Dose escalation with 3D conformal treatment: five year outcomes, treatment optimization, and future directions.

PURPOSE: To report the 5-year outcomes of dose escalation with 3D conformal treatment (3DCRT) of prostate cancer. METHODS AND MATERIALS: Two hundred thirty-two consecutive patients were treated with 3DCRT alone between 6/89 and 10/92 with ICRU reporting point dose that increased from 63 to 79 Gy. The median follow-up was 60 months, and any patient free of clinical or biochemical evidence of disease was termed bNED. Biochemical failure was defined as prostate-specific antigen (PSA) rising on two consecutive recordings and exceeding 1.5 ng/ml. Morbidity was reported by the Radiation Therapy Oncology Group (RTOG) scale, the Late Effects Normal Tissue (LENT) scale, and a Fox Chase modification of the latter (FC-LENT). All patients were treated with a four-field technique with a 1 cm clinical target volume (CTV) to planning target volume (PTV) margin to the prostate or prostate boost; the CTV and gross tumor volume (GTV) were the same. Actuarial rates of outcome were calculated by Kaplan-Meier and cumulative incidence methods and compared using the log rank and Gray's test statistic, respectively. Cox regression models were used to establish prognostic factors predictive of the various measures of outcome. Five-year Kaplan-Meier bNED rates were utilized by dose group to estimate logit response models for bNED and late morbidity.
RESULTS: PSA <10 ng/ml: No dose response was demonstrated using estimated bNED rates or by analysis of PSA nadir vs. dose. PSA 10-19.9 ng/ml: A bNED dose response was demonstrated (p = 0.02) using the log rank test. The logit response model showed 5-year bNED rates of 35% at 70 Gy and 75% at 76 Gy (p = 0.0049) and illustrated the relative ineffectiveness of conventional dose treatment. PSA 20+ ng/ml: A bNED dose response was demonstrated (p = 0.02) using the log rank test. The logit response model indicated a 5-year bNED rate of 10% at 70 Gy and 32% at 76 Gy (p = 0.10). Morbidity: Dose response was demonstrated for FC-LENT grade 2 and grade 3,4 GI morbidity and for LENT grade 2 GU sequelae. RTOG grade 3,4 GI morbidity at 5 years was <1%. Factors associated with bNED, cause-specific survival, and metastasis were studied using Cox multivariate analysis. Pretreatment PSA (p = 0.0001), Gleason score 7-10 (p = 0.0001), and dose (p = 0.017) were significantly predictive of bNED. For each 1 Gy increase in dose, the hazard of bNED failure decreased by 8%. Palpation stage was associated with cause-specific survival (p = 0.002) and distant metastasis (p = 0.0004). Gleason score was also predictive of distant metastasis (p = 0.02).
CONCLUSIONS: A dose response was observed for patients with pretreatment PSA >10 ng/ml based on 5-year bNED results. No dose response was observed for patients with pretreatment PSA < 10 ng/ml. Dose response was observed for FC-LENT grade 2 and grade 3,4 GI sequelae and for LENT grade 2 GU sequelae. Optimization of treatment was made possible by the results in this report. The improvement in 5-year bNED rates for patients with PSA levels > 10 ng/ml strongly suggests that clinical trials employing radiation should investigate the use of 3DCRT and prostate doses of 76-80 Gy.