RESUMEN
Objective: To investigate the value of 3.0T MRI diffusion kurtosis imaging (DKI) quantitative histogram parameters in the differential diagnosis of rectal mucinous adenocarcinoma (MC) and common adenocarcinoma (AC). Methods: One hundred and ten patients from Department of Radiology, the Second Affiliated Hospital of Soochow University between September 2015 and September 2019 with complete magnetic resonance imaging (MRI) and DKI results confirmed by surgery and pathology were retrospectively analyzed, including 16 patients in MC group and 94 patients in AC group. Two physicians outlined the region of interest (ROI) on the DKI image with b=1 000 s/mm(2), and obtained quantitative DKI parameters, including the diffusion coefficient (D value) and kurtosis coefficient (K value) corrected for non-Gaussian distribution. The apparent diffusion coefficient (ADC) values of quantitative parameters of diffusion-weighted imaging (DWI) were obtained through image registration, and histogram analysis was performed to obtain the mean value, 25th percentile, 50th percentile, 75th percentile, skewness and kurtosis of the above parameters, respectively. The difference between the quantitative histogram parameter analysis results of the rectal MC group and the AC group was evaluated, and the main indicators and multivariate comprehensive analysis indicators was screened, and the effectiveness of quantitative histogram parameters related to histopathological classification in the differential diagnosis of rectal MC and AC was evaluated. Results: There was no significant differences in gender, age, lesion location, T stage or N stage between MC group and AC group (all P>0.05). The multivariate binary logistic stepwise regression screening showed that D50th percentile and K25th percentile are statistically significant indicators (B values were 2 966.166 and -4.550, respectively; Wals values were 9.000 and 15.720, respectively; and P values were 0.003 and <0.001, respectively). The combined area under the curve of the two indictors was 0.85, but there was no statistically significant difference in pairwise comparison using DeLong method (P>0.05). The results of histogram analysis of quantitative parameters measured by the two physicians were consistent, and the inter-group correlation coefficient ranged from 0.880 to 0.981. Conclusions: The quantitative parameter histogram analysis of the DKI double-index model is helpful for the differentiation of rectal MC and AC, in which the D50th percentile and K25th percentile have differential diagnosis significance, and are superior to the ADC value of the single-index model.
Asunto(s)
Adenocarcinoma Mucinoso/diagnóstico por imagen , Adenocarcinoma/diagnóstico por imagen , Diagnóstico Diferencial , Femenino , Humanos , Espectroscopía de Resonancia Magnética , Estudios RetrospectivosRESUMEN
Objective: To establish a CT image radiomics-based prediction model for the differential diagnosis of silicosis and tuberculosis nodules. Methods: A total of 53 patients with silicosis and 89 patients with tuberculosis who underwent routine CT scans in Suzhou Fifth People's Hospital from January to August, 2018 were enrolled in this study. AK/ITK software was used to segment the images to obtain 139 silicosis lesions and 119 tuberculosis lesions. For each lesion image, 396 features were extracted, and feature dimension reduction was applied to select the most characteristic feature subset. Support vector machine (SVM) , feedforward back propagation neural network (FNN-BP) , and random forest (RF) were implemented using R software (Rstudio V1.1.463) , and the algorithm that achieved the largest area under of the receiver operating characteristic (ROC) curve (AUC) was selected as the final prediction model. Results: RF was the best prediction model for the differential diagnosis of silicosis and tuberculosis nodules, with an accuracy of 83.1%, a sensitivity of 0.76, a specificity of 0.9, and an AUC of 0.917 (95% confidence interval: 0.8431-0.9758) . RF had a significantly larger AUC than SVM and FNN-BP (P<0.05) . Conclusion: CT image-based RF prediction model can be used to differentially diagnose silicosis and tuberculosis nodules.