Maintaining SUV Accuracy in Low-Count PET with PETfectior: A Deep Learning Denoising Solution

Yamila Rotstein Habarnau; Nicolás Bustos; Paola Corona; Christian González; Sonia Traverso; Federico Matorra; Francisco Funes; Juan Martín Giraut; Laura Pelegrina; Gabriel Bruno; Mauro Namías

arXiv · 2512.02917

Maintaining SUV Accuracy in Low-Count PET with PETfectior: A Deep Learning Denoising Solution

Yamila Rotstein Habarnau·Nicolás Bustos·Paola Corona·Christian González·Sonia Traverso·Federico Matorra·Francisco Funes·Juan Martín Giraut·Laura Pelegrina·Gabriel Bruno·Mauro Namías

View original source ↗Read source document ↗Export metadata ↓

Abstract

Background: Diagnostic PET image quality depends on the administered activity and acquisition time. However, minimizing these variables is desirable to reduce patient radiation exposure and radiopharmaceutical costs. PETfectior is an artificial intelligence-based software that processes PET scans and increases signal-to-noise ratio, obtaining high-quality images from low-count-rate images. We perform an initial clinical validation of PETfectior on images acquired with half of the counting statistics required to meet the most recent EANM quantitative standards for 18F-FDG PET, evaluating lesions detectability, quantitative performance and image quality. Materials and methods: 258 patients referred for 18F-FDG PET/CT were prospectively included. The standard-of-care scans (100% scans) were acquired and reconstructed according to EARL standards 2. Half-counting-statistics versions were generated from list-mode data and processed with PETfecftior (50%+PETfectior scans). All oncologic lesions were segmented on both PET/CT versions, manually or automatically, and lesions detectability was evaluated. The SUVmax of the lesions was measured and the quantitative concordance of 50%+PETfectior and 100% images was evaluated. Subjective image quality was visually assessed by two experienced physicians. Results: 1649 lesions were detected in a total of 198 studies. The 50%+PETfectior images showed high sensitivity for lesion detection (99.9%) and only 1 false positive was detected. The SUVmax measured in 100% and 50%+PETfectior images agreed within 12.5% (95% limits of agreement), with a bias of -1.01%. Image quality of the 50%+PETfectior images was rated equal to or better than the standard-of-care images. Conclusion: PETfectior can safely be used in clinical practice at half counting statistics, with high sensitivity and specificity, low quantitative bias and high subjective image quality.

Explore related subjects

eess.IV physics.med-ph

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Yamila Rotstein Habarnau, Nicolás Bustos, Paola Corona, Christian González, Sonia Traverso, Federico Matorra, Francisco Funes, Juan Martín Giraut, Laura Pelegrina, Gabriel Bruno, Mauro Namías. 2025-12-02. Maintaining SUV Accuracy in Low-Count PET with PETfectior: A Deep Learning Denoising Solution. https://arxiv.org/abs/2512.02917

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

KELP: K-space-conditioned Estimation of Learned Sampling Patterns for Scan-Adaptive Multi-Coil MRI

Deep learning techniques have gained considerable attention for accelerating MRI acquisition while maintaining image quality. In this work, we present a convolutional neural network (CNN)-based framework for predicting scan-adaptive undersampling patterns directly from low-frequency multi-coil $k$-space data. Unlike approaches that optimize sampling patterns during training or rely on nearest-neighbor search at inference, our method is trained using precomputed scan-adaptive optimized masks as supervised labels and predicts a scan-specific sampling pattern in a single forward pass. The training procedure alternates between optimizing the sampling network and a reconstruction network, allowing the learned masks to adapt to reconstruction performance. Experiments on the fastMRI multi-coil knee dataset demonstrate competitive reconstruction quality compared with existing population- and scan-adaptive sampling approaches at $4\times$ and $8\times$ acceleration factors. In addition, the proposed method enables efficient scan-specific mask prediction with inference cost independent of the training dictionary size.

eess.IV↗

arXiv2026

LaminoDiff: Generative Computed Laminography via Near-Isotropic Spectral Supervision and Anisotropic Geometry

Computed Laminography (CL) is widely used for nondestructive inspection of extended planar objects, but its restricted angular coverage produces an anisotropic point spread function, missing-cone spectral incompleteness, and severe aliasing with interlayer leakage. This paper presents LaminoDiff, a physics-constrained diffusion framework for CL reconstruction. During training, a CT-derived near-isotropic supervision target is reconstructed from full-angle Computed Tomography (CT) projections physically degraded to match CL detector noise and focal-spot blur while retaining full angular coverage; it is withheld from inference. At inference, the reverse process uses only the CL observation. An Anisotropic Representation (AR) constructs depth-aware channels from three adjacent slices: a neighbor average, a center-neighbor residual, and the retained current slice, followed by directional in-plane feature extraction. Experiments on simulated and real multilayer printed circuit board data, including ball grid array and high-frequency stub samples, show that LaminoDiff improves artifact suppression, edge preservation, and depth stratification over analytic Feldkamp--Davis--Kress and representative learning-based baselines.

eess.IV↗

arXiv2026

IViT: A Novel Interpretable Visual Transformer for Skin Disease Detection

The clinical diagnosis of skin diseases is susceptible to interference from inter-class similarity of skin lesions, and over-reliance on clinicians'experience easily leads to subjective bias. Although existing deep learning aided diagnosis methods achieve competitive accuracy, they suffer from the black-box opacity of Vision Transformer (ViT) and poor adaptability to medical few-shot scenarios. Moreover, mainstream explainable algorithms generally face the bottleneck of significant accuracy degradation when improving interpretability. This paper proposes an interpretable ViT (IViT) constrained by Quadratic Programming (QP). The introduced pre-trained transfer learning adapts to few-shot feature extraction. A discrete QP feature selection framework is constructed to screen generic and discriminative features consistent with clinical diagnostic logic. A multi-objective loss function is designed to reduce feature redundancy and optimize activation distribution while preserving classification performance. Experimental results on six standard skin disease datasets show that IViT achieves an accuracy of 93.80%, only 0.21% lower than the baseline, with feature redundancy reduced by 29.5%. Its core activation regions are consistent with clinically concerned lesion areas. The proposed model balances accuracy and interpretability, providing a reliable solution for the clinical deployment of few-shot intelligent skin disease diagnosis.

eess.IV↗