Elsevier

Clinical Oncology

Volume 27, Issue 4, April 2015, Pages 197-204
Clinical Oncology

Original Article
Predictors of Respiratory-induced Lung Tumour Motion Measured on Four-dimensional Computed Tomography

https://doi.org/10.1016/j.clon.2014.12.001Get rights and content

Highlights

  • Lower lobe tumour motion in the superior–inferior direction is greater than upper and other lobes.

  • Lower zone tumour motion in the superior–inferior direction is greater than mid- and upper zones.

  • Tumour volume, T-stage and forced expiratory ratio do not predict tumour motion.

Abstract

Aims

The delivery of radical radiotherapy in lung cancer is complicated by respiratory-induced tumour motion. The aim of the study was to correlate tumour motion characteristics with tumour and patient factors, particularly the anatomical lobe and pulmonary zone.

Materials and methods

Lung tumour volumes on four-dimensional computed tomography were delineated by a single observer at maximal expiration and propagated through all 10 phases of the breathing cycle. Movements were tracked in the superior–inferior (SI), anterior–posterior (AP) and medio-lateral (ML) directions by changes in the tumour centroid coordinates. Tumour motion characteristics were correlated with anatomical lobe, pulmonary zone, tumour volume, T-stage, smoking status and spirometry.

Results

In 101 consecutive patients, the median magnitude of tumour motion in the SI direction was significantly larger in tumours located in lower lobes compared with upper lobes and middle/lingular lobes (0.70 cm versus 0.09 cm versus 0.26 cm, P < 0.01). No significant difference was found in median tumour motion between lower, upper and middle/lingular lobes in the AP (0.16 cm versus 0.13 cm versus 0.16 cm, P = 0.45) and ML (0.08 cm versus 0.08 cm versus 0.13 cm, P = 0.32) directions, respectively. When assessed by zone, the median tumour displacement in the SI direction was significantly larger in the lower zones (0.81 cm) as compared with the middle zones (0.30 cm) and upper zones (0.11 cm), P < 0.01. No difference was observed in the AP (P = 0.45) and ML (P = 0.73) directions. Tumour volume, T-stage and forced expiratory ratio were not statistically significant predictors of respiratory-induced tumour motion.

Conclusion

Respiratory-induced tumour motion in the SI direction was significantly greater in lower lobe and lower pulmonary zone tumours compared with apical tumours. Tumour volume, T-stage and spirometry did not correlate with the magnitude or direction of respiratory-induced tumour motion. During curative radiotherapy in lung cancer, attention should be paid to motion management, especially for lower lobe tumours.

Introduction

Respiratory-induced lung tumour motion is a well-established cause of inter-fraction and intra-fraction geometric uncertainty during radiotherapy [1], [2]. Traditional simulation methods rely on three-dimensional computed tomography (CT) for treatment planning, but this captures lung tumours at a random time point within the breathing cycle. Studies have shown that structures within the thorax can be significantly distorted during respiration, giving rise to motion artefact as well as ambiguity in tumour size, shape and position [3].

A uniform margin of tissue is added to the clinical target volume during planning, forming the internal target volume (ITV) that compensates for this phenomenon. The addition of standard respiratory motion margins, however, does not account for individual lung tumour movement. Consequently, ITV margins must be wide enough to prevent geographical miss during the delivery of high dose, radical radiotherapy. This can result in unnecessary irradiation of surrounding normal tissue, increasing the risk of treatment-related toxicities, such as oesophagitis, lung fibrosis and radiation pneumonitis [4], [5], [6].

Numerous motion management techniques have been used during radiotherapy delivery, many of which are described in a recent review article by Cole et al. [7]. They divide these strategies into the broad categories: imaging, breath-hold, abdominal compression, tracking and gating. Specifically, several imaging modalities have been used to characterise respiratory-induced tumour motion, including fluoroscopy, breath-hold CT scans and dynamic magnetic resonance imaging studies [8], [9], [10], [11]. Trials using real-time tracking of fiducial markers have also been conducted, showing the greatest movement in the superior–inferior (SI) direction in tumours located within the lower lobes [12]. In recent years, the advent of four-dimensional CT (4DCT) has further improved our ability to image tumour motion throughout respiration [13], [14], [15]. The clinical role of 4DCT in radiotherapy planning has evolved over the last decade, giving rise to improved gross tumour volume (GTV), ITV and planning target volume delineation [16], [17].

This study was designed to evaluate the magnitude and direction of respiratory-induced lung tumour motion using 4DCT, in patients treated with radical radiotherapy. We tested for associations between tumour motion and patient and tumour characteristics. We hypothesised that the magnitude and direction of lung tumour motion can be predicted based on tumour and patient factors.

Section snippets

Materials and Methods

This was a single-institutional, retrospective study of radical lung cancer patients of the Peter MacCallum Cancer Centre with accessible 4DCT scans between December 2009 and May 2013. This study was approved by the Ethics Committee of the Peter MacCallum Cancer Centre (reference number 13/105).

Results

In total, 609 patients were identified as having 4DCT scans of the chest; 101 of these patients had accessible 4DCT scans of lung cancer primaries treated at the Peter MacCallum Cancer Centre and were therefore eligible. Most of the ineligible patients were excluded due to inaccessible scans; the remaining patients who were ineligible had primary tumours that were not lung origin. Five patients had two separate tumours within the lungs. In these cases, the non-primary lung tumour was not

Discussion

The purpose of this study was to assess predictors of respiratory-induced tumour motion using 4DCT imaging, in patients receiving radical radiotherapy for lung cancer. An analysis of the results from 101 patients showed tumour location within the lung to be the most significant predictor of the magnitude of respiratory-induced tumour motion, in the SI direction.

These findings are consistent with several studies [12], [18], which have similarly shown tumours located in the lower lobes to display

Conclusion

Using 4DCT scans in radical-intent patients with lung cancer, our study showed the anatomical lobe and pulmonary zone to be the most significant predictors of respiratory-induced tumour motion, but only in the SI dimension. In these cases, tumour excursion was greatest in the lower lobes and pulmonary zones. Tumour volume, T-stage and forced expiratory ratio were not significant predictors of respiratory-induced lung tumour motion. For mobile tumours, 4DCT-based scanning remains the standard of

References (27)

Cited by (0)

View full text