Radiotherapy and Oncology 110 (2014) 517–522
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Lung cancer RT
Adaptive radiotherapy of lung cancer patients with pleural effusion or atelectasis Ditte Sloth Møller a,⇑, Azza Ahmed Khalil b, Marianne Marquard Knap b, Lone Hoffmann a a
Department of Medical Physics; and b Department of Oncology, Aarhus University Hospital, Denmark
a r t i c l e
i n f o
Article history: Received 31 May 2013 Received in revised form 20 September 2013 Accepted 3 October 2013 Available online 31 October 2013 Keywords: Adaptive RT IGRT Atelectasis NSCLC SCLC
a b s t r a c t Background and purpose: Changes in lung density due to atelectasis, pleural effusion and pneumonia/ pneumonitis are observed in lung cancer patients. These changes may be an indication for adaptive radiotherapy in order to maintain target coverage and avoid increased risk of normal tissue complications. Material and methods: CBCT scans of 163 patients were reviewed to score lung changes and find the incidence, the impact of geometric and dosimetric changes and the timing of appearance and disappearance of changes. Results: 23% of the patients had changes in the lung related to pleural effusion, atelectasis or pneumonia/ pneumonitis. In 9% of all patients, the appearance or disappearance of a change introduced a shift of the tumor or lymph nodes relative to the spine >5 mm. Only major density changes affected the dose distribution, and 9% of all patients needed adaptive treatment planning due to density changes. In total, 12% of all patients did benefit from an adaptive treatment plan and in 85% of these patients, an atelectasis did change. Conclusions: An adaptive strategy was indicated for 12% of the patients due to atelectasis, pleural effusion or pneumonia/pneumonitis. The predominant cause for adaptation was atelectasis. No systematic pattern in the appearance and disappearance of the changes were observed and hence weekly evaluation is preferable. Ó 2013 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 110 (2014) 517–522
Local control is low in non-small cell lung cancer (NSCLC) as well as small cell lung cancer (SCLC) contributing to poor overall survival [1,2]. Improved local control may be achieved by radiotherapy dose escalation [3–6], even though after the RTOG0617 study it is debated if dose escalation leads to better overall survival [7]. Dose escalation is usually limited by increased treatment related mortality due to the damage to the surrounding normal lung tissue [3] and delivering the prescribed dose to the target without compromising the surrounding normal tissue is the aim of the future dose escalation studies [8]. Traditionally large treatment volumes were used to ensure dose coverage of the target volume during the radiotherapy (RT) course. Large margins were added to account for the different sources of systematic and random errors including intra-fraction breathing motion, interand intra-fraction baseline shift, organ motion and delineation [9]. Much effort has been put into minimizing these errors using 4D-CT [10] and daily setup on tumor [11]. Comparing the errors due to breathing motion and setup errors with anatomical changes during the course of RT showed that the latter had the largest impact on the delivered dose to the tumor
⇑ Corresponding author. Address: Aarhus University Hospital, Department of Medical Physics, Noerrebrogade 44, building 5J, 2. Floor, 8000 Aarhus C, Denmark. E-mail address: [email protected] (D.S. Møller). 0167-8140/$ - see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.radonc.2013.10.013
[12]. The large error that may occur when a patient experiences large changes in the lung tissue cannot be handled by margins, but requires a more individualized adaptive strategy [13]. For lung cancer three typical changes appear in the lung tissue: pleural effusion, atelectasis and pneumonia/pneumonitis. These changes impact the geometry of the target and normal tissue as well as the dosimetry. Adaptive radiotherapy (ART) aims at adjusting the treatment plan during the course of RT to ensure correct target coverage and to avoid normal tissue complications [14]. In the adaptive process, the changes are monitored systematically for the individual patients and when predetermined criteria are violated, an adaptive treatment plan is made. In this paper, we evaluate a large patient group undergoing daily cone beam computed tomography (CBCT) during RT treatment in order to find the incidence of lung density changes as well as the impact of these changes on the geometry of the target and the dose distribution. These numbers are essential when designing a manageable adaptive strategy in a large scale clinic. Material and methods Patient selection One hundred and sixty-three lung cancer patients were treated with thoracic RT between December 2009 and April 2012 at the
518
Adaptive radiotherapy of lung cancer
Department of Oncology at the Aarhus University Hospital, Denmark. All tumors were pathological proven (46 SCLC and 117 NSCLC). The diagnostic work-up included CT imaging of chest and abdomen. Patients underwent routinely 18FDG-PET/CT scan. Treatment and patient characteristics are shown in Table 1.
Treatment All patients underwent a free breathing CT scan with intravenous contrast prior to treatment planning. Twelve patients had a 3DCT scanning and 151 patients had a 4DCT scanning. The diagnostic PET/CT scan supported the target delineation. The gross tumor volume (GTV) including both the primary tumor and pathologically proven lymph nodes were delineated on the midventilation phase of the 4DCT or at the 3D scanning and expanded to a clinical target volume (CTV) by an isotropic 5 mm margin. The CTV was manually corrected for bones, vessels and the opposite lung. A planning target volume (PTV) was created by adding margins of 10 mm in the left–right and ventral–dorsal direction and 13 mm in the cranio–caudal direction. Treatment plans with three to eight 6MV beams were created in the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, CA, version 8.9 or 10.0). The plans were normalized to give a mean PTV dose of 100% of the prescribed dose. The PTV was covered with a homogenous dose of 50–66 Gy in 2 Gy fractions for the NCSLC patients and 45 Gy in 30 fractions or 50 Gy in 25 fractions for the SCLC patients. The dose was delivered using Varian Clinacs and daily CBCTs Table 1 Treatment and patient characteristics. Characteristic
n = 163
Median age
65 (range 35–83)
Gender Male Female
91 72
PS 0 1 2 Median GTV size (154 available) Median PTV size
62 87 14 59 (range 2–623) 465 (range 33–1516) SCLC n = 46 NSCLC = 117
T-stage Tx T1 T2 T3 T4
2 4 17 3 14
3 12 37 21 33
N-stage Nx N0 N1 N2 N3
1 4 2 26 13
2 21 3 62 18
0 45 1
3 108 6 11
0 40 6
8 59 50
M-stage Mx M0 M1 Recurrence Chemotherapy None Concurrent Sequential RT total dose/number of fractions 45 Gy/30 50 Gy/25 60 Gy/30 66 Gy/33
39 7
2 29 86
were acquired for all patients and used to correct the daily rigid setup errors of the bony anatomy in a region of interest that included the spine. CBCT evaluation and geometric effects For each patient the CBCTs of fraction 1, 6, 11, 16, 21, 26 and 31 were evaluated in Offline Review (Varian Medical Systems, version 10.0). Changes in the lung tissue due to pleural effusion, atelectasis or pneumonia/pneumonitis were scored for the selected CBCTs. Pleural effusion was recognized as a well defined, smooth, basal located density change. Only changes in pleural effusion P1 cm were included. Atelectasis was seen as well defined, often smooth and triangular density changes in relation to the tumor area, whereas pneumonia/pneumonitis was diffuse density changes in the lung tissue not specifically related to the tumor area. Tumor volume changes were not included in this study. The incidences of tumor volume changes have previously been discussed in the literature [17]. For patients with a visible change, all daily CBCTs were evaluated to find the time of appearance and disappearance of the lung density change. Furthermore, the fraction with the largest density change between planning CT (pCT) and CBCT was noted. On the CBCT with the largest deviation from CT, the extent of pleural effusion was measured as the largest distance ventral. For both pleural effusion and atelectasis, the change was delineated at the pCT and the CBCT with the largest deviation from the pCT. The volume change of the pleural effusion or atelectasis was measured. In cases where only part of the pleural effusion or atelectasis disappeared, the volumes and distances equated the size of the part that disappeared. If the lung density change induced a baseline shift of the lymph nodes and/or the primary tumor relative to the spine there was a risk of a geometric miss. In this study, we consider baseline shifts above 5 mm to result in a geometric effect on the target coverage and the number of patients with baseline shifts >5 mm was assessed. Dosimetric analysis When the radiation passes through water instead of lung tissue or vice versa, the dose distribution is altered depending on the size and the anatomical position of the lung density change. In the case of pneumonia/pneumonitis the changes were diffuse and had not water-like densities and they were omitted in the dosimetric analysis. The dosimetric effect of a density change was evaluated at the fraction with the largest deviation between the pCT and the daily CBCT. The pleural effusion or atelectasis was contoured on the CBCT scan. A rigid registration based on the bony anatomy was performed online to correct for daily setup errors at each treatment fraction. The same registration was used to propagate the delineated density change from the CBCT to the pCT scan. The structure was cropped to the borders of the lung on pCT and the Hounsfield Units (HU) of the change was set to either lung (HU = 738 corresponding to 0.26 g/cm3) or water (HU = 0) depending on whether the density change had disappeared or appeared on the CBCT relative to the pCT. The dose was recalculated at the pCT with alteration of HU in the delineated structure of the density change and with Monitor Units, field directions and MLC movements obtained from the pCT. The dose distribution obtained from the altered CT scan (denoted altered dose distribution) was compared with the dose distribution of the pCT scan (denoted planned dose distribution). In this study, the altered dose distribution was considered insufficient if: V95PTV (the volume of the PTV covered by 95% of the prescribed dose) decreases more than 3%.
D.S. Møller et al. / Radiotherapy and Oncology 110 (2014) 517–522
PTVmean (the mean dose to the PTV) decreases more than 1%. V107body (the volume receiving more than 107% of the prescribed dose) increases more than 5 cm3. A very fast assessment of the effect of density changes is preferable. Therefore, for all patients with pleural effusion or atelectasis, a quick estimate (denoted Estimate) of the dosimetric effect of the lung density change was made. The dosimetric effect depends on the actual treatment plan. For each field passing through the lung density change, the average thickness of the change was estimated. Moreover, the part of each field passing through the change was estimated. From this, the change in dose from each field was estimated and the total change in dose was compared to the dose calculated by the use of the pCT scan with altered HU. These estimates can be made in a few minutes by a medical physicist based on the treatment plan and the relevant CBCT and facilitates a fast selection of the patients, who were expected to benefit from ART.
Results The lung density changes observed were atelectasis, pleural effusion or pneumonia/pneumonitis. Thirty-seven patients had atelectasis, pleural effusion or pneumonia before treatment or a change evolving during the treatment (Fig. 1). This corresponds to 23% of the patients and the incidence of the three types of lung density changes is shown in the upper part of Fig. 1. Of these, 31 out of 117 (26%) NSCLC patients and 6 out of 46 (13%) SCLC patients experienced changes. Some patients experienced two types of changes. The most frequent lung density change was atelectasis (19 NSCLC + 5 SCLC patients, in total 15%) while pleural effusion (10 NSCLC + 3 SCLC patients, in total 8%) and pneumonia/pneumonitis visible on CBCT was more rare (8 NSCLC + 0 SCLC patients, in total 5%). Patients in whom the lung density change induced geometrical errors recognized as baseline shifts of the primary tumor and/or the malignant lymph nodes relative to the spine >5 mm were counted as visualized in the lower part of Fig. 1. Including all types of lung density changes, 9% (14 patients) experienced a baseline shift. The majority (12 of 14 patients) had an atelectasis causing the shift. One patient had a simultaneous large pleural effusion and small atelectasis, with an undefined boundary between them. The dosimetric effect for this patient was analyzed as being due to
Fig. 1. Diagram showing 37 out of 163 patients with lung density changes. In the upper part, 37 patients were divided into three types: atelectasis (red circle), pneumonia/pneumonitis (blue circle), and pleural effusion (green circle). Some patients have two different types of changes represented by a number in the overlap zone of the two circles, e.g., 5 patients have both an atelectasis and a pleural effusion. The lower part shows the results of the geometric and dosimetric evaluation of the 37 patients. The geometry- and the dosimetry-arrows lead to two Venn diagrams, dividing the changes with geometric or dosimetric impact into the three types. The final Venn diagram shows the total number of patients in the different categories requiring an adaptive treatment plan due to either dosimetry, geometry or both.
519
the pleural effusion solely, and hence only 23 patients were analyzed for the timing and the consequences of atelectasis. The remaining four patients, who had both a pleural effusion and an atelectasis were analyzed for the dosimetric effect of both the issues. Thirteen patients suffered from pleural effusion and Fig. 2a shows the time of appearance and disappearance. Four of 13 patients had a pleural effusion before treatment start. In the remainder of the patients, it appeared during the course of RT, most likely at the beginning of the course. When significant amounts of pleural effusion appeared during RT usually the patient underwent pleuracentesis to make sure that the fluid did not change the dose distribution during RT. Furthermore the fluid was examined for tumor cells. The extent of pleural effusion had a median value of 2.5 cm (range 1–5). Pleural effusion never caused a shift, except when an atelectasis was present simultaneously. Two patients had a stable effusion present at pCT and during the whole course of RT. For the remaining 11 patients, the deviation between the planned dose distribution and the altered dose distribution as calculated from the pCT scan is shown in Fig. 3 (column 1 and 2 for each patient). According to our dosimetric criteria, four patients would benefit from ART (patient number 5, 6, 10 and 12). In these four patients, a pleural effusion appeared during the treatment course and led to a decrease of the V95PTV of more than 3% (column 1) and PTVmean of more than 1% (column 2). Patients 3 and 4 had pleural effusion that disappeared during the course of RT and hence a slight increase of PTVmean as well as V95PTV was observed. The increment in dose did not lead to an increase in V107Body >5 cm3 (not shown). Hence these two patients were not considered for ART. The quick Estimate of the dosimetric effect due to pleural effusion, based on the measurable thickness of the density change, was a nice predictor for the necessity of ART. This is seen from a comparison of the calculated change in mean dose to the PTV based on delineation of the change (column 2) and the Estimate of the change in dose based on the thickness of the density change (column 3). In patients where the Estimate of the dosimetric effect was below 1%, the same was true for the deviation in PTVmean and hence, it is safe to use the Estimate to sort out the patients with changes that do not need ART. The extension of pleural effusion is measurable as a single figure and for patients with no more than 2 cm effusion the dosimetric change is always less than 1%. Twenty-three patients with atelectasis were analyzed and the time of appearance and disappearance is shown in Fig. 2b. Approximately half (13 of 23) of the atelectasis’ appeared before the pCT and the majority of these (10 of 13) dissolved during treatment. Ten atelectasis’ developed during treatment and the majority of these (6) did not dissolve during the course of RT. Half of the patients (12) experienced a baseline shift. The median of the absolute value of the change in volume of the atelectasis’ was 110 cm3 (range 0–314 cm3). There was no difference in the size of the atelectasis’ that caused a shift in the position of the tumor and lymph node (median 114 cm3) and the ones that did not (median 128 cm3). Only one atelectasis was outside the target area and thus, the distribution of atelectasis’ in the lung was therefore similar to the tumor positions. As seen from Fig. 2b, three patients (number 1–3) had an atelectasis present at pCT and during the whole treatment course. One patient (number 23) only had an atelectasis for the last five fractions of treatment. These four patients were excluded from the dosimetric analysis. For the remaining 19 patients, the altered dose distributions due to atelectasis resulted in deviations between the planned and the altered dose distribution (see Fig. 4). For patients 1–13, the atelectasis was present at the pCT and disappeared during the RT course. In these patients, the PTVmean (column 2) increased and the V95PTV slightly improved (mean 0.5% increment,
520
Adaptive radiotherapy of lung cancer
Fig. 2. Time of appearance and disappearance of lung tissue changes. For each patient the bar starts at the fraction, when the atelectasis appeared, while it ends at the fraction where it disappeared. Fraction 0 represents the pCT and the patients where the bar ends at fraction 33 had a lung tissue change that did not disappear during treatment. The patients have been sorted due to the appearance of the lung tissue change (a) pleural effusion. As an example for patient 8, pleural effusion was present from fraction 1 to fraction 7. Patient number 11 had a reoccurring pleural effusion (b) atelectasis.
Fig. 3. Deviation in coverage of the PTV (V95PTV) (gray) and PTVmean (black) between the planned and the altered dose distribution. An estimate of the dose deviation based on the thickness of the density change is shown in white. A limit of PTVmean deviation >1% as well as V95PTV >3% is shown as dotted lines. Deviations above these limits trigger adaptive treatment planning. Patient numbers are identical to those in Fig. 2a.
not shown). The more concerning issue in these patients was a median increasing of V107body (column 1) of 15 cm3 (range 0– 117 cm3). The overall maximum dose for all patients was 115%. Seven patients would benefit from ART (number 4, 5, 6, 9, 11, 12, 13) in order to avoid large hotspots. In the group where the atelectasis appeared during treatment (number 14–22), two patients needed an adaptive plan (21, 22) due to a decrease of V95PTV >3%. Additionally, in one of these patients (22) PTVmean decreased >1%. In total, 70% (17 of 24 patients) of the patients who got an atelectasis would benefit from ART, either due to geometric shifts and/ or dosimetric changes. In patients, where a geometric shift was observed, an adaptive treatment plan was triggered solely by this observation. For patients without a geometric shift, all cases with a dosimetric Estimate showing less than 1% deviation would not benefit from ART (data not shown).
Discussion To our knowledge, this is the first study evaluating a large number of lung cancer patients to find the prevalence of lung tissue
Fig. 4. Deviation between the planned and the altered dose distribution for the patients with atelectasis. For all patients the deviation in PTVmean (in %) is shown with the white bars. Patients are numbered as in Fig. 2b. For patients 4–13 the atelectasis was present at pCT and the increase in volume receiving >107% of the prescribed dose is shown with the gray bars and compared with a 5 cm3 criterion for adaptation. For patients 14–22 the atelectasis appeared during treatment and the decrease in V95PTV was shown with the black bars and compared with a 3% criterion for adaptation.
changes and the effect on the dose distribution. The study shows that 23% of the patients had pleural effusion, atelectasis and/or pneumonia/pneumonitis and that 12% could benefit from ART. This means that only approximately half of the patients with observed changes will actually need an adaptive treatment, diminishing the clinical workload substantially. Moreover, almost all (85%) adaptive treatment plans involved atelectasis, so care must always be taken when an atelectasis appears or disappears. Only 30% of the patients with atelectasis did not benefit from ART. Our data do not support that the decision to adapt the treatment plan can be based on position or size of the atelectasis. Instead, we propose evaluation of daily CBCT’s at least weekly. This evaluation could consist of three easy steps to rule out the majority of patients with changes not benefitting from ART as shown with the presented data. First, a geometric evaluation of shifts and deformations of the tumor and the treated lymph nodes showed if the deviation in target position was systematically above the limit set by the treatment margins. Second, a quick Estimate of the dosimetric error
D.S. Møller et al. / Radiotherapy and Oncology 110 (2014) 517–522
was made for pleural effusion and atelectasis (data not shown) based on the thickness of the effusion/atelectasis and the amount of Monitor Units passing the change. The Estimate was compared with the deviations found in the altered dose distributions. Only cases where the estimated dosimetric error was above 1% were found to benefit from ART. We find this a useful tool for a very quick assessment of the needs for ART due to dosimetric changes for the individual patient. Third, for the remaining patients, we propose to calculate the altered dose distribution based on the changes seen on CBCT in order to select the patients that actually do benefit from ART. Distinguishing between geometric and dosimetric errors is somewhat artificial as baseline shift changes the dose distribution, but the separation guides the adaptive strategy. The geometric errors are severe because they lead to underdosage of part of the target and therefore, systematic baseline shifts demand ART. The limit for acceptable geometrical errors depends on margins [9] and IGRT strategy. In this study, a limit of 5 mm was chosen, based on our clinical margins and IGRT strategy. The baseline shifts were measured based on a bone match using the spine. If the setup strategy is changed to a soft tissue match using the tumor, the number of patients with a geometric error will change. However, in most cases, an atelectasis deforms the mediastinum introducing large shifts of the primary tumor relative to the lymph nodes [16,17], resulting in underdosage of one or both targets. The method used to calculate the altered dose distributions introduces uncertainties. Contouring the lung density change on CBCT introduces a delineation uncertainty. The structure is propagated through a rigid registration and therefore, deformations of the tissue between pCT and CBCT are not included. Furthermore, using fixed Hounsfield Units (HUs) for pleural effusion, atelectasis and lung introduces an uncertainty. Altering the HUs for the largest pleural effusion and the largest atelectasis observed in this study with 100 HU introduces differences below 0.5%, typically much less (data not shown). In the case of pneumonia/pneumonitis the changes are more diffuse and hence the dosimetric analysis was not performed. It would obviously be better to recalculate the dose on the CBCT, but with the current quality of CBCT-scans in the thorax-region [15] and the limited length of the CBCT-scans not covering the entire region, this is not possible. Calculation of the altered dose distribution on a new CT-scan obtained when a change appeared would also have strengthened the data, but this was impossible in this retrospective study. Using the CBCT to select patients profiting from an adaptive strategy using the methods outlined in the present study is feasible and leads to a manageable number of patients selected for adaptation. The incidence of adaptive treatment planning depends very much on the dosimetric criteria chosen to trigger adaptation. In this work, we have chosen rather strict criteria, because the under dosage implied by atelectasis and pleural effusion not only implies underdosage of the PTV boarders but also of the GTV. ART improves the delivered dose to the target, and we expect a positive influence on the low local control in lung cancer [1,2], but in order to set correct criteria for adaptation, we need clinical studies to evaluate how the underdosage influences local control. Furthermore, an adaptive strategy makes it safer to reduce CTV–PTV margins, because patients with large systematic geometric shifts will receive an adaptive treatment plan. When water like tissue disappears during the course of RT, the dose distribution changes and large volumes receive more than 107% of the dose in 6 of 10 patients (see Fig. 4). This potentially increases the risk of complications in e.g. the esophagus [22,23] close to the target. Escalating the dose for NSCLC [3–7] increases this risk and calls for a careful adaptive strategy. The criteria in this paper for the need of ART are not based on clinical data, but are a com-
521
promise between feasibility and safety, to avoid local failures and normal tissue complications. Atelectasis is by far the most severe of the lung tissue changes. The study of appearance or disappearance shows that there is no preferential time of replanning. It is often (17 of 24 patients, 70%) relevant to adapt the treatment plan and adaptation needs to be based on weekly evaluations, starting with an evaluation after the first RT treatment, as the atelectasis disappears/appears between pCT and first fraction in 22% of the patients (5 patients). Presence of an atelectasis at pCT often introduces uncertainty in target delineation [18], because it is located adjacent to the GTV. In our study, only one patient had an atelectasis outside the target area, thus more precise target definition after disappearance of the atelectasis may be an indication for replanning. Several studies have shown that atelectasis often dissolves by RT treatment [18,19]. One large study of SCLC found 22% incidence of atelectasis at treatment start [20], which is much more than we found (3 of 46 patients, 7%, with SCLC have an atelectasis at treatment start). This is due to a different patient selection. In the present study, almost all SCLC patients had limited disease and all of them underwent at least one series of chemotherapy before RT. Vaaler et al. refer to a diagnostic CT before treatment and most of the patients had more advanced disease. Patients were sent to pleuracentesis for symptom relief as well as for diagnostic purposes when large amounts of pleural effusion appeared during the RT course. After pleuracentesis, the fluid may disappear suddenly. Smaller amounts of pleural effusion did not require intervention and sometimes diuretics were used, which did not change the status dramatically. Radiological changes evaluated on the CBCT scan were defined as pneumonia/pneumonitis when diffuse density changes were seen in the lung tissue. Pneumonitis and pneumonia cannot be distinguished based on CBCT. In this cohort of patients, CBCT has revealed major changes in only 2 patients. A radiation oncologist always followed up the development of a pneumonia/pneumonitis and the patients were examined clinically and treated accordingly. Radiological changes attributed to pneumonia/pneumonitis seen in or out of radiation fields on the first diagnostic post RT CT scan can predict the patient outcome [21]. In summary, 23% of the 163 patients evaluated, experience atelectasis, pleural effusion or pneumonia/pneumonitis and 12% of all patients would benefit from an adaptive strategy to avoid inadequate dosage to the target or overdosage of normal tissue. Atelectasis should be evaluated carefully as an adaptive strategy is required in 70% of the cases.
Conflicts of interest None to declare. Acknowledgement None. References [1] Machtay M, Paulus R, Moughan J, et al. Defining local–regional control and its importance in locally advanced non-small cell lung carcinoma. J Thorac Oncol 2012;7:716–22. [2] Larsen AW, Khalil AA, Meldgaard P, Knap MM. 18FDG-PET/CT is a useful tool in staging procedure before chemo-radiotherapy in patients with limited disease small-cell lung cancer. Pattern of failure and survival is analyzed. J Cancer Therapy 2012;3:372–8. [3] Bradley JD, Moughan J, Graham MV, et al. A phase I/II radiation dose escalation study with concurrent chemotherapy for patients with inoperable stages I to III non-small-cell lung cancer: phase I results of RTOG 0117. Int J Radiat Oncol Biol Phys 2010;77:367–72.
522
Adaptive radiotherapy of lung cancer
[4] van Baardwijk A, Wanders S, Boersma L, et al. Mature results of an individualized radiation dose prescription study based on normal tissue constraints in stages I to III non-small-cell lung cancer. J Clin Oncol 2010;28:1380–6. [5] Kong FM, Haken RKT, Schipper MJ, et al. High-dose radiation improved local tumor control and overall survival in patients with inoperable/unresectable non-small-cell lung cancer: long-term results of a radiation dose escalation study. Int J Radiat Oncol Biol Phys 2005;63:324–33. [6] Thirion P, Holmberg O, Collins CD, et al. Escalated dose for non-small-cell lung cancer with accelerated hypofractionated three-dimensional conformal radiation therapy. Radiother Oncol 2004;71:163–6. [7] Cox JD. Are the results of RTOG0617 mysterious? Int J Radiat Oncol Biol Phys 2012;82:1042–4. [8] van Elmpt W, Ruysscher DD, van der Salm A, et al. The PET-boost randomised phase II dose-escalation trial in non-small cell lung cancer. Radiother Oncol 2012;104:67–71. [9] van Herk M, Remeijer P, Rasch C, Lebesque JV. The probability of correct target dosage: dose-population histograms for deriving treatment margins in radiotherapy. Int J Radiat Oncol Biol Phys 2000;47:1121–35. [10] Wolthaus JWH, Sonke JJ, van Herk M, et al. Comparison of different strategies to use four-dimensional computed tomography in treatment planning for lung cancer patients. Int J Radiat Oncol Biol Phys 2008;70:1229–38. [11] Yeung AR, Li JG, Shi W, et al. Tumor localization using cone-beam CT reduces setup margins in conventionally fractionated radiotherapy for lung tumors. Int J Radiat Oncol Biol Phys 2009;74:1100–7. [12] Schmidt ML, Hoffmann L, Kandi M, Møller DS, Poulsen PR. Dosimetric impact of respiratory motion, interfraction baseline shifts, and anatomical changes in radiotherapy of non-small-cell lung cancer. Acta Oncol 2013;52:1490–6. [13] Britton KR, Starkschall G, Liu H, et al. Consequences of anatomic changes and respiratory motion on radiation dose distributions in conformal radiotherapy for locally advanced non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2009;73:94–102.
[14] Yan D, Vicini F, Wong J, Martinez A. Adaptive radiation therapy. Phys Med Biol 2007;42:123–32. [15] Rong Y, Smilowitz J, Tewatia D, Tomé WA, Paliwal B. Dose calculation on kV cone beam CT images: an investigation of the Hu-density conversion stability and dose accuracy using the site-specific calibration. Med Dosim 2010;35:195–207. [16] Roman NO, Shepherd W, Mukhopadhyay N, Hugo GD, Weiss E. Interfractional positional variability of fiducial markers and primary tumors in locally advanced non-small-cell lung cancer during audiovisual biofeedback radiotherapy. Int J Radiat Oncol Biol Phys 2012;83:1566–72. [17] Knap MM, Hoffmann L, Nordsmark M, Vestergaard A. Daily cone-beam computed tomography used to determine tumour shrinkage and localisation in lung cancer patients. Acta Oncol 2010;49:1077–84. [18] Lavrenkov K, Partridge M, Cook G, Brada M. Positron emission tomography for target volume definition in the treatment of non-small cell lung cancer. Radiother Oncol 2005;77:1–4. [19] Chetty KG, Moran EM, Sassoon CS, Viravathana T, Light RW. Effect of radiation therapy on bronchial obstruction due to bronchogenic carcinoma. Chest 1989;95:582–4. [20] Vaaler AK, Forrester JM, Lesar M, Edison M, Venzon D, Johnson BE. Obstructive atelectasis in patients with small cell lung cancer. Incidence and response to treatment. Chest 1997;1997:115–20. [21] Farr KP, Khalil AA, Knap MM, Møller DS, Grau C. Development of radiation pneumopathy and generalised radiological changes after radiotherapy are independent negative prognostic factors for survival in non-small cell lung cancer patients. Radiother Oncol 2013;107:382–8. [22] Uyterlinde W, Chen C, Kwint M, et al. Prognostic parameters for acute esophagus toxicity in intensity modulated radiotherapy and concurrent chemotherapy for locally advanced non-small cell lung cancer. Radiother Oncol 2013;107:392–7. [23] Nijkamp J, Rossi M, Lebesque J, et al. Relating acute esophagitis to radiotherapy dose using FDG-PET in concurrent chemo-radiotherapy for locally advanced non-small cell lung cancer. Radiother Oncol 2013;106:118–23.
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Lung cancer RT
Adaptive radiotherapy of lung cancer patients with pleural effusion or atelectasis Ditte Sloth Møller a,⇑, Azza Ahmed Khalil b, Marianne Marquard Knap b, Lone Hoffmann a a
Department of Medical Physics; and b Department of Oncology, Aarhus University Hospital, Denmark
a r t i c l e
i n f o
Article history: Received 31 May 2013 Received in revised form 20 September 2013 Accepted 3 October 2013 Available online 31 October 2013 Keywords: Adaptive RT IGRT Atelectasis NSCLC SCLC
a b s t r a c t Background and purpose: Changes in lung density due to atelectasis, pleural effusion and pneumonia/ pneumonitis are observed in lung cancer patients. These changes may be an indication for adaptive radiotherapy in order to maintain target coverage and avoid increased risk of normal tissue complications. Material and methods: CBCT scans of 163 patients were reviewed to score lung changes and find the incidence, the impact of geometric and dosimetric changes and the timing of appearance and disappearance of changes. Results: 23% of the patients had changes in the lung related to pleural effusion, atelectasis or pneumonia/ pneumonitis. In 9% of all patients, the appearance or disappearance of a change introduced a shift of the tumor or lymph nodes relative to the spine >5 mm. Only major density changes affected the dose distribution, and 9% of all patients needed adaptive treatment planning due to density changes. In total, 12% of all patients did benefit from an adaptive treatment plan and in 85% of these patients, an atelectasis did change. Conclusions: An adaptive strategy was indicated for 12% of the patients due to atelectasis, pleural effusion or pneumonia/pneumonitis. The predominant cause for adaptation was atelectasis. No systematic pattern in the appearance and disappearance of the changes were observed and hence weekly evaluation is preferable. Ó 2013 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 110 (2014) 517–522
Local control is low in non-small cell lung cancer (NSCLC) as well as small cell lung cancer (SCLC) contributing to poor overall survival [1,2]. Improved local control may be achieved by radiotherapy dose escalation [3–6], even though after the RTOG0617 study it is debated if dose escalation leads to better overall survival [7]. Dose escalation is usually limited by increased treatment related mortality due to the damage to the surrounding normal lung tissue [3] and delivering the prescribed dose to the target without compromising the surrounding normal tissue is the aim of the future dose escalation studies [8]. Traditionally large treatment volumes were used to ensure dose coverage of the target volume during the radiotherapy (RT) course. Large margins were added to account for the different sources of systematic and random errors including intra-fraction breathing motion, interand intra-fraction baseline shift, organ motion and delineation [9]. Much effort has been put into minimizing these errors using 4D-CT [10] and daily setup on tumor [11]. Comparing the errors due to breathing motion and setup errors with anatomical changes during the course of RT showed that the latter had the largest impact on the delivered dose to the tumor
⇑ Corresponding author. Address: Aarhus University Hospital, Department of Medical Physics, Noerrebrogade 44, building 5J, 2. Floor, 8000 Aarhus C, Denmark. E-mail address: [email protected] (D.S. Møller). 0167-8140/$ - see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.radonc.2013.10.013
[12]. The large error that may occur when a patient experiences large changes in the lung tissue cannot be handled by margins, but requires a more individualized adaptive strategy [13]. For lung cancer three typical changes appear in the lung tissue: pleural effusion, atelectasis and pneumonia/pneumonitis. These changes impact the geometry of the target and normal tissue as well as the dosimetry. Adaptive radiotherapy (ART) aims at adjusting the treatment plan during the course of RT to ensure correct target coverage and to avoid normal tissue complications [14]. In the adaptive process, the changes are monitored systematically for the individual patients and when predetermined criteria are violated, an adaptive treatment plan is made. In this paper, we evaluate a large patient group undergoing daily cone beam computed tomography (CBCT) during RT treatment in order to find the incidence of lung density changes as well as the impact of these changes on the geometry of the target and the dose distribution. These numbers are essential when designing a manageable adaptive strategy in a large scale clinic. Material and methods Patient selection One hundred and sixty-three lung cancer patients were treated with thoracic RT between December 2009 and April 2012 at the
518
Adaptive radiotherapy of lung cancer
Department of Oncology at the Aarhus University Hospital, Denmark. All tumors were pathological proven (46 SCLC and 117 NSCLC). The diagnostic work-up included CT imaging of chest and abdomen. Patients underwent routinely 18FDG-PET/CT scan. Treatment and patient characteristics are shown in Table 1.
Treatment All patients underwent a free breathing CT scan with intravenous contrast prior to treatment planning. Twelve patients had a 3DCT scanning and 151 patients had a 4DCT scanning. The diagnostic PET/CT scan supported the target delineation. The gross tumor volume (GTV) including both the primary tumor and pathologically proven lymph nodes were delineated on the midventilation phase of the 4DCT or at the 3D scanning and expanded to a clinical target volume (CTV) by an isotropic 5 mm margin. The CTV was manually corrected for bones, vessels and the opposite lung. A planning target volume (PTV) was created by adding margins of 10 mm in the left–right and ventral–dorsal direction and 13 mm in the cranio–caudal direction. Treatment plans with three to eight 6MV beams were created in the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, CA, version 8.9 or 10.0). The plans were normalized to give a mean PTV dose of 100% of the prescribed dose. The PTV was covered with a homogenous dose of 50–66 Gy in 2 Gy fractions for the NCSLC patients and 45 Gy in 30 fractions or 50 Gy in 25 fractions for the SCLC patients. The dose was delivered using Varian Clinacs and daily CBCTs Table 1 Treatment and patient characteristics. Characteristic
n = 163
Median age
65 (range 35–83)
Gender Male Female
91 72
PS 0 1 2 Median GTV size (154 available) Median PTV size
62 87 14 59 (range 2–623) 465 (range 33–1516) SCLC n = 46 NSCLC = 117
T-stage Tx T1 T2 T3 T4
2 4 17 3 14
3 12 37 21 33
N-stage Nx N0 N1 N2 N3
1 4 2 26 13
2 21 3 62 18
0 45 1
3 108 6 11
0 40 6
8 59 50
M-stage Mx M0 M1 Recurrence Chemotherapy None Concurrent Sequential RT total dose/number of fractions 45 Gy/30 50 Gy/25 60 Gy/30 66 Gy/33
39 7
2 29 86
were acquired for all patients and used to correct the daily rigid setup errors of the bony anatomy in a region of interest that included the spine. CBCT evaluation and geometric effects For each patient the CBCTs of fraction 1, 6, 11, 16, 21, 26 and 31 were evaluated in Offline Review (Varian Medical Systems, version 10.0). Changes in the lung tissue due to pleural effusion, atelectasis or pneumonia/pneumonitis were scored for the selected CBCTs. Pleural effusion was recognized as a well defined, smooth, basal located density change. Only changes in pleural effusion P1 cm were included. Atelectasis was seen as well defined, often smooth and triangular density changes in relation to the tumor area, whereas pneumonia/pneumonitis was diffuse density changes in the lung tissue not specifically related to the tumor area. Tumor volume changes were not included in this study. The incidences of tumor volume changes have previously been discussed in the literature [17]. For patients with a visible change, all daily CBCTs were evaluated to find the time of appearance and disappearance of the lung density change. Furthermore, the fraction with the largest density change between planning CT (pCT) and CBCT was noted. On the CBCT with the largest deviation from CT, the extent of pleural effusion was measured as the largest distance ventral. For both pleural effusion and atelectasis, the change was delineated at the pCT and the CBCT with the largest deviation from the pCT. The volume change of the pleural effusion or atelectasis was measured. In cases where only part of the pleural effusion or atelectasis disappeared, the volumes and distances equated the size of the part that disappeared. If the lung density change induced a baseline shift of the lymph nodes and/or the primary tumor relative to the spine there was a risk of a geometric miss. In this study, we consider baseline shifts above 5 mm to result in a geometric effect on the target coverage and the number of patients with baseline shifts >5 mm was assessed. Dosimetric analysis When the radiation passes through water instead of lung tissue or vice versa, the dose distribution is altered depending on the size and the anatomical position of the lung density change. In the case of pneumonia/pneumonitis the changes were diffuse and had not water-like densities and they were omitted in the dosimetric analysis. The dosimetric effect of a density change was evaluated at the fraction with the largest deviation between the pCT and the daily CBCT. The pleural effusion or atelectasis was contoured on the CBCT scan. A rigid registration based on the bony anatomy was performed online to correct for daily setup errors at each treatment fraction. The same registration was used to propagate the delineated density change from the CBCT to the pCT scan. The structure was cropped to the borders of the lung on pCT and the Hounsfield Units (HU) of the change was set to either lung (HU = 738 corresponding to 0.26 g/cm3) or water (HU = 0) depending on whether the density change had disappeared or appeared on the CBCT relative to the pCT. The dose was recalculated at the pCT with alteration of HU in the delineated structure of the density change and with Monitor Units, field directions and MLC movements obtained from the pCT. The dose distribution obtained from the altered CT scan (denoted altered dose distribution) was compared with the dose distribution of the pCT scan (denoted planned dose distribution). In this study, the altered dose distribution was considered insufficient if: V95PTV (the volume of the PTV covered by 95% of the prescribed dose) decreases more than 3%.
D.S. Møller et al. / Radiotherapy and Oncology 110 (2014) 517–522
PTVmean (the mean dose to the PTV) decreases more than 1%. V107body (the volume receiving more than 107% of the prescribed dose) increases more than 5 cm3. A very fast assessment of the effect of density changes is preferable. Therefore, for all patients with pleural effusion or atelectasis, a quick estimate (denoted Estimate) of the dosimetric effect of the lung density change was made. The dosimetric effect depends on the actual treatment plan. For each field passing through the lung density change, the average thickness of the change was estimated. Moreover, the part of each field passing through the change was estimated. From this, the change in dose from each field was estimated and the total change in dose was compared to the dose calculated by the use of the pCT scan with altered HU. These estimates can be made in a few minutes by a medical physicist based on the treatment plan and the relevant CBCT and facilitates a fast selection of the patients, who were expected to benefit from ART.
Results The lung density changes observed were atelectasis, pleural effusion or pneumonia/pneumonitis. Thirty-seven patients had atelectasis, pleural effusion or pneumonia before treatment or a change evolving during the treatment (Fig. 1). This corresponds to 23% of the patients and the incidence of the three types of lung density changes is shown in the upper part of Fig. 1. Of these, 31 out of 117 (26%) NSCLC patients and 6 out of 46 (13%) SCLC patients experienced changes. Some patients experienced two types of changes. The most frequent lung density change was atelectasis (19 NSCLC + 5 SCLC patients, in total 15%) while pleural effusion (10 NSCLC + 3 SCLC patients, in total 8%) and pneumonia/pneumonitis visible on CBCT was more rare (8 NSCLC + 0 SCLC patients, in total 5%). Patients in whom the lung density change induced geometrical errors recognized as baseline shifts of the primary tumor and/or the malignant lymph nodes relative to the spine >5 mm were counted as visualized in the lower part of Fig. 1. Including all types of lung density changes, 9% (14 patients) experienced a baseline shift. The majority (12 of 14 patients) had an atelectasis causing the shift. One patient had a simultaneous large pleural effusion and small atelectasis, with an undefined boundary between them. The dosimetric effect for this patient was analyzed as being due to
Fig. 1. Diagram showing 37 out of 163 patients with lung density changes. In the upper part, 37 patients were divided into three types: atelectasis (red circle), pneumonia/pneumonitis (blue circle), and pleural effusion (green circle). Some patients have two different types of changes represented by a number in the overlap zone of the two circles, e.g., 5 patients have both an atelectasis and a pleural effusion. The lower part shows the results of the geometric and dosimetric evaluation of the 37 patients. The geometry- and the dosimetry-arrows lead to two Venn diagrams, dividing the changes with geometric or dosimetric impact into the three types. The final Venn diagram shows the total number of patients in the different categories requiring an adaptive treatment plan due to either dosimetry, geometry or both.
519
the pleural effusion solely, and hence only 23 patients were analyzed for the timing and the consequences of atelectasis. The remaining four patients, who had both a pleural effusion and an atelectasis were analyzed for the dosimetric effect of both the issues. Thirteen patients suffered from pleural effusion and Fig. 2a shows the time of appearance and disappearance. Four of 13 patients had a pleural effusion before treatment start. In the remainder of the patients, it appeared during the course of RT, most likely at the beginning of the course. When significant amounts of pleural effusion appeared during RT usually the patient underwent pleuracentesis to make sure that the fluid did not change the dose distribution during RT. Furthermore the fluid was examined for tumor cells. The extent of pleural effusion had a median value of 2.5 cm (range 1–5). Pleural effusion never caused a shift, except when an atelectasis was present simultaneously. Two patients had a stable effusion present at pCT and during the whole course of RT. For the remaining 11 patients, the deviation between the planned dose distribution and the altered dose distribution as calculated from the pCT scan is shown in Fig. 3 (column 1 and 2 for each patient). According to our dosimetric criteria, four patients would benefit from ART (patient number 5, 6, 10 and 12). In these four patients, a pleural effusion appeared during the treatment course and led to a decrease of the V95PTV of more than 3% (column 1) and PTVmean of more than 1% (column 2). Patients 3 and 4 had pleural effusion that disappeared during the course of RT and hence a slight increase of PTVmean as well as V95PTV was observed. The increment in dose did not lead to an increase in V107Body >5 cm3 (not shown). Hence these two patients were not considered for ART. The quick Estimate of the dosimetric effect due to pleural effusion, based on the measurable thickness of the density change, was a nice predictor for the necessity of ART. This is seen from a comparison of the calculated change in mean dose to the PTV based on delineation of the change (column 2) and the Estimate of the change in dose based on the thickness of the density change (column 3). In patients where the Estimate of the dosimetric effect was below 1%, the same was true for the deviation in PTVmean and hence, it is safe to use the Estimate to sort out the patients with changes that do not need ART. The extension of pleural effusion is measurable as a single figure and for patients with no more than 2 cm effusion the dosimetric change is always less than 1%. Twenty-three patients with atelectasis were analyzed and the time of appearance and disappearance is shown in Fig. 2b. Approximately half (13 of 23) of the atelectasis’ appeared before the pCT and the majority of these (10 of 13) dissolved during treatment. Ten atelectasis’ developed during treatment and the majority of these (6) did not dissolve during the course of RT. Half of the patients (12) experienced a baseline shift. The median of the absolute value of the change in volume of the atelectasis’ was 110 cm3 (range 0–314 cm3). There was no difference in the size of the atelectasis’ that caused a shift in the position of the tumor and lymph node (median 114 cm3) and the ones that did not (median 128 cm3). Only one atelectasis was outside the target area and thus, the distribution of atelectasis’ in the lung was therefore similar to the tumor positions. As seen from Fig. 2b, three patients (number 1–3) had an atelectasis present at pCT and during the whole treatment course. One patient (number 23) only had an atelectasis for the last five fractions of treatment. These four patients were excluded from the dosimetric analysis. For the remaining 19 patients, the altered dose distributions due to atelectasis resulted in deviations between the planned and the altered dose distribution (see Fig. 4). For patients 1–13, the atelectasis was present at the pCT and disappeared during the RT course. In these patients, the PTVmean (column 2) increased and the V95PTV slightly improved (mean 0.5% increment,
520
Adaptive radiotherapy of lung cancer
Fig. 2. Time of appearance and disappearance of lung tissue changes. For each patient the bar starts at the fraction, when the atelectasis appeared, while it ends at the fraction where it disappeared. Fraction 0 represents the pCT and the patients where the bar ends at fraction 33 had a lung tissue change that did not disappear during treatment. The patients have been sorted due to the appearance of the lung tissue change (a) pleural effusion. As an example for patient 8, pleural effusion was present from fraction 1 to fraction 7. Patient number 11 had a reoccurring pleural effusion (b) atelectasis.
Fig. 3. Deviation in coverage of the PTV (V95PTV) (gray) and PTVmean (black) between the planned and the altered dose distribution. An estimate of the dose deviation based on the thickness of the density change is shown in white. A limit of PTVmean deviation >1% as well as V95PTV >3% is shown as dotted lines. Deviations above these limits trigger adaptive treatment planning. Patient numbers are identical to those in Fig. 2a.
not shown). The more concerning issue in these patients was a median increasing of V107body (column 1) of 15 cm3 (range 0– 117 cm3). The overall maximum dose for all patients was 115%. Seven patients would benefit from ART (number 4, 5, 6, 9, 11, 12, 13) in order to avoid large hotspots. In the group where the atelectasis appeared during treatment (number 14–22), two patients needed an adaptive plan (21, 22) due to a decrease of V95PTV >3%. Additionally, in one of these patients (22) PTVmean decreased >1%. In total, 70% (17 of 24 patients) of the patients who got an atelectasis would benefit from ART, either due to geometric shifts and/ or dosimetric changes. In patients, where a geometric shift was observed, an adaptive treatment plan was triggered solely by this observation. For patients without a geometric shift, all cases with a dosimetric Estimate showing less than 1% deviation would not benefit from ART (data not shown).
Discussion To our knowledge, this is the first study evaluating a large number of lung cancer patients to find the prevalence of lung tissue
Fig. 4. Deviation between the planned and the altered dose distribution for the patients with atelectasis. For all patients the deviation in PTVmean (in %) is shown with the white bars. Patients are numbered as in Fig. 2b. For patients 4–13 the atelectasis was present at pCT and the increase in volume receiving >107% of the prescribed dose is shown with the gray bars and compared with a 5 cm3 criterion for adaptation. For patients 14–22 the atelectasis appeared during treatment and the decrease in V95PTV was shown with the black bars and compared with a 3% criterion for adaptation.
changes and the effect on the dose distribution. The study shows that 23% of the patients had pleural effusion, atelectasis and/or pneumonia/pneumonitis and that 12% could benefit from ART. This means that only approximately half of the patients with observed changes will actually need an adaptive treatment, diminishing the clinical workload substantially. Moreover, almost all (85%) adaptive treatment plans involved atelectasis, so care must always be taken when an atelectasis appears or disappears. Only 30% of the patients with atelectasis did not benefit from ART. Our data do not support that the decision to adapt the treatment plan can be based on position or size of the atelectasis. Instead, we propose evaluation of daily CBCT’s at least weekly. This evaluation could consist of three easy steps to rule out the majority of patients with changes not benefitting from ART as shown with the presented data. First, a geometric evaluation of shifts and deformations of the tumor and the treated lymph nodes showed if the deviation in target position was systematically above the limit set by the treatment margins. Second, a quick Estimate of the dosimetric error
D.S. Møller et al. / Radiotherapy and Oncology 110 (2014) 517–522
was made for pleural effusion and atelectasis (data not shown) based on the thickness of the effusion/atelectasis and the amount of Monitor Units passing the change. The Estimate was compared with the deviations found in the altered dose distributions. Only cases where the estimated dosimetric error was above 1% were found to benefit from ART. We find this a useful tool for a very quick assessment of the needs for ART due to dosimetric changes for the individual patient. Third, for the remaining patients, we propose to calculate the altered dose distribution based on the changes seen on CBCT in order to select the patients that actually do benefit from ART. Distinguishing between geometric and dosimetric errors is somewhat artificial as baseline shift changes the dose distribution, but the separation guides the adaptive strategy. The geometric errors are severe because they lead to underdosage of part of the target and therefore, systematic baseline shifts demand ART. The limit for acceptable geometrical errors depends on margins [9] and IGRT strategy. In this study, a limit of 5 mm was chosen, based on our clinical margins and IGRT strategy. The baseline shifts were measured based on a bone match using the spine. If the setup strategy is changed to a soft tissue match using the tumor, the number of patients with a geometric error will change. However, in most cases, an atelectasis deforms the mediastinum introducing large shifts of the primary tumor relative to the lymph nodes [16,17], resulting in underdosage of one or both targets. The method used to calculate the altered dose distributions introduces uncertainties. Contouring the lung density change on CBCT introduces a delineation uncertainty. The structure is propagated through a rigid registration and therefore, deformations of the tissue between pCT and CBCT are not included. Furthermore, using fixed Hounsfield Units (HUs) for pleural effusion, atelectasis and lung introduces an uncertainty. Altering the HUs for the largest pleural effusion and the largest atelectasis observed in this study with 100 HU introduces differences below 0.5%, typically much less (data not shown). In the case of pneumonia/pneumonitis the changes are more diffuse and hence the dosimetric analysis was not performed. It would obviously be better to recalculate the dose on the CBCT, but with the current quality of CBCT-scans in the thorax-region [15] and the limited length of the CBCT-scans not covering the entire region, this is not possible. Calculation of the altered dose distribution on a new CT-scan obtained when a change appeared would also have strengthened the data, but this was impossible in this retrospective study. Using the CBCT to select patients profiting from an adaptive strategy using the methods outlined in the present study is feasible and leads to a manageable number of patients selected for adaptation. The incidence of adaptive treatment planning depends very much on the dosimetric criteria chosen to trigger adaptation. In this work, we have chosen rather strict criteria, because the under dosage implied by atelectasis and pleural effusion not only implies underdosage of the PTV boarders but also of the GTV. ART improves the delivered dose to the target, and we expect a positive influence on the low local control in lung cancer [1,2], but in order to set correct criteria for adaptation, we need clinical studies to evaluate how the underdosage influences local control. Furthermore, an adaptive strategy makes it safer to reduce CTV–PTV margins, because patients with large systematic geometric shifts will receive an adaptive treatment plan. When water like tissue disappears during the course of RT, the dose distribution changes and large volumes receive more than 107% of the dose in 6 of 10 patients (see Fig. 4). This potentially increases the risk of complications in e.g. the esophagus [22,23] close to the target. Escalating the dose for NSCLC [3–7] increases this risk and calls for a careful adaptive strategy. The criteria in this paper for the need of ART are not based on clinical data, but are a com-
521
promise between feasibility and safety, to avoid local failures and normal tissue complications. Atelectasis is by far the most severe of the lung tissue changes. The study of appearance or disappearance shows that there is no preferential time of replanning. It is often (17 of 24 patients, 70%) relevant to adapt the treatment plan and adaptation needs to be based on weekly evaluations, starting with an evaluation after the first RT treatment, as the atelectasis disappears/appears between pCT and first fraction in 22% of the patients (5 patients). Presence of an atelectasis at pCT often introduces uncertainty in target delineation [18], because it is located adjacent to the GTV. In our study, only one patient had an atelectasis outside the target area, thus more precise target definition after disappearance of the atelectasis may be an indication for replanning. Several studies have shown that atelectasis often dissolves by RT treatment [18,19]. One large study of SCLC found 22% incidence of atelectasis at treatment start [20], which is much more than we found (3 of 46 patients, 7%, with SCLC have an atelectasis at treatment start). This is due to a different patient selection. In the present study, almost all SCLC patients had limited disease and all of them underwent at least one series of chemotherapy before RT. Vaaler et al. refer to a diagnostic CT before treatment and most of the patients had more advanced disease. Patients were sent to pleuracentesis for symptom relief as well as for diagnostic purposes when large amounts of pleural effusion appeared during the RT course. After pleuracentesis, the fluid may disappear suddenly. Smaller amounts of pleural effusion did not require intervention and sometimes diuretics were used, which did not change the status dramatically. Radiological changes evaluated on the CBCT scan were defined as pneumonia/pneumonitis when diffuse density changes were seen in the lung tissue. Pneumonitis and pneumonia cannot be distinguished based on CBCT. In this cohort of patients, CBCT has revealed major changes in only 2 patients. A radiation oncologist always followed up the development of a pneumonia/pneumonitis and the patients were examined clinically and treated accordingly. Radiological changes attributed to pneumonia/pneumonitis seen in or out of radiation fields on the first diagnostic post RT CT scan can predict the patient outcome [21]. In summary, 23% of the 163 patients evaluated, experience atelectasis, pleural effusion or pneumonia/pneumonitis and 12% of all patients would benefit from an adaptive strategy to avoid inadequate dosage to the target or overdosage of normal tissue. Atelectasis should be evaluated carefully as an adaptive strategy is required in 70% of the cases.
Conflicts of interest None to declare. Acknowledgement None. References [1] Machtay M, Paulus R, Moughan J, et al. Defining local–regional control and its importance in locally advanced non-small cell lung carcinoma. J Thorac Oncol 2012;7:716–22. [2] Larsen AW, Khalil AA, Meldgaard P, Knap MM. 18FDG-PET/CT is a useful tool in staging procedure before chemo-radiotherapy in patients with limited disease small-cell lung cancer. Pattern of failure and survival is analyzed. J Cancer Therapy 2012;3:372–8. [3] Bradley JD, Moughan J, Graham MV, et al. A phase I/II radiation dose escalation study with concurrent chemotherapy for patients with inoperable stages I to III non-small-cell lung cancer: phase I results of RTOG 0117. Int J Radiat Oncol Biol Phys 2010;77:367–72.
522
Adaptive radiotherapy of lung cancer
[4] van Baardwijk A, Wanders S, Boersma L, et al. Mature results of an individualized radiation dose prescription study based on normal tissue constraints in stages I to III non-small-cell lung cancer. J Clin Oncol 2010;28:1380–6. [5] Kong FM, Haken RKT, Schipper MJ, et al. High-dose radiation improved local tumor control and overall survival in patients with inoperable/unresectable non-small-cell lung cancer: long-term results of a radiation dose escalation study. Int J Radiat Oncol Biol Phys 2005;63:324–33. [6] Thirion P, Holmberg O, Collins CD, et al. Escalated dose for non-small-cell lung cancer with accelerated hypofractionated three-dimensional conformal radiation therapy. Radiother Oncol 2004;71:163–6. [7] Cox JD. Are the results of RTOG0617 mysterious? Int J Radiat Oncol Biol Phys 2012;82:1042–4. [8] van Elmpt W, Ruysscher DD, van der Salm A, et al. The PET-boost randomised phase II dose-escalation trial in non-small cell lung cancer. Radiother Oncol 2012;104:67–71. [9] van Herk M, Remeijer P, Rasch C, Lebesque JV. The probability of correct target dosage: dose-population histograms for deriving treatment margins in radiotherapy. Int J Radiat Oncol Biol Phys 2000;47:1121–35. [10] Wolthaus JWH, Sonke JJ, van Herk M, et al. Comparison of different strategies to use four-dimensional computed tomography in treatment planning for lung cancer patients. Int J Radiat Oncol Biol Phys 2008;70:1229–38. [11] Yeung AR, Li JG, Shi W, et al. Tumor localization using cone-beam CT reduces setup margins in conventionally fractionated radiotherapy for lung tumors. Int J Radiat Oncol Biol Phys 2009;74:1100–7. [12] Schmidt ML, Hoffmann L, Kandi M, Møller DS, Poulsen PR. Dosimetric impact of respiratory motion, interfraction baseline shifts, and anatomical changes in radiotherapy of non-small-cell lung cancer. Acta Oncol 2013;52:1490–6. [13] Britton KR, Starkschall G, Liu H, et al. Consequences of anatomic changes and respiratory motion on radiation dose distributions in conformal radiotherapy for locally advanced non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2009;73:94–102.
[14] Yan D, Vicini F, Wong J, Martinez A. Adaptive radiation therapy. Phys Med Biol 2007;42:123–32. [15] Rong Y, Smilowitz J, Tewatia D, Tomé WA, Paliwal B. Dose calculation on kV cone beam CT images: an investigation of the Hu-density conversion stability and dose accuracy using the site-specific calibration. Med Dosim 2010;35:195–207. [16] Roman NO, Shepherd W, Mukhopadhyay N, Hugo GD, Weiss E. Interfractional positional variability of fiducial markers and primary tumors in locally advanced non-small-cell lung cancer during audiovisual biofeedback radiotherapy. Int J Radiat Oncol Biol Phys 2012;83:1566–72. [17] Knap MM, Hoffmann L, Nordsmark M, Vestergaard A. Daily cone-beam computed tomography used to determine tumour shrinkage and localisation in lung cancer patients. Acta Oncol 2010;49:1077–84. [18] Lavrenkov K, Partridge M, Cook G, Brada M. Positron emission tomography for target volume definition in the treatment of non-small cell lung cancer. Radiother Oncol 2005;77:1–4. [19] Chetty KG, Moran EM, Sassoon CS, Viravathana T, Light RW. Effect of radiation therapy on bronchial obstruction due to bronchogenic carcinoma. Chest 1989;95:582–4. [20] Vaaler AK, Forrester JM, Lesar M, Edison M, Venzon D, Johnson BE. Obstructive atelectasis in patients with small cell lung cancer. Incidence and response to treatment. Chest 1997;1997:115–20. [21] Farr KP, Khalil AA, Knap MM, Møller DS, Grau C. Development of radiation pneumopathy and generalised radiological changes after radiotherapy are independent negative prognostic factors for survival in non-small cell lung cancer patients. Radiother Oncol 2013;107:382–8. [22] Uyterlinde W, Chen C, Kwint M, et al. Prognostic parameters for acute esophagus toxicity in intensity modulated radiotherapy and concurrent chemotherapy for locally advanced non-small cell lung cancer. Radiother Oncol 2013;107:392–7. [23] Nijkamp J, Rossi M, Lebesque J, et al. Relating acute esophagitis to radiotherapy dose using FDG-PET in concurrent chemo-radiotherapy for locally advanced non-small cell lung cancer. Radiother Oncol 2013;106:118–23.
