Dear Colleagues
It had been suggested to me that I consider a FROC (Free Response ROC)
analyis
for a study I am involved in. I would be grateful if anyone can direct me
to a reference which will allow me to assess whther this technique is
appropriate or not. The study is decribed below. If there is any related R
or S code around I would also be grateful for any references/links.
Briefly, this will be a multiple reader multiple case study where the
readers are reading a CT scan of a colon and identifying polyps ( possibly
more than 1 per patient). As far as I can assess at present the
specificity will therefore not be available ( on a polyp basis) and so a
standard ROC curve cannot be computed.
many thanks for your help
Susan Charman
|