Biomedical Engineering Reference
In-Depth Information
paring oblimersen sodium plus dacarbazine with dacarbazine alone for
metastatic melanoma patients (Bedikian et al. (2006)). As discussed by
Bhattacharya et al. (2009) and Kay et al. (2011), the FDA expressed
concerns regarding the evaluation-time bias that may be introduced by
imperfect protocol adherence, and sensitivity analysis was conducted to
assess the robustness of results (FDA (2004)).
3. The length of scheduled response assessment intervals may be different
between treatment arms. Assessments can potentially be scheduled dif-
ferently across the treatment arms if the treatment cycles have different
durations in the different treatment arms. For example, in a randomized
phase III study of comparing temozolomide and dacarbazine (DTIC)
in the treatment of advanced metastatic malignant melanoma patients
(Middleton et al. (2000)), temozolomide arm treatment cycle was every
28 days, and DTIC arm treatment cycle was 21 days. The formal radio-
logic evaluation of disease progression was performed every second cycle
in both arms.
Evaluation bias is introduced by subjective determination of progression
based on clinical or nonradiologic evidence. As discussed by Bhattacharya
et al. (2009), such concern is not rare when disease progression is defined as
a composite endpoint that includes both radiological and clinical progressions
in open-label studies. In such cases, the robustness of PFS analysis results is
often a concern. Some examples using both clinical and radiologic evidence
for progression determination include study E2100 (Gray et al. (2009)), an
open-label randomized phase III study of bevacizumab plus paclitaxel com-
pared with paclitaxel alone for metastatic breast cancer, as well as study of
JHQJ (FDA (2006)), an open-label randomized phase III study of compar-
ing gemcitabine plus carboplatin with carboplatin alone for ovarian cancer
patients.
Attrition bias occurs naturally when there is a differential missing rate of
tumor assessments and/or differential treatment discontinuation rate prior to
Search WWH ::
Custom Search