ORIGINALARBEIT
A method for improved 4D-computed tomography data acquisitionMethodik zur verbesserten Datenerfassung bei der 4D-Computertomographie für die stereotaktische ablative Strahlentherapie

https://doi.org/10.1016/j.zemedi.2016.05.001Get rights and content

Abstract

In four-dimensional time-dependent computed tomography (4D-CT) of the lungs, irregularities in breathing movements can cause errors in data acquisition, or even data loss. We present a method based on sending a synthetic, regular breathing signal to the CT instead of the real signal, which ensures 4D-CT data sets without data loss. Subsequent correction of the signal based on the real breathing curve enables an accurate reconstruction of the size and movement of the target volume. This makes it possible to plan radiation treatment based on the obtained data. The method was tested with dynamic thorax phantom measurements using synthetic and real breathing patterns.

Zusammenfassung

Störungen der regelmäßigen Atembewegung des Patienten bzw. der Patientin während der Durchführung einer zeitaufgelösten Computertomographie (4D-CT) der Lunge führen zu Fehlern in den generierten Bilddaten oder sogar zu Datenverlusten. Wird anstelle des realen Atemsignals ein künstlich erzeugtes, regelmäßiges Signal an den CT gesendet, kann eine datenverlustfreie Aufnahme garantiert werden. Nachträgliche Korrektur der Trigger, anhand der Auswertung der realen Atemkurve ermöglicht eine volumenkorrekte Rekonstruktion der Bilddaten, so dass eine Bestrahlungsplanung möglich ist. Die Methode wurde mithilfe eines dynamischen Thorax-Phantoms anhand von künstlichen und realen Atemkurven getestet.

Introduction

Radiation therapy is one of two treatment options for pulmonary carcinoma. Technological progress in this area has led to a significant improvement of results, and one key development has been the ability to model the target volume taking account of the movement of the tumour as the patient breathes, using computed tomography (CT) [1]. This involves generating time-dependent CT data sets which capture the breathing movement as the 4th dimension (4D-CT). Subsequent sorting of the data allows the reconstruction of individual breathing phases [2]. However, this type of 4D-CT image of the lungs is vulnerable to significant errors if the patient's breathing is not regular [2], [3], [4], [5], [6]. Irregularities in breathing have unpredictable effects on the captured data, independent of their magnitude [7]. It is possible to have significant irregularity of breathing throughout the recording period without loss of data; on the other hand a brief episode of coughing with otherwise regular breathing can cause a loss of data. If a loss of data occurs while the scan is passing a tumour region, information on the movement of the tumour is lost and the scan results are suboptimal or completely unusable for radiation therapy treatment planning [8]. Even if no data loss occurs, artefacts arising from irregular breathing can also cause errors in the measured volume of the target [9].

This paper presents a method for completely avoiding data loss and using the captured data to accurately reconstruct the target volume, so that the results can be used for treatment planning and so that repeat scans, which are time-consuming and expose the patient to additional radiation, can be avoided. It should be noted, that repeated scans are not a guarantee for a successful data acquisition. Further, with our method the functioning of the existing systems is not altered and the data processing and data path are not changed. Similar systems using an external gating system [10] should also benefit from this method independent of the CT vendor.

Section snippets

Conventional method

4D-CT acquisition works conventionally with real-time monitoring of breathing movements during the CT scan [3], [11] usually based on detecting periodic features such as a certain amplitude or a particular phase of the respiratory cycle. In case of the lungs a predictive signal at the peak of inspiration preferably in the phase based mode, is sent to the CT scanner, which continuously acquires multiple data at each image position with a low pitch predefined by the patient's breathing cycle.

Results

The average detected total volume of the test object was 7.14 cm3 (6.22–8.15 cm3) using the new method and 6.10 cm3 (5.17–7.60 cm3) using the conventional method (Table 1). The average variation of the total volumes relative to the theoretical value was ±0.59 cm3 in the new method and ±1.63 cm3 in the conventional method. In the reconstruction of the individual phases, average variations were ±0.40 cm3 with the new method and ±0.31 cm3 with the conventional method.

The scans made with the undisturbed

Discussion

One challenge in lung patient's treatment is the generation of 4D-CT images displaying the whole tumour movement. In the conventional method, used at our institution, a signal is sent to the scanner according to the maximum inspiration of the breathing curve so that the data reconstruction includes the breathing phase information. Since patient's breathing is irregular, possibly interrupted by coughs, the signal and the scan vary and are not in phase anymore. Once, this leads to a phase shift,

Conclusions

Our results show that the presented modified method performs clearly better at detecting target volumes than the method currently in clinical use, if irregular breathing occurs. The method should be seen as a temporary solution until CT manufacturers develop integrated systems that can handle irregular breathing movements better. It should be possible to use the modified method in clinical practice, because it complies with the restrictions that apply to medical devices.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

View full text