Medical Dosimetry ] (2015) ]]]–]]]
Medical Dosimetry journal homepage: www.meddos.org
Technique for comprehensive head and neck irradiation using 3-dimensional conformal proton therapy Mark W. McDonald, M.D.,n† Alexander S. Walter, B.S., R.T.(T.),n and Ted A. Hoene, M.S., R.T.(T.), C.M.D.† Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, IN; and †Indiana University Health Proton Therapy Center, Bloomington, IN
*
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 December 2014 Received in revised form 18 February 2015 Accepted 12 April 2015
Owing to the technical and logistical complexities of matching photon and proton treatment modalities, we developed and implemented a technique of comprehensive head and neck radiation using 3dimensional (3D) conformal proton therapy. A monoisocentric technique was used with a 30-cm snout. Cervical lymphatics were treated with 3 fields: a posterior-anterior field with a midline block and a right and a left posterior oblique field. The matchline of the 3 cervical nodal fields with the primary tumor site fields was staggered by 0.5 cm. Comparative intensity-modulated photon plans were later developed for 12 previously treated patients to provide equivalent target coverage, while matching or improving on the proton plans' sparing of organs at risk (OARs). Dosimetry to OARs was evaluated and compared by treatment modality. Comprehensive head and neck irradiation using proton therapy yielded treatment plans with significant dose avoidance of the oral cavity and midline neck structures. When compared with the generated intensity-modulated radiation therapy (IMRT) plans, the proton treatment plans yielded statistically significant reductions in the mean and integral radiation dose to the oral cavity, larynx, esophagus, and the maximally spared parotid gland. There was no significant difference in mean dose to the lesser-spared parotid gland by treatment modality or in mean or integral dose to the spared submandibular glands. A technique for cervical nodal irradiation using 3D conformal proton therapy with uniform scanning was developed and clinically implemented. Use of proton therapy for cervical nodal irradiation resulted in large volume of dose avoidance to the oral cavity and low dose exposure to midline structures of the larynx and the esophagus, with lower mean and integral dose to assessed OARs when compared with competing IMRT plans. & 2015 American Association of Medical Dosimetrists.
Keywords: Proton therapy Head and neck IMRT Dosimetry Nodal irradiation
Introduction Proton therapy is a modality of radiotherapy that differs from xray–based modalities by a finite range of dose deposition within tissue, delivering radiation to the depth of the target and then terminating, with essentially no exit dose radiation to normal tissue beyond the target.1 Proton therapy has been used in the treatment of some uncommon head and neck tumors, particularly paranasal sinus2-5 and nasopharyngeal tumors,6 where the proximity of the optic apparatus can pose challenges to providing adequate target coverage by sufficient dose without undue risk of complications.7,8
Reprint requests to: Mark W. McDonald, M.D., 535 Barnhill Dr, RT 041, Indianapolis, IN 46202-5289. E-mail: [email protected] http://dx.doi.org/10.1016/j.meddos.2015.04.004 0958-3947/Copyright Ó 2015 American Association of Medical Dosimetrists
Patients treated with proton therapy for head and neck tumors where cervical nodal irradiation is indicated have typically received a portion of their treatment with photons during cervical nodal irradiation or received concurrent photon neck radiation matched to the proton therapy plan delivered to the primary tumor site.5,9,10 Sequential use of the 2 modalities may erode the dosimetric advantage of proton therapy in the treatment of the primary tumor site, and concurrent use of both the modalities creates increased uncertainties across the matchline with patient transfer between machines, the possibility of matching through high-risk nodal regions or gross nodal disease, and requirement of treatment room time on 2 machines. At our proton therapy treatment site, a linear accelerator is not available in the same building, leading to the additional complexity of daily transport of patient immobilization devices between 2 treatment sites. Because of these technical and logistical complexities, in 2012,
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
2
we developed and implemented a technique of comprehensive head and neck radiation for patients with sinonasal tumors using proton therapy. Here, we review the treatment technique and dosimetric comparison with intensity-modulated photon therapy. Clinical data on patient tolerance are reported separately.
Methods and Materials In an institution review board–approved retrospective study, we analyzed the dosimetry in 12 patients with sinonasal tumors treated with proton therapy to the bilateral cervical lymphatics and primary tumor site between 2012 and 2014. Patient and treatment characteristics are listed in Table 1.
Patient simulation Patients were immobilized with an Alpha Cradle and a thermoplastic mask attached to a carbon fiber table originally designed at our center for craniospinal irradiation.11 Shoulder retractors (CIVCO, Coralville, IA) were applied with gentle traction to aid in reproducibility of shoulder positioning. A computed tomography scan was acquired with 1-mm slice thickness from the top of the head to the carina. Diagnostic imaging was coregistered to aid in delineation of target volumes and critical normal structures. Treatment planning of both the proton treatment plans and the comparative intensity-modulated radiation therapy (IMRT) treatment plans was performed in Eclipse version 11 (Palo Alto, CA).
Proton treatment planning and delivery Proton treatment was designed for delivery on a 3601 gantry with uniform beam scanning and dose layer stacking.12 Details of our proton therapy treatment system have previously been published.13 A monoisocentric technique was used with a 30-cm snout. The cervical lymphatics were treated with 3 fields: a posterioranterior (PA) field with a midline block over the larynx, esophagus, and spinal cord. The cranial portion of the PA field was left unblocked to provide dose to the medial retropharyngeal lymph nodes. The cervical nodal regions of the right neck were treated with a right posterior oblique (RPO) and of the left neck with a left posterior oblique. The 2 posterior oblique fields were angled 201 to 301 from PA in either direction. On these posterior oblique fields, the medial aperture edge was used to block the spinal cord, while the lateral aperture edge was trimmed close to the body contour to provide some skin sparing on the composite plan (Fig. 1). These cervical nodal fields were weighted differently. For example, when the fraction size was 1.8 Gy, the PA field might deliver 1 Gy and each posterior oblique might deliver 0.8 Gy. When gross nodal disease was present, a shrinking field technique was used to boost the involved nodal region, typically with an anterior and posterior oblique field combination. Table 1 Patient and treatment characteristics Patient age (range)
42.5 y (16 to 71)
Primary site Ethmoids Nasopharynx Maxilla Nasal
9 1 1 1
Histology Esthesioneuroblastoma Sinonasal undifferentiated carcinoma Squamous cell carcinoma Nasopharyngeal undifferentiated carcinoma
5 5 1 1
T category T4 T3 T2
9 2 1
N category N0 N1 N2
8 1 3
Cervical nodal irradiation Elective dose median (range) Definitive dose median (range)
50 Gy (45 to 54) 72.9 Gy (70 to 75.6)
Primary site Postoperative dose median (range) Definitive dose median (range)
66 Gy (63 to 72) 74 Gy (70.2 to 75.6)
At our institution, we elected to use planning target volumes (PTVs) in proton treatment planning, combining concepts of a uniform geometric volume expansion for setup uncertainties with additional beam-specific uncertainties. Above the level of C2, 2 mm of geometric expansion was used to create the PTV, based on institutional assessment of setup error using daily image-guided alignment, which was repeated for each treatment field and checked by a physician before delivery of each field. Below C2, 5 mm of expansion was used owing to increased setup variance expected for the lower cervical neck. An additional 2 mm of proximal and distal margin was added beyond the PTV for proton range uncertainties. A median of 4 mm of compensator smearing was applied (range: 3 to 5 mm). As we incorporated nodal irradiation with proton therapy, an emphasis was placed on sparing the following structures: the oral cavity (an avoidance structure whose lateral and anterior borders were defined by the mandible, cranial by the maxillary bone and inclusive of any air cavity above the oral tongue, inferior including the floor of mouth, and posterior including the base of tongue), the esophagus (drawn from the cricoid to the thoracic inlet), the larynx (inclusive of the thyroid cartilage and postcricoid space abutting the vertebral body), the submandibular glands (unless ipsilateral level IB nodes were treated), and the parotid glands (separated as the “better-spared” and “less-spared” glands). The larynx and the esophagus were spared by the midline block of the PA field and by a combination of aperture edge and distal blocking on the posterior oblique fields. The goal was complete midline sparing, but the lateral edges of the larynx received dose because of the specified range uncertainties. Sparing of the parotid glands was manually adjusted by adjusting the aperture edge on the posterior oblique fields and distal blocking on the PA field to achieve the desired gland sparing, while maintaining acceptable coverage of the lymph node regions medial to the parotid gland. The goal for parotid gland sparing was greater than 50% less than 26 Gy in an uninvolved neck. All the 12 patients had a “better-spared” parotid gland. In an involved neck, the cranial portion of the parotid gland was blocked if target coverage of the nodal regions was not compromised. In 2 patients, regional nodal involvement precluded any ipsilateral parotid gland sparing, so that there were 10 rather than 12 “less-spared” glands. Dose to the oral cavity was limited to the dose received because of range uncertainties. There were 5 submandibular glands that were not spared from radiation owing to gross nodal involvement of either ipsilateral level IB or ipsilateral level II of the neck. For the remaining 19 submandibular glands, the goal of submandibular gland sparing was at least 50% less than 30 Gy. Submandibular gland sparing was obtained by distal blocking in both the PA and the posterior oblique fields, with custom editing of the proton compensator to create the desired dose gradient. All 3 of the cervical nodal fields were treated each day. The number of proton treatment fields directed to the primary tumor site depended on the specific geometry of the primary tumor or tumor bed. Primary site fields were geometrically matched to the cervical neck fields to avoid hot spots in the spinal cord or brainstem, facilitated by the reduced beam divergence of protons at these shallow depths when compared with photons.14 To feather the matchline between the 3 lower cervical nodal fields and the primary tumor site fields, the cranial border of the PA and the cranial border of the 2 posterior oblique fields were staggered by 0.5 cm; the PA nodal field could match to a PA primary site field at one level, whereas the RPO nodal field matched to a right anterior oblique or RPO primary site field at another position 0.5 cm cranial or caudal. The median number of fields directed to the primary tumor site concurrent with cervical nodal irradiation was 4 (range: 2 to 6) and the median number of total fields treated per day during treatment of the neck was 6 (range: 5 to 8). Patients receiving up to 8 fields per day were allocated a 45-minute treatment time slot. Patients were treated with daily kilovoltage image guidance using a robotic patient positioner with 61 of freedom for adjustment in alignment.15
Comparative IMRT treatment plans To evaluate the relative sparing of the previously developed proton plans to what might be achievable with IMRT, comparative treatment plans were developed for each case. Sequential IMRT treatment plans were planned using the same dose prescriptions as were used for the sequential proton treatment plans. In developing the proton plans, manual adjustments in individual fields were made, sometimes selectively compromising PTV coverage, to create the desired dose gradients around organs at risk (OARs) and meet the desired constraints. To create a comparable target for IMRT optimization that reflected the volume actually treated by the proton plan, the 95% isodose line (IDL) of the proton treatment plan prescription dose for PTV1 and PTV2 was converted into a structure and cropped to the respective PTV and additionally cropped within 3 mm of the body contour to create a PTV for IMRT optimization. To create plans with comparable target coverage, the IMRT optimization goal was to cover 100% of this PTV for optimization by 95% of the prescription dose. Dose heterogeneity was limited to 110% to 112%. The dose-volume histogram information from the proton plan for each structure, oral cavity, esophagus, larynx, parotids, and submandibular glands, was used as starting inputs for optimization in the IMRT treatment planning process. The IMRT plans were optimized in an effort to match or improve on the sparing of these structures achieved in the proton plans while also respecting the same constraints placed on critical normal structures (brainstem, spinal cord, optic nerves and chiasm, globes, cochlea, etc) as used in each individual proton plan. Between 9 and 10 coplanar IMRT fields were used with 6-MV photons using dose dynamic multi-leaf collimators.
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
3
Fig. 1. A representative example of the 3 fields used for cervical nodal irradiation. The top row shows the beam's eye view of the aperture openings. It should be noted that the cranial border of the posterior-anterior field is staggered 0.5 cm from the cranial border of the 2 posterior oblique fields to vary the matchline with the fields directed to the primary tumor site. The middle row shows the relative isodose lines for each individual field. The bottom row shows the composite dose distribution with all 3 fields in at 2 representative axial positions and in the coronal plane. Select organs that are at risk are visible in contour including the oral cavity, submandibular glands, parotid, larynx, and spinal cord. (Color version of figure is available online.) Dosimetric comparisons The PTV for optimization was used for all comparisons between modalities. Target volume coverage was compared by measuring the percentage of the PTV covered by the 95% IDL. The conformation number (CN) for each PTV was calculated using the methodology of van Riet et al.,16 where CN equals the product of the percentage of the target volume covered by the 95% IDL and the volume of the target covered by the 95% IDL divided by the volume of the 95% IDL. The CN ranges from 0 to 1, with a value of 1 indicating perfect conformation. A homogeneity index (HI) was calculated as the minimum dose to 5% of the PTV divided by the minimum dose to 95% of the PTV, where the ideal value is 1, and an increasing number indicates a less-homogenous dose distribution.17 Dose-volume histogram information for OARs was exported from the treatment planning system by relative volume to create comparative median dose-volume histograms (DVHs) by modality. Summary statistics for each OAR were recorded, including the mean dose and the volume of the OAR. Mean dose was selected as it is the dosimetric parameter that most consistently correlated with late head and neck toxicity outcomes including parotid dysfunction and dysphagia18-21 as well as acute mucositis and dysphagia.22 An approximation of the integral dose to OARs was calculated as the mean dose in Gy multiplied by the volume in cm3, assuming a uniform density of OARs. Comparison of the CN, HI, mean dose, and integral dose to OARs between treatment modalities was made using the matched-pair Wilcoxon signed rank test. Two-sided p values are reported, and a p o 0.05 was considered statistically significant.
Results Table 2 compares the target coverage for the proton and the IMRT plans as well as the CN and the HI. Although the IMRT plans
were optimized to provide equivalent target coverage, the proton plans had a small statistically significant improvement in target coverage. Statistically significant differences in CN were seen; for PTV1, CN was modestly higher for the IMRT plans, whereas for PTV2, the CN was modestly higher for the proton plans. There was no significant difference in the HI between modalities for PTV1, but for PTV2, the IMRT plans were on average slightly more homogenous. The median DVHs for the previously treated proton plans and the comparative IMRT plans are reviewed in Fig. 2. The median DVH for patients receiving elective nodal irradiation (n ¼ 8) are reported separately from those receiving higher doses for gross nodal disease (n ¼ 4).
Table 2 Median target coverage metrics by modality PTV1
Target coverage factor Conformation number Homogeneity index
PTV2
IMRT
Proton
p Value
IMRT
Proton
p Value
97.33 0.52 1.54
99.96 0.45 1.48
0.002 0.041 0.193
98.27 0.55 1.07
99.87 0.64 1.14
0.028 0.034 0.026
4
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
Fig. 2. A representative example of comparative treatment plans using relative dose color wash. The top row shows 2 representative axial slices and a coronal view from a 3D conformal proton plan, while the bottom row shows the same positions in the comparative intensity-modulated photon plan. (Color version of figure is available online.)
An example of comparative treatment plans is shown in Fig. 3. Table 3 summarizes the comparisons of mean and integral dose to OARs by treatment modality. When compared with the IMRT plans, the proton treatment plan had a statistically significant lower mean dose to all OARs, except the lesser-spared parotid gland and the spared submandibular glands, where there was no statistically significant difference by treatment modality. Comparing the integral dose to OARs by modality, the proton treatment plan had a statistically significant lower integral dose for all OARs, except the spared submandibular gland, for which there was no statistically significant difference by treatment modality.
Discussion Head and neck radiation therapy is associated with significant toxicity. Hospitalization for treatment-related toxicity is common even in the era of IMRT.23 Mucositis in particular is a frequent and often severe toxicity of head and neck radiotherapy that occurs at a relatively low dose.24 Dose-response relationships have been published for the probability of developing grade 3 mucositis during head and neck chemoradiation, suggesting a 15% risk of grade 3 mucositis with a mean oral cavity dose of 10 Gy.22 Many efforts to reduce the toxicities of head and neck radiation have focused on reduction in radiation dose to structures including the parotid glands,25,26 oral cavity,26,27 submandibular glands (when level I lymph nodes are not a target),26 floor of the mouth muscles,28 and esophagus29 among other structures. As more data accumulate tying OARs to clinical toxicity end points, IMRT optimization becomes increasingly complex and is ultimately constrained by the physical properties of photons in that dose reductions to an OAR necessitate a redistribution of dose in the patient. Proton therapy offers opportunities for not only dose attenuation but also complete or partial dose avoidance of OARs. Although both the proton and the IMRT plans generated in our study were clinically acceptable and did not exceed standard critical normal tissue tolerances, our technique of 3-dimensional (3D) conformal proton therapy with posterior-oriented beams for treatment of the cervical lymphatics provided dose avoidance of structures such as the anterior oral cavity and the midline larynx. Areas that receive no radiation dose typically do not manifest radiation toxicity. Given the very low dose:volume exposure of the oral cavity in the proton treatment plans, one would anticipate a
clinical correlation with reduced incidence of oral mucositis and xerostomia. The mean radiation dose to the oral cavity, which harbors innumerable minor salivary glands, has been found to be a significant, independent predictor of xerostomia.30 The mean and integral doses to the “better-spared” parotid gland were also lower in our series with proton therapy. Existing parotid gland normal tissue complication probability curves suggest that these reductions could further reduce the risk of late xerostomia.20 Although predicted by available models, we regard these potential differences in toxicity as hypothetical. Further evaluation of proton therapy to assess for potential reductions in clinical toxicities is warranted. A confounding factor in potential future clinical assessment is the independent risk of mucositis associated with many chemotherapy agents. Reviewing the median DVH curves, the primary comparative advantage of the proton plans lies in the lower-dose regions, echoing other dosimetric comparisons.31 In the higher-dose regions, the IMRT plans were more conformal for some structures. We believe this is due to a combination of factors, the largest of which is the range uncertainty necessarily applied in the proton plans, where the beam intentionally overshoots the target volume to account for uncertainties in the calculated proton range.32 This results in portions of OARs adjacent to targets receiving the prescription dose to ensure that the target is covered, as seen in the representative example of Fig. 2, where a margin of proton dose intentionally extends into the posterior submandibular glands and the lateral larynx. Another consideration reflects the choices made during proton planning. For example, in sparing the parotid glands, the focus was on maximizing a volume spared below 26-Gy dose, whereas the volume of gland receiving greater than 26-Gy dose was not optimized based on the belief that this portion of the gland was already destined to be dysfunctional. This highlights a final consideration, which is the 1-sided nature of the comparative dosimetry. The proton plans were manually edited to achieve desired constraints and were prospectively judged as clinically acceptable without contemporaneous comparison with alternate treatment plans. The IMRT plans were developed with inverse optimization to match or improve on the retrospective proton dosimetry, whereas the proton plans were not reoptimized in response to assess whether further improvements could be made. For the purpose of comparison, the IMRT plans used the same treatment volumes, while in practice, we would apply a larger PTV expansion for IMRT owing to differences in patient
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
5
Fig. 3. The median dose-volume histograms (DVH) for the proton and the IMRT comparative treatment plans for target coverage and select ORAs. The median DVH for patients receiving elective nodal irradiation (n ¼ 8) are reported separately from those receiving higher doses for gross nodal disease (n ¼ 4).
setup at the proton center, including a robotic positioner with 61 of freedom, repeated orthogonal image setup for each treatment field, and real-time treatment setup verification of each field. However, the primary purpose of our study was to report our technique for use of proton therapy in comprehensive head and neck irradiation, with comparative dosimetry generated to provide some reference for assessing the resultant DVHs. Some limitations of our data set must be acknowledged. The midline sparing of the larynx and the esophagus seen in these patients is not expected in different patient cohorts, such as those with grossly involved pretracheal level VI lymph nodes or those with laryngeal, hypopharyngeal, or thyroid cancers where the primary tumor is itself midline. We have not clinically
implemented or formally evaluated the role of proton therapy in these patient populations. Our data set is insufficient to assess the dosimetric effect of varied volumes and locations of cervical nodal disease. Our analysis does not include formal assessment of plan robustness against potential deviations in patient setup and other treatment uncertainties, which are assumed to be adequately accounted for in the PTV margin, proton range uncertainty, and compensator smearing values applied during treatment planning. It is noteworthy that no critical structures (brainstem, spinal cord, and optic apparatus) depended on distal blocking to meet dose constraints. Although distal blocking was used to spare the submandibular glands and portions of the parotid glands, these structures lie in the plane of the incident posterior beams, such
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
6 Table 3 Radiation dose to organs of interest by modality Region of interest
Oral cavity Larynx Esophagus Better-spared parotid (n ¼ 12) Lesser-spared Parotid (n ¼ 10) Spared submandibular gland (n ¼ 19)
Mean of mean dose
Median of integral dose
IMRT
Proton
p Value
IMRT
Proton
p Value
18.48 23.73 13.23 21.09 20.84 23.82
5.98 18.37 8.49 15.22 19.22 21.47
0.002 0.002 0.002 0.004 0.203 0.117
1924.55 1176.30 146.60 506.25 581.50 127.40
532.20 868.75 77.15 352.90 496.95 114.80
0.002 0.002 0.002 0.004 0.028 0.091
Dose in Gy. Integral Dose in Gy cm3.
that uncertainties in vertical patient positioning would have minimal effect on the position of the resultant dose distribution.33 When compared with pencil-beam scanning and intensitymodulated proton therapy, uniform scanning is less susceptible to patient setup and range uncertainties or small misalignments in the proton beam.13 Field matching was used, which is associated with increased dosimetric complexity and quality assurance requirements. A similar concern arises in matching low anterior neck fields to IMRT fields. We mitigated the issues of field matching by an intrafractional feather of the matchline so that each side of the neck was matched to the primary tumor site at 2 matchlines, both treated each day. We have previously reported a formal assessment of the accuracy of field matching using our robotic patient positioner and found good agreement between treatment planning calculations and radiographic film and ionization chamber measurements.34 Most published clinical experiences with proton therapy involved 3D conformal treatment planning in which the desired dosimetry is achieved by manual iterative adjustments in the shape of individual field apertures (blocks) and compensators, or patient-specific devices (PSDs). With uniform scanning, as used in our patients, a uniform dose is delivered across the treatment field with a uniform spread-out Bragg peak (SOBP). This results in 2 limitations to the application of our technique for cervical nodal irradiation. With uniform scanning, the SOBP is chosen to cover the widest part of the target within each field. For example, in the upper cervical neck, the target may require an 8-cm SOBP to cover from the anterior to posterior neck, whereas in the supraclavicular fossa, only 3 cm may be required. With uniform scanning, the SOBP does not vary over the field, and consequently, the larger SOBP required for coverage in the upper neck results in poor posterior dose conformality in the inferior neck, where a smaller SOBP would be preferred. One option to address this would be to subdivide the upper and lower neck, treating through 2 different matched fields that use SOBPs appropriate for the target at each level. We did not elect to do this, as the gain in posterior neck target conformality did not seem to justify the increased complexity of additional matchlines nor the increased cost and treatment time that 2 to 3 additional low neck fields would require. Instead, we accepted that additional radiation would be delivered to deep cervical musculature and lung apices as a trade-off for the improved oral cavity sparing and midline sparing of the esophagus and the larynx. This is not dissimilar to the homogenous dose delivered to the cervical muscles with traditional x-ray radiation using parallel-opposed lateral fields and a matched low anterior neck field. It is possible that sacrificing posterior conformality to the cervical nodal target volumes could increase the risk of neck soft tissue fibrosis. Although we are unaware of any data that suggest that IMRT may reduce the risk of cervical soft tissue fibrosis, it would be desirable to improve on the posterior target conformality in the lower cervical neck using more advanced proton delivery techniques.
A second limitation in uniform scanning is that there are few options to create nonuniform dose intensities, for example, to treat 1 portion of the target at a lower dose per fraction than another portion. A option is to use a field-in-field technique, which is again cumbersome owing to the requirement for additional PSDs, or by selective exclusion of the desired area from some but not all treatment fields, weighted in a fashion to deliver the desired dose per fraction to each area. Given the limitations of these 2 approaches, a simultaneous integrated boost technique to multiple nodal targets is impractical using 3D conformal proton therapy, prompting the use of sequential shrinking field plans. A rapidly emerging proton treatment approach is pencil-beam scanning, in which the proton beam is magnetically steered to cover the target volume without the requirement of PSDs.35 Pencil-beam scanning allows for intensity-modulated proton therapy of various techniques, which have been clinically implemented in treatment of head and neck tumors.36 Using multiple-field optimization, simultaneous integrated boost techniques can be used with proton therapy.37 Proton pencil-beam scanning is therefore anticipated to improve on and replace our approach of 3D planning with uniform scanning. For example, it would have the advantage of varying the SOBP to match the varied depth required at different levels of the cervical neck, addressing the issue of poor posterior target conformality in the low cervical neck seen with our technique, where each beam has a fixed SOBP width. However, our technique of 3D conformal proton therapy with uniform scanning has some advantages when compared with multifield optimization, including being a well-established delivery system that is less susceptible to patient setup and range uncertainties or small misalignments in the proton beam, and has simpler quality assurance.13,38
Conclusions A technique for cervical nodal irradiation using 3D conformal proton therapy with uniform scanning was developed and clinically implemented. Use of proton therapy for cervical nodal irradiation resulted in large volume of dose avoidance to the oral cavity and low dose exposure to midline structures of the larynx and the esophagus, with lower mean and integral dose to assessed OARs when compared with competing IMRT plans.
References 1. Schulz-Ertner, D.; Tsujii, H. Particle radiation therapy using proton and heavier ion beams. J. Clin. Oncol. 25:953–64; 2007. 2. Patel, S.H.; Wang, Z.; Wong, W.W.; et al. Charged particle therapy versus photon therapy for paranasal sinus and nasal cavity malignant diseases: A systematic review and meta-analysis. Lancet Oncol. 15:1027–38; 2014. 3. Zenda, S.; Kohno, R.; Kawashima, M.; et al. Proton beam therapy for unresectable malignancies of the nasal cavity and paranasal sinuses. Int. J. Radiat. Oncol. Biol. Phys. 81:1473–8; 2011.
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]] 4. Truong, M.T.; Kamat, U.R.; Liebsch, N.J.; et al. Proton radiation therapy for primary sphenoid sinus malignancies: Treatment outcome and prognostic factors. Head Neck 31:1297–308; 2009. 5. Fitzek, M.M.; Thornton, A.F.; Varvares, M.; et al. Neuroendocrine tumors of the sinonasal tract. Results of a prospective study incorporating chemotherapy, surgery, and combined proton-photon radiotherapy. Cancer 94:2623–34; 2002. 6. Chan, A.W.; Liebsch, N.J. Proton radiation therapy for head and neck cancer. J. Surg. Oncol. 97:697–700; 2008. 7. Mock, U.; Georg, D.; Bogner, J.; et al. Treatment planning comparison of conventional, 3D conformal, and intensity-modulated photon (IMRT) and proton therapy for paranasal sinus carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 58:147–54; 2004. 8. Lomax, A.J.; Goitein, M.; Adams, J. Intensity modulation in radiotherapy: Photons versus protons in the paranasal sinus. Radiother. Oncol. 66:11–8; 2003. 9. Herr, M.W.; Sethi, R.K.; Meier, J.C.; et al. Esthesioneuroblastoma: An update on the massachusetts eye and ear infirmary and massachusetts general hospital experience with craniofacial resection, proton beam radiation, and chemotherapy. J. Neurol. Surg. B Skull Base 75:58–64; 2014. 10. Fukumitsu, N.; Okumura, T.; Mizumoto, M.; et al. Outcome of T4 (International Union Against Cancer Staging System, 7th edition) or recurrent nasal cavity and paranasal sinus carcinoma treated with proton beam. Int. J. Radiat. Oncol. Biol. Phys. 83:704–11; 2012. 11. Buchsbaum, J.C.; Besemer, A.; Simmons, J.; et al. Supine proton beam craniospinal radiotherapy using a novel tabletop adapter. Med. Dosim. 38:70–6; 2013. 12. Farr, J.B.; Mascia, A.E.; Hsi, W.C.; et al. Clinical characterization of a proton beam continuous uniform scanning system with dose layer stacking. Med. Phys. 35:4945–54; 2008. 13. Anferov, V.A. Scan pattern optimization for uniform proton beam scanning. Med. Phys. 36:3560–7; 2009. 14. Suit, H.; Goldberg, S.; Niemierko, A.; et al. Proton beams to replace photon beams in radical dose treatments. Acta. Oncol. 42:800–8; 2003. 15. Allgower, C.E.; Schreuder, A.N.; Farr, J.B.; et al. Experiences with an application of industrial robotics for accurate patient positioning in proton radiotherapy. Int. J. Med. Robot. 3:72–81; 2007. 16. van't Riet, A.; Mak, A.C.; Moerland, M.A.; et al. A conformation number to quantify the degree of conformality in brachytherapy and external beam irradiation: Application to the prostate. Int. J. Radiat. Oncol. Biol. Phys. 37:731–6; 1997. 17. Semenenko, V.A.; Reitz, B.; Day, E.; et al. Evaluation of a commercial biologically based IMRT treatment planning system. Med. Phys. 35:5851–60; 2008. 18. Roesink, J.M.; Moerland, M.A.; Battermann, J.J.; et al. Quantitative dose-volume response analysis of changes in parotid gland function after radiotherapy in the head-and-neck region. Int. J. Radiat. Oncol. Biol. Phys. 51:938–46; 2001. 19. Batth, S.S.; Caudell, J.J.; Chen, A.M. Practical considerations in reducing swallowing dysfunction following concurrent chemoradiotherapy with intensity-modulated radiotherapy for head and neck cancer. Head Neck 36:291–8; 2014. 20. Dijkema, T.; Raaijmakers, C.P.; Ten Haken, R.K.; et al. Parotid gland function after radiotherapy: The combined Michigan and Utrecht experience. Int. J. Radiat. Oncol. Biol. Phys. 78:449–53; 2010. 21. Dirix, P.; Nuyts, S. Evidence-based organ-sparing radiotherapy in head and neck cancer. Lancet Oncol. 11:85–91; 2010.
7
22. Bhide, S.A.; Gulliford, S.; Schick, U.; et al. Dose-response analysis of acute oral mucositis and pharyngeal dysphagia in patients receiving induction chemotherapy followed by concomitant chemo-IMRT for head and neck cancer. Radiother. Oncol. 103:88–91; 2012. 23. Ling, D.C.; Kabolizadeh, P.; Heron, D.E.; et al. Incidence of hospitalization in patients with head and neck cancer treated with intensity-modulated radiation therapy. Head Neck; 2015. 24. Trotti, A.; Bellm, L.A.; Epstein, J.B.; et al. Mucositis incidence, severity and associated outcomes in patients with head and neck cancer receiving radiotherapy with or without chemotherapy: A systematic literature review. Radiother. Oncol. 66:253–62; 2003. 25. Nutting, C.M.; Morden, J.P.; Harrington, K.J.; et al. Parotid-sparing intensity modulated versus conventional radiotherapy in head and neck cancer (PARSPORT): A phase 3 multicentre randomised controlled trial. Lancet Oncol. 12:127–36; 2011. 26. Little, M.; Schipper, M.; Feng, F.Y.; et al. Reducing xerostomia after chemo-IMRT for head-and-neck cancer: Beyond sparing the parotid glands. Int. J. Radiat. Oncol. Biol. Phys. 83:1007–14; 2012. 27. Schwartz, D.L.; Hutcheson, K.; Barringer, D.; et al. Candidate dosimetric predictors of long-term swallowing dysfunction after oropharyngeal intensity-modulated radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 78:1356–65; 2010. 28. Kumar, R.; Madanikia, S.; Starmer, H.; et al. Radiation dose to the floor of mouth muscles predicts swallowing complications following chemoradiation in oropharyngeal squamous cell carcinoma. Oral Oncol. 50:65–70; 2014. 29. Eisbruch, A.; Kim, H.M.; Feng, F.Y.; et al. Chemo-IMRT of oropharyngeal cancer aiming to reduce dysphagia: Swallowing organs late complication probabilities and dosimetric correlates. Int. J. Radiat. Oncol. Biol. Phys. 81:e93–9; 2011. 30. Eisbruch, A.; Kim, H.M.; Terrell, J.E.; et al. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 50:695–704; 2001. 31. Baumert, B.G.; Norton, I.A.; Lomax, A.J.; et al. Dose conformation of intensitymodulated stereotactic photon beams, proton beams, and intensity-modulated proton beams for intracranial lesions. Int. J. Radiat. Oncol. Biol. Phys. 60:1314–24; 2004. 32. Paganetti, H. Range uncertainties in proton therapy and the role of Monte Carlo simulations. Phys. Med. Biol. 57:R99–117; 2012. 33. Thomas, S.J. Margins for treatment planning of proton therapy. Phys. Med. Biol. 51:1491–501; 2006. 34. Farr, J.B.; O'Ryan-Blair, A.; Jesseph, F.; et al. Validation of dosimetric field matching accuracy from proton therapy using a robotic patient positioning system. J. Appl. Clin. Med. Phys 11:3015; 2010. 35. Schippers, J.M.; Lomax, A.J. Emerging technologies in proton therapy. Acta Oncol. 50:838–50; 2011. 36. Frank, S.J.; Cox, J.D.; Gillin, M.; et al. Multifield optimization intensity modulated proton therapy for head and neck tumors: A translation to practice. Int. J. Radiat. Oncol. Biol. Phys. 89:846–53; 2014. 37. Zhu, X.R.; Poenisch, F.; Li, H.; et al. A single-field integrated boost treatment planning technique for spot scanning proton therapy. Radiat. Oncol. 9:202; 2014. 38. Liu, W.; Mohan, R.; Park, P.; et al. Dosimetric benefits of robust treatment planning for intensity modulated proton therapy for base-of-skull cancers. Pract. Radiat. Oncol. 4:384–91; 2014.
Medical Dosimetry journal homepage: www.meddos.org
Technique for comprehensive head and neck irradiation using 3-dimensional conformal proton therapy Mark W. McDonald, M.D.,n† Alexander S. Walter, B.S., R.T.(T.),n and Ted A. Hoene, M.S., R.T.(T.), C.M.D.† Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, IN; and †Indiana University Health Proton Therapy Center, Bloomington, IN
*
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 December 2014 Received in revised form 18 February 2015 Accepted 12 April 2015
Owing to the technical and logistical complexities of matching photon and proton treatment modalities, we developed and implemented a technique of comprehensive head and neck radiation using 3dimensional (3D) conformal proton therapy. A monoisocentric technique was used with a 30-cm snout. Cervical lymphatics were treated with 3 fields: a posterior-anterior field with a midline block and a right and a left posterior oblique field. The matchline of the 3 cervical nodal fields with the primary tumor site fields was staggered by 0.5 cm. Comparative intensity-modulated photon plans were later developed for 12 previously treated patients to provide equivalent target coverage, while matching or improving on the proton plans' sparing of organs at risk (OARs). Dosimetry to OARs was evaluated and compared by treatment modality. Comprehensive head and neck irradiation using proton therapy yielded treatment plans with significant dose avoidance of the oral cavity and midline neck structures. When compared with the generated intensity-modulated radiation therapy (IMRT) plans, the proton treatment plans yielded statistically significant reductions in the mean and integral radiation dose to the oral cavity, larynx, esophagus, and the maximally spared parotid gland. There was no significant difference in mean dose to the lesser-spared parotid gland by treatment modality or in mean or integral dose to the spared submandibular glands. A technique for cervical nodal irradiation using 3D conformal proton therapy with uniform scanning was developed and clinically implemented. Use of proton therapy for cervical nodal irradiation resulted in large volume of dose avoidance to the oral cavity and low dose exposure to midline structures of the larynx and the esophagus, with lower mean and integral dose to assessed OARs when compared with competing IMRT plans. & 2015 American Association of Medical Dosimetrists.
Keywords: Proton therapy Head and neck IMRT Dosimetry Nodal irradiation
Introduction Proton therapy is a modality of radiotherapy that differs from xray–based modalities by a finite range of dose deposition within tissue, delivering radiation to the depth of the target and then terminating, with essentially no exit dose radiation to normal tissue beyond the target.1 Proton therapy has been used in the treatment of some uncommon head and neck tumors, particularly paranasal sinus2-5 and nasopharyngeal tumors,6 where the proximity of the optic apparatus can pose challenges to providing adequate target coverage by sufficient dose without undue risk of complications.7,8
Reprint requests to: Mark W. McDonald, M.D., 535 Barnhill Dr, RT 041, Indianapolis, IN 46202-5289. E-mail: [email protected] http://dx.doi.org/10.1016/j.meddos.2015.04.004 0958-3947/Copyright Ó 2015 American Association of Medical Dosimetrists
Patients treated with proton therapy for head and neck tumors where cervical nodal irradiation is indicated have typically received a portion of their treatment with photons during cervical nodal irradiation or received concurrent photon neck radiation matched to the proton therapy plan delivered to the primary tumor site.5,9,10 Sequential use of the 2 modalities may erode the dosimetric advantage of proton therapy in the treatment of the primary tumor site, and concurrent use of both the modalities creates increased uncertainties across the matchline with patient transfer between machines, the possibility of matching through high-risk nodal regions or gross nodal disease, and requirement of treatment room time on 2 machines. At our proton therapy treatment site, a linear accelerator is not available in the same building, leading to the additional complexity of daily transport of patient immobilization devices between 2 treatment sites. Because of these technical and logistical complexities, in 2012,
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
2
we developed and implemented a technique of comprehensive head and neck radiation for patients with sinonasal tumors using proton therapy. Here, we review the treatment technique and dosimetric comparison with intensity-modulated photon therapy. Clinical data on patient tolerance are reported separately.
Methods and Materials In an institution review board–approved retrospective study, we analyzed the dosimetry in 12 patients with sinonasal tumors treated with proton therapy to the bilateral cervical lymphatics and primary tumor site between 2012 and 2014. Patient and treatment characteristics are listed in Table 1.
Patient simulation Patients were immobilized with an Alpha Cradle and a thermoplastic mask attached to a carbon fiber table originally designed at our center for craniospinal irradiation.11 Shoulder retractors (CIVCO, Coralville, IA) were applied with gentle traction to aid in reproducibility of shoulder positioning. A computed tomography scan was acquired with 1-mm slice thickness from the top of the head to the carina. Diagnostic imaging was coregistered to aid in delineation of target volumes and critical normal structures. Treatment planning of both the proton treatment plans and the comparative intensity-modulated radiation therapy (IMRT) treatment plans was performed in Eclipse version 11 (Palo Alto, CA).
Proton treatment planning and delivery Proton treatment was designed for delivery on a 3601 gantry with uniform beam scanning and dose layer stacking.12 Details of our proton therapy treatment system have previously been published.13 A monoisocentric technique was used with a 30-cm snout. The cervical lymphatics were treated with 3 fields: a posterioranterior (PA) field with a midline block over the larynx, esophagus, and spinal cord. The cranial portion of the PA field was left unblocked to provide dose to the medial retropharyngeal lymph nodes. The cervical nodal regions of the right neck were treated with a right posterior oblique (RPO) and of the left neck with a left posterior oblique. The 2 posterior oblique fields were angled 201 to 301 from PA in either direction. On these posterior oblique fields, the medial aperture edge was used to block the spinal cord, while the lateral aperture edge was trimmed close to the body contour to provide some skin sparing on the composite plan (Fig. 1). These cervical nodal fields were weighted differently. For example, when the fraction size was 1.8 Gy, the PA field might deliver 1 Gy and each posterior oblique might deliver 0.8 Gy. When gross nodal disease was present, a shrinking field technique was used to boost the involved nodal region, typically with an anterior and posterior oblique field combination. Table 1 Patient and treatment characteristics Patient age (range)
42.5 y (16 to 71)
Primary site Ethmoids Nasopharynx Maxilla Nasal
9 1 1 1
Histology Esthesioneuroblastoma Sinonasal undifferentiated carcinoma Squamous cell carcinoma Nasopharyngeal undifferentiated carcinoma
5 5 1 1
T category T4 T3 T2
9 2 1
N category N0 N1 N2
8 1 3
Cervical nodal irradiation Elective dose median (range) Definitive dose median (range)
50 Gy (45 to 54) 72.9 Gy (70 to 75.6)
Primary site Postoperative dose median (range) Definitive dose median (range)
66 Gy (63 to 72) 74 Gy (70.2 to 75.6)
At our institution, we elected to use planning target volumes (PTVs) in proton treatment planning, combining concepts of a uniform geometric volume expansion for setup uncertainties with additional beam-specific uncertainties. Above the level of C2, 2 mm of geometric expansion was used to create the PTV, based on institutional assessment of setup error using daily image-guided alignment, which was repeated for each treatment field and checked by a physician before delivery of each field. Below C2, 5 mm of expansion was used owing to increased setup variance expected for the lower cervical neck. An additional 2 mm of proximal and distal margin was added beyond the PTV for proton range uncertainties. A median of 4 mm of compensator smearing was applied (range: 3 to 5 mm). As we incorporated nodal irradiation with proton therapy, an emphasis was placed on sparing the following structures: the oral cavity (an avoidance structure whose lateral and anterior borders were defined by the mandible, cranial by the maxillary bone and inclusive of any air cavity above the oral tongue, inferior including the floor of mouth, and posterior including the base of tongue), the esophagus (drawn from the cricoid to the thoracic inlet), the larynx (inclusive of the thyroid cartilage and postcricoid space abutting the vertebral body), the submandibular glands (unless ipsilateral level IB nodes were treated), and the parotid glands (separated as the “better-spared” and “less-spared” glands). The larynx and the esophagus were spared by the midline block of the PA field and by a combination of aperture edge and distal blocking on the posterior oblique fields. The goal was complete midline sparing, but the lateral edges of the larynx received dose because of the specified range uncertainties. Sparing of the parotid glands was manually adjusted by adjusting the aperture edge on the posterior oblique fields and distal blocking on the PA field to achieve the desired gland sparing, while maintaining acceptable coverage of the lymph node regions medial to the parotid gland. The goal for parotid gland sparing was greater than 50% less than 26 Gy in an uninvolved neck. All the 12 patients had a “better-spared” parotid gland. In an involved neck, the cranial portion of the parotid gland was blocked if target coverage of the nodal regions was not compromised. In 2 patients, regional nodal involvement precluded any ipsilateral parotid gland sparing, so that there were 10 rather than 12 “less-spared” glands. Dose to the oral cavity was limited to the dose received because of range uncertainties. There were 5 submandibular glands that were not spared from radiation owing to gross nodal involvement of either ipsilateral level IB or ipsilateral level II of the neck. For the remaining 19 submandibular glands, the goal of submandibular gland sparing was at least 50% less than 30 Gy. Submandibular gland sparing was obtained by distal blocking in both the PA and the posterior oblique fields, with custom editing of the proton compensator to create the desired dose gradient. All 3 of the cervical nodal fields were treated each day. The number of proton treatment fields directed to the primary tumor site depended on the specific geometry of the primary tumor or tumor bed. Primary site fields were geometrically matched to the cervical neck fields to avoid hot spots in the spinal cord or brainstem, facilitated by the reduced beam divergence of protons at these shallow depths when compared with photons.14 To feather the matchline between the 3 lower cervical nodal fields and the primary tumor site fields, the cranial border of the PA and the cranial border of the 2 posterior oblique fields were staggered by 0.5 cm; the PA nodal field could match to a PA primary site field at one level, whereas the RPO nodal field matched to a right anterior oblique or RPO primary site field at another position 0.5 cm cranial or caudal. The median number of fields directed to the primary tumor site concurrent with cervical nodal irradiation was 4 (range: 2 to 6) and the median number of total fields treated per day during treatment of the neck was 6 (range: 5 to 8). Patients receiving up to 8 fields per day were allocated a 45-minute treatment time slot. Patients were treated with daily kilovoltage image guidance using a robotic patient positioner with 61 of freedom for adjustment in alignment.15
Comparative IMRT treatment plans To evaluate the relative sparing of the previously developed proton plans to what might be achievable with IMRT, comparative treatment plans were developed for each case. Sequential IMRT treatment plans were planned using the same dose prescriptions as were used for the sequential proton treatment plans. In developing the proton plans, manual adjustments in individual fields were made, sometimes selectively compromising PTV coverage, to create the desired dose gradients around organs at risk (OARs) and meet the desired constraints. To create a comparable target for IMRT optimization that reflected the volume actually treated by the proton plan, the 95% isodose line (IDL) of the proton treatment plan prescription dose for PTV1 and PTV2 was converted into a structure and cropped to the respective PTV and additionally cropped within 3 mm of the body contour to create a PTV for IMRT optimization. To create plans with comparable target coverage, the IMRT optimization goal was to cover 100% of this PTV for optimization by 95% of the prescription dose. Dose heterogeneity was limited to 110% to 112%. The dose-volume histogram information from the proton plan for each structure, oral cavity, esophagus, larynx, parotids, and submandibular glands, was used as starting inputs for optimization in the IMRT treatment planning process. The IMRT plans were optimized in an effort to match or improve on the sparing of these structures achieved in the proton plans while also respecting the same constraints placed on critical normal structures (brainstem, spinal cord, optic nerves and chiasm, globes, cochlea, etc) as used in each individual proton plan. Between 9 and 10 coplanar IMRT fields were used with 6-MV photons using dose dynamic multi-leaf collimators.
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
3
Fig. 1. A representative example of the 3 fields used for cervical nodal irradiation. The top row shows the beam's eye view of the aperture openings. It should be noted that the cranial border of the posterior-anterior field is staggered 0.5 cm from the cranial border of the 2 posterior oblique fields to vary the matchline with the fields directed to the primary tumor site. The middle row shows the relative isodose lines for each individual field. The bottom row shows the composite dose distribution with all 3 fields in at 2 representative axial positions and in the coronal plane. Select organs that are at risk are visible in contour including the oral cavity, submandibular glands, parotid, larynx, and spinal cord. (Color version of figure is available online.) Dosimetric comparisons The PTV for optimization was used for all comparisons between modalities. Target volume coverage was compared by measuring the percentage of the PTV covered by the 95% IDL. The conformation number (CN) for each PTV was calculated using the methodology of van Riet et al.,16 where CN equals the product of the percentage of the target volume covered by the 95% IDL and the volume of the target covered by the 95% IDL divided by the volume of the 95% IDL. The CN ranges from 0 to 1, with a value of 1 indicating perfect conformation. A homogeneity index (HI) was calculated as the minimum dose to 5% of the PTV divided by the minimum dose to 95% of the PTV, where the ideal value is 1, and an increasing number indicates a less-homogenous dose distribution.17 Dose-volume histogram information for OARs was exported from the treatment planning system by relative volume to create comparative median dose-volume histograms (DVHs) by modality. Summary statistics for each OAR were recorded, including the mean dose and the volume of the OAR. Mean dose was selected as it is the dosimetric parameter that most consistently correlated with late head and neck toxicity outcomes including parotid dysfunction and dysphagia18-21 as well as acute mucositis and dysphagia.22 An approximation of the integral dose to OARs was calculated as the mean dose in Gy multiplied by the volume in cm3, assuming a uniform density of OARs. Comparison of the CN, HI, mean dose, and integral dose to OARs between treatment modalities was made using the matched-pair Wilcoxon signed rank test. Two-sided p values are reported, and a p o 0.05 was considered statistically significant.
Results Table 2 compares the target coverage for the proton and the IMRT plans as well as the CN and the HI. Although the IMRT plans
were optimized to provide equivalent target coverage, the proton plans had a small statistically significant improvement in target coverage. Statistically significant differences in CN were seen; for PTV1, CN was modestly higher for the IMRT plans, whereas for PTV2, the CN was modestly higher for the proton plans. There was no significant difference in the HI between modalities for PTV1, but for PTV2, the IMRT plans were on average slightly more homogenous. The median DVHs for the previously treated proton plans and the comparative IMRT plans are reviewed in Fig. 2. The median DVH for patients receiving elective nodal irradiation (n ¼ 8) are reported separately from those receiving higher doses for gross nodal disease (n ¼ 4).
Table 2 Median target coverage metrics by modality PTV1
Target coverage factor Conformation number Homogeneity index
PTV2
IMRT
Proton
p Value
IMRT
Proton
p Value
97.33 0.52 1.54
99.96 0.45 1.48
0.002 0.041 0.193
98.27 0.55 1.07
99.87 0.64 1.14
0.028 0.034 0.026
4
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
Fig. 2. A representative example of comparative treatment plans using relative dose color wash. The top row shows 2 representative axial slices and a coronal view from a 3D conformal proton plan, while the bottom row shows the same positions in the comparative intensity-modulated photon plan. (Color version of figure is available online.)
An example of comparative treatment plans is shown in Fig. 3. Table 3 summarizes the comparisons of mean and integral dose to OARs by treatment modality. When compared with the IMRT plans, the proton treatment plan had a statistically significant lower mean dose to all OARs, except the lesser-spared parotid gland and the spared submandibular glands, where there was no statistically significant difference by treatment modality. Comparing the integral dose to OARs by modality, the proton treatment plan had a statistically significant lower integral dose for all OARs, except the spared submandibular gland, for which there was no statistically significant difference by treatment modality.
Discussion Head and neck radiation therapy is associated with significant toxicity. Hospitalization for treatment-related toxicity is common even in the era of IMRT.23 Mucositis in particular is a frequent and often severe toxicity of head and neck radiotherapy that occurs at a relatively low dose.24 Dose-response relationships have been published for the probability of developing grade 3 mucositis during head and neck chemoradiation, suggesting a 15% risk of grade 3 mucositis with a mean oral cavity dose of 10 Gy.22 Many efforts to reduce the toxicities of head and neck radiation have focused on reduction in radiation dose to structures including the parotid glands,25,26 oral cavity,26,27 submandibular glands (when level I lymph nodes are not a target),26 floor of the mouth muscles,28 and esophagus29 among other structures. As more data accumulate tying OARs to clinical toxicity end points, IMRT optimization becomes increasingly complex and is ultimately constrained by the physical properties of photons in that dose reductions to an OAR necessitate a redistribution of dose in the patient. Proton therapy offers opportunities for not only dose attenuation but also complete or partial dose avoidance of OARs. Although both the proton and the IMRT plans generated in our study were clinically acceptable and did not exceed standard critical normal tissue tolerances, our technique of 3-dimensional (3D) conformal proton therapy with posterior-oriented beams for treatment of the cervical lymphatics provided dose avoidance of structures such as the anterior oral cavity and the midline larynx. Areas that receive no radiation dose typically do not manifest radiation toxicity. Given the very low dose:volume exposure of the oral cavity in the proton treatment plans, one would anticipate a
clinical correlation with reduced incidence of oral mucositis and xerostomia. The mean radiation dose to the oral cavity, which harbors innumerable minor salivary glands, has been found to be a significant, independent predictor of xerostomia.30 The mean and integral doses to the “better-spared” parotid gland were also lower in our series with proton therapy. Existing parotid gland normal tissue complication probability curves suggest that these reductions could further reduce the risk of late xerostomia.20 Although predicted by available models, we regard these potential differences in toxicity as hypothetical. Further evaluation of proton therapy to assess for potential reductions in clinical toxicities is warranted. A confounding factor in potential future clinical assessment is the independent risk of mucositis associated with many chemotherapy agents. Reviewing the median DVH curves, the primary comparative advantage of the proton plans lies in the lower-dose regions, echoing other dosimetric comparisons.31 In the higher-dose regions, the IMRT plans were more conformal for some structures. We believe this is due to a combination of factors, the largest of which is the range uncertainty necessarily applied in the proton plans, where the beam intentionally overshoots the target volume to account for uncertainties in the calculated proton range.32 This results in portions of OARs adjacent to targets receiving the prescription dose to ensure that the target is covered, as seen in the representative example of Fig. 2, where a margin of proton dose intentionally extends into the posterior submandibular glands and the lateral larynx. Another consideration reflects the choices made during proton planning. For example, in sparing the parotid glands, the focus was on maximizing a volume spared below 26-Gy dose, whereas the volume of gland receiving greater than 26-Gy dose was not optimized based on the belief that this portion of the gland was already destined to be dysfunctional. This highlights a final consideration, which is the 1-sided nature of the comparative dosimetry. The proton plans were manually edited to achieve desired constraints and were prospectively judged as clinically acceptable without contemporaneous comparison with alternate treatment plans. The IMRT plans were developed with inverse optimization to match or improve on the retrospective proton dosimetry, whereas the proton plans were not reoptimized in response to assess whether further improvements could be made. For the purpose of comparison, the IMRT plans used the same treatment volumes, while in practice, we would apply a larger PTV expansion for IMRT owing to differences in patient
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
5
Fig. 3. The median dose-volume histograms (DVH) for the proton and the IMRT comparative treatment plans for target coverage and select ORAs. The median DVH for patients receiving elective nodal irradiation (n ¼ 8) are reported separately from those receiving higher doses for gross nodal disease (n ¼ 4).
setup at the proton center, including a robotic positioner with 61 of freedom, repeated orthogonal image setup for each treatment field, and real-time treatment setup verification of each field. However, the primary purpose of our study was to report our technique for use of proton therapy in comprehensive head and neck irradiation, with comparative dosimetry generated to provide some reference for assessing the resultant DVHs. Some limitations of our data set must be acknowledged. The midline sparing of the larynx and the esophagus seen in these patients is not expected in different patient cohorts, such as those with grossly involved pretracheal level VI lymph nodes or those with laryngeal, hypopharyngeal, or thyroid cancers where the primary tumor is itself midline. We have not clinically
implemented or formally evaluated the role of proton therapy in these patient populations. Our data set is insufficient to assess the dosimetric effect of varied volumes and locations of cervical nodal disease. Our analysis does not include formal assessment of plan robustness against potential deviations in patient setup and other treatment uncertainties, which are assumed to be adequately accounted for in the PTV margin, proton range uncertainty, and compensator smearing values applied during treatment planning. It is noteworthy that no critical structures (brainstem, spinal cord, and optic apparatus) depended on distal blocking to meet dose constraints. Although distal blocking was used to spare the submandibular glands and portions of the parotid glands, these structures lie in the plane of the incident posterior beams, such
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]]
6 Table 3 Radiation dose to organs of interest by modality Region of interest
Oral cavity Larynx Esophagus Better-spared parotid (n ¼ 12) Lesser-spared Parotid (n ¼ 10) Spared submandibular gland (n ¼ 19)
Mean of mean dose
Median of integral dose
IMRT
Proton
p Value
IMRT
Proton
p Value
18.48 23.73 13.23 21.09 20.84 23.82
5.98 18.37 8.49 15.22 19.22 21.47
0.002 0.002 0.002 0.004 0.203 0.117
1924.55 1176.30 146.60 506.25 581.50 127.40
532.20 868.75 77.15 352.90 496.95 114.80
0.002 0.002 0.002 0.004 0.028 0.091
Dose in Gy. Integral Dose in Gy cm3.
that uncertainties in vertical patient positioning would have minimal effect on the position of the resultant dose distribution.33 When compared with pencil-beam scanning and intensitymodulated proton therapy, uniform scanning is less susceptible to patient setup and range uncertainties or small misalignments in the proton beam.13 Field matching was used, which is associated with increased dosimetric complexity and quality assurance requirements. A similar concern arises in matching low anterior neck fields to IMRT fields. We mitigated the issues of field matching by an intrafractional feather of the matchline so that each side of the neck was matched to the primary tumor site at 2 matchlines, both treated each day. We have previously reported a formal assessment of the accuracy of field matching using our robotic patient positioner and found good agreement between treatment planning calculations and radiographic film and ionization chamber measurements.34 Most published clinical experiences with proton therapy involved 3D conformal treatment planning in which the desired dosimetry is achieved by manual iterative adjustments in the shape of individual field apertures (blocks) and compensators, or patient-specific devices (PSDs). With uniform scanning, as used in our patients, a uniform dose is delivered across the treatment field with a uniform spread-out Bragg peak (SOBP). This results in 2 limitations to the application of our technique for cervical nodal irradiation. With uniform scanning, the SOBP is chosen to cover the widest part of the target within each field. For example, in the upper cervical neck, the target may require an 8-cm SOBP to cover from the anterior to posterior neck, whereas in the supraclavicular fossa, only 3 cm may be required. With uniform scanning, the SOBP does not vary over the field, and consequently, the larger SOBP required for coverage in the upper neck results in poor posterior dose conformality in the inferior neck, where a smaller SOBP would be preferred. One option to address this would be to subdivide the upper and lower neck, treating through 2 different matched fields that use SOBPs appropriate for the target at each level. We did not elect to do this, as the gain in posterior neck target conformality did not seem to justify the increased complexity of additional matchlines nor the increased cost and treatment time that 2 to 3 additional low neck fields would require. Instead, we accepted that additional radiation would be delivered to deep cervical musculature and lung apices as a trade-off for the improved oral cavity sparing and midline sparing of the esophagus and the larynx. This is not dissimilar to the homogenous dose delivered to the cervical muscles with traditional x-ray radiation using parallel-opposed lateral fields and a matched low anterior neck field. It is possible that sacrificing posterior conformality to the cervical nodal target volumes could increase the risk of neck soft tissue fibrosis. Although we are unaware of any data that suggest that IMRT may reduce the risk of cervical soft tissue fibrosis, it would be desirable to improve on the posterior target conformality in the lower cervical neck using more advanced proton delivery techniques.
A second limitation in uniform scanning is that there are few options to create nonuniform dose intensities, for example, to treat 1 portion of the target at a lower dose per fraction than another portion. A option is to use a field-in-field technique, which is again cumbersome owing to the requirement for additional PSDs, or by selective exclusion of the desired area from some but not all treatment fields, weighted in a fashion to deliver the desired dose per fraction to each area. Given the limitations of these 2 approaches, a simultaneous integrated boost technique to multiple nodal targets is impractical using 3D conformal proton therapy, prompting the use of sequential shrinking field plans. A rapidly emerging proton treatment approach is pencil-beam scanning, in which the proton beam is magnetically steered to cover the target volume without the requirement of PSDs.35 Pencil-beam scanning allows for intensity-modulated proton therapy of various techniques, which have been clinically implemented in treatment of head and neck tumors.36 Using multiple-field optimization, simultaneous integrated boost techniques can be used with proton therapy.37 Proton pencil-beam scanning is therefore anticipated to improve on and replace our approach of 3D planning with uniform scanning. For example, it would have the advantage of varying the SOBP to match the varied depth required at different levels of the cervical neck, addressing the issue of poor posterior target conformality in the low cervical neck seen with our technique, where each beam has a fixed SOBP width. However, our technique of 3D conformal proton therapy with uniform scanning has some advantages when compared with multifield optimization, including being a well-established delivery system that is less susceptible to patient setup and range uncertainties or small misalignments in the proton beam, and has simpler quality assurance.13,38
Conclusions A technique for cervical nodal irradiation using 3D conformal proton therapy with uniform scanning was developed and clinically implemented. Use of proton therapy for cervical nodal irradiation resulted in large volume of dose avoidance to the oral cavity and low dose exposure to midline structures of the larynx and the esophagus, with lower mean and integral dose to assessed OARs when compared with competing IMRT plans.
References 1. Schulz-Ertner, D.; Tsujii, H. Particle radiation therapy using proton and heavier ion beams. J. Clin. Oncol. 25:953–64; 2007. 2. Patel, S.H.; Wang, Z.; Wong, W.W.; et al. Charged particle therapy versus photon therapy for paranasal sinus and nasal cavity malignant diseases: A systematic review and meta-analysis. Lancet Oncol. 15:1027–38; 2014. 3. Zenda, S.; Kohno, R.; Kawashima, M.; et al. Proton beam therapy for unresectable malignancies of the nasal cavity and paranasal sinuses. Int. J. Radiat. Oncol. Biol. Phys. 81:1473–8; 2011.
M.W. McDonald et al. / Medical Dosimetry ] (2015) ]]]–]]] 4. Truong, M.T.; Kamat, U.R.; Liebsch, N.J.; et al. Proton radiation therapy for primary sphenoid sinus malignancies: Treatment outcome and prognostic factors. Head Neck 31:1297–308; 2009. 5. Fitzek, M.M.; Thornton, A.F.; Varvares, M.; et al. Neuroendocrine tumors of the sinonasal tract. Results of a prospective study incorporating chemotherapy, surgery, and combined proton-photon radiotherapy. Cancer 94:2623–34; 2002. 6. Chan, A.W.; Liebsch, N.J. Proton radiation therapy for head and neck cancer. J. Surg. Oncol. 97:697–700; 2008. 7. Mock, U.; Georg, D.; Bogner, J.; et al. Treatment planning comparison of conventional, 3D conformal, and intensity-modulated photon (IMRT) and proton therapy for paranasal sinus carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 58:147–54; 2004. 8. Lomax, A.J.; Goitein, M.; Adams, J. Intensity modulation in radiotherapy: Photons versus protons in the paranasal sinus. Radiother. Oncol. 66:11–8; 2003. 9. Herr, M.W.; Sethi, R.K.; Meier, J.C.; et al. Esthesioneuroblastoma: An update on the massachusetts eye and ear infirmary and massachusetts general hospital experience with craniofacial resection, proton beam radiation, and chemotherapy. J. Neurol. Surg. B Skull Base 75:58–64; 2014. 10. Fukumitsu, N.; Okumura, T.; Mizumoto, M.; et al. Outcome of T4 (International Union Against Cancer Staging System, 7th edition) or recurrent nasal cavity and paranasal sinus carcinoma treated with proton beam. Int. J. Radiat. Oncol. Biol. Phys. 83:704–11; 2012. 11. Buchsbaum, J.C.; Besemer, A.; Simmons, J.; et al. Supine proton beam craniospinal radiotherapy using a novel tabletop adapter. Med. Dosim. 38:70–6; 2013. 12. Farr, J.B.; Mascia, A.E.; Hsi, W.C.; et al. Clinical characterization of a proton beam continuous uniform scanning system with dose layer stacking. Med. Phys. 35:4945–54; 2008. 13. Anferov, V.A. Scan pattern optimization for uniform proton beam scanning. Med. Phys. 36:3560–7; 2009. 14. Suit, H.; Goldberg, S.; Niemierko, A.; et al. Proton beams to replace photon beams in radical dose treatments. Acta. Oncol. 42:800–8; 2003. 15. Allgower, C.E.; Schreuder, A.N.; Farr, J.B.; et al. Experiences with an application of industrial robotics for accurate patient positioning in proton radiotherapy. Int. J. Med. Robot. 3:72–81; 2007. 16. van't Riet, A.; Mak, A.C.; Moerland, M.A.; et al. A conformation number to quantify the degree of conformality in brachytherapy and external beam irradiation: Application to the prostate. Int. J. Radiat. Oncol. Biol. Phys. 37:731–6; 1997. 17. Semenenko, V.A.; Reitz, B.; Day, E.; et al. Evaluation of a commercial biologically based IMRT treatment planning system. Med. Phys. 35:5851–60; 2008. 18. Roesink, J.M.; Moerland, M.A.; Battermann, J.J.; et al. Quantitative dose-volume response analysis of changes in parotid gland function after radiotherapy in the head-and-neck region. Int. J. Radiat. Oncol. Biol. Phys. 51:938–46; 2001. 19. Batth, S.S.; Caudell, J.J.; Chen, A.M. Practical considerations in reducing swallowing dysfunction following concurrent chemoradiotherapy with intensity-modulated radiotherapy for head and neck cancer. Head Neck 36:291–8; 2014. 20. Dijkema, T.; Raaijmakers, C.P.; Ten Haken, R.K.; et al. Parotid gland function after radiotherapy: The combined Michigan and Utrecht experience. Int. J. Radiat. Oncol. Biol. Phys. 78:449–53; 2010. 21. Dirix, P.; Nuyts, S. Evidence-based organ-sparing radiotherapy in head and neck cancer. Lancet Oncol. 11:85–91; 2010.
7
22. Bhide, S.A.; Gulliford, S.; Schick, U.; et al. Dose-response analysis of acute oral mucositis and pharyngeal dysphagia in patients receiving induction chemotherapy followed by concomitant chemo-IMRT for head and neck cancer. Radiother. Oncol. 103:88–91; 2012. 23. Ling, D.C.; Kabolizadeh, P.; Heron, D.E.; et al. Incidence of hospitalization in patients with head and neck cancer treated with intensity-modulated radiation therapy. Head Neck; 2015. 24. Trotti, A.; Bellm, L.A.; Epstein, J.B.; et al. Mucositis incidence, severity and associated outcomes in patients with head and neck cancer receiving radiotherapy with or without chemotherapy: A systematic literature review. Radiother. Oncol. 66:253–62; 2003. 25. Nutting, C.M.; Morden, J.P.; Harrington, K.J.; et al. Parotid-sparing intensity modulated versus conventional radiotherapy in head and neck cancer (PARSPORT): A phase 3 multicentre randomised controlled trial. Lancet Oncol. 12:127–36; 2011. 26. Little, M.; Schipper, M.; Feng, F.Y.; et al. Reducing xerostomia after chemo-IMRT for head-and-neck cancer: Beyond sparing the parotid glands. Int. J. Radiat. Oncol. Biol. Phys. 83:1007–14; 2012. 27. Schwartz, D.L.; Hutcheson, K.; Barringer, D.; et al. Candidate dosimetric predictors of long-term swallowing dysfunction after oropharyngeal intensity-modulated radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 78:1356–65; 2010. 28. Kumar, R.; Madanikia, S.; Starmer, H.; et al. Radiation dose to the floor of mouth muscles predicts swallowing complications following chemoradiation in oropharyngeal squamous cell carcinoma. Oral Oncol. 50:65–70; 2014. 29. Eisbruch, A.; Kim, H.M.; Feng, F.Y.; et al. Chemo-IMRT of oropharyngeal cancer aiming to reduce dysphagia: Swallowing organs late complication probabilities and dosimetric correlates. Int. J. Radiat. Oncol. Biol. Phys. 81:e93–9; 2011. 30. Eisbruch, A.; Kim, H.M.; Terrell, J.E.; et al. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 50:695–704; 2001. 31. Baumert, B.G.; Norton, I.A.; Lomax, A.J.; et al. Dose conformation of intensitymodulated stereotactic photon beams, proton beams, and intensity-modulated proton beams for intracranial lesions. Int. J. Radiat. Oncol. Biol. Phys. 60:1314–24; 2004. 32. Paganetti, H. Range uncertainties in proton therapy and the role of Monte Carlo simulations. Phys. Med. Biol. 57:R99–117; 2012. 33. Thomas, S.J. Margins for treatment planning of proton therapy. Phys. Med. Biol. 51:1491–501; 2006. 34. Farr, J.B.; O'Ryan-Blair, A.; Jesseph, F.; et al. Validation of dosimetric field matching accuracy from proton therapy using a robotic patient positioning system. J. Appl. Clin. Med. Phys 11:3015; 2010. 35. Schippers, J.M.; Lomax, A.J. Emerging technologies in proton therapy. Acta Oncol. 50:838–50; 2011. 36. Frank, S.J.; Cox, J.D.; Gillin, M.; et al. Multifield optimization intensity modulated proton therapy for head and neck tumors: A translation to practice. Int. J. Radiat. Oncol. Biol. Phys. 89:846–53; 2014. 37. Zhu, X.R.; Poenisch, F.; Li, H.; et al. A single-field integrated boost treatment planning technique for spot scanning proton therapy. Radiat. Oncol. 9:202; 2014. 38. Liu, W.; Mohan, R.; Park, P.; et al. Dosimetric benefits of robust treatment planning for intensity modulated proton therapy for base-of-skull cancers. Pract. Radiat. Oncol. 4:384–91; 2014.
