- Email: [email protected]
Patterns of care studies: Dose-response observations for local control of adenocarcinoma of the prostate
0360-3016/85 $03.00 + .I0 Copyright Q 1985 Pergamon Res Ltd.
Inr. J. Rdiarwn Oncology Bid. Phw.. Vol I I, pp. 153-157 Printed in the U.S.A. All rights reserved.
0 Clinical Radiation Therapy PATTERNS
OF CARE STUDIES: DOSE-RESPONSE OBSERVATIONS FOR LOCAL CONTROL OF ADENOCARCINOMA OF THE PROSTATE GERALD E. HANKS,
M.D.
Radiation Oncology Center, Sutter Community Hospitals, Sacramento, CA 958 19 STEVEN
A.
LEIBEL,
M.D.
Department of Radiation Oncology, University of California Medical Center, San Francisco, CA 94143 JOHN
M.
KRALL,
PH.D.
American College of Radiology, Philadelphia, PA 19 107 AND SIMON
KRAMER,
M.D.
Department of Radiation Therapy and Nuclear Medicine, Thomas Jefferson University Hospital, Philadelphia, PA 19107 Five hundred seventy-four patients with prostate cancer treated by external beam radiation therapy in the United States in 1973 to 1975 have been analyzed comparing radiation dose with in-field recurrence. Dose-response effects are observed for all cases (p = c.05) and T-2 and T-3 tumors, but not for T-O, T-l and T-4 tumors. For doses calculated at the center of the prostate, these observations suggest optimal control is obtained at no more than 6000 rad for T-O and T-l tumors; 6000-6500 rad for T-2 tumors; 6500-7000 rad for T-3 tumors; and that greater than 7000 rad is required only for T-4 tumors. The paraprostatic dose calculated at a point 4 cm lateral to the center of the prostate also shows a correlation of dose with infield failure for all cases (p = .Ol). Observations in individual T states suggest optimal control is obtained at no more than 6000 rad for T-O, T-l and T-2 tumors, 6500-6999 rad for T-3 and ~7000 rad for T-4. These data suggest that for T-2 and T-3 cancers, extension in the periprostatic region must be treated. A comparison of central dose vs. stage indicates institutional policy rather than cancer volume determines the radiation dose used in treating prostate cancer. A change in institutional policies to treat with optimal doses as indicated by this study would result in an overall increase in local control and a decrease in complications. Prostate cancer, Dose-response,
Patterns of Care, Complications.
INTRODUCIION
necessary to demonstrate an effect of dose; when variations do occur, they are not random, but determined by special patient circumstances, which may invalidate their use in determining dose-response relationships. Second, reports of the results of treatment of prostate cancer in single institutions illustrate the slow accumulation of cases. In 7 recent series, this varied from 10 to 27 patients per year, with total patient populations reported of 96 to 310 patients.3~8~‘0-12~‘4~‘ Changes 5 in patient characteristics and treatment programs over these long time spans make conclusions about dose response observations suspect and obviously none of these reports
Radiation therapy is commonly used in the curative management of localized adenocarcinoma of the prostate.‘1,‘2*‘4q’5Although the tumor volume may vary by as much as 20-fold, there is minimal evidence indicating what dose is necessary to obtain optimal local control for the various stages or T extents that characterize these changes in tumor volume. Two situations have resulted in this information gap: First, tumor doses delivered in an individual institution are tightly controlled and do not provide the variation
Funded by the National Cancer Institute, Division of Cancer Prevention and Control, Grant CA 15978, Simon Kramer, M.D., Principal Investigator. Reprint requests to: American College of Radiology, PCS
Publications Dep’t., 925 Chestnut Street, Philadelphia, PA 19107. Accepted for publication 17 Aug 1984.
153
154
Radiation Oncology 0 Biology 0 Physics Table 1. Patterns of Care Outcome Study: Adenocarcinoma
January 1985, Volume 11, Number 1 of the prostate (N-O, N-X, M-0)-Radiation
dose vs. T state
Stage
<6000
%
6000-6999
%
27000
%
Number of patients
T-O
6 26 24 20 20
11 18 18 12 26
32 73 68 82 33
58 50 51 50 43
17 48 41 61 23
31 33 31 37 30
55 147 133 163 76
Total
574
T-l T-2 T-3 T-4
is a random sample to which eration can be applied. METHODS
AND
proper
statistical
consid-
MATERIALS
This analysis was performed on the Patterns of Care Outcome Survey data base for cancer of the prostate treated in the United States in 1973-75. Six hundred eighty-two patient records were reviewed at 163 facilities randomly selected from the 1000 present in the United States. One hundred eight patients were excluded (37 with positive lymph nodes, 14 with a missing dose and 57 with unknown stage), leaving 574 patients in the current analysis. The methodology and statistical control of these surveys have been previously published, as have our national benchmarks for prostate cancer.2,5-7 The endpoint used in this report is 4 year actuarial in-field failure rate by the Kaplan-Meier method.4 Statistical significance is calculated both for whole curve comparisons and for linear trend.‘,13 Doses stated were recalculated by the surveying physicists at the center of the prostate gland and at a point 4 cm lateral to the center of the prostate. In-field failure is the clinical impression recorded in the record by the physician following the patient. Biopsy confirmation was uncommon. RESULTS Table 1 lists the 574 patients who have been analyzed for in-field failure by stage and center of prostate radia-
Table 2. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-O, 1, 2, 3, 4, N-O, N-X, M-O Primary dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients
4 year free of recurrence (%)
10 21 65 93 195 190
75 76 80 84 89 90
tion dose. Fifty percent of patients received between 6000 and 7000 rad, 33% more than 7000 rad and 17% less than 6000 rad. A significant dose response relationship is observed for the central prostate point with the pooled data as shown in Table 2. Tables 3 a, b, c, d and e present the dose-response observations by individual T state for this central prostate point. No significant effect is observed for T-O or T-l tumors. There is a significant doseresponse effect for T-2 and T-3 tumors and a trend is apparent for T-4 tumors but it is not significant. A strong dose-response relationship for in-field recurrence for the combined T state paraprostatic dose point is illustrated in Table 4. In-field recurrences are presented
Table 3a. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-O, N-O, N-X, M-O Primary dose @ad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
patients
4 year free of recurrence (%)
0
100
1
100
5 9 23 17
82 87 100 100
55 Not significant.
Table 3b. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-l, N-O, N-X, M-O Primary dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
574 Whole curve p = .07; Linear trend p < .Ol.
Number of
Number of patients 1 6 19 21 52 48 147
Not significant.
4 year free of recurrence (%) 100 100 95 95 96 92
155
Control of prostate adenocarcinoma 0 G. E. HANKS ef al. Table 3c. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-2, N-O, N-X, M-O Primary dose (rad)
Number of patients
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
4 year free of recurrence (W)
1 6 17 17 51 41
100 43 80 87 87 90
Table 4. Patterns of Care Oucome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-O, 1, 2, 3, 4, N-O, N-X, M-O Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Table 3d. Patterns of Care Outcome Study: Carcinoma of the Prostate-In-field recurrence vs. center of prostate dose T-3, N-O, N-X, M-O 4 year free of recurrence (%) 44 100 54 75 88 89
3 3 14 29 53 61
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
81 83 94 105 112 88
79 84 85 83 94 93
Whole curve p = .Ol; Linear trend p < .Ol.
Whole curve JJ < .05; Linear trend p < .05.
Number of patients
4 year free of recurrence (%)
563
133
Primary dose (rad)
Number of patients
Stage C patients, only 13 received less than 6000 rad and the curve is dependent on the 5 failures observed in that group. Neglia et al. I’ have reported no difference in local control for their Stage C-l and C-2 patients when less than 6750 rad was given as compared to et al.’ have reported greater than 6750 rad. Duttenhaver no difference in local control when comparing a group of patients who received 6000-6800 rad to a group boosted with protons to 7000-7650 rad. These are the only reports of dose-response effects free of major questions of validity.
163
Table 5a. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-O, N-O, N-X, M-O
Whole curve p < .05; Linear trend p < .Ol.
Table 3e. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-4, N-O, N-X, M-O Primary dose (rad)
Number of patients
4 year free of recurrence (%)
5 5 10 17 16 23
55 52 84 69 64 84
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Para-prostatic dose (rad)
Number of patients
<5000 5000-5499 5500-5999
9 6 7
6000-6499 6500-6999 27000
9 14 7
4 year free of recurrence (%) 88 83 100 100 100 100
52 Not significant.
76 Table 5b. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T- 1, N-O, N-X, M-O
Not significant.
for each T state in Tables response trend is strongly tumors.
5 a, b, c, d and e. A dosesuggested for T-2 and T-3
DISCUSSION Perez et al. I2 have reported dose-response observations for Stage C prostate cancer. They report a significant trend to decreased recurrence with increased dose (p = 0.05) as radiation dose increases from less than 6000 rad to greater than 7000 rad. In their group of 139
Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients
25 25 21 30 22 143
Not significant.
4 year free of recurrence (%) 96 93 90 95 96 96
156
Radiation Oncology 0 Biology 0 Physics
Table 5c. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-2. N-O. N-X. M-O Para-prostatic dose (rad)
Number of natients
4 year free of recurrence (%)
24 16 21 26 23 21
68 83 90 86 96 90
131 Whole curve p >
10 (not significant);
Linear trend p < .05.
Table 5d. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-3, N-O, N-X, M-O Para-prostatic dose (rad) (5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients 14 24 29 32 36 26
4 year free of recurrence (%) 73 80 83 79 94 87
161 Whole curve p > 10 (not significant);
Linear trend p = .05.
Table 5e. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-4, N-O, N-X, M-O Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 17000
Number of patients
4 year free of recurrence (%)
9 12 16 18 9 12
64 89 57 53 81 100
76 Whole curve p = < .lO; Linear trend p = not significant.
Our data illustrate dose-response relationships for each T state of prostate cancer for dose calculated at the center of the gland and in the paraprostatic tissue, and may be used to optimize therapy by tumor extent. The failure to observe a dose-response effect for T-O and T-l tumors suggests that 6000 rad is adequate. For T-2, T-3 and T-4 tumors, dose-response relationships are observed and suggest the optimum dose is 60006500 rad for T-2 tumors; 6500-7000 rad for T-3 tumors;
January 1985, Volume
1 I,
Number 1
and that only for T-4 tumors is it necessary to exceed 7000 rad. For T-2 and T-3 cancers, significant doseresponse relationships are observed in the paraprostatic region, suggesting that in addition to clinically noted extracapsular extension in T-3 patients, T-2 tumors have subclinical paraprostatic extension that must be treated. Doses required for optimal control are similar to those at the center of the prostate. The pooled data for central prostate dose indicate 96 patients were treated to less than 6000 rad. These patients are an extremely valuable resource available in this statistically appropriate retrospective review of all strata of practice in the United States, as they allow the demonstration of dose-response effect. These data could not be obtained in a prospective random trial as ethical considerations would prohibit treating prostate cancer with these inadequate doses. Only 6% of in-field recurrences were noted in T-O and T-l tumors treated to this low dose, but 34% of T-2, T-3 and T-4 tumors treated with these low doses had in-field recurrences. When T-2, 3 or 4 tumors were treated to the optimum described in this report, the in-field failure rate was lo15%. Clearly, greater local control can be obtained with appropriate dose selection. The pooled data also indicate 190 patients received greater than 7000 rad to the central prostate dose point. When 7000 rad was exceeded, the complication rate doubled from 3.5% to 7%.12 Only T-4 tumors (23 of the 190 patients receiving 27000 rad) required 7000 rad to obtain optimal in-field control indicating that in 167 patients (3 1% of T-O, 33% of T-l, 3 1% of T-2, and 37% of T-3) the increased risk of complications associated with 7000 rad could have been avoided without a loss in local control. These data also indicate that institutional policy rather than tumor extent is the determinant of dose in prostate cancer in the United States. This is illustrated in Table 1 by the percentages of patients receiving a given dose interval by T state. T-O, T-l, T-2 and T-3 patient groups are remarkably similar with the ~6000 rad dose interval comprising 1 l-18% of each T state; the 6000-6999 rad interval comprising 50-58% of each T state and the greater than 7000 rad group comprising 3 l-37% of each T state. There is an indication of a shift between the <6000 rad group and the 6000-6999 rad group for T-4 tumors, perhaps representing a stage related dose decision in patients with T-4 tumors. Nonetheless, the fraction of T-4 tumors receiving 7000 rad or more is identical to the other T states. This dominance of institutional policy is important as this paper indicates an alteration of treatment policy from a single dose program for the treatment of all stages of prostate cancer, to a treatment policy that adjusts for individual tumor size would result in improved patient outcome for tumor control and patient complications.
Control of prostate adenocarcinoma 0 G.
E. HANKS et
al.
157
REFERENCES 1. Duttenhaver, J.R., Shipley, W.U., Perrone, T., Verhey, L.J., Goitein, M., Munzemider, J.E., Prout, G.R., Parkburst, E.C., Suit, H.D.: Protons or megavoltage X-rays as boost therapy for patients irradiated for localized prostatic carcinoma. Cancer 51: 1599-1604, 1983. 2. Hanks, G.E., Kramer, S., Diamond, J.J., Herring, D.F.: Patterns of Care Outcome Survey: National outcome data for six disease sites. Am. J. Clin. Oncol. 5: 349-353, 1982. 3. Harisiadis, L., Veenema, R.J., Senyszym, J.J., Puchner, P.J., Tretter, P., Romas, N.A., Chang, C.H., Lattimer, J.K., Tannenbaum, M.: Carcinoma of the prostate: Treatment with external radiotherapy. Cancer 41: 2131-2142, 1978. 4. Kaplan, E.L., Meier, P.: Nonparametric estimation from J. Am. Stat. Assoc. 53: 457-481, incomplete observations.
1958. 5. Kramer,
S.: The Patterns of Care Study: A nationwide evaluation of the practice of radiation therapy in cancer management. Int. J. Radiat. Oncol. Biol. Phys. 1: 12311236, 1976. 6. Kramer, S.: The study of patterns of cancer care in radiation therapy. Cancer 39: 780-787, 1977. 7. Leibel, S.A., Hanks, G.E., Kramer, S.: Patterns of Care Outcome Studies: Results of the National Practice in adenocarcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 10: 401-409, 1984. 8. Lipsette, A.L., Cosgrove, M.D., Green, N., Casagrande, MS., Melbye, R.W., George, F.W., III: Factors influencing
prognosis in the radiotherapeutic management of carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 1: 1049-1058, 1976. 9. Mantel, N.: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother. Rep. 50: 163-170, 1966. 10. McGowan, D.G.: Radiation therapy in the management of localized carcinoma of the prostate. Cancer 39: 98-103, 1977. 11. Neglia, W.J., Hussey, D.H., Johnson, D.E.: Megavoltage radiation therapy for carcinoma of the prostate. Znt. J.
Radiat. Oncol. Biol. Phys. 2: 873-882, 1977. 12. Perez, C.A., Walz, B.J., Zivnuska, F.R., Pilepich,
M., Prasad, K., Bauer, W.: Irradiation of carcinoma of the prostate localized to the pelvis: Analysis of tumor response and prognosis. Int. J. Radiat. Oncol. Biol. Phys. 6: 555563, 1980. 13. Peto, R., Pike, M.C., Armitage, P., Breslow, N.E., Cox, D.R., Howard, S.V., Mantel, N., McPherson, K., Peto, J., Smith, P.G.: Design and analysis of randomized clinical trials requiring prolonged observation of each patient II. Analysis and Examples. Br. J. Cancer 35: l-39, 1977. 14. Ray, G.R., Cassady, J.R., Bagshaw, M.A.: Definitive radiation therapy of carcinoma of the prostate. Radiology
106: 47-48, 1973. 15. Taylor, W.J., Richardson, diation
therapy
for localized
1123-l 127, 1979.
R.G., Hafermann, prostate cancer.
M.D.:
Ra-
Cancer 43:
Inr. J. Rdiarwn Oncology Bid. Phw.. Vol I I, pp. 153-157 Printed in the U.S.A. All rights reserved.
0 Clinical Radiation Therapy PATTERNS
OF CARE STUDIES: DOSE-RESPONSE OBSERVATIONS FOR LOCAL CONTROL OF ADENOCARCINOMA OF THE PROSTATE GERALD E. HANKS,
M.D.
Radiation Oncology Center, Sutter Community Hospitals, Sacramento, CA 958 19 STEVEN
A.
LEIBEL,
M.D.
Department of Radiation Oncology, University of California Medical Center, San Francisco, CA 94143 JOHN
M.
KRALL,
PH.D.
American College of Radiology, Philadelphia, PA 19 107 AND SIMON
KRAMER,
M.D.
Department of Radiation Therapy and Nuclear Medicine, Thomas Jefferson University Hospital, Philadelphia, PA 19107 Five hundred seventy-four patients with prostate cancer treated by external beam radiation therapy in the United States in 1973 to 1975 have been analyzed comparing radiation dose with in-field recurrence. Dose-response effects are observed for all cases (p = c.05) and T-2 and T-3 tumors, but not for T-O, T-l and T-4 tumors. For doses calculated at the center of the prostate, these observations suggest optimal control is obtained at no more than 6000 rad for T-O and T-l tumors; 6000-6500 rad for T-2 tumors; 6500-7000 rad for T-3 tumors; and that greater than 7000 rad is required only for T-4 tumors. The paraprostatic dose calculated at a point 4 cm lateral to the center of the prostate also shows a correlation of dose with infield failure for all cases (p = .Ol). Observations in individual T states suggest optimal control is obtained at no more than 6000 rad for T-O, T-l and T-2 tumors, 6500-6999 rad for T-3 and ~7000 rad for T-4. These data suggest that for T-2 and T-3 cancers, extension in the periprostatic region must be treated. A comparison of central dose vs. stage indicates institutional policy rather than cancer volume determines the radiation dose used in treating prostate cancer. A change in institutional policies to treat with optimal doses as indicated by this study would result in an overall increase in local control and a decrease in complications. Prostate cancer, Dose-response,
Patterns of Care, Complications.
INTRODUCIION
necessary to demonstrate an effect of dose; when variations do occur, they are not random, but determined by special patient circumstances, which may invalidate their use in determining dose-response relationships. Second, reports of the results of treatment of prostate cancer in single institutions illustrate the slow accumulation of cases. In 7 recent series, this varied from 10 to 27 patients per year, with total patient populations reported of 96 to 310 patients.3~8~‘0-12~‘4~‘ Changes 5 in patient characteristics and treatment programs over these long time spans make conclusions about dose response observations suspect and obviously none of these reports
Radiation therapy is commonly used in the curative management of localized adenocarcinoma of the prostate.‘1,‘2*‘4q’5Although the tumor volume may vary by as much as 20-fold, there is minimal evidence indicating what dose is necessary to obtain optimal local control for the various stages or T extents that characterize these changes in tumor volume. Two situations have resulted in this information gap: First, tumor doses delivered in an individual institution are tightly controlled and do not provide the variation
Funded by the National Cancer Institute, Division of Cancer Prevention and Control, Grant CA 15978, Simon Kramer, M.D., Principal Investigator. Reprint requests to: American College of Radiology, PCS
Publications Dep’t., 925 Chestnut Street, Philadelphia, PA 19107. Accepted for publication 17 Aug 1984.
153
154
Radiation Oncology 0 Biology 0 Physics Table 1. Patterns of Care Outcome Study: Adenocarcinoma
January 1985, Volume 11, Number 1 of the prostate (N-O, N-X, M-0)-Radiation
dose vs. T state
Stage
<6000
%
6000-6999
%
27000
%
Number of patients
T-O
6 26 24 20 20
11 18 18 12 26
32 73 68 82 33
58 50 51 50 43
17 48 41 61 23
31 33 31 37 30
55 147 133 163 76
Total
574
T-l T-2 T-3 T-4
is a random sample to which eration can be applied. METHODS
AND
proper
statistical
consid-
MATERIALS
This analysis was performed on the Patterns of Care Outcome Survey data base for cancer of the prostate treated in the United States in 1973-75. Six hundred eighty-two patient records were reviewed at 163 facilities randomly selected from the 1000 present in the United States. One hundred eight patients were excluded (37 with positive lymph nodes, 14 with a missing dose and 57 with unknown stage), leaving 574 patients in the current analysis. The methodology and statistical control of these surveys have been previously published, as have our national benchmarks for prostate cancer.2,5-7 The endpoint used in this report is 4 year actuarial in-field failure rate by the Kaplan-Meier method.4 Statistical significance is calculated both for whole curve comparisons and for linear trend.‘,13 Doses stated were recalculated by the surveying physicists at the center of the prostate gland and at a point 4 cm lateral to the center of the prostate. In-field failure is the clinical impression recorded in the record by the physician following the patient. Biopsy confirmation was uncommon. RESULTS Table 1 lists the 574 patients who have been analyzed for in-field failure by stage and center of prostate radia-
Table 2. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-O, 1, 2, 3, 4, N-O, N-X, M-O Primary dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients
4 year free of recurrence (%)
10 21 65 93 195 190
75 76 80 84 89 90
tion dose. Fifty percent of patients received between 6000 and 7000 rad, 33% more than 7000 rad and 17% less than 6000 rad. A significant dose response relationship is observed for the central prostate point with the pooled data as shown in Table 2. Tables 3 a, b, c, d and e present the dose-response observations by individual T state for this central prostate point. No significant effect is observed for T-O or T-l tumors. There is a significant doseresponse effect for T-2 and T-3 tumors and a trend is apparent for T-4 tumors but it is not significant. A strong dose-response relationship for in-field recurrence for the combined T state paraprostatic dose point is illustrated in Table 4. In-field recurrences are presented
Table 3a. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-O, N-O, N-X, M-O Primary dose @ad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
patients
4 year free of recurrence (%)
0
100
1
100
5 9 23 17
82 87 100 100
55 Not significant.
Table 3b. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-l, N-O, N-X, M-O Primary dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
574 Whole curve p = .07; Linear trend p < .Ol.
Number of
Number of patients 1 6 19 21 52 48 147
Not significant.
4 year free of recurrence (%) 100 100 95 95 96 92
155
Control of prostate adenocarcinoma 0 G. E. HANKS ef al. Table 3c. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-2, N-O, N-X, M-O Primary dose (rad)
Number of patients
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
4 year free of recurrence (W)
1 6 17 17 51 41
100 43 80 87 87 90
Table 4. Patterns of Care Oucome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-O, 1, 2, 3, 4, N-O, N-X, M-O Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Table 3d. Patterns of Care Outcome Study: Carcinoma of the Prostate-In-field recurrence vs. center of prostate dose T-3, N-O, N-X, M-O 4 year free of recurrence (%) 44 100 54 75 88 89
3 3 14 29 53 61
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
81 83 94 105 112 88
79 84 85 83 94 93
Whole curve p = .Ol; Linear trend p < .Ol.
Whole curve JJ < .05; Linear trend p < .05.
Number of patients
4 year free of recurrence (%)
563
133
Primary dose (rad)
Number of patients
Stage C patients, only 13 received less than 6000 rad and the curve is dependent on the 5 failures observed in that group. Neglia et al. I’ have reported no difference in local control for their Stage C-l and C-2 patients when less than 6750 rad was given as compared to et al.’ have reported greater than 6750 rad. Duttenhaver no difference in local control when comparing a group of patients who received 6000-6800 rad to a group boosted with protons to 7000-7650 rad. These are the only reports of dose-response effects free of major questions of validity.
163
Table 5a. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-O, N-O, N-X, M-O
Whole curve p < .05; Linear trend p < .Ol.
Table 3e. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. center of prostate dose T-4, N-O, N-X, M-O Primary dose (rad)
Number of patients
4 year free of recurrence (%)
5 5 10 17 16 23
55 52 84 69 64 84
<5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Para-prostatic dose (rad)
Number of patients
<5000 5000-5499 5500-5999
9 6 7
6000-6499 6500-6999 27000
9 14 7
4 year free of recurrence (%) 88 83 100 100 100 100
52 Not significant.
76 Table 5b. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T- 1, N-O, N-X, M-O
Not significant.
for each T state in Tables response trend is strongly tumors.
5 a, b, c, d and e. A dosesuggested for T-2 and T-3
DISCUSSION Perez et al. I2 have reported dose-response observations for Stage C prostate cancer. They report a significant trend to decreased recurrence with increased dose (p = 0.05) as radiation dose increases from less than 6000 rad to greater than 7000 rad. In their group of 139
Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients
25 25 21 30 22 143
Not significant.
4 year free of recurrence (%) 96 93 90 95 96 96
156
Radiation Oncology 0 Biology 0 Physics
Table 5c. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-2. N-O. N-X. M-O Para-prostatic dose (rad)
Number of natients
4 year free of recurrence (%)
24 16 21 26 23 21
68 83 90 86 96 90
131 Whole curve p >
10 (not significant);
Linear trend p < .05.
Table 5d. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-3, N-O, N-X, M-O Para-prostatic dose (rad) (5000 5000-5499 5500-5999 6000-6499 6500-6999 27000
Number of patients 14 24 29 32 36 26
4 year free of recurrence (%) 73 80 83 79 94 87
161 Whole curve p > 10 (not significant);
Linear trend p = .05.
Table 5e. Patterns of Care Outcome Study: Carcinoma of the prostate-In-field recurrence vs. para-prostatic dose T-4, N-O, N-X, M-O Para-prostatic dose (rad) <5000 5000-5499 5500-5999 6000-6499 6500-6999 17000
Number of patients
4 year free of recurrence (%)
9 12 16 18 9 12
64 89 57 53 81 100
76 Whole curve p = < .lO; Linear trend p = not significant.
Our data illustrate dose-response relationships for each T state of prostate cancer for dose calculated at the center of the gland and in the paraprostatic tissue, and may be used to optimize therapy by tumor extent. The failure to observe a dose-response effect for T-O and T-l tumors suggests that 6000 rad is adequate. For T-2, T-3 and T-4 tumors, dose-response relationships are observed and suggest the optimum dose is 60006500 rad for T-2 tumors; 6500-7000 rad for T-3 tumors;
January 1985, Volume
1 I,
Number 1
and that only for T-4 tumors is it necessary to exceed 7000 rad. For T-2 and T-3 cancers, significant doseresponse relationships are observed in the paraprostatic region, suggesting that in addition to clinically noted extracapsular extension in T-3 patients, T-2 tumors have subclinical paraprostatic extension that must be treated. Doses required for optimal control are similar to those at the center of the prostate. The pooled data for central prostate dose indicate 96 patients were treated to less than 6000 rad. These patients are an extremely valuable resource available in this statistically appropriate retrospective review of all strata of practice in the United States, as they allow the demonstration of dose-response effect. These data could not be obtained in a prospective random trial as ethical considerations would prohibit treating prostate cancer with these inadequate doses. Only 6% of in-field recurrences were noted in T-O and T-l tumors treated to this low dose, but 34% of T-2, T-3 and T-4 tumors treated with these low doses had in-field recurrences. When T-2, 3 or 4 tumors were treated to the optimum described in this report, the in-field failure rate was lo15%. Clearly, greater local control can be obtained with appropriate dose selection. The pooled data also indicate 190 patients received greater than 7000 rad to the central prostate dose point. When 7000 rad was exceeded, the complication rate doubled from 3.5% to 7%.12 Only T-4 tumors (23 of the 190 patients receiving 27000 rad) required 7000 rad to obtain optimal in-field control indicating that in 167 patients (3 1% of T-O, 33% of T-l, 3 1% of T-2, and 37% of T-3) the increased risk of complications associated with 7000 rad could have been avoided without a loss in local control. These data also indicate that institutional policy rather than tumor extent is the determinant of dose in prostate cancer in the United States. This is illustrated in Table 1 by the percentages of patients receiving a given dose interval by T state. T-O, T-l, T-2 and T-3 patient groups are remarkably similar with the ~6000 rad dose interval comprising 1 l-18% of each T state; the 6000-6999 rad interval comprising 50-58% of each T state and the greater than 7000 rad group comprising 3 l-37% of each T state. There is an indication of a shift between the <6000 rad group and the 6000-6999 rad group for T-4 tumors, perhaps representing a stage related dose decision in patients with T-4 tumors. Nonetheless, the fraction of T-4 tumors receiving 7000 rad or more is identical to the other T states. This dominance of institutional policy is important as this paper indicates an alteration of treatment policy from a single dose program for the treatment of all stages of prostate cancer, to a treatment policy that adjusts for individual tumor size would result in improved patient outcome for tumor control and patient complications.
Control of prostate adenocarcinoma 0 G.
E. HANKS et
al.
157
REFERENCES 1. Duttenhaver, J.R., Shipley, W.U., Perrone, T., Verhey, L.J., Goitein, M., Munzemider, J.E., Prout, G.R., Parkburst, E.C., Suit, H.D.: Protons or megavoltage X-rays as boost therapy for patients irradiated for localized prostatic carcinoma. Cancer 51: 1599-1604, 1983. 2. Hanks, G.E., Kramer, S., Diamond, J.J., Herring, D.F.: Patterns of Care Outcome Survey: National outcome data for six disease sites. Am. J. Clin. Oncol. 5: 349-353, 1982. 3. Harisiadis, L., Veenema, R.J., Senyszym, J.J., Puchner, P.J., Tretter, P., Romas, N.A., Chang, C.H., Lattimer, J.K., Tannenbaum, M.: Carcinoma of the prostate: Treatment with external radiotherapy. Cancer 41: 2131-2142, 1978. 4. Kaplan, E.L., Meier, P.: Nonparametric estimation from J. Am. Stat. Assoc. 53: 457-481, incomplete observations.
1958. 5. Kramer,
S.: The Patterns of Care Study: A nationwide evaluation of the practice of radiation therapy in cancer management. Int. J. Radiat. Oncol. Biol. Phys. 1: 12311236, 1976. 6. Kramer, S.: The study of patterns of cancer care in radiation therapy. Cancer 39: 780-787, 1977. 7. Leibel, S.A., Hanks, G.E., Kramer, S.: Patterns of Care Outcome Studies: Results of the National Practice in adenocarcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 10: 401-409, 1984. 8. Lipsette, A.L., Cosgrove, M.D., Green, N., Casagrande, MS., Melbye, R.W., George, F.W., III: Factors influencing
prognosis in the radiotherapeutic management of carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 1: 1049-1058, 1976. 9. Mantel, N.: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother. Rep. 50: 163-170, 1966. 10. McGowan, D.G.: Radiation therapy in the management of localized carcinoma of the prostate. Cancer 39: 98-103, 1977. 11. Neglia, W.J., Hussey, D.H., Johnson, D.E.: Megavoltage radiation therapy for carcinoma of the prostate. Znt. J.
Radiat. Oncol. Biol. Phys. 2: 873-882, 1977. 12. Perez, C.A., Walz, B.J., Zivnuska, F.R., Pilepich,
M., Prasad, K., Bauer, W.: Irradiation of carcinoma of the prostate localized to the pelvis: Analysis of tumor response and prognosis. Int. J. Radiat. Oncol. Biol. Phys. 6: 555563, 1980. 13. Peto, R., Pike, M.C., Armitage, P., Breslow, N.E., Cox, D.R., Howard, S.V., Mantel, N., McPherson, K., Peto, J., Smith, P.G.: Design and analysis of randomized clinical trials requiring prolonged observation of each patient II. Analysis and Examples. Br. J. Cancer 35: l-39, 1977. 14. Ray, G.R., Cassady, J.R., Bagshaw, M.A.: Definitive radiation therapy of carcinoma of the prostate. Radiology
106: 47-48, 1973. 15. Taylor, W.J., Richardson, diation
therapy
for localized
1123-l 127, 1979.
R.G., Hafermann, prostate cancer.
M.D.:
Ra-
Cancer 43: