Purpose: To demonstrate the use of radiomics in extracting features from cone beam computed tomography (CT) images of prostate cancer patients to quantify the amount of gas in the rectum and the probability of the patient experiencing diarrhea following radiotherapy.
Methods: Two prostate cancer patients (65 and 83 years old) undergoing radiotherapy were monitored for 20 days over the course of their treatment. Periodic CT scans and region of interest contour structures were recorded in DICOM files. Python programming was used to read into the DICOM files and extract hounsfield units (HU) found within the rectum. HU histogram distribution, statistical HU values, and gas fluctuations over the treatment course were assessed.
Results: There was a strong correlation between the average HU values and the percentage of gas found within the rectum. The 83 year old patient had significantly greater gas fluctuations than the 65 year old patient. However, the peaks and troughs of gas levels in both patients over the course of treatment showed slight correlation.
Conclusion: Extracting HUs from the rectum of a prostate cancer patient has shown the potential to quantifying the amount of gas in the rectum over the course of radiotherapy treatment and thus, determine the probability of the patient experiencing diarrhea as a side-effect of treatment.