Time Analysis in Magnetic Resonance Image-Guided Radiotherapy Workflow

Abstract

Magnetic resonance (MR) image-guided radiotherapy (MRIgRT) has been recently implemented in routine clinical practice in our department. Treatment workflow for MRIgRT differs substantially from other radiotherapy systems and consists of set-up, MR imaging, re-contouring, re-optimization & plan QA and treatment delivery. We hypothesized that delivery of total on table treatment time will be less than 60 minutes for >75% of fractions. Aforementioned treatment components were recorded in 44 patients (304 fractions), treated between September 2018 and January 2019 with our linear accelerator based MRIgRT. Effect of treatment parameters (treatment site, dose per fraction, being first or last fraction, being treated with adapted plan or not, using breath hold technique or not) on total treatment time were analyzed. Median age was 66 years (28-83). Upper abdomen (21 patients, 41.1%) and pelvic (17 patients, 33.3%) regions were most common sites treated. Most common diagnosis was prostate cancer (14 patients - 31.8%). Majority of patients (40 patients, 91%) were treated with stereotactic body radiotherapy (SBRT), with fraction number 8 or less. Twenty-nine patients (56.8 %) were treated with audio-visually coached breath-hold gating method. Median total dose was 36.25 Gy (24 – 70 Gy) and median fraction number is 5 (3 – 28). Median total treatment time for all fractions was 43.5 minutes (range, 21-125 ) with 85% of patients being treated in less than 1 hour. Median patient setup time was 10 minutes (range 5-31), contouring time was 8 minutes (range 1-45), re-optimization + QA time was 4,5 minutes (range 1-16). Median time lost due to technical problems was 7 minutes (range 2-49) and occurred in 86 fractions. The longest component of the workflow was found to be radiotherapy treatment delivery; median 18 minutes (range 8-76). There was correlation with the treatment site and the total treatment time (p<0.0001) Median total treatment durations for upper abdomen, pelvis, thorax and spine were 48.3, 40.9, 42.2 and 53 minutes, respectively. As expected, lower fraction doses (< vs > 6 Gy) resulted in shorter treatment time (median 39.4 vs 47.2, p<0.0001). Breath hold technique resulted in longer overall treatment time compared to non-breath hold techniques (median 41.7 vs. 45.1, p=0.01). Similar difference was noticed in total treatment time between fractions treated with the original plan or the adapted plan (median 38.9 vs. 45.3, p<0,0001). There was a gradual decrease in treatment time between the first and last fraction (median 51 vs 41.7, p<0.0001). Total treatment time was found to be longer for SBRT compared with non-SBRT fractions (median 45 vs 40.5, p=0.001). MRIgRT can be easily delivered in less than 1 hour in 85% of fractions if appropriate precautions are taken. Radiotherapy delivery time remains the longest component of overall treatment. Optimization of the workflow and future hardware changes are expected to further reduce overall delivery times.