What does this research mean for the field?

SwinPix, a Swin Transformer-based framework, significantly improves the quality of low-dose PET imaging by utilizing multi-level inputs, achieving a 13% increase in PSNR and an 82-86% reduction in RMSE at 4% dose levels compared to single-input models. Novelty: ClaimNovelty.NOVEL_FINDING. Consensus alignment: ConsensusAlignment.SUPPORTS_CONSENSUS.

What question did this study set out to answer?

To evaluate the effectiveness of multi-level low-dose PET inputs for predicting standard-dose PET images using the SwinPix framework.

March 12, 2026Open Access

SwinPix: A Swin Transformer-based Pix2Pix Framework for Low-Dose PET Denoising Using Multi-level Inputs Toward Standard-Dose Quality

Key Points

To evaluate the effectiveness of multi-level low-dose PET inputs for predicting standard-dose PET images using the SwinPix framework.
Developed two models: single-input and multi-input using low-dose PET data at different dose levels (4%, 6%, 10%)
Evaluated performance using structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and root mean square error (RMSE)
Conducted comparisons with Pix2Pix and Swin Transformer for image reconstruction quality.
Multi-input model showed superior performance over single-input models across all dose levels
At 4% dose, PSNR improved by 13% and SSIM increased from 0.97 to 0.99
RMSE decreased by 82-86% for lesions
SUV<sub>mean</sub> and SUV<sub>max</sub> biases reduced by up to 78% and 58% respectively
All results were statistically significant (p < 0.01).

Abstract

This study introduces SwinPix, a novel network architecture designed to explore the effectiveness of multi-level low-dose (LD) PET inputs as prior knowledge for standard-dose (SD) PET image prediction. By employing SwinPix architecture, we assess the performance of single-input and multi-input models trained with PET data at 4%, 6%, and 10% dose levels. Two models were developed: the first was trained using a single input corresponding to 4%, 6%, and 10% LD PET images, while the second considered a multi-input approach, utilizing three lower-dose inputs to predict the corresponding SD PET images. The performance of the six models was evaluated using structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), mean standardized uptake value (SUVmean) bias, SUVmax bias, and root mean square error (RMSE) within the entire head region and malignant lesions. The SwinPix multi-input model outperformed single-input versions across all dose levels. At 4%, PSNR increased by 13%, SSIM improved from 0.97 to 0.99, while RMSE (lesion/head) dropped by 82-86%. Similarly, SUVmean and SUVmax biases decreased by 78% and 58%, respectively. At 6% and 10% dose levels, SwinPix showed comparable improvements, reducing RMSE by over 40% and SUV biases by up to 53%. Compared to Pix2Pix and Swin Transformer, SwinPix consistently achieved the best reconstruction quality. All improvements were statistically significant (p < 0.01), supporting the effectiveness of multi-input SwinPix for accurate LD PET imaging. SwinPix offers a hybrid transformer-based solution for LD PET reconstruction. By incorporating multi-level LD PET inputs through a Generative adversarial network (GAN) framework, it enhances image quality and lesion quantification while maintaining computational efficiency, supporting its potential for clinical deployment.

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper