To assess the diagnostic value of standard uptake values(SUVs)in lung lesions and to explore the optimal threshold

the biological influencing factors for SUV in lung cancer were researched.84 patients

all of whom had lung lesions

experienced PET-CT scanning(malignant lesions 59

benign lesions 30

respectively)．Every lung lesion was measured by the technique of region of interest(ROI)

and each lesion’S maximal SUV(SUVmax) and average SUV(SUVave)values were recorded.The diagnostic efficiencies of SUVmax and SUVave were compared and the optimal threshold was established according to the analysis of the ROC curve.The relationships between SUVave and pathologic type and tumor size were analyzed through the analysis of correlation and variance with covariates.There is no significant difference between the areas under curves of SUVmax(O.8975)and SUVave(O.8966)

so the diagnostic values of these two methods for lung lesions are similar.When SUVmax>5.2 or SUVave>4.4 was taken as the cut-off value

the diagnostic accuracy was the highest

and the sensitivity

specificity and accuracy is 89.8%

80.O%

86.5%

respectively.Squamous cell carcinoma had higher SUV(8.5±2.9)and bigger tumor size(4.I±I.5 cm)than adenocarcinoma(5.3±2.2

2.5±1.0 cm)(P<O.05).There is a significant positive correlation between tumor size and SUV either in squmnous cell carcinoma or in adenocareinoma(P<0.05).After eliminating the influence of tumor size

SUV of squamous cell carcinoma(7.2±0.5)is still higher than that of adenocarcinoma(6.1±0.4)

but there is no statistic significance(P>0.05).It is concluded that SUV is useful for differentiating lung lesions.When SUVmax>5.2 or SUVave>4.4 is taken as the cut-off value

the diagnostic accuracy is highest．The SUV of lung cancer is more likely influenced by tumor size than by pathologic type.