Nina Fukunaga-Johnson, M.D.*, Howard M. Sandler, M.D.*, P. William McLaughlin, M.D.**, Myla S. Strawderman, M.S.**, Katherine H. Grijalva, M.A.**, Kathleen E. Kish, B.A.* and Allen S. Lichter, M.D.*

*Department of Radiation Oncology, University of Michigan Medical Center, Ann Arbor, MI and **Providence Hospital, Southfield, MI

---

Purpose

3D conformal radiotherapy (3D CRT) has been shown to decrease acute morbidity in the treatment of prostate cancer. Therapeutic outcome and late morbidity data have been accumulating. To evaluate the results of 3D CRT for the treatment of prostate cancer, we analyzed the outcome of a large series of patients treated with conformal techniques.

Material and Methods

From January 1987 through June 1994, 707 patients with localized prostate cancer were treated with 3D CRT. Patients with pathologicaly confirmed pelvic lvmph node metastasis, treated with pre-irradiation (preRT) androgen ablation, or treated post-prostatectomn, were excluded. All had CT obtained specifically for treatment planning, multiple structures contoured on the axial images, and beam's-eye view conformal beams edited to provide 3D dose coverage. Median follow-up is 36 mos; 70 patients have been followed longer than 5.5 vears. Six hundred three had T1-T2 tumors. PreRT prostate specific antigen (PSA) was available for 649 patients: median preRT PSA was 12.9 ng/ml, 209 patients had preRT PSA > 20 ng/ml. The median dose of radiation was 69 Gy; 102 patients received >= 69 Gy. Biochemical failure was defined as: 1) two consecutive PSA rises over 2.0 ng/ml if nadir PSA <= 2.0 ng/ml, 2) two consecutive PSA rises over nadir if nadir PSA > 2.0 ng/ml, or 3) initiation of hormonal therapy after RT. Complications were graded using the RTOG system.

Results

PreRT PSA and Gleason score emerged as independent indicators of biochemical control (bNED). Patients with preRT PSA > 10 had a significantly worse bNED at 5 years than patients with preRT PSA <= 10. Five-year bNED was determined according to preRT PSA: PSA <= 4, 88%; PSA > 4 <=10, 72%; PSA > IO <= 20, 43%; and PSA > 20, 30%. Patients with Gleason score >= 7 also had a significantly worse bNED than patients with Gleason score < 7. Patients were divided into two prognostic groups: a favorable group with PSA <= 10, Gleason score < 7, and T1-T2 tumors, and an unfavorable group with PSA > IO, Gleason score >= 7 or T3-T4 tumor and studied for the effect of dose on bNED status. The bNED at 5 years was 75% for the favorable group and 37% for the unfavorable group. In addition, a group that might be considered a surgical subset was reviewed: patients with PSA <= 10, Gleason score <= 7, and T1-T2 tumors who were <70 years old. This subset had an 84% 5-year bNED rate and 98% 5-year overall survival. Complications with the techniques used here are very low: 3% risk at 7 years of Grade 3-4 complications and 1% risk at 7 years of Grade 3 bladder complications (no Grade 4).

Conclusion

3D CRT allows for treatment of prostate cancers with a very low risk of complications. Patients with relatively, eariv disease as defined bv preRT PSA, Gleason score < 7, and T1-2 tumors and patients who are candidates for radical prostatectomv have excellent 5-year bNED rates. Patients with adverse prognostic factors have a high risk of biochemical recurrence and are candidates for innovative therapy.

Introduction

Radiation therapy (RT) has been accepted as a standard treatment of localized prostate cancer. It is becoming more apparent that accurate localization is necessary to encompass the target volume with the intended radiation treatment and maybe essential for improving outcome¹⁶. Conformal radiotherapy (CRT) has been developed to encompass the target volume with the prescribed dose while minimizing the dose to surrounding normal tissue¹⁸. CRT utilizes the creation of 3D dose distributions based on anatomic images and requires various tools including Beam's Eye View (BEV) block design and dose calculation algorithms that facilitate full dose coverage of target structures. The standard planning of prostate cancer has included 3D CRT at the University of Michigan and its affiliate, Providence Hospital, since 1987.

CRT has been shown to decrease acute morbidity in the treatment of patients with localized prostate cancer^2,9,15,17. That is, the use of CRT results in a decreased dose to the bladder¹¹ and rectum while providing a complete high-dose distribution to the target volume. This is accompanied by an improvement in acute radiotherapy reactions. More recently there are data which support the hypothesis that CRT improves outcome, as measured by PSA normalization, over conventional techniques¹. In the past it has been difficult to assess the efficacy of prostate cancer treatment because of the long natural history of the disease. The use of PSA as a marker of determining failure following radiation has allowed investigators to assess the efficacy of treatment of prostate cancer relatively early⁵. The current study analyzes 707 patients with localized prostate cancer treated with CRT for biochemical control (bNED) as determined by PSA and overall survival.

Methods and Materials

Patient characteristics
Between January 1987 and June 1994, 707 patients with localized prostate cancer were treated at University of Michigan and Providence Hospital, Department of Radiation Oncolocy, using 3D CRT, and have been reviewed for this study. Patients with pathologically confirmed pelvic lymph node metastasis, treated with pre-irradiation (preRT) androcen ablation, or treated post-prostatectomy were excluded. Table 1 shows the patient characteristics of the entire group. There were 603 patients with T1-T2 disease. PreRT PSA was available for 649 patients with median preRT PSA of 12.9 ng/ml (range 0.2-257.1). Two hundred nine patients had preRT PSA > 20 ng/ml. Median age of the patients was 72 years (range 44-87). One hundred eighty-seven patients did not have a Gleason score. Primarily, these patients were treated at Providence Hospital early in the analysis period when Gleason score was not routinely assigned. All patients were seen in follow-up with PSA determination and digital rectal exam every 3-6 months following treatment. There has been up to 8 years of follow-up with median follow-up time of 36 months. Ten percent of patients have been followed beyond 5.5 years.

Radiotherapy technique
The 3D CRT techniques used at the University of Michigan have been previously described^7,13. Briefly, all patients underwent initial simulation and in ceneral, were immobilized using an alpha cradle foam immobilization device. This was followed by a planning computed tomographic (CT) imaging scan in the treatment position using 5 mm cuts throuch the prostate. The CT data were entered into a 3D planning system and various structures including the prostate, seminal vesicles, bladder, rectum and nodal structures, when nodes were to be treated, were outlined on each slice by the physician. The treatment targets were expanded in 3D to account for patient setup and microscopic tumor uncertainty. Custom cerrobend blocks were designed using BEV and marcins were adjusted to provide full target coverage.

At the University of Michigan a treatment philosophy has been to treat the pelvis and prostate for patients with high risk of nodal metastasis (e.g., Gleason >= 7, Stages T3, T4) to 45 Gy followed by a boost to the prostate and seminal vesicles. All other patients have been treated to small fields encompassing only the prostate and seminal vesicles. Since 1989, the seminal vesicles have been excluded from the final target after a dose of 55 Gy, except in T3-T4 lesions. Field arrangements were planned using BEV. Treatment plans were acceptable if the target volume was encompassed by the 95% isodose surface. Doses in BEV planned fields were prescribed to the 100% isodose surface. Field arrangements for the boost treatment have evolved over time. Initially, a four-field box technique using custom, shaped fields was used for the prostate boost. This is still the beam arrangement being used at Providence Hospital. At the University, there was a shift from the four-field box technique to an axial, six-field technique¹⁶ and more recently to nonaxial, four-field oblique technique⁶.

Treatment was delivered at 1.8-2.0 Gy daily fractions 5 days/week with continuous course radiotherapy. The doses and volumes treated during this study period are given in Table 1. Total doses of radiation ranged from 49-80 Gy (median dose 69 Gy), with 51% receiving >= 69 Gy.

Complications were scored using the RTOG toxicity scale as follows: Grade 1, rectal bleeding, no therapy; Grade 2, rectal bleeding, simple therapy; Grade 3, minor outpatient surgery; Grade 4, major surgery or prolonged hospitalization. Generally, rectal morbidity was manitested by painless hematochezia and, generally, an endoscopic examination was performed to rule out other sources of bleeding.

Statistics

Biochemical failure was defined as 1 ) two consecutive PSA rises over 2.0 ng/ml if nadir PSA <= 2.0 ng/ml, 2) two consecutive rises in PSA over nadir if nadir PSA > 2.0 ng/ml, or 3) initiation of hormonal therapy after RT. The time of PSA failure was documented as the date of the confirmatory PSA rise. The biochemical survival was measured from the date RT ended to the date of PSA failure or last PSA measurement for censored patients. The distribution of bNED survival was estimated non-parametrically by the Kaplan and Meier method. The 95% confidence interval for the Kaplan-Meier estimate at 5 years was also provided. The length of bNED survival was compared between patient groups with the log rank test. These univariate analyses are based on all patients with available data. Multivariate analyses were performed using a Cox regression model. In these models, the independent association of predictors with the rate of chemical failure was evaluated. Initially, the results were estimated for the model including all factors which were significant by the univariate logrank test. A parsimonious model was attained when factors which were not significant by the likelihood ratio test were removed, beginning with the least significant, until only statistically significant factors remained. All two-way interactions were examined among the factors in the parsimonious model. These models were based only on patients with complete information on all factors included in the initial model.

Results

Biochemical Control (bNED)
PreRT PSA were available in 649 of 707 patients. Median time that preRT PSA was obtained was 2.3 months prior to RT. Table 1 summarizes the results of univariate analysis of the patient factors. T-stage (T3-4), Gleason score (7, 8-10), initial PSA value (>10), and use of pelvic RT were all indicators of poor biochemical control at 5 years by the log rank test (Table 1). The estimated proportion of patients chemically NED at 5 years was determined according to 1 ) preRT PSA: PSA <= 4, 88%; PSA > 4 <= 10, 72%; PSA > 10 <= 20, 43%; and PSA > 20, 30% (Fig. 1) and 2) Gleason score: Gleason 2-6, 61%; Gleason 7, 35%, Gleason 8 -10; 34% (Fig. 2). Age, race, and boost technique did not effect biochemical failure. Interestingly, patients who received > 69 Gy did worse than patients who received <= 69 Gy in a univariate analvsis. This is due to the fact that patients with favorable disease received less radiation. The dose of radiation was no longer a factor in biochemical failure in a multivariate analysis.

A multivariate analysis was performed to determine if factors associated with biochemical survival by univariate analysis were also independent indicators of time to biochemical failure. These models are based only on patients which have complete information (473 patients) of all factors included in the model. The univariate factors preRT PSA, T-stage, and Gleason score emerged as significant independent predictors of biochemical failure. The univariate tests were calculated among the subset of 473 patients, which had complete information, and the entire population and were analyzed in the Cox multivariate analysis model. The results were similar in the subset of patients with complete information compared to the analyses using all patients wtih information on a specific characteristic. Further evaluation revealed that there was a statistically significant interaction between T-stage and preRT PSA (Table 2). This model indicates that, after adjusting for Gleason score, the effect of PSA > 10 on time to biochemical failure is of a different magnitude but same direction, depending on the T-stage (Fig. 3). In the presence of a significant interaction, the risk ratios in Table 2 for each factor involved in the interaction should not be interpreted alone. Table 3 reports the estimated risk's ratio for the four groups defined by PreRT PSA and T-stage after adjusting for Gleason score. The risk's ratios in Table 3 are derived from Table 2 by multiplying the risk ratios for each factor. For example, the risk of failure for patients with PreRT PSA >= 10 and T Stage 3 or 4 versus the risk of failure for a patient with a PSA < 10 and T Stage 1 or 2 is (5.3)(6.86)(0.24) = 8.73. The risk of biochemical failure for having a PreRT PSA > 10 is an estimated 5.3 times higher than if the preRT PSA is 1O, if the T Stage is 1 or 2; however, this baseline PSA factor confers only a 1.27 (8.73/6.86) times higher risk, if the T Stage is 3 or 4.

Patients were then divided into two prognostic groups: a favorable group with PSA <= 10 ng/ml, Gleason score < 7, and T1-T2 tumors, and an unfavorable croup with PSA > 10 ng/ml, Gleason score >= 7 or T3-T4 tumors. There was a significant improvement in bNED (p < 0.0001) for patients in the favorable group (Fig. 4). For the favorable group, 75% were bNED at 5 years. Only 8 of 133 failed. For the unfavorable group, 33% were bNED at 5 years.

In addition, a group that might be considered a surgical subset was reviewed: patients < 70 years old with T1-T2 tumors, PSA <= IO, and Gleason score <= 7. There were 85 patients in this group. For the 79 patients without chemical failure, the median follow-up was 27.9 months (range 3.2-93.7 months). This subset had an 84% 5-year bNED rate.

Overall survival
All 707 patients were available for overall survival analysis. Five-year survival for the favorable group was 86%, compared to 79% for the unfavorable group. Patients who were considered a surgical subset had a statistically significant improvement in survival (p = .0110) with an actuarial survival of 98% at 5 years versus 78% for those patients not considered to be surgical candidates.

Toxicity
Complications were graded using the RTOG scale. Complications with the techniques used here are very low: 3% actuarial risk at 7 years of Grade 3-4 rectal complications and 1% actuarial risk at 7 years of Grade 3 bladder complications (no Grade 4).

Discussion

Three dimensional CRT aids in radiotherapeutic treatment planning and dose delivery of prostate cancer bv improving tumor localization. This allows conformal treatment of the target volume while sparing the critical normal structures such as the bladder and the rectum. This in turn will allow increased doses of radiation to the tumor while maintaining low doses to these critical structures. This report is the largest series of patients with prostate cancer to be treated with 3D CRT.

Although randomized trials comparing 3D CRT to conventional radiation therapy have yet to be performed, it appears that treatment technique is an important factor in achieving local control. Corn et al.¹ retrospectively compared CRT to conventional therapy and showed an improvement in biochemical control for patients treated with CRT. There were no differences in dose between patients treated with CRT compared to the conventional group which implies that treatment technique with precise target localization alone is important. In our retrospective review, patients who received >69 Gy had a worse bNED than patients who received <=69 Gy. In ceneral, favorable patients received lower doses of RT, thus dose was not a factor in a multivariate analysis.

It is also important to understand whether increasing the dose to the prostate will improve local control and outcome. We have shown that it is tolerable to give a high dose of radiation to the prostate using CRT with minimal chronic rectal morbidity^12,14 and long term complications in this series were minimal, with 3% actuarial risk of Grade 3 or 4 rectal and Grade 3 bladder (no Grade 4) toxicity. Recent data from Hanks et al.⁴ demonstrate that patients with elevated PSA ( >10 ng/ml) may benefit from dose escalation above 71 and 73 Gy. In our series, in the unfavorable aroup of patients (T3-4 or Gleason score >= 7 or PSA > 10), radiation therapy alone did not provide a high degree of disease control with only 37% bNED at 5 years. Clearly additional therapy is needed in this unfavorable group of patients. Androgen ablation with radiation therapy has been shown to increase local control and disease free survival in these patients with locally advanced prostate cancer^3,10. However, long term surveillance is required to assess effects on overall survival.

In this series, we evaluated a favorable subset of patients which might be compared against a modem prostatectomy series (T1-2, Gleason score <= 7, PSA <= 10, age < 70). Our results show a 75% bNED and 98% survival at 5 years. Although surgical series may be composed of a more highly select group of patients, these radiation results are comparable to surgical series published by Scardino et al.⁸ and Walshet al.¹⁹

It is apparent that preRT PSA is the most critical predictor of success; however, disappointing biochemical outcome results have caused skepticism regarding the efficacy of radiotherapy. It is important; however, how one defines biochemical failure. To date, there is no definite criteria for biochemical failure. In this study, we defined biochemical failure as: 1) two consecutive PSA rises over 2.0, if PSA <= 2.0 ng/ml, 2) two consecutive rises in PSA over nadir, if nadir PSA >= 2.0 ng/ml, or 3) initiation of hormonal therapy after RT. Hanks et al. defines biochemical failure as a confirmed rise in PSA over 1.5 ng/ml¹ whereas Massachusetts General defines failure as two sequential rises in PSA or a PSA > 1 ng/ml two or more years after radiation²⁰. When different criteria of failure are applied, a difference in outcome has been noted²⁰. It is clear that a uniform criteria to evaluate biochemical failure is necessary and there will be consensus conference to determine this criteria to allow better comparison of future data.

3D CRT allows for treatment of prostate cancers with a very low risk of complications. Patients with relatively early disease as defined by preRT PSA and patients who are candidates for radical prostatectomy have excellent 5-year bNED rates. Patients with adverse prognostic factors have a high risk of biochemical recurrence and are candidates for innovative therapy. The use of 3D CRT allows for better localization of tumor while sparing normal structures. and may allow for dose escalation in a select group of patients.

References

1. Corn, B.W.; Hanks. G. E.; Schultheiss, T.E.; Hunt, M.A.; Lee. W.R.; Coia, L.R. Conformal treatment of prostate cancer with improved targeting: Superior prostate-specific antigen response compared to standard treatment. Int. J. Radiat. Oncol. Biol. Phys. 32:325-330; 1995.
2. Epstein, B.E.; Peter. R.; Martin, E.; Hanks, G.E. Low complication rate with conformal radiotherapy for cancer of the prostate. Radiother. Oncol. (Suppl.) 24:S100; 1992.
3. Hanks, G.E.; Hanlon, A.; Schultheiss, T.; Corn, B.; Shipley, W.U. ; Lee, W.R. Early prostate cancer: The national results of radiation treatment from the Patterns of Care and Radiation Therapy Oncology Group studies with prospects for improvement with conformal radiation and adjuvant androgen deprivation. J. Urol. 152:1775-1780; 1994.
4. Hanks. G.E.; Lee. W.R.; Hanlon, A. L.; Kaplan, E.; Epstein, B.; Schultheiss, T. Conformal technique dose escalation in prostate cancer: Improved cancer control with higher doses in patients with pretreatment PSA > 10 ng/ml. (Abstr.) Proc of 37th Annual ASTRO Meeting. 141; 1995.
5. Kuban. D.A.; El-Mahdi, A.M.; Schellhammer, P.F. Prostate-specific antiaen for pretreatment prediction and post-treatment evaluation of outcome after definitive irradiation for prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 32:307-316: 1995.
6. Marsh, L.H.; TenHaken, R.K.; Sandler, H.M. A customized nonaxial external beam technique for treatment of prostate carcinomas. Med. Dosim. 17:123-127; 1992.
7. McShan. D. L.; Fraass, B.A.; Lichter, A.S. Full integration of the beam's eye view concept into computerized treatment planning. Int. J. Radiat. Oncol. Biol. Phys. 18:1485-1494; 1990.
8. Ohori, M.; Wheeler, T.M.; Kattan, M.W.; Goto, Y.; Scardino, P. Proanostic significance of positive surgical margins in radical prostatectoiiiy specimens. J. Urol. 154:1818-1824; 1995.
9. Pickett, B.; Roach, M. 3rd; Verhey, L.; Horine, P.; Malfatti, C.; Akazawa, C.; Dea, D.; Varad, B.; Rathbun, C.; Phillips, T.L. The value of nonuniform margins for six-field conformal irradiation of localized prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 32:211-8; 1995.
10. Pilepich, M.V.; Krall, J.M.; al-Sarraf, M.; John, M. J. ; Doggett, R.L.; Sause, W.T.; Lawton, C.A.; Abrams. R.A.; Rotman, M.; Rubin, P. Androgen deprivation with radiation therapy compared with radiation therapy alone for locally advanced prostatic carcinoma: A randomized comparative trial of the Radiation Therapy Oncology Group. Urolouy. 45:616-23; 1995.
11. Pollack, A.; Zagars, G.K.; Starkschall, G.; Childress, C.H.; Kopplin, S.; Boyer, A.L.; Rosen, I.I. Conventional vs. conformal radiotherapy for prostate cancer: Preliminary results of dosimetry and acute toxicity. Int. J. Radiat. Oncol. Biol. Phys. 34:555-564: 1996.
12. Sandier, H.M.; McLaughlin, P.W.; TenHaken, R.K.; Addison, H.; Fonnan, J.; Lichter, A. Three dimensional conformal radiotherapy for the treatment of prostate cancer: Low risk of chronic rectal morbidity observed in a large series of patients. Int. J. Radiat. Oncol. Biol. Phys. 33:797-801; 1995.
13. Sandier, H.M.; McShan, D.L.; Lichter, A.S. Potential improvement in the results of irradiation for prostate carcinoma using improved dose distribution. Int. J. Radiat. Oncol. Biol. Phys. 22:361-367; 1992.
14. Sandier, H.M.; Perez-Tamayo, C.; TenHaken, R.K.; Lichter, A.S. Dose escalation for stage C (T3) prostate cancer: Minimal rectal toxicity observed using conformal therapy. Radiat. Oncol. 23:53-54; 1992.
15. Soffen, E.M.; Hanks, G.E.; Hunt, M.A.; Epstein, B.E. Conformal static field radiation therapy treatment of early prostate cancer vs. nonconformal techniques: A reduction in acute morbidity. Int. J. Radiat. Oncol. Biol. Phys. 24:485-488; 1992.
16. TenHaken, R.K.; Perez-Tamayo, D.; Tesser, R.J.; McShan, D.L.; Fraass, B.A.; Lichter, A.S. Boost treatment of the prostate using shaped, fixed fields. Int. J. Radiat. Oncol. Biol Phys. 16:193-200; 1989.
17. Vijayakumar, S.; Awan, A.; Karrison, T.; Culbert, H.; Chan, S.; Kolker, J.; Low, N.; Halpern, H.; Rubin, S.; Chen, G.T.; Weichselbaum, R.R. Acute toxicity during external-beam radiotherapy for localized prostate cancer: Comparison of different techniques. Int. J. Radiat. Oncol. Biol. Phys. 25:359-371; 1993.
18. Vijayakumar, S.; Chen, G. T. Implementafion of three dimensional conformal radiation therapy: Prospects, opportunities, and challenges. Int. J. Radiat. Oncol. Biol. Phvs. 33:979-983; 1995.
19. Walsh, P.C.; Parfin, A.W.; Epstein, J.I. Cancer control and quality of life following anatomical radical retropubic prostatectomy: Results at 10 years. J. Urol. 152:1831-1836; 1994.
20. Zeitman, A.L.; Coen, J.J.; Dallow, K.C.; Shipley, W.U. The treatment of prostate cancer by conventional radiation therapy: An analvsis of long-term outcome. Int. J. Radiat. Oncol. Biol. Phys. 32:287-292; 1995.

Int. J. Radiation Oncology, Vol 38, #2, Jan 1997