RESEARCH ARTICLE For reprint orders, please contact: [email protected]
Chest wall and rib irradiation and toxicities of early-stage lung cancer patients treated with CyberKnife stereotactic body radiotherapy Tarun Podder‡,1, Tithi Biswas‡,1, Min Yao1, Yuxia Zhang1, Ellen Kim1, Rodney J Ellis1, Simon S Lo*,1 & Mitchell Machtay1
ABSTRACT Aim: The aim of the study is to evaluate the chest wall and rib toxicities in primary lung cancer patients treated with CyberKnife-based stereotactic body radiotherapy. Materials & methods: In this study, data were collected from the 118 patients, of which 25 patients who had longer follow-up (mean: 21.9 months) were considered. Studied parameters were maximum point dose, doses to 1–100 cm3 of chest wall and 1–10 cm3 of ribs. Results: Three patients developed chest wall pain (grade I). 25 studied patients, on average, received 27.7 Gy to 30 cm3 of chest wall and 50.4 Gy to 1 cm3 of rib. Nine patients had more than 30 Gy dose to 30 cm3 of chest wall. No rib bone fracture was found. Conclusion: No correlations of chest wall pain and volume of irradiation were found. Stereotactic body radiotherapy (SBRT) has become an established modality to treat early stage medically inoperable non-small-cell lung cancer (NSCLC). With SBRT, excellent local control of greater than 90% may be achieved. Since the total treatment is being delivered in a few fractions, the toxicity of such treatment is expected to be remarkably different compared with regular fractionated daily radiotherapy treatment. Over the recent years, one of the common toxicities that have been reported in the published literature is the chronic chest wall pain along with rib fractures, which could range from grade I to quite debilitating and difficult to manage even with aggressive medical treatment. From the available literature, reported rates of grade III chest wall pain ranged from 10 to 44% for lesions close to chest wall [1–5] . Several authors have identified dose–volume parameters predictive of chronic chest wall pain of which volume receiving 30 Gy (V30) is most frequently described [1] . However, the threshold volume receiving 30 Gy as a useful predictive marker has not been uniformly confirmed by different authors [6] . It is also to be noted that having a very tight constraint for any SBRT plan can minimize the effective dose distribution to the target and thereby may lower the tumor control rate. Therefore, more data and information are crucial to establish a robust predictive model for chest wall pain and or rib fractures while making clinical decision for treating primary lung cancer patients. CyberKnife is the only robotic system currently available for real-time active tracking of moving tumors for SBRT treatment. At the University Hospitals Seidman Cancer Center, we have been treating medically inoperable early stage NSCLC for several years. We undertook this retrospective study to evaluate the incidence of chest wall pain, to identify important dose–volume parameters correlating chest wall pain and to compare the differences in dose computation, if any exists, between the two different dose planning algorithms.
KEYWORDS
• chest pain • chest wall toxicity • CyberKnife • rib fracture • stereotactic body
radiotherapy
Department of Radiation Oncology, University Hospitals Seidman Cancer Center, Case Western Reserve University, 11100 Euclid Avenue, LTR B181, Cleveland, OH 44106, USA *Author for correspondence: Tel.: +1 216 286 6740; Fax: +1 216 844 2005; [email protected] ‡ Authors contributed equally 1
10.2217/FON.14.158 © 2014 Future Medicine Ltd
Future Oncol. (2014) 10(15), 2311–2317
part of
ISSN 1479-6694
2311
Research Article Podder, Biswas, Yao et al. Materials & methods SBRT with CyberKnife is one of the most common modalities of treating early stage medically inoperable lung cancers in our clinic. Ray Tracing Algorithm (RTA; also known as effective path length algorithm) was the dose computational method in early years until 2012 when Monte Carlo Algorithm (MCA) was adopted for dose computation. Because of update in software, patients treated prior to 2010 could not be included in this analysis as re-computation was not feasible. An institutional review board approval was obtained prior to implementation of this research. A total of 118 lung patients treated between January 2010 and December 2011 were considered for this study (median age: 73.4 years; range: 48.5–95.0 years; male: 50). Out of 118 patients, 28 patients (23.7%) with 29 tumors abutting the chest wall were identified. Out of 28 patients having solitary tumor abutting chest wall, three patients had follow-up less than 6 months and were excluded from the final analysis. These patients (n = 25) had on average 21.9 months of post-treatment follow-up (range: 6.2–38.9 months; median: 23.2 months). A uniform 2-cm thickness of chest wall and additionally rib were contoured for evaluating dose distribution and associated toxicities (chest wall pain and rib fracture). Treatment plans were first computed using RTA and used for patient treatment; however, for this study, those plans were recomputed using MCA of Accuray’s MultiPlan software (version 3.5.2). Taking the clinically used plans computed using RTA, we set up the MCA uncertainty at 1% and put the computation in high resolution mode of MCA computation without changing the total monitor unit (MU) of the RTA plan. To verify the treatment plan’s accuracy, one needs to create quality
assurance plan using a representative lung phantom. The quality assurance plans when delivered in the phantom (Standard Imaging IMRT Phantom 91230) for measurement, the mean difference between the measured and Monte Carlo-calculated doses was 2.3 ± 1.6% and that for RTA was 1.9 ± 1.3%. Range of prescription dose was 50–60 Gy (mean: 54.6 Gy) in three to five fractions. We collected dosimetric parameters including maximum point dose and doses to 1, 5, 10, 30, 50, 70 and 100 cm 3 of chest wall. These parameters were considered as a wide range of clinically relevant predictors of the studied toxicities. All 118 patients were followed clinically every 3 months with radiographic evaluation every 3–4 months for at least 2 years. Incidence of chest wall pain or rib fracture was documented from the followup notes of each patient. CT scans were most commonly used for follow-up with occasional PET-CT scans for inconclusive impression based on the CT scans. The typical setup for diagnostic CT was as follows: voltage was 120 kV, ampere was 35–40 mA, thickness was 3–5 mm (reconstructed at 2 mm as deemed appropriate), window and level for lung were approximately 1200 and -600 and those for ribs on the range of 1000–3000 and 500–600, respectively. These patients had an average of 15.1 months of posttreatment follow-up (median: 13.9 months). Chest wall pain was graded based on common toxicity criteria (version 3). Statistical significance was considered when p-value was less than 0.05 using two-tailed Student’s t-test. Results Dosimetric statistics of these 25 patients (median age: 76.3 years; range: 44.5–93.3 years; 61% female) are provided in the Table 1. Dosimetric
Table 1. Dose to chest wall and rib when computed using Ray Tracing Algorithm. Structure Chest wall (n = 25): – Mean – Median – Range – SD Rib (n = 25): – Mean – Median – Range – SD
Maximum 1 cm3 (Gy) dose (Gy)
2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy)
61.0 62.7 44.7–74 6.7
54.1 55.0 38.5–69.2 6.6
– – – –
46.2 46.9 33.8–58.5 6.5
39.9 40.2 30.4–51.6 6.2
– – – –
– – – –
– – – –
– – – –
59.8 61.7 30.9–73.6 9.4
27.7 26.1 20–39.9 5.8 50.4 51.2 25–63 9.2
21.4 20.2 13.4–34.6 5.5 45.8 48.5 23–57.1 9.1
17.5 17.5 10.5–30.4 5.4 34.2 36.0 18.3–48.1 8.6
13.5 13.9 5.7–25.7 5.1 22.3 21.7 10.6–39.7 7.6
SD: Standard deviation.c
2312
Future Oncol. (2014) 10(15)
future science group
Chest wall toxicities from stereotactic body radiotherapy
Research Article
Table 2. Difference in chest wall and rib dose distribution computed using Ray Tracing Algorithm and Monte Carlo Algorithm. Structure
Maximum 1 cm3 (Gy) 2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy) dose (Gy)
Chest wall (n = 25): – Maximum difference (%) – Mean difference (%) – p-value Rib (n = 25): – Maximum difference (%) – Mean difference (%) – p-value
28.4 11.1 0.0007
26.9 11.9 0.0010
– – –
– – –
– – –
– – –
22.1 10.4 0.0081 31.4 13.3 0.0016
22.3 10.2 0.0112
20.3 9.8 0.0447
18.4 8.0 0.2519
27.5 7.5 0.3681
26.8 13.0 0.0055
25.4 13.0 0.0087
23.2 12.0 0.0329
23.9 10.0 0.1899
18.7 6.7 0.5495 – – –
Monte Carlo algorithm-computed dose was the reference for the comparison.
computation results revealed that RTA overestimated dose to both the chest wall and the ribs significantly as compared with MCA for all the studied parameters (Table 2) . A representative case depicting dose distribution using RTA and MCA has been shown in Figures 1 & 2. Increase in average dose for RTA was on the range of 6.7–11.9% (p < 0.05, except 50–100 cm 3 p > 0.05), and 10.0–13.0% (p < 0.05) for chest wall and ribs, respectively. The lower the volume of interest, the larger was the difference in dose for the two algorithms. Average difference in maximum point dose for the chest wall was 11.2% (p < 0.001) and was 13.4% for the ribs (p < 0.002). Because our prescription dose varied between three and five fractions, we have also calculated EQD2 using α/β ratio of 3 for normal tissue and 10 for tumor (Table 3) . Nine of the 25 patients, having nine tumors on or adjacent to chest wall, had more than 30 Gy dose to 30 cm3 of chest wall, mean 34.2 Gy (range: 30.1–39.9 Gy). The mean of maximum point dose was 61.0 Gy (range: 44.7–74 Gy) to the chest wall and 59.8 Gy (range: 30.9–73.6 Gy) to the ribs. The corresponding EQD2 values can be found in Table 3. Out of 118 patients, only three patients (2.5%) developed grade I pain in chest wall. One of these three patients (age: 69.3 years) received 30.2 Gy to 30 cm3 of chest wall (maximum point dose to chest wall was 65 Gy [EQD2: 78.9 Gy] and 70 cm3 of the chest wall received 19.7 Gy [EQD2: 37.7 Gy] for a prescription of 60 Gy in three fractions). The second patient (age: 74.7 years) received 65.7 Gy maximum dose (EQD2: 212.1 Gy) and 21.3 Gy (EQD2: 30.9 Gy) to 30 cm3 of the chest wall and 10.5 Gy (EQD2: 10.7 Gy) to 70 cm3 of chest wall, while the third patient (age: 75.3 years) having bilateral right lower lung and left upper lung lesions (dosimetrically non-overlapping)
future science group
received lower maximum dose: 53.5 Gy (maximum; EQD2: 146.6 Gy), 52.0 Gy (maximum; EQD2: 139.4 Gy); 23.5 Gy (30 cm3 of chest wall; EQD2: 36.2 Gy), 29.7 Gy (30 cm3 of chest wall; EQD2: 53.1 Gy); 12.5 Gy (70 cm3 of chest wall; EQD2: 13.8 Gy), 24.8 Gy (70 cm3 of chest wall; EQD2: 44 Gy), respectively. Mean dose to 1 cm3 of rib for the 25 patients was 50.4 Gy (EQD2: 191.0 Gy). No rib bone fracture was found in follow-up CT scans. The EQD2 differences for two algorithms (RTA and MCA) are provided in Table 4. We were unable to establish any correlation between the incidence of chest wall pain and dose–volume parameters. Overall, the incidence of reported chest wall toxicity was low in the entire population under this study. Discussion Chest wall pain and rib fractures are commonly reported toxicities of SBRT for early stage lung cancer. From the available literature, the reported rates of chest wall toxicity (including grade I–III chest wall pain) ranged from 10 to 44%. Spontaneous rib fracture rates were usually reported to be under 10% [1] , but one series reported a rate of as high as 21% [7] . In the series from Indiana University, only 19% of the patients who developed chest wall pain after SBRT actually had associated rib fractures, implicating that the mechanism of chest wall pain might be intercostal nerve injury in a majority of cases [8] . The median time to develop chest wall toxicity is longer than 1 year. Several groups reported various dose–volume histogram parameters trying to develop predictive dose–volume model for clinical practice. However, the dose–volume histogram parameters are not uniform to predict such toxicity. While Dunlap et al. [1] reported 30 cm3 receiving 30 Gy as a significant predictor of chest wall toxicity, Mutter et al. [5] found 70 cm3 receiving greater than 30 Gy is a strong
www.futuremedicine.com
2313
Research Article Podder, Biswas, Yao et al.
A
B
Figure 1. Dose distribution when computed using Ray Tracing Algorithm. (A) 3D beam distribution, dose–volume histogram, axial 2D view of isodose lines and dose tabulation for various regions of interest. (B) 3D distribution of beams, axial, sagittal and coronal view of isodose lines.
predictor for grade 2 chest wall toxicity compared with 30 cm3. Creach et al. noted percentage of chest wall volume receiving 30 Gy rather than absolute cm3 volume as a strong predictor for chest wall pain [9] . Andolino et al. from Indiana University reported Dmax of greater than
A
50 Gy as a risk factor for chest wall toxicity [8] . However, they did not find any threshold volume to predict chest wall toxicity, but a gradual increase in risk as the volume receiving a particular dose increased. A separate report from Cleveland Clinic found a strong correlation with
B
Figure 2. Dose distribution when computed using Monte Carlo Algorithm. (A) 3D beam distribution, dose–volume histogram, axial 2D view of isodose lines and dose tabulation for various regions of interest. (B) 3D distribution of beams, axial, sagittal and coronal view of isodose lines.
2314
Future Oncol. (2014) 10(15)
future science group
Chest wall toxicities from stereotactic body radiotherapy
Research Article
Table 3. EQD2 of distributed dose to chest wall and rib when computed using Ray Tracing Algorithm. Structure Chest wall (n = 25): – Mean dose – Median – Range – SD Rib (n = 25): – Mean dose – Median – Range – SD
Maximum dose (Gy)
1 cm3 (Gy)
2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy)
259.8 276.9 139.4–409.5 74.1
208.9 228.4 103.4–360.8 63.0
– – – –
156.7 162.5 70.3–263.3 49.5
120.1 118.0 55.2–208.5 39.8
64.2 55.1 28.2–128.9 22.3
41.9 37.9 15.2–81.3 19.4
30.4 27.5 10.7–64.4 15.4
20.5 18.8 4.7–48.4 11.6
260.6 276.9 82.2–405.3 82.6
191.0 201.1 56.7–302.4 63.7
161.6 172.1 49.1–251.6 56.9
96.7 108.0 29.3–173.4 41.3
47.1 44.4 10.9–100.1 26.4
– – – –
– – – –
– – – –
– – – –
SD: Standard deviation.
increased chest wall pain with prescription of 60 Gy in three fractions compared with 50 Gy in five fractions [6] . We report only 2.5% incidence of grade I chest wall pain and no rib fracture with a median follow-up of 13.9 months. This incidence rate is on the lower side compared with other published reports. Because of very low incidence of actual events, we could not establish any dosimetric correlation between dose–volume parameters and chest wall toxicity. We also calculated the EQD2 for chest wall and ribs and even though the mean maximum point dose to chest wall was 259.8 Gy (range: 139.4–409.5 Gy), again we were unable to establish any correlation between the two. It is expected to have longer follow-up (more than 12 months) for having better analysis of occurrence of rib fracture after SBRT. Out of 25 patients in this study, only three patients had less than 12 months of follow-up (6.2, 10.1 and 10.4 months). The median follow-up for all patients (n = 25) was 23.2 months (mean: 21.9 months). We also compared commonly used RTA that was used for original dose calculation to MCA. Prior dosimetric studies showed that RTA was an inferior treatment planning algorithm in chest locations, where there is significant heterogeneity
because of the lungs, compared with MCA [10,11] . In our study, as compared with MCA, RTA algorithm significantly overestimated dose to chest wall and ribs. It is plausible that the actual dose to these critical structures was lower as shown by MCA. As often the decision becomes very critical between adequate coverage of the target while sparing the critical structures, it becomes of utmost importance to use the appropriate dose computation technique for accurate dose calculation. Although treatment planning with MCA is more labor-intensive, the latter should be used because of its accuracy in dose computation and significant difference in dose distribution in chest wall and ribs. Currently, we are in the process of comparing target coverage using RTA and MCA. Conclusion & future perspective From this study at our center, we observed low incidence of chest wall and rib toxicities using CyberKnife SBRT technique for treatment of medically inoperable early stage lung cancer. Because of RTA’s significant overestimation in dose distribution to chest wall and rib, MCA should be used for more accurate dose calculation. The more accurate estimate may facilitate not only better control of toxicity, but also allow
Table 4. Difference in EQD2 for chest wall and rib computed using Ray Tracing Algorithm and Monte Carlo Algorithm. Structure
Maximum 1 cm3 (Gy) 2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy) dose (Gy)
Chest wall (n = 25): – Mean difference (%) – p-value Rib (n = 25): – Mean difference (%) – p-value
21.9 0.0097
23.2 0.0138
– –
20.2 0.0448
19.3 0.0439
17.7 0.0049
14.2 0.28915
12.8 0.3782
11.0 0.4863
26.5 0.0019
25.3 0.0033
25.0 0.0048
22.3 0.0180
17.5 0.1378
– –
– –
– –
– –
Monte Carlo algorithm-computed dose was the reference for the comparison.
future science group
www.futuremedicine.com
2315
Research Article Podder, Biswas, Yao et al. use of higher doses to target tumor volumes, which may yield better clinical outcomes. More studies are needed to formulate the dose–volume risk factors to predict future chest wall toxicity.
or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.
Financial & competing interests disclosure
Ethical conduct
SS Lo is a member of the Elekta Oligometastasis Core Group (Elekta AB) and has received travel funds and honorarium from Varian Medical System for educational presentation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter
The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
EXECUTIVE SUMMARY Purpose ●●
To examine chest wall toxicities from CyberKnife-based stereotactic body radiotherapy (SBRT).
Materials & methods ●●
A total of 25 patients with lung tumors abutting the chest wall treated with SBRT were examined:
●●
Median follow-up was 23.2 months; mean was 21.9 months.
●●
Prescribed dose ranged from 50 Gy in five fractions to 60 Gy in three fractions:
●●
EQD2 was computed to consider the variation in prescribed dose.
●●
Dosimetric parameters were studied and correlated with chest wall toxicities.
Results ●●
Three patients developed grade 1 chest wall pain.
●●
No rib fractions were observed.
●●
A correlation of the incidence of chest wall pain and dose–volume parameters could not be established.
Conclusion ●●
No correlations of chest wall pain and volume of irradiation (or maximum point dose) were found.
●●
All these patients received CyberKnife-based SBRT treatments that were planned based on Ray Tracing Algorithm and this had overestimated the dose delivered to the chest wall.
References
•
Papers of special note have been highlighted as: • of interest; •• of considerable interest
Important study of chest wall toxicity from SBRT.
3
Welsh J, Thomas J, Shah D et al. Obesity increases the risk of chest wall pain from thoracic stereotactic body radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 81(1), 91–96 (2011).
•
Important study of chest wall pain from SBRT.
4
Ong CL, Palma D, Verbakel WF, Slotman BJ, Senan S. Treatment of large stage I–II lung tumors using stereotactic body radiotherapy (SBRT): planning considerations and early toxicity. Radiother. Oncol. 97(3), 431–436 (2010).
1
Dunlap NE, Cai J, Biedermann GB et al. Chest wall volume receiving >30 Gy predicts risk of severe pain and/or rib fracture after lung stereotactic body radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 76(3), 796–801 (2010).
•• Seminal study of chest wall toxicity from stereotactic body radiotherapy (SBRT). 2
Stephans KL, Djemil T, Tendulkar RD, Robinson CG, Reddy CA, Videtic GM. Prediction of chest wall toxicity from lung stereotactic body radiotherapy (SBRT). Int. J. Radiat. Oncol. Biol. Phys. 82(2), 974–980 (2012).
2316
Future Oncol. (2014) 10(15)
5
Mutter RW, Liu F, Abreu A, Yorke E, Jackson A, Rosenzweig KE. Dose–volume parameters predict for the development of chest wall pain after stereotactic body radiation for lung cancer. Int. J. Radiat. Oncol. Biol. Phys. 82(5), 1783–1790 (2012).
6
Woody NM, Videtic GM, Stephans KL, Djemil T, Kim Y, Xia P. Predicting chest wall pain from lung stereotactic body radiotherapy for different fractionation schemes. Int. J. Radiat. Oncol. Biol. Phys. 83(1), 427–434 (2012).
•
Important study of chest wall pain from SBRT.
7
Voroney JP, Hope A, Dahele MR et al. Chest wall pain and rib fracture after stereotactic
future science group
Chest wall toxicities from stereotactic body radiotherapy radiotherapy for peripheral non-small cell lung cancer. J. Thorac. Oncol. 4(8), 1035–1037 (2009). 8
Andolino DL, Forquer JA, Henderson MA et al. Chest wall toxicity after stereotactic body radiotherapy for malignant lesions of the lung and liver. Int. J. Radiat. Oncol. Biol. Phys. 80(3), 692–697 (2011).
future science group
•
Important study of chest wall toxicity from SBRT.
9
Creach KM, El Naqa I, Bradley JD et al. Dosimetric predictors of chest wall pain after lung stereotactic body radiotherapy. Radiother. Oncol. 104(1), 23–27 (2012).
10 Sharma SC, Ott JT, Williams JB, Dickow D.
Clinical implications of adopting Monte
Research Article
Carlo treatment planning for CyberKnife. J. Appl. Clin. Med. Phys. 11(1), 3142 (2011). 11 Wu VW, Tam KW, Tong SM. Evaluation of
the influence of tumor location and size on the difference of dose calculation between Ray Tracing Algorithm and Fast Monte Carlo algorithm in stereotactic body radiotherapy of non-small cell lung cancer using CyberKnife. J. Appl. Clin. Med. Phys. 14(5), 68–78 (2013).
www.futuremedicine.com
2317
Chest wall and rib irradiation and toxicities of early-stage lung cancer patients treated with CyberKnife stereotactic body radiotherapy Tarun Podder‡,1, Tithi Biswas‡,1, Min Yao1, Yuxia Zhang1, Ellen Kim1, Rodney J Ellis1, Simon S Lo*,1 & Mitchell Machtay1
ABSTRACT Aim: The aim of the study is to evaluate the chest wall and rib toxicities in primary lung cancer patients treated with CyberKnife-based stereotactic body radiotherapy. Materials & methods: In this study, data were collected from the 118 patients, of which 25 patients who had longer follow-up (mean: 21.9 months) were considered. Studied parameters were maximum point dose, doses to 1–100 cm3 of chest wall and 1–10 cm3 of ribs. Results: Three patients developed chest wall pain (grade I). 25 studied patients, on average, received 27.7 Gy to 30 cm3 of chest wall and 50.4 Gy to 1 cm3 of rib. Nine patients had more than 30 Gy dose to 30 cm3 of chest wall. No rib bone fracture was found. Conclusion: No correlations of chest wall pain and volume of irradiation were found. Stereotactic body radiotherapy (SBRT) has become an established modality to treat early stage medically inoperable non-small-cell lung cancer (NSCLC). With SBRT, excellent local control of greater than 90% may be achieved. Since the total treatment is being delivered in a few fractions, the toxicity of such treatment is expected to be remarkably different compared with regular fractionated daily radiotherapy treatment. Over the recent years, one of the common toxicities that have been reported in the published literature is the chronic chest wall pain along with rib fractures, which could range from grade I to quite debilitating and difficult to manage even with aggressive medical treatment. From the available literature, reported rates of grade III chest wall pain ranged from 10 to 44% for lesions close to chest wall [1–5] . Several authors have identified dose–volume parameters predictive of chronic chest wall pain of which volume receiving 30 Gy (V30) is most frequently described [1] . However, the threshold volume receiving 30 Gy as a useful predictive marker has not been uniformly confirmed by different authors [6] . It is also to be noted that having a very tight constraint for any SBRT plan can minimize the effective dose distribution to the target and thereby may lower the tumor control rate. Therefore, more data and information are crucial to establish a robust predictive model for chest wall pain and or rib fractures while making clinical decision for treating primary lung cancer patients. CyberKnife is the only robotic system currently available for real-time active tracking of moving tumors for SBRT treatment. At the University Hospitals Seidman Cancer Center, we have been treating medically inoperable early stage NSCLC for several years. We undertook this retrospective study to evaluate the incidence of chest wall pain, to identify important dose–volume parameters correlating chest wall pain and to compare the differences in dose computation, if any exists, between the two different dose planning algorithms.
KEYWORDS
• chest pain • chest wall toxicity • CyberKnife • rib fracture • stereotactic body
radiotherapy
Department of Radiation Oncology, University Hospitals Seidman Cancer Center, Case Western Reserve University, 11100 Euclid Avenue, LTR B181, Cleveland, OH 44106, USA *Author for correspondence: Tel.: +1 216 286 6740; Fax: +1 216 844 2005; [email protected] ‡ Authors contributed equally 1
10.2217/FON.14.158 © 2014 Future Medicine Ltd
Future Oncol. (2014) 10(15), 2311–2317
part of
ISSN 1479-6694
2311
Research Article Podder, Biswas, Yao et al. Materials & methods SBRT with CyberKnife is one of the most common modalities of treating early stage medically inoperable lung cancers in our clinic. Ray Tracing Algorithm (RTA; also known as effective path length algorithm) was the dose computational method in early years until 2012 when Monte Carlo Algorithm (MCA) was adopted for dose computation. Because of update in software, patients treated prior to 2010 could not be included in this analysis as re-computation was not feasible. An institutional review board approval was obtained prior to implementation of this research. A total of 118 lung patients treated between January 2010 and December 2011 were considered for this study (median age: 73.4 years; range: 48.5–95.0 years; male: 50). Out of 118 patients, 28 patients (23.7%) with 29 tumors abutting the chest wall were identified. Out of 28 patients having solitary tumor abutting chest wall, three patients had follow-up less than 6 months and were excluded from the final analysis. These patients (n = 25) had on average 21.9 months of post-treatment follow-up (range: 6.2–38.9 months; median: 23.2 months). A uniform 2-cm thickness of chest wall and additionally rib were contoured for evaluating dose distribution and associated toxicities (chest wall pain and rib fracture). Treatment plans were first computed using RTA and used for patient treatment; however, for this study, those plans were recomputed using MCA of Accuray’s MultiPlan software (version 3.5.2). Taking the clinically used plans computed using RTA, we set up the MCA uncertainty at 1% and put the computation in high resolution mode of MCA computation without changing the total monitor unit (MU) of the RTA plan. To verify the treatment plan’s accuracy, one needs to create quality
assurance plan using a representative lung phantom. The quality assurance plans when delivered in the phantom (Standard Imaging IMRT Phantom 91230) for measurement, the mean difference between the measured and Monte Carlo-calculated doses was 2.3 ± 1.6% and that for RTA was 1.9 ± 1.3%. Range of prescription dose was 50–60 Gy (mean: 54.6 Gy) in three to five fractions. We collected dosimetric parameters including maximum point dose and doses to 1, 5, 10, 30, 50, 70 and 100 cm 3 of chest wall. These parameters were considered as a wide range of clinically relevant predictors of the studied toxicities. All 118 patients were followed clinically every 3 months with radiographic evaluation every 3–4 months for at least 2 years. Incidence of chest wall pain or rib fracture was documented from the followup notes of each patient. CT scans were most commonly used for follow-up with occasional PET-CT scans for inconclusive impression based on the CT scans. The typical setup for diagnostic CT was as follows: voltage was 120 kV, ampere was 35–40 mA, thickness was 3–5 mm (reconstructed at 2 mm as deemed appropriate), window and level for lung were approximately 1200 and -600 and those for ribs on the range of 1000–3000 and 500–600, respectively. These patients had an average of 15.1 months of posttreatment follow-up (median: 13.9 months). Chest wall pain was graded based on common toxicity criteria (version 3). Statistical significance was considered when p-value was less than 0.05 using two-tailed Student’s t-test. Results Dosimetric statistics of these 25 patients (median age: 76.3 years; range: 44.5–93.3 years; 61% female) are provided in the Table 1. Dosimetric
Table 1. Dose to chest wall and rib when computed using Ray Tracing Algorithm. Structure Chest wall (n = 25): – Mean – Median – Range – SD Rib (n = 25): – Mean – Median – Range – SD
Maximum 1 cm3 (Gy) dose (Gy)
2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy)
61.0 62.7 44.7–74 6.7
54.1 55.0 38.5–69.2 6.6
– – – –
46.2 46.9 33.8–58.5 6.5
39.9 40.2 30.4–51.6 6.2
– – – –
– – – –
– – – –
– – – –
59.8 61.7 30.9–73.6 9.4
27.7 26.1 20–39.9 5.8 50.4 51.2 25–63 9.2
21.4 20.2 13.4–34.6 5.5 45.8 48.5 23–57.1 9.1
17.5 17.5 10.5–30.4 5.4 34.2 36.0 18.3–48.1 8.6
13.5 13.9 5.7–25.7 5.1 22.3 21.7 10.6–39.7 7.6
SD: Standard deviation.c
2312
Future Oncol. (2014) 10(15)
future science group
Chest wall toxicities from stereotactic body radiotherapy
Research Article
Table 2. Difference in chest wall and rib dose distribution computed using Ray Tracing Algorithm and Monte Carlo Algorithm. Structure
Maximum 1 cm3 (Gy) 2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy) dose (Gy)
Chest wall (n = 25): – Maximum difference (%) – Mean difference (%) – p-value Rib (n = 25): – Maximum difference (%) – Mean difference (%) – p-value
28.4 11.1 0.0007
26.9 11.9 0.0010
– – –
– – –
– – –
– – –
22.1 10.4 0.0081 31.4 13.3 0.0016
22.3 10.2 0.0112
20.3 9.8 0.0447
18.4 8.0 0.2519
27.5 7.5 0.3681
26.8 13.0 0.0055
25.4 13.0 0.0087
23.2 12.0 0.0329
23.9 10.0 0.1899
18.7 6.7 0.5495 – – –
Monte Carlo algorithm-computed dose was the reference for the comparison.
computation results revealed that RTA overestimated dose to both the chest wall and the ribs significantly as compared with MCA for all the studied parameters (Table 2) . A representative case depicting dose distribution using RTA and MCA has been shown in Figures 1 & 2. Increase in average dose for RTA was on the range of 6.7–11.9% (p < 0.05, except 50–100 cm 3 p > 0.05), and 10.0–13.0% (p < 0.05) for chest wall and ribs, respectively. The lower the volume of interest, the larger was the difference in dose for the two algorithms. Average difference in maximum point dose for the chest wall was 11.2% (p < 0.001) and was 13.4% for the ribs (p < 0.002). Because our prescription dose varied between three and five fractions, we have also calculated EQD2 using α/β ratio of 3 for normal tissue and 10 for tumor (Table 3) . Nine of the 25 patients, having nine tumors on or adjacent to chest wall, had more than 30 Gy dose to 30 cm3 of chest wall, mean 34.2 Gy (range: 30.1–39.9 Gy). The mean of maximum point dose was 61.0 Gy (range: 44.7–74 Gy) to the chest wall and 59.8 Gy (range: 30.9–73.6 Gy) to the ribs. The corresponding EQD2 values can be found in Table 3. Out of 118 patients, only three patients (2.5%) developed grade I pain in chest wall. One of these three patients (age: 69.3 years) received 30.2 Gy to 30 cm3 of chest wall (maximum point dose to chest wall was 65 Gy [EQD2: 78.9 Gy] and 70 cm3 of the chest wall received 19.7 Gy [EQD2: 37.7 Gy] for a prescription of 60 Gy in three fractions). The second patient (age: 74.7 years) received 65.7 Gy maximum dose (EQD2: 212.1 Gy) and 21.3 Gy (EQD2: 30.9 Gy) to 30 cm3 of the chest wall and 10.5 Gy (EQD2: 10.7 Gy) to 70 cm3 of chest wall, while the third patient (age: 75.3 years) having bilateral right lower lung and left upper lung lesions (dosimetrically non-overlapping)
future science group
received lower maximum dose: 53.5 Gy (maximum; EQD2: 146.6 Gy), 52.0 Gy (maximum; EQD2: 139.4 Gy); 23.5 Gy (30 cm3 of chest wall; EQD2: 36.2 Gy), 29.7 Gy (30 cm3 of chest wall; EQD2: 53.1 Gy); 12.5 Gy (70 cm3 of chest wall; EQD2: 13.8 Gy), 24.8 Gy (70 cm3 of chest wall; EQD2: 44 Gy), respectively. Mean dose to 1 cm3 of rib for the 25 patients was 50.4 Gy (EQD2: 191.0 Gy). No rib bone fracture was found in follow-up CT scans. The EQD2 differences for two algorithms (RTA and MCA) are provided in Table 4. We were unable to establish any correlation between the incidence of chest wall pain and dose–volume parameters. Overall, the incidence of reported chest wall toxicity was low in the entire population under this study. Discussion Chest wall pain and rib fractures are commonly reported toxicities of SBRT for early stage lung cancer. From the available literature, the reported rates of chest wall toxicity (including grade I–III chest wall pain) ranged from 10 to 44%. Spontaneous rib fracture rates were usually reported to be under 10% [1] , but one series reported a rate of as high as 21% [7] . In the series from Indiana University, only 19% of the patients who developed chest wall pain after SBRT actually had associated rib fractures, implicating that the mechanism of chest wall pain might be intercostal nerve injury in a majority of cases [8] . The median time to develop chest wall toxicity is longer than 1 year. Several groups reported various dose–volume histogram parameters trying to develop predictive dose–volume model for clinical practice. However, the dose–volume histogram parameters are not uniform to predict such toxicity. While Dunlap et al. [1] reported 30 cm3 receiving 30 Gy as a significant predictor of chest wall toxicity, Mutter et al. [5] found 70 cm3 receiving greater than 30 Gy is a strong
www.futuremedicine.com
2313
Research Article Podder, Biswas, Yao et al.
A
B
Figure 1. Dose distribution when computed using Ray Tracing Algorithm. (A) 3D beam distribution, dose–volume histogram, axial 2D view of isodose lines and dose tabulation for various regions of interest. (B) 3D distribution of beams, axial, sagittal and coronal view of isodose lines.
predictor for grade 2 chest wall toxicity compared with 30 cm3. Creach et al. noted percentage of chest wall volume receiving 30 Gy rather than absolute cm3 volume as a strong predictor for chest wall pain [9] . Andolino et al. from Indiana University reported Dmax of greater than
A
50 Gy as a risk factor for chest wall toxicity [8] . However, they did not find any threshold volume to predict chest wall toxicity, but a gradual increase in risk as the volume receiving a particular dose increased. A separate report from Cleveland Clinic found a strong correlation with
B
Figure 2. Dose distribution when computed using Monte Carlo Algorithm. (A) 3D beam distribution, dose–volume histogram, axial 2D view of isodose lines and dose tabulation for various regions of interest. (B) 3D distribution of beams, axial, sagittal and coronal view of isodose lines.
2314
Future Oncol. (2014) 10(15)
future science group
Chest wall toxicities from stereotactic body radiotherapy
Research Article
Table 3. EQD2 of distributed dose to chest wall and rib when computed using Ray Tracing Algorithm. Structure Chest wall (n = 25): – Mean dose – Median – Range – SD Rib (n = 25): – Mean dose – Median – Range – SD
Maximum dose (Gy)
1 cm3 (Gy)
2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy)
259.8 276.9 139.4–409.5 74.1
208.9 228.4 103.4–360.8 63.0
– – – –
156.7 162.5 70.3–263.3 49.5
120.1 118.0 55.2–208.5 39.8
64.2 55.1 28.2–128.9 22.3
41.9 37.9 15.2–81.3 19.4
30.4 27.5 10.7–64.4 15.4
20.5 18.8 4.7–48.4 11.6
260.6 276.9 82.2–405.3 82.6
191.0 201.1 56.7–302.4 63.7
161.6 172.1 49.1–251.6 56.9
96.7 108.0 29.3–173.4 41.3
47.1 44.4 10.9–100.1 26.4
– – – –
– – – –
– – – –
– – – –
SD: Standard deviation.
increased chest wall pain with prescription of 60 Gy in three fractions compared with 50 Gy in five fractions [6] . We report only 2.5% incidence of grade I chest wall pain and no rib fracture with a median follow-up of 13.9 months. This incidence rate is on the lower side compared with other published reports. Because of very low incidence of actual events, we could not establish any dosimetric correlation between dose–volume parameters and chest wall toxicity. We also calculated the EQD2 for chest wall and ribs and even though the mean maximum point dose to chest wall was 259.8 Gy (range: 139.4–409.5 Gy), again we were unable to establish any correlation between the two. It is expected to have longer follow-up (more than 12 months) for having better analysis of occurrence of rib fracture after SBRT. Out of 25 patients in this study, only three patients had less than 12 months of follow-up (6.2, 10.1 and 10.4 months). The median follow-up for all patients (n = 25) was 23.2 months (mean: 21.9 months). We also compared commonly used RTA that was used for original dose calculation to MCA. Prior dosimetric studies showed that RTA was an inferior treatment planning algorithm in chest locations, where there is significant heterogeneity
because of the lungs, compared with MCA [10,11] . In our study, as compared with MCA, RTA algorithm significantly overestimated dose to chest wall and ribs. It is plausible that the actual dose to these critical structures was lower as shown by MCA. As often the decision becomes very critical between adequate coverage of the target while sparing the critical structures, it becomes of utmost importance to use the appropriate dose computation technique for accurate dose calculation. Although treatment planning with MCA is more labor-intensive, the latter should be used because of its accuracy in dose computation and significant difference in dose distribution in chest wall and ribs. Currently, we are in the process of comparing target coverage using RTA and MCA. Conclusion & future perspective From this study at our center, we observed low incidence of chest wall and rib toxicities using CyberKnife SBRT technique for treatment of medically inoperable early stage lung cancer. Because of RTA’s significant overestimation in dose distribution to chest wall and rib, MCA should be used for more accurate dose calculation. The more accurate estimate may facilitate not only better control of toxicity, but also allow
Table 4. Difference in EQD2 for chest wall and rib computed using Ray Tracing Algorithm and Monte Carlo Algorithm. Structure
Maximum 1 cm3 (Gy) 2 cm3 (Gy) 5 cm3 (Gy) 10 cm3 (Gy) 30 cm3 (Gy) 50 cm3 (Gy) 70 cm3 (Gy) 100 cm3 (Gy) dose (Gy)
Chest wall (n = 25): – Mean difference (%) – p-value Rib (n = 25): – Mean difference (%) – p-value
21.9 0.0097
23.2 0.0138
– –
20.2 0.0448
19.3 0.0439
17.7 0.0049
14.2 0.28915
12.8 0.3782
11.0 0.4863
26.5 0.0019
25.3 0.0033
25.0 0.0048
22.3 0.0180
17.5 0.1378
– –
– –
– –
– –
Monte Carlo algorithm-computed dose was the reference for the comparison.
future science group
www.futuremedicine.com
2315
Research Article Podder, Biswas, Yao et al. use of higher doses to target tumor volumes, which may yield better clinical outcomes. More studies are needed to formulate the dose–volume risk factors to predict future chest wall toxicity.
or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.
Financial & competing interests disclosure
Ethical conduct
SS Lo is a member of the Elekta Oligometastasis Core Group (Elekta AB) and has received travel funds and honorarium from Varian Medical System for educational presentation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter
The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
EXECUTIVE SUMMARY Purpose ●●
To examine chest wall toxicities from CyberKnife-based stereotactic body radiotherapy (SBRT).
Materials & methods ●●
A total of 25 patients with lung tumors abutting the chest wall treated with SBRT were examined:
●●
Median follow-up was 23.2 months; mean was 21.9 months.
●●
Prescribed dose ranged from 50 Gy in five fractions to 60 Gy in three fractions:
●●
EQD2 was computed to consider the variation in prescribed dose.
●●
Dosimetric parameters were studied and correlated with chest wall toxicities.
Results ●●
Three patients developed grade 1 chest wall pain.
●●
No rib fractions were observed.
●●
A correlation of the incidence of chest wall pain and dose–volume parameters could not be established.
Conclusion ●●
No correlations of chest wall pain and volume of irradiation (or maximum point dose) were found.
●●
All these patients received CyberKnife-based SBRT treatments that were planned based on Ray Tracing Algorithm and this had overestimated the dose delivered to the chest wall.
References
•
Papers of special note have been highlighted as: • of interest; •• of considerable interest
Important study of chest wall toxicity from SBRT.
3
Welsh J, Thomas J, Shah D et al. Obesity increases the risk of chest wall pain from thoracic stereotactic body radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 81(1), 91–96 (2011).
•
Important study of chest wall pain from SBRT.
4
Ong CL, Palma D, Verbakel WF, Slotman BJ, Senan S. Treatment of large stage I–II lung tumors using stereotactic body radiotherapy (SBRT): planning considerations and early toxicity. Radiother. Oncol. 97(3), 431–436 (2010).
1
Dunlap NE, Cai J, Biedermann GB et al. Chest wall volume receiving >30 Gy predicts risk of severe pain and/or rib fracture after lung stereotactic body radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 76(3), 796–801 (2010).
•• Seminal study of chest wall toxicity from stereotactic body radiotherapy (SBRT). 2
Stephans KL, Djemil T, Tendulkar RD, Robinson CG, Reddy CA, Videtic GM. Prediction of chest wall toxicity from lung stereotactic body radiotherapy (SBRT). Int. J. Radiat. Oncol. Biol. Phys. 82(2), 974–980 (2012).
2316
Future Oncol. (2014) 10(15)
5
Mutter RW, Liu F, Abreu A, Yorke E, Jackson A, Rosenzweig KE. Dose–volume parameters predict for the development of chest wall pain after stereotactic body radiation for lung cancer. Int. J. Radiat. Oncol. Biol. Phys. 82(5), 1783–1790 (2012).
6
Woody NM, Videtic GM, Stephans KL, Djemil T, Kim Y, Xia P. Predicting chest wall pain from lung stereotactic body radiotherapy for different fractionation schemes. Int. J. Radiat. Oncol. Biol. Phys. 83(1), 427–434 (2012).
•
Important study of chest wall pain from SBRT.
7
Voroney JP, Hope A, Dahele MR et al. Chest wall pain and rib fracture after stereotactic
future science group
Chest wall toxicities from stereotactic body radiotherapy radiotherapy for peripheral non-small cell lung cancer. J. Thorac. Oncol. 4(8), 1035–1037 (2009). 8
Andolino DL, Forquer JA, Henderson MA et al. Chest wall toxicity after stereotactic body radiotherapy for malignant lesions of the lung and liver. Int. J. Radiat. Oncol. Biol. Phys. 80(3), 692–697 (2011).
future science group
•
Important study of chest wall toxicity from SBRT.
9
Creach KM, El Naqa I, Bradley JD et al. Dosimetric predictors of chest wall pain after lung stereotactic body radiotherapy. Radiother. Oncol. 104(1), 23–27 (2012).
10 Sharma SC, Ott JT, Williams JB, Dickow D.
Clinical implications of adopting Monte
Research Article
Carlo treatment planning for CyberKnife. J. Appl. Clin. Med. Phys. 11(1), 3142 (2011). 11 Wu VW, Tam KW, Tong SM. Evaluation of
the influence of tumor location and size on the difference of dose calculation between Ray Tracing Algorithm and Fast Monte Carlo algorithm in stereotactic body radiotherapy of non-small cell lung cancer using CyberKnife. J. Appl. Clin. Med. Phys. 14(5), 68–78 (2013).
www.futuremedicine.com
2317
