of two MR sequences with spiral CT. Fortschr Röntgenstr 2002;174:1107–14.[CrossRef]
- Plathow C, Fink C, Ley S, Puderbach M, Eichinger M, Zuna I, et al. Measurement of tumor-diameter dependent mobility of lung tumors by dynamic MRI. Radiother Oncol 2004;73:349–54.[Medline]
- Madore B, Pelc NJ. SMASH and SENSE: experimental and numerical comparisons. Magn Reson Med 2001;45:1103–11.[CrossRef][Medline]
- Plathow C, Fink C, Sandner M, Ley S, Puderbach M, Eichinger M, et al. Comparison of relative forced expiratory volume of one second with dynamic magnetic resonance imaging parameters in healthy subjects and patients with lung cancer. J Magn Reson Imaging 2005;21:212–8.[Medline]
- Plathow C, Schoebinger M, Fink C, Ley S, Puderbach M, Eichinger M, et al. Evaluation of lung volumetry using dynamic three-dimensional magnetic resonance imaging. Invest Radiol 2005;40:173–9.[CrossRef][Medline]
Top
Abstract
Introduction
