Objective To observe the value of 18F-D3FSP PET/CT machine learning(ML)models for evaluating subjective cognitive decline(SCD).Methods Thirty-two SCD patients(SCD group)and 16 healthy volunteers(control group)who received 18F-D3FSP PET/CT were selected from Sino Longitudinal Study on Cognitive Decline(SILCODE)and divided into training set(n=34)and test set(n=14)at a ratio of 7∶3.Support vector machine(SVM),random forest(RF)and logistic regression(LR)models were constructed based on Hamilton anxiety scale(HAMA)and standard uptake value ratio(SUVR)of brain regions being significantly different between groups to evaluate SCD.Then PET/CT data were amplified by format conversion and divided into training set(including 8 775 CT images and 1 833 PET images)and test set(including 2 025 CT images and 423 PET images)at the ratio of 8∶2.VGG16 models were constructed based on CT and PET images to evaluate SCD,respectively.Results The area under the receiver operating characteristic curve(AUC)of SVM,RF and LR model for evaluating SCD in training set was all 1.000,while was 0.863,0.872 and 1.000 in test set,respectively.LR model was overfitting,and RF model had better performance.The accuracy of VGG16 model for evaluating SCD based on CT and PET images tended to be stable after the 175th and 150th iterations,with the highest accuracy of 67.11％ and 65.35％ in training set,which tended to be stable after the 165th and 145th iterations,with the highest accuracy of 62.43％ and 59.16％ in test set,respectively.Conclusion 18F-D3FSP PET/CT ML models could be used to evaluate SCD.

Objective To observe the value of 18F-D3FSP PET/CT machine learning(ML)models for evaluating subjective cognitive decline(SCD).Methods Thirty-two SCD patients(SCD group)and 16 healthy volunteers(control group)who received 18F-D3FSP PET/CT were selected from Sino Longitudinal Study on Cognitive Decline(SILCODE)and divided into training set(n=34)and test set(n=14)at a ratio of 7∶3.Support vector machine(SVM),random forest(RF)and logistic regression(LR)models were constructed based on Hamilton anxiety scale(HAMA)and standard uptake value ratio(SUVR)of brain regions being significantly different between groups to evaluate SCD.Then PET/CT data were amplified by format conversion and divided into training set(including 8 775 CT images and 1 833 PET images)and test set(including 2 025 CT images and 423 PET images)at the ratio of 8∶2.VGG16 models were constructed based on CT and PET images to evaluate SCD,respectively.Results The area under the receiver operating characteristic curve(AUC)of SVM,RF and LR model for evaluating SCD in training set was all 1.000,while was 0.863,0.872 and 1.000 in test set,respectively.LR model was overfitting,and RF model had better performance.The accuracy of VGG16 model for evaluating SCD based on CT and PET images tended to be stable after the 175th and 150th iterations,with the highest accuracy of 67.11％ and 65.35％ in training set,which tended to be stable after the 165th and 145th iterations,with the highest accuracy of 62.43％ and 59.16％ in test set,respectively.Conclusion 18F-D3FSP PET/CT ML models could be used to evaluate SCD.

Objective:To compare the differences of actual absorbed doses of liver malignant tumors after 90Y-selective internal radiation therapy (SIRT) evaluated by 90Y PET/CT and 90Y bremsstrahlung (BRS) SPECT/CT imaging, and to compare the image quality of the 2 imaging methods. Methods:Twenty-one patients (15 males and 6 females; age: (52.4±15.4) years) with liver malignant tumors (15 cases of primary liver cancer, 6 cases of liver metastases; 39 lesions) between September 2021 and August 2022 were retrospectively analyzed. All patients underwent both 90Y PET/CT imaging and 90Y BRS SPECT/CT imaging in the Department of Nuclear Medicine, Hainan Cancer Hospital. The ratios of the actual absorbed doses based on 90Y PET/CT imaging and 90Y BRS SPECT/CT imaging to the lowest standard absorbed dose(100 Gy) for tumor response were calculated. The image contrast and distinguishability of the two imaging methods were scored. Wilcoxon signed rank test and Wilcoxon rank sum test were used for data analysis. Results:The tumor absorbed doses evaluated by 90Y PET/CT and 90Y BRS SPECT/CT were 143.94(55.91, 233.48) Gy and 107.82(53.59, 157.53) Gy respectively. The doses evaluated by 90Y PET/CT were higher than the standard threshold in 24 lesions, while 19 lesions showed higher evaluated doses by 90Y BRS SPECT/CT than the standard threshold. Compared with 90Y PET/CT, 90Y BRS SPECT/CT underestimated the tumor absorbed dose of -24.25%(-32.32%, -12.14%). The ratio of dose evaluated by 90Y PET/CT to the lowest standard threshold was 1.33(0.56, 1.91), which was higher than that of dose evaluated by 90Y BRS SPECT/CT to the lowest standard threshold (0.97(0.47, 1.25); z=0.04, P<0.001). PET/CT image contrast was scored 0, 1, 2, 3 in 2, 2, 12, 23 lesions respectively, and SPECT/CT image contrast was scored 0, 1, 2, 3 in 2, 3, 16, 18 lesions respectively ( z=-1.29, P=0.199). The distinguishability scores of 0, 1, 2 based on PET/CT images were found in 3, 15, 21 lesions, while those based on SPECT/CT images were found in 4, 32, 3 lesions respectively ( z=-2.79, P=0.005). Conclusion:90Y PET/CT imaging is superior to 90Y BRS SPECT/CT imaging in radiation dose evaluation and tumor focus differentiation in patients with liver malignant tumors after 90Y-SIRT.