- Email: [email protected]
The role of radiotherapy in treatment of stage I non-small cell lung cancer
Lung Cancer (2003) 41, 1 /11
www.elsevier.com/locate/lungcan
The role of radiotherapy in treatment of stage I non-small cell lung cancer ´na, Xueying Qiaoa, Owe Tullgrena, Ingmar Laxa, Florin Sirze Rolf Lewensohna,b,* a
Department of Oncology and Pathology, Radiumhemmet, Karolinska Hospital, SE-171 76 Stockholm, Sweden b Department of Medical Radiation Biology, Cancer Centre Karolinska R8:00, Karolinska Hospital, SE-171 76 Stockholm, Sweden Received 4 December 2002; received in revised form 3 March 2003; accepted 6 March 2003
KEYWORDS NSCLC stage I; Radiotherapy; Local control; Regional control; Prophylactic nodal irradiation; Survival
Summary Most information on results with radiotherapy (RT) for stage I non-small cell lung cancer (NSCLC) is based on retrospective studies on RT-treated inoperable NSCLC cases. Thus, the role of RT for stage I NSCLC, as a curative modality, has not yet been established. A literature search for studies on stage I non-small cell lung carcinoma (NSCLC) treated by RT alone resulted in 18 papers published between 1988 and 2000. The majority of stage I patients received RT treatment because they were medically inoperable. The main contraindications for surgery were grave impairment of pulmonary function and serious cardiovascular disease. Local recurrence was the most common reason for treatment failure (median value 40%) but varied highly between the studies, ranging from 6.4 to 70%. In contrast with local recurrence, regional failure was not a major problem (0 /3.2%). Generally, smaller tumour size, low T-stage and increased dose had a favourable impact on local control and increased local control was followed by increased survival. No serious treatment complications were recorded in the majority of these studies. Overall treatment results were, however, disappointing. The median survival in these studies ranged from 18 to 33 months. The 3- and 5-year overall survival was 349/9 and 219/8% (mean value9/1 S.E.), respectively. The cause-specific survival at 3 and 5 years was 399/10 and 259/9% (mean value9/1 S.E.), respectively. Dose escalation, in a setting with conformal RT using involved field or stereotactic RT, should be the focus of developmental therapeutic strategies with inoperable stage I NSCLC to improve local control and survival. – 2003 Elsevier Science Ireland Ltd. All rights reserved.
1. Introduction
*Corresponding author. Tel.: /46-8-5177-5092; fax: /46-85177-6630. E-mail address: [email protected] (R. Lewensohn).
Surgery is, at present, standard treatment for stage I non-small cell lung cancer (NSCLC). Most studies on long-term survival have shown a more than 50% 5-year survival after radical resection [1 / 3]. It is unclear whether radiotherapy (RT) may
0169-5002/03/$ - see front matter – 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0169-5002(03)00152-1
2
emerge as an alternative method of treatment to surgery considering efficacy parameters. RT has been offered to patients who are not suitable for surgery due to medical problems and to patients who refuse surgery. This paper presents a review of publications on results from studies of conventional RT for stage I NSCLC and addresses the following questions: (1) Can these patients be cured by RT alone? (2) What is the pattern of failure? (3) What are the factors affecting prognosis of the patients with stage I NSCLC after treatment with RT alone? (4) Is RT well tolerated by these patients? (5) Which are the future directions for the treatment of inoperable stage I NSCLC?
2. Material This review is based on 18 studies reporting results on RT for stage I NSCLC published between 1988 and 2000. The majority of these studies (13/ 18) included less than 100 patients (ranging from 31 to 93), four studies included between 100 and 150 patients and one larger study included 347 patients. The majority of these studies are retrospective analyses. Data selected from publications concerning treatment for early stage NSCLC by RT alone are listed in Table 1A and B. The percentage of medically inoperable patients varies from 59.2 to 100%. Two publications show a lower percentage of medically inoperable patients (59.2 and 64%) which is due to the inclusion of patients refusing surgery [4,5]. The main medical problems of the medically inoperable patients were poor pulmonary function, serious cardiovascular disease, poor general condition etc. It is also apparent from Table 1A that most of the patients treated by RT were elderly. Radiation dose and fraction size varied in different studies. For comparative reasons, the biologically effective dose (BED) was calculated (acute effect a /b /10) for each study using time correction whenever it was possible [6]. The BED values were calculated in order to make possible a comparison of the biological response to different fractionation schemes, rather than the physical doses given [7]. In some studies, different total doses and different doses/fractions were delivered to different patients. RT usually started with opposed AP and PA fields, followed by boost to the primary tumour after 40/50 Gy [4,8 /15]. Opposed AP/PA fields and oblique fields (avoiding the spinal cord) were also used [5,6,16]. There was no consistent use of prophylactic node irradiation (Table 1B). Notably, there are studies were only some of the
X. Qiao et al.
patients are reported to have received prophylactic node irradiation.
3. Results and discussion 3.1. Factors of importance for local and regional control When using RT one of the main issues is local control. Local recurrence was the most common reason for treatment failure with conventional RT of stage I NSCLC, but the frequency of recurrence varied considerably according to different reports (between 6.4 and 70%) [4,6,10,14 /20] as can also be seen in Table 3. Distant metastases should, however, not be underestimated, since in some series distant failure was more common than local failure [9,10,21]. On the other hand, regional node failure alone occurred alone only in 0 /3.2% of the cases and is, therefore, not a major problem [4,6,9,10,14 /21]. Follow up of treatment-results in lung cancer is often done on a 2-, 3- and 5-year basis. Since the reporting at these time intervals varies between different authors it is not possible to extract the median survival for the patients in all studies (Table 2). Median time to relapse varied from 21 to 30 months. In Fig. 3, the cumulative probability of recurrence is plotted according to the data obtained from two publications [6,9]. The 3-year recurrence rate in these two studies were 60 [9] and 67.1% [6]. The majority of recurrences occurred within 3 years after treatment. Thus, according to these two studies a 3-year follow up is needed to estimate the recurrence rate after RT of stage I NSCLC. 3.1.1. Tumour parameters having an impact on local control When using RT it is generally thought that the size of a tumour will predict the possibility of obtaining complete remission (CR). When reviewing the papers that we refer to wide range of CR rates are reported, ranging from 17 to 61% [4,8,9,14 /16]. Out of a few papers one may conclude that smaller tumours, as expected, had a higher CR rate than larger ones. Smaller tumours, as expected, had a higher CR rate than larger ones [14 /16]. It should be understood that the time window for evaluating response is short in RT of the lung since longer follow up of remission will be hampered by radiation caused fibrosis. The evaluations of response, therefore, refer to those performed at the end of the RT treatment.
Radiotherapy, inoperable stage I NSCLC
3
Table 1 Author (year)
Age (years) Mean/median
(A ) Patients characteristics Slotman et al., 1996 75/ / [9] Zhang et al., 1989 [7] 57.1/ / Hayakawa et al., 1999 n.s [8] Haffty et al., 1988 64/ / [19] Kupelian et al., 1996 n.s [10] Slotman et al., 1994 75 [17] Jeremic et al., 1997 //63 [4] Noordijk et al., 1988 74/ / [13] Kaskowitz et al., 1993 //73 [16] Talton et al., 1990 65/ / [11] Sandler et al., 1990 //72 [18] Cheung et al., 2000 //71.5 [6] Graham et al., 1995 //67 [22] Dosoretz et al., 1992 //74 [12] Krol et al., 1996 [15] 74.2/ / Sibley et al., 1998 [20] //70 Morita et al., 1997 74.2/74.6 [14] Gauden et al., 1995 //70 [5] Total Total dose (Gy)
Number of patients
Medically inoperable patients (%)
Range
Stage I
All
54 /85
31
31
81
37 /70 n.s
35 36
44 36
72.7 88.9
51 /83
43
43
86
n.s
45
71
100
57 /93
47
47
n.s
51 /70
49
49
59.2
n.s
50
50
84
44 /86
53
53
81.8
40 /89
60
77
n.s
n.s
77
77
78
45 /86
92
102
76.5
43 /89
93
103
70
51 /88
106
152
88.8
58 /88 46 /95 50 /89
108 141 149
108 141 149
88 100 83
34 /90
347
347
64
Fractions (F)
1562 Acute BED (Gy)
1680 D2 Prophylactic node irradiation yes/no
In general the term local control is used by different authors to describe whether or not there is tumour progression in the irradiated volume of the chest. This, of course, must mean that there are some uncertainties in the determination of local control. Such uncertainties may influence the reported frequencies of local control. Another way of assessing local control is trying to understand the frequency of local failure in relation to e.g. tumour size. A general relation between tumour size and local failure seems to be apparent [11,15,18]. The authors presenting such data generally use arbitrary cut-offs in tumour size of which
one example is the work by Slotman et al. who shows that total local failure rate was 2.4 and 47.8% for tumours 5/4 and /4 cm, respectively [18]. Interestingly, Slotman et al. found no correlation between tumour size and loco-regional node failure [10]. Another way of assessing the relation between local control and local tumour extent in stage I disease is to use T-stage. Again in three papers [4,11,13] a better local control was shown to be obtained with T1 as compared with T2 tumours. In one of these, the Kupelian et al. study the difference was not statistically significant, a find-
4
X. Qiao et al. Table 1 (Continued )
Author (year)
Age (years) Mean/median
Number of patients Range
Stage I
Medically inoperable patients (%) All
(B ) Compilation of publications on treatment of early stage NSCLC by RT alone Slotman et al., 1996 48 12 F 76.4 [9] 2 Gy/F 62.8 Zhang et al., 1989 [7] 55 /70 Hayakawa et al., 1999 60 /81 2 Gy/F 68.1 67.1 [8] Haffty et al., 1988 54 6 weeks/split 59 62.2 [19] course 59 6 /7 weeks/continuous Kupelian et al., 1996 63.23 65 60 [10] Slotman et al., 1994 32 Gy/6 F, 40 Gy/ 28/19 [17] 10 F 48 Gy/12 F, 56 Gy/ 20 F Jeremic et al., 1997 69.6 1.2 Gy bid 71 58.5 [4] Noordijk et al., 1988 60 Split course, 3 Gy/ 63.4 72 [13] F 2 Gy/F 65.1 63 Kaskowitz et al., 1993 63.2a [16] Talton et al., 1990 60 30 F 62.8 60 [11] 2 Gy/F 62.8 60 Sandler et al., 1990 60a [18] Cheung et al., 2000 52.5 20 F 67.8 59 [6] Graham et al., 1995 60 30 F 62.8 60 [22] 1.8 /2 Gy/F 62.8 60 Dosoretz et al., 1992 45 /75 [12] Krol et al., 1996 [15] 60 7 weeks, 3 Gy bid 63.4/74.3 72 65/5.2 weeks, 2.5 qd Sibley et al., 1998 [20] 64a 1.2 bid, 3 qd Morita et al., 1997 64.7a 2 Gy/F 65.3 64.7 [14] Gauden et al., 1995 50 20 F/4 weeks 62.5 55 [5] a
0/31 44/0 10/26 38/5
51/20
49/0 0/50 n.s. 77/0 69/8 0/102 82/21 n.s. 0/108
103/38 66/88 347/0
Median, D2 is the total dose given with 2 Gy per fraction, calculated with the BED formalism.
ing which may be explained by an overlap in tumour size between the different T-stages [11]. Looking at the data there is a definite trend for a lowered local control with increasing tumour size. The lack of a statistical difference between tumour sizes in relation to local control may be based on common problems encountered when analysing retrospective studies such as limited number of
patients and radiation given with conventional dose delivery, i.e. larger tumours may have received higher radiation doses. The dependence of local control on histology is in general not handled by the different papers. Most likely this relates to improperties in the technology of differentiating between different histologies of NSCLC using sputum cytology etc.
Radiotherapy, inoperable stage I NSCLC
5
23
48
41
33
32
41 28
5-year 3
32 32
3.1.2.1. RT dose There is support for better local control when total dose is increased [11,17]. This conclusion is based on dividing the patient groups into those receiving less than 60 /65 Gy and those receiving higher doses. In addition, intrathoracic recurrence rate was significantly lower in patients who received the high dose schedule (2/24: 8.3%) compared with the low dose group (10/23: 43.5%), with a statistically significant difference (P B/0.01) in the Slotman et al. series [18]. We were able to calculate acute BED for nine different studies (Table 4). As can be seen in Fig. 4 there is a trend for better local control with increased BED [4,9,10,14,15,17,20]. The relatively limited span of doses (BED 59 /76 Gy) delivered does not enable a clear-cut conclusion of the role of dose in obtaining local control. In addition, we cannot exclude the possibility that in those studies presenting the highest BED-values, tumours treated were of a larger size. Taken together the data indicate that better local control is obtained with the higher doses used in these studies.
54
34.2 40
21 16 22.2 27 77.9
70
39
56
33 17 36
43 72
20.9 33 33 18 20 20 28 27 27.2 27.9
81
31.4 15 23 6 8 30 13 15
Zhang et al. [7] Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Gauden et al. [5]
91.4 75
57.1 31 42 19 42
90
81
60
54 93
53 22
42 56 33
31 39 13
2 5-year 3 2 1 5-year 3 2 1
Median survival (month) Authors
Table 2
Survival after RT for stage I NSCLC
Overall survival (%)
Cause-specific survival (%)
Disease-free survival (%)
3.1.2. Modalities of radiotherapy and local control
3.1.2.2. RT volume (issue of margins, elective nodal irradiation) One obvious question when irradiating lung tumours is which field margins to accept. For most series, the margin was at least 1.5 cm between the gross tumour and field border. For groups of patients with margins 5/1.5 cm, 48% of the tumours recurred at their primary location, while 35% of the patients had local recurrence when the margin were /1.5 cm [17]. In Cheung et al.’s series, 90.3% of the margins were larger than 1.5 cm and could not be considered a prognostic factor for recurrence-free survival [6]. Thus, there are no conclusive data to indicate optimal margins for local control. Regional failure alone is not a common cause of failure of stage I NSCLC treated by RT only. Locoregional relapse occurred only in 0 /3.2% of the patients treated without prophylactic node irradiation (see Table 3 and [10,14,16,18]). It was, however, reported to be higher but not statistically significant in a non-prophylactic mediastinal irradiation group than that of a prophylactic mediastinal irradiation group [21].
6
X. Qiao et al.
Table 3 The patterns of failure after RT for stage I NSCLC Author
Local failure (%) Alone
Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Weighted mean a b
27.8 11.1 41.5 0 n.s. 16.3 19.1 42.8 39 n.s. n.s. 17.2
a
Regional failure alone (%) b
Total 65.7 19.4 43.4 6.4 45 19.1 25.2 42.8 39 70 44.3 28.6
1.8 2.8 0 3.2 0 2.8 n.s. n.s. n.s. 0 0 1.0
Distant failure (%) Alone
All
2.8 33.3 n.s. 3.2 n.s. 14.9 21.3 n.s. 11.6 26 n.s. 9.3
33.3 36.1 32 16.1 25 19.8 21.3 n.s. 11.6 26 12.1 15.7
Single site failure. Any local failure including combined, regional and distant failure.
Fig. 1 Overall survival of stage I NSCLC (the estimated 3, 5-year overall survival rate is 349/9 and 219/8%, respectively). The data were extracted from the studies presented in Table 2.
3.2. Which factors are prognostic for survival The reports on treatment results with stage I NSCLC present large variations of patient survival. The survival after RT treatment when considered only for stage I NSCLC of different studies is shown in Table 2. The median survival varied between 18 and 33 months. The overall survival rate of 1-, 2-, 3- and 5-year survival rate were 70 /91.4, 39 /72, 16 /57.1 and 6 /31.4%, respectively. In this heavily compromised group of patients with short life expectancy it is more appropriate to use data on cause specific survival rather than overall survival. The 2-, 3- and 5-year cause-specific survivals were 54 /93, 22/56, and 13/39%. The 5-year disease-
Fig. 2 Cause-specific survival of stage I NSCLC (the estimated 3, 5-year cause-specific survival is 399/10 and 259/9%, respectively). The data were extracted from the studies presented in Table 2.
free survival was 23 /41%. The overall survival rates and cause-specific survival rates are presented in Figs. 1 and 2. The estimated 3- and 5-year overall survival was 349/9 and 219/8% (mean value9/1 S.E.), respectively. The mean value was calculated considering the number of patients and the S.E. values were calculated for a group of 100 patients. The estimated cause-specific survival at 3 and 5 years was 399/10 and 259/9%, respectively. This can be compared with the roughly 60% 5-year survival rate of cases that undergo surgery [22]. 3.2.1. Is tumour volume in RT treated stage I NSCLC of importance for survival When evaluating treatment results, survival may be another end point not entirely linked to local or loco-regional control and should, therefore, be evaluated separately. Thus, the importance of
Radiotherapy, inoperable stage I NSCLC
Fig. 3 Cumulative probability of recurrence after RT for early stage NSCLC (the data with ‘‘'’’ and ‘‘I’’ was from reference [7,18], respectively).
tumour size was analysed with respect to survival. The median survival time for patients with a tumour size 5/4 cm was longer than for those with tumours /4 cm (32 vs. 12 months; P /0.007) [14]. In Sandler et al.’s series [19], the overall survival for tumours of B/3, 3 /6 and /6 cm was 25, 13 and 0% at 3 years (P /0.009). The causespecific survival at 3 years was 30, 17 and 0%, respectively (P /0.0001). No relationship was, however, found between tumour size and survival in Kaskowitz et al.’s report [17], although the material is composed of similar tumour sizes as in e.g. Sandlers report. Krol et al. found that the cause-specific survival for tumours of 5/4 or /4 cm was 40 and 13%, respectively, at 3 years.
Fig. 4 Local-component failure (including local failure alone, local plus regional failure and local plus distant failure) vs. BED.
7
Similarly to the analysis of which factors are important for local control we used T-stage as surrogate for tumour volume when analysing the relation to survival. Zhang et al. reported that 66.7% of T1N0 patients but only 24% of T2N0 patients survived more than 5 years [8]. Gauden et al. found that the disease-free survival at 5 years was 32% for T1N0 tumours compared with 18% for T2N0 tumours, P B/0.01 [5]. Others have reported similar results [4,18]. There was, however, no significant relation between T-stage and overall survival in one study [10]. It is hard to draw firm conclusions on the exact importance of tumour size for survival. The general trend, however, is that the prognosis worsens with increasing tumour size. It is our contention that the analysis is more strict when using direct measurements of tumour size compared with the T-staging procedure. 3.2.2. The role of prophylactic nodal irradiation Only one study reported better overall survival in a patient group treated with prophylactic nodal irradiation compared with those that did not receive this treatment [15]. However, the authors did not report an increased rate of regional failure in the non-irradiated group. The results may be influenced by the imbalance of non-squamous histology between the groups (51% in the nonirradiated and 31% in the irradiated group) which may explain the higher rate of distant metastatic events in the non-irradiated group. In four other studies, no improvement of overall and causespecific survival was observed after prophylactic irradiation of mediastinal nodal groups [4,11,18,21]. In studies not using elective nodal irradiation the reported regional failure rates were remarkably low [10,14,16,18]. Taken together these data argue against the use of prophylactic RT of the mediastinum in stage I NSCLC. 3.2.3. Total radiation dose and survival Total dose is an important prognostic factor for stage I NSCLC. Dosoretz et al. [13] reported that the 2 year disease-free survival of patients receiving radiation doses of 70 Gy or more, between 60 and 70 Gy and between 50 and 60 Gy was 50, 33 and 22%, respectively. Patients presenting T1 tumours and who received radiation doses of ]/65 Gy experienced a disease-free survival at 2 years in approximately 73% of the cases compared with 43% at 2 years for patients receiving less than 65 Gy. Similar dose dependence was also noted for T2 patients [13]. The disease-free survival at 3 years for patients having tumours less than 3 cm in diameter and treated with a dose higher than 65 Gy was found to be 90% [13]. Zhang et al. observed
8
X. Qiao et al.
Table 4 BED in relation to local failure, 5-year overall and cause-specific survival after RT for stage I NSCLC Author
BED (acute)
Local failure alone (%)
Any local failure (%)
Overall survival (%)
Cause-specific survival (%)
Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Gauden et al. [5]
/ 68.1
27.8 11.1
65.7 19.4
15 23
31 39
65.1
41.5
43.4
6
13
76.4 71 / / 62.8 59 63.4 65.3 62.5
0 / 16.3 19.1 42.8a 39 / / /
6.4 45 19.1 25.2 42.8a 39 70 44.3 /
8 30 13 15 17a 21 16 22.2 27
76b / 32 32 22a / / / /
a b
3-year survival. 4-year survival.
that a higher dose was associated with good prognosis [8]. In that study the 5 year overall survival of the patients receiving 69 /70 Gy was higher than that of patients receiving 55 /61 Gy (36 vs. 27%). While the two previous reports favour dose-relationship, the latter study does not. It is, however, important to understand that the lack of correlation in Sandlers study may, as pointed out by the author himself, very well be related to the narrow range of doses used. Also, another factor adding to the problem of obtaining a linear dose response with survival is the possibility that larger tumours by themselves confer less favourable prognosis as compared with smaller tumours, in spite of the possibility that they may have received a higher dose compared with smaller tumours.
3.2.4. Is survival linked to local control There was an improved 5-year cause-specific survival of patients in whom the primary tumour was controlled (46% survival) compared with those with local failure (12% survival), P B/0.03 [21]. On the other hand a trend but not a statistically significant correlation was found for local control in relation to cause specific survival when analysing the articles listed in Table 4 [9,16 /18,21]. Patients achieving CR after RT had a high overall survival rate and local control [8]. In Krol et al.’s report [16], out of the 50 patients who achieved CR, 31(60%) had long-term local control. The actuarial local relapse free survival at 1, 3 and 5 years were 96, 71 and 66%, respectively. All 58 patients not achieving CR had local progression. Taken together, these data indicate that the chance of long term
survival is linked to the achievement of local control.
3.3. Comparison between radiotherapy and surgery No randomised study using modern RT addresses this question. One study presented a comparison of survival for elderly ( ]/70 years) NSCLC patients treated with curative RT or surgery, but no statistical difference was, however, recorded between these groups [14]. This study must then, be judged with caution since it was a retrospective analysis. Due to medical problems such as impairment of lung function and cardiac diseases affecting survival of patients treated with RT it is difficult to compare results between RT and surgery [8]. Interestingly, it has been reported that patients refusing surgery had a better chance for long-term survival when treated by RT compared with expected surgery outcome [8].
3.4. Side effects of radiotherapy alone Analysis of the relationship between the dose to the organs at risk and the side effects was not possible because no data about dose (BED) to the organs at risk were available in the publications. Table 5 shows that there is no significant relationship between the dose (BED) prescribed to target and the side effects. This lack of relationship may partly be due to the different criteria and different methods utilised in the different studies and partly due to the fact that the majority of the patients received treatment with similar BED values. In
Radiotherapy, inoperable stage I NSCLC
9
Table 5 Side effects of RT for stage I NSCLC Author
Number of patients
BED acute
Slotman et al. [9] Zhang et al. [7] Hayakawa et al. [8] Haffty et al. [19] Kupelian et al. [10] Slotman et al. [17] Jeremic et al. [4] Noordijk et al. [13] Kaskowitz et al. [16] Talton et al. [11] Sandler et al. [18] Cheung et al. [6] Graham et al. [22] Dosoretz et al. [12] Krol et al. [15]
31
76.4
Well tolerated
112
44 36
62.8 68.1
18% G2 peumonitis No serious complication
100 111.7
43
59
No
71
65
G2 pneumonitis 9.3% G2 esopha- 103.5 gitis 18.6% 7% G2 pneumonitis 100
47
/
Moderate esophagitis
/
2.1% G2 pneumonitis
49
71
97.4
6.1% G3 pneumonitis 6.1% G3 esophagitis
50
63.4
4.1% G3 pneumonitis 4.1% G3 esophagitis No serious side effects
53
65.1
77
Sibley et al. [20] Morita et al. [14] Gauden et al. [5]
BED late No serious complication
120 105
5.7% Moderate pneumonitis
62.8
Mild 64.2 esophagitis 3.8% moderate esophagitis 20% G2 Esophagitis
100
Pulmonary fibrosis 1.3%
77
62.8
Rare
100
102
67.8
22.5% esophagitis
98.4
103
62.8
100
152
62.8
Pneumonitis G3:1% Esophagitis G1:37.9%, G2:2.9% No major complication
100
108
Mild
120
141
63.4/ 73.4 /
1.5% G 3 /5 pneumonitis
/
149
65.3
No reported
107.8
347
62.5
No serious complication
91.5
several of the studies the RTOG scale for toxicity assessment was used. The lack of RTOG conformed toxicity scaling in some of the studies presents an additional difficulty in evaluating these results. Unfortunately no dose volume histograms were presented in these studies, which is something warranted for the future. In most series, however, there were no serious complications. The most frequent early side effect was acute esophagitis. Mild radiation esophagitis occurred in about 22.5 / 64.2% [6,17,23], moderate in 2.9 /20% [12,17,20,23] and severe in about 4.1% of the cases [4]. Serious acute radiation pneumonitis was noted in 1 /4.1% of the patients. Severe late radiation esophagitis occurred rarely and was reported in one article only, as occurring in 6.1% of the patients [4]. In patients with prophylactic irradiation of mediastinum, moderate esophagitis was noted, but
4.9% Pneumonitis, 2.9% G2 fibrosis Peumonitis G2:1.9%
not described in detail [18]. In the Cheung series, esophagitis was also recorded in 22.5% of the cases [6]. The differences in this type of complication may be due to differences in field arrangements even if involved field technique was used. Moderate late radiation pneumonitis was found in 1.9 / 18% of the cases [6,8,17,18,23] and severe late radiation pneumonitis (grade 3 or more) in 6.1% of the cases [4,21]. Late complication in the form of lung fibrosis was seen in 1.3 and 2.9% of the cases [6,12]. These complication frequencies are surprisingly low and probably related to whether or not follow-up X-rays were performed. Sibley et al. reported two patients with grade 3 /5 pneumonitis out of 141 patients (1.5%), both of which received prophylactic mediastinal irradiation [21]. Small fields (7.5 /7.5 cm) did not interfere with the respiratory capacity [13]. There was no difference
10
between complications in patients receiving radiation to the primary tumour only and those patients receiving radiation to the tumour and lymph nodes [11]. Also, there is no evidence to suggest that the use of larger portal fields correlated with a higher incidence of pneumonitis [17]. All in all, since these studies are retrospective, the exact frequency of complications is hard to assess. Only prospective well-controlled clinical studies may give us exact figures on the incidence of complications with RT.
X. Qiao et al.
Considering the high local failure and the low regional failure after conventional RT treatment, dose escalation using conformal RT or stereotactic RT with ‘‘involved field technique’’ or light ion therapy is warranted in stage I NSCLC with the aim to improve local control and survival.
Acknowledgements This study was supported by grants from the Stockholm Cancer Society and the Swedish Heart Lung Foundation.
4. Conclusions and future directions
References
A minority of patients diagnosed with stage I NSCLC can be cured by conventional RT alone. The most common reason for treatment failure is local recurrence alone. Regional failure in the mediastinum is surprisingly uncommon indicating that prophylactic irradiation of the mediastinum may not be necessary in these patients. In general, smaller tumour-size, lower T stage and higher dose, correlates with better local control and survival. The 5-year overall survival and cause-specific survival were 219/8 and 259/9%, respectively, in cases treated with conventional RT. Thus, whilst these results compare unfavourably with surgery, one must keep in mind that the majority of patients who have been offered RT for stage I NSCLC are heavily compromised by their medical condition. This review indicates that higher RT doses correlate with better local control and survival. It is our belief that conformal 3-D RT or stereotactic RT may better conform to the prescribed dose to the actual target volume and thereby spare normal tissues permitting dose escalation. In stage I /III NSCLC dose escalation using conformal 3-D RT has been feasible [24]. Attempts have now been made to use stereotactic radiation therapy in primary lung cancer. With an acute BED of at least 75 Gy delivered by stereotactic RT, with or without conventional RT, the rates of overall survival, cause-specific survival and local control at 3 years for early stage NSCLC has been reported to be 68, 90 and 96%, respectively [25]. This is higher than the results obtained by conventional RT and also by surgery. The use of charged particle beams is another option. In patients with stage I NSCLC treated with a carbon ion beam using involved field technique, tumour control rate increased linearly with increasing carbon ion dose up to 100% local control [26].
[1] Martini N, Rusch VW, Bains MS, Kris MG, Downey RJ, Flehinger BJ, et al. Factors influencing 10-year survival in resected stages I to IIIa non-small cell lung cancer. J Thorac Cardiovasc Surg 1999;117:32 /6 (Discussion 7 /8). [2] Williams DE, Pairolero PC, Davis CS, Bernatz PE, Payne WS, Taylor WF, et al. Survival of patients surgically treated for stage I lung cancer. J Thorac Cardiovasc Surg 1981;82:70 / 6. [3] Naruke T, Goya T, Tsuchiya R, Suemasu K. Prognosis and survival in resected lung carcinoma based on the new international staging system. J Thorac Cardiovasc Surg 1988;96:440 /7. [4] Jeremic B, Shibamoto Y, Acimovic L, Milisavljevic S. Hyperfractionated radiotherapy alone for clinical stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1997;38:521 /5. [5] Gauden S, Ramsay J, Tripcony L. The curative treatment by radiotherapy alone of stage I non-small cell carcinoma of the lung. Chest 1995;108:1278 /82. [6] Cheung PC, Mackillop WJ, Dixon P, Brundage MD, Youssef YM, Zhou S. Involved-field radiotherapy alone for earlystage non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2000;48:703 /10. [7] Fowler JF. Brief summary of radiobiological principles in fractionated radiotherapy. Semin Radiat Oncol 1992;2:16 / 22. [8] Zhang HX, Yin WB, Zhang LJ, Yang ZY, Zhang ZX, Wang M, et al. Curative radiotherapy of early operable non-small cell lung cancer. Radiother Oncol 1989;14:89 /94. [9] Hayakawa K, Mitsuhashi N, Saito Y, Nakayama Y, Furuta M, Sakurai H, et al. Limited field irradiation for medically inoperable patients with peripheral stage I non-small cell lung cancer. Lung Cancer 1999;26:137 /42. [10] Slotman BJ, Antonisse IE, Njo KH. Limited field irradiation in early stage (T1-2N0) non-small cell lung cancer. Radiother Oncol 1996;41:41 /4. [11] Kupelian PA, Komaki R, Allen P. Prognostic factors in the treatment of node-negative non-small cell lung carcinoma with radiotherapy alone. Int J Radiat Oncol Biol Phys 1996;36:607 /13. [12] Talton BM, Constable WC, Kersh CR. Curative radiotherapy in non-small cell carcinoma of the lung. Int J Radiat Oncol Biol Phys 1990;19:15 /21. [13] Dosoretz DE, Katin MJ, Blitzer PH, Rubenstein JH, Salenius S, Rashid M, et al. Radiation therapy in the management of medically inoperable carcinoma of the lung: results and implications for future treatment strategies. Int J Radiat Oncol Biol Phys 1992;24:3 /9.
Radiotherapy, inoperable stage I NSCLC [14] Noordijk EM, vd Poest Clement E, Hermans J, Wever AM, Leer JW. Radiotherapy as an alternative to surgery in elderly patients with resectable lung cancer. Radiother Oncol 1988;13:83 /9. [15] Morita K, Fuwa N, Suzuki Y, Nishio M, Sakai K, Tamaki Y, et al. Radical radiotherapy for medically inoperable non-small cell lung cancer in clinical stage I: a retrospective analysis of 149 patients. Radiother Oncol 1997;42:31 /6. [16] Krol AD, Aussems P, Noordijk EM, Hermans J, Leer JW. Local irradiation alone for peripheral stage I lung cancer: could we omit the elective regional nodal irradiation. Int J Radiat Oncol Biol Phys 1996;34:297 /302. [17] Kaskowitz L, Graham MV, Emami B, Halverson KJ, Rush C. Radiation therapy alone for stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1993;27:517 /23. [18] Slotman BJ, Njo KH, Karim AB. Curative radiotherapy for technically operable stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1994;29:33 /7. [19] Sandler HM, Curran WJ, Jr, Turrisi AT, III. The influence of tumor size and pre-treatment staging on outcome following radiation therapy alone for stage I nonsmall cell lung cancer. Int J Radiat Oncol Biol Phys 1990;19:9 /13.
.
11 [20] Haffty BG, Goldberg NB, Gerstley J, Fischer DB, Peschel RE. Results of radical radiation therapy in clinical stage I, technically operable non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1988;15:69 /73. [21] Sibley GS, Jamieson TA, Marks LB, Anscher MS, Prosnitz LR. Radiotherapy alone for medically inoperable stage I nonsmall-cell lung cancer: the Duke experience. Int J Radiat Oncol Biol Phys 1998;40:149 /54. [22] Mountain CF. Surgical treatment of lung cancer. Crit Rev Oncol Hematol 1991;11:179 /207. [23] Graham PH, Gebski VJ, Langlands AO. Radical radiotherapy for early non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1995;31:261 /6. [24] Robertson JM, Ten Haken RK, Hazuka MB, Turrisi AT, Martel MK, Pu AT, et al. Dose escalation for non-small cell lung cancer using conformal radiation therapy. Int J Radiat Oncol Biol Phys 1997;37:1079 /85. [25] Kono M, Hama Y, Kusano S. CT guided stereotactic radiation therapy for lung carcinomas using a fusion of CT and linac (FOCAL) unit. In: The fourth Japanese body stereotactic irradiation meeting. Tokyo, 2001. [26] Miyamoto T, Nishimura H, Matzuoka Y, Tujii H, Yamaguchi Y. Carbon ion radiotherapy for inoperable non-small cell lung cancer at stage I, Lung Cancer 2000;29(Suppl. 2):182.
www.elsevier.com/locate/lungcan
The role of radiotherapy in treatment of stage I non-small cell lung cancer ´na, Xueying Qiaoa, Owe Tullgrena, Ingmar Laxa, Florin Sirze Rolf Lewensohna,b,* a
Department of Oncology and Pathology, Radiumhemmet, Karolinska Hospital, SE-171 76 Stockholm, Sweden b Department of Medical Radiation Biology, Cancer Centre Karolinska R8:00, Karolinska Hospital, SE-171 76 Stockholm, Sweden Received 4 December 2002; received in revised form 3 March 2003; accepted 6 March 2003
KEYWORDS NSCLC stage I; Radiotherapy; Local control; Regional control; Prophylactic nodal irradiation; Survival
Summary Most information on results with radiotherapy (RT) for stage I non-small cell lung cancer (NSCLC) is based on retrospective studies on RT-treated inoperable NSCLC cases. Thus, the role of RT for stage I NSCLC, as a curative modality, has not yet been established. A literature search for studies on stage I non-small cell lung carcinoma (NSCLC) treated by RT alone resulted in 18 papers published between 1988 and 2000. The majority of stage I patients received RT treatment because they were medically inoperable. The main contraindications for surgery were grave impairment of pulmonary function and serious cardiovascular disease. Local recurrence was the most common reason for treatment failure (median value 40%) but varied highly between the studies, ranging from 6.4 to 70%. In contrast with local recurrence, regional failure was not a major problem (0 /3.2%). Generally, smaller tumour size, low T-stage and increased dose had a favourable impact on local control and increased local control was followed by increased survival. No serious treatment complications were recorded in the majority of these studies. Overall treatment results were, however, disappointing. The median survival in these studies ranged from 18 to 33 months. The 3- and 5-year overall survival was 349/9 and 219/8% (mean value9/1 S.E.), respectively. The cause-specific survival at 3 and 5 years was 399/10 and 259/9% (mean value9/1 S.E.), respectively. Dose escalation, in a setting with conformal RT using involved field or stereotactic RT, should be the focus of developmental therapeutic strategies with inoperable stage I NSCLC to improve local control and survival. – 2003 Elsevier Science Ireland Ltd. All rights reserved.
1. Introduction
*Corresponding author. Tel.: /46-8-5177-5092; fax: /46-85177-6630. E-mail address: [email protected] (R. Lewensohn).
Surgery is, at present, standard treatment for stage I non-small cell lung cancer (NSCLC). Most studies on long-term survival have shown a more than 50% 5-year survival after radical resection [1 / 3]. It is unclear whether radiotherapy (RT) may
0169-5002/03/$ - see front matter – 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0169-5002(03)00152-1
2
emerge as an alternative method of treatment to surgery considering efficacy parameters. RT has been offered to patients who are not suitable for surgery due to medical problems and to patients who refuse surgery. This paper presents a review of publications on results from studies of conventional RT for stage I NSCLC and addresses the following questions: (1) Can these patients be cured by RT alone? (2) What is the pattern of failure? (3) What are the factors affecting prognosis of the patients with stage I NSCLC after treatment with RT alone? (4) Is RT well tolerated by these patients? (5) Which are the future directions for the treatment of inoperable stage I NSCLC?
2. Material This review is based on 18 studies reporting results on RT for stage I NSCLC published between 1988 and 2000. The majority of these studies (13/ 18) included less than 100 patients (ranging from 31 to 93), four studies included between 100 and 150 patients and one larger study included 347 patients. The majority of these studies are retrospective analyses. Data selected from publications concerning treatment for early stage NSCLC by RT alone are listed in Table 1A and B. The percentage of medically inoperable patients varies from 59.2 to 100%. Two publications show a lower percentage of medically inoperable patients (59.2 and 64%) which is due to the inclusion of patients refusing surgery [4,5]. The main medical problems of the medically inoperable patients were poor pulmonary function, serious cardiovascular disease, poor general condition etc. It is also apparent from Table 1A that most of the patients treated by RT were elderly. Radiation dose and fraction size varied in different studies. For comparative reasons, the biologically effective dose (BED) was calculated (acute effect a /b /10) for each study using time correction whenever it was possible [6]. The BED values were calculated in order to make possible a comparison of the biological response to different fractionation schemes, rather than the physical doses given [7]. In some studies, different total doses and different doses/fractions were delivered to different patients. RT usually started with opposed AP and PA fields, followed by boost to the primary tumour after 40/50 Gy [4,8 /15]. Opposed AP/PA fields and oblique fields (avoiding the spinal cord) were also used [5,6,16]. There was no consistent use of prophylactic node irradiation (Table 1B). Notably, there are studies were only some of the
X. Qiao et al.
patients are reported to have received prophylactic node irradiation.
3. Results and discussion 3.1. Factors of importance for local and regional control When using RT one of the main issues is local control. Local recurrence was the most common reason for treatment failure with conventional RT of stage I NSCLC, but the frequency of recurrence varied considerably according to different reports (between 6.4 and 70%) [4,6,10,14 /20] as can also be seen in Table 3. Distant metastases should, however, not be underestimated, since in some series distant failure was more common than local failure [9,10,21]. On the other hand, regional node failure alone occurred alone only in 0 /3.2% of the cases and is, therefore, not a major problem [4,6,9,10,14 /21]. Follow up of treatment-results in lung cancer is often done on a 2-, 3- and 5-year basis. Since the reporting at these time intervals varies between different authors it is not possible to extract the median survival for the patients in all studies (Table 2). Median time to relapse varied from 21 to 30 months. In Fig. 3, the cumulative probability of recurrence is plotted according to the data obtained from two publications [6,9]. The 3-year recurrence rate in these two studies were 60 [9] and 67.1% [6]. The majority of recurrences occurred within 3 years after treatment. Thus, according to these two studies a 3-year follow up is needed to estimate the recurrence rate after RT of stage I NSCLC. 3.1.1. Tumour parameters having an impact on local control When using RT it is generally thought that the size of a tumour will predict the possibility of obtaining complete remission (CR). When reviewing the papers that we refer to wide range of CR rates are reported, ranging from 17 to 61% [4,8,9,14 /16]. Out of a few papers one may conclude that smaller tumours, as expected, had a higher CR rate than larger ones. Smaller tumours, as expected, had a higher CR rate than larger ones [14 /16]. It should be understood that the time window for evaluating response is short in RT of the lung since longer follow up of remission will be hampered by radiation caused fibrosis. The evaluations of response, therefore, refer to those performed at the end of the RT treatment.
Radiotherapy, inoperable stage I NSCLC
3
Table 1 Author (year)
Age (years) Mean/median
(A ) Patients characteristics Slotman et al., 1996 75/ / [9] Zhang et al., 1989 [7] 57.1/ / Hayakawa et al., 1999 n.s [8] Haffty et al., 1988 64/ / [19] Kupelian et al., 1996 n.s [10] Slotman et al., 1994 75 [17] Jeremic et al., 1997 //63 [4] Noordijk et al., 1988 74/ / [13] Kaskowitz et al., 1993 //73 [16] Talton et al., 1990 65/ / [11] Sandler et al., 1990 //72 [18] Cheung et al., 2000 //71.5 [6] Graham et al., 1995 //67 [22] Dosoretz et al., 1992 //74 [12] Krol et al., 1996 [15] 74.2/ / Sibley et al., 1998 [20] //70 Morita et al., 1997 74.2/74.6 [14] Gauden et al., 1995 //70 [5] Total Total dose (Gy)
Number of patients
Medically inoperable patients (%)
Range
Stage I
All
54 /85
31
31
81
37 /70 n.s
35 36
44 36
72.7 88.9
51 /83
43
43
86
n.s
45
71
100
57 /93
47
47
n.s
51 /70
49
49
59.2
n.s
50
50
84
44 /86
53
53
81.8
40 /89
60
77
n.s
n.s
77
77
78
45 /86
92
102
76.5
43 /89
93
103
70
51 /88
106
152
88.8
58 /88 46 /95 50 /89
108 141 149
108 141 149
88 100 83
34 /90
347
347
64
Fractions (F)
1562 Acute BED (Gy)
1680 D2 Prophylactic node irradiation yes/no
In general the term local control is used by different authors to describe whether or not there is tumour progression in the irradiated volume of the chest. This, of course, must mean that there are some uncertainties in the determination of local control. Such uncertainties may influence the reported frequencies of local control. Another way of assessing local control is trying to understand the frequency of local failure in relation to e.g. tumour size. A general relation between tumour size and local failure seems to be apparent [11,15,18]. The authors presenting such data generally use arbitrary cut-offs in tumour size of which
one example is the work by Slotman et al. who shows that total local failure rate was 2.4 and 47.8% for tumours 5/4 and /4 cm, respectively [18]. Interestingly, Slotman et al. found no correlation between tumour size and loco-regional node failure [10]. Another way of assessing the relation between local control and local tumour extent in stage I disease is to use T-stage. Again in three papers [4,11,13] a better local control was shown to be obtained with T1 as compared with T2 tumours. In one of these, the Kupelian et al. study the difference was not statistically significant, a find-
4
X. Qiao et al. Table 1 (Continued )
Author (year)
Age (years) Mean/median
Number of patients Range
Stage I
Medically inoperable patients (%) All
(B ) Compilation of publications on treatment of early stage NSCLC by RT alone Slotman et al., 1996 48 12 F 76.4 [9] 2 Gy/F 62.8 Zhang et al., 1989 [7] 55 /70 Hayakawa et al., 1999 60 /81 2 Gy/F 68.1 67.1 [8] Haffty et al., 1988 54 6 weeks/split 59 62.2 [19] course 59 6 /7 weeks/continuous Kupelian et al., 1996 63.23 65 60 [10] Slotman et al., 1994 32 Gy/6 F, 40 Gy/ 28/19 [17] 10 F 48 Gy/12 F, 56 Gy/ 20 F Jeremic et al., 1997 69.6 1.2 Gy bid 71 58.5 [4] Noordijk et al., 1988 60 Split course, 3 Gy/ 63.4 72 [13] F 2 Gy/F 65.1 63 Kaskowitz et al., 1993 63.2a [16] Talton et al., 1990 60 30 F 62.8 60 [11] 2 Gy/F 62.8 60 Sandler et al., 1990 60a [18] Cheung et al., 2000 52.5 20 F 67.8 59 [6] Graham et al., 1995 60 30 F 62.8 60 [22] 1.8 /2 Gy/F 62.8 60 Dosoretz et al., 1992 45 /75 [12] Krol et al., 1996 [15] 60 7 weeks, 3 Gy bid 63.4/74.3 72 65/5.2 weeks, 2.5 qd Sibley et al., 1998 [20] 64a 1.2 bid, 3 qd Morita et al., 1997 64.7a 2 Gy/F 65.3 64.7 [14] Gauden et al., 1995 50 20 F/4 weeks 62.5 55 [5] a
0/31 44/0 10/26 38/5
51/20
49/0 0/50 n.s. 77/0 69/8 0/102 82/21 n.s. 0/108
103/38 66/88 347/0
Median, D2 is the total dose given with 2 Gy per fraction, calculated with the BED formalism.
ing which may be explained by an overlap in tumour size between the different T-stages [11]. Looking at the data there is a definite trend for a lowered local control with increasing tumour size. The lack of a statistical difference between tumour sizes in relation to local control may be based on common problems encountered when analysing retrospective studies such as limited number of
patients and radiation given with conventional dose delivery, i.e. larger tumours may have received higher radiation doses. The dependence of local control on histology is in general not handled by the different papers. Most likely this relates to improperties in the technology of differentiating between different histologies of NSCLC using sputum cytology etc.
Radiotherapy, inoperable stage I NSCLC
5
23
48
41
33
32
41 28
5-year 3
32 32
3.1.2.1. RT dose There is support for better local control when total dose is increased [11,17]. This conclusion is based on dividing the patient groups into those receiving less than 60 /65 Gy and those receiving higher doses. In addition, intrathoracic recurrence rate was significantly lower in patients who received the high dose schedule (2/24: 8.3%) compared with the low dose group (10/23: 43.5%), with a statistically significant difference (P B/0.01) in the Slotman et al. series [18]. We were able to calculate acute BED for nine different studies (Table 4). As can be seen in Fig. 4 there is a trend for better local control with increased BED [4,9,10,14,15,17,20]. The relatively limited span of doses (BED 59 /76 Gy) delivered does not enable a clear-cut conclusion of the role of dose in obtaining local control. In addition, we cannot exclude the possibility that in those studies presenting the highest BED-values, tumours treated were of a larger size. Taken together the data indicate that better local control is obtained with the higher doses used in these studies.
54
34.2 40
21 16 22.2 27 77.9
70
39
56
33 17 36
43 72
20.9 33 33 18 20 20 28 27 27.2 27.9
81
31.4 15 23 6 8 30 13 15
Zhang et al. [7] Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Gauden et al. [5]
91.4 75
57.1 31 42 19 42
90
81
60
54 93
53 22
42 56 33
31 39 13
2 5-year 3 2 1 5-year 3 2 1
Median survival (month) Authors
Table 2
Survival after RT for stage I NSCLC
Overall survival (%)
Cause-specific survival (%)
Disease-free survival (%)
3.1.2. Modalities of radiotherapy and local control
3.1.2.2. RT volume (issue of margins, elective nodal irradiation) One obvious question when irradiating lung tumours is which field margins to accept. For most series, the margin was at least 1.5 cm between the gross tumour and field border. For groups of patients with margins 5/1.5 cm, 48% of the tumours recurred at their primary location, while 35% of the patients had local recurrence when the margin were /1.5 cm [17]. In Cheung et al.’s series, 90.3% of the margins were larger than 1.5 cm and could not be considered a prognostic factor for recurrence-free survival [6]. Thus, there are no conclusive data to indicate optimal margins for local control. Regional failure alone is not a common cause of failure of stage I NSCLC treated by RT only. Locoregional relapse occurred only in 0 /3.2% of the patients treated without prophylactic node irradiation (see Table 3 and [10,14,16,18]). It was, however, reported to be higher but not statistically significant in a non-prophylactic mediastinal irradiation group than that of a prophylactic mediastinal irradiation group [21].
6
X. Qiao et al.
Table 3 The patterns of failure after RT for stage I NSCLC Author
Local failure (%) Alone
Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Weighted mean a b
27.8 11.1 41.5 0 n.s. 16.3 19.1 42.8 39 n.s. n.s. 17.2
a
Regional failure alone (%) b
Total 65.7 19.4 43.4 6.4 45 19.1 25.2 42.8 39 70 44.3 28.6
1.8 2.8 0 3.2 0 2.8 n.s. n.s. n.s. 0 0 1.0
Distant failure (%) Alone
All
2.8 33.3 n.s. 3.2 n.s. 14.9 21.3 n.s. 11.6 26 n.s. 9.3
33.3 36.1 32 16.1 25 19.8 21.3 n.s. 11.6 26 12.1 15.7
Single site failure. Any local failure including combined, regional and distant failure.
Fig. 1 Overall survival of stage I NSCLC (the estimated 3, 5-year overall survival rate is 349/9 and 219/8%, respectively). The data were extracted from the studies presented in Table 2.
3.2. Which factors are prognostic for survival The reports on treatment results with stage I NSCLC present large variations of patient survival. The survival after RT treatment when considered only for stage I NSCLC of different studies is shown in Table 2. The median survival varied between 18 and 33 months. The overall survival rate of 1-, 2-, 3- and 5-year survival rate were 70 /91.4, 39 /72, 16 /57.1 and 6 /31.4%, respectively. In this heavily compromised group of patients with short life expectancy it is more appropriate to use data on cause specific survival rather than overall survival. The 2-, 3- and 5-year cause-specific survivals were 54 /93, 22/56, and 13/39%. The 5-year disease-
Fig. 2 Cause-specific survival of stage I NSCLC (the estimated 3, 5-year cause-specific survival is 399/10 and 259/9%, respectively). The data were extracted from the studies presented in Table 2.
free survival was 23 /41%. The overall survival rates and cause-specific survival rates are presented in Figs. 1 and 2. The estimated 3- and 5-year overall survival was 349/9 and 219/8% (mean value9/1 S.E.), respectively. The mean value was calculated considering the number of patients and the S.E. values were calculated for a group of 100 patients. The estimated cause-specific survival at 3 and 5 years was 399/10 and 259/9%, respectively. This can be compared with the roughly 60% 5-year survival rate of cases that undergo surgery [22]. 3.2.1. Is tumour volume in RT treated stage I NSCLC of importance for survival When evaluating treatment results, survival may be another end point not entirely linked to local or loco-regional control and should, therefore, be evaluated separately. Thus, the importance of
Radiotherapy, inoperable stage I NSCLC
Fig. 3 Cumulative probability of recurrence after RT for early stage NSCLC (the data with ‘‘'’’ and ‘‘I’’ was from reference [7,18], respectively).
tumour size was analysed with respect to survival. The median survival time for patients with a tumour size 5/4 cm was longer than for those with tumours /4 cm (32 vs. 12 months; P /0.007) [14]. In Sandler et al.’s series [19], the overall survival for tumours of B/3, 3 /6 and /6 cm was 25, 13 and 0% at 3 years (P /0.009). The causespecific survival at 3 years was 30, 17 and 0%, respectively (P /0.0001). No relationship was, however, found between tumour size and survival in Kaskowitz et al.’s report [17], although the material is composed of similar tumour sizes as in e.g. Sandlers report. Krol et al. found that the cause-specific survival for tumours of 5/4 or /4 cm was 40 and 13%, respectively, at 3 years.
Fig. 4 Local-component failure (including local failure alone, local plus regional failure and local plus distant failure) vs. BED.
7
Similarly to the analysis of which factors are important for local control we used T-stage as surrogate for tumour volume when analysing the relation to survival. Zhang et al. reported that 66.7% of T1N0 patients but only 24% of T2N0 patients survived more than 5 years [8]. Gauden et al. found that the disease-free survival at 5 years was 32% for T1N0 tumours compared with 18% for T2N0 tumours, P B/0.01 [5]. Others have reported similar results [4,18]. There was, however, no significant relation between T-stage and overall survival in one study [10]. It is hard to draw firm conclusions on the exact importance of tumour size for survival. The general trend, however, is that the prognosis worsens with increasing tumour size. It is our contention that the analysis is more strict when using direct measurements of tumour size compared with the T-staging procedure. 3.2.2. The role of prophylactic nodal irradiation Only one study reported better overall survival in a patient group treated with prophylactic nodal irradiation compared with those that did not receive this treatment [15]. However, the authors did not report an increased rate of regional failure in the non-irradiated group. The results may be influenced by the imbalance of non-squamous histology between the groups (51% in the nonirradiated and 31% in the irradiated group) which may explain the higher rate of distant metastatic events in the non-irradiated group. In four other studies, no improvement of overall and causespecific survival was observed after prophylactic irradiation of mediastinal nodal groups [4,11,18,21]. In studies not using elective nodal irradiation the reported regional failure rates were remarkably low [10,14,16,18]. Taken together these data argue against the use of prophylactic RT of the mediastinum in stage I NSCLC. 3.2.3. Total radiation dose and survival Total dose is an important prognostic factor for stage I NSCLC. Dosoretz et al. [13] reported that the 2 year disease-free survival of patients receiving radiation doses of 70 Gy or more, between 60 and 70 Gy and between 50 and 60 Gy was 50, 33 and 22%, respectively. Patients presenting T1 tumours and who received radiation doses of ]/65 Gy experienced a disease-free survival at 2 years in approximately 73% of the cases compared with 43% at 2 years for patients receiving less than 65 Gy. Similar dose dependence was also noted for T2 patients [13]. The disease-free survival at 3 years for patients having tumours less than 3 cm in diameter and treated with a dose higher than 65 Gy was found to be 90% [13]. Zhang et al. observed
8
X. Qiao et al.
Table 4 BED in relation to local failure, 5-year overall and cause-specific survival after RT for stage I NSCLC Author
BED (acute)
Local failure alone (%)
Any local failure (%)
Overall survival (%)
Cause-specific survival (%)
Krol et al. [15] Hayakawa et al. [8] Kaskowitz et al. [16] Slotman et al. [9] Jeremic et al. [4] Sibley et al. [20] Slotman et al. [17] Sandler et al. [18] Haffty et al. [19] Noordijk et al. [13] Morita et al. [14] Gauden et al. [5]
/ 68.1
27.8 11.1
65.7 19.4
15 23
31 39
65.1
41.5
43.4
6
13
76.4 71 / / 62.8 59 63.4 65.3 62.5
0 / 16.3 19.1 42.8a 39 / / /
6.4 45 19.1 25.2 42.8a 39 70 44.3 /
8 30 13 15 17a 21 16 22.2 27
76b / 32 32 22a / / / /
a b
3-year survival. 4-year survival.
that a higher dose was associated with good prognosis [8]. In that study the 5 year overall survival of the patients receiving 69 /70 Gy was higher than that of patients receiving 55 /61 Gy (36 vs. 27%). While the two previous reports favour dose-relationship, the latter study does not. It is, however, important to understand that the lack of correlation in Sandlers study may, as pointed out by the author himself, very well be related to the narrow range of doses used. Also, another factor adding to the problem of obtaining a linear dose response with survival is the possibility that larger tumours by themselves confer less favourable prognosis as compared with smaller tumours, in spite of the possibility that they may have received a higher dose compared with smaller tumours.
3.2.4. Is survival linked to local control There was an improved 5-year cause-specific survival of patients in whom the primary tumour was controlled (46% survival) compared with those with local failure (12% survival), P B/0.03 [21]. On the other hand a trend but not a statistically significant correlation was found for local control in relation to cause specific survival when analysing the articles listed in Table 4 [9,16 /18,21]. Patients achieving CR after RT had a high overall survival rate and local control [8]. In Krol et al.’s report [16], out of the 50 patients who achieved CR, 31(60%) had long-term local control. The actuarial local relapse free survival at 1, 3 and 5 years were 96, 71 and 66%, respectively. All 58 patients not achieving CR had local progression. Taken together, these data indicate that the chance of long term
survival is linked to the achievement of local control.
3.3. Comparison between radiotherapy and surgery No randomised study using modern RT addresses this question. One study presented a comparison of survival for elderly ( ]/70 years) NSCLC patients treated with curative RT or surgery, but no statistical difference was, however, recorded between these groups [14]. This study must then, be judged with caution since it was a retrospective analysis. Due to medical problems such as impairment of lung function and cardiac diseases affecting survival of patients treated with RT it is difficult to compare results between RT and surgery [8]. Interestingly, it has been reported that patients refusing surgery had a better chance for long-term survival when treated by RT compared with expected surgery outcome [8].
3.4. Side effects of radiotherapy alone Analysis of the relationship between the dose to the organs at risk and the side effects was not possible because no data about dose (BED) to the organs at risk were available in the publications. Table 5 shows that there is no significant relationship between the dose (BED) prescribed to target and the side effects. This lack of relationship may partly be due to the different criteria and different methods utilised in the different studies and partly due to the fact that the majority of the patients received treatment with similar BED values. In
Radiotherapy, inoperable stage I NSCLC
9
Table 5 Side effects of RT for stage I NSCLC Author
Number of patients
BED acute
Slotman et al. [9] Zhang et al. [7] Hayakawa et al. [8] Haffty et al. [19] Kupelian et al. [10] Slotman et al. [17] Jeremic et al. [4] Noordijk et al. [13] Kaskowitz et al. [16] Talton et al. [11] Sandler et al. [18] Cheung et al. [6] Graham et al. [22] Dosoretz et al. [12] Krol et al. [15]
31
76.4
Well tolerated
112
44 36
62.8 68.1
18% G2 peumonitis No serious complication
100 111.7
43
59
No
71
65
G2 pneumonitis 9.3% G2 esopha- 103.5 gitis 18.6% 7% G2 pneumonitis 100
47
/
Moderate esophagitis
/
2.1% G2 pneumonitis
49
71
97.4
6.1% G3 pneumonitis 6.1% G3 esophagitis
50
63.4
4.1% G3 pneumonitis 4.1% G3 esophagitis No serious side effects
53
65.1
77
Sibley et al. [20] Morita et al. [14] Gauden et al. [5]
BED late No serious complication
120 105
5.7% Moderate pneumonitis
62.8
Mild 64.2 esophagitis 3.8% moderate esophagitis 20% G2 Esophagitis
100
Pulmonary fibrosis 1.3%
77
62.8
Rare
100
102
67.8
22.5% esophagitis
98.4
103
62.8
100
152
62.8
Pneumonitis G3:1% Esophagitis G1:37.9%, G2:2.9% No major complication
100
108
Mild
120
141
63.4/ 73.4 /
1.5% G 3 /5 pneumonitis
/
149
65.3
No reported
107.8
347
62.5
No serious complication
91.5
several of the studies the RTOG scale for toxicity assessment was used. The lack of RTOG conformed toxicity scaling in some of the studies presents an additional difficulty in evaluating these results. Unfortunately no dose volume histograms were presented in these studies, which is something warranted for the future. In most series, however, there were no serious complications. The most frequent early side effect was acute esophagitis. Mild radiation esophagitis occurred in about 22.5 / 64.2% [6,17,23], moderate in 2.9 /20% [12,17,20,23] and severe in about 4.1% of the cases [4]. Serious acute radiation pneumonitis was noted in 1 /4.1% of the patients. Severe late radiation esophagitis occurred rarely and was reported in one article only, as occurring in 6.1% of the patients [4]. In patients with prophylactic irradiation of mediastinum, moderate esophagitis was noted, but
4.9% Pneumonitis, 2.9% G2 fibrosis Peumonitis G2:1.9%
not described in detail [18]. In the Cheung series, esophagitis was also recorded in 22.5% of the cases [6]. The differences in this type of complication may be due to differences in field arrangements even if involved field technique was used. Moderate late radiation pneumonitis was found in 1.9 / 18% of the cases [6,8,17,18,23] and severe late radiation pneumonitis (grade 3 or more) in 6.1% of the cases [4,21]. Late complication in the form of lung fibrosis was seen in 1.3 and 2.9% of the cases [6,12]. These complication frequencies are surprisingly low and probably related to whether or not follow-up X-rays were performed. Sibley et al. reported two patients with grade 3 /5 pneumonitis out of 141 patients (1.5%), both of which received prophylactic mediastinal irradiation [21]. Small fields (7.5 /7.5 cm) did not interfere with the respiratory capacity [13]. There was no difference
10
between complications in patients receiving radiation to the primary tumour only and those patients receiving radiation to the tumour and lymph nodes [11]. Also, there is no evidence to suggest that the use of larger portal fields correlated with a higher incidence of pneumonitis [17]. All in all, since these studies are retrospective, the exact frequency of complications is hard to assess. Only prospective well-controlled clinical studies may give us exact figures on the incidence of complications with RT.
X. Qiao et al.
Considering the high local failure and the low regional failure after conventional RT treatment, dose escalation using conformal RT or stereotactic RT with ‘‘involved field technique’’ or light ion therapy is warranted in stage I NSCLC with the aim to improve local control and survival.
Acknowledgements This study was supported by grants from the Stockholm Cancer Society and the Swedish Heart Lung Foundation.
4. Conclusions and future directions
References
A minority of patients diagnosed with stage I NSCLC can be cured by conventional RT alone. The most common reason for treatment failure is local recurrence alone. Regional failure in the mediastinum is surprisingly uncommon indicating that prophylactic irradiation of the mediastinum may not be necessary in these patients. In general, smaller tumour-size, lower T stage and higher dose, correlates with better local control and survival. The 5-year overall survival and cause-specific survival were 219/8 and 259/9%, respectively, in cases treated with conventional RT. Thus, whilst these results compare unfavourably with surgery, one must keep in mind that the majority of patients who have been offered RT for stage I NSCLC are heavily compromised by their medical condition. This review indicates that higher RT doses correlate with better local control and survival. It is our belief that conformal 3-D RT or stereotactic RT may better conform to the prescribed dose to the actual target volume and thereby spare normal tissues permitting dose escalation. In stage I /III NSCLC dose escalation using conformal 3-D RT has been feasible [24]. Attempts have now been made to use stereotactic radiation therapy in primary lung cancer. With an acute BED of at least 75 Gy delivered by stereotactic RT, with or without conventional RT, the rates of overall survival, cause-specific survival and local control at 3 years for early stage NSCLC has been reported to be 68, 90 and 96%, respectively [25]. This is higher than the results obtained by conventional RT and also by surgery. The use of charged particle beams is another option. In patients with stage I NSCLC treated with a carbon ion beam using involved field technique, tumour control rate increased linearly with increasing carbon ion dose up to 100% local control [26].
[1] Martini N, Rusch VW, Bains MS, Kris MG, Downey RJ, Flehinger BJ, et al. Factors influencing 10-year survival in resected stages I to IIIa non-small cell lung cancer. J Thorac Cardiovasc Surg 1999;117:32 /6 (Discussion 7 /8). [2] Williams DE, Pairolero PC, Davis CS, Bernatz PE, Payne WS, Taylor WF, et al. Survival of patients surgically treated for stage I lung cancer. J Thorac Cardiovasc Surg 1981;82:70 / 6. [3] Naruke T, Goya T, Tsuchiya R, Suemasu K. Prognosis and survival in resected lung carcinoma based on the new international staging system. J Thorac Cardiovasc Surg 1988;96:440 /7. [4] Jeremic B, Shibamoto Y, Acimovic L, Milisavljevic S. Hyperfractionated radiotherapy alone for clinical stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1997;38:521 /5. [5] Gauden S, Ramsay J, Tripcony L. The curative treatment by radiotherapy alone of stage I non-small cell carcinoma of the lung. Chest 1995;108:1278 /82. [6] Cheung PC, Mackillop WJ, Dixon P, Brundage MD, Youssef YM, Zhou S. Involved-field radiotherapy alone for earlystage non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2000;48:703 /10. [7] Fowler JF. Brief summary of radiobiological principles in fractionated radiotherapy. Semin Radiat Oncol 1992;2:16 / 22. [8] Zhang HX, Yin WB, Zhang LJ, Yang ZY, Zhang ZX, Wang M, et al. Curative radiotherapy of early operable non-small cell lung cancer. Radiother Oncol 1989;14:89 /94. [9] Hayakawa K, Mitsuhashi N, Saito Y, Nakayama Y, Furuta M, Sakurai H, et al. Limited field irradiation for medically inoperable patients with peripheral stage I non-small cell lung cancer. Lung Cancer 1999;26:137 /42. [10] Slotman BJ, Antonisse IE, Njo KH. Limited field irradiation in early stage (T1-2N0) non-small cell lung cancer. Radiother Oncol 1996;41:41 /4. [11] Kupelian PA, Komaki R, Allen P. Prognostic factors in the treatment of node-negative non-small cell lung carcinoma with radiotherapy alone. Int J Radiat Oncol Biol Phys 1996;36:607 /13. [12] Talton BM, Constable WC, Kersh CR. Curative radiotherapy in non-small cell carcinoma of the lung. Int J Radiat Oncol Biol Phys 1990;19:15 /21. [13] Dosoretz DE, Katin MJ, Blitzer PH, Rubenstein JH, Salenius S, Rashid M, et al. Radiation therapy in the management of medically inoperable carcinoma of the lung: results and implications for future treatment strategies. Int J Radiat Oncol Biol Phys 1992;24:3 /9.
Radiotherapy, inoperable stage I NSCLC [14] Noordijk EM, vd Poest Clement E, Hermans J, Wever AM, Leer JW. Radiotherapy as an alternative to surgery in elderly patients with resectable lung cancer. Radiother Oncol 1988;13:83 /9. [15] Morita K, Fuwa N, Suzuki Y, Nishio M, Sakai K, Tamaki Y, et al. Radical radiotherapy for medically inoperable non-small cell lung cancer in clinical stage I: a retrospective analysis of 149 patients. Radiother Oncol 1997;42:31 /6. [16] Krol AD, Aussems P, Noordijk EM, Hermans J, Leer JW. Local irradiation alone for peripheral stage I lung cancer: could we omit the elective regional nodal irradiation. Int J Radiat Oncol Biol Phys 1996;34:297 /302. [17] Kaskowitz L, Graham MV, Emami B, Halverson KJ, Rush C. Radiation therapy alone for stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1993;27:517 /23. [18] Slotman BJ, Njo KH, Karim AB. Curative radiotherapy for technically operable stage I non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1994;29:33 /7. [19] Sandler HM, Curran WJ, Jr, Turrisi AT, III. The influence of tumor size and pre-treatment staging on outcome following radiation therapy alone for stage I nonsmall cell lung cancer. Int J Radiat Oncol Biol Phys 1990;19:9 /13.
.
11 [20] Haffty BG, Goldberg NB, Gerstley J, Fischer DB, Peschel RE. Results of radical radiation therapy in clinical stage I, technically operable non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1988;15:69 /73. [21] Sibley GS, Jamieson TA, Marks LB, Anscher MS, Prosnitz LR. Radiotherapy alone for medically inoperable stage I nonsmall-cell lung cancer: the Duke experience. Int J Radiat Oncol Biol Phys 1998;40:149 /54. [22] Mountain CF. Surgical treatment of lung cancer. Crit Rev Oncol Hematol 1991;11:179 /207. [23] Graham PH, Gebski VJ, Langlands AO. Radical radiotherapy for early non-small cell lung cancer. Int J Radiat Oncol Biol Phys 1995;31:261 /6. [24] Robertson JM, Ten Haken RK, Hazuka MB, Turrisi AT, Martel MK, Pu AT, et al. Dose escalation for non-small cell lung cancer using conformal radiation therapy. Int J Radiat Oncol Biol Phys 1997;37:1079 /85. [25] Kono M, Hama Y, Kusano S. CT guided stereotactic radiation therapy for lung carcinomas using a fusion of CT and linac (FOCAL) unit. In: The fourth Japanese body stereotactic irradiation meeting. Tokyo, 2001. [26] Miyamoto T, Nishimura H, Matzuoka Y, Tujii H, Yamaguchi Y. Carbon ion radiotherapy for inoperable non-small cell lung cancer at stage I, Lung Cancer 2000;29(Suppl. 2):182.