- Email: [email protected]
Effect of interfraction interval in hyperfractionated radiotherapy with or without concurrent chemotherapy for Stage III nonsmall cell lung cancer
lnt. J. Radiation
Oncology
Biol.
Phys.. Vol. 34, No. 2, pp. 303-308, 1996 Copyright 10 1996 Elsevier Science Inc. Prinrcd in Ihe USA. All rights reserved
0360-30
I h/96
$ IS .oO + .(X1
ELSEVIER
0360-3016( 95)02158-2
l
Clinical Original Contribution EFFECT OF INTERFRACTION INTERVAL IN HYPERFRACTIONATED RADIOTHERAPY WITH OR WITHOUT CONCURRENT CHEMOTHERAPY FOR STAGE III NONSMALL CELL LUNG CANCER BRANISLAV
JEREMIC,
M.D.,
PH.D. * AND YUTA
SHIBAMOTO,
M.D.,
D.M.Sc.“f
*Department of Oncology, University Hospital, Kragujevac, Yugoslavia and TDepartment of Oncology, Chest Disease Research Institute, Kyoto University, Kyoto 606-01, Japan Purpose: To explore the influence of interfraction interval in hyperfractionated radiotherapy (HFX RT) withwithout concurrent chemotherapy for Stage III nonsmall cell lung cancer. Methods and Materials: One hundred sixty-nine patients treated in a randomized study were retrospectively analyzed. Group I patients were treated by HFX RT with 1.2 Gy twice daily with a total dose of 64.8 Gy in 27 treatment days, while Groups II and III patients were treated by the same HFX RT and concurrent chemotherapy with carboplatin and etoposide (every week in Group II and every other week in Group III). Interfraction intervals of either 4.5-5 h or 5.5-6 h were used for each patient. Results: Patients treated with shorter interfraction intervals (4.5-5 h) had a better prognosis than those treated with longer intervals (5.5-6 h) (median survival: 22 vs. 7 months; 5-year survival rate: 27% vs. 0%, p = 0.00000). This phenomenon was observed in all treatment groups. Patients 2 60 years of age, with Stage IIIA disease, or with previous weight loss 5 5% were treated more often with the shorter intervals than those < 60 years, with Stage IIIB disease, or with weight loss > 5%, respectively, but in all of these subgroups of patients, the shorter intervals were associated with a better prognosis. Multivariate analysis showed that the interfraction interval was an independent prognostic factor, together with sex, age, performance status, and stage. The shorter intervals were associated with an increased incidence of acute high grade toxicity, but not with an increase in late toxicity. Conclusion: Patients treated with shorter interfraction intervals (4.5-5 h) appeared to have a better survival than those treated with longer intervals (5.5-6 h). Prospective randomized studies are warranted to further investigate the influence of interfraction interval in HFX RT. Radiotherapy,
Hyperfractionation,
Interfraction
interval, Nonsmall cell lung cancer. Locally advanced nonsmall cell lung cancer (NSCLC) has been frequently treated with MFD RT (1-3, 15-17). All reports analyzed survival and toxicity but no reports considered the role of interfraction interval in treatment outcome. We carried out a randomized study comparing hyperfractionated (HFX) RT alone and HFX RT with concurrent chemotherapy for Stage ill NSCLC. ln this trial, various interfraction intervals in the range of 4.5-6 h were used but the interval was kept constant for each patient. Therefore, we performed a retrospective analysis of these patients to explore the influence of interfraction interval on treatment outcome.
INTRODUCTION Multiple fraction-per-day ( MFD) radiation therapy (RT) is usually given in twice daily fractions and interfraction intervals of 4-8 h have been used in most clinical trials. These intervals have been chosen based on the repair kinetics of normal tissues (but not tumor). Biological data suggest that 3-4 h intervals are not sufficient to allow for full repair of radiation injury in many systems (5, 7, 18) and this is already borne out by several clinical studies in which intervals shorter than 4.5 h produced more morbidity than did the longer intervals ( 11, 12, 15). Based on these findings on normal tissues, interfraction intervals of 6 h or longer are generally recommended currently ( 14). However, it is not known what range of interfraction interval is optimal in terms of tumor response.
METHODS
AND MATERIALS
One hundred sixty-nine patients with Stage ill NSCLC who entered a randomized trial comparing HFX RT alone
Reprint requests to: Yuta Shibamoto, M.D., Department of Oncology, Chest Disease Research Institute, Kyoto University, Kyoto 606-01, Japan.
Accepted for publication 19 September 1995.
303
304
1. J. Radiation Oncology 0 Biology 0 Physics Table 1. Distribution
of pretreatment
Volume 34, Number 2, 1996 prognostic
Group I 4.5-5
h
factors by interfraction
Group II
5.5-6
h
4.5-5
h
interval
Group III
5.5-6
h
4.5-5
h
Total
5.5-6
h
4.5-5
h
5.5-6
h
Sex M:F
22~7
24:8
15:14
14:18
6:19
5:24
9:23
3122
Stage 1IIA:IIIB
19:lO
13:19
Weight loss s5%:>5%
16:13
7:25
Age
<60:260
Karnofsky performance ~70:r80
17:8
2217
21:6
61:22 p = 0.39
68:18
18:9
12:17
16:ll
33:50 p = 0.037
48:38
status 7:20
5:24
7:20
13:70 p = 0.079
23~63
14:ll
11:16
16:13
12:15
49134 p = 0.026
36:50
15:lO
4~23
15:14
6:21
and HFX RT plus chemotherapy at University Hospital, Kragujevac in 1988 and 1989 were’ analyzed. The details of this study were described previously (9). Briefly, adult patients with Stage III NSCLC and a Karnofsky performance status (KPS) score of at least 50% were treated with: HFX RT alone to a total dose of 64.8 Gy with 1.2 Gy b.i.d. fractions in 27 treatment days (Group I, 62 patients) ; with the same HFX RT and concurrent chemotherapy consisting of 100 mg of carboplatin on Days 1 and 2, and 100 mg of etoposide on Days l-3 of each week during the RT course (Group II, 52 patients) ; and with the same HFX RT and concurrent chemotherapy consisting of 200 mg of carboplatin on Days 1 and 2, and 100 mg of etoposide on Days l-5 of the first, third, and fifth weeks during the RT course (Group III, 56 patients). The drugs were administered between the two daily fractions: 3-4 h after the first one and l-2 h before the second one. The patient characteristics are shown in Table 1. The RT method was also described previously (9). Briefly, the initial target volume including tumor plus a 2-cm margin and a major part of mediastinum was treated to 50.4 Gy, after which the RT field was reduced to cover detectable tumors. Interfraction intervals of 4.5-5 h or 5.5-6 h were nonrandomly assigned to each patient and under no circumstances was it allowed to change from the shorter interval to the longer interval and vice versa. The avearge interfraction interval was 4.5 h in 40 patients, 5 h in 43, 5.5 h in 40, and 6 h in 46. Radiation-induced effects on normal tissue were assessed as either acute or late phenomena, according to the criteria of the Radiation Therapy Oncology Group ( 3). Differences in the patient characteristics and incidence of toxicity between groups were evaluated by the &i-square test. Survival rates were calculated from the date of randomization by
’ HALBAU,
23:4
Gendaisuugakusha,
Kyoto, Japan.
46:37 17:69 p = o.OOOOO2
the Kaplan-Meier method and differences between survival curves were analyzed by the generalized Wilcoxon test. The interaction of each prognostic factor and their effect on survival were analyzed using the Cox proportional hazards model. All statistical analyses were carried out using a computer program.’
RESULTS All 169 patients were evaluated for this analysis. Survival data for the three treatment groups were reported previously (9). Briefly, Group II patients had a better prognosis than Group I patients, but there were no differences between Groups 11 and III and Groups I and III (Table 2). As a whole, patients treated with the shorter intervals had a better prognosis than those treated with the longer intervals (p = 0.00000) (Fig. 1). They also had a longer median time to first (local or distant) relapse (MTFR) and a better first relapse-free survival (FRFS ) (MTFR: 20 vs. 5 months; l-, 3-, and 5-year FRFS: 64%, 28%, and 26% vs. 31%, O%, and 0%; p = 0.00000) (Fig. 2). Because differences between survival and FRFS were constant throughout the study, all further analyses were performed with regard to survival. The difference in survival due to interfraction interval was also seen in all treatment groups (Fig. 3) and it was the most striking in Group II patients receiving concurrent weekly chemotherapy. The survival rates of the patients treated with 5.5-6 h intervals were similar among the three groups (I vs. II, p = 0.083; I vs. III, p = 0.78; and II vs. III, p = 0.33), whereas the survival of Group II patients treated with 4.5-5 h intervals was significantly better than that of Group I or III patients treated with
305
Effect of interfraction interval in hyperfractionated radiotherapy l B. JEREMIC AND Y. SHIBAMOTO
Table 2. Survival data and statistical analysis for various potential prognostic factors
Variable
n
Sex Male Female Age <60 260
Survival fate (%)
MST (months)
1 yr
3 yr
5 Yr
Univariate p
Multivariate p
0.00012
O.ooooOl
0.00000
0.000081
O.OOOOO
O.OOOOO1
31 2.8
o.ooooo
0.080
27 2.4
24
O.OOOOO
o.OoOo59
39 73 50
6.6 23 16
4.9 21 16
0.0028* 0.33' 0.061'
0.33
71 36
30
27
O.OOOOO
O.OOOOOO
129
10
40
25
46 78
7.7 38
38
81 88
7 18
35 70
3.7 25
24
133 36
16 5
65 11
19
17
0
0
63 106
24 8
87 33
33 3.8
85 84
23 7
73 33
61 52 56
8 18 13
83 86
22 7
-
Karnofsky performance status 80-100 SO-70
Weight loss 55% >5%
Stage IIIA IIIB Treatment group I II III Interfraction interval 4.5-5 h 5.5-6 h
0
-
0
MST = median survival time; yr = year; * Group I vs. II; ’ Group I vs. III; t Group II vs. III. 4.5-5 h intervals (p = 0.0010 and 0.033, respectively). The survival rates of Groups I and III patients treated with the shorter intervals were not significantly different (p = 0.29).
Distribution of some pretreatment prognostic factors was not equal between the shorter and longer interval groups. Sex and KPS were balanced, but patients > 60 years of age, those with Stage IIIA disease, and those with previous weight loss < 5% were treated more often with the shorter intervals than those < 60 years, those with Stage IIIB disease, and those with weight loss > 5%, respectively (Table 1) . Therefore, survival data were analyzed for each of the subgroups of patients (age: < 60 vs. > 60; Stage: IIIA vs. IIIB; weight loss: < 5% vs. > 5%). As shown in Figs. 4-6, patients treated with the shorter intervals had a better prognosis than those treated with the longer intervals in all of these subgroups.
4.5-5 h
L,
5.5-6 I?‘:., c
..I
To determine whether the observed great differences were due to the effect of interfraction interval or imbalance of prognostic factors, multivariate analysis was performed. Survival data and univariate and multivariate p values for potential prognostic factors including interfraction interval are shown in Table 2. Multivariate analysis showed that interfraction interval was an independent prognosticator of treatment outcome together with sex, age, KPS, and stage. Treatment group was not significant when all three groups were analyzed together. The same effect of interfraction interval was seen in separate multivariate analyses performed for each treatment group (data not shown). Acute and late toxicities observed in this study are shown in Tables 3 and 4, respectively. Overall, Grade 3 or 4 acute toxicity was seen in 2 1 of 83 (25% ) patients treated with the shorter intervals and in 11 of 86 ( 13%)
-‘-& , 4,5-5 \L. 5.5-6 h ‘-I.,.
-\..., 30
60
Months
Fig. 1. Survival for all patients according to interfraction interval (p = 0.00000).
Fig. 2. First relapse-free survival for all patients according to interfraction interval (p = O.OOOOO).
I. J. Radiation Oncology l Biology 0 Physics
306
Group
I
4.5-5
h
Volume 34, Number 2, 1996
‘.
5.5-6
h .::-,
-.-
b
Group
5oj
” :: .h
I
5.5-6
III
4.5-5 h
I
60
Fig. >. Survival according to interfraction interval in Stages IIIA and IIIB. (a) p = 0.00002; (b) p = O.OOOOO.
h ‘:.
30
60
treated with the longer intervals, and there was no differ-rice (p = 0.68).
Months Fig. 3. Survival accordingto interfraction interval in eachtreatment group (Group I, p = 0.00023; Group II, p = O.OOOOO; Group III, p = 0.00027). patients treated with the longer intervals. This difference was significant (p = 0.038). Grade 3 or 4 late toxicity was seen in 8 of 83 (9.6%) patients treated with the shorter intervals and 10 of 86 (12%) patients
DISCUSSION There was a great difference in prognosis between patients treated with 4.5-5 h interfraction intervals and those treated with 5.5-6 h intervals, but this difference is considered to be partly due to imbalance of prognostic factors. Younger age, lower KPS, Stage IIIB, and weight loss of > 5% were associatedwith a poor prognosis and
Weight
loss
sz 5 %
Age < 60
4.5-5
$
5.5-6
h
.-.
h :: :.
L
b
b
I
60
30
>5%
4.5-5 5.5-6
h ---. . . . . .
h 60
60
M::hs
Fig. 4. Survival accordingto interfraction interval in patients < and > 60 yearsold. (a) p = O.OOOOO;(b) p = O.OOOOO.
Fig. 6. Survival accordingto interfraction interval in patients with previousweightlossof 5 5% and > 5%. (a) p = 0.00073; (b) p = 0.00001.
Effect
of inferfraction
interval
in hyperfractionated
Table 3. Incidence of Grade 3 + 4 acute toxicity 4.5-5 h intervals
5.5-6 h intervals
Treatment group
I
II
III
I
II
III
Bronchopulmonary Hematologic Esophageal Gastric Cutaneous
2 0 2 1 0
3 (1) 2 (1) 1 0 0
2(l) 5 (2) 1 0 2(l)
1 0 1 100 0
1 1 1
2 (1) 2 (1) 1
0
0
Subtotal
5129 6125
Total
10129 3132 3127 5127
21/83 (25%)
11/86 (13%)
( ) = Grade 4 toxicity. patients with these characteristics tended to have been treated with the longer intervals. However, multivariate analysis indicated that the inter-fraction interval was an independent prognostic factor. A moderate increase in the incidence of acute Grade 3 or 4 toxicity was also observed in the shorter interval group, although the incidence of late toxicity was similar to that in the longer interval group. These findings were somewhat in contrast with those reported by Cox et al. (4) for HFX RT of carcinomas of the upper respiratory and digestive tracts; interfraction intervals of < 4.5 h were associated with an increase in late toxicity as compared with intervals of > 4.5 h, but tumor control, survival, and acute toxicity were similar between the two groups. Marcia1 et al. ( 11) also found a trend towards more late tissue reactions when the interfraction interval was 4.5 h or less in HI% RT for squamous cell carcinoma of the oral cavity, pharynx, larynx, and sinuses. The major differences between these two studies and ours are the treatment site and range of inter-fraction intervals (4.5-5 h/5.5-6 h vs. < 4.5 h/> 4.5 h), which might account for the discrepancy. Some experimental data suggested a faster repair in acutely responding tissues than that in late-responding tissues (5, 18), but Rojas and Joiner (14) later summarized that the repair half times for most tissues were between 1 and 2 h, irrespective of their proliferation kinetics. Because of the relatively small number of patients and concurrent use of chemotherapy, our data appear to be insufficient to affirm the difference in the incidence of acute and late toxicity due to the shorter interfraction intervals. With regard to the influence of interfraction interval on treatment outcome, there are two reports analyzing patients with malignant glioma. Nelson et al. ( 13) compared interfraction intervals of 4-4.4 h and 4.5-8 h in HFX RT using 1.2 Gy b.i.d. fractions and found that survival was better for patients treated with 4-4.4 h intervals than for those treated with 4.5-8 h intervals. The improved survival for the shorter interval was evident in each treatment arm (64.8 Gy, 72.0 Gy, 76.8 Gy, and 81.6 Gy, respectively). However, no information was avail-
able to assess differences between 4.5-6 h and > 6 h
radiotherapy
0
B. JEREMIC
Y.
AND
307
SHIEAMOTO
(or < 4.5 h) regarding either toxicity or survival rates. Jeremic et al. (8) reported on a similar study that used 1.2 Gy b.i.d. fractions to a total dose of 72 Gy followed by multiagent chemotherapy. Inter-fraction intervals were 4.5-5 h vs. 5.5-6 h. Again, an improved survival was observed in patients treated with the shorter intervals, which was supported by multivariate analyses. These two studies as well as ours were not randomized and these findings should be confirmed by prospective randomized trials in the near future. Nevertheless, if it is assumed that shorter intervals (4-5 h) could produce a beneficial effect, what could be the reasons for this? It is controversial whether tumor cells complete repair earlier than lateresponding normal tissue cells (6, 14), and it is possible that 4-5 h intervals are still insufficient for malignant glioma and NSCLC cells to undergo full repair. In the present study, the best survival was observed in Group II patients treated with the shorter intervals. These patients received combination chemotherapy every week (from Day 1 to 3) between the two daily radiation fractions. Such an intensive course of treatment given within 4.5-5 h periods might produce a better outcome by inhibiting tumor cell repair.
The cell cycle effect might possibly be another reason for the improved survival with the shorter interfraction intervals. In an experimental HFX RT for rat rhabdomyosarcoma reported by Kleineidam et al. ( lo), interfraction intervals of 5-6 h were more efficient than a 2-h interval. They speculated that this was in part owing to the cell cycle redistribution and reoxygenation. Reoxygenation may not be responsible for the results of the present study, but redistribution of tumor cells into a more radiosensitive cell cycle phase could be a reason, although it would be very difficult to prove it. In conclusion, our analysis suggested that shorter interfraction intervals (4.5-5 h) were associated with a better treatment outcome in patients with Stage III NSCLC treated with HFX RT with and without concurrent chemotherapy. In designing a trial of MFD RT, it seems important to fix interfraction interval. Also, prospective randomized studies are warranted to further explore the role of interfraction interval in MFD RT.
Table 4. Incidence of Grade 3 + 4 late toxicity 4.5-5 h intervals
5.5-6 h intervals
Treatment group
I
II
III
I
II
Bronchopulmonary Esophageal Osseous Cardiac
0 0 1 00
1 2 (1) 1
2(l) 1 (1) 0 0
0 0 1 0
2(l) 1 1 0
2(l) 1 (1) 1 1 (1)
1129 4125
3129
II32
4f27
5127
Subtotal Total
( ) = Grade 4 toxicity.
8183 (9.6%)
III
10/86 (12%)
308
I. J. RadiationOncology 0 Biology 0 Physics
Volume 34, Number 2, 1996
REFERENCES 1. Ball. D.: Bishon, J.: Smith, J.: Crennan. E.: O’Brien. P.; Davis, S.; Ryan,‘G.; Joseph, D.; Walker, Q. A phase III study of accelerated radiotherapy with and without carboplatin in nonsmall cell lung cancer: An interim toxicity analysis of the first 100 patients. Int. J. Radiat. Oncol. Biol. Phys. 31:267-272; 1995. 2. Byhardt, R. W.; Pajak, T. F.; Emami, B.; Herskovic, A.; Doggett, R. S.; Olsen, L. A. A phase I/II study to evaluate accelerated fractionation via concomitant boost for squamous, adeno, and large cell carcinoma of the lung: Report of Radiation Therapy Oncology Group 84-07. Int. J. Radiat. Oncol. Biol. Phys. 26:459-468; 1993. 3. Cox, J. D.; Azamia, N.; Byhardt, R. W.; Shin, K. H.; Emami, B.; Pajak, T. F. A randomizedphaseI/II trial of hyperfractionatedradiationtherapywith total dosesof 60.0 Gy to 79.2 Gy: Possiblesurvival benefit with > 69.6 Gy in favorablepatientswith RadiationTherapy Oncology Group StageIII nonsmallcell lungcarcinoma:Reportof Radiation Therapy Oncology Group 83-11. J. Clin. Oncol. 8: 15431555; 1990. 4. Cox, J. D.; Pajak,T. F.; Martial, V. A.; Coia, L.; Mohiuddin, M.; Fu, K. K.; Selim, H.; Rubin, P.; Ortiz, H. andthe RadiationTherapy Oncology Group. Interfraction interval is a majordeterminantof late effectswith hyperfractionated radiation therapy of carcinomasof upper respiratory and digestivetracts: Resultsfrom RadiationTherapy Oncology Group protocol 83-13. Int. J. Radiat. Oncol. Biol. Phys. 20:1191-1195; 1991. 5. Fowler, J. F. Intervals betweenmultiple fractions per day. Differencesbetweenearly andlateradiationreactions.Acta Oncol. 27:181-183; 1988. 6. Fowler, J. F. Are half-times of repair reliably shorterfor tumors than for late normal-tissueeffects?Int. J. Radiat. Oncol. Biol. Phys. 31:189-190; 1995. Hopewell, J. W.; van den Aardweg, G. J. M. J. Current conceptsof dose-fractionationin radiotherapy:Normal tissuetolerance.Br. J. Radiol. (Suppl. 22): 88-94; 1988. Jeremic,B.; Grujicic, D.; Antunovic, V.; Djuric, L.; Shibamoto, Y. Hyperfractionatedradiation therapy (HFX RT) followed by multiagent chemotherapy(CHT) in patients with malignantglioma. A phaseII study. Int. J. Radiat. Oncol. Biol. Phys. 30:1179- 1185; 1994. Jeremic,B.; Shibamoto,Y.; Acimovic, L.; Djuric, L. Randomized trial of hyperfractionatedradiation therapy with
or without concurrentchemotherapyfor StageIII nonsmallcell lung cancer.J. Clin. Oncol. 13:452-458; 1995. 10. Kleineidam,M.; Pieconka,A.; Beck-Bomholdt, H. P. Radiotherapyof the rhabdomyosarcoma RlH of the rat: Influenceof thetime interval betweentwo daily fractionsduring hyperfractionatedradiotherapy.Radiother.Oncol. 30:128132; 1994. 11. Martial, V. A.; Pajak, T. F.; Chang, C.; Tupchong, L.; Stetz, J. Hyperfractionatedphotonradiation therapy in the treatmentof advancedsquamous cell carcinomaof the oral cavity, pharynx, larynx, andsinuses, usingradiationtherapy asthe only plannedmodality: (Preliminary report) by the RadiationTherapy Oncology Group (RTOG) _ Int. I. Radiat. Oncol. Biol. Phys. 13: 41-47; 1987. 12. Morgan, D. A. L.; Bradley, P. J.; MacLennan,K. A. Radiotherapy of advancedlaryngeal cancerusingthree fractions per day. Br. J. Radiol. 60:607, 1987. 13. Nelson, D. F.; Curran, W. J.; Scott, C.; Nelson, J. S.; Weinstein, A. S.; Ahmad, K.; Constine,L.; Murray, K.; Powlis, W. D.; Mohiuddin, M.; Fischbach,J. Hyperfractionatedradiationtherapy and bis-chlorethylnitrosoureain thetreatmentof malignantglioma-Possible advantageobserved at 72.0 Gy in 1.2 b.i.d. fractions: Report of the RadiationTherapy Oncology Group protocol 8302. Int. J. Radiat.Oncol. Biol. 25:193-207; 1993. 14. Rojas,A.; Joiner,M. C. The influenceof doseper fraction on repair kinetics. Radiother.Oncol. 14329-336; 1989. 15. Saunders,M. I.; Dische,S. Continuous,hyperfractionated, acceleratedradiotherapy(CHART) in nonsmallcell carcinomaof the bronchus.Int. J. Radiat. Oncol. Biol. Phys. 19:1211-1215; 1990. 16. Seagren,S. L.; Hemdon,J. E.; Baeker, J. R.; Boles, M.; Chung,C.; Green,M. R. Alternating irradiationandchemotherapy in Stage III A and B nonsmallcell lung cancer: Reportof a CancerandLeukemiaGroup B phaseII study. Int. J. Radiat. Oncol. Biol. Phys. 29:1085-1089; 1994. 17. Shaw,E. G.; Deming,R. L.; Creagan,E. T.; Nair, S.; SU, J. Q.; Levitt, R.; Steen,P. D.; Wiesenfeld,M.; Mailliard, J. A. Pilot study of humanrecombinantinterferon gamma and acceleratedhyperfractionatedthoracic radiation therapy in patientswith unresectableStage IIIAlB nonsmall cell lung cancer.Int. J. Radiat.Oncol. Biol. Phys. 31:827831; 1995. 18. Thames,H. D.; Withers, H. R.; Peters,L. J. Tissuerepair capacity andrepairkineticsdeducedfrom multifractionated or continuousregimenswith incompleterepair.Br. J. Cancer 49 (Suppl. VI):263-269; 1984.
Oncology
Biol.
Phys.. Vol. 34, No. 2, pp. 303-308, 1996 Copyright 10 1996 Elsevier Science Inc. Prinrcd in Ihe USA. All rights reserved
0360-30
I h/96
$ IS .oO + .(X1
ELSEVIER
0360-3016( 95)02158-2
l
Clinical Original Contribution EFFECT OF INTERFRACTION INTERVAL IN HYPERFRACTIONATED RADIOTHERAPY WITH OR WITHOUT CONCURRENT CHEMOTHERAPY FOR STAGE III NONSMALL CELL LUNG CANCER BRANISLAV
JEREMIC,
M.D.,
PH.D. * AND YUTA
SHIBAMOTO,
M.D.,
D.M.Sc.“f
*Department of Oncology, University Hospital, Kragujevac, Yugoslavia and TDepartment of Oncology, Chest Disease Research Institute, Kyoto University, Kyoto 606-01, Japan Purpose: To explore the influence of interfraction interval in hyperfractionated radiotherapy (HFX RT) withwithout concurrent chemotherapy for Stage III nonsmall cell lung cancer. Methods and Materials: One hundred sixty-nine patients treated in a randomized study were retrospectively analyzed. Group I patients were treated by HFX RT with 1.2 Gy twice daily with a total dose of 64.8 Gy in 27 treatment days, while Groups II and III patients were treated by the same HFX RT and concurrent chemotherapy with carboplatin and etoposide (every week in Group II and every other week in Group III). Interfraction intervals of either 4.5-5 h or 5.5-6 h were used for each patient. Results: Patients treated with shorter interfraction intervals (4.5-5 h) had a better prognosis than those treated with longer intervals (5.5-6 h) (median survival: 22 vs. 7 months; 5-year survival rate: 27% vs. 0%, p = 0.00000). This phenomenon was observed in all treatment groups. Patients 2 60 years of age, with Stage IIIA disease, or with previous weight loss 5 5% were treated more often with the shorter intervals than those < 60 years, with Stage IIIB disease, or with weight loss > 5%, respectively, but in all of these subgroups of patients, the shorter intervals were associated with a better prognosis. Multivariate analysis showed that the interfraction interval was an independent prognostic factor, together with sex, age, performance status, and stage. The shorter intervals were associated with an increased incidence of acute high grade toxicity, but not with an increase in late toxicity. Conclusion: Patients treated with shorter interfraction intervals (4.5-5 h) appeared to have a better survival than those treated with longer intervals (5.5-6 h). Prospective randomized studies are warranted to further investigate the influence of interfraction interval in HFX RT. Radiotherapy,
Hyperfractionation,
Interfraction
interval, Nonsmall cell lung cancer. Locally advanced nonsmall cell lung cancer (NSCLC) has been frequently treated with MFD RT (1-3, 15-17). All reports analyzed survival and toxicity but no reports considered the role of interfraction interval in treatment outcome. We carried out a randomized study comparing hyperfractionated (HFX) RT alone and HFX RT with concurrent chemotherapy for Stage ill NSCLC. ln this trial, various interfraction intervals in the range of 4.5-6 h were used but the interval was kept constant for each patient. Therefore, we performed a retrospective analysis of these patients to explore the influence of interfraction interval on treatment outcome.
INTRODUCTION Multiple fraction-per-day ( MFD) radiation therapy (RT) is usually given in twice daily fractions and interfraction intervals of 4-8 h have been used in most clinical trials. These intervals have been chosen based on the repair kinetics of normal tissues (but not tumor). Biological data suggest that 3-4 h intervals are not sufficient to allow for full repair of radiation injury in many systems (5, 7, 18) and this is already borne out by several clinical studies in which intervals shorter than 4.5 h produced more morbidity than did the longer intervals ( 11, 12, 15). Based on these findings on normal tissues, interfraction intervals of 6 h or longer are generally recommended currently ( 14). However, it is not known what range of interfraction interval is optimal in terms of tumor response.
METHODS
AND MATERIALS
One hundred sixty-nine patients with Stage ill NSCLC who entered a randomized trial comparing HFX RT alone
Reprint requests to: Yuta Shibamoto, M.D., Department of Oncology, Chest Disease Research Institute, Kyoto University, Kyoto 606-01, Japan.
Accepted for publication 19 September 1995.
303
304
1. J. Radiation Oncology 0 Biology 0 Physics Table 1. Distribution
of pretreatment
Volume 34, Number 2, 1996 prognostic
Group I 4.5-5
h
factors by interfraction
Group II
5.5-6
h
4.5-5
h
interval
Group III
5.5-6
h
4.5-5
h
Total
5.5-6
h
4.5-5
h
5.5-6
h
Sex M:F
22~7
24:8
15:14
14:18
6:19
5:24
9:23
3122
Stage 1IIA:IIIB
19:lO
13:19
Weight loss s5%:>5%
16:13
7:25
Age
<60:260
Karnofsky performance ~70:r80
17:8
2217
21:6
61:22 p = 0.39
68:18
18:9
12:17
16:ll
33:50 p = 0.037
48:38
status 7:20
5:24
7:20
13:70 p = 0.079
23~63
14:ll
11:16
16:13
12:15
49134 p = 0.026
36:50
15:lO
4~23
15:14
6:21
and HFX RT plus chemotherapy at University Hospital, Kragujevac in 1988 and 1989 were’ analyzed. The details of this study were described previously (9). Briefly, adult patients with Stage III NSCLC and a Karnofsky performance status (KPS) score of at least 50% were treated with: HFX RT alone to a total dose of 64.8 Gy with 1.2 Gy b.i.d. fractions in 27 treatment days (Group I, 62 patients) ; with the same HFX RT and concurrent chemotherapy consisting of 100 mg of carboplatin on Days 1 and 2, and 100 mg of etoposide on Days l-3 of each week during the RT course (Group II, 52 patients) ; and with the same HFX RT and concurrent chemotherapy consisting of 200 mg of carboplatin on Days 1 and 2, and 100 mg of etoposide on Days l-5 of the first, third, and fifth weeks during the RT course (Group III, 56 patients). The drugs were administered between the two daily fractions: 3-4 h after the first one and l-2 h before the second one. The patient characteristics are shown in Table 1. The RT method was also described previously (9). Briefly, the initial target volume including tumor plus a 2-cm margin and a major part of mediastinum was treated to 50.4 Gy, after which the RT field was reduced to cover detectable tumors. Interfraction intervals of 4.5-5 h or 5.5-6 h were nonrandomly assigned to each patient and under no circumstances was it allowed to change from the shorter interval to the longer interval and vice versa. The avearge interfraction interval was 4.5 h in 40 patients, 5 h in 43, 5.5 h in 40, and 6 h in 46. Radiation-induced effects on normal tissue were assessed as either acute or late phenomena, according to the criteria of the Radiation Therapy Oncology Group ( 3). Differences in the patient characteristics and incidence of toxicity between groups were evaluated by the &i-square test. Survival rates were calculated from the date of randomization by
’ HALBAU,
23:4
Gendaisuugakusha,
Kyoto, Japan.
46:37 17:69 p = o.OOOOO2
the Kaplan-Meier method and differences between survival curves were analyzed by the generalized Wilcoxon test. The interaction of each prognostic factor and their effect on survival were analyzed using the Cox proportional hazards model. All statistical analyses were carried out using a computer program.’
RESULTS All 169 patients were evaluated for this analysis. Survival data for the three treatment groups were reported previously (9). Briefly, Group II patients had a better prognosis than Group I patients, but there were no differences between Groups 11 and III and Groups I and III (Table 2). As a whole, patients treated with the shorter intervals had a better prognosis than those treated with the longer intervals (p = 0.00000) (Fig. 1). They also had a longer median time to first (local or distant) relapse (MTFR) and a better first relapse-free survival (FRFS ) (MTFR: 20 vs. 5 months; l-, 3-, and 5-year FRFS: 64%, 28%, and 26% vs. 31%, O%, and 0%; p = 0.00000) (Fig. 2). Because differences between survival and FRFS were constant throughout the study, all further analyses were performed with regard to survival. The difference in survival due to interfraction interval was also seen in all treatment groups (Fig. 3) and it was the most striking in Group II patients receiving concurrent weekly chemotherapy. The survival rates of the patients treated with 5.5-6 h intervals were similar among the three groups (I vs. II, p = 0.083; I vs. III, p = 0.78; and II vs. III, p = 0.33), whereas the survival of Group II patients treated with 4.5-5 h intervals was significantly better than that of Group I or III patients treated with
305
Effect of interfraction interval in hyperfractionated radiotherapy l B. JEREMIC AND Y. SHIBAMOTO
Table 2. Survival data and statistical analysis for various potential prognostic factors
Variable
n
Sex Male Female Age <60 260
Survival fate (%)
MST (months)
1 yr
3 yr
5 Yr
Univariate p
Multivariate p
0.00012
O.ooooOl
0.00000
0.000081
O.OOOOO
O.OOOOO1
31 2.8
o.ooooo
0.080
27 2.4
24
O.OOOOO
o.OoOo59
39 73 50
6.6 23 16
4.9 21 16
0.0028* 0.33' 0.061'
0.33
71 36
30
27
O.OOOOO
O.OOOOOO
129
10
40
25
46 78
7.7 38
38
81 88
7 18
35 70
3.7 25
24
133 36
16 5
65 11
19
17
0
0
63 106
24 8
87 33
33 3.8
85 84
23 7
73 33
61 52 56
8 18 13
83 86
22 7
-
Karnofsky performance status 80-100 SO-70
Weight loss 55% >5%
Stage IIIA IIIB Treatment group I II III Interfraction interval 4.5-5 h 5.5-6 h
0
-
0
MST = median survival time; yr = year; * Group I vs. II; ’ Group I vs. III; t Group II vs. III. 4.5-5 h intervals (p = 0.0010 and 0.033, respectively). The survival rates of Groups I and III patients treated with the shorter intervals were not significantly different (p = 0.29).
Distribution of some pretreatment prognostic factors was not equal between the shorter and longer interval groups. Sex and KPS were balanced, but patients > 60 years of age, those with Stage IIIA disease, and those with previous weight loss < 5% were treated more often with the shorter intervals than those < 60 years, those with Stage IIIB disease, and those with weight loss > 5%, respectively (Table 1) . Therefore, survival data were analyzed for each of the subgroups of patients (age: < 60 vs. > 60; Stage: IIIA vs. IIIB; weight loss: < 5% vs. > 5%). As shown in Figs. 4-6, patients treated with the shorter intervals had a better prognosis than those treated with the longer intervals in all of these subgroups.
4.5-5 h
L,
5.5-6 I?‘:., c
..I
To determine whether the observed great differences were due to the effect of interfraction interval or imbalance of prognostic factors, multivariate analysis was performed. Survival data and univariate and multivariate p values for potential prognostic factors including interfraction interval are shown in Table 2. Multivariate analysis showed that interfraction interval was an independent prognosticator of treatment outcome together with sex, age, KPS, and stage. Treatment group was not significant when all three groups were analyzed together. The same effect of interfraction interval was seen in separate multivariate analyses performed for each treatment group (data not shown). Acute and late toxicities observed in this study are shown in Tables 3 and 4, respectively. Overall, Grade 3 or 4 acute toxicity was seen in 2 1 of 83 (25% ) patients treated with the shorter intervals and in 11 of 86 ( 13%)
-‘-& , 4,5-5 \L. 5.5-6 h ‘-I.,.
-\..., 30
60
Months
Fig. 1. Survival for all patients according to interfraction interval (p = 0.00000).
Fig. 2. First relapse-free survival for all patients according to interfraction interval (p = O.OOOOO).
I. J. Radiation Oncology l Biology 0 Physics
306
Group
I
4.5-5
h
Volume 34, Number 2, 1996
‘.
5.5-6
h .::-,
-.-
b
Group
5oj
” :: .h
I
5.5-6
III
4.5-5 h
I
60
Fig. >. Survival according to interfraction interval in Stages IIIA and IIIB. (a) p = 0.00002; (b) p = O.OOOOO.
h ‘:.
30
60
treated with the longer intervals, and there was no differ-rice (p = 0.68).
Months Fig. 3. Survival accordingto interfraction interval in eachtreatment group (Group I, p = 0.00023; Group II, p = O.OOOOO; Group III, p = 0.00027). patients treated with the longer intervals. This difference was significant (p = 0.038). Grade 3 or 4 late toxicity was seen in 8 of 83 (9.6%) patients treated with the shorter intervals and 10 of 86 (12%) patients
DISCUSSION There was a great difference in prognosis between patients treated with 4.5-5 h interfraction intervals and those treated with 5.5-6 h intervals, but this difference is considered to be partly due to imbalance of prognostic factors. Younger age, lower KPS, Stage IIIB, and weight loss of > 5% were associatedwith a poor prognosis and
Weight
loss
sz 5 %
Age < 60
4.5-5
$
5.5-6
h
.-.
h :: :.
L
b
b
I
60
30
>5%
4.5-5 5.5-6
h ---. . . . . .
h 60
60
M::hs
Fig. 4. Survival accordingto interfraction interval in patients < and > 60 yearsold. (a) p = O.OOOOO;(b) p = O.OOOOO.
Fig. 6. Survival accordingto interfraction interval in patients with previousweightlossof 5 5% and > 5%. (a) p = 0.00073; (b) p = 0.00001.
Effect
of inferfraction
interval
in hyperfractionated
Table 3. Incidence of Grade 3 + 4 acute toxicity 4.5-5 h intervals
5.5-6 h intervals
Treatment group
I
II
III
I
II
III
Bronchopulmonary Hematologic Esophageal Gastric Cutaneous
2 0 2 1 0
3 (1) 2 (1) 1 0 0
2(l) 5 (2) 1 0 2(l)
1 0 1 100 0
1 1 1
2 (1) 2 (1) 1
0
0
Subtotal
5129 6125
Total
10129 3132 3127 5127
21/83 (25%)
11/86 (13%)
( ) = Grade 4 toxicity. patients with these characteristics tended to have been treated with the longer intervals. However, multivariate analysis indicated that the inter-fraction interval was an independent prognostic factor. A moderate increase in the incidence of acute Grade 3 or 4 toxicity was also observed in the shorter interval group, although the incidence of late toxicity was similar to that in the longer interval group. These findings were somewhat in contrast with those reported by Cox et al. (4) for HFX RT of carcinomas of the upper respiratory and digestive tracts; interfraction intervals of < 4.5 h were associated with an increase in late toxicity as compared with intervals of > 4.5 h, but tumor control, survival, and acute toxicity were similar between the two groups. Marcia1 et al. ( 11) also found a trend towards more late tissue reactions when the interfraction interval was 4.5 h or less in HI% RT for squamous cell carcinoma of the oral cavity, pharynx, larynx, and sinuses. The major differences between these two studies and ours are the treatment site and range of inter-fraction intervals (4.5-5 h/5.5-6 h vs. < 4.5 h/> 4.5 h), which might account for the discrepancy. Some experimental data suggested a faster repair in acutely responding tissues than that in late-responding tissues (5, 18), but Rojas and Joiner (14) later summarized that the repair half times for most tissues were between 1 and 2 h, irrespective of their proliferation kinetics. Because of the relatively small number of patients and concurrent use of chemotherapy, our data appear to be insufficient to affirm the difference in the incidence of acute and late toxicity due to the shorter interfraction intervals. With regard to the influence of interfraction interval on treatment outcome, there are two reports analyzing patients with malignant glioma. Nelson et al. ( 13) compared interfraction intervals of 4-4.4 h and 4.5-8 h in HFX RT using 1.2 Gy b.i.d. fractions and found that survival was better for patients treated with 4-4.4 h intervals than for those treated with 4.5-8 h intervals. The improved survival for the shorter interval was evident in each treatment arm (64.8 Gy, 72.0 Gy, 76.8 Gy, and 81.6 Gy, respectively). However, no information was avail-
able to assess differences between 4.5-6 h and > 6 h
radiotherapy
0
B. JEREMIC
Y.
AND
307
SHIEAMOTO
(or < 4.5 h) regarding either toxicity or survival rates. Jeremic et al. (8) reported on a similar study that used 1.2 Gy b.i.d. fractions to a total dose of 72 Gy followed by multiagent chemotherapy. Inter-fraction intervals were 4.5-5 h vs. 5.5-6 h. Again, an improved survival was observed in patients treated with the shorter intervals, which was supported by multivariate analyses. These two studies as well as ours were not randomized and these findings should be confirmed by prospective randomized trials in the near future. Nevertheless, if it is assumed that shorter intervals (4-5 h) could produce a beneficial effect, what could be the reasons for this? It is controversial whether tumor cells complete repair earlier than lateresponding normal tissue cells (6, 14), and it is possible that 4-5 h intervals are still insufficient for malignant glioma and NSCLC cells to undergo full repair. In the present study, the best survival was observed in Group II patients treated with the shorter intervals. These patients received combination chemotherapy every week (from Day 1 to 3) between the two daily radiation fractions. Such an intensive course of treatment given within 4.5-5 h periods might produce a better outcome by inhibiting tumor cell repair.
The cell cycle effect might possibly be another reason for the improved survival with the shorter interfraction intervals. In an experimental HFX RT for rat rhabdomyosarcoma reported by Kleineidam et al. ( lo), interfraction intervals of 5-6 h were more efficient than a 2-h interval. They speculated that this was in part owing to the cell cycle redistribution and reoxygenation. Reoxygenation may not be responsible for the results of the present study, but redistribution of tumor cells into a more radiosensitive cell cycle phase could be a reason, although it would be very difficult to prove it. In conclusion, our analysis suggested that shorter interfraction intervals (4.5-5 h) were associated with a better treatment outcome in patients with Stage III NSCLC treated with HFX RT with and without concurrent chemotherapy. In designing a trial of MFD RT, it seems important to fix interfraction interval. Also, prospective randomized studies are warranted to further explore the role of interfraction interval in MFD RT.
Table 4. Incidence of Grade 3 + 4 late toxicity 4.5-5 h intervals
5.5-6 h intervals
Treatment group
I
II
III
I
II
Bronchopulmonary Esophageal Osseous Cardiac
0 0 1 00
1 2 (1) 1
2(l) 1 (1) 0 0
0 0 1 0
2(l) 1 1 0
2(l) 1 (1) 1 1 (1)
1129 4125
3129
II32
4f27
5127
Subtotal Total
( ) = Grade 4 toxicity.
8183 (9.6%)
III
10/86 (12%)
308
I. J. RadiationOncology 0 Biology 0 Physics
Volume 34, Number 2, 1996
REFERENCES 1. Ball. D.: Bishon, J.: Smith, J.: Crennan. E.: O’Brien. P.; Davis, S.; Ryan,‘G.; Joseph, D.; Walker, Q. A phase III study of accelerated radiotherapy with and without carboplatin in nonsmall cell lung cancer: An interim toxicity analysis of the first 100 patients. Int. J. Radiat. Oncol. Biol. Phys. 31:267-272; 1995. 2. Byhardt, R. W.; Pajak, T. F.; Emami, B.; Herskovic, A.; Doggett, R. S.; Olsen, L. A. A phase I/II study to evaluate accelerated fractionation via concomitant boost for squamous, adeno, and large cell carcinoma of the lung: Report of Radiation Therapy Oncology Group 84-07. Int. J. Radiat. Oncol. Biol. Phys. 26:459-468; 1993. 3. Cox, J. D.; Azamia, N.; Byhardt, R. W.; Shin, K. H.; Emami, B.; Pajak, T. F. A randomizedphaseI/II trial of hyperfractionatedradiationtherapywith total dosesof 60.0 Gy to 79.2 Gy: Possiblesurvival benefit with > 69.6 Gy in favorablepatientswith RadiationTherapy Oncology Group StageIII nonsmallcell lungcarcinoma:Reportof Radiation Therapy Oncology Group 83-11. J. Clin. Oncol. 8: 15431555; 1990. 4. Cox, J. D.; Pajak,T. F.; Martial, V. A.; Coia, L.; Mohiuddin, M.; Fu, K. K.; Selim, H.; Rubin, P.; Ortiz, H. andthe RadiationTherapy Oncology Group. Interfraction interval is a majordeterminantof late effectswith hyperfractionated radiation therapy of carcinomasof upper respiratory and digestivetracts: Resultsfrom RadiationTherapy Oncology Group protocol 83-13. Int. J. Radiat. Oncol. Biol. Phys. 20:1191-1195; 1991. 5. Fowler, J. F. Intervals betweenmultiple fractions per day. Differencesbetweenearly andlateradiationreactions.Acta Oncol. 27:181-183; 1988. 6. Fowler, J. F. Are half-times of repair reliably shorterfor tumors than for late normal-tissueeffects?Int. J. Radiat. Oncol. Biol. Phys. 31:189-190; 1995. Hopewell, J. W.; van den Aardweg, G. J. M. J. Current conceptsof dose-fractionationin radiotherapy:Normal tissuetolerance.Br. J. Radiol. (Suppl. 22): 88-94; 1988. Jeremic,B.; Grujicic, D.; Antunovic, V.; Djuric, L.; Shibamoto, Y. Hyperfractionatedradiation therapy (HFX RT) followed by multiagent chemotherapy(CHT) in patients with malignantglioma. A phaseII study. Int. J. Radiat. Oncol. Biol. Phys. 30:1179- 1185; 1994. Jeremic,B.; Shibamoto,Y.; Acimovic, L.; Djuric, L. Randomized trial of hyperfractionatedradiation therapy with
or without concurrentchemotherapyfor StageIII nonsmallcell lung cancer.J. Clin. Oncol. 13:452-458; 1995. 10. Kleineidam,M.; Pieconka,A.; Beck-Bomholdt, H. P. Radiotherapyof the rhabdomyosarcoma RlH of the rat: Influenceof thetime interval betweentwo daily fractionsduring hyperfractionatedradiotherapy.Radiother.Oncol. 30:128132; 1994. 11. Martial, V. A.; Pajak, T. F.; Chang, C.; Tupchong, L.; Stetz, J. Hyperfractionatedphotonradiation therapy in the treatmentof advancedsquamous cell carcinomaof the oral cavity, pharynx, larynx, andsinuses, usingradiationtherapy asthe only plannedmodality: (Preliminary report) by the RadiationTherapy Oncology Group (RTOG) _ Int. I. Radiat. Oncol. Biol. Phys. 13: 41-47; 1987. 12. Morgan, D. A. L.; Bradley, P. J.; MacLennan,K. A. Radiotherapy of advancedlaryngeal cancerusingthree fractions per day. Br. J. Radiol. 60:607, 1987. 13. Nelson, D. F.; Curran, W. J.; Scott, C.; Nelson, J. S.; Weinstein, A. S.; Ahmad, K.; Constine,L.; Murray, K.; Powlis, W. D.; Mohiuddin, M.; Fischbach,J. Hyperfractionatedradiationtherapy and bis-chlorethylnitrosoureain thetreatmentof malignantglioma-Possible advantageobserved at 72.0 Gy in 1.2 b.i.d. fractions: Report of the RadiationTherapy Oncology Group protocol 8302. Int. J. Radiat.Oncol. Biol. 25:193-207; 1993. 14. Rojas,A.; Joiner,M. C. The influenceof doseper fraction on repair kinetics. Radiother.Oncol. 14329-336; 1989. 15. Saunders,M. I.; Dische,S. Continuous,hyperfractionated, acceleratedradiotherapy(CHART) in nonsmallcell carcinomaof the bronchus.Int. J. Radiat. Oncol. Biol. Phys. 19:1211-1215; 1990. 16. Seagren,S. L.; Hemdon,J. E.; Baeker, J. R.; Boles, M.; Chung,C.; Green,M. R. Alternating irradiationandchemotherapy in Stage III A and B nonsmallcell lung cancer: Reportof a CancerandLeukemiaGroup B phaseII study. Int. J. Radiat. Oncol. Biol. Phys. 29:1085-1089; 1994. 17. Shaw,E. G.; Deming,R. L.; Creagan,E. T.; Nair, S.; SU, J. Q.; Levitt, R.; Steen,P. D.; Wiesenfeld,M.; Mailliard, J. A. Pilot study of humanrecombinantinterferon gamma and acceleratedhyperfractionatedthoracic radiation therapy in patientswith unresectableStage IIIAlB nonsmall cell lung cancer.Int. J. Radiat.Oncol. Biol. Phys. 31:827831; 1995. 18. Thames,H. D.; Withers, H. R.; Peters,L. J. Tissuerepair capacity andrepairkineticsdeducedfrom multifractionated or continuousregimenswith incompleterepair.Br. J. Cancer 49 (Suppl. VI):263-269; 1984.