Hepatocellular carcinoma (HCC) is a common malignancy that occurs in humans, and only 30–40% of patients are diagnosed at an early stage of the disease [1]. Although treatment options exist for primary HCC, radiotherapy has become an effective strategy for patients with unresectable HCC [1, 2]. Three-dimensional conformal radiotherapy (3D-CRT) and intensity-modulated radiotherapy (IMRT) has been applied to HCC radiotherapy, and high-dose local radiotherapy was feasible and effective for treating unresectable HCC [4, 5, 6, 7]. IMRT can spare more organs at risk (OARs) with highly conformal and homogeneous dose distribution than 3D-CRT. However, some problems exist with IMRT, such as a long delivery time with more monitor units per fraction, which may affect the accuracy of treatment because of increased intrafractional patient motion and damage to normal tissue [8, 9].

Volumetric modulated arc therapy with the RapidArc (RA, Medical Systems, Palo Alto, CA, USA) and VMAT™ (Elekta Group, Crawley, UK) technique has been used in clinical practice for the treatment of various cancers [10, 11, 12, 13, 14, 15, 16, 17, 18, 19], including HCC [14]. For radiotherapy of HCC, target volume delineation is greatly affected by respiratory motion [20]. A larger margin (internal target volume) than needed is commonly added based on the gross tumor volume (GTV) to assure adequate coverage, which could cause radiation-induced injury of organs at risk (OAR). However, respiratory motion was widely variable in different patients, making it difficult to determine a completely safe margin [13]. For precise positioning of the tumor target volume, an active breathing coordinator (ABC) device has been used in 3D-CRT for the treatment of HCC [20, 21, 22]. The 3D-CRT associated with ABC in HCC radiotherapy spared more of the normal liver tissue and assured the accuracy of the target volume. However, the disadvantage of the relatively long beam-on time of each field in IMRT limits its combination with ABC, and few reports have studied the combination of IMRT and ABC.

Compared to conventional IMRT, RA achieves intensity-modulated radiotherapy with continuous rotation of the gantry combined with the dynamic multiple leaf collimator (MLC) [23]. Volumetric modulated arc therapy achieves better planning target volume (PTV) coverage, and the dose distribution becomes noticeably smoother with fewer monitor units and a shorter treatment time [10, 11, 12, 13, 14, 15, 16, 17, 24]. Because the MLC of static gantry IMRT required carriage movement to cover a large volume from left to right, while the MLC of RapidArc moves back and forth repeatedly during rotation of the gantry, IMRT requires an excessive number of monitor units (MU). Using RA in conjunction with ABC for HCC has not been reported to date.

In this study, 3D-CRT, IMRT, and RA plans using three breathing techniques (free-breathing, (FB), end inspiration hold (EIH), or end expiration hold (EEH)) were designed for patients with HCC. We investigated the feasibility and the dosimetric features of RA associated with ABC for HCC radiotherapy.

Materials and methods

Patients

Twelve randomly selected patients (2 females and 10 males, median age 56 years, range 52–60 years) with pathologically confirmed HCC who were treated at our hospital from January 2010 to January 2011 were included in the study. The cardiopulmonary function of each patient was examined before the study was performed to ensure that they could coordinate with the ABC. The breathing hold time was over 30 s [22, 25]. All patients accepted iodized oil transcatheter arterial chemoembolization (TACE) treatment that was well deposited in the tumor region. The study was approved by the Research Ethics Board of Shandong Cancer Hospital and informed consent was obtained from all patients.

CT simulation and PTV acquisition

Simulations were performed with a Philips Brilliance CT Big Bore (Phillips Medical Systems, 96 Highland Heights, OH, USA) associated with the Elekta active breathing coordinator™ (ABC) system (Synergy 102™, Elekta, Crawley, UK). Before simulation, breathing training was conducted. The patients were immobilized by vacuum pillow with their hands above their head. The CT scanning region extended from 4 cm of the upper edge of the diaphragm to 4 cm of the lower edge of the right kidney, with 3 mm reconstruction slice thickness. After the CT scans were obtained for FB (CTFB), the EIH CT scans (CTEIH) and EEH CT scans (CTEEH) associated with ABC were performed. All CT scans were obtained in a spiral scan model (pitch = 0.938, table speed = 30 mm/s, reconstruction slice thickness = 3.0 mm) [26, 27]. The CT images were transmitted to the treatment planning system Varian Eclipse V8.6.15 (Varian Medical Systems, Palo Alto, CA, USA) for target volume contouring and plan designing.

GTVs were contoured under the same window width (200 Hu) and level (40 Hu) on CTFB, CTEIH, and CTEEH, and labeled as GTVFB, GTVEIH, and GTVEEH, respectively. PTVEIH and PTVEEH were obtained from GTVEIH and GTVEEH, respectively, plus 8 mm margins isotropically [20]. PTVFB was obtained using 2 cm margins at the craniocaudal direction and 1.5 cm margins in the left–right and anterior–posterior directions (all margins contained the breath motion and setup errors) [28, 29, 30, 31, 32]. The liver, stomach, and duodenum were also delineated. Normal liver was defined as the volume of liver minus the PTV.

Radiotherapy plan design requirements

3D-CRT, IMRT, and RA plans were designed based on CTFB, CTEIH, and CTEEH, respectively. The tumor dose was 50 Gy (2.0 Gy/fraction × 25 fractions) administered at the isocenter with inhomogeneity correction. The requirements for tumor dose coverage were as follows: the PTV had to be covered by the prescription dose, and inhomogeneity had to be less than 10% [18]. The dose constraints for OAR were as follows: the mean dose of normal liver was limited to 23 Gy, and the dose–volume histogram (DVH) of normal liver was within the tolerance area (i.e., V5 < 86%, V10 < 68%, V20 < 49%, V30 < 28%, and V40 < 20%) [34]; for the stomach and duodenum, the maximum dose was limited to 45 Gy, and the volume receiving > 25 Gy was limited to < 5 cm3 [35].

Radiotherapy plans

  1. 1.

    3D-CRT plans: 4–7 coplanar radiation fields were used, and the weight and gantry angle as well as the shape, size, and angle of the multiple leaf collimator (MLC) of every field were adjusted to meet the dose requirements. The dose rate was set at 300 MU/min.

  2. 2.

    IMRT plans: 5 coplanar radiation fields were designed using the step-static approach, and the gantry angles were not divided equally. The dose rate was set at 400 MU/min.

  3. 3.

    RA plans: Three 135 ° arc coplanar fields were applied and optimized simultaneously. Two arcs overlapped in the liver region. The maximum dose rate was set at 600 MU/min.

The planning objectives were as follows: 98% of the volume of PTV reached 95% of the prescription dose, with 10% of the volume of PTV not exceeding 110% of the prescription dose. The 100% prescription dose of all plans was normalized to the PTV mean dose. All plans were optimized for Varian Trilogy equipped with MLC with a leaf width of 5 mm at the isocenter in the inner 20 cm, and 10 mm for the outer 2 × 10 cm for a 15 MV photon beam. Dose distributions were computed with the Analytical Anisotropic Algorithm (AAA) implemented in the Eclipse 8.6.15 treatment planning system with a maximum calculation grid resolution of 2.5 mm [16].

Planning evaluation

For PTV, the D1% and D99% (doses to 1% and 99% volume of the PTV, respectively) were the maximum and minimum dose, respectively. The conformality index (CI) was calculated as the Van’t Riet definition:

(Equ1)

where TVRI  is the target volume covered by the reference isodose, TV is the target volume, and VRI is the volume of the reference isodose. The homogeneity index (HI) was defined as

(Equ2)

The CI and HI ranges from 0 to an ideal value of 1 [33]. The mean irradiated dose of normal liver (Dmean), V5, V10, V20, V30, and V40, where Vx represents the percentage of the volume of x Gy in the volume of normal liver [34], and the maximum doses (Dmax) of the stomach and duodenum and the irradiated dose received by 5cm 3 of volume (D5cm 3) were also recorded and compared [35].

Statistics analysis

SPSS 16.0 (IBM, Chicago, IL, USA) was used for the statistical analyses. The Friedman test was used to compare three groups of data. The paired Wilcoxon test was used to compare the pairwise data. A p  < 0.05 represented statistical significance.

Results

Target volume comparison

All patients completed the CT simulation associated with ABC. The diaphragmatic mobility measured by CTEIH and CTEEH images averaged 1.3 cm (0.71–1.90 cm). No significant difference occurred in the volume of liver, normal liver, and GTVs between EIH, EEH, and FB (p > 0.05). PTVFB was significantly larger than the PTVEIH and PTVEEH (p < 0.05). The mean value of PTVFB/PTVEIH and PTVFB/PTVEEH were 1.94 and 1.835, respectively, while no significant difference between PTVEIH and PTVEEH was observed (p > 0.05) (Tab. 1).

Tab. 1 The volume difference of liver, normal liver, GTV, and PTV between EIH, EEH, and FB techniques
Full size table

Target coverage and dose homogeneity

All plans met the requirements for tumor dose coverage and OAR dose limitation (Fig. 1). The CI and HI of RA were significantly better than IMRT and 3D-CRT (p < 0.05; Tab. 2). The maximum dose of PTV in 3D-CRT was greater than RA and IMRT (p  < 0.05). Moreover, there was no significant difference in the minimum dose of target volume among the three plans (p > 0.05). No significant difference in CI, HI, and the minimum dose and maximum dose of PTV was observed among RA plans in FB, EIH, and EEH (p > 0.05; Tab. 2 and Fig. 2).

Fig. 1
figure 1

The dose distribution features of the different plans (100 cGy = 1 Gy). All of the plans met the clinical requirements, and PTVEEH and PTVEIH were smaller than PTVFB. The RAEEH and RAEIH plans significantly reduce the normal liver radiation dose compared to RAFB (p < 0.05)

Full size image
Fig. 2
figure 2

The dose–volume relationship of normal liver between three-dimensional conformal radiotherapy (3D-CRT), intensity-modulated radiotherapy (IMRT), and RapidArc (RA) (100 cGy = 1 Gy). RA and IMRT plans significantly reduce the V30 of normal liver (p < 0.05)

Full size image
Tab. 2 The target coverage and dose homogeneity difference between 3D-CRT, IMRT, and RA; as well as the differences between RA plans in FB, EEH, and EIH
Full size table

Organs at risk

The Dmean, V10, V20, V30, and V40 of normal liver in 3D-CRT was greater than RA and IMRT (p < 0.05), while the V5 and V10 in RA were higher than in IMRT. In contrast, no statistically significant difference was found in V5, V10 of normal liver, Dmax, and D5cm 3 of the stomach and duodenum between 3D-CRT, IMRT, and RA (p > 0.05) (Tab. 3).

Tab. 3 The dose–volume difference of organs at risk between 3D-CRT, IMRT, and RA
Full size table

The Dmean, V5, V10, V20, V30, and V40 of RA in FB were larger than those in EEH and EIH (p < 0.05), but there was no statistically significance difference in V5 and V40 among the three respiration techniques. No significant difference was found in all of these indices between RA in EEH and EIH, and D5cm 3 of stomach and the Dmax of duodenum did not differ significantly among the three respiration techniques (p > 0.05). The stomach Dmax of RA plans in FB was significantly larger than EEH and EIH (p < 0.05), while the duodenum D5cm 3 of the RA plans in EEH was significantly smaller than FB and EIH (p < 0.05) (Tab. 4, Fig. 3).

Tab. 4 The dose–volume difference of organs at risk among RA plans in FB, EEH, and EIH
Full size table
Fig. 3
figure 3

The dose–volume relationship of normal liver among RapidArc(RA) plans in free breathing (FB), end expiration hold (EEH), and end inspiration hold (EIH). The RA plans in EEH and EIH significantly reduce the normal liver radiation dose (p < 0.05). 100 cGy = 1 Gy

Full size image

Monitor units and delivery time

The monitor unit in the IMRT plans (626.33 ± 113.97 MU) was significantly greater than RA (550.28 ± 122.56 MU) and 3D-CRT (254.06 ± 18.59 MU) (p < 0.05). The treatment time (from first filed beam-on to the last filed beam-off) in RA (130 ± 10 s) was similar to 3D-CRT (135 ± 10 s), but was significantly less than IMRT (540 ± 45 s) (p < 0.05).

Discussion

Volumetric modulated arc radiotherapy has been investigated for the treatment of various cancers [11, 12, 13, 14, 15, 16, 17, 18, 19], and RA has been shown to achieve better dose coverage than traditional IMRT using a markedly shorter treatment time. We compared RA with IMRT plans and showed that RA achieved conformal degrees with shorter treatment times and fewer monitor units [14]. Here, RA associated with ABC in HCC radiotherapy achieved better target coverage and spared more OARs using shorter treatment times.

The main advantage of incorporating ABC with RA in HCC radiotherapy is to reduce the negative effects on OARs. We compared dose–volume parameters for normal liver, stomach, and duodenum and found that normal liver tissue had the highest risk of damage during radiotherapy [28]. Radiotherapy associated with ABC reduced the volume of the PTV with an accurate determination of the position and target volume. RA and IMRT significantly reduced the Dmean, V20, V30, and V40 of normal liver compared to 3D-CRT, and RA plans in EIH and EEH significantly reduced the Dmean, V5, V10, V20, V30, and V40 compared to RA plans in FB. RA associated with ABC can reduce the Dmean and V30 of normal liver with perfect PTV dose coverage, which may lower the incidence of RT-induced liver disease (RILD) [36]. The stomach and duodenum are also at risk for damage during HCC radiotherapy. No significant difference in the Dmax and D5cm 3 of these organs between 3D-CRT, IMRT, and RA was observed, possibly due to the smaller tumor volume.

When ABC is implemented into the RA process, the size of the arcs at different rotations is critical. Most RA plans and treatments use single or two 358° whole rotation arcs or rotation arcs over 180°. The whole arc was used in designing RA plans, and the segmenting treatment was considered at the beginning of our study. However, during treatment, if the beam were suddenly beam-offed, the gantry of the linear accelerator will move forward due to inertia, which may cause mechanical failure. Moreover, if the treatment was interrupted during the entire arc, the gantry would return to the initial position and run empty to the location of the interruption. The gantry would then re-administer the beams, so the total treatment time would be prolonged. Three 135° small arcs were used in this study, and the linear accelerator used had an angular velocity of 4.8°/s. The treatment time for each arc was approximately 28 s. Breaks for patient rest and gantry preparation take approximately 30 s, so the treatment could be completed within 115 s, which is equivalent to the time for 3D-CRT and much shorter than IMRT. The breath holding time limited the size of each arc for RapidArc in combination with ABC. It was previously demonstrated that “there was no measurable diaphragm or hepatic microcoil movement during 30 s ABC breath holds in any of the patients treated with ABC” [20]. Here, the breath holding time of patients reached 30 s after training. We believe the breath training is very important for the progress of radiotherapy using ABC, and is a simple and effective way to assure reproducibility and stability of respiration motion. However, patients should not hold their breath for too long, because it could fatigue respiratory muscles, which could affect the accuracy of dose distribution.

The breath function and cooperation of each patient were important when the breathing techniques were selected for radiotherapy associated ABC. We compared FB, EIH, and EEH hold without using the end of deep inspiration hold (EDIH) or end of deep expiration hold (EDEH). First, greater mobility was investigated in the EDIH and EDEH, which may affect the location relationship between the liver and surrounding organs, thereby affecting the dose–volume relationship of OARs [37]. Second, the deep inspiration and expiration required good respiratory functions and the coordination ability of the patients. The respiratory ranges of each patient in EIH or EEH had to spontaneously reach the trigger threshold and exhibit good reproducibility. Here, the difference in the D5cm 3 of the duodenum may be related to the target volume and duodenum position becoming distant in EEH. For selection of EIH and EEH, the respiratory functions, coordination ability, and other specific conditions should be considered.

Our study provides guidance for radiotherapy using RA for tumors that are affected by respiratory motion. Unfortunately, we did not use an image-guided system, and some reports suggest that image-guided radiotherapy (IGRT) achieved by cone beam computer tomography and ultrasound can spare more OARs and ensure an accurate target volume [38, 39, 40, 41]. Applying the IGRT system could evaluate the irradiation dose of the PTV, which may be efficacious against cancers without increasing toxicity [40, 41, 42]. Sometimes, the IGRT system requires implanted fiducials in the tumor [40], and lipiodol is a direct surrogate for CBCT image guidance for radiotherapy of HCC after TACE [43]. Image-guided system associated with ABC can increase the overall precision of target volume [43, 44, 45]. Image-guided RA in combination with ABC may become an effective and accurate way to administer HCC radiotherapy.

Conclusion

The use of RA with three arcs of 135° associated with ABC is a feasible method for HCC radiotherapy and provides equal or better target coverage than IMRT. It also provides better protection of normal tissue and improves treatment efficiency, and may become an effective method for administering HCC radiotherapy with perfect dose delivery. However, the advantage of this strategy should be explored in additional studies.