Evaluating the impact of lead-time bias on the observed efficacy of early salvage radiation therapy in prostate cancer: A post-hoc analysis of the RTOG 9601 trial
Abdollah F, Dalela D, Sood A, Arora S, Tang HJ, Keeley J, Alanee S, Rogers CG, Peabody JO, and Menon M. Evaluating the impact of lead-time bias on the observed efficacy of early salvage radiation therapy in prostate cancer: A post-hoc analysis of the RTOG 9601 trial. Eur Urol Suppl 2019; 18(1):e861.
Eur Urol Suppl
Introduction & Objectives: To evaluate the potential impact of lead-time bias on the reported beneficial role of early salvage radiotherapy (sRT), defined as delivering radiation at a relatively low pre-sRT prostate-specific antigen (PSA)value, in treating prostate cancer patients with biochemical recurrence after radical prostatectomy (RP). Materials & Methods: All available demographic, tumor-specific, and overall survival data from 760 men who participated in the RTOG 9601 trial were extracted using the Project Data Sphere platform. Patients were stratified based on pre-sRT PSA (<0.7 [n=405], 0.7-1.5 [n=237], and >1.5-4.0 ng/mL [n=118])as reported in the original trial. Cox regression analysis assessed the impact of pre-sRT PSA on overall mortality, after controlling for covariates. To ascertain the role of lead-time bias, survival time zero was set to the time of (1)initiation of sRT, and (2)RP. Results: There were no statistically significant differences amongst men within the three groups, except for a greater proportions of African-American patients and men with post-RP PSA nadirs 30.5 ng/mL within the higher pre-sRT PSA groups. For men with pre-sRT PSA <0.7, 0.7-1.5, and >1.5-4 ng/mL, estimated 15-year overall mortality was 39%, 45%, and 50%, respectively (p=0.005)when calculated from time since sRT, and 23%, 29%, and 36% respectively (p=0.08)when assessed from time since RP. On multivariable regression analyses, pre-sRT PSA >1.5-4.0 ng/mL was significantly associated with overall mortality only when measured from time of initiation of sRT (HR 1.61, 95% confidence interval [CI]1.13-2.28; p=0.008), but not from time since RP (HR 1.24, 95% CI 0.87-1.76; p=0.2). Conclusions: Our findings suggest that the putative survival benefit with early sRT instituted at lower PSA thresholds is mitigated when measured from time since surgery, highlighting the role of lead-time bias. These findings need further validation given their significant clinical implications.