Synthetic PET From CT Improves Precision Diagnosis and Treatment of Lung Cancer: a Prospective, Observational, Multicenter Study

This is a core outcome measure evaluating the degree of resemblance between the structural characteristics of synthetic PET images and reference PET images. It reflects whether the synthetic PET retains the key structural information of the original image, which is a fundamental indicator for confirming the structural relevance of synthetic PET.

A commonly used objective evaluation index in image quality assessment, calculated based on the mean square error between the synthetic PET image and the reference image. It quantifies the ratio of the maximum possible signal value in the image to the noise power that affects image quality. Higher PSNR values indicate that the synthetic PET image has less noise interference and better consistency with the reference image in terms of signal characteristics, thereby reflecting better retention of structural information.

An index designed to simulate human visual perception to evaluate image structural similarity, which comprehensively considers three aspects: brightness, contrast, and structural consistency between the synthetic PET image and the reference image. The value range is between 0 and 1; a value closer to 1 means that the synthetic PET image has a higher degree of structural consistency with the reference image, indicating that the structural relevance is well retained.

A key outcome measure for verifying the biological relevance of synthetic PET, which evaluates whether the metabolic characteristic parameters derived from synthetic PET are consistent with those from reference PET. Metabolic parameters are directly related to the biological functions of tissues/organs, so their consistency is crucial to ensuring that synthetic PET can be used for accurate biological activity assessment.

Validation study of indicators demonstrating the added value of synthetic PET in lung cancer diagnosis: Accuracy in distinguishing benign from malignant lesions, Accuracy in diagnosing lymph node metastasis, Accuracy in diagnosing distant metastasis, Including corresponding diagnostic sensitivity, specificity, positive predictive value, negative predictive value, area under the curve (AUC), etc.