Radiotherapy respiratory motion management in hepatobiliary and pancreatic malignancies: a systematic review of patient factors influencing effectiveness of motion reduction with abdominal compression

Abstract

Background

The effectiveness of abdominal compression for motion management in hepatobiliary-pancreatic (HPB) radiotherapy has not been systematically evaluated.

Methods & materials

A systematic review was carried out using PubMed/Medline, Cochrane Library, Web of Science, and CINAHL databases up to 1 July 2021. No date restrictions were applied. Additional searches were carried out using the University of Manchester digital library, Google Scholar and of retrieved papers’ reference lists. Studies conducted evaluating respiratory motion utilising imaging with and without abdominal compression in the same patients available in English were included. Studies conducted in healthy volunteers or majority non-HPB sites, not providing descriptive motion statistics or patient characteristics before and after compression in the same patients or published without peer-review were excluded. A narrative synthesis was employed by tabulating retrieved studies and organising chronologically by abdominal compression device type to help identify patterns in the evidence.

Results

The inclusion criteria were met by 6 studies with a total of 152 patients. Designs were a mix of retrospective and prospective quantitative designs with chronological, non-randomised recruitment. Abdominal compression reduced craniocaudal respiratory motion in the majority of patients, although in four studies there were increases seen in at least one direction. The influence of patient comorbidities on effectiveness of compression, and/or comfort with compression was not evaluated in any study.

Conclusion

Abdominal compression may not be appropriate for all patients, and benefit should be weighed with potential increase in motion or discomfort in patients with small initial motion (<5 mm). Patient factors including male sex, and high body mass index (BMI) were found to impact the effectiveness of compression, however with limited evidence. High-quality studies are warranted to fully assess the clinical impact of abdominal compression on treatment outcomes and toxicity prospective in comparison to other motion management strategies.

Keywords:

• Abdominal compression
• motion management
• SABR
• hepatobiliary cancer
• pancreatic cancer

Introduction

Despite advances in management, hepato-pancreatic-biliary (HPB) malignancies including pancreatic cancer, hepatocellular carcinoma (HCC) and cholangiocarcinoma carry poor prognoses. In England the one-year survival rates for pancreatic cancer, HCC and cholangiocarcinoma are 25% [1], 38% [2] and 30% [3], respectively.

Over the last two decades, radiotherapy techniques have evolved allowing increased precision: stereotactic ablative radiotherapy (SABR) entails the delivery of highly precise ablative doses whilst minimising dose to surrounding normal tissue [4–7]. Reduction of dose to nearby organs at risk (OAR), including duodenum and normal liver reduces the risk of complications including nausea and bleeding [8], as well as radiation-induced liver disease (RILD) [9]. Liver SABR has been used since the early 2000s [10,11], and is showing increasingly favourable results [12]. SABR allows treatment of patients refractory to other local therapies, ineligible for radical resection, or to improve resectability [13]. SABR as primary treatment is evolving, similar survival rates to resection have been seen in small (≤5 cm) primary HCC [14]. SABR is demonstrated to be feasible as a bridge to liver transplant [4,15] and for inoperable pancreatic cancer [5,6], and can be delivered across treatment platforms.

Respiratory motion is largest in the craniocaudal direction and is a significant challenge for abdominal SABR, as the target can move outside of the beam [7,16,17], reducing dose to the target and increasing OAR dose due to steep dose gradients [18–22]. Respiratory motion over 1.0 cm significantly affects dosimetry [22], particularly for GTV and normal liver [23]. Motion reduction may allow for dose escalation and, therefore increase the biologically effective dose of SABR, leading to improved local control [24–30].

Respiratory motion management strategies include breath-hold, gating, and abdominal compression [18,21]. Abdominal compression devices exert pressure on the abdomen, forcing shallow breathing and restricting abdominal excursion [31]. Abdominal compression is recommended by the UK SABR Consortium for use in liver SABR [32]. There are two main types of compression device: an arch with a plate that sits on the upper abdomen, and a belt that is positioned around the abdomen. Belts can be made of firm material, or inflated using a sphygmomanometer (pneumatic). Abdominal compression has, however, demonstrated inter-patient variation and even motion increases in other anatomical planes [33–35].

To the best of the authors’ knowledge, the effectiveness of abdominal compression for respiratory motion management in hepatobiliary-pancreatic (HPB) radiotherapy has not been systematically evaluated. The aim of this systematic review was to critically evaluate the effectiveness of motion reduction with abdominal compression in comparison to without, and to determine which patient factors impact effectiveness of compression, for HPB radiotherapy.

Methods

Inclusion and exclusion criteria

The inclusion criteria used were defined using the Population, Intervention, Control, Outcome (PICO, Table 1(A)) criteria. Article titles and abstracts were screened for relevance. The full inclusion and exclusion criteria are outlined in Table 1(B). Studies were required to contain information on compression device and imaging modality used for motion quantification. Descriptive statistics (i.e. mean, standard deviation, and range) of craniocaudal motion magnitude with and without abdominal compression for radiotherapy localisation was required as a minimum. Studies conducted with volunteers were excluded from this review, due to the potential influence of confounding factors that do not represent this patient population. Studies conducted using two separate cohorts with and without compression were excluded due to the lack of comparative analysis of motion reduction within the same individuals. Studies comprised solely of patients of non-HPB sites (i.e. lung, stomach) were also excluded.

Table 1. (A) Population, intervention, control, outcome (PICO) values. (B) Inclusion and exclusion criteria.

A
Population	Adult patients with HPB malignancies (pancreas, bile duct, liver) of any stage receiving radiotherapy
Intervention	Abdominal compression using an arch or belt device on pre-treatment imaging
Control	Self-control: before and after abdominal compression
Outcome	Tumour (or surrogate) or organ motion magnitude
B
	Inclusion	Exclusion
Publication type	Full length Peer-reviewed journal articles Available in English	Conference abstracts Non-peer-reviewed articles Grey Literature No English translation available
Publication Date Range	No limits	No limits
Population	Adults (aged 18+ years) Pancreatic, liver or gallbladder cancer patients	Children aged 0–18 years Non-patient volunteers Non-hepatobiliary or pancreatic tumour
Motion quantification method	Studies with and without abdominal compression in the same individuals	Grouped cohort studies not evaluating motion with and without compression in the same individuals
Descriptive data	Patient characteristics (i.e. gender, age, weight, BMI, Child-Pugh score)	Articles providing no information on patient characteristics
	Descriptive statistics (i.e. mean, standard deviation, and range) of craniocaudal motion magnitude	Articles only providing motion reduction in % or similar Articles only providing reproducibility/inter-fraction motion information

Search strategy and data sources

A systematic review was carried out of published literature, according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines [36]. Searches were conducted by the primary author (MD) using databases such as PubMed/Medline, Cochrane Library, Web of Science, and CINAHL. The search included studies published up to 17 March 2021 and was updated on 1 July 2021. No publication date restrictions were applied. Multiple known synonyms for key search terms were used, and the following search strategy was employed, with minor adjustments made for specific databases: (‘abdominal compression [All Fields]’ OR corset [All Fields]) AND (‘radiotherapy [MeSH]’ OR ‘radiation therapy’ [All Fields] OR ‘sabr’ [All Fields] OR ‘sbrt’ [All Fields] OR ‘stereotactic’ [All Fields] OR proton [All Fields]) AND (‘liver’ [All Fields] OR ‘liver neoplasms’ [MeSH] OR ‘hepat*’ [All Fields] OR ‘pancrea*’ [All Fields] OR ‘pancreatic neoplasms’ [MeSH] OR ‘hcc’ OR ‘gallbladder’ OR cholangiocarcinoma OR ‘gallbladder neoplasms’ [MeSH]). Additional searches were carried out using the University of Manchester digital library, Google Scholar and of reference lists of retrieved papers to ensure completeness. The search keyword ‘corset’ was added after a hand reference search retrieved articles with this synonym for an abdominal compression belt.

Data extraction and synthesis

The key outcomes of patient factors (BMI, age, sex), motion measurement methods (i.e. fluoroscopy; 4D computed tomography, 4DCT, 4D/cine MRI), and motion amplitude with and without abdominal compression were extracted from included studies by the primary author (MD) and captured in a spreadsheet. To analyse the results of this review, a narrative approach was used as meta-analysis or pooling of data were not possible due to heterogeneities in study design, motion measurement methods, and patient populations. Studies were tabulated and organised chronologically by abdominal compression device type (i.e. arch, belt) to help identify patterns in the evidence [37].

Results

The initial search yielded 143 results (Figure 1). Duplicates were removed (n = 24), and titles screened to eliminate those not relevant to the search (n = 58). Studies were excluded based on abstract review (n = 23). Abstract-only studies (n = 23) were also excluded, due to lack of patient detail. Finally, studies were excluded based on eligibility in-text (n = 9). Five studies were excluded after full-text review despite otherwise apparently fulfilling eligibility criteria and were referred to in the discussion instead. First, the format of reported motion data precluded comparison with other studies [38], the second and third due to primary analysis being conducted on CBCT images [39,40]. The final two were excluded because the with or without compression did not provide any intra-patient motion reduction comparisons [41,42]. Six studies with a total of 152 patients were included in the final analysis (Table 2).

Figure 1. Flowchart describing search and exclusion strategy, based on the PRISMA guidelines [36].

Table 2. Overview of included studies including choice of abdominal compression (AC) device, number of included patients, and radiotherapy modality.

Reference	Study design	N (site)	AC Device	Modality
Wunderink, 2008 [33]	Prospective non-randomised, non-blinded	12 Liver (10 metastases, 2 HCC)	Arch	SABR
Eccles, 2011 [35]	Prospective non-randomised (part of Phase I/II study)	60 Liver (30 metastases, 24 HCC, 6 CCA)	Arch	SABR
Lovelock, 2014 [43]	Retrospective non-randomised, non-blinded	42 (30 liver, 6 adrenal, 3 pancreas, 3 nodes)	In-house developed pneumatic belt	SABR
Heerkens, 2017 [34]	Prospective non-randomised (part of Phase II feasibility trial)	10 Pancreas, (9 LAPC, 1 refused surgery) (8 head, 2 tail)	In-house developed ‘corset’ belt	SABR
Van Gelder, 2018 [44]	Retrospective non-randomised, non-blinded	15 (13 Liver, 2 pancreas)	Pneumatic belt	SABR
West, 2018 [45]	Retrospective non-randomised, non-blinded	13 (12 liver, 1 kidney)	Pneumatic belt	SABR

HCC: hepatocellular carcinoma; CCA: cholangiocarcinoma.

Three studies included a single treatment site cohort of either liver or pancreas [33–35], whereas three included mixed sites [43–45]. One study [43] was included in this evaluation despite inclusion of some non-HPB sites (i.e. adrenal gland, abdominal node), as the majority (71%) were HPB patients.

Study design and risk of bias

Sample sizes ranged between 10 and 60 patients. Two studies were conducted as Phase I or II studies [34,35], sample size rationale were not provided for any other studies. Participant recruitment to studies was pragmatic, primarily based on consecutively referred patients in a single-centre setting. Additionally, none of the studies included randomisation of participants, which may have contributed a degree of confounding. One study reported exclusion criteria of presence of ascites or colostomy [35]. No studies included blinding of participants or staff, leading to a risk of performance and detection bias.

Motion quantification

Measurement methods are outlined in Table 3. Reduction in mean craniocaudal motion with abdominal compression ranged from 0.7 mm to 6.6 mm.

Table 3. Changes in craniocaudal (CC) motion amplitude with and without abdominal compression (AC) for studies included in this systematic review.

	Reference	Measurement method	Measurement Point	CC Motion Without AC	CC Motion With AC
Arch	Wunderink, 2008 [33]	Fluoroscopy	Fiducial markers	Mean 10.8 mm Range 4.4–25.3 mm	Mean 4.2 mm Range 1.7–8.5 mm
	Eccles, 2011 [35]	T2-weighted cine-MRI	Tumour	Mean 11.7 mm Range 4.8–23.3 mm	Mean 9.4 mm Range 1.6–23.5 mm
Belt	Lovelock, 2014 [43]	Fluoroscopy	Surgical clips or Fiducial markers	Mean 11.4 mm Range 5–20 mm	Mean 4.4 mm Range 1–8 mm
	Heerkens, 2017 [34]	T2-weighted Cine MRI, 4DCT	Tumour	Mean 11.3 mm Range 7.5 − 22.1 mm	Mean 7.2 mm Range 4.1 − 12.1 mm
	Van Gelder, 2018 [44]	4DCT	Liver dome	Mean 8.7 ± 3.0 mm	Mean 8.0 ± 3.8 mm
	West, 2018 [45]	4DCT	Right diaphragm	Median: 12.0 ± 4.4 mm	Median: 8.0 ± 5.1 mm

Change in motion amplitude

All studies reported a reduction in craniocaudal respiratory motion magnitude (Table 3). However, four studies demonstrated motion increases in one or more planes in a minority of patients with abdominal compression [33–35,44]. While overall mean tumour motion reduced in the study by Eccles et al. [35], there was negligible reduction in the overall maximum with abdominal compression, and an increase in one or more directions was seen in 28% of patients. Similarly, maximal increases in the left-right direction of 1.6 mm in 8 patients [33], 0.4 mm in 10 patients [34] were seen.

Patient characteristics

Only one study provided all five areas of patient characteristics (sex, age, performance status, weight (or BMI) and Child-Pugh score) [31]. The age range for all studies was 28-89 years (Table 4). HPB cancers are more prevalent in males [1,2], which is reflected in the larger proportion of men in included studies. Only one study evaluated the impact of BMI, finding no correlation between BMI and percentage craniocaudal motion reduction [43].

Table 4. Patient characteristics for studies included in the systematic review.

	Reference	Sex (M: F)	Age	PS	Weight/BMI	CP score
Arch	Wunderink, 2008 [33]	10:2	–	KPS > 80 s	–	A–B
	Eccles, 2011 [35]	32:28	Median: 64.2 years Range 30–85 years	KPS >60	–	A
Belt	Lovelock, 2014 [43]	–	Median: 60 years Range 38 − 84 years	–	No correlation of BMI with % motion reduction	–
	Heerkens, 2017 [34]	–	–	ECOG ≤2	–	–
	Van Gelder, 2018 [44]	10:5	Mean age 67.9 years Range 46–89 years	–	–	–
	West, 2018 [45]	100% Male	Median age 62 years Range 28–83 years	–	–	–

Dashes signify this information was not provided in the study. PS: performance status; KPS: Karnofsky performance status; ECOG: Eastern Cooperative Oncology Group; CP: Child–Pugh.

Comorbidities were not described in any study, although factors such as prior surgery or abdominal pain were thought to cause large variations in achievable compression pressure in one [44]. Patient comfort was mentioned in five studies [33,34,43–45], four using a belt [34,43–45]. Lovelock et al. [43] reported use of lower belt inflation pressures patients were uncomfortable. Patient tolerance was cited by Van Gelder et al. [44] limiting factor for inflation which may have limited compression effectiveness. Anti-anxiety medication was administered in two studies [33,43]. Mean motion reduction with and without lorazepam was 5.8 and 7.4 mm, respectively [43]. One study reported excluding patients with colostomy from evaluation with compression [35].

Discussion

This review is, to our knowledge, the first to evaluate the effectiveness of abdominal compression in reducing target and organ motion including the impact of patient factors for radiotherapy to HPB malignancies. Six studies measuring respiratory motion reduction with abdominal compression in HPB radiotherapy are summarised here. This review shows that abdominal compression adequately reduces craniocaudal respiratory motion in the majority of included patients, although motion increases in other anatomical planes are possible and may be due to changes from abdominal to chest breathing. Some inter-patient variability in motion reduction is evident, however there is a paucity of evidence on the association of patient-specific factors on compression effectiveness.

Care must be taken when interpreting results from non-randomised observational studies due to the risk of potential bias [46]. All included studies were single-centre setting, lacking external validity [47], the results of which therefore cannot be assumed to be valid for all populations. Blinding was not possible due to the nature of abdominal compression devices. Sample size selection rationale was not clear in most studies within this review, which may be due to a pragmatic approach considering clinical capacity and funding. Randomised-controlled trials (RCTs) have always been regarded as the highest level of clinical evidence due to bias limitation and methodological rigour [48]. However, due to the rapid incremental nature of technical radiotherapy advances, these are not often performed in the setting of immobilisation devices [49,50]. There has not been a systematic evaluation of abdominal compression devices, having been developed incrementally since the early 1990s [51]. The current evidence has not yet reached a stage where disease control, toxicity and patient-reported outcome measures (PROMs) are evaluated [52]. There is a lack of equipoise in this setting, as any reduction in respiratory motion is presumably preferable to none [53,54].

No discernable difference between device type was seen in this review. However, variation was seen with different points of measurement (i.e. delineated contours, visible tumour, diaphragm), as studies reporting small motion increases in anterior-posterior or left-right planes measured liver tumour or intratumoral fiducial marker motion directly. This suggests that small motion changes may not be apparent at the whole-organ level due to deformation [31]. Whole-organ, or surrounding structures, are therefore not appropriate surrogates for accurate motion monitoring. Abdominal compression can cause day-to-day variation in deformation of OARs such as the stomach and duodenum [55], which should also monitored to reduce significant changes in plan dosimetry.

Accurate motion characterisation is vital for HPB radiotherapy, as over- or under-estimation of target and OAR motion can lead to under- and over-dose, respectively. The motion quantification methods varied across studies in this review, which may have influenced motion magnitude results. 4DCT is commonly used for motion quantification of abdominal tumours where discretely visible [32,56,57], or with intravenous contrast [32]. 4DCT scans are retrospectively sorted into a series of image phase bins based on breathing signals, typically measured by an external surrogate [58]. MRI, however, typically provides improved soft tissue contrast over CT-based imaging [59–61]. 4DCT has also been shown to underestimate three-dimensional liver tumour motion amplitude in comparison to cine-MRI [62], and liver fiducial marker motion in comparison to fluoroscopy [63].

Cheng et al [64] saw diaphragm motion reduction on 4DMRI images, although not all volunteers demonstrated reduction under abdominal compression. Evaluation by Case et al. [65] of 29 patients, 14 with compression and 15 without, showed reduced average liver motion amplitude in craniocaudal and anterior-posterior planes, however a small (<1 mm) increase in the left-right plane. Motion increases in planes other than craniocaudal may be a result of changes from diaphragmatic to chest wall breathing. Nevertheless, as craniocaudal motion tends to be greatest, small changes in left-right motion may not pose a significant issue for radiotherapy planning as long as the total 3-dimensional vector of motion remains small. Dolde et al. [66] found using 4DMRI that the mean magnitude of pancreatic deformation was reduced in the majority of volunteers using an abdominal corset in comparison to no immobilisation. Eccles and colleagues [31] found similarly promising results, with liver deformation of <5 mm in the majority of patients. A dosimetric evaluation of an abdominal compression arch device found that AC did not significantly reduce OAR doses in pancreas SABR, and slightly increased dose to some such as the right kidney, and duodenum [67]. However, this evaluation was conducted in seven lung patients, so may not be representative for the pancreatic cancer patient population.

Tumour location may impact motion magnitude; however, this requires validation. Cranial and anterior liver segments have demonstrated largest displacements [68] without compression. Similarly, tumours within the peripheral liver segments were associated with target motion in 145 CyberKnife patients [69]. Evidence on the impact of liver cirrhosis and magnitude of motion is conflicting, with cirrhotic patients previously found to have significantly larger motion in free breathing [70], however a more recent study showed no association [69]. Patients with cirrhosis have previously been shown to have more tumour motion in free breathing than those without using tracking of a single gold FM [69], although a more recent study using tracking of three to eight FM found no association [70]. Reduced liver strain, a measure of tissue elasticity, has been demonstrated in the presence of cirrhosis [71–73]. Exploration of the relationship between cirrhosis and liver motion is required, particularly as most HCC patients have some degree of cirrhosis [74]. Post-hepatectomy patients have been shown to have reduced respiratory motion in free-breathing than patients who had not undergone resection [75,76]. This may be due to detachment of liver ligaments used in respiration [75] or postoperative tissue adhesion [76]. There currently is no evidence available for abdominal compression in patients with stoma such as colostomy, due to frequent exclusion from having abdominal compression in studies.

Male sex and high BMI were predictive factors for reduced effectiveness of an arch system [42]. In comparison, no correlation between BMI and motion reduction was found with a belt, although patients were administered lorazepam in this study which may have ameliorated the results [43]. BMI does not take into account fat distribution [77], and increased abdominal adipose tissue may act like a cushion, attenuating the pressure of abdominal compression [42]. Belt devices may therefore be superior in larger patients, due to circumferential pressure application. The subxiphoid area is the ideal position for abdominal compression, with benefit reducing inferiorly [41,44]. If the device composition requires positioning more inferiorly to avoid the treatment beam, then an alternative motion management method should be considered.

Patient discomfort with abdominal compression is frequently cited but has not yet been measured. Physical discomfort increases the risk of the patient adjusting their position mid-treatment [65,78]. Arch systems may be particularly uncomfortable due to the pressure exertion from one point [31]. Due to the nature of abdominal compression, all devices may be uncomfortable for patients, particularly in patients with pre-existing pain [44,65]. Claustrophobia can also be distressing for patients, manifesting in feelings of restriction or suffocation [79], both considerations in abdominal compression. Anti-anxiety medications (i.e. lorazepam) may improve the compression tolerance or even augment motion reduction [80]. Patients with advanced HPB cancer often present with upper abdominal or back pain, which alone can increase free-breathing intrafraction respiratory amplitude changes [40,65]. Pain can limit optimal pressure application [44], therefore an appropriate pain management plan should be formulated for patients if required.

Despite not being the aim of the present review, positional reproducibility with abdominal compression should be considered. Device positioning can vary slightly, which can increase day-to-day deformation of OARs [55]. Interfraction liver deformation with compression has been demonstrated to be <5 mm in most patients for SABR fractionations [31,40,81–83]. Abdominal compression may reduce treatment setup error and mean image matching times [84], which is advantageous for effective IGRT to minimise the risk of intrafraction changes [21,85].

OAR motion with digestion and cardiac function may not be addressed by abdominal compression. The stomach size and shape varies considerably from day to day in free breathing [86], even with fasting [55,87]. Motility-associated craniocaudal motion of 10 mm occurs independent of pancreatic head motion [88], and small bowel diameter can vary by 10.8 mm with peristalsis [89], although anti-spasmodic medications (e.g. hyoscine butylbromide) can reduce peristaltic motion [90,91]. It is assumed that abdominal compression worsens dosimetry in pancreatic radiotherapy by reducing the distance from OARs to the PTV [92], however this has not been proven.

Abdominal compression systems require a cost outlay, which may be prohibitive for some clinics. Their size, shape, and composition can limit their use on treatment platforms such as Tomotherapy or MR-Linac, There are additional considerations on these platforms, including MR-safety and compatibility with linac bore size [93]. Additionally, attenuation of the treatment beam causes artefacts on imaging or uncertainties in plan dosimetry [23,43], and is dependent on device composition, although this is not evaluated in the present review. This sometimes requires positioning inferior to the treatment beam, which is suboptimal for motion management. Abdominal compression devices can deform the external contour [94], or cause day-to-day variations in tissue density along the beam path, potentially leading to significant errors in proton dosimetry [95]. Therefore, their use on emerging treatment platforms must be fully evaluated.

Based on this review, there is no clear difference between either abdominal compression device type for respiratory motion management. Despite the advantages of compression, the reduction in respiratory motion can be inconsistent and demonstrate inter-patient variation. Alternative motion management with breath hold or gating should be considered in cases that require device positioning inferiorly to avoid the treatment beam. While compression can reduce the magnitude and variability of craniocaudal respiratory motion, residual motion from digestion and other processes is unavoidable. Therefore, tumour and OAR motion and deformation should be monitored on treatment using intrafraction imaging. There is currently no high-quality evidence with which to stratify patients for abdominal compression effectiveness, however quantification of initial motion and identification of the benefit of abdominal compression on a patient-specific basis would reduce unnecessary use. Measurement and reporting of patient comfort are required in future studies to help develop patient-centred processes, as tolerance will be limited by the patient’s specific circumstances. Further study into developing robust motion management strategies that promote patient comfort, and usable across radiotherapy platforms is warranted.

This systematic review has several limitations. The review was carried out according to a pre-determined unpublished protocol but was not registered to a database. Articles were coded for relevance and inclusion by a single author (MD), which may have introduced a risk of selection bias. Measurement of motion in included studies was carried out using heterogenous imaging methods, some which show real-time motion (cine-MRI, fluoroscopy), and 4DCT which is retrospectively sorted based on respiratory phase, which may underestimate motion. Another limitation of the current systematic review is the lack of a systematic evaluation of the risk of bias of included studies. However, due to the nature of the type of evidence included, all studies are considered for the purpose of this review as having a moderate to serious risk of bias particularly with regard to selection, performance, and detection bias.

Conclusion

Respiratory motion is patient-specific and therefore requires patient-specific management strategies. While an effective motion management strategy in HPB radiotherapy, abdominal compression may not be appropriate or effective for all patients. Patient-specific factors including BMI, or levels of abdominal subcutaneous adipose tissue, may impact effectiveness of motion magnitude reduction. In cases of small respiratory motion magnitude (< 5 mm), the benefit of abdominal compression should be weighed with the risk of potential increased motion and patient discomfort. In cases where target motion magnitude is not sufficiently reduced, alternative motion management methods, such as breath hold, should be considered where available. Monitoring of tumour and OAR motion during treatment delivery is vital, even when using motion management strategies such as abdominal compression. Comparison of abdominal compression with other motion management techniques in adequately powered prospective randomised studies, evaluating long term tumour control, toxicity, and PROMs, is required.

Acknowledgements

Mairead Daly acknowledges support from the NIHR Manchester Biomedical Research Centre and ARTNET.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.

Additional information

Funding

Mairead Daly is supported by Manchester CRUK RADNET. Ananya Choudhury is supported by NIHR Manchester Biomedical Research Centre. Cynthia Eccles is supported by NIHR Manchester Biomedical Research Centre. This work is also supported by The Christie Charity; Cancer Research UK Manchester Centre.

Mairead Daly is supported by Manchester Cancer Research UK RadNet [C1994/A28701], the Advanced Radiotherapy Technologies Network (ART-NET) [C309/A21993], the NIHR Manchester Biomedical Research Centre, and The Christie Hospital Charitable Fund. Ananya Choudhury and Cynthia Eccles are supported by NIHR Manchester Biomedical Research Centre. This work is also supported by Cancer Research UK Manchester Centre.