Differential cyclooxygenase expression levels and survival associations in type I and type II ovarian tumors

Table 3 Associations between COX expression levels and ovarian cancer survival, the Vanderbilt TROC

	Disease-free survival (DFS), hazard ratio (95% CI), P-value						Overall survival (OS), hazard ratio (95% CI), P-value
	IHC¹		H-Score²		QPCR³		IHC¹		H-Score²		QPCR³
COX-1 / PTGS1
N cases / events^a	47 / 15 vs 71 / 42		60 / 26 vs 58 / 31		15 / 12 vs 17 / 13		77 / 43 vs 113 / 77		95 / 57 vs 95 / 63		24 / 19 vs 25 / 18
Unadjusted	2.34 (1.30–4.24)	0.005	1.49 (0.88–2.50)	0.138	1.77 (0.79–3.98)	0.168	1.13 (0.78–1.65)	0.515	1.09 (0.76–1.56)	0.647	1.17 (0.61–2.23)	0.641
Adjusted^b	1.58 (0.85–2.94)	0.148	1.44 (0.85–2.43)	0.180	4.12 (1.51–11.26)	0.006	0.79 (0.54–1.16)	0.227	0.97 (0.68–1.40)	0.882	2.02 (0.96–4.22)	0.063
Adjusted^c	1.55 (0.81–2.98)	0.187	1.30 (0.75–2.34)	0.356	6.37 (1.84–22.01)	0.003	0.87 (0.59–1.30)	0.495	1.14 (0.78–1.65)	0.504	2.26 (1.04–4.91)	0.039
COX-2 / PTGS2
N cases / events	50 / 33 vs 68 / 24		46 / 29 vs 72 / 28		10 / 10 vs 14 / 9		103 / 80 vs 87 / 40		95 / 72 vs 95 / 48		18 / 16 vs 18 / 11
Unadjusted	0.51 (0.30–0.87)	0.013	0.61 (0.36–1.02)	0.061	0.51 (0.20–1.29)	0.154	0.52 (0.36–0.77)	< 0.001	0.57 (0.39–0.82)	0.002	0.59 (0.27–1.27)	0.173
Adjusted^b	1.53 (0.85–2.76)	0.160	1.93 (1.08–3.46)	0.028	0.86 (0.30–2.45)	0.772	0.85 (0.57–1.27)	0.420	0.89 (0.61–1.31)	0.559	0.77 (0.35–1.71)	0.519
Adjusted^c	1.54 (0.84–2.80)	0.161	1.92 (1.06–3.49)	0.033	0.87 (0.31–2.47)	0.798	1.05 (0.69–1.59)	0.818	1.05 (0.71–1.56)	0.802	1.40 (0.56–3.51)	0.477

¹COX protein measured by immunohistochemistry: moderate & strong vs weak
²COX protein measured by H-Score: dichotomized at median
³COX gene expression measured by quantitative real-time PCR: dichotomized at median
^aN cases / N events, reference vs high COX
^bAdjusted for age at diagnosis (continuous), stage (early, late), and tumor type (I, II)
^cAdditionally adjusted for debulking (optimal, suboptimal, missing) and platinum sensitivity (sensitive, resistant, missing) Bold values denote signficant associations; column percentages may not sum to 100% due to rounding error

ISSN: 1757-2215