Introduction
Brain metastases occur in 20–40% of cancer patients [
1], and with improvements in local control and new effective systemic treatments, this incidence continues to increase [
2]. The overall prognosis of cancer cases with brain metastases is poor. Whole brain radiotherapy (WBRT) is a fundamental radiation modality used specifically for patients with extensive brain metastases [
3‐
7]. However, WBRT is associated with a short local control time and side effects including neurocognition dysfunction [
8] and hearing deficits [
9]. It is known that radiation-induced damage to the hippocampus plays an important role in the cognitive dysfunction [
4,
8,
10‐
12], and hearing impairment can be linked to damage to the inner ear [
13]. The threshold cochlear dose for hearing loss with cisplatin-based chemotherapy and radiotherapy was predicted to be 10 Gy [
14]. Encouragingly, the results of the RTOG 0933 trial have shown that hippocampus avoidance can effectively reduce the cognitive impairment caused by WBRT [
8].
WBRT with simultaneous integrated boost (SIB) for brain metastases has been proven to be advantageous in prolonging local control time and overall survival (OS) with a short treatment time [
15,
16]. With the development of advanced radiation technologies such as step-and-shoot intensity-modulated radiotherapy (sIMRT), dynamic intensity-modulated radiation therapy (dIMRT), volumetric-modulated arc therapy (VMAT), and helical tomotherapy (Tomo), it is possible to perform WBRT with organs at risk (OARs) avoidance and SIB for brain metastases. Ferro M et al. suggested that IMRT could deliver a SIB-WBRT treatment in patients with more than 3 brain metastases [
17]. Sumit Sood et al. demonstrated the feasibility of WBRT using VMAT not only to spare hippocampus, but also significantly reduce dose to OARs, including the scalp, ear canal, cochleae, and parotid glands [
18]. Tomo provides a homogeneous dose distribution to the whole brain and conformally avoids the hippocampus in a single treatment plan [
19]. However, it is not clear which is the most appropriate WBRT approach with hippocampus and inner ear avoidance and simultaneous integrated boost for limited brain metastases. In the present study, a dosimetric comparison between sIMRT, dIMRT, VMAT and Tomo was performed to address the technical advantages of the four treatment modalities.
Materials and methods
Simulation and contouring
Ten consecutive patients with 1–5 brain metastases confirmed by magnetic resonance imaging (MRI), who were treated between September 2017 and October 2017, were included. All patients had histologic evidence of non-small cell lung cancer (NSCLC). All patients were in supine position wearing a thermoplastic mask (CIVCO, Orange City, IA, USA) before computed tomography (CT) simulation. CT scans of the entire head region were acquired using a large Bore CT simulator (Philips, Cleveland, OH, USA) with a slice thickness of 1.5 mm. The 3D e-THRIVE brain MRI axial T1-weighted sequences and double gadolinium contrast-enhanced sequences were obtained with a Philips Ingenia 3 .0T MRI scanner (Philips, Cleveland, OH, USA) after CT simulation scanning. The slice thickness for MRI scans was 1.5 mm. The CT and MRI datasets were registered in the Eclipse Version 11.0 planning system (Varian Medical Systems, Palo Alto, CA, USA) for target volume delineation.
Contouring was made using 2D brush on axial images. The gross tumor volume (GTV) was defined as the contrast-enhanced lesion. The planning tumor volume for metastases (PTVmets) was generated by adding to the GTV a 3 mm margin to correct for positional inaccuracies. The whole brain volume was contoured as the clinical target volume (CTV). The hippocampus was contoured on the MRI axial T1-weighted image sequence according to the RTOG 0933 protocol [
8]. The hippocampus with a 5 mm expansion was defined as the area for hippocampal avoidance. The inner ear (cochlea, vestibule, semicircular canal) was delineated on CT bone window image datasets. To construct the planning target volume for brain (PTVbrain), a 3 mm margin was added to the CTV and adjusted to not overlap the inner ear volume. Then, the area for hippocampal avoidance and the GTV volume plus 5 mm (area for dose fall-off) were subtracted from the whole brain volume. Other OARs were outlined, including optic nerves, eyes, lenses, and brainstem. All contours were delineated by the same radiation oncologist and peer-reviewed.
Prescription and planning
The prescription dose was 45–50 Gy in 10 fractions to the PTVmets and 30 Gy in 10 fractions to the PTVbrain simultaneously. The 30 Gy in 10 fractions scheme is a common scheme used in WBRT [
1,
20]. The equivalent dose in 2 Gy/fraction (EQD2) of our fractionation scheme for PTVmets was calculated to be 54.4–62.5 Gy
10 based on the LQ model [
21]. All plans aimed to encompass at least 95% of PTVmets by 100% of the prescribed dose.
We constrained the unilateral or bilateral hippocampi (subgranular zone) dose according to findings of the RTOG 0933 trial [
8]. The maximum dose (Dmax) and minimum dose (D100%) limit for the hippocampal structures were 17 Gy and 10 Gy in 10 fractions, respectively. In two cases the metastasis was located close to the ipsilateral hippocampus (< 1.5 cm), we met the dosimetric constraint criteria for the contralateral hippocampus. The mean dose (Dmean) constraints for inner ear structures were 15 Gy in 10 fractions. For other OARs, the maximal doses allowed were 9 Gy in 10 fractions for lens, 37.5 Gy in 10 fractions for brainstem and optic nerves, and 40 Gy in 10 fractions for normal brain. PTVmets coverage was the first aim in the planning process. Achieving homogenous and conformal PTVbrain dose distribution without exceeding the dose constraints of OARs was the second aim. The PTVbrain coverage could be reduced to 95% or even < 90% to meet the restrictions to the OARs.
For each case, three separate treatment plans (sIMRT, dIMRT, and VMAT) were optimized with 6 MV photon beams for a linear accelerator (Trilogy, Varian Medical Systems) in the Eclipse Version 11.0 planning system. This linear accelerator is equipped with Millenium 120 multi-leaf collimator. The width of the central leaf pairs is 0.5 cm. A Analytical Anisotropic Algorithm algorithm was used for dose calculation with a spatial resolution of 2.5 mm. Beam arrangement for sIMRT and dIMRT was done using nine coplanar fields with a collimator angle of 0°. Two 360° non-coplanar full arcs were used in the VMAT plans. Tomo planning was performed using the Tomotherapy treatment planning system (Accuray® Planning Station, Hi Art® Version 5.1.3, Madison, WI, USA). Compared to 2 .5cm, 1 .0cm field width significantly improved the the whole brain homogeneity by 32% [
18]. So, The field width and pitch were chosen as 1.05 cm and 0.5 cm, respectively. Each plan was calculated to achieve the optimal deliverable plan with an acceptable target coverage while not exceeding the dose constraints to OARs. All plans were designed by the same medical physicist who had 5 years of planning experience and reviewed by other experts.
Dosimetric evaluation
Plans were finally reviewed and analyzed based on dose–volume histogram (DVH). For each patient, the target coverage (TC), conformity index (CI), and homogeneity index (HI) for PTVmets and PTVbrain were evaluated. Subsequently, the following dose–volume parameters for OARs were calculated: 1) hippocampal maximum dose, 2) hippocampal mean dose, 3) hippocampal minimum dose, 4) inner ear mean dose, 5) eyes maximum dose, 6) lens maximum dose, 7) optic nerve maximum dose, and 8) brainstem maximum dose.
According to the conformity index reported by Paddick [
22], the CI was defined as:
$$ CI=\frac{{V^2}_{Tpres}}{TV\times {V}_{pres}} $$
Where VTpres represents the volume within the target receiving a dose greater than or equal to the prescription dose. TV represents the target volume. Vpres represents the volume receiving a dose greater than or equal to the prescription dose. The value of CI can range from 0 to 1, and values closer to 1.0 are usually considered optimal.
The HI was defined as:
$$ HI=\frac{D_{2\%}-{D}_{98\%}}{D_{median}} $$
This is the most commonly used formula in the literature [
23], where
D2% represents the maximum dose delivered to 2% of the target volume and
D98% represents the minimum dose delivered to 98% of the target volume.
Dmedian represents the median dose to the target volume. HI is an objective tool to indicate the uniformity of dose distribution, and values close to 0 are considered optimal.
TC was defined as:
$$ TC=\frac{V_{Tpres}}{TV}\times 100\% $$
TC describes the fraction of the target volume receiving at least the prescription dose. For perfect TV coverage, TC equals 100%.
Statistical analysis
Statistical comparisons among results obtained with the four treatment plan modalities were performed in SPSS, version 18.0 (SPSS Inc., Chicago, IL, USA) using one-way analysis of variance (ANOVA) with Tukey’s multiple comparison post-hoc tests. All statistical tests were two-sided, and differences were considered to be statistically significant if p < 0.05.
Discussion
WBRT is associated with short local control and side effects, including neurocognition dysfunctions [
8] and inner ear deficits [
9]. The potential radiation-induced damage to OARs have been underestimated. With the development of sIMRT, dIMRT, VMAT and Tomo, it is feasible to reduce neurocognitive toxicity and inner ear deficits by using advanced techniques. It is possible that SIB-WBRT with hippocampus and inner ear avoidance can achieve the best balance between intracranial tumor control and cognitive function protection. In the present study, we evaluated the dosimetric advantages of four SIB-WBRT radiotherapy modalities (sIMRT, dIMRT, VMAT and Tomo) with hippocampus and inner ear avoidance for the treatment of limited brain metastases. To summarize the results, while all four modalities achieved PTVmets target coverage and met the basic dosimetric compliance criteria, we found that TomoTherapy provided the most homogeneous PTVbrain target dose (mean HI = 0.39).
Recent reports on SIB-WBRT showed promising outcomes with a shorter treatment period [
15,
16]. Both median intracranial progression-free survival and median survival were increased up to 10 months [
24]. The SIB technique was also found to achieve good hippocampus sparing [
25]. In the present study, SIB-WBRT with hippocampus and inner ear avoidance was replanned and compared. All four planning modalities provided good mean PTVmets target prescription coverage (range, 97.1–99.1%) and similar D2%, D98%, Dmedian, and CI values. The HI value for VMAT was higher than those with the other three modalities (0.091 vs 0.065–0.071,
p < 0.05), but this difference is small and has little clinical significance considering that brain metastases are commonly surrounded by healthy brain.
For the whole brain volume, good target coverage, CI and dose homogeneity are very important. In our study, the best target dose homogeneity was achieved by Tomo (0.39, p < 0.05), followed by VMAT (0.57, p < 0.05). The two fixed gantry IMRT modalities (step and shot, dynamic) provided similar dose homogeneity (0.76, respectively). Tomo plans could achieve better dose distribution in the PTVbrain over the other techniques, as already confirmed by Cozzi et al. [
26] No statistically significant differences in TC and CI were found across modalities. Thus, the four modalities yielded similar results in target volume coverage and conformity.
A phase II multi-institutional trial (RTOG 0933) showed that hippocampus avoidance during WBRT is associated with preservation of memory and quality of life in patients with brain metastases [
8]. The development of advanced techniques such as VMAT and Tomo is increasing the feasibility of this treatment paradigm [
27‐
29]. In the present study, we demonstrated that all four planning modalities (sIMRT, dIMRT, VMAT and Tomo) met the RTOG 0933 protocol dose compliance criteria for hippocampal sparing. Similar mean doses to the hippocampi were achieved with the four planning modalities with an acceptable mean PTVbrain coverage (88.2%–92.6). Compared to healthy brain, the dose to the hippocampi was lower by approximately two-thirds. The entire inner ear system is also vulnerable to RT injury, and dose reduction to this region must be considered even at the relatively low dose of WBRT, as many patients receive some form of cytotoxic therapy. We also demonstrated that the mean inner ear dose was significantly reduced to Dmean ≤15 Gy. None of the four modalities produced any notable dose difference in the dose to the brainstem, since the brainstem contour overlapped the PTVbrain volume in all cases. With the use of two full non-coplanar arcs, the average maximum dose to the lenses was highest with VMAT planning.
Although our results show that the four SIB-WBRT modalities can produce acceptable treatment plans with good avoidance of the hippocampus and inner ear, it is important to note that some limitations of the work exsist. For instance, the four investigated techniques are not only delivery techniques but also associated to treatment planning as well, in particular to plan optimization. In addition, treatments using the CyberKnife or GammaKnife are also very common for brain metastases [
30,
31], which are not mentioned in the study. Moreover, only one physicist was involved in the treatment planning. This procedure could bring some bias to our result. The application of automatic treatment planning approaches promises to reduce this bias.
Conclusions
Our data has demonstrated that all four SIB-WBRT modalities (sIMRT, dIMRT, VMAT and Tomo) can produce acceptable treatment plans with good avoidance of the hippocampus and inner ear. Tomo can provide satisfactory PTVbrain coverage with the highest HI but seems the least efficient, whereas VMAT can provide a sufficient dose distribution with remarkably reduced MUs.
Acknowledgements
Not applicable.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.