Abstract
Background/Aim: To investigate dosimetric differences in organs at risk (OARs) and cardiac substructures in patients with locally advanced non-small cell lung cancer (NSCLC) between the adaptive radiotherapy (ART) and non-ART groups. Patients and Methods: Thirty patients were treated with definitive radiotherapy +/− chemotherapy. Cardiac substructures including the left anterior descending coronary artery (LAD) and large vessels, were contoured. Eight patients experienced tumor shrinkage and were replanned (ART). Cumulative plans after ART were compared to the original plans (not considering volume reduction) in terms of dosimetric parameters. The cumulative plans of the ART group (n=8) and non-ART group (n=22) were compared in terms of the same dosimetric parameters. Results: Within the ART group, the following parameters were found to be significantly improved after re-planning: mean lung dose (MLD) (13.79 Gy vs. 15.6 Gy), V20Gy both lungs (17.88% vs. 27.38%), ipsilateral MLD (20.87 Gy vs. 24.44 Gy), and esophagus mean dose (20.79 Gy vs. 24.2 Gy). No dosimetric differences were observed in heart substructures. Dosimetric parameters, particularly LAD, were significantly worse in the ART group than in the non-ART group. This is probably because this OAR was not considered in the plan optimization after re-planning, because it was not routinely contoured as an OAR. Conclusion: Our analysis showed an improvement in dosimetric parameters in the lungs and esophagus in the ART group. This approach may lead to a possible reduction in toxicity. Contouring of cardiac substructures could lead to a plan optimization of their parameters and eventually reduce the risk of cardiac toxicities in these patients.
Cardiac events occur in 5.8% of patients with locally advanced lung cancer treated with radiochemotherapy (RCHT) a median of 20 months after therapy (1). The mean heart dose (threshold ≥10 Gy) has been shown to increase the risk of coronary heart disease (1, 2). In the RTOG-9410 trial (3), 2.4% of grade ≥3 heart toxicities occurred after a sequential/combined regimen of radiation and chemotherapy in advanced non-small cell lung cancer (NSCLC) patients. However, these events remain unclear and cannot be attributed solely to thoracic radiation.
In fact, cardiac toxicities may also be related to the dose reaching the heart substructures (that is, anterior coronary artery, left ventricle, and others) and not only directly to the whole heart volume. Heart events could take years or decades to develop after therapy in patients with malignancies who have a good prognosis (4). The damage mechanism may involve the pericardium, heart muscles, valves, electric conduction system or vascular structures regarding mainly the small vessels.
Similar to other malignancies, such as breast cancer, the dose to the left anterior descending coronary artery (LAD) in the adjuvant setting is a well-established organ at risk (OAR) with defined constraints to avoid the risk of coronary events (5).
In recent years, the prognosis of advanced oligometastatic NSCLC has significantly improved mainly due to the introduction of more effective systemic therapies and their combination with local therapies (6-8). Technological developments have played a role in reducing toxicity rates and improving local control due to more accurate dose distribution and better OAR sparing. For example, the use of intensity-modulated radiotherapy (IMRT) or volumetric modulated arc-therapy (VMAT) compared to three-dimensional radiotherapy (3DRT) was associated with lower rates of severe pneumonitis and cardiac doses in locally advanced NSCLC (9), although the rates of grade 3 esophagitis and cardiovascular toxicity were not significantly different between IMRT and 3DRT. However, cardiac toxicity is progressively acquiring clinical interest in the context of locally advanced NSCLC as well as in planning optimization due to the relatively long survival of these patients (10).
The evidence of tumor volume reduction over the RCHT course in locally advanced NSCLC has led to the question of how to account for target changes (11). In fact, shrunken tumor targets might expose OARs to unwanted higher doses, which could lead to increased toxicity. On the other hand, this can eventually improve the visualization of the atelectatic lung, thus, improving tumor definition (11, 12). Adaptive radiotherapy (ART) and four-dimensional computed tomography (4DCT)-planning might represent a possible solution for this problem and can also help in further sparing of OARs and dosimetric optimization in case of tumor volume reduction during radiation. To date, there are only limited studies regarding adaptive radiotherapy for lung malignancies and the majority have focused on lung sparing (13). Cardiac toxicity in patients with advanced NSCLC remains an open issue and the constraints in radiation oncology regarding heart substructures remain unvalidated.
This study aimed to investigate the differences between dosimetric parameters in the lung, heart, and esophagus within a group of patients with locally advanced NSCLC receiving ART and to compare the same dosimetric parameters to the non-ART group. Moreover, dosimetric parameters in cardiac substructures retrospectively contoured were evaluated in both groups to investigate their importance in plan optimization with or without ART in patients with advanced NSCLC.
To the best of our current knowledge, this is the first study reporting detailed dosimetric data including cardiac substructures regarding ART in NSCLC planned with 4DCT-VMAT, also in comparison with a non-ART group.
Patients and Methods
Patients and treatment/contouring. Thirty patients treated with definitive radiotherapy (66 Gy/33 fractions) with or without chemotherapy, were selected for the analysis. All patients had previously been staged and had a histologically confirmed diagnosis of advanced stage IIIA/B NSCLC. Twenty-five patients presented with nodal involvement, 15 of whom presented with cN2-3 nodal status. In addition, 26 patients had large primary tumors in cT3-4 stage. All patients underwent a 4D Planning-CT and VMAT treatment planning.
Gross tumor volume (GTV) encompassed the positive mediastinal lymph nodes and the primary tumor. GTVs were contoured in different respiratory phases (0-100% in 10% steps, CT average, CT maximum intensity projection) for a median of 7-8 phases depending on tumor movement. Internal margin volumes (ITV) were generated from the GTVs. ITVs were expanded by 5 mm for primary tumors and by 0-2 mm for the involved mediastinal nodes in all directions to generate clinical target volumes (CTVs). Afterwards, CTVs were expanded by 3-5 mm in all directions to generate planning target volumes (PTVs). Plan optimization was driven with the aim of delivering the prescribed dose to at least 95% of the PTV, while maintaining the maximum dose (Dmax) to the PTV at 105% of the prescribed dose with heterogeneity corrections.
Dose constraints for OARs were as follows: lungs V20 <35% and V5 <65%, mean lung dose <20 Gy, V5 <65%; heart V40 <80%, V45 <60%, V60 <30%, mean heart dose <25 Gy; spinal canal + 3 mm (PRV) max dose <45 Gy. Heart substructures or great vessels were not routinely contoured and not considered for further plan optimization or the administration of radiation therapy.
A second 4DCT was performed at the administered dose of 40 Gy to evaluate the response during radiotherapy. The two 4D-CTs were fused to verify the tumor shrinkage. In case of response of the primary and/or mediastinal lymph nodes during radiation, the tumor and/or nodal volume including OARs were newly contoured. An adaptive radiation plan was implemented using the same constraints. In the case of non-response/irrelevant tumor response no adaptive planning was performed, and the patients continued to be treated with the original (first) plan.
For the purpose of this study, 38 4DCTs were revised including all patients (16 in the ART group and 22 in the non-ART group). In all 4DCTs, cardiac substructures (right and left ventricle, LAD) and large vessels (ascending/arch, descending aorta, superior vena cava, and pulmonary artery) were retrospectively contoured using a heart atlas (14) by an expert radiation oncologist, in order to evaluate further dosimetric parameters, previously not taken into account for the planning of radiation treatment (Figure 1 and Figure 2).
Contouring of organs at risk and cardiac substructures in axial images of an average 4DCT: heart, right and left ventricle.
Contouring of organs at risk and cardiac substructures in axial images of an average 4DCT: great vessels and left anterior coronary artery.
Endpoints and statistical analysis. The primary endpoint of this study was to investigate the differences between dosimetric parameters in the principal OARs lung, heart, and esophagus in a group of patients with locally advanced NSCLC after replanning. Additionally, dosimetric parameters in the heart substructures within the ART-group were also evaluated to determine possible parameters, which can be included as constraints in future planning optimization, since they are not usually considered in routine radiation planning. The secondary endpoint was to compare the difference between the same dosimetric parameters in OARs and heart substructures between patients treated in the ART group and those treated in the non-ART group to observe possible benefits for good responders.
Dosimetric parameters of OARs and newly contoured heart structures were extracted (9). In eight patients, a significant reduction in tumor volume was detected and they were replanned using ART. In this group, the cumulative plans after ART (before and after replanning) and the original plans (first plans) were compared in terms of dosimetric parameters using Wilcoxon signed rank test. In addition, the cumulative plans of the ART group (n=8) and the non-ART group (n=22) were compared in terms of the same dosimetric parameters using the Mann–Whitney test.
Results
In the entire population of patients (n=30), the median value of mean heart dose was 9.87 Gy and the heart V5 Gy, V30 Gy, and V50 Gy were 44.5%, 7.48%, and 2.36% respectively. The median values of mean LAD dose, V15 Gy and Dmax 1 cc to the LAD were 8 Gy, 21.9% and 13.14 % respectively. The median values of MLD, V5 Gy, and V20 Gy to the lungs were 16.38 Gy, 61.9% and 26.79%, respectively (Table I).
Target and constraints characteristics for the entire patient population (n=30).
ART group. Within the ART group the following parameters were found to be significantly improved after replanning as a result of tumor shrinkage compared to original plan (first plan) (Figure 2): MLD (13.79 Gy vs. 15.6 Gy, p=0.038), V20 Gy both lungs (17.88% vs. 27.38%, p=0.011), ipsilateral MLD (20.87 Gy vs. 24.44 Gy, p=0.021) and esophagus mean dose (20.79 Gy vs. 24.2 Gy, p=0.028). A trend of significance (p=0.051) was observed in the improvement of mean dose to the left ventricle. No difference in dosimetric parameters was found for the whole heart volume and other cardiac substructures (Table II).
Treatment volume and constraint comparison between the same patient group (ART-group) with or without adaptive planning strategy (n=8).
ART vs. no-ART group. The mean heart dose was 11.31 Gy and 10.46 Gy in no-ART and ART group, respectively (p=0.086). The following dosimetric parameters were found to be significantly worse in the ART-group compared to the non-ART-group: LAD maximum dose 1 cc (31.57 Gy vs. 15.56, p=0.009) and 0.03 cc (42.89 Gy vs. 26.57 Gy, p=0.028), LAD mean dose (16.03 Gy vs. 10.38 Gy, p=0.049) and LAD V15 Gy (41.35% vs. 20.24%, p=0.015), mean descendent aorta dose (27.99 Gy vs. 16.59 Gy, p=0.032), and also maximal descendent aorta dose 0.03 cc (66.09 Gy vs. 63.10 Gy, p=0.049) and mean pulmonary artery dose (40.60 Gy vs. 31.94 Gy, p=0.049). Only maximal dose 0.03 cc to the ascending aorta/arch was found to be significantly better (59.43 Gy vs. 60.65 Gy, p=0.044) in the ART-group compared to the non-ART-group (Table III and Figure 3).
Target and constraints parameters for the adaptive (n=8) and non-adaptive (n= 22) population (total n=30).
Boxplots showing constraint significant results between adaptive plan and no adaptive plan (n=8). A) mean lung dose (MLD) of the lungs (p=0.038), B) lungs V20 (p=0.011), C) ipsilateral MLD (p=0.021), D) esophagus mean dose (p=0.028).
Discussion
With the improved prognosis of advanced NSCLC and increased survival over time, an improvement in the quality of radiotherapy needs to be considered. In advanced NSCLC, little improvement has been made in determining the optimal dose and volume as well as in RT planning optimization. Therefore, 3D radiation treatments are often reported in the literature, and IMRT/VMAT has not yet become routine in the clinical practice, so that irradiated volumes are still very large, as well as treatment margins.
Technology development in radiation oncology, the use of IMRT or VMAT planning, image-guided radiotherapy (IGRT), and modern online adaptive radiotherapy (15-17) allow to safely deliver treatments (9). As a result, the rates of severe lung toxicity and the mean heart dose could be reduced (4). An adaptive radiotherapy workflow for good responders during treatment could also allow a dose escalation developing SIB (simultaneous integrated boost) radiation regimens and simultaneously reduce toxicity over time (18). A recent dosimetric study of 13 patients with NSCLC demonstrated that ART can reduce lung doses without reducing safety margins by utilizing synthetic CTs generated from converted cone beam CTs (19).
We showed that the use of adaptive radiotherapy by tumor shrinkage and VMAT planning significantly improved MLD (p=0.038), V20 Gy of both lungs (p=0.011), ipsilateral MLD (p=0.021), and esophagus mean dose (p=0.028). Our results are in accordance with those of a dosimetric analysis conducted by Hoegen et al. (13) on 10 patients who received IGRT ART for locally advanced NSCLC, obtaining a significantly decreased V20Gy and mean lung dose at high levels of plan conformity.
A longitudinal study investigated the volume and dose variability in 15 patients with lung cancer undergoing conventionally fractionated 4DCT-planned radiochemotherapy (20). Repeated dose estimations of the heart and lung revealed that anatomical and positioning variations during radiotherapy induced changes and was predicted for normal tissue toxicity.
The use of 4DCT in reducing margins and escalating the dose and also the use of VMAT-IGRT in improving dosimetric parameters could be considered as a potential strategy in the clinical practice for advanced NSCLC (15). However, even if preliminary data report a dosimetric advantage of ART some questions remain unanswered, such as which tumor volume reduction parameters should be used to efficiently adapt the plan, what toxicity improvement could be expected, and whether tumor shrinkage should be observed in order to find the candidates for an adaptive radiotherapy.
A recent analysis of 2,031 patients from Canada (21) showed that median survival in patients with NSCLC who had a cardiac event after 3D-RCHT was significantly shorter compared to those treated with modern techniques (13.7 vs. 23.4 months, p=0.012). Moreover, there was an improvement in cumulative cardiac toxicity with IMRT or VMAT techniques.
Additionally, an increase in cardiac events has generally been observed after combined definitive treatment in patients with NSCLC, possibly because a relatively prolonged overall survival (OS) might reveal more frequently late toxicity. A re-analysis of RTOG 0617, reported that LAD V15 Gy ≥10% was associated with a significantly increased risk of all-cause mortality (p=0.037) in patients with locally advanced NSCLC treated with thoracic radiotherapy, whereas mean heart dose was not. The median OS for patients with LAD V15 Gy ≥10% vs. <10% was 20.2 vs. 25.1 months, respectively, and the estimated 2-year OS was 47% vs. 67% (p=0.004), respectively (22). In our analysis, dosimetric parameters in great vessels, and particularly in the LAD (maximum dose, mean dose and V15Gy), were significantly worse in the ART group compared to those in the non-ART group. This is probably because these OARs were not considered for further plan optimization after re-planning, because they are not routinely contoured as OARs. Dose constraints regarding this anatomical structure need to be validated in large series or prospective data.
Heart dosimetric parameters, such as the percentage of heart volume receiving ≥5 Gy and ≥30 Gy are important predictors of survival (23) and should be considered in the optimization of radiation plans. Moreover, Dess et al. reported, that the 2-year incidence of grade ≥3 cardiac events primarily consisting of acute coronary syndrome exceeded 10% among patients with locally advanced NSCLC treated with definitive thoracic radiation; pre-existing heart disease and higher mean heart dose were significantly associated with higher cardiac event rates (24). The same group advises to reduce heart doses in order to decrease the risk of radiation-associated heart injury. In our analysis, the median values of V5 Gy and V30 Gy for the whole heart were about 45% and 12%, respectively; however, they were not significantly better after ART. In contrast, the lung constraints were significantly better in the ART group. In our study, the median mean heart dose for all patients was 9.9 Gy; however, this parameter was reported to be 11 Gy in the study by Dess et al.
Our study reported an improvement in dosimetric parameters of OARs, particularly in the lung and esophagus, in patients with advanced NSCLC undergoing definitive radiotherapy using 4DCT and ART after tumor shrinkage. Dosimetric parameters of cardiac substructures probably related to the risk of heart injuries were also investigated. These parameters worsened after ART compared with the non-ART group since they were not considered in the optimization planning process. However, these structures could be considered in the future as risk organs for cardiac disorders and further optimization of radiation planning should be evaluated. The existing literature on cardiac toxicity in patients with NSCLC undergoing thoracic radiotherapy consists of small, retrospective series, mostly single-center analyses with several endpoints (25). More effort is required in this field. Ongoing prospective studies (NCT03978377, NCT04305613, and NCT03645317) including multidisciplinary collaboration with cardiologists aim to investigate heart toxicities after thoracic radiotherapy and to correlate outcomes to possible biomarkers.
In addition, a planning with high tech radiotherapy and a refinement of volumes including also OAR for locally advanced NSCLC should also be introduced in clinical practice, since the survival of these patients is also increasing due to the development of new systemic therapies (10). Hence, studies with a larger number of patients are needed in order to increase expertise in this field.
Conclusion
Our analysis showed an improvement in dosimetric parameters in the lungs and esophagus of patients affected by locally advanced NSCLC treated with ART. This approach could lead to a reduction in pulmonary and esophageal toxicities. New research studies, particularly regarding radiotherapy in advanced NSCLC, should include cardiological evaluations to determine the risk and mechanisms of heart events due to therapy. Contouring of cardiac substructures could lead to the optimization of their parameters and eventually reduce the risk of cardiac toxicities in these patients.
Footnotes
Authors’ Contributions
Investigation, L.A. and D.H.; formal analysis, L.N.; writing original draft, L.A: and L.N.; methodology, LA. and Z.E.; writing–review & editing, L.A. and D.H.; data curation, Z.E. and L.N.; research conceptualization and final manuscript revision L.A.; project administration, L.A. All Authors have read and agreed to the published version of the manuscript.
Conflicts of Interest
The Authors have no conflicts of interest to declare in relation to this study.
- Received September 28, 2023.
- Revision received October 28, 2023.
- Accepted October 31, 2023.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
