Managed C++ Interoperability

Has anyone successfully created a CUDA program in a managed C++ application that uses the .NET runtime? If so, are there any special compiler options that need to be used. I have a file form1.cpp that calls a functions in the file imageProcess.cu. The imageProcess.cu function issues a kernel call. I get this error message:
imageProcess.cu(43) : error C2059: syntax error : ‘<’ when compiling form1.cpp.