A Dosimetric Comparison of Whole-Lung Treatment Techniques. in the Pediatric Population

A Dosimetric Comparison of Whole-Lung Treatment Techniques in the Pediatric Population Corresponding Author: Christina L. Bosarge, B.S., R.T. (R) (T) Indiana University School of Medicine Department of Radiation Oncology 535 Barnhill Drive Indianapolis, Indiana 46202-5289 Phone: 317-948-6701 Fax: 317-974-2486 cbosarge@umail.iu.edu Co-Author: Marvene M. Ewing, B.S., C.M.D. Indiana University School of Medicine Department of Radiation Oncology 535 Barnhill Drive Indianapolis, Indiana 46202-5289 Co-Author: Jeffrey Buchsbaum, M.D., Ph.D. Indiana University School of Medicine Department of Radiation Oncology 535 Barnhill Drive Indianapolis, Indiana 46202-5289

Abstract Introduction: To demonstrate the dosimetric advantages and disadvantages of standard anteroposterior-posteroanterior (S-AP/PA), inverse-planned AP/PA (IP- AP/PA), and volumetric-modulated arc (VMAT) radiotherapies in the treatment of children undergoing whole-lung irradiation. Each technique was evaluated by means of target coverage and normal tissue sparing, including data regarding low doses. A historical approach with and without tissue heterogeneity corrections is also demonstrated. Materials and Methods: CT scans of ten children scanned from the neck to the reproductive organs were used. Six plans were created for each: (1) S-AP/PA using the Anisotropic Analytical Algorithm (AAA), (2) IP-AP/PA, (3) VMAT, (4) S-AP/PA without heterogeneity corrections (S-NO), (5) S-AP/PA utilizing the Pencil-Beam algorithm (S- PB) and enforcing monitor units from technique 4, and (6) S-AP/PA with the Anisotropic Analytical Algorithm (S-AAA) also with fixed monitor units. The first three plans compare modern methods and were evaluated based on target coverage and normal tissue sparing. Body maximum and lower body doses (50% and 30%) were also analyzed. Plans 4 through 6 provide a historic view on the progression of heterogeneity algorithms and elucidate what was actually delivered in the past. Results: Averages of each comparison parameter were calculated for all techniques. The S-AP/PA technique resulted in superior target coverage but had the highest maximum dose to every normal tissue structure except the reproductive organs. The IP- AP/PA technique provided the lowest dose to the esophagus, stomach, reproductive organs, and lower body doses. VMAT excelled at body maximum dose and maximum 2

doses to the heart, spine, and spleen, but resulted in the highest reproductive organ and 30% body doses. It was, however, superior to the S-AP/PA approach in the 50% range. Conclusion: Both the VMAT and IP-AP/PA methods provide optimum normal tissue sparing in the high dose regions. VMAT, however, delivers a wider region of low dose. Keywords: Whole-Lung Irradiation, VMAT, Inverse-Planned, Wilms Introduction A remarkable improvement in the survival of Wilms tumor patients has been witnessed over the past four decades. Whole-lung irradiation (WLI) has been a crucial component in the treatment of these children, with current multimodality methods yielding a 90% survival outcome. Prior to this improvement, nearly all of these children died from their disease 1. Given this greater percentage of survivors, the potential for late toxicities resulting from a combination of radiotherapy and cardiac toxic chemotherapy is increasingly becoming a concern for health-care providers. In fact, many reports have shown that WLI with or without doxorubicin has led to a higher prevalence of various cardiac complications including congestive heart failure, myocardial infarction, pericardial disease, and valvular heart disease in these pediatric survivors 2. This study serves to demonstrate the optimal treatment approach with regard to target coverage as well as normal tissue sparing by means of three treatment techniques: (1) standard anteroposterior-posteroanterior (S-AP/PA) radiotherapy, (2) inverse-planned AP/PA (IP-AP/PA) radiotherapy, and (3) volumetric-modulated arc radiotherapy (VMAT). It will also evaluate each technique at the lower dose regions, specifically 50% and 30% doses. A brief historical timeline of WLI techniques with the 3

use and non-use of tissue heterogeneity correction algorithms will also be demonstrated. Literature Review Previous research has been conducted which compared the dosimetric advantages of intensity-modulated radiotherapy (IMRT) over S-AP/PA radiotherapy. John A. Kalapurakal and others at Northwestern University gathered CT scans of 22 children and created plans using these two methods with the finding that IMRT was superior to S-AP/PA in terms of cardiac protection, target coverage, dose uniformity, and heart dose when flank radiotherapy followed WLI 2. However, as is often a concern in the pediatric population, on-table time becomes a major factor and was therefore part of the criteria in choosing treatment techniques for this particular study. Although they discovered great results when using nine beams for their IMRT treatments, information was sought on a simpler AP/PA inverse-planned approach. Comparison of this technique versus the standard AP/PA static field approach was desired, as well as the newer technology known as VMAT. Methods and Materials Treatment plans were created in order to compare the dosimetric advantages and disadvantages of S-AP/PA (using AAA), IP-AP/PA, and VMAT radiotherapy techniques. Each plan was generated using Eclipse version 11 (Varian Medical Systems; Palo Alto, CA). Plans were also created to provide historical insight to the WLI approach of the past involving S-AP/PA without tissue heterogeneity corrections and how it compared to the evolving Pencil-Beam and Anisotropic Analytical algorithms, each with fixed monitor units originating from the original no tissue heterogeneity plan. 4

Normal Tissue Contouring The normal tissues contoured for dose-volume histogram (DVH) evaluation in this study included the esophagus, heart and pericardium (contoured using the RTOG contouring atlas), spinal canal, spleen, stomach, and reproductive organ region. Specifically, the reproductive organ region was contoured by adjusting the volume of interest box with the following borders: superiorly at the level between the first and second sacral segments, inferiorly to include at least a centimeter of external genitalia, and laterally at the outer edges of the femoral heads. Contours drawn outside the body were then cropped to remove any parts that extended outside the body with a 0 cm margin. Target Structure Contouring The gross tumor volume (GTV) consisted of both lungs in their entirety utilizing the acquisition window/level setting. The clinical target volume (CTV) was a 1.5 cm expansion from the GTV and was cropped 0.3 cm within the body, and the planning target volume (PTV) was an additional 0.3 cm extension. A PTV_DVH structure which was cropped 3 mm beneath the skin was created in order to better evaluate PTV coverage on the DVH, as many of the PTV contours on these smaller statures were right at the skin surface where the dose appears to be quite lower. Treatment Planning Ten CT scans of pediatric patients (mean age 3.8 years, range 1-12 years) who were simulated between the years 2010-2014 were used for this study. All scans included the neck superiorly and the reproductive organ region inferiorly. AAA version 5

11.0.31 was used for the S-AP/PA, IP-AP/PA, and VMAT plans. For the historic portion, the No heterogeneity, Pencil-Beam version 10.0.28, and AAA options were applied. All plans used 6-MV photon beams with a total dose of 1,500 cgy at 150 cgy delivered per fraction. For both the S-AP/PA and IP-AP/PA techniques, gantry angles of 0 and 180 were used. The collimator was optimized in the S-AP/PA technique to provide the best MLC conformation to the PTV (90 for all patients). A 0.7 cm margin around the PTV was set. The IP-AP/PA field sizes were not fixed in the optimization page and multiple static segments with the smoothing levels set to thirteen were used. For the VMAT technique, nine out of ten plans consisted of two full arcs. The first was a clockwise arc (Varian IEC scale) from 181 to 179. The second arc was counterclockwise traveling from 179 to 181. The one subject s plan that did not follow this method was actually the oldest (and therefore the largest) and required more arcs and two isocenters. For this patient, four half arcs were used consisting of the following rotations: (1) a clockwise arc traveling from 0 to 179, (2) counter-clockwise from 179 to 0, (3) counter-clockwise from 0 to 181, and finally (4) a clockwise arc going from 181 to 0. A collimator angle of 30 for the clockwise arcs and 330 for the counterclockwise arcs were utilized. The plans were created to provide the best target coverage as was achievable with the heart structure designated as the most important normal tissue organ to be spared. For the historic portion, a S-AP/PA plan was created as described above but was set so that the tissue heterogeneity correction was turned off (S-NO). This is indicative of the type of plan that would have been created in the 1970s, 1980s, and into the 1990s. As can be recalled, this method produced a homogenous isodose distribution, 6

but is not realistic due to its lack of tissue density consideration. This plan was copy and pasted and the calculation model switched to the Pencil-Beam algorithm (S-PB) and is indicative of plans generated in the 1990s and 2000s. Monitor units from the S-NO plan were noted and fixed for this new plan. The same process was repeated for the last technique (S-AAA) with the exception of switching the calculation model to AAA. This provides insight into what was really treated back in the 1970s, 1980s, and 1990s. Results Parameters used for comparison involved target coverage, normal tissue sparing, body maximum dose, and doses in the lower isodose regions (50% and 30%). Target comparison was accomplished by evaluating the percent dose that 95% of the GTV, CTV, PTV, and PTV_DVH volumes received by calculating an average for each technique from the ten scans. The average maximum dose to all normal tissue structures was calculated as well as the average body maximum dose. Evaluation of the lower dose region was determined by computing the percentage of the body receiving 50% and 30% of the prescription dose. This was achieved by dividing the number of cubic centimeters receiving 50% (or 30%) of the prescription dose by the number of cubic centimeters in the body structure. An average was then determined for each set of plans and used for assessment. Target Coverage The S-AP/PA technique provided the best coverage for all target structures. The results of the 95% volume coverage follow: the GTV was covered by 101.8% of the dose, CTV by 98.1%, PTV by 85.7%, and the PTV_DVH by 97.1% of the dose. The VMAT technique came in second, with 95% of the GTV, CTV, PTV, and PTV_DVH 7

volumes having been covered by 97.3%, 93.0%, 84.6%, and 91.9% of the dose, respectively. Finally, the IP-AP/PA produced the lowest coverage, with 95% of the GTV, CTV, PTV, and PTV_DVH volumes covered by 92.3%, 89.0%, 78.6%, and 87.7% of the dose, respectively (see Table 1). Normal Tissue Sparing Maximum doses to the esophagus, heart and pericardium, spinal canal, spleen, stomach, and reproductive organs were collected and the mean calculated for each technique. Additionally, the mean dose to the reproductive region between the ten scans was averaged. Of the six structures, the VMAT technique best protected the heart and pericardium, the spinal canal, and the stomach. The IP-AP/PA technique offered the best protection to the esophagus, stomach, and reproductive organs. The S- AP/PA plan was the least effective at normal tissue sparing. Table 2 reveals the calculated average doses for each organ per treatment approach. Other Doses Body maximum doses were averaged and compared as well. The VMAT technique delivered the lowest body maximum dose (109.6%), followed by the IP- AP/PA (113.0%) and S-AP/PA techniques (115.1%). Lower doses were of concern due to higher integral doses often being associated with dynamic MLC plans. This particular study actually found the IP-AP/PA plan with dynamic MLCs to be the optimum technique in regards to the percentage of the body receiving 50% and 30% of the prescription dose. The VMAT fared better than the S- AP/PA technique at the 50% dose region, and the reverse was true for the 30% region. Specific percentages can be found in Table 3. 8

Historic Segment: Heterogeneity Corrections This portion of the study provides a timeline approach and involved the creation of the following plans: (1) a S-AP/PA plan without any tissue heterogeneity corrections set, (2) the same plan with identical (fixed) monitor units but applying the Pencil-Beam algorithm, and finally (3) repeating the same methodology only with AAA. The results of this particular study show that the targets were actually covered slightly better than was presumed in the past. Additionally, the S-AAA plan demonstrated that the body maximum, esophageal, and spinal canal doses were all lower than was thought in the past. However, the maximum doses to the heart, spleen, stomach, and reproductive organs were all higher than was supposed. Averaged doses to the reproductive organs (mean dose), 50% body dose, and 30% body dose were identical between the two plans. Table 4 provides the average calculated doses with inclusion of data from the S- PB plan. Discussion Wilms tumor is the most common renal mass in children, with radiotherapy remaining a crucial therapeutic component 3. In fact, statistics emanating from the National Wilms Tumor Study show that children diagnosed with favorable histology have a relapse-free and overall survival rate of 72% and 78%, respectively 2. With cardiac toxic chemotherapeutic agents being an important component of their treatment, a larger number of children surviving these days require a greater emphasis to be placed on their possible late tissue effects. For example, reports from the National Wilms Tumor Study suggest that the 20-year congestive heart failure rate was 4.4% after their initial treatment and 17.4% after their first or successive relapse 2. 9

Normal tissue sparing is one of the more essential principles of management in radiotherapy to the pediatric population 2. As such, emphasis should be placed on normal tissue sparing when planning these patients. In this study, however, an emphasis was placed on target coverage so that each technique could be compared side-by-side. If optimization was required, as was the case for the IP-AP/PA and VMAT approaches, the heart was given the highest priority when compared to the other normal structures. IP-AP/PA and VMAT techniques should be the treatment of choice when normal tissue sparing is of utmost concern with VMAT excelling in terms of heart dose. However, these methods do not remain problem-free. Because both require the use of dynamic MLCs, motion interplay effect (especially in the lung) is a possible issue. Fortunately, however, reports have not shown any major negative concerns 4. Another common topic when discussing intensity-modulated treatments is the potential for secondary malignancy production. Although this study found the IP-AP/PA technique to be more effective in terms of lower body doses, reports exist suggesting that they are actually a legitimate issue where secondary malignancies are concerned 5. Additionally, this particular study revealed that the IP-AP/PA technique required over twice the number of monitor units (462 MUs) as compared to the VMAT technique (219 MUs), and over 3.5 times the number of monitor units required for the S-AP/PA technique (130 MUs). Again, with more children surviving, the risk of secondary malignancies remains a legitimate concern. Conclusion 10

VMAT is the optimal approach allowing for the optimization of dose constraints, however there is the concern of scatter and head leakage when this technique is used. Both the VMAT and IP-AP/PA methods provide more normal tissue sparing in the high dose regions than the more conventional S-AP/PA method. However, the VMAT technique delivered a wider region of low dose throughout the patient, raising the risk potential for late effects and the probability of a second malignancy. 11

References 1. Paulino, A.C.; Wen, B.C.; Brown, C.K.; et al. Late Effects in Children Treated with Radiation Therapy for Wilms Tumor. Int. J. Radiat. Oncol. Biol. Phys. 46:1239-1246; 2000. 2. Kalapurakal, J.A.; Zhang, Y.; Kepka, A.; et al. Cardiac-Sparing Whole Lung IMRT in Children with Lung Metastasis. Int. J. Radiat. Oncol. Biol. Phys. 85:761-767; 2013. 3. Van Dijk, I.W.; Oldenburger, F.; Cardous-Ubbink, M.C.; et al. Evaluation of Late Adverse Events in Long-Term Wilms Tumor Survivors. Int. J. Radiat. Oncol. Biol. Phys. 78:370-378; 2010. 4. Seco, J.; Sharp, G.C.; Wu, Z.; et al. Dosimetric Impact of Motion in Free-Breathing and Gated Radiotherapy: a 4D Monte Carlo study of Intrafraction and Interfraction Effects. Med Phys. 35:356-366; 2008. 5. Hall, E.J. Intensity-Modulated Radiation Therapy, Protons and the Risk of Second Cancers. Int. J. Radiat. Oncol. Biol. Phys. 65:1-7; 2006. 12

Table 1. Comparison of target coverage between the S-AP/PA, IP-AP/PA, and VMAT plans TARGET S-AP/PA IP-AP/PA VMAT GTV * 101.8% 92.3% 97.3% CTV * 98.1% 89.0% 93.0% PTV * 85.7% 78.6% 84.6% PTV_DVH * 97.1% 87.7% 91.9% * values are the percentage of the prescription dose that 95% of the target volumes received Table 2. Maximum doses (and mean dose to the reproductive organs) averaged between the ten scans per treatment technique ORGAN S-AP/PA IP-AP/PA VMAT Esophagus 16.63 Gy 14.69 Gy 14.78 Gy Heart and Pericardium 16.72 Gy 15.87 Gy 14.97 Gy Spinal Canal 16.51 Gy 15.1 Gy 14.62 Gy Spleen 16.4 Gy 16.12 Gy 15.96 Gy Stomach 16.0 Gy 15.87 Gy 15.94 Gy Reproductive Organ Max Reproductive Organ Mean 23.6 Gy 19.6 Gy 34.0 Gy 0.66 Gy 0.53 Gy 2.9 Gy 13

Table 3. Table indicating body maximum doses between the three techniques, as well as the percentage of the body receiving 50% and 30% of the prescription dose PARAMETER S-AP/PA IP-AP/PA VMAT Body Maximum 115.1% 113.0% 109.6% % of Body Receiving 50% Rx % of Body Receiving 30% Rx % Variation from IP-AP/PA (50%/30%) * 30.4% 29.0% 29.4% 32.3% 31.1% 32.8% 1.4% / 1.2% 0% / 0% 0.4% / 1.7% * percent variation from the IP-AP/PA plan is listed on the last row 14

Table 4. Comparison of target coverage between the S-NO, S-PB, and S-AAA plans TARGET/ORGAN * values are the percentage of the prescription dose that 95% of the target volumes received S-NO 1970s 1980s maximum doses (and mean dose to the reproductive organs) averaged between the ten scans for each treatment technique body maximum doses between the three algorithms, as well as the percentage of the body receiving 50% and 30% of the prescription dose are listed; percent variation from the IP-AP/PA plan is recorded on the final row 15 S-PB 1990s 2000s S-AAA Present GTV * 99.3% 98.3% 99.4% CTV * 95.2% 94.4% 95.8% PTV * 92.8% 92.0% 83.8% PTV_DVH * 94.2% 93.2% 94.8% Esophagus 16.31 Gy 15.90 Gy 16.26 Gy Heart and Pericardium 15.83 Gy 16.08 Gy 16.33 Gy Spinal Canal 16.21 Gy 15.90 Gy 16.15 Gy Spleen 15.88 Gy 15.70 Gy 16.04 Gy Stomach 15.5 Gy 15.45 Gy 15.61 Gy Reproductive Organ Max Reproductive Organ Mean 22.7 Gy 8.1 Gy 23.2 Gy 0.7 Gy 3.3 Gy 0.7 Gy Body Maximum 111.3% 109.2% 111.0% % of Body Receiving 50% Rx % of Body Receiving 30% Rx % Variation from S-PB (50%/30%) 30.3% 30.9% 30.3% 32.2% 32.4% 32.2% 1.3% / 1.1% 1.9% / 1.3% 1.3% / 1.1%