Computational protocol: Clinical impact and network of determinants of tumour necrosis in colorectal cancer

Similar protocols

Protocol publication

[…] To evaluate the accuracy of the visual assessment of tumour necrosis, one H&E stained section per case of 50 randomly selected CRC cases was scanned using the Aperio AT2 image-capturing device (Leica Biosystems, Wetzlar, Germany). Imagescope software 11.2 (Aperio Technologies, Vista, CA, USA) was used to view the scanned images, and one image was captured per each case at × 8 magnification. Two researchers independently performed visual necrosis percentage estimations for each image and after this, tumour necrosis area was demarcated by hand and its proportion was calculated with ImageJ (US National Institutes of Health, Bethesda, MD, USA), utilising the analyse particles tool. [...] Normally distributed continuous variables are presented as mean (s.d.) while other continuous variables are presented as median (IQR). The statistical analyses were carried out using IBM SPSS Statistics 22.0 (IBM, Chicago, IL, USA). To test the statistical significance between two categorical variables, χ2 or Fisher's exact test was used. The statistical significance of the associations of tumour necrosis percentage with clinicopathological characteristics was analysed by Mann–Whitney U-test (two classes) or Kruskal–Wallis test (three or more classes). Pearson's coefficients (r) were applied in examining correlations between continuous variables, and logarithmic transformation was applied to the variables with positive skewness to accomplish a more normal distribution. Cytoscape, an open source software platform for visualising complex networks, was used in creating a 2D visualisation of the relationships between tumour necrosis and clinicopathological variables with the Prefuse force directed algorithm weighted by the statistical significances of the associations between individual variables (). Receiver operating characteristics (ROC) analysis was used to define an optimal cutoff scores, with the shortest distance to the coordinate (0,1), for tumour necrosis and immunohistochemical markers (CD31, vWF, CD105 and Ki-67) in discriminating survivors from nonsurvivors (). DFS, CSS and OS were measured and univariate survival analysis was performed according to the Kaplan–Meier method. Cox's proportional hazards regression model was conducted to analyse the independent prognostic significance of tumour necrosis. Pearson's r and κ coefficients were used to evaluate intra- and interobserver agreement. A two-tailed P-value <0.05 was considered statistically significant. […]

Pipeline specifications

Software tools ImageScope, ImageJ, SPSS, Prefuse
Applications Miscellaneous, Whole slide imaging, Microscopic phenotype analysis
Diseases Neoplasms, Colorectal Neoplasms