Factors Influencing The False Negative Of CT-guided Transthoracic Core Needle Biopsy

Purpose:The purpose of this study was to investigate the factors influencing the false negative rate in CT-guided transthoracic core needle biopsy (TTNB) for the diagnosis of benign and malignant lesions in thorax.Materials and Methods:From January 2008 to December 2013,1339 consecutive TTNB with final diagnosis were included from Sir Run Run Shaw Hospital. we retrospectively reviewed the CT images, clinical data and pathological results of 65 false negative and 106 true positive cases. The SPSS 20 software was used for statistical analysis. Mann-Whitney U test was used to compare the age, lesion size, lesion depth, needle path length, needle-pleural angle between two group; the gender, lesion types, lesion location, cavity, necrosis, patient position, coaxial biopsy, needle types, the number of punctures and pneumothorax between two groups were compared by Fisher’s exact test. The same statistical method was also used to compare the groups with different lesion size, lesion depth, needle path length and needle-pleural angle between two group. Subsequently, the variables found to be significantly different by univariate analyses were subjected to multivariate logistic regression analysis to determine the independent factors. p< 0.05 was defined as statistical significance.Results:1237 were true-positive,31 were true-negative,4 were false-positive, and 67(including 2 missing data) were false-negative. The overall diagnostic accuracy was 94.7%(1268 of 1339 lesions). The sensitivity and specificity for the diagnosis of malignancy and benign lesions was 94.9%(1237 of 1304 lesions) and 88.6% (31 of 35 lesions), respectively. Positive and negative predictive values were 99.7%(1237 of 1241 lesions) and 31.6% (31 of 98 lesions), respectively. There is statistical significance for lesion size (p=0.035), location (p=0.005), necrosis (p=0.026), coaxial biopsy (p=0.017), the number of punctures (p=0.014), pneumothorax (p=0.020) between two groups. However, no significance was found for age (p=0.071) and gender (p=0.631), lesion types (p=0.812), lesion depth (p=0.584), patient position (p=0.717), needle path length (p=0.370), needle-pleural angle (p=0.156), needle types (p=0.382). In multivariate logistic regression analysis, predictive factors for a false-negative result were lesion size(OR,2.773; 95%CI,1.395-5.509; p=0.004), location(OR,0.770; 95%CI,0.651-0.911; p=0.002), coaxial biopsy(OR,0.245; 95%CI,0.101-0.595; p=0.002).Conclusions:The false negative results of TTNB were significantly associated with the lesion location, lesion size, necrosis, coaxial biopsy, the number of punctures, pneumothorax; the independent risk factors for the false negative results were lesion size, location, coaxial biopsy.