The site of visceral metastases (mets) to predict overall survival (OS) in castration-resistant prostate cancer (CRPC) patients (pts): A meta-analysis of five phase III trials.

Genitourinary (Prostate) Cancer
Session Type and Session Title: 
Oral Abstract Session, Genitourinary (Prostate) Cancer
Abstract Number: 
J Clin Oncol 32:5s, 2014 (suppl; abstr 5002)
Susan Halabi, William Kevin Kelly, Haojin Zhou, Andrew J. Armstrong, David Quinn, Karim Fizazi, Nicole C Solomon, Ian Tannock, Daniel Peter Petrylak, Michael J. Morris, Eric Jay Small; Duke University Medical Center, Durham, NC; Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA; Duke University, Durham, NC; University of Southern California, Los Angeles, CA; Department of Cancer Medicine, Gustave Roussy, University of Paris Sud, Cancer Campus, Grand Paris, Villejuif, France; Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada; Yale University Medical Center, New Haven, CT; Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY; University of California, San Francisco, San Francisco, CA

Abstracts that were granted an exception in accordance with ASCO's Conflict of Interest Policy are designated with a caret symbol (^).

Abstract Disclosures


Background: Recent studies have shown that metastatic site is an important predictor of overall survival, but the reports were based on small sample sizes and as such the estimates were unstable. We sought to test two hypotheses: 1) CRPC pts with lung mets have worse survival than pts with non-visceral mets; and 2) pts with liver mets have worse survival than pts with lung mets. Methods: We combined individual patient data from 3,993 chemotherapy-naive mets CRPC pts randomized to receive docetaxel (D) based therapy on 5 phase III trials: CALGB 90401 (D +/- Bevacizumab): SWOG 0421(D +/-atrasentan),ENTHUSE 33 (D +/- zibotentan), TAX327 (D 3 wks, D weekly) and SWOG 9916 (D + estramustine). Site of mets at baseline was categorized as: lymph node (LN) only, bone +/- LN with no visceral mets, any lung mets (but no liver), any liver mets, and other visceral. We used fixed-effects meta-analysis to estimate the pooled hazard ratios (pHR) and 95% confidence intervals (CI) for comparing pts with lung mets vs. non-visceral mets and liver mets vs. lung mets. Results: The pHR for death for pts with lung mets compared to pts with non-visceral mets was 1.3 (95% CI= 1.1-1.5, p<0.001) and the pHR for pts with any liver mets compared with pts with lung mets was 1.4 (1.2-1.7, p<0.001). The median overall survival by site of mets is presented in table below. f mets at baseline was categorized as: lymph node (LN) only, bone +/- LN with no visceral mets, any lung mets (but no liver), any liver mets, and other visceral. Conclusions: As anticipated, CRPC patients with liver mets had the worst OS (12.1 m). While pts with lung mets had better OS (16.5 months) compared to liver mets pts, they had significantly worse survival than pts with non-visceral bone mets (20 months). These data may help in treatment decisions and in the design of future clinical trials in mCRPC pts. Clinical trial information: NCT00110214.

Metastatic Site N (%) Median OS in months (95% CI)
LN only 187 (5) 27.0 (24.7-32.4)
Bone/Bone with LN 3334 (83) 20.3 (19.7-21.0)
Lung (+/- bone), no liver 300 (7) 16.5 (14.8-18.4)
Liver (LM) (+/- bone) 280 (7) 12.1 (10.1-13.5)
Other Visceral (adrenal, brain) 55 (1) 14.4 (12.6-19.1)