Physics contribution
Four-dimensional CT scans for treatment planning in stereotactic radiotherapy for stage I lung cancer

https://doi.org/10.1016/j.ijrobp.2004.07.665Get rights and content

Purpose

Hypofractionated stereotactic radiotherapy (SRT) for Stage I non–small-cell lung cancer requires that meticulous attention be paid toward ensuring optimal target definition. Two computed tomography (CT) scan techniques for defining internal target volumes (ITV) were evaluated.

Methods and materials

Ten consecutive patients treated with SRT underwent six “standard” rapid multislice CT scans to generate an ITV6 CT and one four-dimensional CT (4DCT) scan that generated volumetric datasets for 10 phases of the respiratory cycle, all of which were used to generate an ITV4DCT. Geometric and dosimetric analyses were performed for (1) PTV4DCT, derived from the ITV4DCT with the addition of a 3-mm margin; (2) PTV6 CT, derived from the ITV6 CT with the addition of a 3-mm margin; and (3) 6 PTV10 mm, derived from each separate GTV6 CT, to which a three-dimensional margin of 10 mm was added.

Results

The ITV4DCT was not significantly different from the ITV6 CT in 8 patients, but was considerably larger in 2 patients whose tumors exhibited the greatest mobility. On average, the ITV6 CT missed on average 22% of the volume encompassing both ITVs, in contrast to a corresponding mean value of only 8.3% for ITV4DCT. Plans based on PTV4DCT resulted in coverage of the PTV6 CT by the 80% isodose in all patients. However, plans based on use of PTV6 CT led to a mean PTV4DCT coverage of only 92.5%, with a minimum of 77.7% and 77.5% for the two most mobile tumors. PTVs derived from a single multislice CT expanded with a margin of 10 mm were on average twice the size of PTVs derived using the other methods, but still led to an underdosing in the two most mobile tumors.

Conclusions

Individualized ITVs can improve target definition for SRT of Stage I non–small-cell lung cancer, and use of only a single CT scan with a 10-mm margin is inappropriate. A single 4D scan generates comparable or larger ITVs than are generated using six unmonitored rapid CT scans, a finding related to the ability to account for all respiration-correlated mobility.

Introduction

The local control rates after curative radiotherapy in patients with medically inoperable Stage I non–small-cell lung cancer (NSCLC) have been disappointing (1). Even the use of three-dimensional (3D) conformal radiotherapy to doses of up to 70 Gy has failed to improve on these results, with local failure rates still as high as 40–60% at 3 years (2). Approaches for dose-escalation using conventional, hyperfractionated and hypofractionated schemes have been investigated as a means of improving local control. Recent studies have reported unacceptable toxicity with conventionally fractionated doses exceeding 90 Gy (3, 4). In contrast, several Phase I/II studies using hypofractionated stereotactic radiotherapy (SRT) for Stage I NSCLC have shown that high local control rates of 90% or more can be obtained with only limited toxicity, although the follow-up is relatively short in some series (5, 6, 7, 8, 9, 10, 11, 12).

The application of high-dose radiotherapy, however, requires meticulous attention toward ensuring an optimal target definition. Recent studies have shown that individualized (i.e., “patient-based”) margins, as opposed to standard “population-based” margins, are required for radiotherapy of lung tumors (13, 14, 15). Approaches used for determining individualized margins include “slow” computed tomography (CT) scans (16, 17), and the coregistration of expiratory and inspiratory target volumes (13, 18, 19, 20).

When the extracranial SRT program for Stage I NSCLC commenced at our center in April 2003, the option of performing “slow” CT scans on the CT scanner was not available. Therefore, target definition was based on the use of six rapid unmonitored spiral multislice CT (MSCT) scans to generate an internal target volume (ITV), a lengthy approach that requires laborious coregistration. Recently, a 16-slice CT scanner became available, and it allowed for four-dimensional (4D) or respiration-correlated CT scans to be performed. 4DCT scans generate spatial and temporal information on mobility in a single investigation and represent a major breakthrough in imaging for radiotherapy planning (21, 22). In this technique described as retrospective gating, the respiratory waveform is synchronously recorded during CT acquisition, and multiple CT slices are acquired at each table position for at least the duration of one full respiratory cycle (23). This yields CT datasets for 10 phases of the respiratory cycle.

We evaluated the use of 4DCT scans as the sole technique for generating ITVs for peripheral lung tumors and compared this with target volumes generated using our routine six-scan approach. The dosimetric consequences of treatment planning using both methods of target definition were evaluated, and compared with the use of standard planning target volumes (PTVs).

Section snippets

Methods and materials

Ten consecutive patients with peripheral Stage I NSCLC, in whom both CT scanning techniques were performed for planning hypofractionated SRT, were included in this analysis. All patients underwent staging using an (18F)fluorodeoxyglucose-Positron Emission Tomography (FDG-PET)scan and only involved-field radiotherapy was performed. Tumor characteristics are summarized in Table 1. Five patients had lesions located in the lower lobes.

Volumetric analysis

The mean GTVs derived using both CT techniques were comparable, ranging from 1.1 to 36.8 cc for multiple MSCT scans, and 1.5 to 38.8 cc for 4DCT scans (Table 2). To estimate the extent of tumor mobility, the ratio between the mean GTV6 CT and ITV6 CT was calculated for each patient. For completely immobile tumors, this ratio will be 1. The mean ratio GTV6 CT/ITV6 CT was 0.58 ± 0.17, and was particularly low in Patients A (0.36) and G (0.26), both of whom had tumors of the lower lobe.

The volume

Discussion

Far higher rates of local control have been reported in early-stage NSCLC with the use of hypofractionated SRT than was achieved using other radiotherapy techniques (5, 6, 7, 8, 9, 10, 11, 12). Most authors used standard, population-based margins for target definition (5, 6, 9, 11). However, an accurate individualized 3D definition of target volumes is a more appropriate prerequisite in view of the high doses per fraction and steep dose gradients characteristic of SRT. This is particularly the

References (27)

Cited by (263)

View all citing articles on Scopus
View full text