- Email: [email protected]
Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy
Radiotherapy and Oncology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy Yi Pan a,b, Carsten Brink b,c, Marianne Knap d, Azza A. Khalil d, Christa H. Nyhus e, Tine McCulloch f, Bente Holm g, Yi-long Wu h, Tine Schytte i, Olfred Hansen b,i,⇑ a Department of Radiation Oncology, Guangdong General Hospital & Guangdong Academy of Medical Science, PR China; b Institute of Clinical Research, University of Southern Denmark, Odense; c Laboratory of Radiation Physics, Odense University Hospital; d Department of Oncology, Aarhus University Hospital; e Department of Oncology, Vejle Hospital, Sygehus Lillebaelt; f Department of Oncology, Aalborg University Hospital; g Department of Oncology, Herlev University Hospital, Denmark; h Guangdong Lung Cancer Institute, PR China; and i Department of Oncology, Odense University Hospital, Denmark
a r t i c l e
i n f o
Article history: Received 4 October 2015 Received in revised form 8 January 2016 Accepted 10 January 2016 Available online xxxx Keywords: Acute esophagitis Non small cell lung cancer IMRT Concurrent chemoradiotherapy
a b s t r a c t Purpose: Esophagitis is common in patients treated with definitive radiotherapy for local–regional advanced non small cell lung cancer (NSCLC). The purpose of this study was to estimate the dose–effect relationship using clinical and dosimetric parameters in patients receiving intensity modulated radiotherapy (IMRT) and concomitant chemotherapy (CCT). Methods: Between 2009 and 2013, 117 patients with stages IIB–IIIB NSCLC were treated in a multicenter randomized phase II trial with 2 cycles of induction chemotherapy followed by IMRT and CCT. The esophagitis was prospectively scored using the Common Toxicity Criteria 3.0. Clinical and dosimetric variables were analyzed for the correlation with grade P2 esophagitis through logistic regression. Results: Grade 2 esophagitis was experienced by 31 (27%). All models including gender, institution, a dosimetric parameter and a position parameter were significantly associated with esophagitis. The two models using the relative esophagus volume irradiated above 40 Gy (V40, OR = 2.18/10% volume) or the length of esophagus irradiated above 40 Gy (L40, OR = 4.03/5 cm) were optimal. The upper part of esophagus was more sensitive and females experienced more toxicity than men. Conclusion: V40 and L40 were most effective dosimetric predictors of grade P2 esophagitis. The upper part of esophagus was more sensitive. Ó 2016 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2016) xxx–xxx
Thirty-five percent of patients with non small cell lung cancer (NSCLC) are irresectable at diagnosis due to being in locoregional advanced stage [1]. Definitive concomitant chemoradiotherapy (CCRT) is the standard of care for this group of patients if they are in good performance status. Overall survival is improved by 5% by CCRT compared to sequential chemoradiotherapy, which may be caused by an increased local–regional control [1]. Although CCRT has a better therapeutic effect, the risk and severity of acute toxicity, especially the acute esophagitis may increase. In a metaanalysis performed by the NSCLC Collaborative Group containing 6 randomized clinical trials, the incidence of grade 3–4 esophagitis in CCRT (18%) was higher than observed in sequential chemoradiotherapy (4%), with a relative ratio of 4.9 (p < 0.01) [1]. For patients treated with radiotherapy alone, the incidence of esophagitis Pgrade 3 is 1%, but it is markedly increased to 6–24% with CCT [2]. ⇑ Corresponding author at: Department of Oncology, Odense University Hospital, DK-5000 Odense, Denmark. E-mail address: [email protected] (O. Hansen).
The symptoms of esophagitis include retrosternal pain, dysphagia and odynophagia. It often occurs 2–4 week after the start of radiation therapy and may persist 2–7 weeks [3,4]. Severe esophagitis may induce hospitalization, and require parenteral nutrition, naso-gastric tube and analgetics. It may lead to treatment interruptions, which may have a negative impact on the quality of life of the patients, lower the chance of local tumor control, and thus impact survival negatively. Intensity modulated radiotherapy (IMRT) facilitates a more conform radiation dose distribution allowing for shaping the dose of the tumor and sparing esophagus more than 3-dimensional radiation (3DCRT) may be able to do. For NSCLC, IMRT has the potential to escalate target dose without a corresponding increase to the lung and esophagus [5–7]. In many centers IMRT has become the principal radiation technique for NSCLC to reduce local relapse and protect critical organ at risk (OAR). Although a number of clinical and dosimetric risk factors for esophagitis have been reported in previous studies [8–13], no dosimetric parameter has been generally accepted as the best predictor. In addition, most of these studies were based on 3DCRT. The
http://dx.doi.org/10.1016/j.radonc.2016.01.007 0167-8140/Ó 2016 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
2
Esophagitis in IMRT and CCT for NSCLC
objective of this study was to estimate the dose–effect relation between esophagitis and clinical as well as dosimetric parameters in patients with local–regional advanced NSCLC receiving IMRT and CCT. Materials and methods Patients Between May 2009 and August 2013, 117 patients with NSCLC were enrolled in the Danish multicenter randomized phase II trial (NARLAL). The inclusion criteria were histologically or cytologically confirmed NSCLC, clinical American Joint Committee of Cancer stage (the seventh edition) IIB–IIIB, performance status (PS) <2 on the ECOG scale, weight loss 610%, adequate hepatic and renal function, forced expiratory volume in 1 s (FEV1) P1.0. Patients with pleural effusion, prior chemotherapy for lung cancer and any other active malignancy within the last 5 years were excluded. Each patient underwent basic laboratory studies, computed tomography (CT) scan of the chest and upper abdomen with contrast and whole body positron emission tomography (PET) before treatment. Approval was granted by the Institutional Review Board for conducting this study. The NARLAL clinic cohort study was filed at clinicaltrials.gov (No. NCT00887783). The medium age was 66 years (44–82 years). The patient characteristics are presented in Table 1. NARLAL protocol All patients were treated with induction chemotherapy followed by CCRT. The induction chemotherapy consisted of 2 cycles of Carboplatin (AUC = 5, day 1) and Vinorelbine oral (60–80 mg/m2, day 1 and 8). After completing induction chemotherapy, patients were randomized into two groups: 60 Gy in 30 fractions (Arm A) and 66 Gy in 33 fractions (Arm B), once daily, 5 times per week. The concomitant chemotherapy regimen consisted of Vinorelbine oral 50 mg 3 times a week for 6—6 12 weeks. Eighty-three percent of the patients received the scheduled doses of concomitant chemotherapy (Supplementary Table 1). Radiotherapy preparation IMRT was planned for all patients. Gross tumor volume including the lung tumor and involved mediastinal lymph node was defined on CT of the chest and PET-CT. A treatment planning CT scan was performed with 2.5–3 mm slice thickness from upper neck to mid-abdomen. Radiotherapy was delivered with a linear accelerator using energy of 6 MV X-rays. The planning technique was based on ICRU62 recommendations [14]. Critical organ dose tolerances were defined as maximum dose of esophagus 6 66 Gy, lung V20 6 40%, spinal cord 6 45 Gy, heart V50 6 20%. Esophageal contours and dosimetric data For the consistency, the esophagus was re-delineated in all patients by the same oncologist (Y.P.). The external surface of esophagus was contoured on each axial plane of the planning CT scan from the inferior margin of cricoid cartilage to the gastroesophageal junction. Dose-volumetric values were calculated from Dose Volume Histogram (DVH) for each patient. The following esophageal dosimetric parameters were extracted from the dose data set in treatment planning system: maximum, median and mean dose, the entire length and volume, series of dosevolumetric values from DVH such as: the percentage of the volume
Table 1 Patient characteristics (N = 117). Characteristics
Number
Percentage
Gender Male Female
68 49
58 42
Age (years) Median (range)
65.5 (44–82)
Smoking Never and former smokers Current smoker
89 28
76 24
Histology SCC NSCC
47 70
40 60
Performance status 0 1
62 55
53 47
T stage T1 + 2 T3 + 4
74 43
63 37
N stage N0–1 N2 N3
18 85 14
15 73 12
Stage IIB IIIA IIIB
6 81 30
5 69 26
Prescribed Radiation Dose (Gy) 66 60
57 60
49 51
Length of esophagus (cm) Median (range)
26 (20–31)
Volume of esophagus (cm3) Median (range)
35 (19–141)
Mean esophagus dose (Gy) Median (range)
23 (3–42)
Maximum esophagus dose (Gy) Median (range)
63 (9–72)
Median esophagus dose (Gy) Median (range)
10 (1–60)
Esophagus V20 (%) Median (range)
43 (0–70)
Esophagus V30 (%) Median (range)
36 (0–70)
Esophagus V40 (%) Median (range)
32 (0–68)
Esophagus V50 (%) Median (range)
26 (0–66)
Esophagus V60 (%) Median (range)
8 (0–50)
Abbreviations: SCC = squamous cell carcinoma; NSCC = non squamous cell carcinoma; V20–V60: percentage of esophageal volume receiving at least 20–60 Gy.
of esophagus receiving from P20 Gy (V20) and in 10 Gy increments up to P60 Gy (V60). Based on the hypothesis that length and positional information, which is discarded in traditional DVH, might be important to predict grade 2 or greater esophagitis, the following metrics were calculated: the length of the part of esophagus for which the median dose within a transversal CT slices was above 20 Gy, 30 Gy, 40 Gy, 50 Gy, or 60 Gy (L20, L30, L40, L50, or L60) and mean relative position of the volume related to L20–L60 (P20, P30, P40, P50, and P60) defined as 0 at the most cranial located esophagus delineation and 1 at the most caudal located delineation.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
3
Y. Pan et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx
Follow-up and esophagitis scoring Acute esophagitis was defined according to version 3.0 of National Cancer Institute’s Common Toxicity Criteria for Adverse Events version 3.0 (NCI-CTCAE 3.0) from the start of RT until 3 months afterward. During RT, patients underwent evaluation for side effects once a week and more often if interventions were indicated. After RT, follow up monitoring was at regular intervals. First time was 30–45 days after radiation and then every 3 month. The clinical esophagitis grades were available for all patients. Statistics analyses/normal tissue complication probability (NTCP) modeling Patients and treatment characteristics were summarized by descriptive statistics. Firstly, univariate analyses were used to identify the significant risk factors among gender, age, institution, smoking status, PS, histology, tumor dose, and dosimetric parameters including maximum dose, median dose, mean dose, volumetric metrics, length and position metrics. Due to large inter variable correlation within the dosimetric variables and within the position variables, only one of the positions and one of the dosimetric variables were included in pairs for each multivariable analysis. Backward selection was performed in the binary logistic regression analyses. Akaike information criterion (AIC) was used for model selection, which balances goodness of fit with the complexity of the model. AIC is defined as AIC = 2 * k 2 * ln(L), where k is the number of parameters in the model and L is the maximized value of the likelihood function. The preferred model was taken as the one resulting in the lowest AIC value. Model calibration was estimated by classifying patients into 6 risk groups according to predicted probability of esophagitis and plotting the mean observed probability versus the mean predicted probability for each group. Goodness of fit was estimated by the Hosmer–Lemeshow test. To further validate the model performance area under the receiver operating curve (AUC), Brier score, and cross validation of the models by bootstrap was performed and is presented as Supplementary material. To validate the impact of the fixed dose of Vinorelbine as concurrent chemotherapy, the body surface area (BSA) was also tested within the models. BSA is defined as BSA = 0.007184 * weight0.425 * height0.725 [15]. Value of p < 0.05 was considered statistically significant. Statistical analysis was performed with SPSS (Statistics Package for Social Science). No patient was lost to follow-up.
Fig. 1. Incidence of acute esophagitis.
maximum dose, mean dose, median dose, all relative volumetric variables (V20, V30, V40, V50, and V60), absolute length (L20, L30, L40, L50, or L60) and position variables (P20, P30, P40, P50, and P60) were all significant. An Anova analysis did not show any associations between dosimetric/position variables and gender/ institution. Each model in binary logistic regression contained 4 independent variables: gender, institution, one position parameter, and one dosimetric parameter including maximum dose, mean dose, median dose, volumetric (V20–V60), and length metrics (L20–L60). All test models demonstrated a significant correlation between esophagitis and the predictor variables. The models with mean position for the length of esophagus receiving P20 Gy (P20) variable had lower AIC values than models with P30–P60. Hence, P20 was selected to estimate the effects of irradiation of specific parts of the esophagus (relative position along esophagus). Table 2 shows the AIC of logistic models with gender, institution, P20 and one dosimetric parameter. For grade P2 esophagitis, the models containing V40 and V50 had the lowest AIC value. The AIC value for V40 was only slightly lower than the corresponding value for V50. L40 was very close to V40 and V50 and was further investigated since L40 is conceptual different from V40 in term of the need of delineating the entire esophagus. Table 3 shows the results of multivariable analysis. Female, longer length of esophagus receiving doses above 40 Gy, large relative volume of esophagus receiving 40 Gy or more, and
Results Of 117 patients treated with IMRT and CCT, 91 (78%) patients developed esophagitis including 50 (43%) patients with grade 1 esophagitis, 31 (27%) with grade 2 and 10 (8%) with grade 3. No grade 4 or 5 esophagitis was observed (Fig. 1). The medium onset time was 18 days (1–44 days) after the start of the RT. The incidences of grade P2 esophagitis in 5 institutions were 33% (12/36) in OUH, 40% (8/20) in Vejle Hospital, 8% (1/12) in Aalborg Hospital, 46% (19/41) in Aarhus Hospital, and 13% (1/8) in Herlev hospital, respectively. For the total study population, the average length of esophagus was 26 cm (20–31 cm). The dosimetric data of the 117 patients are shown in Table 1. Among the clinical parameters: gender, age, institution, smoking status, performance status (PS), histology, tumor dose, gender was the only significant factor for grade P2 esophagitis (p = 0.001, Supplementary Table 2). The association between esophagitis and T, N, and clinical stage were not analyzed since they were highly associated with dosimetric parameters which in turn correlate with esophagitis. The dosimetric parameters,
Table 2 Akaike information criterion of the models. Dosimetric parameters
AIC
Mean dose Maximum dose Median dose L20 L30 L40 L50 L60 V20 V30 V40 V50 V60
110.902 113.296 122.52 114.811 113.438 109.8648a 111.825 119.02 115.312 112.243 108.844a 109.313a 116.937
Abbreviations: AIC = Akaike information criterion. All models included 4 variables: gender, institution, P20 and one of above dosimetric parameters. a Minimum Akaike information criterion.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
4
Esophagitis in IMRT and CCT for NSCLC
Table 3 Results of multivariate analyses.
Model A
Model B
Variable
OR
95% CI
p-Value
Constant Female gender L40 (per 5 cm) P20 (per 0.1 relative unit) Institution
0.38 12.74 4.03 0.52
3.99–40.64 1.84–8.79 0.29–0.95
0.598 <0.001 <0.001 0.032 0.020
Constant Female gender V40 (per 10% relative volume) P20 (per 0.1 relative unit) Institution
0.91 12.33 2.18 0.44
3.89–39.06 1.44–3.30 0.24–0.81
0.957 <0.001 <0.001 0.008 0.011
Abbreviations: OR = odds ratio; CI = confidence interval; L40 = length of 50% esophagus within a slice receiving P40 Gy; P20 = mean position for the length of which 50% esophagus within a slice received P 20 Gy.
cranial position of esophageal receiving more than 20 Gy significantly increased the risk of grade P2 esophagitis. BSA was not significant in a multivariable analysis neither together with the gender variable, nor within the two genders separately. Model performances for the V40 and L40 model is illustrated in Fig. 2, comparing the mean predictive probability with the mean observed probability of esophagitis. The p values for goodness-of-fit were 0.440 and 0.410 – reflecting the relative good correlations seen Fig. 2. AUC, Brier score, and cross validation of models are shown in Supplementary Table 3. As a result of multivariable logistic regression analysis, the NTCP model of developing grade P2 esophagitis was expressed as:
NTCP ¼ ð1 þ expðSÞÞ1 where S was defined as:
S ¼ 0:97 þ 2:55 Sex þ 0:28 cm1 L40 6:45 P20 þ b4 Institution S ¼ 0:09 þ 2:51 Sex þ 7:79 V40 8:19 P20 þ b4 Institution Sex: female = 1, male = 0; P20: cranial position = 0, caudal position = 1; V40: 0% = 0, 100% = 1. The constant term in the S equation (0.97 and 0.09) depended on which institution is used as reference for calculation of b4. As standard one of the institutions was reference institution during the binary regression, the b4 values for the remaining institutions was ranging from 2.4 to 1.4 and 2.5 to 1.6 for the L40 and V40 models, respectively.
Stratification for institutions was performed in order to correct for differences in reporting of the toxicity level. Without stratification all the parameters in the above models were still significant except for P20 in the L40 model in which the p-value was 12%. To summarize: larger V40, longer L40, the upper part of esophagus and female gender were significant risk factors for developing grade 2 or greater acute esophagitis. Discussion Numerous dosimetric parameters have previously been investigated for the association with acute esophagitis. However, these results were mostly based on 3DCRT and conventional radiotherapy. Only few studies reported esophagitis in patients treated with IMRT and the same CCT regimen until now [16,18]. As IMRT has been introduced as a treatment option, previous data has not been sufficient to estimate esophageal toxicities in IMRT. Kwint et al. showed that historical NTCP model was unsuitable to predict esophagitis in IMRT [16]. Furthermore, some parameters make no sense in IMRT such as length of esophagus included in the radiation field. With the use of IMRT in patients with NSCLC, further study was needed to get a better model of prediction. CCRT were reported to result in a 29–37% rate of grade 2 acute esophagitis and a 4% and 20% rate of grade 3 acute esophagitis [3,17,18]. In our study, the incidences of grade 2 and 3 esophagitis were 27% and 8%, respectively, which are similar to those reported by Rodriquez et al. [17] and Wijsman et al. [18]. Kwint et al. reported higher incidences of esophagitis in IMRT (38% and 22%) [16]. One possible reason is that the radiation dose per fraction was 2.75 Gy (66 Gy in 24 fractions) in their study, whereas it was only 2 Gy in our study [19]. Within the previous published studies there is no consensus on a possible association between gender and esophagitis. A number of studies report not to be able to detect a statistically significant association between gender and esophagitis [12,20]. However, Dehing-Oberije et al. [8] and Wijsman et al. [18] found as in the current study, a statistically significant correlation between gender and esophagitis although with less OR than in the current study. Since the concomitant chemotherapy regimen in our study was a fixed dose of Vinorelbine, the high OR for gender could be somewhat confounded by the BSA, which is less for the female gender. However, confounder analyses were not able to support the existence of a confounding effect. Previous published data have demonstrated that volumes receiving P20–80 Gy (V20–V80) is significantly correlated with esophagitis [8–10], especially V30–V50 [9,10,21]. Our results showed V40 and V50 were most strongly associated with grade
Fig. 2. Calibration plot of L40 model (A) and V40 model (B) for grade P2 esophagitis in NSCLC patients treated with IMRT and CCT, illustrating the relation between the predicted and observed probability. Patients have been grouped by increasing predicted probability.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
Y. Pan et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx
P2 esophagitis in NSCLC patients treated with IMRT compared to other volumetric parameters. Their values on predicting esophagitis were almost equivalent. Based on the AIC values V40 might be slightly better than V50. In some studies, length parameters have been found to be significant in univariate analysis, but not in multivariable analysis [11–13]. In our study, length of esophagus within a slice receiving P40 Gy (L40) was identified to be a strong predictor for grade P2 esophagitis. Comparing the models with V40 and L40, it is difficult to decide which logistic model is the better to predict esophagitis as AIC of the two models were very close. It is very convenient for clinicians to estimate esophagitis by volume metrics as they can be directly obtained through dosevolume histogram in the treatment planning systems while this is not easily done with length metrics. However, V40 is the relative volume of esophagus and it depends on delineation of how much of the esophagus is contoured. By contrast, L40 is the absolute esophageal length and independent of esophagus delineation within regions of low dose. Which parameter to use, V40 or L40, depends on the clinicians’ choice of the accessibility to the information and the robustness against delineation inaccuracies. The maximum dose, mean dose, and V60 were reported to have a significantly association with esophagitis in patients treated with 3DCRT [12,22,23]. Mean dose recommendation was adopted in the National Comprehensive Cancer Network (NCCN) guidelines [24]. The univariable dose results from Wijsman et al. are in close agreement with the univariable result in the current study (Supplementary Table 2). However, very strong correlations exist between many of the dose parameters, and a model reduction needs to be performed for a multivariable analyses. Wijsman et al.’s chose the mean dose to be the only dose parameter used within their multivariable models in their retrospective study of 149 NSCLC IMRT treated patients [18]. Their choice using the mean was based on LKB result with a volume parameter close to the value one. The present study uses the statistical concept Akaike information criterion (AIC) to select between the multivariable models including the different dose parameters. The current results indicate that logistic models with maximum dose, mean dose, or V60 were not the optimal choice for evaluating grade P2 esophagitis, even though these dose parameters in 3D CRT studies and in Wijsman’s IMRT study have shown associations with esophagitis. IMRT is likely to have as lower association between dose to target and to normal tissue as 3DCRT [5] and may cause a difference between IMRT and 3D CRT toxicity models. It may therefore be difficult to extrapolate normal tissue complication probability models based on 3DCRT treatment to treatments based on IMRT. Furthermore, it is hard to identify the association between esophagitis and dosimetric parameters in multivariable analysis when these parameters are highly inter-correlated. The inter-correlation is likely stronger in 3D CRT studies than in IMRT studies. In the current study, caudal position of esophagus irradiated above 20 Gy decreased the risk of grade P2 esophagitis. To the best of our knowledge, this is the first time position parameter is reported as a predictive factor for esophagitis. One possible reason is the distributions of sensory nerves vary in upper, middle, and lower esophagus [25]. Sensory axons are abundant in cervical esophagus and progressively spares in thoracic and abdominal esophagus. Patients may be more sensitive to injuries in the upper part of esophagus than in the middle and lower part and have more severe symptoms. In this multicenter study, the incidence of grade P2 esophagitis in the participating institutions was different and institution was a significant factor to esophagitis in multivariable analysis. At present, the scoring system which is most commonly used for assessment of esophagitis is CTC scale. It evaluates the occurrence and severity of esophagitis mainly by symptoms reported by patients and treatments used by doctors, which are convenient for clinical
5
management. However, patients’ subjective perception and doctors’ judgment vary from person to person. This may result in overestimation or underestimation of the incidence of esophagitis. This shows that evaluation of toxicity is inherently difficult and calls on continuous education and discussion of how to score toxicity as precise as possible. Conclusion This study demonstrated in a cohort of patients with NSCLC receiving IMRT and concurrent chemotherapy that V40 and L40 were the most effective dosimetric predictors of grade 2 or greater acute esophagitis. The application of V40 or L40 depends on clinicians’ preference for either the accessibility to the information or robustness against delineation inaccuracies. Mean relative position of the part of esophagus within a slice receiving P20 Gy (P20) is found to be significantly correlated with acute esophagitis and should be included when modeling acute esophageal toxicity. In IMRT, maximum dose and mean dose was not as informative as they are in conventional radiotherapy and 3DCRT. While the current findings present good performance of the models, OR for gender may be somewhat confounded as fixed dose CCT was used in our study. Validation in a large independent patient cohort is needed to confirm this result. Conflict of interest statement No conflict of interest. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2016.01. 007. References [1] Auperin A, Le Pechoux C, Rolland E, Curran WJ, Furuse K, Fournel P, et al. Metaanalysis of concomitant versus sequential radiochemotherapy in locally advanced non-small-cell lung cancer. J Clin Oncol 2010;28:2181–90. [2] Werner-Wasik M, Yorke E, Deasy J, Nam J, Marks LB. Radiation dose-volume effects in the esophagus. Int J Radiat Oncol Biol Phys 2010;76:S86–93. [3] Uyterlinde W, Chen C, Kwint M, de Bois J, Vincent A, Sonke JJ, 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. [4] Werner-Wasik M, Pequignot E, Leeper D, Hauck W, Curran W. Predictors of severe esophagitis include use of concurrent chemotherapy, but not the length of irradiated esophagus: a multivariate analysis of patients with lung cancer treated with nonoperative therapy. Int J Radiat Oncol Biol Phys 2000;48:689–96. [5] Grills IS, Yan D, Martinez AA, Vicini FA, Wong JW, Kestin LL. Potential for reduced toxicity and dose escalation in the treatment of inoperable non-smallcell lung cancer: a comparison of intensity-modulated radiation therapy (IMRT), 3D conformal radiation, and elective nodal irradiation. Int J Radiat Oncol Biol Phys 2003;57:875–90. [6] Schwarz M, Alber M, Lebesque JV, Mijnheer BJ, Damen EM. Dose heterogeneity in the target volume and intensity-modulated radiotherapy to escalate the dose in the treatment of non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2005;62:561–70. [7] Nielsen TB, Hansen O, Schytte T, Brink C. Inhomogeneous dose escalation increases expected local control for NSCLC patients with lymph node involvement without increased mean lung dose. Acta Oncol 2014;53:119–25. [8] Dehing-Oberije C, De Ruysscher D, Petit S, Van Meerbeeck J, Vandecasteele K, De Neve W, et al. Development, external validation and clinical usefulness of a practical prediction model for radiation-induced dysphagia in lung cancer patients. Radiother Oncol 2010;97:455–61. [9] Rose J, Rodrigues G, Yaremko B, Lock M, D’Souza D. Systematic review of dosevolume parameters in the prediction of esophagitis in thoracic radiotherapy. Radiother Oncol 2009;91:282–7. [10] Kim TH, Cho KH, Pyo HR, Lee JS, Han JY, Zo JI, et al. Dose-volumetric parameters of acute esophageal toxicity in patients with lung cancer treated with threedimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2005;62:995–1002.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
6
Esophagitis in IMRT and CCT for NSCLC
[11] Belderbos J, Heemsbergen W, Hoogeman M, Pengel K, Rossi M, Lebesque J. Acute esophageal toxicity in non-small cell lung cancer patients after high dose conformal radiotherapy. Radiother Oncol 2005;75:157–64. [12] Ahn SJ, Kahn D, Zhou S, Yu X, Hollis D, Shafman TD, et al. Dosimetric and clinical predictors for radiation-induced esophageal injury. Int J Radiat Oncol Biol Phys 2005;61:335–47. [13] Rosenman JG, Halle JS, Socinski MA, Deschesne K, Moore DT, Johnson H, et al. High-dose conformal radiotherapy for treatment of stage IIIA/IIIB non-smallcell lung cancer: technical issues and results of a phase I/II trial. Int J Radiat Oncol Biol Phys 2002;54:348–56. [14] Baggins B. Prescribing, recording and reporting photon beam therapy. Report 62; supplement to report 50. Bethesda: International Commission Radiation Units and Measurements; 1999. [15] Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989;5:303–11. discussion 12–3. [16] Kwint M, Uyterlinde W, Nijkamp J, Chen C, de Bois J, Sonke JJ, et al. Acute esophagus toxicity in lung cancer patients after intensity modulated radiation therapy and concurrent chemotherapy. Int J Radiat Oncol Biol Phys 2012;84: e223–8. [17] Rodriguez N, Algara M, Foro P, Lacruz M, Reig A, Membrive I, et al. Predictors of acute esophagitis in lung cancer patients treated with concurrent threedimensional conformal radiotherapy and chemotherapy. Int J Radiat Oncol Biol Phys 2009;73:810–7. [18] Wijsman R, Dankers F, Troost EG, Hoffmann AL, van der Heijden EH, de GeusOei LF, et al. Multivariable normal-tissue complication modeling of acute esophageal toxicity in advanced stage non-small cell lung cancer patients
[19]
[20]
[21]
[22]
[23]
[24]
[25]
treated with intensity-modulated (chemo-)radiotherapy. Radiother Oncol 2015;117:49–54. Bar-Ad V, Leiby B, Witek M, Xiao Y, Cui Y, Dai Y, et al. Treatment-related acute esophagitis for patients with locoregionally advanced non-small cell lung cancer treated with involved-field radiotherapy and concurrent chemotherapy. Am J Clin Oncol 2013. Watkins JM, Wahlquist AE, Shirai K, Garrett-Mayer E, Aguero EG, Fortney JA, et al. Factors associated with severe acute esophagitis from hyperfractionated radiotherapy with concurrent chemotherapy for limited-stage small-cell lung cancer. Int J Radiat Oncol Biol Phys 2009;74:1108–13. Takeda K, Nemoto K, Saito H, Ogawa Y, Takai Y, Yamada S. Dosimetric correlations of acute esophagitis in lung cancer patients treated with radiotherapy. Int J Radiat Oncol Biol Phys 2005;62:626–9. Singh AK, Lockett MA, Bradley JD. Predictors of radiation-induced esophageal toxicity in patients with non-small-cell lung cancer treated with threedimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2003;55:337–41. Wei X, Liu HH, Tucker SL, Liao Z, Hu C, Mohan R, et al. Risk factors for acute esophagitis in non-small-cell lung cancer patients treated with concurrent chemotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2006;66:100–7. Ettinger DS, Akerley W, Borghaei H, Chang AC, Cheney RT, Chirieac LR, et al. Non-small cell lung cancer, version 2.2013. J Natl Compr Canc Netw 2013;11:645–53. quiz 53. Collman PI, Tremblay L, Diamant NE. The central vagal efferent supply to the esophagus and lower esophageal sphincter of the cat. Gastroenterology 1993;104:1430–8.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy Yi Pan a,b, Carsten Brink b,c, Marianne Knap d, Azza A. Khalil d, Christa H. Nyhus e, Tine McCulloch f, Bente Holm g, Yi-long Wu h, Tine Schytte i, Olfred Hansen b,i,⇑ a Department of Radiation Oncology, Guangdong General Hospital & Guangdong Academy of Medical Science, PR China; b Institute of Clinical Research, University of Southern Denmark, Odense; c Laboratory of Radiation Physics, Odense University Hospital; d Department of Oncology, Aarhus University Hospital; e Department of Oncology, Vejle Hospital, Sygehus Lillebaelt; f Department of Oncology, Aalborg University Hospital; g Department of Oncology, Herlev University Hospital, Denmark; h Guangdong Lung Cancer Institute, PR China; and i Department of Oncology, Odense University Hospital, Denmark
a r t i c l e
i n f o
Article history: Received 4 October 2015 Received in revised form 8 January 2016 Accepted 10 January 2016 Available online xxxx Keywords: Acute esophagitis Non small cell lung cancer IMRT Concurrent chemoradiotherapy
a b s t r a c t Purpose: Esophagitis is common in patients treated with definitive radiotherapy for local–regional advanced non small cell lung cancer (NSCLC). The purpose of this study was to estimate the dose–effect relationship using clinical and dosimetric parameters in patients receiving intensity modulated radiotherapy (IMRT) and concomitant chemotherapy (CCT). Methods: Between 2009 and 2013, 117 patients with stages IIB–IIIB NSCLC were treated in a multicenter randomized phase II trial with 2 cycles of induction chemotherapy followed by IMRT and CCT. The esophagitis was prospectively scored using the Common Toxicity Criteria 3.0. Clinical and dosimetric variables were analyzed for the correlation with grade P2 esophagitis through logistic regression. Results: Grade 2 esophagitis was experienced by 31 (27%). All models including gender, institution, a dosimetric parameter and a position parameter were significantly associated with esophagitis. The two models using the relative esophagus volume irradiated above 40 Gy (V40, OR = 2.18/10% volume) or the length of esophagus irradiated above 40 Gy (L40, OR = 4.03/5 cm) were optimal. The upper part of esophagus was more sensitive and females experienced more toxicity than men. Conclusion: V40 and L40 were most effective dosimetric predictors of grade P2 esophagitis. The upper part of esophagus was more sensitive. Ó 2016 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2016) xxx–xxx
Thirty-five percent of patients with non small cell lung cancer (NSCLC) are irresectable at diagnosis due to being in locoregional advanced stage [1]. Definitive concomitant chemoradiotherapy (CCRT) is the standard of care for this group of patients if they are in good performance status. Overall survival is improved by 5% by CCRT compared to sequential chemoradiotherapy, which may be caused by an increased local–regional control [1]. Although CCRT has a better therapeutic effect, the risk and severity of acute toxicity, especially the acute esophagitis may increase. In a metaanalysis performed by the NSCLC Collaborative Group containing 6 randomized clinical trials, the incidence of grade 3–4 esophagitis in CCRT (18%) was higher than observed in sequential chemoradiotherapy (4%), with a relative ratio of 4.9 (p < 0.01) [1]. For patients treated with radiotherapy alone, the incidence of esophagitis Pgrade 3 is 1%, but it is markedly increased to 6–24% with CCT [2]. ⇑ Corresponding author at: Department of Oncology, Odense University Hospital, DK-5000 Odense, Denmark. E-mail address: [email protected] (O. Hansen).
The symptoms of esophagitis include retrosternal pain, dysphagia and odynophagia. It often occurs 2–4 week after the start of radiation therapy and may persist 2–7 weeks [3,4]. Severe esophagitis may induce hospitalization, and require parenteral nutrition, naso-gastric tube and analgetics. It may lead to treatment interruptions, which may have a negative impact on the quality of life of the patients, lower the chance of local tumor control, and thus impact survival negatively. Intensity modulated radiotherapy (IMRT) facilitates a more conform radiation dose distribution allowing for shaping the dose of the tumor and sparing esophagus more than 3-dimensional radiation (3DCRT) may be able to do. For NSCLC, IMRT has the potential to escalate target dose without a corresponding increase to the lung and esophagus [5–7]. In many centers IMRT has become the principal radiation technique for NSCLC to reduce local relapse and protect critical organ at risk (OAR). Although a number of clinical and dosimetric risk factors for esophagitis have been reported in previous studies [8–13], no dosimetric parameter has been generally accepted as the best predictor. In addition, most of these studies were based on 3DCRT. The
http://dx.doi.org/10.1016/j.radonc.2016.01.007 0167-8140/Ó 2016 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
2
Esophagitis in IMRT and CCT for NSCLC
objective of this study was to estimate the dose–effect relation between esophagitis and clinical as well as dosimetric parameters in patients with local–regional advanced NSCLC receiving IMRT and CCT. Materials and methods Patients Between May 2009 and August 2013, 117 patients with NSCLC were enrolled in the Danish multicenter randomized phase II trial (NARLAL). The inclusion criteria were histologically or cytologically confirmed NSCLC, clinical American Joint Committee of Cancer stage (the seventh edition) IIB–IIIB, performance status (PS) <2 on the ECOG scale, weight loss 610%, adequate hepatic and renal function, forced expiratory volume in 1 s (FEV1) P1.0. Patients with pleural effusion, prior chemotherapy for lung cancer and any other active malignancy within the last 5 years were excluded. Each patient underwent basic laboratory studies, computed tomography (CT) scan of the chest and upper abdomen with contrast and whole body positron emission tomography (PET) before treatment. Approval was granted by the Institutional Review Board for conducting this study. The NARLAL clinic cohort study was filed at clinicaltrials.gov (No. NCT00887783). The medium age was 66 years (44–82 years). The patient characteristics are presented in Table 1. NARLAL protocol All patients were treated with induction chemotherapy followed by CCRT. The induction chemotherapy consisted of 2 cycles of Carboplatin (AUC = 5, day 1) and Vinorelbine oral (60–80 mg/m2, day 1 and 8). After completing induction chemotherapy, patients were randomized into two groups: 60 Gy in 30 fractions (Arm A) and 66 Gy in 33 fractions (Arm B), once daily, 5 times per week. The concomitant chemotherapy regimen consisted of Vinorelbine oral 50 mg 3 times a week for 6—6 12 weeks. Eighty-three percent of the patients received the scheduled doses of concomitant chemotherapy (Supplementary Table 1). Radiotherapy preparation IMRT was planned for all patients. Gross tumor volume including the lung tumor and involved mediastinal lymph node was defined on CT of the chest and PET-CT. A treatment planning CT scan was performed with 2.5–3 mm slice thickness from upper neck to mid-abdomen. Radiotherapy was delivered with a linear accelerator using energy of 6 MV X-rays. The planning technique was based on ICRU62 recommendations [14]. Critical organ dose tolerances were defined as maximum dose of esophagus 6 66 Gy, lung V20 6 40%, spinal cord 6 45 Gy, heart V50 6 20%. Esophageal contours and dosimetric data For the consistency, the esophagus was re-delineated in all patients by the same oncologist (Y.P.). The external surface of esophagus was contoured on each axial plane of the planning CT scan from the inferior margin of cricoid cartilage to the gastroesophageal junction. Dose-volumetric values were calculated from Dose Volume Histogram (DVH) for each patient. The following esophageal dosimetric parameters were extracted from the dose data set in treatment planning system: maximum, median and mean dose, the entire length and volume, series of dosevolumetric values from DVH such as: the percentage of the volume
Table 1 Patient characteristics (N = 117). Characteristics
Number
Percentage
Gender Male Female
68 49
58 42
Age (years) Median (range)
65.5 (44–82)
Smoking Never and former smokers Current smoker
89 28
76 24
Histology SCC NSCC
47 70
40 60
Performance status 0 1
62 55
53 47
T stage T1 + 2 T3 + 4
74 43
63 37
N stage N0–1 N2 N3
18 85 14
15 73 12
Stage IIB IIIA IIIB
6 81 30
5 69 26
Prescribed Radiation Dose (Gy) 66 60
57 60
49 51
Length of esophagus (cm) Median (range)
26 (20–31)
Volume of esophagus (cm3) Median (range)
35 (19–141)
Mean esophagus dose (Gy) Median (range)
23 (3–42)
Maximum esophagus dose (Gy) Median (range)
63 (9–72)
Median esophagus dose (Gy) Median (range)
10 (1–60)
Esophagus V20 (%) Median (range)
43 (0–70)
Esophagus V30 (%) Median (range)
36 (0–70)
Esophagus V40 (%) Median (range)
32 (0–68)
Esophagus V50 (%) Median (range)
26 (0–66)
Esophagus V60 (%) Median (range)
8 (0–50)
Abbreviations: SCC = squamous cell carcinoma; NSCC = non squamous cell carcinoma; V20–V60: percentage of esophageal volume receiving at least 20–60 Gy.
of esophagus receiving from P20 Gy (V20) and in 10 Gy increments up to P60 Gy (V60). Based on the hypothesis that length and positional information, which is discarded in traditional DVH, might be important to predict grade 2 or greater esophagitis, the following metrics were calculated: the length of the part of esophagus for which the median dose within a transversal CT slices was above 20 Gy, 30 Gy, 40 Gy, 50 Gy, or 60 Gy (L20, L30, L40, L50, or L60) and mean relative position of the volume related to L20–L60 (P20, P30, P40, P50, and P60) defined as 0 at the most cranial located esophagus delineation and 1 at the most caudal located delineation.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
3
Y. Pan et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx
Follow-up and esophagitis scoring Acute esophagitis was defined according to version 3.0 of National Cancer Institute’s Common Toxicity Criteria for Adverse Events version 3.0 (NCI-CTCAE 3.0) from the start of RT until 3 months afterward. During RT, patients underwent evaluation for side effects once a week and more often if interventions were indicated. After RT, follow up monitoring was at regular intervals. First time was 30–45 days after radiation and then every 3 month. The clinical esophagitis grades were available for all patients. Statistics analyses/normal tissue complication probability (NTCP) modeling Patients and treatment characteristics were summarized by descriptive statistics. Firstly, univariate analyses were used to identify the significant risk factors among gender, age, institution, smoking status, PS, histology, tumor dose, and dosimetric parameters including maximum dose, median dose, mean dose, volumetric metrics, length and position metrics. Due to large inter variable correlation within the dosimetric variables and within the position variables, only one of the positions and one of the dosimetric variables were included in pairs for each multivariable analysis. Backward selection was performed in the binary logistic regression analyses. Akaike information criterion (AIC) was used for model selection, which balances goodness of fit with the complexity of the model. AIC is defined as AIC = 2 * k 2 * ln(L), where k is the number of parameters in the model and L is the maximized value of the likelihood function. The preferred model was taken as the one resulting in the lowest AIC value. Model calibration was estimated by classifying patients into 6 risk groups according to predicted probability of esophagitis and plotting the mean observed probability versus the mean predicted probability for each group. Goodness of fit was estimated by the Hosmer–Lemeshow test. To further validate the model performance area under the receiver operating curve (AUC), Brier score, and cross validation of the models by bootstrap was performed and is presented as Supplementary material. To validate the impact of the fixed dose of Vinorelbine as concurrent chemotherapy, the body surface area (BSA) was also tested within the models. BSA is defined as BSA = 0.007184 * weight0.425 * height0.725 [15]. Value of p < 0.05 was considered statistically significant. Statistical analysis was performed with SPSS (Statistics Package for Social Science). No patient was lost to follow-up.
Fig. 1. Incidence of acute esophagitis.
maximum dose, mean dose, median dose, all relative volumetric variables (V20, V30, V40, V50, and V60), absolute length (L20, L30, L40, L50, or L60) and position variables (P20, P30, P40, P50, and P60) were all significant. An Anova analysis did not show any associations between dosimetric/position variables and gender/ institution. Each model in binary logistic regression contained 4 independent variables: gender, institution, one position parameter, and one dosimetric parameter including maximum dose, mean dose, median dose, volumetric (V20–V60), and length metrics (L20–L60). All test models demonstrated a significant correlation between esophagitis and the predictor variables. The models with mean position for the length of esophagus receiving P20 Gy (P20) variable had lower AIC values than models with P30–P60. Hence, P20 was selected to estimate the effects of irradiation of specific parts of the esophagus (relative position along esophagus). Table 2 shows the AIC of logistic models with gender, institution, P20 and one dosimetric parameter. For grade P2 esophagitis, the models containing V40 and V50 had the lowest AIC value. The AIC value for V40 was only slightly lower than the corresponding value for V50. L40 was very close to V40 and V50 and was further investigated since L40 is conceptual different from V40 in term of the need of delineating the entire esophagus. Table 3 shows the results of multivariable analysis. Female, longer length of esophagus receiving doses above 40 Gy, large relative volume of esophagus receiving 40 Gy or more, and
Results Of 117 patients treated with IMRT and CCT, 91 (78%) patients developed esophagitis including 50 (43%) patients with grade 1 esophagitis, 31 (27%) with grade 2 and 10 (8%) with grade 3. No grade 4 or 5 esophagitis was observed (Fig. 1). The medium onset time was 18 days (1–44 days) after the start of the RT. The incidences of grade P2 esophagitis in 5 institutions were 33% (12/36) in OUH, 40% (8/20) in Vejle Hospital, 8% (1/12) in Aalborg Hospital, 46% (19/41) in Aarhus Hospital, and 13% (1/8) in Herlev hospital, respectively. For the total study population, the average length of esophagus was 26 cm (20–31 cm). The dosimetric data of the 117 patients are shown in Table 1. Among the clinical parameters: gender, age, institution, smoking status, performance status (PS), histology, tumor dose, gender was the only significant factor for grade P2 esophagitis (p = 0.001, Supplementary Table 2). The association between esophagitis and T, N, and clinical stage were not analyzed since they were highly associated with dosimetric parameters which in turn correlate with esophagitis. The dosimetric parameters,
Table 2 Akaike information criterion of the models. Dosimetric parameters
AIC
Mean dose Maximum dose Median dose L20 L30 L40 L50 L60 V20 V30 V40 V50 V60
110.902 113.296 122.52 114.811 113.438 109.8648a 111.825 119.02 115.312 112.243 108.844a 109.313a 116.937
Abbreviations: AIC = Akaike information criterion. All models included 4 variables: gender, institution, P20 and one of above dosimetric parameters. a Minimum Akaike information criterion.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
4
Esophagitis in IMRT and CCT for NSCLC
Table 3 Results of multivariate analyses.
Model A
Model B
Variable
OR
95% CI
p-Value
Constant Female gender L40 (per 5 cm) P20 (per 0.1 relative unit) Institution
0.38 12.74 4.03 0.52
3.99–40.64 1.84–8.79 0.29–0.95
0.598 <0.001 <0.001 0.032 0.020
Constant Female gender V40 (per 10% relative volume) P20 (per 0.1 relative unit) Institution
0.91 12.33 2.18 0.44
3.89–39.06 1.44–3.30 0.24–0.81
0.957 <0.001 <0.001 0.008 0.011
Abbreviations: OR = odds ratio; CI = confidence interval; L40 = length of 50% esophagus within a slice receiving P40 Gy; P20 = mean position for the length of which 50% esophagus within a slice received P 20 Gy.
cranial position of esophageal receiving more than 20 Gy significantly increased the risk of grade P2 esophagitis. BSA was not significant in a multivariable analysis neither together with the gender variable, nor within the two genders separately. Model performances for the V40 and L40 model is illustrated in Fig. 2, comparing the mean predictive probability with the mean observed probability of esophagitis. The p values for goodness-of-fit were 0.440 and 0.410 – reflecting the relative good correlations seen Fig. 2. AUC, Brier score, and cross validation of models are shown in Supplementary Table 3. As a result of multivariable logistic regression analysis, the NTCP model of developing grade P2 esophagitis was expressed as:
NTCP ¼ ð1 þ expðSÞÞ1 where S was defined as:
S ¼ 0:97 þ 2:55 Sex þ 0:28 cm1 L40 6:45 P20 þ b4 Institution S ¼ 0:09 þ 2:51 Sex þ 7:79 V40 8:19 P20 þ b4 Institution Sex: female = 1, male = 0; P20: cranial position = 0, caudal position = 1; V40: 0% = 0, 100% = 1. The constant term in the S equation (0.97 and 0.09) depended on which institution is used as reference for calculation of b4. As standard one of the institutions was reference institution during the binary regression, the b4 values for the remaining institutions was ranging from 2.4 to 1.4 and 2.5 to 1.6 for the L40 and V40 models, respectively.
Stratification for institutions was performed in order to correct for differences in reporting of the toxicity level. Without stratification all the parameters in the above models were still significant except for P20 in the L40 model in which the p-value was 12%. To summarize: larger V40, longer L40, the upper part of esophagus and female gender were significant risk factors for developing grade 2 or greater acute esophagitis. Discussion Numerous dosimetric parameters have previously been investigated for the association with acute esophagitis. However, these results were mostly based on 3DCRT and conventional radiotherapy. Only few studies reported esophagitis in patients treated with IMRT and the same CCT regimen until now [16,18]. As IMRT has been introduced as a treatment option, previous data has not been sufficient to estimate esophageal toxicities in IMRT. Kwint et al. showed that historical NTCP model was unsuitable to predict esophagitis in IMRT [16]. Furthermore, some parameters make no sense in IMRT such as length of esophagus included in the radiation field. With the use of IMRT in patients with NSCLC, further study was needed to get a better model of prediction. CCRT were reported to result in a 29–37% rate of grade 2 acute esophagitis and a 4% and 20% rate of grade 3 acute esophagitis [3,17,18]. In our study, the incidences of grade 2 and 3 esophagitis were 27% and 8%, respectively, which are similar to those reported by Rodriquez et al. [17] and Wijsman et al. [18]. Kwint et al. reported higher incidences of esophagitis in IMRT (38% and 22%) [16]. One possible reason is that the radiation dose per fraction was 2.75 Gy (66 Gy in 24 fractions) in their study, whereas it was only 2 Gy in our study [19]. Within the previous published studies there is no consensus on a possible association between gender and esophagitis. A number of studies report not to be able to detect a statistically significant association between gender and esophagitis [12,20]. However, Dehing-Oberije et al. [8] and Wijsman et al. [18] found as in the current study, a statistically significant correlation between gender and esophagitis although with less OR than in the current study. Since the concomitant chemotherapy regimen in our study was a fixed dose of Vinorelbine, the high OR for gender could be somewhat confounded by the BSA, which is less for the female gender. However, confounder analyses were not able to support the existence of a confounding effect. Previous published data have demonstrated that volumes receiving P20–80 Gy (V20–V80) is significantly correlated with esophagitis [8–10], especially V30–V50 [9,10,21]. Our results showed V40 and V50 were most strongly associated with grade
Fig. 2. Calibration plot of L40 model (A) and V40 model (B) for grade P2 esophagitis in NSCLC patients treated with IMRT and CCT, illustrating the relation between the predicted and observed probability. Patients have been grouped by increasing predicted probability.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
Y. Pan et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx
P2 esophagitis in NSCLC patients treated with IMRT compared to other volumetric parameters. Their values on predicting esophagitis were almost equivalent. Based on the AIC values V40 might be slightly better than V50. In some studies, length parameters have been found to be significant in univariate analysis, but not in multivariable analysis [11–13]. In our study, length of esophagus within a slice receiving P40 Gy (L40) was identified to be a strong predictor for grade P2 esophagitis. Comparing the models with V40 and L40, it is difficult to decide which logistic model is the better to predict esophagitis as AIC of the two models were very close. It is very convenient for clinicians to estimate esophagitis by volume metrics as they can be directly obtained through dosevolume histogram in the treatment planning systems while this is not easily done with length metrics. However, V40 is the relative volume of esophagus and it depends on delineation of how much of the esophagus is contoured. By contrast, L40 is the absolute esophageal length and independent of esophagus delineation within regions of low dose. Which parameter to use, V40 or L40, depends on the clinicians’ choice of the accessibility to the information and the robustness against delineation inaccuracies. The maximum dose, mean dose, and V60 were reported to have a significantly association with esophagitis in patients treated with 3DCRT [12,22,23]. Mean dose recommendation was adopted in the National Comprehensive Cancer Network (NCCN) guidelines [24]. The univariable dose results from Wijsman et al. are in close agreement with the univariable result in the current study (Supplementary Table 2). However, very strong correlations exist between many of the dose parameters, and a model reduction needs to be performed for a multivariable analyses. Wijsman et al.’s chose the mean dose to be the only dose parameter used within their multivariable models in their retrospective study of 149 NSCLC IMRT treated patients [18]. Their choice using the mean was based on LKB result with a volume parameter close to the value one. The present study uses the statistical concept Akaike information criterion (AIC) to select between the multivariable models including the different dose parameters. The current results indicate that logistic models with maximum dose, mean dose, or V60 were not the optimal choice for evaluating grade P2 esophagitis, even though these dose parameters in 3D CRT studies and in Wijsman’s IMRT study have shown associations with esophagitis. IMRT is likely to have as lower association between dose to target and to normal tissue as 3DCRT [5] and may cause a difference between IMRT and 3D CRT toxicity models. It may therefore be difficult to extrapolate normal tissue complication probability models based on 3DCRT treatment to treatments based on IMRT. Furthermore, it is hard to identify the association between esophagitis and dosimetric parameters in multivariable analysis when these parameters are highly inter-correlated. The inter-correlation is likely stronger in 3D CRT studies than in IMRT studies. In the current study, caudal position of esophagus irradiated above 20 Gy decreased the risk of grade P2 esophagitis. To the best of our knowledge, this is the first time position parameter is reported as a predictive factor for esophagitis. One possible reason is the distributions of sensory nerves vary in upper, middle, and lower esophagus [25]. Sensory axons are abundant in cervical esophagus and progressively spares in thoracic and abdominal esophagus. Patients may be more sensitive to injuries in the upper part of esophagus than in the middle and lower part and have more severe symptoms. In this multicenter study, the incidence of grade P2 esophagitis in the participating institutions was different and institution was a significant factor to esophagitis in multivariable analysis. At present, the scoring system which is most commonly used for assessment of esophagitis is CTC scale. It evaluates the occurrence and severity of esophagitis mainly by symptoms reported by patients and treatments used by doctors, which are convenient for clinical
5
management. However, patients’ subjective perception and doctors’ judgment vary from person to person. This may result in overestimation or underestimation of the incidence of esophagitis. This shows that evaluation of toxicity is inherently difficult and calls on continuous education and discussion of how to score toxicity as precise as possible. Conclusion This study demonstrated in a cohort of patients with NSCLC receiving IMRT and concurrent chemotherapy that V40 and L40 were the most effective dosimetric predictors of grade 2 or greater acute esophagitis. The application of V40 or L40 depends on clinicians’ preference for either the accessibility to the information or robustness against delineation inaccuracies. Mean relative position of the part of esophagus within a slice receiving P20 Gy (P20) is found to be significantly correlated with acute esophagitis and should be included when modeling acute esophageal toxicity. In IMRT, maximum dose and mean dose was not as informative as they are in conventional radiotherapy and 3DCRT. While the current findings present good performance of the models, OR for gender may be somewhat confounded as fixed dose CCT was used in our study. Validation in a large independent patient cohort is needed to confirm this result. Conflict of interest statement No conflict of interest. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2016.01. 007. References [1] Auperin A, Le Pechoux C, Rolland E, Curran WJ, Furuse K, Fournel P, et al. Metaanalysis of concomitant versus sequential radiochemotherapy in locally advanced non-small-cell lung cancer. J Clin Oncol 2010;28:2181–90. [2] Werner-Wasik M, Yorke E, Deasy J, Nam J, Marks LB. Radiation dose-volume effects in the esophagus. Int J Radiat Oncol Biol Phys 2010;76:S86–93. [3] Uyterlinde W, Chen C, Kwint M, de Bois J, Vincent A, Sonke JJ, 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. [4] Werner-Wasik M, Pequignot E, Leeper D, Hauck W, Curran W. Predictors of severe esophagitis include use of concurrent chemotherapy, but not the length of irradiated esophagus: a multivariate analysis of patients with lung cancer treated with nonoperative therapy. Int J Radiat Oncol Biol Phys 2000;48:689–96. [5] Grills IS, Yan D, Martinez AA, Vicini FA, Wong JW, Kestin LL. Potential for reduced toxicity and dose escalation in the treatment of inoperable non-smallcell lung cancer: a comparison of intensity-modulated radiation therapy (IMRT), 3D conformal radiation, and elective nodal irradiation. Int J Radiat Oncol Biol Phys 2003;57:875–90. [6] Schwarz M, Alber M, Lebesque JV, Mijnheer BJ, Damen EM. Dose heterogeneity in the target volume and intensity-modulated radiotherapy to escalate the dose in the treatment of non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2005;62:561–70. [7] Nielsen TB, Hansen O, Schytte T, Brink C. Inhomogeneous dose escalation increases expected local control for NSCLC patients with lymph node involvement without increased mean lung dose. Acta Oncol 2014;53:119–25. [8] Dehing-Oberije C, De Ruysscher D, Petit S, Van Meerbeeck J, Vandecasteele K, De Neve W, et al. Development, external validation and clinical usefulness of a practical prediction model for radiation-induced dysphagia in lung cancer patients. Radiother Oncol 2010;97:455–61. [9] Rose J, Rodrigues G, Yaremko B, Lock M, D’Souza D. Systematic review of dosevolume parameters in the prediction of esophagitis in thoracic radiotherapy. Radiother Oncol 2009;91:282–7. [10] Kim TH, Cho KH, Pyo HR, Lee JS, Han JY, Zo JI, et al. Dose-volumetric parameters of acute esophageal toxicity in patients with lung cancer treated with threedimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2005;62:995–1002.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007
6
Esophagitis in IMRT and CCT for NSCLC
[11] Belderbos J, Heemsbergen W, Hoogeman M, Pengel K, Rossi M, Lebesque J. Acute esophageal toxicity in non-small cell lung cancer patients after high dose conformal radiotherapy. Radiother Oncol 2005;75:157–64. [12] Ahn SJ, Kahn D, Zhou S, Yu X, Hollis D, Shafman TD, et al. Dosimetric and clinical predictors for radiation-induced esophageal injury. Int J Radiat Oncol Biol Phys 2005;61:335–47. [13] Rosenman JG, Halle JS, Socinski MA, Deschesne K, Moore DT, Johnson H, et al. High-dose conformal radiotherapy for treatment of stage IIIA/IIIB non-smallcell lung cancer: technical issues and results of a phase I/II trial. Int J Radiat Oncol Biol Phys 2002;54:348–56. [14] Baggins B. Prescribing, recording and reporting photon beam therapy. Report 62; supplement to report 50. Bethesda: International Commission Radiation Units and Measurements; 1999. [15] Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989;5:303–11. discussion 12–3. [16] Kwint M, Uyterlinde W, Nijkamp J, Chen C, de Bois J, Sonke JJ, et al. Acute esophagus toxicity in lung cancer patients after intensity modulated radiation therapy and concurrent chemotherapy. Int J Radiat Oncol Biol Phys 2012;84: e223–8. [17] Rodriguez N, Algara M, Foro P, Lacruz M, Reig A, Membrive I, et al. Predictors of acute esophagitis in lung cancer patients treated with concurrent threedimensional conformal radiotherapy and chemotherapy. Int J Radiat Oncol Biol Phys 2009;73:810–7. [18] Wijsman R, Dankers F, Troost EG, Hoffmann AL, van der Heijden EH, de GeusOei LF, et al. Multivariable normal-tissue complication modeling of acute esophageal toxicity in advanced stage non-small cell lung cancer patients
[19]
[20]
[21]
[22]
[23]
[24]
[25]
treated with intensity-modulated (chemo-)radiotherapy. Radiother Oncol 2015;117:49–54. Bar-Ad V, Leiby B, Witek M, Xiao Y, Cui Y, Dai Y, et al. Treatment-related acute esophagitis for patients with locoregionally advanced non-small cell lung cancer treated with involved-field radiotherapy and concurrent chemotherapy. Am J Clin Oncol 2013. Watkins JM, Wahlquist AE, Shirai K, Garrett-Mayer E, Aguero EG, Fortney JA, et al. Factors associated with severe acute esophagitis from hyperfractionated radiotherapy with concurrent chemotherapy for limited-stage small-cell lung cancer. Int J Radiat Oncol Biol Phys 2009;74:1108–13. Takeda K, Nemoto K, Saito H, Ogawa Y, Takai Y, Yamada S. Dosimetric correlations of acute esophagitis in lung cancer patients treated with radiotherapy. Int J Radiat Oncol Biol Phys 2005;62:626–9. Singh AK, Lockett MA, Bradley JD. Predictors of radiation-induced esophageal toxicity in patients with non-small-cell lung cancer treated with threedimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2003;55:337–41. Wei X, Liu HH, Tucker SL, Liao Z, Hu C, Mohan R, et al. Risk factors for acute esophagitis in non-small-cell lung cancer patients treated with concurrent chemotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2006;66:100–7. Ettinger DS, Akerley W, Borghaei H, Chang AC, Cheney RT, Chirieac LR, et al. Non-small cell lung cancer, version 2.2013. J Natl Compr Canc Netw 2013;11:645–53. quiz 53. Collman PI, Tremblay L, Diamant NE. The central vagal efferent supply to the esophagus and lower esophageal sphincter of the cat. Gastroenterology 1993;104:1430–8.
Please cite this article in press as: Pan Y et al. Acute esophagitis for patients with local–regional advanced non small cell lung cancer treated with concurrent chemoradiotherapy. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.01.007