Room: Exhibit Hall | Forum 8
Purpose: To explore the prognostic value of diffusion weighted magnetic resonance imaging (DWI) 3D texture features in esophageal squamous cell carcinoma (ESCC) patients undergoing concurrent chemo-radiotherapy (CRT). We constructed a nomogram to predict the over survival of patients diagnosed ESCC.
Methods: We prospectively enrolled 82 cases with ESCC into a cohort study which underwent DWI before CRT. Two groups of tumor features were examined: (1) clinical features (eg, TNM stage, age and gender) and demographics; (2) spatial texture features of apparent diffusion coefficient (ADC), which characterize tumor intensity range, spatial patterns and distribution and associated changes resulting from CRT. A reproducible and no redundant feature set was statistically filtered and validations. Pretreatment clinical parameters were used to construct a nomogram for predicting over survival of ESCC.
Results: The radiomics features ( IHIST_energy, m_contrast_1, m_clustershade_2, Diff_ClusetrTendency_2, Diff_homogeneity_2, m_lnversevariance_2, HISE and LILE) were associated significantly with the survival. The multivariate analysis showed seven 3D texture parameters extracted from ADC maps could distinguish high, median and low risk groups (Log-rank ï?£2=9.6, P=0.00824). The nomogram showed an area under the receiver operating characteristic curve (AUC) of 0.68 (95% CI, 0.68 to 0.71) in the training set.
Conclusion: The ADC 3D texture features can be useful biomarkers to predict the survival of ESCC patients underwent CRT. The combination of DWI texture and conventional prognostic factors can be used to generate robust predictive nomogram models for survival rate.
Not Applicable / None Entered.