CDX2 expression as a predictive and prognostic biomarker of 5-FU response in cancer of unknown primary

Cancer of unknown primary (CUP) remains a diagnostic and therapeutic challenge with limited treatment guidance. Caudal type homeobox 2 (CDX2), a transcription factor indicative of intestinal differentiation, may define a distinct subset of CUP with potential sensitivity to 5-fluorouracil (5-FU)-based chemotherapy. This study examines outcomes in patients with CDX2-positive CUP and the predictive value of CDX2 expression for response to 5-FU-based regimens.

We analyzed data from CDX2-positive CUP patients in the Mayo Clinic Cancer of Unknown Primary registry. All patients underwent comprehensive diagnostic evaluations and received first-line systemic therapy, categorized as 5-FU-based or non-5-FU-based. Primary endpoints were objective response rate (ORR) and clinical benefit rate; secondary endpoints included progression-free survival at first progression (PFS1) and overall survival. Logistic and Cox regression models were used to identify predictors of ORR and survival.

A total of 209 CDX2-positive CUP patients met inclusion criteria. The median age was 64 years (range 18-89 years), with 115 (55%) female patients. Adenocarcinoma was the predominant histology in 147 patients (74.7%), while only 39 (18.7%) exhibited the classical colorectal-like immunoprofile (CK7-negative/CK20-positive/CDX2-positive). Most patients (203, 97.1%) had multiple metastatic sites, and 98 (46.9%) received 5-FU-based chemotherapy regimens. 5-FU-Based treatment was associated with a significantly higher ORR compared with non-5-FU regimens [68 of 98 (69.4%) versus 53 of 111 (47.7%), P = 0.002], as well as higher clinical benefit rate [84 of 98 (85.7%) versus 80 of 111 (72.1%), P = 0.017]. Multivariable logistic regression identified 5-FU use as the only independent predictor of response (OR 2.28, 95% confidence interval 1.23-4.23, P = 0.009). Median overall survival was longer in the 5-FU group compared with non-5-FU (21.0 versus 14.0 months, P = 0.008), as was median PFS1 (15.2 versus 6.3 months, P < 0.001). Multivariable Cox regression confirmed 5-FU therapy as an independent prognostic factor for longer PFS1 (hazard ratio 0.50, 95% confidence interval 0.31-0.80, P = 0.004).

CDX2-positive CUP represents a distinct subgroup with favorable response to 5-FU-based therapy, supporting CDX2 as a predictive biomarker for treatment selection.