A prospective study on radiation pneumonitis following conformal radiation therapy in non-small-cell lung cancer: clinical and dosimetric factors analysis

Radiotherapy and Oncology 71 (2004) 175–181 www.elsevier.com/locate/radonline

A prospective study on radiation pneumonitis following conformal radiation therapy in non-small-cell lung cancer: clinical and dosimetric factors analysisq Line Claudea, David Pe´rolb, Chantal Ginesteta, Lionel Falcheroc, Dominique Arpind, Michel Vincente, Isabelle Martela, Ste´phane Hominalf, Jean-Franc¸ois Cordierg, Christian Carriea,* a

Department of Radiation Oncology, Centre Le´on Be´rard-28, rue Lae¨nnec 69373 Lyon Cedex 08, France b Department of Public Health (UBET), Centre Le´on Be´rard, Lyon, France c Department of Respiratory Medicine, Centre Hospitalier, Villefranche/s, France d Department of Respiratory Medicine, Hoˆpital des Chanaux, Maˆcon, France e Department of Respiratory Medicine, Hoˆpital Saint Joseph, Lyon, France f Department of Respiratory Medicine, Centre Hospitalier de la Croix-Rousse, Lyon, France g Department of Respiratory Medicine, Hoˆpital Louis Pradel, Lyon, France Received 27 June 2003; received in revised form 27 January 2004; accepted 5 February 2004

Abstract Background and purpose: Clinical and dosimetric prognostic factors for radiation pneumonitis (RP) have been reported after threedimensional conformal radiotherapy (3D-CRT) in patients with non-small cell lung cancer (NSCLC). Patients and methods: Ninety-six patients who received 3D-CRT for stage IA to IIIB NSCLC were evaluated prospectively. Surgery was performed before radiation in 51% of the patients ðn ¼ 49Þ: RP was diagnosed six-eight weeks after 3D-CRT using the Lent-Soma classification. Factors evaluated included treatment factors such as total mean lung dose (MLD), and dose-volume histogram (DVH) thresholds for several radiation dose steps. These thresholds were originally determined from the median of the irradiated lung volume at each step. Results: Six patients could not be evaluated for RP six weeks after 3D-CRT. Of the 90 remaining patients, 40 (44%) had RP (i.e. grade $ 1) at 6 weeks, including 7 patients (7.8%) with severe RP (grade $2). Regarding the whole toxicity (grade $ 1), age ($ 60 years), MLD, V20 and V30 were significantly related to RP. DVH thresholds determined for radiation doses from 20 to 40 Gy were also predictive of RP. Considering only severe RP (grade $ 2), only MLD, V20 and V30 remained associated with increased acute pulmonary toxicity. Conclusions: In this study, dosimetric factors (MLD, V20, V30) and age ($ 60 years) were predictive of RP regarding the whole pulmonary toxicity (grade $1). In addition, thresholds from 20 to 40 Gy, based on a stratification according to the median of the percentage of irradiated lung volume, were also predictive factors. They may, therefore, help discriminate patients at high and low risk for RP. However, only MLD, V20 and V30 remained associated with severe RP (grade $ 2), probably due to the small number of severe events in our series. q 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Non-small cell lung cancer; Conformal radiation therapy; Dose-volume-histogram; Radiation pneumonitis

1. Introduction Radiotherapy plays a major role in the treatment of nonsmall cell lung cancer (NSCLC), either as adjuvant therapy after surgery or as radical local therapy in patients with q Supported by grants from the French Health Ministry (PHRC No 2701), the Ligue Contre le Cancer de l’Ain (France) and the Ligue Contre le Cancer du Rhoˆne (France). * Corresponding author.

0167-8140/$ - see front matter q 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2004.02.005

unresectable cancers. High radiation doses are correlated with improved local control [18]. However, radiationinduced pulmonary complications, in particular acute radiation pneumonitis (RP) are common side-effects that can lead to chronic respiratory insufficiency and sometimes death. These complications are strongly correlated with the dose delivered to normal tissues. Specific dose-volume histogram (DVH) parameters were previously reported to be risk factors for radiotherapy-related pneumonitis in NSCLC

176

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181

patients treated with conformal radiation therapy (3DCRT). The percentage of irradiated lung volume exceeding, in particular, 20 and 30 Gy (V20 and V30) has been identified as a useful parameter to reduce the risk of radiation-related pneumonitis [6,7]. In this study, 3D-CRT was performed according to previously published data on V20 and V30, i.e. less than 30 and 20% of irradiated lung volume exceeding 20 and 30 Gy, respectively, as often as possible [6,7]. Under these conditions, we prospectively report the incidence of acute RP using the Lent-Soma classification, as well as predictive factors for RP.

2. Methods and treatment 2.1. Eligibility From November 1996 to September 2000, 96 patients with histologically proven NSCLC were enrolled in a prospective study to evaluate pulmonary consequences following 3D-CRT. Patients were eligible for the study if they had histologically or cytologically confirmed diagnosis of non metastatic NSCLC, if they were at least eighteen years old, had a Karnofsky performance status (KPS) $ 60%, and a life expectancy of at least six months. 3D-CRT with curative intent was delivered in all patients. Pulmonary function tests (PFTs) involved the determination of: The ratio of forced expiratory volume in one second (FEV1) to the vital capacity (VC) $ 50% The diffusion capacity for carbon monoxide (DLCO) or a transfer coefficient for carbon monoxide (KCO) $ 50% of predicted. The patients had received no radiotherapy prior to inclusion. PFTs were performed immediately before 3DCRT in the event of surgery, or 28 days (range 4– 134) before 3D-CRT if surgery was not performed. Written informed consent was obtained from all the patients before inclusion in the study. 2.2. Conformal radiation therapy 2.2.1. Treatment modalities Patients were immobilised in a customised alpha cradle with their arms supported above their head for planning computed tomography (CT) scan (5-mm spacing) and treatment. Patients received conventional fractionated radiation therapy (2 Gy per fraction, 5 days per week). The total irradiation dose ranged from 46 to 72 Gy, with a median of 66 Gy. The use of electrons concerned only the supraclavicular field. For this reason, the contribution of the energy delivered by the electrons to the lungs was considered as insignificant and was not taken into account in DVH generation. For supraclavicular radiation,

X-ray energy of 10 – 18 MV and electron energy of 9– 12 MV were used. Target volumes were defined using the ICRU-50 [9]. Gross tumour volume (GTV), defined as disease visible on CT, was contoured on each CT slice. When no surgery was performed, it included the tumour and involved nodes. After surgery, GTV included the residual tumour when appropriate, and/or initial macroscopic or histologically involved nodes. Clinical target volume (CTV) 1 was obtained by 5-mm 3D expansion from the GTV and the ipsilateral hilum. The ipsilateral mediastinum and supraclavicular fossa were included if the tumour was stage II or III and located in an upper lobe. CTV2 corresponded to the GTV without margin, encompassing all identifiable disease (tumour and nodal involvement if surgery was not performed; areas with initial macroscopic or histological lymph node involvement after surgery). Planning target volume (PTV) 1 and PTV2 were obtained by 5-mm 3D expansion from CTV1 and CTV2. We did not include any specific margin for breathing. The first part of radiation (PTV1) used six portal entrances (A/P, P/A and 4 oblique beams) for a total prescribed dose of 50 Gy. An additional dose of 16 –22 Gy was prescribed on PTV2 using six portal entrances (4 other oblique beams and 2 lateral beams). Multileaf collimation fields were used. Six portals were treated each day. Beam’s eye views were used to control the position of the isocenter collimator and leaves at the beginning of treatment and compared with digitally reconstructed radiography. Orthogonal X-rays were performed weekly. Corrections were applied and confirmed by a physicist when the deviation was more than 5 mm. 2.2.2. DVH parameters The structures of interest such as GTV, CTV and normal structures were contoured on the multiple CT pictures. The lungs were automatically contoured, excluding the GTV. Doses were calculated to reflect tissue density heterogeneity, and the DVHs of the lungs (treated as a combined paired organ) were calculated based on CT-defined lung volumes. Lung DVHs were calculated from the physical dose distribution (Fig. 1). Mean lung dose, V20, V30 were calculated from the lung DVHs. Total mean lung dose (MLD) was calculated from the lung DVHs using: MLD ¼ [(right lung volume £ mean dose to right lung) þ (left lung volume £ mean dose to left lung)]/(left lung volume þ right lung volume). In addition, irradiated lung volumes at 10 to 50 Gy (with 10-Gy steps) were extracted from the lung DVHs for each patient. 3D-CRT was performed according to previously published data on V20 and V30, i.e. less than 30 and 20% of irradiated lung volume exceeding 20 and 30 Gy, respectively [6,7].

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181

Fig. 1. Lung dose-volume histogram.

2.3. Concomitant chemotherapy Twenty-four of the 96 patients (25%) received a cisplatin-based chemotherapy regimen during radiation. 2.4. Evaluation of radiotherapy-related pneumonitis The severity of the radiotherapy-related pneumonitis was determined six weeks after the end of 3D-CRT using the Lent-Soma scale defined by the Radiation Therapy Oncology Group (RTOG) and the European Organization for the Research and Treatment of Cancer (EORTC) [14]. RP was scored on clinical symptoms, radiological abnormalities and loss of pulmonary function. This includes three subjective scales and two objective scales: Subjective scales: (1) cough; (2) dyspnea; (3) thoracic pain. Objective scales: (1) chest-X ray read by an independent panel of experts (pneumologists, radiologists and radiation oncologists); (2) PFTs (reduction of VC and/or DLCO). All of the single-scale measures ranged in score from grade 0 to 4. The final pneumonitis grading was equal to an average of the five scores. For example, if one patient was scored: cough: grade 1, dyspnea: grade 1, thoracic pain: grade 0, chest-X ray: grade 2, PFTs (reduction of VC and/or DLCO): grade 1, his final grade was (1 þ 1 þ 0 þ 2 þ 1)/5 ¼ 1. RP was defined as the development of a pulmonary toxicity of grade $ 1. Toxicity was considered as severe if the grade was $ 2. 2.5. Statistical analysis Risk factors for RP were tested in univariate and multivariate analyses using the SPSS statistical software

177

package (SPSS Inc., IL). The relationships between any clinical or treatment parameter and the incidence of RP were analysed using and the Pearson’s x 2 test or Fischer’s exact test for categorical variables, and the nonparametric Mann – Whitney exact test for quantitative variables, as appropriate. A logistic regression analysis, including the parameters studied in the univariate analysis, was performed using a backward regression procedure with a P value # 0.10 for entry. Several clinical factors, i.e. gender, age, KPS, smoking status, tumour stage, previous surgery, previous chemotherapy, PFTs at baseline (VC, KCO, ratio of FEV1 to VC) were tested. Since surgery was performed before 3D-CRT in some cases and not in others, KCO was included instead of DLCO to take into account patients’ alveolar volumes before radiation. Treatment factors included concomitant chemotherapy, concomitant corticotherapy, and total MLD. In addition, irradiated lung volumes were included in the analysis as dichotomous variables. These variables were defined by determining the median of the percentage of irradiated lung volume to each dose, from 10 to 50 Gy with 10 Gy steps. For each dose step, the first subgroup included the half of the patients with the lowest values of irradiated lung volume, and the second subgroup included the other half with the highest values. This way, the irradiated lung volume thresholds considered in the analysis of prognostic factors were defined independently from the presence or absence of RP.

3. Results Patients’ characteristics are summarised in Table 1. Of the 96 patients enrolled, 6 could not be evaluated for RP after 3D-CRT. Two patients died before the 6-week evaluation due to evolution of their disease. PFTs were not performed at the 6-week evaluation for 4 patients. There was no statistically significant difference between the 90 patients evaluated for RP and the other 6 patients, except for gender (P ¼ 0:025; Fisher’s exact test). Previous surgery was performed in 49 patients (51%): 17 (35%) had a pneumonectomy, 27 (55%) a lobectomy and 5 (10%) a bilobectomy. Sixty-three patients (66%) had previously received one line of chemotherapy (cisplatin-based regimens for all patients except one who received gemcitabine alone). Eleven patients (11%) had received two lines of chemotherapy (cisplatin-based regimens except for 1 patient who received navelbine only), and none had received more than two lines of treatment before radiation. RP grading is shown in Table 2. Forty of the 90 evaluated patients (44%) had RP (i.e. grade $ 1) at 6 weeks. Among these, severe RP (i.e. grade $ 2) was observed in 7 patients (7.8%). In univariate analysis, none of the clinical or functional factors was statistically significantly associated with RP (Tables 3a and b), except age at the time of inclusion ðP ¼ 0:01Þ: In addition, a trend was observed for previous surgery ðP ¼ 0:10Þ and KCO ðP ¼ 0:08Þ: The

178

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181

Table 1 Patients’ characteristics

Table 3a Univariate analysis of risk factors for radiation pneumonitis: clinical factors at baseline

Characteristics

Group with radiotherapy-related pneumonitis evaluated ðN ¼ 90Þ

Total ðN ¼ 96Þ

Gender—n (%) Male Female Age (years)—median (range)

81 (90) 9 (10) 59 (27–77)

84 (88) 12 (12) 60 (27–77)

Karnofsky performance status—n (%)a 90–100% 80%

81 (91) 8 (9)

87 (92) 8 (8)

Smoking status—n (%)b Current Previous Never

34 (41) 46 (55) 3 (4)

37 (42) 47 (53) 4 (5)

Histology—n (%) Squamous cell carcinoma Adenocarcinoma Others

53 (59) 32 (36) 5 (5)

53 (55) 34 (36) 9 (9)

Stage—n (%) I –II IIIa IIIb Unknown

21 (23) 43 (48) 23 (26) 3 (3)

22 (23) 44 (46) 26 (27) 4 (4)

Previous surgery—n (%) Yes No

47 (52) 43 (48)

49 (51) 47 (49)

Previous chemotherapy—n (%) Yes No

58 (64) 32 (36)

63 (66) 33 (34)

Factors

No of patients with RP (%)

P-value

Gender Male Female

37 (46) 3 (33)

0.73a

Age (years) , 60 $ 60

14 (31) 26 (58)

0.01

Karnofsky performance status (%) 80 5 (62) 90–100 34 (42)

0.29a

Current smoker Yes No

16 (47) 20 (43)

0.75

Tumour stage I –II III

12 (57) 27 (41)

0.19

Previous surgery Yes No

23 (53) 17 (36)

0.10

Previous chemotherapy Yes 26 (45) No 14 (44) a

Pulmonary function tests (baseline)—mean ^ SD Vital capacity (%) 81 (21.1) 88 (12.1) Ratio of FEV1 to vital capacity (%) DLCO (%)c 75 (25.3) KCO (%)c 84 (21.1)

81 (20.8) 88 (12.2) 75 (24.9) 83 (21.2)

FEV1: forced expiratory volume in one second; DLCO: diffusion capacity for carbon monoxide; KCO: transfer coefficient for carbon monoxide. a Data were missing for one patient. b Data were missing for eight patients. c Data were missing for three patients.

Table 2 Grading of the 90 patients evaluated for radiotherapy-related pneumonitis Grade

No of patients

%

0 1a 2a 3a 4a

50 33 6 1 0

55.6 36.7 6.7 1.1 –

a Radiation pneumonitis was defined as the development of pulmonary toxicity of grade $1.

0.92

RP: radiation pneumonitis. Fisher’s exact test.

Table 3b Univariate analysis of risk factors for radiation pneumonitis: functional factors at baseline Factors

Vital capacity (%) Ratio of FEV1 to vital capacity (%) KCO (%)a

Mean (SD)

P-value

RP

No RP

84 (18.8) 86 (12.0) 79 (20.8)

77 (22.6) 90 (12.0) 87 (20.9)

0.13 0.16 0.08

RP: radiation pneumonitis; FEV1: forced expiratory volume in one second; KCO: transfer coefficient for carbon monoxide. a Data were missing for three patients (RP: 36 patients, no RP: 51 patients).

analysis of previously defined irradiated lung volume thresholds showed that more than 18, 13 and 10% of lung volume irradiated to 20, 30, and 40 Gy respectively, were statistically significant thresholds predictive of RP ðP , 0:05Þ (Table 4). MLD was also significantly associated with RP in the univariate analysis ðP ¼ 0:01Þ: Regarding only severe RP (i.e. grade $ 2), only MLD, V20 and V30 remained associated with increased toxicity. In contrast, neither age nor DVH thresholds were discriminant for toxicity (data not shown). As a consequence, no multivariate analysis was performed for this endpoint.

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181 Table 4 Univariate analysis of risk factors for radiation pneumonitis: treatment factors Factors

No of patients with RP (%)

0.09

LV irradiated to 20 Gy . 18% Yes 25 (56) No 15 (33)

0.03

LV irradiated to 30 Gy . 13% Yes 25 (56) No 15 (33)

0.03

LV irradiated to 40 Gy . 10% Yes 25 (56) No 15 (33)

0.03

LV irradiated to 50 Gy . 5% Yes 24 (53) No 16 (36)

0.09

Mean (SD) RP 13 (4.0) 17 (6.5) 13 (5.5)

P-value No RP 10 (3.9) 12 (7.7) 9 (5.9)

0.01 0.008 0.009

RP: radiation pneumonitis; LV: lung volume; V20: percentage volume of total lung exceeding 20 Gy; V30: percentage volume of total lung exceeding 30 Gy. a Data were missing for six patients (RP: 37 patients, no RP: 47 patients).

Multivariate analysis, including age, previous surgery, KCO and MLD identified two independent risk factors for whole RP (grade $ 1): age ($ 60 years vs. , 60 years) and MLD (P ¼ 0:036 and P ¼ 0:045; respectively) (Table 5). These results were not modified if relative lung volumes exceeding 20, 30 or 40 Gy (evaluated as discrete thresholds used in the univariate analysis) were chosen instead of MLD to represent the effect of 3D-CRT on RP.

Table 5 Multivariate analysis of factors predicting radiation pneumonitis Factorsa

Logistic relative risk (95% CI)b

Age $60 years Mean lung dose

2.73 (1.07–6.98) 1.11 (1.01–1.27)

a

4. Discussion 4.1. Definition of RP

P-value

LV irradiated to 10 Gy . 33% Yes 24 (53) No 16 (36)

Mean lung dose (Gy)a V20 V30

179

The following factors, which were significantly associated with radiation pneumonitis according to the univariate analysis, were included in the model: age (,60 vs. $60 years); previous surgery (yes vs. no); transfer coefficient for carbon monoxide (KCO); and mean lung dose (MLD). b The odds ratio provided from the logistic model is defined as the exponential value of the coefficient. CI denotes confidence interval.

In this article we report data from one of the largest prospective series of patients with NSCLC treated with 3D-CRT in the same centre. The lack of consensus to define uniform criteria for RP makes it difficult to evaluate and compare the incidence and severity of RP between publications. Some authors used their own scoring system [7,10], others used the Southwest Oncology Group classification [16], in which grade 2 toxicity is defined as the need for steroids, and grade 3 as the need for oxygen. However, most other authors have preferred to use the RTOG classification [4] which defines grade 2 as mild or moderate symptoms with no steroid requirement, and grade 3 as intermittent use of steroids or oxygen [2,3]. We decided to use the Lent-Soma (Late Effect Normal Tissues; Subjective, Objective, Management and Analytic) classification defined by the RTOG and EORTC, that has become the official scale in Europe [14]. The main objective of this choice was to take into account subjective (clinical) but also objective (radiological and functional evaluation) criteria to score toxicities. The other classifications being mainly based on subjective criteria, it was highly difficult to use them for scoring RP, which has lead to creating the Lent-Soma classification. We cannot present our results with both RTOG or SWOG scales because the date of corticosteroid introduction was not collected for all our patients. Using the Lent-Soma classification, 36.7% of the patients had RP grade 1 or higher, while 7.8% had grade 2 or higher, which is quite lower than usually described. In a meta-analysis of 27 prospective studies (1911 patients), Roach et al. reported 14% of NSCLC patients with RP grade 2 or over, using the RTOG classification [19]. However, in contrast with these older studies, we used 3D-CRT that may significantly reduce the volume of lung incidentally irradiated and consequently the risk of RP, as previously described [16]. A prospective phase I –II trial of hyperfractionated radiation therapy from the RTOG on 848 patients reported that 2.6– 8.1% of patients had severe or life-threatening acute pneumonitis (, 90 days) according to the total radiation doses [3] while we observed only 1.1% of patients with grade 3 RP, and no patient with grade 4. However, they used hyperfractionated radiation therapy delivering high total doses (from 60 to 79.2 Gy with daily doses of 2.4 Gy in 2 fractions of 1.2 Gy) without 3D-CRT, which makes it highly difficult to compare results with our report. Several studies have reported the incidence of RP using 3D-CRT: a large prospective study (99 patients) reported that 14% of the patients had grade 2 or higher RP (RTOG classification) at six-months [6], considering both acute and late toxicity, in contrast with our early evaluation (2 months). Martel et al. reported that 20% of the patients included in a prospective study on conformal radiation in patients with NSCLC had severe RP with a long follow-up of 2 years, including again

180

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181

acute and late toxicity [15]. In contrast, a recent report showed that 19% of patients (39/201) treated with 3D-CRT for lung cancer showed a moderate to severe RP using the common toxicity criteria (including only clinical factors) [7]. These results are better than ours (19% of RP vs. 36%) but we have taken into account clinical symptoms, radiological criteria and loss of pulmonary function, which probably lead to diagnose more cases of RP. 4.2. Prognostic factors 4.2.1. Regarding whole radiation pneumonitis (grade $ 1) Karnofsky performance status [17,20], tumour localisation [6,22], no previous surgery [17], concomitant chemotherapy [13,22] and gender [20] have been reported to be independent prognostic factors for RP. No statistically significant association was found between clinical factors and RP in our study, except for age. In contrast with most of the other studies that did not use 3D dosimetry data, our study used 3D-CRT, thus making comparison and interpretation difficult. A statistically significant positive association between age and RP had already been reported by Gagliardi et al. but their study considered breast cancers [5]. Another study reported that while there was no difference in the development of RP after radiotherapy for lung cancer between younger and older patients, pneumonitis was inclined to be more severe in the elderly group [10]. However, the association between age and RP was not found in several previous studies and it still remains controversial [7,17,20]. Low baseline pulmonary function test results (FEV1 and VC) have been associated with RP [17]. We observed no association between VC, KCO or ratio of FEV1 to VC at baseline and RP at six weeks in this study; however, patients with significantly impaired lung function tests were not eligible for inclusion. The difference between patient PFT values may have been to small too show any significant difference between patients with and without RP. Correlation with dosimetric factors was considered in view of previously published data. We found, in agreement with Graham et al. [6], that V20 was a prognostic factor for RP, as well as V30, as reported by Hernando et al. [7]. MLD was also correlated with the incidence of RP as described in other studies [6,7,11]. The MLD (13 Gy) and the V20 and V30 values reported in our study were lower than those reported by Hernando et al. (21 Gy) [7]. This may be due to the smaller fields used in our study, because of conformal radiation performed with 12 fields. In addition, surgery was performed in 51% of the patients, 35% of them undergoing pneumonectomy, leading to a lower MLD in our study. Previous studies have reported that dosimetric factors, such as planning target volume [21], V20 [6,7], V25 [1], V30 [7], and MLD [7,11,12], were strongly correlated with RP. Since physicians now attempt to treat their patients respecting dosimetric rules (V20 , 30% and V30 , 20%), we need to identify other prognostic factors for RP under these conditions. In this study, all except three patients had a

V30 , 20%. We found new DVH thresholds, based on the median of the percentage of irradiated lung volume, and not based on methods such as the optimised cut-off point, which is statistically incorrect [8]. The latter method considers all possible divisions of the population in two groups obtained by changing the cut-off level for irradiated lung volume, and then selects the best division leading to the largest difference in RP between the two groups. The optimised cut-off point thus varies with the different groups of patients, without any standardisation, and is difficult to interpret clinically or statistically. In our study, the limits used to define the groups were independent from the observation of RP in the groups. We used 12 radiation fields to perform the conformal radiation treatment, leading to a pattern of DVHs. This pattern of DVHs would be different if, for example, only 8 fields were used. The availability of a large panel of thresholds, from low to high doses (20 –40 Gy), might provide an advantage because this may reflect DVH patterns and be more independent from radiation modalities than would be a punctual value. There may be an advantage in looking at more than only one or two threshold values, to decide if there is an acceptable risk of complications or not. Our threshold values were significant prognostic factors for RP in the univariate analysis, as was the loss of pulmonary function, considered alone. This could mean that the loss of pulmonary function may play an important role in RP diagnosis. 4.2.2. Regarding severe RP (grade $ 2 only) Regarding only grade 2 or higher RP, the univariate analysis shows that a high MLD, V20 and V30 remain associated with more RP. In contrast, neither age nor DVH thresholds are discriminant for toxicity. The main reason is obviously the too small number of patients. There were only 7 cases of severe (grade $ 2) RP and, as a consequence, the difference between groups is very difficult to show. What we may conclude is that age over 60 years is probably an independent prognostic factor for RP due to the pulmonary susceptibility of the oldest patients. However, the grade of RP in these patients remains moderate and age over 60 years does not induce more severe toxicities (grade $ 2) than in younger patients. Regarding our thresholds, we think they should be considered whenever possible to avoid occurrence of RP (including moderate complications). However, their capacity to predict severe pulmonary toxicity remains uncertain, probably due to the small number of severe events in our series. In conclusion, the incidence and severity of RP remain difficult to evaluate. Using PFTs for the diagnosis of RP probably leads to more patients being diagnosed than if only clinical and radiological events are used. In this study, regarding the whole pulmonary toxicity (grade $ 1), dosimetric factors (MLD, V20, V30) and age ($ 60 years) were predictive for RP. In addition, thresholds from 20 to 40 Gy, based on stratification according to the median of the percentage of irradiated lung volume, were also predictive

L. Claude et al. / Radiotherapy and Oncology 71 (2004) 175–181

factors. They may, therefore, be used to distinguish between patients at high and low risk for RP. However, only MLD, V20, and V30 remained associated with severe RP (grade $ 2), probably due to the small number of severe events in our series.

Acknowledgements Thanks to Marie-Dominique Reynaud and Franc¸oise Chirat for assistance with manuscript preparation.

References [1] Armstrong J, Raben A, Zelefsky M, et al. Promising survival with three-dimensional conformal radiation therapy for non-small cell lung cancer. Radiother Oncol 1997;44:17–22. [2] Byhardt RW, Martin L, Pajak TF, Shin KH, Emami B, Cox JD. The influence of field size and other treatment factors on pulmonary toxicity following hyperfractionated irradiation for inoperable nonsmall cell lung cancer (NSCLC)—analysis of a radiation therapy oncology group (RTOG) protocol. Int J Radiat Oncol Biol Phys 1993; 27:537–44. [3] Cox DJ, Azarnia N, Byhardt RW, Shin KH, Emami B, Pajak TF. A randomized phase I/II trial of hyperfractionated radiation therapy with total doses of 60 Gy to 79.2 Gy: possible survival benefit with .69.6 Gy in favorable patients with radiation Therapy Oncology Group Stage III Non-Small-Cell lung Carcinoma: report of Radiation of Radiation Therapy Oncology Group 83-11. J Clin Oncol 1990;8: 1543–55. [4] Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Radiat Oncol Biol Phys 1995;31:1341–6. [5] Gagliardi G, Bjohle J, Lax I, et al. Radiation pneumonitis after breast cancer irradiation: analysis of the complication probability using the relative seriality model. Int J Radiat Oncol Biol Phys 2000;46: 373–81. [6] Graham MV, Purdy JA, Emami B, et al. Clinical dose-volume histogram analysis for pneumonitis after 3D treatment for non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1999;45:323 –9. [7] Hernando ML, Marks LB, Bentel GC, et al. Radiation-induced pulmonary toxicity: a dose-volume histogram analysis in 201 patients with lung cancer. Int J Radiat Oncol Biol Phys 2001;51:650–9. [8] Hilsenbeck SG, Clark GM, Mc Guire WL. Why do so many

[9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

181

prognostic factors fail to pan out? Breast Cancer Res Treat 1992;22: 197–206. ICRU-50, Prescribing, recording, reporting, photon beam therapy. Washington, DC: International Commission on Radiation Units and Measurements; 1994. Koga K, Kusumoto S, Watanabe K, Nishikawa K, Harada K, Ebihara H. Age factor relevant to the development of radiation pneumonitis in radiotherapy of lung cancer. Int J Radiat Oncol Biol Phys 1988;14: 367–71. Kwa SLS, Lebesque JV, Theuws JCM, et al. Radiation pneumonitis as a function of mean lung dose: an analysis of pooled data of 540 patients. Int J Radiat Oncol Biol Phys 1998;42:1–9. Lebesque JV, Seppenwoolde Y, Belderbos JS, et al. Radiation pneumonitis and NTCP models. Int J Radiat Oncol Biol Phys 2001; 51(Suppl 1). abstr. 154:86. Lee JS, Scott Ch, Komaki R, et al. Concurrent chemoradiation therapy with oral etoposide and cisplatin for locally advanced inoperable nonsmall-cell lung cancer: radiation therapy oncology group protocol 91-06. J Clin Oncol 1996;14:1055–64. LENT SOMA Tables, Radiother Oncol 1995;35:17–60. Martel M, Graham M. Dose volume histograms (DVHs) and normal tissue complication probabilities (NTCPs) and their clinical applications. Treatment optimisation for lung cancer: from classical to innovative procedures, Paris: Elsevier; 1998. p. 125–38. Martel MK, Ten Haken RK, Hazuka MB, Turrisi AT, Fraass BA, Lichter AS. Dose-volume histogram and 3-D treatment planning evaluation of patients with pneumonitis. Int J Radiat Oncol Biol Phys 1994;28:575 –81. Monson JM, Stark P, Reily JJ, et al. Clinical radiation pneumonitis and radiographic changes after thoracic radiation therapy for lung carcinoma. Cancer 1998;82:842–50. Perez CA, Stanley K, Grundy G, et al. Impact of irradiation technique and tumor extent in tumor control and survival of patients with unresectable non oat cell carcinoma of the lung: report by the RTOG. Cancer 1982;50:1091 –9. Roach III M, Gandara DR, You HS, et al. Radiation pneumonitis following combined modality therapy for lung cancer: analysis of prognostic factors. J Clin Oncol 1995;13:2606 –12. Robnett TJ, Machtay M, Vines EF, McKenna MG, Algazy KM, McKenna WG. Factors predicting severe radiation pneumonitis in patients receiving definitive chemoradiation for lung cancer. Int J Radiat Oncol Biol Phys 2000;48:89– 94. Sunyach MP, Falchero L, Pommier P, et al. Prospective evaluation of early lung toxicity following three-dimensional conformal radiation therapy in non-small-cell-lung cancer: preliminary results. Int J Radiat Oncol Biol Phys 2000;48:459– 63. Yamada M, Kudoh S, Hirata K, Nakajima T, Yoshikawa J. Risk factors of pneumonitis following chemoradiotherapy for lung cancer. Eur J Cancer 1998;34:71–5.