At the moment I am investigating the relation between a number of lymph nodes dissected during esophagectomy. Critics say the relation is spurious because of stage migrational effects when positive lymph nodes in the extended fields change change the N-stage towards a higher category. Will correction for stage in the multivariate analysis then introduce a bias?
Very interesting question Dr Talsma,
I think we need to see how much of a real difference stage migration results in in the current settings of high quality surgery and diagnostic/staging work-up.
When we look at treatment and follow-up using high quality diagnostic and surveillance tools and compare it to historical controls, we definitely see a very significant difference.
Similarly when we compare two regions with different patient sets and different standards of care, there is very likely to be a significant difference (case in point D2 gastrectomy: Japanese vs Western data).
However, if we want to look at how stage migration affects our current set of patients where standard diagnostic and therapeutic modalities are used, we will have to identify a group of patients we want to compare against to see the difference..
An important philosophical question is which stage we want to look at as our reference, the lower stage allocated using lesser information or the higher stage that we obtain after better scrutinization. i tend to incline towards more information but this is my personal opinion only.
regards, raza
Very difficult to make allowances for this in analyses. The literature is conflicting on the subject. There are so many confounders, initial stage, chemo / radiotherapy reaponse and surgica / pathological factors