Address Sanitizer in Mixed C++ and CUDA Code

I am debugging a Windows heap corruption error in a program that involves a mix of .cpp and .cu files. To address this, I am attempting to use the Address Sanitizer in Visual Studio. However, I encounter an issue when a CUDA API call, such as cudaMallocHost, is present in a .cpp file. This results in an access violation due to the inability to allocate pinned memory. Conversely, if the program is implemented solely with .cu files, it runs smoothly. Unfortunately, compiling all files with nvcc is not feasible because many components of the program are implemented in C++20, and I am currently using CUDA 11.4. Is there a workaround to mitigate this issue?