Access the full text.
Sign up today, get DeepDyve free for 14 days.
Background: For many decades, the standard of care radiotherapy regimen for medulloblastoma has been photon (megavoltage x-rays) craniospinal irradiation (CSI). The late effects associated with CSI are well-documented in the literature and are in-part attributed to unwanted dose to healthy tissue. Recently, there is growing interest in using proton therapy for CSI in pediatric and adolescent patients to reduce this undesirable dose. Previous comparisons of dose to target and non-target organs from conventional photon CSI and passively scattered proton CSI have been limited to small populations (n ≤ 3) and have not considered the use of age-dependent target volumes in proton CSI. Methods: Standard of care treatment plans were developed for both photon and proton CSI for 18 patients. This cohort included both male and female medulloblastoma patients whose ages, heights, and weights spanned a clinically relevant and representative spectrum (age 2–16, BMI 16.4–37.9 kg/m2). Differences in plans were evaluated using Wilcoxon signed rank tests for various dosimetric parameters for the target volumes and normal tissue. Results: Proton CSI improved normal tissue sparing while also providing more homogeneous target coverage than photon CSI for patients across a wide age and BMI spectrum. Of the 24 parameters (V ,V ,V , and V in the 5 10 15 20 esophagus, heart, liver, thyroid, kidneys, and lungs) Wilcoxon signed rank test results indicated 20 were significantly higher for photon CSI compared to proton CSI (p ≤ 0.05) . Specifically, V and V in all six organs and V ,V in the 15 20 5 10 esophagus, heart, liver, and thyroid were significantly higher with photon CSI. Conclusions: Our patient cohort is the largest, to date, in which CSI with proton and photon therapies have been compared. This work adds to the body of literature that proton CSI reduces dose to normal tissue compared to photon CSI for pediatric patients who are at substantial risk for developing radiogenic late effects. Although the present study focused on medulloblastoma, our findings are generally applicable to other tumors that are treated with CSI. Keywords: Proton, Photon, Craniospinal irradiation, CSI, Medulloblastoma * Correspondence: firstname.lastname@example.org Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA Graduate School of Biomedical Sciences, The University of Texas at Houston, Houston, TX, USA Full list of author information is available at the end of the article © 2012 Howell et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Howell et al. Radiation Oncology 2012, 7:116 Page 2 of 12 http://www.ro-journal.com/content/7/1/116 Background differences and variations in target volume definition Medulloblastoma is the most common malignant child- and dose delivered between photon and proton therapy. hood brain tumor. In recent decades, the 5-year survival rate for this cancer has improved from 60% to between Methods 80% and 85% for average-risk patients and from 35% to Study patients between 60% and 70% for high-risk patients [1-3]. The This study was carried out under a protocol for retro- primary tumor normally originates in midline cerebellar spective treatment planning studies approved by our in- structures with infiltration of surrounding posterior stitution (University of Texas at M.D. Anderson Cancer fossa and may disseminate throughout the neuroaxis via Center, UTMDACC). We compared therapeutic dose cerebrospinal fluid (CSF) pathways [4-6]. Treatment for distributions for photon and proton CSI for a group of medulloblastoma thus often includes chemotherapy and 18 consecutive patients (8 girls and 10 boys). The inclu- craniospinal irradiation (CSI) [7-9], including a boost to sion criteria were that the patients be between 2 and the posterior fossa or the surgical bed with a margin. 18 years old at the time of treatment and were treated Because of the high survival rate, the fact that patients with proton CSI at our institution between 2007 and require radiotherapy, and the fact that children and ado- 2009. The patients in this study had a mean age of lescents are more likely to develop radiation-related late 9.5 years (range, 2–16 years). Patient age, sex, height, effects than adults, late (>5 years after treatment) effects weight and BMI are listed in Table 1. from radiation are a major concern for medulloblastoma Patients underwent computed tomography (CT) simu- patients [10-13]. The late effects associated with CSI are lation while in the supine position with their heads well-documented and may include (but are not limited immobilized using an Aquaplast face mask (WFR/Aqua- to) impaired growth , endocrine abnormalities [15- plast Corp. and Qfix Systems, LLC, Avondale, PA) and a 17], hearing loss [17,18], diminished fertility , neuro- plastic head holder to reduce kyphotic neck curvature. psychological dysfunction [17,19], cardiac diseases The CT images were acquired on a multi-slice CT scan- [17,20-24], and second cancers [17,22-28]. For many ner (General Electric (GE) LightSpeed RT16, GE Health- decades, the standard of care radiotherapy regimen for care, Waukesha, WI) and had a 2.5-mm slice thickness. CSI has been photon (megavoltave x-rays) therapy that Both photon and proton treatment planning were car- included opposed lateral cranial fields and either single ried out according to the standards of care at our insti- or multiple posterior spinal fields . Late effects are, tution (UTMDACC). We streamlined plan comparisons in part, a consequence of dose from the CSI treatment by using the same commercial treatment planning sys- fields to various non-target organs. Compared to tem (TPS) for both modalities (Eclipse version 8.9, photons, protons have substantially lower entrance dose Table 1 Patient characteristics and almost no exit dose and thus can significantly re- Index Age Sex Height (cm) Weight (kg) BMI (kg/m ) duce the dose to all organs situated outside the cra- 1 2 female 85.0 11.9 16.5 niospinal axis which are irradiated unnecessarily. 2 4 female 111.7 20.5 16.4 Consequently, there is growing interest in using proton therapy for CSI in pediatric and adolescent patients. 3 6 female 115.2 26.9 20.3 Passively scattered proton CSI has been shown to im- 4 8 female 142.0 37.5 18.6 prove dose uniformity along the spinal canal and de- 5 10 female 130.6 24.2 14.2 crease dose to non-target organs compared with photon 6 2 male 109.2 18.9 15.8 CSI [30-33]. However, each of these studies was limited 7 4 male 128.0 31.3 19.1 to a very small number of patients—there were a total of 8 6 male 144.8 24.9 11.9 7 patients, 6 of whom were under the age of 5, in all the studies combined—limiting the results’ applicability and 9 8 male 123.4 20.3 13.3 understanding of dosimetric differences across a wide 10 10 male 133.0 28.2 15.9 spectrum of patient ages and body sizes. Finally, none of 11 12 female 146.0 28.9 13.6 these reports addressed the differences in target volumes 12 13 female ————data not available--—— used in planning proton and photon CSI, e.g., the age- 13 16 female 162.0 62.0 23.6 specific target volumes used in proton CSI. Thus, we 14 12 male 166.3 66.5 24.0 sought to carry out a detailed comparison of the current treatment standards for photon and proton CSI for a 15 13 male 173.0 57.5 19.2 population of both male and female medulloblastoma 16 14 male 162.5 58.6 22.2 patients whose ages, heights, and weights spanned a 17 15 male 172.1 73.3 24.7 clinically relevant and representative spectrum (age 2– 18 16 male 191.0 138.2 37.9 16, BMI 16.4–37.9 kg/m ) with a focus on the Howell et al. Radiation Oncology 2012, 7:116 Page 3 of 12 http://www.ro-journal.com/content/7/1/116 Varian Medical Systems, Palo Alto, CA). All treatment plans were calculated using a 2.5-mm calculation grid with heterogeneity corrections. Dose distributions in the photon and proton plans were respectively calculated using anisotropic analytical and pencil beam algorithms. The proton calculation algorithm was previously vali- dated using the methodology described by Newhauser et al.  and the photon algorithm was commissioned following methodologies described in the literature [35,36]. The beam arrangements for the photon and pro- ton treatment plans were similar. Both included two opposed lateral oblique cranial fields, which were angled so that they avoided ocular structures, and postero- anterior spinal field(s). The proton plans used one to three spinal fields while the photon plans used either one or two spinal fields to cover the entire length of the spinal canal through the inferior extent of the thecal sac, typically at the level of the S2/S3 vertebral junction. The spinal fields were matched at the posterior edge of the vertebral canal (not on the vertebral body). The total prescribed dose was 23.4 Gy relative bio- logical effectiveness (RBE) (i.e., 21.3 Gy × 1.1 to reflect the biological effectiveness of protons relative to photons) and 23.4 Gy for the proton and photon CSI treatment plans, respectively. Hereafter, dose units will be simply be referred to as Gy and Gy or Gy-RBE for photons and protons, respectively. The use of the gen- Figure 1 Age-specific target volumes for proton treatment planning eric RBE factor of 1.1 is in accordance with the recom- (red contour) and CTVs (blue color wash) in proton and photon mendations on dose prescription and reporting in treatment planning for representative patients. (a) Volumes for a International Commission on Radiation units and Mea- 4-year-old patient and (b) volumes for a 15-year-old patient. surements (ICRU) Report 78  and consistent with the clinical practice at our institution. However, it noted a board-certified radiation oncologist who specializes in that the recommended RBE value has never been mea- pediatric radiotherapy (A. Mahajan). sured in humans who received proton therapy . The prescription dose of 23.4 Gy was selected for this study because it the most commonly used dose for moderate Proton therapy treatment planning risk patients and is the dose used at our institution for In this study (and in accordance with clinical practice at such patients. However, for high risk patients the CSI our institution), we used age specific target volumes for dose can be as high as 36 to 39.6 Gy and but may also proton CSI treatment planning. For all proton CSI be as low as 18 Gy, which is currently being evaluated patients, the CTV included the entire CSF space (the by some institutions. The fractionation schedule was brain and spinal canal through the cauda equina to the 1.8 Gy per fraction for 13 fractions with 2 junction shifts level of the S2/S3 vertebral junction (Figure 1) and was (initial and 2 shifted positions), which is a common dose equivalent to the photon CTV. Additionally, for patients and fractionation pattern for patients with average-risk under the age of 15 years there was an additional normal medulloblastoma. The clinical target volume (CTV) for tissue target volume (NTTV), which included the entire both the photon and proton treatment plans included vertebral bodies. The rationale for this was to avoid sharp the entire CSF space (the brain and spinal canal through dose gradients in the vertebral bodies in patients whose the cauda equina to the level of the S2/S3 vertebral junc- skeletons were still maturing. More specifically, proton tion (Figure 1). Additionally for patients under the age treatments that are designed to irradiate only the spinal of 15 years there was an additional target volume which canal have high dose gradients distal to the spinal canal was also treated to the full prescription dose (discussed and lead to non-uniform irradiation of the vertebral bod- below in the section on proton therapy planning). All ies. Uniformly irradiating a larger target volume that fully treatment plans were reviewed by a board certified med- encompasses the vertebral bodies is thought to reduce ical physicist (R. Howell) and reviewed and approved by the risk of asymmetric growth of the vertebral body in Howell et al. Radiation Oncology 2012, 7:116 Page 4 of 12 http://www.ro-journal.com/content/7/1/116 patients whose skeletons are still maturing [33,38] i.e., geometries [29,44] were defined, multiple lower- those under the age of 15 years. weighted reduction fields within the primary cranial and Adequate uncertainty margins are especially important spinal fields were added to minimize dosimetric hetero- in proton therapy because proton fields are especially geneities (reduce hot spots in thinner regions of the sensitive to patient positioning due to several factors in- anatomy and cold spots in thicker regions of the anat- cluding: 1) proton fields have a sharp distal fall-off, but omy). The reduction fields contained blocked segments the location of that fall-off is dependent on the beam strategically placed to reduce the highest dose areas to range which is determined by the composition of tissues force greater homogeneity and conformity in the target in the beam path; thus, lateral or superior/inferior shifts volume. This planning technique is commonly referred in patient position, relative to the field’s isocenter, can to as intensity-modulated field-in-field planning and was change the location of the distal field edge relative to described in detail by Yom et al. . Photon treatment specific organs of interest and 2) proton fields are plans were normalized so that the 100% isodose line shaped by field specific apertures and tissue compensa- covered the CTV and allowed for setup-up uncertainty. tors, so lateral or superior/inferior shifts in patient pos- ition, relative to the field’s isocenter, can shift patient Comparison of photon and proton treatment plans anatomy from its optimal alignment to these devices. To We compared three dosimetric parameters for the CTV: ensure the proton treatment fields had appropriate un- the maximum dose (D ), the conformity index (CI), max certainty margins we used the methodology of the ICRU and the heterogeneity index (HI). The CI is defined as Report 78 . As a result, field parameters were deter- Rx mined using the CTV, rather than the PTV, and the bur- CI ¼ ð1Þ EV den of applying the parameters was placed on the computer algorithm. That is, values for compensator where V is the volume receiving the prescribed dose Rx smear, lateral, proximal, and distal margins were manu- and the V the total CTV and HI is defines as EV ally calculated for each beam using a methodology simi- lar to that used in our previous studies [39,40] and 5% HI ¼ ð2Þ following the methods originally outlined by Urie et al. 95%  and Moyers and Miller  and Moyers et al. . Once calculated those values were entered into the TPS where D is the dose delivered to the hottest 5% of the 5% as planning parameters. Then, the TPS selected the cor- CTV and D is the minimum dose received by 95% of 95% responding machine parameters (beam energy, range the CTV. modulation, and range shifter settings), designed the The HI was used to quantify dosimetric homogeneity compensator, and sized the apertures. For patients older within the CTV. A lower HI indicated a more uniform than 15, these uncertainty margins were designed to en- dose distribution. The CI was used to quantify how well sure coverage of the CTV. Similarly, for patients the prescribed dose conformed to the CTV. A lower CI younger than 15 the uncertainty margins were designs indicated a more conformal dose distribution. such that the CTV as well as the entire vertebral bodies In addition to the CTV, we contoured the following (NTTV) received the full prescription dose. normal tissues so we could compare photon and proton Beam energies for the proton plans were patient and doses in organs that were within or near the treatment field specific and included energies of 140 MeV, fields: spinal cord, optic chiasm, cochlea, brainstem, 160 MeV, 180 MeV, 200 MeV, and 225 MeV. The mean esophagus, heart, kidneys, liver, lungs, and thyroid. A cranial and spinal field energies were 198 MeV dose volume histogram (DVH) was calculated for each (SD = 12 MeV) and 163 MeV (SD = 17 MeV) for the cra- of these structures. Then, we quantitatively compared nial and spinal fields, respectively. The mean range was the photon and proton DVH data for each structure by 17 cm (SD = 1 cm) and 11 cm (SD = 2 cm) for the cranial comparing the mean percent volume (V) receiving vari- and spinal fields, respectively. The mean Spread out ous specified dose levels in units of gray (Gy). V and 23.4 Bragg peak was 16 cm (SD = 1 cm) and 5 cm (SD = 1 cm) V were compared for the CTV and organs that were for the cranial and spinal fields, respectively. A more entirely within the treatment fields. V ,V ,V ,V , 5 10 15 20 comprehensive and detailed description of the proton and V were compared for partially in-field and out- 23.4 CSI treatment planning technique used in this study is of-field organs. reported in the literature by Giebeler et al. (in review). Statistical methods Photon treatment planning Statistical analyses were performed to compare the vari- The photon CSI plans were calculated using a beam en- ous dosimetric parameters for the CTV and the normal ergy of 6 MV. After the cranial and spinal field organs. We used the Wilcoxon signed rank test with a Howell et al. Radiation Oncology 2012, 7:116 Page 5 of 12 http://www.ro-journal.com/content/7/1/116 null hypothesis that the differences between the various significant at P ≤ 0.05 were then evaluated for signifi- dosimetric parameters for photon and proton therapy cance at P ≤ 0.01. The sequential Bonferroni-type pro- come from a continuous, symmetric distribution with cedure, as described by Benjamini and Hochberg , zero median. For the CTV and organs entirely within was then used to test for false positives in the independ- the CTV (optic chiasm, cochleas, brainstem, spinal ent Wilcoxon sign ranked tests. cord), we used a two-tailed Wilcoxon signed rank test to compare these values. The alternative hypothesis for this Results two-tailed test was that the differences between the vari- Isodose distributions (Figure 2) and DVHs (Figure 3) for ous dosimetric parameters for photon and proton ther- the photon and proton treatment plans for a representa- apy come from a continuous, symmetric distribution tive patient under the age of 15 are shown (index 2). In with a positive or negative median. For partially in-field Figure 2, the 100% isodose line indicates the intended and out-of-field organs (esophagus, heart, kidneys, liver, treatment region. Qualitatively, several observations can lungs, and thyroid), we used a one-tailed Wilcoxon be made: (1) the prescribed dose covers all the vertebral signed rank test. The alternative hypothesis for this one- bodies in the proton plan but covers only the spinal tailed test was that the differences between the various canal in the photon plan; (2) the proton dose rapidly dosimetric parameters for photon and proton therapy decreases beyond the target volume, whereas the photon come from a continuous distribution with a median dose gradually decreases; and (3) the normal organs and greater than zero. Differences that were found to be tissues in close proximity to the treatment volume Figure 2 Photon and proton treatment plans for a representative patient under the age of 15 (this patient was 4 years old, index 2). (a) Proton dose distribution in the sagittal plane. (b) Photon dose distribution in sagittal plane. (c) Proton dose distribution in axial planes from the cervical spine to the sacral spine in 5-cm increments. (d) Photon dose distribution shown in for axial planes from the cervical spine to the sacral spine in 5-cm increments. (e) Isodose scale for both photon and proton treatment plans. Howell et al. Radiation Oncology 2012, 7:116 Page 6 of 12 http://www.ro-journal.com/content/7/1/116 Figure 3 Photon and proton dose volume histograms (DVHs) for a representative patient (age 4, index 2) under the age of 15. Proton and photon DVHs are indicated by dashed and solid lines, respectively. The absolute dose values shown on the horizontal axis of 250, 500, 750, 1000, 1250, 1500, 1750, 2000, 2250, and 2750 cGy correspond to percent dose values of 11, 21, 32, 43, 53, 64, 75, 85, 96, 107, and 118%, respectively. receive substantially lower doses from the proton plan spinal canal, whereas the photon dose distribution had than from the photon plan. Isodose distributions hot and cold spots on either side of the spine field junc- (Figure 4) and DVHs (Figure 5) for photon and proton tions when more than 1 spinal field was required (as in treatment plans for a representative patient over the age Figure 4). While the dosimetric impact of field junctions of 15 (index 13) are also shown. As in the younger pa- in CSI was previously known to clinicians, this study tient, the photon target volume in this patient included highlights the difference in field junction dosimetry be- the craniospinal canal. However, because this patient was tween photon and proton CSI. older than 15, the proton target volume was the same as Quantitative dose-volume results are summarized in the photon target volume. The qualitative observations Tables 2, 3, and 4 for the photon and proton treatment for this older patient were similar to those for the plans. In the next three subsections, we detail the results younger patient, except that the normal tissue sparing in from our analysis of the modalities’ coverage of the CTV, the proton plan was even greater for this patient because sparing of in-field organs, sparing of partially in-field the sharp dose fall-off began at the anterior end of the organs, and sparing of out-of-field organs. As mentioned spinal canal rather than at the anterior end of the verte- in the methods, that while we chose to use a prescribed bral bodies. The dose distributions and DVHs for these dose of 23.4 Gy in this study, prescribed doses as low as two representative patients (ages 4 and 16) highlight the 18 Gy and as high at 39 Gy have been used for CSI. differences in photon and proton dose distributions that Therefore, percent dose is given in parenthesis next to result from age-specific treatment volumes. each parameter that is reported in Gy so that our data When comparing the dose distributions in Figures 2 can be easily translated to any prescription dose. Simi- and 4, there is another another age/size effect due to the larly, percent doses are given in the table captions for number of fields required to cover the spinal canal. For each table where absolute doses are reported. the younger patient (Figure 2), the proton and photon treatments could both be delivered using a single spinal CTV coverage field. For the older patient (Figure 4), 3 proton fields and No significant difference was observed between the 2 photon fields were required to cover the spinal canal. photon and proton plans in the mean values of the For both the older and younger patients, the proton dose V for the CTV (Table 2). For both modalities, 23.4(100%) distributions were homogeneous along the spine, regard- the mean V value was greater than 99%. Simi- 23.4(100%) less of the number of spinal fields required to treat the larly, no significant difference in the CI was observed Howell et al. Radiation Oncology 2012, 7:116 Page 7 of 12 http://www.ro-journal.com/content/7/1/116 Figure 4 Photon and proton treatment plans for a representative patient over the age of 15 (this patient was 16 years old, index 13). (a) Proton dose distribution in the sagittal plane. (b) Photon dose distribution in the sagittal plane. (c) Proton dose distribution in axial planes from the cervical spine to the sacral spine in 5-cm increments. (d) Photon dose distribution in axial planes from the cervical spine to the sacral spine in 5-cm increments. (e) Isodose scale for both photon and proton treatment plans. between photon and proton treatment plans, which heterogeneous than the proton dose distributions. In was greater than 0.99 for both modalities, indicating summary, the photon and proton treatment plans both that the dose distribution conformed well to the CTV provided very good coverage and conformed well to (Table 3). In contrast, statistically significant differences the craniospinal axis, but in general, the photon plans were observed in the D ,V , and HI values were (approximately 8%) hotter than the proton plans. max 25(107%) (Tables 2 and 3). Both the mean D (P = 1.60E-05) max and mean V values (P =1.04E-03) were greater Tissue sparing of in-field organs 25(107%) for the photon plans, indicating higher maximum The cochleae, brainstem, spinal cord, and optic chiasm doses and higher doses to a larger percentage of the were entirely within the 100% isodose region in the pho- volume. The mean HI was greater for the photon ton and proton plans for all patients. We observed no plans than the proton plans (P = 4.87E-04), indicating significant difference between the mean V values 23.4(100%) that the photon dose distributions were more from the photon and proton plans value for the Howell et al. Radiation Oncology 2012, 7:116 Page 8 of 12 http://www.ro-journal.com/content/7/1/116 Figure 5 Photon and proton dose volume histograms for a representative patient (age 16, index 13) over the age of 15. Proton and photon DVHs are indicated by dashed and solid lines, respectively. The absolute dose values shown on the horizontal axis of 250, 500, 750, 1000, 1250, 1500, 1750, 2000, 2250, and 2750 cGy correspond to percent dose values of 11, 21, 32, 43, 53, 64, 75, 85, 96, 107, and 118%, respectively. Table 2 Dose volume histogram (DVH) analysis for brainstem and spinal cord. For the cochleae, there photon and proton craniospinal irradiation (n = 18) was a significant difference (P = 1.56E-2) in the mean Structure, Photons Protons P value, Significance V values, with the mean photon plan value 23.4(100%) DVH dose Wilcoxian level Mean SD Mean SD being approximately 1% greater than for the mean level (Gy) Signed (P ≤ 0.05) proton plan value (99.96 ± 0.16% for photons versus Rank test 99.00% ± 3.09% for protons, P = 5.57E-14, Table 2). No CTV significant difference between the mean V values 25(106.8%) 23.4 99.36 1.04 99.23 0.88 5.42E-01 NS from the photon and proton plans value for the optic 25 12.01 9.87 1.72 4.01 8.63E-04 <0.01 chiasm and spinal cord. In addition, for brainstem the Spinal cord mean V value was lower in the photon plans 25(106.8%) 23.4 99.16 0.85 98.91 1.46 9.83E-01 NS than in the proton plans (P = 3.12E-02) but the mean 25 59.02 9.15 4.64 7.52 1.96E-04 <0.01 values for both treatment techniques were less than 2% (0.68 ± 1.39% for photons versus 1.70% ± 7.61% for pro- Optic chiasm tons, P = 5.57E-14, Table 2). In contrast, for the spinal 23.4 100.00 0.00 100.00 0.00 n/a N/A cord the mean V value was much higher in the 25(106.8%) 25 3.48 4.29 3.26 3.36 5.69E-01 NS photon plans than in the proton plans (59.0% ± 9.2% for Cochlea 23.4 99.96 0.16 99.00 3.09 1.56E-02 < 0.05 Table 3 Comparison of parameters to evaluate dose 25 2.08 2.96 4.26 6.95 6.36E-01 NS variation with in the target for photon and proton craniospinal irradiation (n = 18) Brain stem CTV Photons Protons P value, Significance 23.4 98.51 3.20 98.96 1.76 8.39E-01 NS dosimetric Wilcoxian level Mean SD Mean SD 25 0.68 1.39 1.70 7.61 3.12E-02 < 0.05 parameters Signed Rank test Note: The table provides mean V and V and standard deviation (SD) 23.4 25 values for the CTV and in-field organs for proton and photon therapy. CI 0.99 0.01 0.99 0.009 5.28E-01 NS Statistical results from Wilcoxon signed rank test (i.e., two-tailed P-value for HI 1.05 0.009 1.04 0.012 2.47E-03 <0.01 Wilcoxon signed rank test) are also listed. For this study 23.4 and 25 Gy is equivalent to 100% and 106.8% of the D 28.13 15.21 26.05 7.868 1.96E-04 <0.01 max prescribed dose, respectively. Howell et al. Radiation Oncology 2012, 7:116 Page 9 of 12 http://www.ro-journal.com/content/7/1/116 Table 4 Dose volume histogram (DVH) analysis for photon and proton craniospinal irradiation (n = 18) Structure, Photons Protons P value, Significance Sequential DVH dose Wilcoxian level Bonferroni Mean SD Mean SD level (Gy) Signed Rank test Esophagus 20 65.87 23.54 3.89 7.68 9.82E-05 <0.01 † 15 96.68 5.12 8.73 12.53 9.80E-05 <0.01 † 10 98.09 3.26 14.76 16.59 9.75E-05 <0.01 † 5 99.61 0.98 24.67 21.28 9.80E-05 <0.01 † Heart 20 2.80 4.58 0.03 0.08 2.51E-04 <0.01 † 15 42.49 16.98 0.15 0.23 9.82E-05 <0.01 † 10 56.77 11.12 0.53 0.62 9.82E-05 <0.01 † 5 60.68 11.30 1.31 1.28 9.82E-05 <0.01 † Kidneys 20 2.03 1.91 0.60 0.82 3.15E-03 <0.01 † 15 4.11 2.81 2.49 2.34 1.07E-03 <0.01 † 10 5.92 3.68 5.53 4.70 2.50E-01 NS — 5 8.89 4.95 10.58 7.95 6.24E-01 NS — Liver 20 3.09 3.17 0.08 0.15 9.82E-05 <0.01 † 15 14.69 4.22 0.27 0.30 9.82E-05 <0.01 † 10 22.55 1.99 0.61 1.05 9.82E-05 <0.01 † 5 24.78 3.95 1.10 0.75 9.82E-05 <0.01 † Lungs 20 3.07 2.14 2.27 1.92 3.54E-02 < 0.05 * 15 6.03 2.92 4.87 3.24 6.14E-03 < 0.01 † 10 8.35 3.52 7.66 4.32 7.23E-02 NS — 5 11.69 4.51 11.31 5.52 1.53E-01 NS — Thyroid 20 11.91 21.19 0.00 0.00 6.10E-05 <0.01 † 15 66.16 30.19 0.00 0.00 9.80E-05 <0.01 † 10 80.97 21.53 0.00 0.00 9.48E-05 <0.01 † 5 92.50 10.68 0.51 0.76 9.65E-05 <0.01 † Note: The table lists mean V ,V ,V , and V and standard deviation (SD) values for the esophagus, heart, kidneys, liver, lungs, and thyroid. Statistical results 20 15 10 5 from the Wilcoxon signed rank test (i.e., p-value for 1-tailed Wilcoxon signed rank test) are listed. Results from the sequential Bonferroni procedure are also included; † and * indicate differences were significant at the 0.01 and 0.05 levels, respectively. For this study 5, 10, 15, and 20 Gy are equivalent to 21.4%, 42.7%, 64.1%, and 85.5% of the prescribed dose, respectively. photons versus 4.6% ± 7.5% for protons, P = 5.57E-14, six partially in-field and out-of-field organs). The Wil- Table 2). The spinal cord was part of the CTV and these coxon sign ranked test results indicated that 20 of the data parallel those that were observed for the CTV, i.e., 24 parameters (83%) had effects that were significantly photon plans resulted in more heterogeneous dose distri- different between the proton and photon treatments at butions and had larger hot spots than the proton plans. the 0.05 level, Table 4. Results of the sequential-type Bonferroni procedure were consistent with those from the Wilcoxon sign ranked tests and did not find any Tissue sparing of partially in-field organs and out-of-field organs false positives. Results for individual organs are summarized in In summary, we evaluated 24 individual dosimetric para- Table 4. For the esophagus, heart, liver, and thyroid, meters (V ,V ,V , and V ) for 5(21.4%) 10(42.7%) 15(64.1%) 20(85.5%) Howell et al. Radiation Oncology 2012, 7:116 Page 10 of 12 http://www.ro-journal.com/content/7/1/116 there was a significant difference observed between 64.1%, and 85.5%). The kidneys and lungs are bilateral the photon and proton plans for V ,V , organs situated to the right and left of the spinal fields. 5(21.4%) 10(42.7%) V , and V with the values all being higher They received higher dose from the proton plans com- 15(64.1%) 20(85.5%) for photons than for protons. For the kidneys and lungs, pared to the organs that were anterior to the target vol- there were significant differences observed between the ume due to the lateral margins used for planning. This photon and proton plans for V , and V , effect was more pronounced for the younger patients 15(64.1%) 20(85.5%) again with the values higher for photons than for pro- (Figure 2), whose treatment volumes included the entire tons. However, a similar difference was not observed at vertebral bodies and whose proton plans required the lower dose levels of 5 and 10 Gy (21.4% and 42.7%). greater distal margins. As a consequence of the lateral and distal margins we observed that similar percentages Discussion the kidney and lung volumes receiving 5 and 10 Gy In this study, we compared proton and photon CSI for (21.4% and 42.7%) for proton and photon CSI. Like the 18 patients. It is important to note that this cohort lungs and kidneys, part of the liver is also lateral to the included both male and female medulloblastoma spinal field, but it is not a bilateral organ. Therefore, patients whose ages, heights, and weights spanned a compared to the lungs and kidneys, a smaller percentage clinically relevant and representative spectrum (age 2– of the liver volume received 5 and 10 Gy (21.4% and 16, BMI 16.4–37.9 kg/m ) and that we compared the 42.7%) in the proton plans than in the photon plans. current standard of care at our institution (UTMDACC) Recently, Brodin et al.  reported differences be- for each modality. Furthermore, our patient cohort is tween photon and proton CSI plans for 10 patients the largest, to date, in which CSI with proton and pho- whose ages also spanned the range of medulloblastoma ton therapies have been compared, a feature that consti- patients. However, they considered intensity-modulated tutes this study’s major strength. Finally, this study proton therapy (IMPT), volumetric-modulated arc pho- addressed differences in the various dosimetric para- ton therapy (VMAT), and conventional photon therapy meters associated with variations in target volume defin- without modulation. Their findings are limited in their ition, (i.e., that proton volumes were age-dependent, clinical meaningfulness, however, because neither IMPT whereas photon target volumes were the same for all nor VMAT is routinely used for CSI, and conventional patients). In the end, we found that proton CSI improves photon therapy has very heterogeneous dose distribu- normal tissue sparing while also providing more homo- tions compared to the field-in-field photon therapy tech- geneous target coverage than photon CSI for patients nique studied here. Another advantage of our work is across a wide age and BMI spectrum. that we considered current standards of care for photon For this population of patients, we found that proton and proton therapies that are currently in use. Thus, our CSI provided similar CTV coverage to that of photon findings are directly relevant to clinicians who have the CSI but allowed for a statistically significant reduction in option of treating patients with photon or proton CSI. doses to non-target organs in close proximity to the cra- Despite the differences in study design, there is niospinal axis. Moreover, proton treatment plans had consistency between the major findings of our study and greater dosimetric homogeneity along the craniospinal those of Brodin et al., i.e., that proton CSI improves nor- axis than photon treatment plans. Our results thus indi- mal tissue sparing while also providing more homoge- cate that proton CSI is superior to photon CSI over the neous target coverage than photon CSI. entire age range of children and adolescents affected by One limitation of this study is that we only focused on medulloblastoma. These results are consistent with therapeutic dose and did not consider stray dose. For those from earlier studies of fewer patients [30-33]. photon therapy, the stray dose would comprise only The differences that were observed between the pho- photons (patient scatter and scatter/leakage from treat- ment head) because all the treatment plans used beams ton and proton treatment plans were primarily due to the differences in the physical properties of photon and with an energy of 6 MV, which is below the threshold proton beams and the physical location of the organs for photoneutron production. In a previous study, we examined the accuracy of the TPS used in this study to relative to the intended target volume. The esophagus, heart, and thyroid were anterior to the treatment volume predict dose outside of the treatment field, where stray and thus were located in a high dose gradient for the dose is the main component; we found that the TPS was accurate at doses of approximately 5% or more of the photon plans, leading to a higher percentage of the structures receiving 5, 10, 15, and 20 Gy (21.4%, 42.7%, prescribed dose , which would be 1.17 Gy in the 64.1%, and 85.5%). In contrast, for the proton plans, present study, with its prescribed dose of 23.4 Gy. The lowest dosimetric parameter considered here was the V these organs were beyond the distal edge of the Bragg , peak, leading to a substantially lower percentage of the and the photon dose at this level was accurate, as organs receiving 5, 10, 15, and 20 Gy (21.4%, 42.7%, reported by the TPS. For proton therapy, stray dose is Howell et al. Radiation Oncology 2012, 7:116 Page 11 of 12 http://www.ro-journal.com/content/7/1/116 composed almost entirely of secondary neutrons. Dose Received: 26 March 2012 Accepted: 24 July 2012 Published: 24 July 2012 from stray neutrons was not calculated by the TPS. However, previous Monte Carlo studies [13,48,49] have References reported neutron organ doses (for the same proton 1. Dhall G: Medulloblastoma. J Child Neurol 2009, 24(11):1418–1430. treatment apparatus used in this work) between 0.83 2. Gajjar A, Chintagumpala M, Ashley D, Kellie S, Kun LE, Merchant TE, Woo S, and 61 mSv/Gy for proton CSI, which in this study cor- Wheeler G, Ahern V, Krasin MJ, Fouladi M, Broniscer A, Krance R, Hale GA, Stewart CF, Dauser R, Sanford RA, Fuller C, Lau C, Boyett JM, Wallace D, responds to between 0.0194 Sv and 1.43 Sv for the pre- Gilbertson RJ: Risk-adapted craniospinal radiotherapy followed by high- scribed dose of 23.4 Gy. As discussed above, the lowest dose chemotherapy and stem-cell rescue in children with newly dosimetric parameter considered here was the V , and diagnosed medulloblastoma (st jude medulloblastoma-96): Long-term results from a prospective, multicentre trial. Lancet Oncol 2006, inclusion of the stray neutron doses would not have 7(10):813–820. changed the V values. Therefore, neglecting stray neu- 3. Gilbertson RJ: Medulloblastoma: signalling a change in treatment. Lancet tron dose was not a serious limitation of this study. Oncol 2004, 5(4):209–218. 4. Deutsch M, Reigel DH: The value of myelography in the management of Nevertheless, stray dose would be an important compo- childhood medulloblastoma. Cancer 1980, 45(8):2194–2197. nent of a full comparison of photon and proton therapy 5. Allen JC, Epstein F: Medulloblastoma and other primary malignant for CSI, especially for stochastic late effects such as sec- neuroectodermal tumors of the cns - the effect of patients age and extent of disease on prognosis. J Neurosurg 1982, 57(4):446–451. ond cancers, and is therefore part of our ongoing re- 6. Fouladi M, Gajjar A, Boyett JM, Walter AW, Thompson SJ, Merchant TE, search in radiogenic late effects. Jenkins JJ, Langston LW, Liu AY, Kun LE, Heideman RL: Comparison of csf cytology and spinal magnetic resonance imaging in the detection of leptomeningeal disease in pediatric medulloblastoma or primitive Conclusions neuroectodermal tumor. J Clin Oncol 1999, 17(10):3234–3237. In conclusion, this study demonstrated that proton CSI 7. Freeman CR, Taylor RE, Kortmann RD, Carrie C: Radiotherapy for medulloblastoma in children: a perspective on current international improved normal tissue sparing while also providing clinical research efforts. Medical and Pediatric Oncology 2002, 39(2):99–108. more homogeneous target coverage than photon CSI. 8. Packer RJ, Gajjar A, Vezina G, Rorke-Adams L, Burger PC, Robertson PL, Bayer Although the present study focused on medulloblas- L, LaFond D, Donahue BR, Marymont MH, Muraszko K, Langston J, Sposto R: Phase iii study of craniospinal radiation therapy followed by adjuvant toma, our findings are generally applicable to other chemotherapy for newly diagnosed average-risk medulloblastoma. J Clin tumors that are treated with CSI. Future work should Oncol 2006, 24(25):4202–4208. calculate organ equivalent doses, which include both 9. Packer RJ, Vezina G: Management of and prognosis with medulloblastoma therapy at a crossroads. Arch Neurol 2008, therapeutic and stray doses. 65(11):1419–1424. 10. Jarlskog CZ, Lee C, Bolch WE, Xu XG, Paganetti H: Assessment of organ- Competing interests specific neutron equivalent doses in proton therapy using The authors declare that they have no competing interests. computational whole-body age-dependent voxel phantoms. Phys Med Biol 2008, 53(3):693–717. Authors’ contributions 11. Jarlskog CZ, Paganetti H: Risk of developing second cancer from neutron RH conceived of the study, drafted (and edited) the manuscript, participated dose in proton therapy as function of field characteristics, organ, and in creating the FIF photon plans, and performed the statistical analysis. AG patient age. Int J Radiat Oncol Biol Phys 2008, 72(1):228–235. created the proton plans and helped draft the manuscript. WK created the 12. NRC: Health risks from exposure to low levels of ionizing radation: Beir vii - photon FIF treatment plans and specifically was involved in adapting the CSI phase 2. Washington, D.C: Nation Research Council of the National FIF technique for the TPS used in this study. AM provided guidance on Academies; 2006. treatment plan design and reviewed all treatment plans, and reviewed/ 13. Taddei PJ, Mirkovic D, Fontenot JD, Giebeler A, Zheng YS, Kornguth D, edited the manuscript. WN assisted with reviewing/editing the manuscript Mohan R, Newhauser WD: Stray radiation dose and second cancer risk for and provided funding for the study. CE provided guidance for the statistical a pediatric patient receiving craniospinal irradiation with proton beams. analysis and reviewed all statistical results in detail. AD participated in the Phys Med Biol 2009, 54(8):2259–2275. statistical analysis and wrote code to perform the Bonferroni procedure. KH 14. Silber JH, Littman PS, Meadows AT: Stature loss following skeletal wrote codes to extract DVH data from TPS and format data for statistical irradiation for childhood-cancer. J Clin Oncol 1990, 8(2):304–312. analysis; he also assisted with data analysis and in reviewing manuscript. All 15. Chin D, Sklar C, Donahue B, Uli N, Geneiser N, Allen J, Nirenberg A, David R, authors read and approved the final manuscript. Kohn B, Oberfield SE: Thyroid dysfunction as a late effect in survivors of pediatric medulloblastoma primitive neuroectodermal tumors - a Acknowledgments comparison of hyperfractionated versus conventional radiotherapy. This work was funded in part by a grant Northern Illinois University through Cancer 1997, 80(4):798–804. a Department of Defense subcontract (award W81XWH-08-1-0205) and by 16. Constine LS, Woolf PD, Cann D, Mick G, Mccormick K, Raubertas RF, Rubin P: the National Cancer Institute (award 1 R01 CA13146301A1). Hypothalamic pituitary dysfunction after radiation for brain-tumors. N Engl J Med 1993, 328(2):87–94. Author details 17. Oeffinger KC, Mertens AC, Sklar CA, Kawashima T, Hudson MM, Department of Radiation Physics, The University of Texas MD Anderson Meadows AT, Friedman DL, Marina N, Hobbie W, Kadan-Lottick NS, Cancer Center, Houston, TX, USA. Graduate School of Biomedical Sciences, Schwartz CL, Leisenring W, Robison LL, Study CCS: Chronic health The University of Texas at Houston, Houston, TX, USA. Department of conditions in adult survivors of childhood cancer. N Engl J Med Radiation Oncology, The Emory Clinic, Atlanta, GA, USA. Department of 2006, 355(15):1572–1582. Radiation Oncology, The University of Texas MD Anderson Cancer Center, 18. Grau C, Overgaard J: Postirradiation sensorineural hearing loss: a Houston, TX, USA. Department of Epidemiology, The University of Texas MD common but ignored late radiation complication. Int J Radiat Oncol Biol Anderson Cancer Center, Houston, TX, USA. Department of Radiation Phys 1996, 36(2):515–517. Physics, Unit 094, The University of Texas MD Anderson Cancer Center, 1515 19. Mulhern RK, Kepner JL, Thomas PR, Armstrong FD, Friedman HS, Kun LE: Holcombe Blvd, Houston, TX 77030, USA. Present Address: Louisiana State Neuropsychologic functioning of survivors of childhood University, Department of Physics and Astronomy, Baton Rouge, LA, USA. medulloblastoma randomized to receive conventional or reduced-dose Howell et al. Radiation Oncology 2012, 7:116 Page 12 of 12 http://www.ro-journal.com/content/7/1/116 craniospinal irradiation: a pediatric oncology group study. J Clin Oncol 39. Newhauser WD, Giebeler A, Langen KM, Mirkovic D, Mohan R: Can 1998, 16(5):1723–1728. megavoltage computed tomography reduce proton range uncertainties 20. Bowers DC, Liu Y, Leisenring W, McNeil E, Stovall M, Gurney JG, Robison LL, in treatment plans for patients with large metal implants? Phys Med Biol Packer RJ, Oeffinger KC: Late-occurring stroke among long-term survivors 2008, 53(9):2327–2344. of childhood leukemia and brain tumors: a report from the childhood 40. Zhang XD, Dong L, Lee AK, Cox JD, Kuban DA, Zhu RX, Wang XC, Li YP, cancer survivor study. J Clin Oncol 2006, 24(33):5277–5282. Newhauser WD, Gillin M, Mohan R: Effect of anatomic motion on proton 21. Jakacki RI, Goldwein JW, Larsen RL, Barber G, Silber JH: Cardiac dysfunction therapy dose distributions in prostate cancer treatment. Int J Radiat following spinal irradiation during childhood. J Clin Oncol 1993, Oncol Biol Phys 2007, 67(2):620–629. 11(6):1033–1038. 41. Urie M, Goitein M, Wagner M: Compensating for heterogeneities in 22. Mertens AC: Cause of mortality in 5-year survivors of childhood cancer. proton radiation-therapy. Phys Med Biol 1984, 29(5):553–566. Pediatr Blood Cancer 2007, 48(7):723–726. 42. Moyers MF, Miller DW: Range, range modulation, and field radius requirements for proton therapy of prostate cancer. Technol Cancer Res 23. Mertens AC, Liu Q, Neglia JP, Wasilewski K, Leisenring W, Armstrong GT, Treat 2003, 2(5):445–447. Robison LL, Yasui Y: Cause-specific late mortality among 5-year survivors 43. Moyers MF, Miller DW, Bush DA, Slater JD: Methodologies and tools for of childhood cancer: the childhood cancer survivor study. J Natl Cancer proton beam design for lung tumors. Int J Radiat Oncol Biol Phys 2001, Inst 2008, 100(19):1368–1379. 49(5):1429–1438. 24. Reulen RC, Winter DL, Frobisher C, Lancashire ER, Stiller CA, Jenney ME, 44. Vann A, Dasher B, Chestnut S, Wiggers N: Portal design in radiation therapy. Skinner R, Stevens MC, Hawkins MM, Survivor BCC: Long-term cause- Columbia, S.C: R.L. Bryan Co.; 2006. specific mortality among survivors of childhood cancer. JAMA 2010, 45. Yom SS, Frija EK, Mahajan A, Chang E, Klein K, Shiu A, Ohrt J, Woo S: Field- 304(2):172–179. in-field technique with intrafractionally modulated junction shifts for 25. Inskip PD, Curtis RE: New malignancies following childhood cancer in the craniospinal irradiation. Int J Radiat Oncol Biol Phys 2007, 69(4):1193–1198. united states, 1973–2002. Int J Cancer 2007, 121(10):2233–2240. 46. Benjamini Y, Hochberg Y: "Controlling the false discovery rate: A practical 26. Inskip PD, Robison LL, Stovall M, Smith SA, Hammond S, Mertens AC, and powerful approach to multiple testing. Journal of the Royal Statistical Whitton JA, Diller L, Kenney L, Donaldson SS, Meadows AT, Neglia JP: Society Series B (Methodological) 1995, 57(1):289–300. Radiation dose and breast cancer risk in the childhood cancer survivor 47. Howell RM, Scarboro SB, Kry SF, Yaldo DZ: Accuracy of out-of-field dose study. J Clin Oncol 2009, 27(24):3901–3907. calculations by a commercial treatment planning system. Phys Med Biol 27. Meadows AT, Friedman DL, Neglia JP, Mertens AC, Donaldson SS, Stovall M, 2010, 55(23):6999–7008. Hammond S, Yasui Y, Inskip PD: Second neoplasms in survivors of 48. Newhauser WD, Fontenot JD, Mahajan A, Kornguth D, Stovall M, Zheng YS, childhood cancer: Findings from the childhood cancer survivor study Taddei PJ, Mirkovic D, Mohan R, Cox JD, Woo S: The risk of developing a cohort. J Clin Oncol 2009, 27(14):2356–2362. second cancer after receiving craniospinal proton irradiation. Phys Med 28. Neglia JP, Friedman DL, Yasui Y, Mertens AC, Hammond S, Stovall M, Biol 2009, 54(8):2277–2291. Donaldson SS, Meadows AT, Robison LL: Second malignant neoplasms in 49. Taddei PJ, Mahajan A, Mirkovic D, Zhang R, Giebeler A, Kornguth D, Harvey five-year survivors of childhood cancer: Childhood cancer survivor study. M, Woo S, Newhauser WD: Predicted risks of second malignant neoplasm J Natl Cancer Inst 2001, 93(8):618–629. incidence and mortality due to secondary neutrons in a girl and boy 29. Bentel G: Radiation therapy planning: Including problems and solutions. receiving proton craniospinal irradiation. Phys Med Biol 2010, Health Professions Division, New York: McGraw-Hill; 1996. 55(23):7067–7080. 30. Lee CT, Bilton SD, Famiglietti RM, Riley BA, Mahajan A, Chang EL, Maor MH, Woo SY, Cox JD, Smith AR: Treatment planning with protons for pediatric doi:10.1186/1748-717X-7-116 retinoblastoma, medulloblastoma, and pelvic sarcoma: How do protons Cite this article as: Howell et al.: Comparison of therapeutic dosimetric compare with other conformal techniques? Int J Radiat Oncol Biol Phys data from passively scattered proton and photon craniospinal 2005, 63(2):362–372. irradiations for medulloblastoma. Radiation Oncology 2012 7:116. 31. Miralbell R, Lomax A, Bortfeld T, Rouzaud M, Carrie C: Potential role of proton therapy in the treatment of pediatric medulloblastoma primitive neuroectodermal tumors: Reduction of the supratentorial target volume. Int J Radiat Oncol Biol Phys 1997, 38(3):477–484. 32. Slater JD, Yuh GE, Loredo LN, Yonemoto LT, Bush DA, Shahnazi K, Preston W, Slater JM: Reducing toxicity from craniospinal irradiation: Using proton beams to treat medulloblastoma in young children. Cancer J 2004, 10(6):386–390. 33. St Clair WH, Adams JA, Bues M, Fullerton BC, La Shell S, Kooy HM, Loeffler JS, Tarbell NJ: Advantage of protons compared to conventional x-ray or IMRT in the treatment of a pediatric patient with medulloblastoma. Int J Radiat Oncol Biol Phys 2004, 58(3):727–734. 34. Newhauser W, Fontenot J, Zheng YS, Polf J, Titt U, Koch N, Zhang X, Mohan R: Monte carlo simulations for configuring and testing an analytical proton dose-calculation algorithm. Phys Med Biol 2007, 52(15):4569–4584. 35. Das IJ, Cheng CW, Watts RJ, Ahnesjo A, Gibbons J, Li XA, Lowenstein J, Mitra RK, Simon WE, Zhu TC: Accelerator beam data commissioning equipment and procedures: Report of the TG-106 of the therapy physics committee Submit your next manuscript to BioMed Central of the AAPM. Medical Physics 2008, 35(9):4186–4215. and take full advantage of: 36. Fraass B, Doppke K, Hunt M, Kutcher G, Starkschall G, Stern R, Van Dyke J: American association of physicists in medicine radiation therapy • Convenient online submission committee task group 53: Quality assurance for clinical radiotherapy • Thorough peer review treatment planning. Medical Physics 1998, 25(10):1773–1829. 37. ICRU: International comission on radiation units and measurements • No space constraints or color ﬁgure charges report 78: Prescribing, recording, and reporting proton-beam therapy • Immediate publication on acceptance (ICRU report 78). J ICRU 2007, 7(2):1–210. 38. Brodin NP, Rosenschold PMAF, Aznar MC, Kiil-Berthelsen A, Vogelius IR, • Inclusion in PubMed, CAS, Scopus and Google Scholar Nilsson P, Lannering B, Bjork-Eriksson T: Radiobiological risk estimates of • Research which is freely available for redistribution adverse events and secondary cancer for proton and photon radiation therapy of pediatric medulloblastoma. Acta Oncol 2011, 50(6):806–816. Submit your manuscript at www.biomedcentral.com/submit
Radiation Oncology – Springer Journals
Published: Jul 24, 2012
Access the full text.
Sign up today, get DeepDyve free for 14 days.