Biomedical Engineering Reference
In-Depth Information
to patients sooner. Hence, in such settings, PFS is often used as a primary
endpoint if it is known to be a surrogate endpoint for OS.
On the other hand, the use of PFS is still debatable due to its own short-
comings by nature (Tuma (2009) and Dodd et al. (2008)). In solid tumor
oncology studies, response is usually assessed by CT/MRI scans at prespeci-
fied intervals. The date of disease progression is not the exact date on which
a patient progressed because the progression date is determined based on pe-
riodical assessments of scans due to disease latency. Hence, the exact date of
true disease progression is never known, and the date of progression is usually
the date of first documented progression after true progression, as shown in
Figure 10.1 (Bhattacharya et al. (2009)). From this point of view, PFS should
be viewed as interval-censored time-to-event data.
FIGURE 10.1: Comparison of event determination in OS and PFS analyses.
Search WWH ::
Custom Search