The Measure Methods Of Density Of The Subsolid Lung Nodules And The Mass Research

Objective:1.To study the reproducibility of density measurement of subsolid nodules by manual measurement and automatic measurement by advanced lung analysis(ALA)software.2.To explore the correlation and consistency of the change trend of of subsolid nodules density between manual and automatic measurement.3.To draw growth curves of volume and mass of subsolid nodules with different characteristicsand compare the growth patterns in malignant and low-risk pulmonary subsolid nodules.Materials and Methods:The CT data of 55 patients from July 2009 to February 2017 with at least one subsolid noudue were retrospectively analyzed.All the patients underwent at least twice CT test w.The size of the nodules at first CT scan ranged from 5mm to 3cm.Malignant and benign nodules were confirmed by histopathology.No change in volume(less than 25%)in two years were difined as low-risk nodules.Two independent observers performed manual and auto-matic softwar-generated measurements and repeated all measurements after 1 week.The Bland-Altman analysis was used to estimate the intra-and inter-observer reproducibility of manual measurements vs auto-software measurements.The correlation between manual and automatic density measurement of subsolid nodules was evaluated using pearson test.Mcnemar test was used to compare the consistency of density change trends between manual of software-generated measurement.Calculate the sensitivity and specific degrees of the two methods.The mass was calculated as the product of volume and density.The volume and mass growth curve of benign and malignant nodules were drawn and the growth patter was described.Results:1.The RD value of density measured by manual and automatic was 0% and 0.2% for intra-observer,0.2% and-0.3% for inter-observer respectively.The 95% CI for RD of manual and automatic methods was-5.7% ~ 5.8% and-3.7% ~ 4.2% for intra-observer respectively,-13.9% ~14.4% and-5.0% ~ 4.4% for inter-observer respectively.There was no significant difference between RDs measured by manual and automatic for intra-observer(P>0.05)so as for inter-observervariability(P>0.05).2.The correlation coefficient between the manual and the automatic measurement of the density of the subsolid nodules is 0.893(P<0.001),either for different size type or different density types(All P<0.001).The results of McNemar Test showed the consistency of two methods is poor on density change evaluation(χ2=6.936;P<0.001).3.Of 15 subsolid cancers,the volume growth curves of 4 cases(26.7%)manifested as steep ascendant,8 cases(53.3%)as slowly ascendant,1(6.70%)as slowly descendant and 2 cases(13.3%)as descendant-ascendant;the mass growth curves of 5 cases(33.3%)manifested as steep ascendant,6 cases(40%)as slowly ascendant,1(6.70%)as slowly descendant and 3 cases(20%)as descendant-ascendant.Of 40 low-risk subsolid nodules,the volume growth curves of 5 cases(12.5%)manifested as descendant,18 cases(45%)as flat,6 cases(15%)as slowly ascendant and 11 cases(27.5%)as undulate;the mass growth curves of 4 cases(10%)manifested as descendant,19 cases(47.5%)as flat,5 cases(12.5%)as slowly ascendant and 12 cases(30%)as undulate.Conclusion:1.The intra-and inter-observer repeatability of the manual measurement and semi-automatic software measurement is comparable.Semi-automatic measurement showed a slightly better reproducibity,There is no significant difference between the mean RD of two methods..2.A significantly correlation was detected between the manual and semi-automatic measurement of the density of subsolid nodules.However,the evaluation of the density changing trend by these two methods exist a significant difference by the results of Mcnemar test.Therefore,both methods should be combined in the purpose of density change evaluation.3.The mass and volume growth curve both can be used to test the subsolid nodules growth trend,the mass growth curve may be can detect the earlier growth of the nodules in minority cases.A steep ascendant is the main type for lung cancer with the exception of flat,slowly ascendant or even descendant.A slowly ascendant cannot exclude the diagnosis of lung cancer.