iPETcertum (v1.0)

iPETcertum (v1.0) is image processing software designed for radiologists and nuclear medicine physicians to enhance positron emission tomography (PET) images through noise reduction, sharpening, and resolution improvement. It can optionally segment lesions based on uptake values and supports PET, PET/CT, and PET/MRI images with any radionuclide. Enhanced images are saved in DICOM, NIfTI, or ECAT formats alongside the original images.

Both the subject device (iPETcertum) and predicate device (Claritas iPET) use the same core 3D non-local means filtering algorithm optionally guided by CT or MR data. The subject device adds optional variance stabilization before filtering and automated lesion segmentation based on activity range and voxel connectivity, whereas the predicate used manual segmentation. Both operate on Windows/Linux and support the same modalities and file format inputs.

Development followed ISO 13485 (quality system regulations), IEC 62304:2006 (software development lifecycle), and ISO 14971:2019 (risk management). Performance testing measured signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), root mean square error (RMSE), and DICE index for lesion delineation accuracy.

The devices share identical core technology and intended use for PET image enhancement. Performance testing demonstrated that iPETcertum with variance stabilization disabled produces equivalent SNR, PSNR, and RMSE values as the predicate, and with variance stabilization enabled, actually improves these metrics, particularly on noisy low-dose scans. The automated lesion segmentation feature met acceptance criteria (≥50% DICE overlap with manual ground truth) without raising new safety or effectiveness concerns, as it operates on the same enhanced images and does not alter the underlying algorithm or introduce new risks.