What are the main causes of interfraction motion of the uterine fundus and cervix?

Abstract

Purpose or Objective Interfraction motion of uterus and cervix can be large and often exceeds applied CTV-PTV margins. There is variation between patients and different parts of the uterus move differently. Most strategies accounting for this are based on bladder filling. However, we hypothesise that there are other causes of uterine motion. Our study aims to provide a qualitative assessment of causes of motion of the uterine fundus and cervix in a large cohort of patients. Material and Methods Anonymised scans were retrospectively obtained for 83 patients who received radical radiotherapy for cervical cancer, with imaging at 3 time points (Fig. 1); diagnostic MRI scan (1), planning CT, ~2 weeks later (2) and final week MRI scan, ~6 weeks later (3). Scans were registered on bony anatomy to the diagnostic MRI for each patient. The uterus was contoured by a single observer for all 249 scans on a single sagittal slice identified as mid of the uterus on scan 1. Motion at the cervix and uterine fundus was evaluated between scans 1-2 (S1-2) and 1-3 (S1-3). The main cause of motion and direction of motion was recorded based on visual interpretation. Results In S1-2, large motion (>1cm) was seen in 44 cases (53%) at the cervix level and in 65 (78%) at the fundus level. In S1- 3, this was 57 (69%) at the cervix level and 64 (77%) at the fundus level. Large motion at the cervix rarely occurred without large motion at the fundus: in S1-2, 3 cases (4%) and S1-3, 9 cases (11%). The main causes and direction of motion are summarised in figure 2. For the cervix, the most common cause of motion in S1-2 was rectal change with 30 cases (36%) and in S1-3 it was tumour regression, also with 30 cases (36%). Bladder filling differences only accounted for cervix motion in 5 cases (6%) in S1-2 and 1 case (1%) in S1-3. Main drivers of motion at the fundus were bladder filling with 23 cases (28%) in S1-2 and 18 cases (22%) in S1-3. However, motion was also related to rectal changes in 13 cases (16%) in S1-2 and S1-3, to bowel changes in 21 cases (25 %) in S1-2 and 12 cases (15%) in S1- 3, and to tumour regression in 16 cases (19%) in S1-3. At the cervix, in S1-2, there was a superior/inferior component of motion in 20 cases (24%), mainly related to rectal changes. In S1-3, there was a superior/inferior component of motion in 38 cases (45%), mainly related to tumour regression. Conclusion The main causes of cervical motion in our cohort are changes in rectal filling and tumour regression, with bladder filling playing a limited role. Motion at the uterine fundus is affected by bladder filling but other factors also have an important role. Rectal motion can lead to changes in the superior/inferior position of the cervix, which should be considered if developing an ITV. Our study suggests that current radiotherapy motion management strategies based on bladder filling may not account for the most important causes of cervix motion. Alternative approaches such as online adaption may be beneficial.