| Objective The purpose of this study is to develop a new method to evaluate the nature of solitary pulmonary nodules.The method is based on pixel histogram analysis of solitary pulmonary nodules on computed tomographic(CT) images to retrospectively compare the sensitivity,specificity and diagnostic accuracy of an alternative method used to distinguish between benign and malignant pulmonary nodules with that of the hot-spot method,with clinical and pathological results as reference standard.Preliminary studying CT value distribution features of benign and malignant pulmonary nodules,comparing between two methods of observers/inner agreement..Materials and Methods Collect 51 cases of patients with solitary pulmonary nodules(SPNs,diameter≤3cm) in our hospital from 2007.5-2008.11. Nine of them were excluded because there were no clinical and pathologically confirmed.42 cases(24 men,18 women;age range:29~78,mean age:52 years) with pulmonary nodules were studied.These nodules were divided into three groups: malignant nodules(22 cases,including phosphorus carcinoma,adenocarcinoma, undifferentiated carcinoma and metastatic carcinoma,etc.);benign nodules(15 cases, including hamartoma,tuberculoma and granuloma,etc.);inflammatory nodules(5 cases,mainly acute inflammatory nodule,inflammatory pseudotumor).42 cases of solitary pulmonary nodules were performed with plain and dynamic contrast-enhanced CT examinations by 16-SCT repectively.Scanning conditions: 120Kv,280mA;scanning slice thickness:2.5mm,;reconstruction slice thickness:2.5mm.CT scans through SPNs were obtained at 30s and 120s after onset of injection.The CT value of 42 SPNs was measured before and after enhancement. We determined the largest increased CT value.The raw data of 42 cases were transmitted to GE AW4.2 workstation.Additionally,we segmented each solitary pulmonary nodule by CT value and obtained the constituent ratio of each segment, the segmental standard was-120~0HU,1~15HU,16~30HU,31~45HU,46~60HU, 61~75HU and 76~120HU.Two radiologists independently analyzed histograms of solitary pulmonary nodules and recorded the results of histogram analysis. Interobserver agreement of the two observer were analyzed by statistics.Results 1.For the histogram method,interobserver agreement and withinobserver agreement were almost perfect.2.The distribution of SPN' CT value in three groups(malignant,benign, inflammatory) had significant difference in the first and the second phases of dynamic enhancement(p<0.0001~p<0.0098).3.Using the histogram method,subtract the percentage of CT value in the threshold of 46~120HU from the percentage of CT value in the threshold of-120~30HU minus,take the result≤45%as the point to determine the nature of tumors,that is to distinguish benign and malignant SPN,and the sensitivity,specificity and accuracy of this method was 0.818,0.667,0.757.4.In our study,histogram method can not distinguish inflammatory SPN and malignant SPN.Conclusion For the histogram method,interobserver agreement and withinobserver agreement are more satisfactory.The Diagnostic performance of the histogram method is as the same as the hot method in a group of small samples.And the histogram method provides a new way and a new method for the diagnosis and differential diagnosis of SPN.These conclusions need to be further confirmed in large samples. |