Background

Respiratory induced tumor motion is the major complicating factor in radiotherapy of thoracic and upper abdominal targets. A variety of techniques have been developed for the clinical management of organ motion, each with distinct advantages and drawbacks [1–5]. These techniques can be generally categorized in order of approximate increased technical complexity as motion encompassing irradiation, breath-hold methods, compression methods, gating methods, and dynamic tracking methods. Among these, gating methods have gained clinical traction as they limit the volume of normal tissue irradiated relative to motion encompassing irradiation, yet provide a reliable and a technical feasible alternative to continuous tracking irradiation [6–18]. In general, the dosimetric variation of DG-VMAT versus conventional single gated plans will depend on the patient anatomy on EOE and EOI phases, and the quality of the plans for each phase.Figure 3 displays the dosimetric comparison of delivery, relative to the planned DG-VMAT plan using the diode array. The deviations as analyzed using the a 3 mm/3% gamma-test criterion; Figure 3a,b displays the passing points in the two planes of the diode array, while the dose profiles and the gamma histogram is shown in Figure 3c-e. For the dual-gated delivery, 97.5% of the measurement points pass the gamma-test criterion a gamma < 1. As indicated by Figure 3a,b, the failed points were at the primarily at periphery of the field in low dose regions. In addition to the diode measurements, the pinpoint chamber absolute dose measurement agreed with the dose calculation within 0.7%.

Figure 2
figure 2

Isodose distributions of individual (a) EOE, (b) EOI, and (c) summed dual-gated plan. (d) Dose-volume histograms of the single EOE gating plan (solid line) and deformed dose summation of the dual-gated plan (dashed line).

Full size image
Figure 3
figure 3

Dosimetric validation of DG-VMAT using a diode array (a), (b) Display of the diodes passing the 3%/3 mm gamma test for the two crossed plane diode arrays (c) Horizontal dose profile of dose plane in (a), (d) Vertical dose profile of dose plane in (b), (e) Histogram of the gamma distribution.

Full size image

The delivery time reduction was assessed through comparison of the DG-VMAT delivery time with that of the conventional EOE plan scaled to the same dose. The conventional EOE gated VMAT delivery requires 346 seconds for the studied case, while the proposed DG-VMAT technique is delivered in 177 seconds per fraction. Thus, for this particular case, dual-gated VMAT provides a 95.5% improvement in delivery efficiency compared to the corresponding single-gated delivery.

Discussion

Implementation of DG-VMAT requires the synchronization of the gantry motion and MLC with two phases of the respiratory cycle. As such, for half the transitions between exhale and inhale phases, the gantry is required to roll back between the phases, as depicted in Figure 1c,d. This is shown to be possible with the TrueBeam™ STx, which has a gantry rotation speed of 6 degrees/second, and a MLC leaf speed of 2.5 cm/second at isocenter. The DG-VMAT delivery required a roll-back of an average of 2.05 degrees, which is achieved in 0.34 seconds. Since the transition time between exhale and inhale gating windows was 1.5 seconds, there was more than sufficient time for the gantry and MLC to move to the planned positions in preparation for the subsequent nodal delivery. Considering an average breathing cycle of 4–6 seconds, such a motion is within the limits of current linacs as demonstrated in this first experimental demonstration.

While the results indicate that the treatment time may be reduced by nearly a factor two for an ideal breathing pattern, a number of issues must be considered for implementation in a clinical setting. Most importantly is the variability in breathing patterns of human subjects. Specifically, it is known that under free breathing, subjects may spend more time in the exhale phase than inhale. If un-coached, a reduction of the magnitude of the efficiency enhancement with dual gating may be expected. In the current implementation of DG-VMAT, it is explicitly assumed that there is 1:1 ratio between the EOI and EOE gating window. To achieve this during a patient treatment, coaching via audio-visual guidance must be used. Specifically, the patient will be directed to briefly hold their breath at inhale and exhale for equal time intervals that are known from a simulation study to be comfortably tolerable for the patient. Through audio-visual guidance, the patient will be instructed on when to exit the EOI or EOE phase to proceed to the next delivery node. Such a technique has been experimentally shown to effectively equalize the inhale and exhale phases in healthy human subjects by Geneser et al.[19], as shown in Figure 4.

Figure 4
figure 4

Dual-gated dynamics of a healthy individual under free breathing (top) and coached inhale and exhale breath-holds (bottom). The inhale and exhale gating windows are indicated by the light and dark green windows, respectively.

Full size image

In this initial work, treatment planning was performed with the inhale and exhale phase optimized independently of each other. 4D treatment planning [20–22] may be adapted for a more cohesive optimization of the two phases through which DVH parameters are simultaneously optimized.

Several observations on the limitations and advantages of DG-VMAT can be made in relation to other respiratory management techniques. DG-VMAT is technically more complex than breath-hold techniques. Deep inspiration breath hold may achieve more advantageous anatomical separation for normal tissue sparing and has become more feasible with the use of high dose-rate flattening-filter-free beams [23, 24]. However, prolonged breath holding, as required to deliver SBRT doses, may not be tolerated by portions of the patient population, specifically those with already compromised lung function. Gating presents an alternative solution for such patients. Gating however, inherently results in significantly higher total treatment times due the fact the beam is conventionally activated for one phase of the breathing cycle. Dual gating, aims to enhance the efficiency of gating. While the technical complexity for such a delivery is higher than conventional gating, it represents significant simplification of alternate dynamic tracking proposals. More importantly, relative to tracking, dual gating only utilizes the stable portions of the respiratory cycle, and thereby eliminates intermediate irradiation between exhale and inhale which is known to be unstable and unpredictable.

Conclusions

To enhance the delivery efficiency of gated VMAT, a technique for dual-gated delivery, leveraging the natural pauses that occur at peak-inspiration and exhalation for irradiation, has been proposed. The technique which necessarily coordinates the gantry rotation and MLC modulation with two different phases of respiratory cycle was experimentally implemented using custom XML programing in TrueBeam™ STx Developer Mode. The results presented herein demonstrate the first successful delivery of DG-VMAT which is shown to result in nearly a doubling of treatment delivery efficiency for ideal sinusoidal respiratory motion. For clinical implementation on patients, audio-visual guidance may be used to coordinate the breathing with the delivery. Dual-gated delivery efficiency can be further improved with additional linac hardware and software modifications to enable implementation in clinical mode. As compared to the existing respiratory-gating VMAT technique, a major advantage of DG-VMAT is that it substantially reduces treatment duration with a modest but practically achievable increase in complexity of the treatment delivery processes. DG-VMAT can potentially provide a compromise between breath-hold, gating, and tracking techniques by increasing the tolerability relative to breath-hold, reducing technical demand and potential inaccuracies associated with irradiation of variable portions of the respiratory cycle relative to tracking techniques, and increasing the efficiency of treatment relative to conventional single window gating.

Authors’ information

Benjamin Fahimian and Junqing Wu are co-first author.