Home » Posts tagged 'tumour heterogeneity'
Tag Archives: tumour heterogeneity
Feedback PLC – ICENI Centre & ISLCRS study to investigate whether textural features of rectal cancer on Magnetic Resonance can predict long term survival
Magnetic resonance based texture parameters as potential imaging biomarkers for predicting long term survival in locally advanced rectal cancer patients.
Omer Jalil, Asim Afaq, Balaji Ganeshan, Uday Patel, Raymond Endozo, Darren Boone, Ashley Groves, David Humber, Tan Arulampalam
Colchester University Hospital
Heterogeneity is a well-known feature of malignancy and is associated with adverse outcomes. Tumour heterogeneity can be quantified on imaging using textural analysis (TA), which assesses the distribution of pixel grey-level intensity, coarseness and regularity and have shown to predict survival outcomes in a number of oncological applications. The main objective of this study was to investigate whether textural features of rectal cancer on MR can predict long term survival.
Patients with primary non-metastatic locally advanced rectal adenocarcinoma treated with long course chemoradiotherapy (LCRT) with curative intent from 01/2006 to 06/2011 in our institution were included. TA of T2-weighted pre and post LCRT MR images was undertaken using TexRAD, a proprietary software algorithm (TexRAD Ltd www.texrad.com, part of Feedback Plc). Tumour region of interest (ROI) delineating the largest cross-sectional area underwent MRTA comprising an image filtration-histogram technique. Filtration step extracted features of different sizes (fine, medium, coarse – texture scales) followed by quantification of statistical features (mean intensity, standard-deviation, entropy, skewness, kurtosis and mean of positive pixels – MPP) using histogram analysis. Univariate Kaplan Meier analysis was used to assess the ability of the biomarkers (textures and other clinically employed radiological and histological features) to predict survival. Cox-multiple regression analysis determined which significant univariate markers were independent predictors of survival.
For overall survival (OS), pre-treatment texture features (MPP at fine-texture, p=0.008, Mean at medium-texture, p=0.03) were significant univaraite markers of OS. Using multi-variate analysis, these pre-treatment textures (MPP at fine-texture, HR: 6.5, 95% CI: 1.9 – 22.0, p=0.002, Mean at medium-texture, HR: 5.6, 95% CI: 1.4 – 21.7, p=0.013) and post-treatment EMVI positive status (HR: 5.2, 95% CI: 1.6 – 16.8, p=0.006) were the only independent predictors of OS.
For disease-free survival (DFS), pre-treatment texture (Mean at medium-texture, p=0.007) and post-treatment texture (Kurtosis at medium-texture, p=0.009) were significant univaraite markers of DFS. Using multi-variate analysis, the post-treatment textures (Kurtosis at medium-texture, HR: 3.5, 95% CI: 1.3 – 9.5, p=0.013) and pre-treatment CRM involvement (HR: 5.2, 95% CI: 1.5 – 18.1, p=0.010) were the only independent predictors of DFS
Non-invasive imaging features derived during the filtration histogram method of MR Textural analysis can predict patients with poorer prognosis before undergoing surgery, and could lead to selection of patients for more intensive treatment before undergoing surgery.
Table1- Significant features on univariate and multivariate analysis-OS
|OS: Significant pre-treatment texture parameters||Filter value||Univariate analysis P value||Multivariate analysis p value|
|OS: Significant post-treatment texture parameters|
|OS: Significant Clinical Variables|
Table 2: Significant features on univariate and multivariate analysis- DFS
|DFS: Significant pre-treatment texture parameters||Filter Value||Univariate Analysis P value||Multivariate Analysis P value|
|DFS: Significant post-treatment texture parameters|
|DFS- Significant Clinical Variables|