Eliminating the Need for CT Scans In Abdominal MR-Linac Treatments: Organ-Specific Electron Density Characterization
Abstract
Purpose
To eliminate the need for CT scans in abdominal MR-Linac treatments by determining and tabulating organ electron densities (EDs) based on patient age and sex. Resultantly, these results promote simpler workflow, costing less time, patient dose, and expenditure from both the patient and treatment facility.
Methods
One hundred abdominal cancer patients (aged 41 – 98) treated with an MR-Linac were retrospectively analyzed. ED data from contoured CT scans were extracted from thirteen structures (external, esophagus, heart, left and right lungs, liver, spinal canal, stomach, left and right kidneys, large bowel, small bowel, and duodenum). Median values were calculated and tabulated under patient sex and age. Five liver SBRT treatment plans were recalculated using median values from an atlas to determine deviations in dose to PTVs and organs at risk.
Results
Male and female patients exhibit highly overlapping organ ED distributions with similar medians for many structures. Females demonstrate a lower median ED for the external body contour (0.968) compared to males (0.979). Organs with the most variability include the stomach (0.48-1.15), large bowel (0.52-1.01), and small bowel (0.71-1.04), due to the presence or absence of gases. Across all patients and organ structures, dose calculations using tabulated median electron density values demonstrate minimal deviation from CT-based plans, with percent differences remaining within ±2.1% for all evaluated targets and organs at risk.
Conclusion
This study demonstrates that averaged electron density values yield comparable dose calculations, supporting the feasibility of using tabulated data to override electron density values. These findings support clinical changes to omit the CT on MR-Linac patients, thus reducing patient dose, time, and expenditure. Future work will involve further categorizing this data using a larger sample size and patient BMI information to better calculate average values across populations.