Five Year Results of US Intergroup/RTOG 9704 with Postoperative Ca 19-9 #90 and Comparison to the Conko-001 Trial

Reviewer: Lara Bonner Millar, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 3, 2010

Authors: A. C. Berger1,K. Winter2,J. P. Hoffman3,W. F. Regine4,R. A. Abrams5,H. Safran6,G. M. Freedman3,A. B. Benson7,J. MacDonald8,C. Willet9
1 Thomas Jefferson University, Philadelphia, PA
2 Radiation Therapy Oncology Group, Philadelphia, PA
3 Fox Chase Cancer Center, Philadelphia, PA
4 University of Maryland School of Medicine, Baltimore, MD
5 Rush University Medical Center, Chicago, IL
6 The Miriam Hospital, Providence, RI
7 Northwestern Memorial Hospital, Chicago, IL
8 St. Vincent's Comprehensive Cancer Center, New York, NY
9 Duke University Medical Center, Durham, NC


  • RTOG 9704 was the largest randomized trial utilizing adjuvant chemoradiation (CRT) therapy for patients with resected pancreatic cancer (PC). In this trial, all patients received 5-FU CRT with adjuvant 5-FU vs. Gemcitabine (Gem). An RTOG 9704 secondary endpoint included the prospective evaluation of the ability of post-resection CA 19-9 to predict overall survival.
  • CONKO-001(Oettle H, et al. JAMA. 2007;297(3):267-77) evaluated adjuvant gemcitabine alone vs observation after resected pancreatic cancer. The potential benefit of radiation therapy in the adjuvant setting was not addressed by the CONKO-001 trial, since neither arm included radiation therapy. CONKO found that adjuvant Gemcitabine resulted in improved disease free survival but no change in overall survival among patients with pancreatic adenocarcinoma. Results of CONKO were not available at the time RTOG 97-04 was accruing.
  • The CONKO-001 trial required post-resection CA 19-9 levels to be less than 2.5 times the upper limit of normal (approximately 90 U/ml), while RTOG did not have an exclusion criteria based on CA 19-9.
  • In this study, the authors analyzed the 5-year survival results of RTOG 97-04 according to CA 19-9 ?90 vs. >90 and compared it to 5-year results reported for the CONKO-001 trial (adjuvant Gemzar alone).

Materials and Methods

  • CA 19-9 expression was analyzed as a dichotomized variable (?90 vs. >90), based on the exclusion criteria of the CONKO trial. Lewis antigen negative patients (who do not synthesize CA 19-9, about 5-10% of the population) were grouped with ?90.
  • Cox proportional hazard models were utilized to identify the impact of CA 19-9 value on overall survival (OS). The following variables were included in the multivariate analyses:
    • Treatment, tumor stage, nodal involvement, tumor diameter, margin status, and RT quality assurance (QA) score (per protocol vs. not per protocol).
  • Actuarial estimates of OS were calculated using the Kaplan-Meier method.


  • 538 patients accrued to the trial with 385 having analyzable CA 19-9.
    • 332 (86%) had baseline CA 19-9 ?90.
    • 53 (14%) had values >90.
  • Both univariate (HR = 3.2 (95% CI = 2.32-4.29) p < 0.0001) and multivariate (HR = 3.1 (95% CI = 2.24-4.17) p<0.0001) analyses demonstrated a highly statistically significant decrease in OS for CA 19-9 >90.
  • For patients treated on the adjuvant Gem arm with a CA 19-9 ?90, the median survival was 21 months with 5-year survival of 24% (95% CI = 17%-30%).
    • For patients with CA 19-9 >90, these numbers dropped dramatically to 10 months and 4% (95% CI = 0.3%-18%).
    • In patients on the Gem arm with prospective RTQA per protocol, median and 5-year survivals were 23 months and 32% (95% CI = 22%-43%) for those with CA 19-9 ? 90.
  • In comparison, the median and 5-year OS for patients in the Gem arm of CONKO were 22.8 months and 21% .
    • Of note, 35% of patients with CA 19-9 ? 90 in the RTOG trial had a known R1 resection. This is twice the number seen in the CONKO trial. Margin status was unknown in 25% on RTOG 97-04.
  • Finally, for patients with pancreatic head tumors with CA 19-9 ?90, and treated with Gem (n = 145), the median and 5 year survivals were 22 months and 25% (95% CI = 18%-32%), respectively.

Author's Conclusions

  • This analysis demonstrates that patients with post-resection CA 19-9 values >90 had a significantly worse survival.
  • All patients and those with pancreatic head tumors treated with Gem with CA 19-9 ?90 and per protocol RT have favorable survival compared to that seen on the CONKO trial despite having twice as many R1 resections in RTOG 97-04 compared to CONKO.
  • CA 19-9 is a stratification factor for the current RTOG adjuvant pancreas trial (0848), a phase III trial evaluating both erlotinib and chemoradiation as adjuvant treatment for patients with resected head of pancreas adenocarcinoma.

Clinical Implications

  • CA 19-9 is an important prognostic factor, with post-resection values >90 U/ml correlating with significantly worse survival. Elevated CA 19-9 after resection is a marker of either residual disease in the pancreatic bed and nodal region or subclinical metastatic disease.
  • The RTOG 97-04 and CONKO-001 trials evaluated gemcitabine-based treatment in the adjuvant setting. Unfortunately, the results of RTOG 97-04 cannot be directly compared with the CONKO-001 results because of fundamental differences in treatment design and patient characteristics.
    • CONKO-001 had more favorable patients than RTOG 97-04, with 83% of patients undergoing R0 resections (vs 40%) and all patients having a CA19-9 < 2.5 times the upper limit of normal (vs no upper limit).
    • Nevertheless, when comparing outcomes between the two studies according to a CA 19-9 cut-off of 90, the patients on RTOG fared similarly to those on CONKO, despite a lower R0 resection rate. This indicates that radiation therapy should not be dismissed in adjuvant treatment of pancreatic cancer.
  • In practice, there is variation in radiation treatment fields for pancreatic cancer in terms of dose and decision to treat elective nodal regions and margin size on the primary tumor. The group on RTOG that underwent RTQA had slightly improved median and 5 year OS. Future adjuvant therapy trials for patients with pancreatic adenocarcinoma should stratify patients according to post-resection CA 19-9 levels and should include RTQA with well defined treatment parameters.