Effect of split-course radiotherapy on survival and local control in advanced localized prostatic carcinoma

Im. J Radialron Oncology Biol Phys Vol Printed in the U.S.A. All rights reserved.

0360.3016/93 $6.00 + .oO Copyright Q 1993 Pergamon Press Ltd.

26. pp. 21 l-216

??Clinical Original Contribution

EFFECT OF SPLIT-COURSE RADIOTHERAPY ON SURVIVAL AND LOCAL CONTROL IN ADVANCED LOCALIZED PROSTATIC CARCINOMA MIKAEL KAJANTI,

Department

M.D., LARS R. HOLSTI, M.D., PAUL HOLSTI, AND KARI MUYKKYNEN, M.D.

of Radiotherapy and Oncology, Helsinki University Central Hospital, Haartmaninkatu

M.D.*

4, SF-00290 Helsinki, Finland

Purpose: to analyze the effect of overall treatment time of radiotherapy on survival and local control in locally advanced prostatic cancer in a split-course treatment setting. Methods and Materials: 168 patients with Stage C prostatic cancer treated during 1979-1989 by the split-course method where the overall treatment time is protracted. Treatment consisted of whole pelvis irradiation of 40 Gy in 4 weeks, followed by a planned 3-week interruption and an additional 26 Gy by the reduced field technique to a total dose of 66 Gy in 9 weeks and 30-33 fractions. The overall treatment time varied from 55 to 100 days. Thirtyeight percent (63) of the patients were treated primarily with radiotherapy, while the rest (105) had received androgen ablative therapy during 2 to 4.5 years before radiotherapy. To examine the effect of treatment time on local control, the patients were divided into three groups (5 63 days, 64-70 days, and > 70 days) by treatment time. Results: the 5-year actuarial survival rates, calculated from the date of diagnosis, were 91% for the hormonally slated patients and 69% for the patients treated with radiotherapy alone. The 5-year actuarial local control rates, counted from the start of radiotherapy, were 84% for radiotherapy and 80% for the hormonally manipulated group. Overall, no significant effect of treatment time could be seen, either for radiotherapy alone or for the hormonally manipulated group. The results were similar when the material was further divided by T category and histologic grade. Conclusions: no significant effect of overall treatment time (55 to 100 days) on survival or local control was found in either group. The survival time from diagnosis was longer in the hormonally pretreated group. Apparently, with adequate doses (2 65 Gy) the overall treatment time becomes less important for local control of advanced prostatic cancer, even in a split-course treatment setting. Prostate carcinoma, Split-course

radiotherapy, Overall treatment time.

INTRODUCTION

METHODS

Carcinoma of the prostate occurs most commonly in the seventh and eighth decades of life and is characterized by a long natural history. During the last 20 years, radiation therapy has assumed an important role in the definitive treatment of this malignancy. However, questions concerning optimal radiotherapy for carcinoma of the prostate like the value of continuous versus split-course treatment have remained unanswered (7, 25). Recently published results ( 1,20) have also sparked discussion about the effect of overall treatment time in radiotherapy for localized prostatic cancer (13). Survival and local control rates of locally advanced carcinoma of the prostate with pelvic radiation followed by a planned 3-week interruption, and local irradiation of the tumor carried out at the Department of Radiotherapy and Oncology, Helsinki University Central Hospital, during 1979- 1989 are presented.

AND MATERIALS

Patients The present study comprised 168 consecutive patients with locally advanced (T3-T4NXMO or Stage C) adenocarcinoma of the prostate. The ages of the patients at diagnosis ranged from 46 to 84 years, mean 69 years. The diagnosis of prostatic carcinoma was confirmed by fine needle biopsy (FNB) in 114 (68%) patients and by transurethral resection (TURP) in 54 (32%) patients. Histologic grading according to the WHO classification was available for all patients. Before the commencement of radiotherapy patients underwent complete physical and rectal examination, routine blood count, urinalysis, serum prostatic acid phosphatase and alkaline phosphatase determinations, x-ray skeletal survey, isotope bone scan, liver function tests and chest X ray. Sixty-three (38%) patients were treated with radiotherapy alone, started within 2 months

* Deceased. Reprint requests to: Mikael Kajanti, M.D.

Accepted for publication 211

I December 1992.

212

I. J. Radiation Oncology 0 Biology 0 Physics

of diagnosis, 38 (23%) patients were referred to radiotherapy after failure to orchiectomy (2.5 + 0.5 years after diagnosis), 40 (24%) after estrogen (diethylstilbestrol or estradurin; 2.7 f 0.4 years after diagnosis), and 27 (16%) patients after orchiectomy and estrogen (4.0 f 0.7 years after diagnosis). In all these patients androgen suppression therapy was started within 1 month of diagnosis. All tumors were classified at onset of radiotherapy. For further analysis, the three hormonally treated groups were pooled. The distribution of the patients by tumor classification and histopathological grade is presented in Table 1. Radiation treatment All patients were treated with megavoltage irradiation using 6, 8 or 24 MeV photons from linear accelerators. Radiotherapy was planned on a simulator, and verification films were taken through each treatment portal. All patients had split-course radiotherapy beginning with two opposed AP and PA pelvic fields ( 14 X 14 cm to 18 X 18 cm). The daily mid-plane dose was 2 or 2.2 Gy; the target absorbed dose was 40 Gy, delivered in 18 to 20 fractions, 5 fractions per week over 4 weeks. The treatment was interrupted for 3 weeks after 30 Gy. This break was compensated for by a 10% increase in the total dose. After 40 Gy, a three-field technique (two oblique fields, one posterior field) was used to irradiate the prostate. The fraction size was the same, 2 or 2.2 Gy, and the dose was 26 Gy. The field sizes were 6 X 8 cm to 10 X 12 cm. The planned total dose delivered to the prostate was 66 Gy in 30-33 fractions over 9 weeks. The dose was calculated at an isodose that encompassed the prostate gland within a I2 cm margin. There was no difference between the four groups in median total dose (66 Gy, range 65-72 Gy) given over a median overall time of 9.2 weeks (range 8.714.2 weeks). For further analysis, patients were divided into three groups according to treatment time: s 63 days (5 1 patients, 30%), 64-70 days (99 patients, 59%), and > 70 days ( 18 patients, 11%). The variation in the overall treatment time was attributable to prolonged interruptions mainly due to acute side effects. The reasons for protracting the overall treatment time beyond 70 days were as Table 1. Distribution of patients by tumor classification and histopathological grade Grade

1

2

3-4

Total

Radiotherapy alone T3NXMO T4NXMO Total

25 (44%) 1(17%) 26 (41%)

22 (39%) 2 (33%) 24 (38%)

10 (17%) 3 (50%) 13 (21%)

57 (100%) 6 (100%) 63 (100%)

Previous hormonal manipulation + radiotherapy T3NXMO T4NXMO Total

9 (12%) 2 (6%) ll(ll%)

33 (45%) 8 (25%) 41 (39%)

31 (43%) 22 (69%) 53 (50%)

73 (100%) 32 (100%) 105 (100%)

Volume 26, Number 2, 1993

follows: prolonged acute mucosal reactions (nine patients), heart attack (two patients) and sudden failure of the general condition (two patients). In addition, five patients prolonged their therapy to over 12 weeks. The reasons were as follows: urosepsis and grave hematuria in two patients (the tumor had invaded the bladder), car accident (one patient), alcoholism (one patient) and work on the farm (one patient, who requested a 6-week interval for that purpose). However, after a prolonged break, all patients completed their therapy. Follow-up All patients were followed up for at least 2 years after treatment (median 4.7 years, maximum 11 years). Patients were evaluated 1 month post-treatment, quarterly for 3 years, biannually for the next 2 years, and annually thereafter. Tumor status was assessed by periodic rectal examination, prostatic acid phosphatase determination, and when indicated, ultrasound of the prostate, and radiographic and/or radionuclide procedures. Local failure (the presence of tumor in the prostate and/or pelvis within the treated volume after radiotherapy) was determined clinically (i.e., evidence of progressive prostatic enlargement or nodularity on examination, pelvic mass, ureteral obstruction, or reappearance of obstructive symptoms). Distant metastases were defined as any nodal failure outside the treatment volume, or as hematogenous spread. Patients were considered to have died without cancer if they were examined, and found to be free of malignancy, 6 months or less before death from causes other than cancer, or if they had died of an unrelated cause found at autopsy or entered on their death certificate. Statistics All statistical analyses were performed using the BMDP88 program (8). Survival rates were calculated by the product limit method, without correction for age or intercurrent disease, from the date of diagnosis and, to evaluate the effect of split-course radiotherapy, from the date of initiation of radiotherapy until the date of the last follow-up examination or death, respectively. The MantelCox and Tarone-Ware tests were used to evaluate the difference between survival curves. A pvalue of less than 0.05 was considered statistically significant. For the life-table calculation of local control, patients who remained continuously free of local recurrence were excluded from the analysis either on the date of death from intercurrent disease or metastasis, or on the date of the latest follow-up. Patients who were placed on hormone therapy for metastatic disease were excluded on the date that hormone therapy began. In addition, actuarial local control was compared at the three cut-off-points (I 63 days, 64-70 days, and > 70 days) in the overall treatment time range. For differences in frequencies the chi-square test was used. The simultaneous effect on survival of several factors was investigated using Cox’s proportional hazards regression analysis (BMDP-88 program 2L).

213

Split-course therapy for advanced prostate cancer 0 M. KAJANTI etal

0) 0

2

4

6

10

a Time (years)

Fig. 1. Uncorrected actuarial survival rates from the diagnosis for patients treated with radiotherapy alone (N = 63) (- 0 -) and those with previous hormonal manipulation (N = 105) (- 0 -) (p < 0.01).

RESULTS

Survival The 2- and 5-year actuarial survival rates in the different therapy groups, calculated from the date of diagnosis, were as follows: 80% and 51% (radiotherapy alone), 84% and 65% (orchiectomy + radiotherapy), 97% and 77% (estrogen + radiotherapy), and 96% and 63% (orchiectomy + estrogen + radiotherapy), respectively. When the three groups of hormonally manipulated patients were pooled and compared with the radiotherapy alone group, the corresponding figures were 91% and 69% (p < 0.01) (Fig. I). The 2- and 5-year survival rates from the start of radiotherapy were as follows: 79% and 47% (radiotherapy alone), 74% and 32% (orchiectomy + radiotherapy), 79% and 44% (estrogen + radiotherapy), and 63% and 32% (orchiectomy + estrogen + radiotherapy), respectively. There was no statistically significant overall difference between the group receiving radiotherapy alone and the group with previous hormonal manipulation in uncor-

Table 2. 2- and 5-year actuarial survival and local control rates (from start of radiotherapy) Previous hormonal manipulation + radiotherapy (%)

Radiotherapy alone (%)

2-year survival 5-year survival 5-year local control

T3

T4

T3

T4

79

83

80

56

49

33

38

30

87

67

77

47

Note: The hormonally manipulated anaplastic (Grade 3-4) tumors.

group contained

more

rected actuarial survival for the T3NXMO tumors from the start of radiotherapy (p = 0.287). The same was true for the T4NXMO tumors (p = 0.6 19) (Table 2). According to the proportional hazards regression analysis, the most important factors favoring survival were: low histologic grade, absence of anemia at onset of radiotherapy, and normal alkaline phosphatase levels. Local control There was no statistically significant difference between the group receiving radiotherapy alone and the hormonally manipulated group in overall actuarial local control calculated from the start of radiotherapy for the T3NXMO tumors (Q = 0.657). The same holds for the T4NXMO tumors (p = 0.248) (Table 2). The influence of overall treatment time on local control The three groups divided by treatment time matched for age, Kamofsky index, tumor category and histologic grade, method of diagnosis, various laboratory parameters, duration and type of previous hormonal treatment, and total dose of radiation. Overall, no significant effect of treatment time on local control could be seen (p = 0.795) either for the radiotherapy alone group (p = 0.478) or the hormonally manipulated group (p = 0.577). The results

Table 3. Analysis of 5-year local control rates (%) by T category and tumor grade (chi-square test) Overall treatment time (days)

T category T3NXMO T4NXMO Grade Gl G2 G3-4

5 63

64-70

> 70

p-value

85 83

88 47

91 64

0.436 0.146

82 84 86

87 85 67

91 92 81

0.823 0.771 0.279

214

I. J. Radiation Oncology 0 Biology 0 Physics

were similar when the material was further divided by T category and histologic grade (Table 3).

Failures During follow-up, failures were observed in 83 (49%) patients: recurrence in the primary site in 11 (7%), recurrence in the primary site and distant metastasis in 22 (13%), and distant metastasis in 50 patients (30%). According to the methodology of competing risks (9), the probability of local failure (in competition with metastasis and death from intercurrent diseases) was 13.5%. If metastasis and death from other causes could be eliminated, the lifetime probability of local failure would rise to about 33.5%. Thirty percent of all local failures were detected within 1 year, 48% within 2 years, 75% within 5 years and 100% within 7 years. Palliative radiotherapy was given to 33 patients and hormone therapy to 30 patients. Fortyone patients died of intercurrent diseases, the majority of cardiovascular disease, 11 of local recurrence, 47 of metastatic disease, 22 of local and metastatic disease. At the end of the study, 44 patients were alive with no disease, whereas three were alive with metastatic disease.

Complications The incidence of major complications was 4% (seven patients): persistent urinary incontinence in one patient, proctitis requiring colostomy in three, intestinal perforation in one, and small bowel obstruction in two patients. Minor complications developed in 13 patients (8%): hematuria (chronic cystitis) in four, rectal stricture (no colostomy) in two, proctitis not requiring colostomy in three, persistent diarrhea in two, and scrotal/penile edema in two patients. No patients had laparotomy before radiotherapy. The incidence of late complications was independent of overall treatment time, and the daily and total dose. No reliable data on sexual activity before or after the completion of radiotherapy was available. DISCUSSION

This study attempts to address the issue of prolonged overall treatment time and its potential effect on local control of prostate cancer. Sixty-three patients were treated primarily with radiotherapy, whereas 105 patients received radiotherapy after failure to hormonal manipulations. The average time between initial diagnosis and irradiation in these 105 patients was about 3 years. The actuarial survival rates, recurrence-free time and local control rates, as counted from the start of radiotherapy, were similar irrespective of previous therapy. The disease-free survival does flatten with time, and the majority of failures become manifest within 7 years. With the exception of the Stanford (6) and some other series (19, 28), most reports give 5year (1, 2, 11) or shorter (25) survival rates. The patients with previous hormonal manipulation lived significantly longer than the patients treated with radiotherapy alone. This is in agreement with the VA-

Volume 26, Number 2, 1993

CURG studies showing that patients with higher grade tumors may benefit from hormone therapy beginning at diagnosis (5). Green et al. ( 14) received encouraging results using estrogen prior and concurrently with irradiation in bulky prostate carcinoma. The early use of hormonal therapy before resistant cells become prevalent could also theoretically be justified, particularly if androgen sensitivity could be predicted by simple tests (10). Local control is expected to be a more sensitive indicator of radiobiological differences than the survival rate ( 13). According to RTOG, the overall treatment time had no impact on overall survival, NED survival, or local/ regional control when the overall treatment time was < 49 days, 50 to 63 days, or > 64 days (20). We did not find any significant effect of overall treatment time either on survival or on local control even when the split-course therapy was protracted beyond 70 days, which is in accordance with previous results ( 19). Several reports (Table 4) and our own results show that local control can be achieved in the majority of patients with locally advanced prostatic carcinoma, although the overall treatment time varies between 6 and 10 weeks. There is a distinct advantage in extending therapy to 8 weeks, as the isoeffective dose required in 6 weeks would be expected to be associated with a relative increase in local complications. It is indeed the practice in the USA to protract treatment time, and this policy can now be retrospectively justified (15). The use of split-course radiotherapy in prostatic cancer has resulted in fewer (25) or more (24) complications than continuous therapy. Lai et al. (19) reported no influence of the duration of radiotherapy on the incidence of moderate or severe complications.

Table 4. 5-year actuarial local control rates in clinical Stage C adenocarcinoma of the prostate treated with external heam radiation therapy

Number of patients 551 26 30 10 13 60 117 110 64 47 68 27 82 130 38

(Cl) (Cl) (C2) (C2)

(Cl) (C2) (Cl) (C2)

5-year local control (%) 88 73 50 80 54 69 2 77 + 77 + 77 f 69 87 63 88 82 60

6 5 5 6

Overall treatment time (wk) 6-7 <8 >8 <8 >8 <8 9 10 > 10 9 6-7 6-7 6-7 9 9

Total dose (GY) 60-70 65 65* 65 64* 70’ 70+ 707 70+ 6St 60-70 60-70 65-70 66* 66*

Ref. 28 1

1 19 19 19 19 25 16 16 2 Present study

* Includes an interruption of 2 weeks. + Includes unplanned treatment related interruptions. * Includes an interruption of 3 weeks. In studies 2, 16 and 25, follow-up time was < 5 years.

Split-course therapy for advanced prostate cancer 0 M. KAJANTI etal.

The lack of influence of duration of radiation therapy on treatment outcome is not synonymous with lack of dose dependence on local control (19). Hanks et al. (16) suggest that optimal local control of adenocarcinoma of the prostate is obtained with 65-70 Gy for T3 tumors, and that 70 Gy is required only for T4 tumors. The selection of a central prostate dose from iso-effect plots for time-dose optimization in T3 tumors resulted in approximately a 10% increase in total dose when the overall treatment time was extended from 6 to 8 weeks (15). Within the dose range (65 to 72 Gy) used in this analysis, there was no dependence of local control on the treatment time. Zagars et al. (28) adopted 66 Gy in 33 fractions over 6.5 weeks as the standard dose for most patients with Stage C disease. Different human tumors have different proliferation rates (12) and therefore the risk of proliferation during prolonged treatment probably also varies (17). The Sphase fractions of prostatic cancer have been proven to be extremely low (1 .O%to 10.7%; average 3.5%) (23). From the pretreatment labeling indices (22,26) alone, one might expect prostate cancers to proliferate 2 or 3 times more slowly than head and neck tumors (13). This can also be seen clinically, as most head and neck cancers, if they fail, they fail during the first 2-3 years after treatment (21), whereas the local failures in prostate cancer can be detected much later (1). Withers et al. (27) suggested in oropharyngeal cancer that a dose increment of about 60 cGy per day is required to compensate for accelerated repopulation during 4 * 1 weeks after initiation of therapy, whereas the regression line for pelvic failure in Stage C prostatic cancer varies by less than 3 cGy for each day of prolongation of treatment, suggesting that the tumor dose has been at least 63 Gy ( 19). Increasing tumor volume results in increased heterogeneity and the more aggressive tumor lines are genetically more unstable and therefore more heterogeneous on a volume-for-volume basis as compared to less aggressive

215

tumors (3). Prostatic tumors that respond to hormonal therapy contain a significant number of androgen-sensitive cells. Progression of disease after hormonal therapy results from the emergence of androgen-resistant cell clones, which may arise due to clonal selection or cell adaptation (3). On the other hand, Brawn (4) stated on the basis of serial histologic specimens in 54 patients (47 of them had received androgen ablative therapy) that the usual course of prostate carcinoma is dedifferentiation. One might speculate the androgen resistant clones to double faster and perhaps to be more resistant and susceptible to the effects of overall treatment time. In this case, more anaplastic tumors even in prostatic cancer may benefit from a shorter treatment time, as has been posited for other tumors (12). Treatment is very often prolonged because of complications, poor general condition, etc., factors which as such may influence the outcome ( 18), as was seen in this study. It is quite possible that multiple subgroup analyses can create a false impression of a time effect when indeed there really is none.

CONCLUSIONS The present retrospective material was heterogeneous in that a part of the patients had primary split-course radiotherapy, whereas another group had radiotherapy after failure to hormonal therapy, about 3 years later than the primarily irradiated group. No significant effect of overall treatment time (55 to 100 days) on survival or local control was found in either group. The survival time from diagnosis was longer in the hormonally pretreated group. Apparently, accelerated repopulation of clonogenic cells does not occur in adenocarcinoma of the prostate, and the overall treatment time is less important for local control even in a split-course treatment setting with adequate doses.

REFERENCES 1. Amdur, R. J.; Parsons, J. T.; Fitzgerald, L. T.; Million, R. R. The effect of overall treatment time on local control

prostatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 12:172 I 1727;1986.

in patients with adenocarcinoma of the prostate treated with

7. Cupps, R. E.; Utz, D. C.; Fleming, T. T.; Carson, C. C.; Zincke, H.; Myers, R. P. Definitive radiation therapy for prostatic carcinoma: Mayo Clinic experience. J. Urol. 124: 855-859;1980. 8. Dixon, W. J., ed. BMDP statistical software 1988. Berkeley, Los Angeles, London: University of California Press, 1988. 9. Dong, M. H.; Dubey, S. D.; O’Neill, R. T.; Tsong, Y. Competing riskanalysis of life table data: Application to lifetime risk computation*. J. Clin. Epidemiol. 43( 12): 135 l1359;1990. 10. Einstein, A. B., Jr. Hormonal therapy for locally advanced prostate cancer. NC1 Monogr. 7: 171-174;1988. 11. Forman, J. D.; Zinreich, E.; Lee, D. J.; Wharam, M. D.; Baumgartner, R. A.; Order, S. E. Improving the therapeutic ratio of external beam irradiation for carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 11:2073-2080;1985.

2.

3.

4. 5.

6.

radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 19: 13771382;1980. Aristizabal, S. A.; Steinbronn, D.; Heusinkveld, R. S. External beam radiotherapy in cancer of the prostate. Radiother. Oncol. 1:309-3 15; 1984. Benson, M. C.; Coffey, D. S. Prostate cancer research: current concepts and controversies. Semin. Urol. 8:323330;1983. Brawn, P. N. The dedifferentiation of prostate carcinoma. Cancer 52:246-251;1983. Byar, D. P.; Corle, D. K. Hormone therapy for prostate cancer: Results of the Veterans Administration Cooperative Urological Research Group Studies. NC1 Monogr. 7:165170;1988. Bagshaw, M. A. Current conflicts in the management of

216

I. J. Radiation Oncology 0 Biology 0 Physics

Volume 26, Number 2, 1993

12. Fowler, J. F. Potential for increasing the differential response

13.

14.

15.

16.

17.

18.

19.

20.

between tumors and normal tissues. Can proliferation rate be used? Int. J. Radiat. Oncol. Biol. Phys. 12:64 l-645; 1986. Fowler, J. F. The effect of overall treatment time in radiotherapy for localized prostate carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 2 I : 1097- 1098; 199 1. Green, N.; Bodner, H.; Broth, E.; Chiang, C.; Garrett, J.; Goldstein, A.; Goldberg, H.; Gualtieri, V.; Gray, R.; Jaffe, J.; Kaplan, R.; Polse, S.; Ross, S.; Skaist, L.; Treible, D.; Vatz, A.; Wallack, H. Improved control of bulky prostate carcinoma with sequential estrogen and radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 10:971-976;1984. Hanks, G. E. Optimizing the radiation treatment and outcome of prostatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 11:1235-1245;1985. Hanks, G. E.; Leibel, S. A.; Krall, J. M.; Kramer, S. Patterns of care studies: Dose-response observations for local control of carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 11:153-157;1985. Holsti, L. R.; Mlntyll, M. Split-course versus continuous radiotherapy. Analysis of a randomized trial from 1964 to 1967. Acta Oncol. 27:153-161;1988. Kajanti, M.; Holsti, L. R.; Holsti, P. Radical surgery and postoperative split-course radiotherapy in squamous cell carcinoma of the mobile tongue: Factors influencing local control and time to recurrence. Radiother. Oncol. 22: 174179;1991. Lai, P. P.; Perez, C. A.; Shapiro, S. J.; Lockett, M. A. Carcinoma of the prostate stage B and C: Lack of influence of duration on tumor control and treatment morbidity. Int. J. Radiat. Oncol. Biol. Phys. 19:56 l-568; 1990. Lai, P. P.; Pilepich, M. V.; Krall, J. M.; Asbell, S. 0.; Hanks, G. E.; Perez, C. A.; Rubin, P.; Sause, W. T.; Cox, J. D. The

21.

22.

23.

24.

25.

26.

27.

28.

effect of overall treatment time on the outcome ofdefinitive radiotherapy for localized prostate carcinoma: The Radiation Therapy Oncology Group 75-06 and 77-06 experience. Int. J. Radiat. Oncol. Biol. Phys. 2 1:925-933; 199 1. Mar&l, V. A.; Amato, D. A.; Pajak, T. F. Patterns of failure after treatment for cancer of upper respiratory and digestive tracts: A Radiation Therapy Oncology Group report. Cancer Treat. Symp. 2:33-40; 1983. Meyer, J. S. Potential value of cell kinetics in management of cancer of unknown origin. Semin. Oncol. 9:513521;1982. Nemoto, R.; Hattori, K.; Uchida, K.; Shimazui, K.; Nishijima, Y.; Koiso, K.; Harada, M. S-phase fraction of human prostate adenocarcinoma studied with in vivo bromodeoxyuridine labeling. Cancer 66:509-5 14; 1990. Parsons, J. T.; Thar, T. L.; Bova, F. J.; Million, R. R. An evaluation of split-course irradiation for pelvic malignancies. J. Radiat. Oncol. Biol. Phys. 6: 175-I 8 1; 1980. Torres, J. L.; Lee, D. J.; Leibel, S. A.; Wharam, M. D.; Cantrell, B. B.; Brass, D. S.; Order, S. E. Local control and reduced complications in split-course irradiation of prostatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 7:43-47;1981. Rew, D. A.; Thomas, D. J.; Coptcoat, M.; Wilson, G. D. Measurement of in vivo urological tumor cell kinetics using multiparameter flow cytometry. Br. J. Urol. 6844-48; 199 1. Withers, H. R.; Taylor, J. M. G.; Maciejewski, B. The hazard of accelerated tumor clonogen repopulation during radiotherapy. Acta Oncol. 2:131-146;1988. Zagars, G. K.; von Eschenbach, A. C.; Johnson, D. E.; OSwald, M. J. Stage C adenocarcinoma of the prostate. An analysis of 55 1 patients treated with external beam radiation. Cancer 60:1489-1499;1987.