Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA.
Cancer. 2012 Feb 1;118(3):839-47. doi: 10.1002/cncr.26307. Epub 2011 Jul 12.
During the first 3 years after prostate cancer treatment with radiation therapy, benign prostate-specific antigen (PSA) bounces are difficult for clinicians to distinguish from a biochemical recurrence, which can result in unnecessary interventions and erroneous predictions of outcomes. The objective of this study was to evaluate a commonly used PSA failure definition in a multinational, multi-institutional study after monotherapy with prostate brachytherapy.
Participants were selected from 2919 men who underwent permanent prostate brachytherapy at the University Medical Center Utrecht, Princess Margaret Hospital, or Seattle Prostate Institute between 1998 and 2006. Inclusion required not having received androgen-deprivation therapy and having at least 30 months of follow-up. Failure was defined as any post-treatment use of hormone therapy, clinical relapse, or prostogram-defined biochemical (PSA) failure. Cases in which the nomogram predicted biochemical failure were evaluated at each institution to verify biochemical status over time and the actual clinical outcome at 5 years.
The median follow-up for the 1816 patients was 5.2 years. Concordance between the prostogram-predicted and actual outcomes, as measured by the Harrell c statistic, was 0.655 (95% confidence interval [CI], 0.536-0.774; P = .010) for the Princess Margaret group, 0.493 (95% CI, 0.259-0.648; P = .955) for the Seattle group, and 0.696 (95% CI, 0.648-0.744, P < .001) for the Utrecht group. The overall mean difference in biochemical recurrence-free survival at 5 years between actual outcomes and prostogram-defined outcomes was 9.2% (95% CI, 7.7%-10.6%). The total numbers of prostogram-defined and actual biochemical failures were 312 and 157, respectively (P = .001).
The widely used prostogram could not adequately distinguish a benign PSA bounce from a biochemical recurrence after prostate brachytherapy and could not be used to counsel patients about their predicted outcomes after treatment. The authors conclude that, to avoid unnecessary active interventions after treatment, clinicians should monitor PSA levels for at least 3 years and provide reassurance to patients that a PSA rise during this time is common and may not indicate a treatment failure.
在前列腺癌放射治疗后的头 3 年，良性前列腺特异性抗原（PSA）反弹很难被临床医生与生化复发区分开来，这可能导致不必要的干预和错误的预后预测。本研究的目的是在前列腺近距离放射治疗的单药治疗后，在一项多国家、多机构的研究中评估一种常用的 PSA 失败定义。
参与者选自 1998 年至 2006 年期间在乌得勒支大学医学中心、玛格丽特公主医院或西雅图前列腺研究所接受永久性前列腺近距离放射治疗的 2919 名男性。纳入标准为未接受雄激素剥夺治疗且随访时间至少 30 个月。失败定义为任何治疗后使用激素治疗、临床复发或前列腺程序定义的生化（PSA）失败。在每个机构，对预测为生化失败的诺莫图病例进行评估，以验证随时间推移的生化状态和 5 年的实际临床结果。
1816 例患者的中位随访时间为 5.2 年。乌得勒支组、玛格丽特公主组和西雅图组 Harrell c 统计的诺莫图预测结果与实际结果的一致性分别为 0.655（95%置信区间，0.536-0.774；P =.010）、0.493（95%置信区间，0.259-0.648；P =.955）和 0.696（95%置信区间，0.648-0.744，P <.001）。5 年时，实际结果与诺莫图定义结果的生化无复发生存率的平均差异为 9.2%（95%置信区间，7.7%-10.6%）。诺莫图定义的生化失败和实际生化失败的总数分别为 312 例和 157 例（P =.001）。
广泛使用的诺莫图不能充分区分前列腺近距离放射治疗后的良性 PSA 反弹和生化复发，也不能用于告知患者治疗后的预测结果。作者得出结论，为避免治疗后不必要的积极干预，临床医生应至少监测 3 年 PSA 水平，并向患者保证在此期间 PSA 升高很常见，并不一定表示治疗失败。
