Keywords:
Toxicity, Cancer, Radiation therapy / Oncology, CT, Oncology, Cardiac
Authors:
D. Capra1, C. B. Monti2, M. Melloni2, C. Gumina2, A. Luporini3, F. Lombardi2, A. Sironi2, E. Asti2, L. Bonavina2, F. Sardanelli3, F. Secchi3; 1Segrate/IT, 2Milano/IT, 3Milan/IT
Methods and Materials
Study population
Patients with thoracic esophageal cancer, who had undergone CT examinations at our institution were retrieved from the picture archiving and communication sistem (PACS) of our institution. Patients who did not undergo radiation therapy were excluded, as well as patients who did not have either a staging CT before treatment, or a follow up CT after treatment. Patients who had cardiac comorbidities were also excluded, to avoid any possible confounding factor altering the attenuation measurement of the myocardial tissue. Fig. 1
Image analysis
A reader with 1 year of experience in cardiac CT, with the aid of an experienced radiologist, chose the slice which best allowed the visualization of the left ventricle in the long axis. A round, as large as possible region of interest (ROI) was placed in the mid-level ventricular septum, to obtain its attenuation measurement. Since the utilized staging and follow-up CT scans were not gated, cardiac movement was considered in the decision of placing the ROIs in the middle of the septum, thus excluding the areas close to the intraventricular blood pool, prone to more movement. Similarly, a second ROI was placed in the intraventricular blood pool at the same level, avoiding the papillary muscles. Fig. 2 This was first done on the enhanced scans, as myocardium and blood pool were clearly recognizable, and then ROIs were placed in the same positions on unenhanced scans and adjusted following attenuation value patterns. Fig. 3
Septal ECV was calculated using the formula first proposed by Bandula et al.:
ECV = (1 - Haematocrit) x [(HUMyopost - HUMyopre)/(HUBloodpost - HUBloodpre)]
where: myo = myocardium; blood = intraventricular blood pool; pre = pre-contrast; post = post-contrast.
Statistical analysis
Shapiro-Wilk tests were conducted to assess data distribution. Normal data were reported as mean ± standard deviation. Non-parametric data were reported as median and interquartile range (IQR). Measurements were compared using Student’s t test for paired data for normal distributions, or Wilcoxon test for non-normal distributions. Pearson and Spearman’s correlation test were used according to data distribution. Statistical analysis was performed with R v3.5.3. P-values <0.05 were considered as significant.