Preliminary Results of a Phase II Study of Cyclooxygenase-2 (Cox-2) Inhibition in Rectal Cancer

Reviewer: Neha Vapiwala, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 5, 2004

Presenter: B. Chakravarthy
Presenter's Affiliation: Vanderbilt University
Type of Session: Scientific


Manipulation of arachidonic acid metabolism via the cyclooxygenase-2 (COX-2) pathway is an active area of oncology research. Studies have established that COX-2 leads to tumor growth via prostaglandin E2 (PGE2) production and vascular endothelial growth factor (VEGF) induction. Studies have also demonstrated that selective blockade of COX-2 can improve the anti-tumor efficacy of radiation therapy and chemotherapy.  Furthermore, the use of COX-2 inhibitors has been found to reduce colon polyp formation in patients with familial adenosis polyposis (FAP). Stemming from this pathophysiologic observation with colon adenoma cells, the investigators of this trial sought to explore the efficacy of COX-2 inhibition in patients with clinically resectable rectal cancer. The COX-2 inhibitor was celecoxib, and it was tested alone as well as in combination with chemoradiation. 

Materials and Methods

  • Phase II trial
  • Eligible pts had histologically-positive primary rectal adenocarcinoma that was considered surgically resectable 
  • No history of NSAID, sulfonamide,  or celecoxib allergy was allowed
  • 27 pts enrolled
  • Treatment was as follows: 
    • Stage I (8 pts): celecoxib 400 mg po bid x 5 days --> surgery
    • Stage II/III (19 pts): celecoxib bid x 5 days --> preoperative chemoradiation with continuous 7-day infusional 5-FU @ 225 mg/m2/day + 5040 cGy radiation in 180 cGy fractions --> surgery
  • Tumor and surrounding tissue specimens were obtained from stage I pts a) pre-tx and b) post-5 days of celecoxib
  • Tumor and surrounding tissue specimens were obtained from stage II/III pts a) pre-tx, b) post-5 days of celecoxib, c) post-7 days of 5-FU/radiation, and d) at surgery
  • Samples were used to measure microarray gene expression, mass spectroscopy, COX-2 and CD-31 immunostaining, serum VEGF levels, and urinary PGE2 metabolite (PGE-M) levels    


  • All 8 stage I pts completed the planned tx.
  • Of the 19 stage II/III pts, at the time of this report, there was:
    • one pt with grade 2 hypokalemia, nausea/dehydration,
    • one pt with grade 2 thrombocytopenia,
    • one pt with small bowel perforation requiring surgery during the preoperative chemoradiotherapy, 
    • two pts still receiving the preoperative tx and not yet gone to surgery,
    • and finally two pts who suffered unexpected deaths
  • Of the 12 pts who have undergone resection to date, there were 3 pathologic complete responses (25%).
  • The use of celcecoxib was associated with a significant decrease in urine PGE-M levels. 
  • There was also a decrease in serum VEGF levels with celecoxib use in those pts who had excpetionally high pre-tx VEGF levels.

Author's Conclusions

  • The use of celecoxib with concurrent chemoradiation in the preoperative setting in rectal cancer pts is safe and tolerable.
  • The use of short-term celecoxib is associated with promising reponse rates, including complete pathologic responses.
  • The correlation of decreased urinary PGE-M levels to favorable pathologic responses with celecoxib therapy  suggests a possible role for PGE-M as a predictive indicator.
  • The establishment of useful non-invasive biomarkers for response to COX-2 inhibitor therapy among cancer patients deserves further investigation.

Clinical/Scientific Implications
Attempts at optimizing cancer treatments continue to investigate and incorporate "targeted agents" as part of a multimodality approach. These agents focus on various pathways with known involvement in tumor growth and proliferation. Selective inhibition of the COX-2 enzyme process is an example of this, and this important phase II study from Vanderbilt University offers promising data on the efficacy of COX-2 inhibition in achieving meaningful pathologic tumor responses for resectable rectal cancer patients. Furthermore, the data shown here introduce a new, reliable, and easily attainable biomarker, such as urinary levels of a COX-2 pathway metabolite, to help follow and predict a given patient's response to treatment. Whether or not such a marker could actually help avoid surgery altogether in that subset of patients who are predicted to have a pathological complete response based on the biomarker remains hypothetical at this time. However, the potential use of biomarkers to custom-tailor treatment for certain patients (ie: those who might not be able to tolerate an extensive surgical procedure) makes further study of this arena very important. and worthwhile.

From the National Cancer Institute