Receiver operating curves and dose-volume analysis of late toxicity with stereotactic body radiation therapy for prostate cancer.

PURPOSE: The purpose of this study was to evaluate a receiver operating characteristic (ROC) curve method to determine dose thresholds with late genitourinary (GU) toxicity after stereotactic body radiation therapy for prostate cancer.

METHODS AND MATERIALS: Seventy-eight patients diagnosed with low- to intermediate-risk prostate cancer and treated with stereotactic body radiation therapy alone were reviewed retrospectively. All patients received a total dose of 38 Gy in 4 fractions with a planning target volume expansion of 2 mm. GU toxicity was documented according to the Common Terminology Criteria for Adverse Events, version 4. ROC analysis applied on a logistic regression model was used to determine optimal dosimetric parameters for GU toxicity.

RESULTS: The median age at treatment was 69 years with a median prostate volume of 46.2 mL. The median prescription isodose line was 67% (interquartile range, 65, 70). The median clinical follow-up was 35.49 months. Late grade 1, 2, and 3 GU toxicity occurred in 21.8%, 19.2%, and 2.6% of cases, respectively. Late grade 2+ GU toxicity was associated with prescription to isodose line (P = .009) and normalized volumes for heterogeneity ≥46 Gy. The ROC method successfully produced thresholds for dose-volume recommendations for both prostate and urethra, including normalized prostate volumes from 46 to 50 Gy, such as volume of target tissue receiving 46% of the prescribed dose (V46) Gy of 36.7% (sensitivity, 71%; specificity, 61%; area under the curve, 0.67) with an associated probability of late GU grade 2+ toxicity of 21%.

CONCLUSIONS: Intraprostatic heterogeneity should be controlled with potential thresholds at V46 Gy <36.7%, V48 Gy <21%, and V50 Gy <9.5% of the normalized prostate volume to keep late grade 2+ GU toxicity ≤20% with 4-fraction schemes. This may be facilitated with a higher prescription isodose line (>69%).