Given these variables, two main questions come to mind in envisioning a follow-up study (such as a prospective RCT or a retrospective analysis of an RCT) to confirm the role of CDX2 as a prognostic biomarker in stage II disease.
First, how much sub-stratification based on tumor characteristics (i.e. other molecular markers and clinicopathologic features) is necessary? How much sub-stratification is likely feasible?
Second, how much sub-stratification based on the adjuvant chemotherapy regimen is necessary? Is surgery alone vs. adjuvant chemotherapy an appropriate comparison or should surgery alone be compared to a specific chemotherapy regimen? How much sub-stratification is likely feasible or ethical?
Essentially, in a randomized study to examine the prognostic implications of a CDX2-negative tumor, how would you negotiate the trade-off between increased applicability and decreased sample size?