Does Microenema Administration Improve the Quality of DWI Sequences in Rectal MRI?

Abstract

Purpose:

To determine whether the administration of a microenema immediately prior to rectal magnetic resonance imaging (MRI) decreases the level of gas-related artifacts on diffusion-weighted imaging (DWI) sequences.

Methods:

This retrospective analysis included 492 (183 baseline and 309 post total neoadjuvant treatment [TNT]) consecutive MRI scans for rectal cancer from January 2019–January 2020. Scan-related factors were identified including microenema use (yes or no), field of view (FOV) in DWI (b=800 or b=1500), and magnet strength (1.5T or 3T). Two readers scored DWI studies for gas-related artifacts and T2-weighted sequences for the amount of intraluminal gas on a 5-point scale. Fisher’s exact test and the Rao–Scott Chi-squared test were used to examine associations between microenema use and other factors. Generalized estimating equation and multivariable regression models were performed to examine the effect of microenema use in subgroups for each reader. Cohen’s κ was used to assess inter-reader agreement.

Results:

Gas-related artifact levels decreased in scans with microenema (P < 0.001) as well as when stratified by FOV (P ≤ 0.003). For both readers, post-TNT scans with microenema showed lower artifact levels overall (P < 0.014 and P < 0.001), and in post-TNT subgroups of axial DWI (P ≤ 0.006 and P < 0.001) and scans acquired with a 3T magnet (P ≤ 0.001 for both FOV). No evidence of decreased artifact level was found for baseline studies. Decreased gas was seen with microenema use (P < 0.001 for both readers). Inter-reader agreement on artifact level and gas level assessments ranged from slight to substantial (κ=0.273–0.685).

Conclusion:

Microenema use prior to rectal MRI reduces gas-related artifacts on DWI, including both large and small FOV sequences and particularly on post-TNT scans performed at 3T, and offers a viable solution to improve DWI quality.

Introduction

Colorectal cancer represents the third leading cause of new cancer cases and the second leading cause of cancer-related deaths in men and women combined. In 2020, approximately 43,340 new cases of rectal cancers will be diagnosed in the United States. The estimated overall number of people to die from colorectal malignancies in 2020 is 53,200 [1]. Currently, management plans are tailored for patients upon diagnosis, based on clinical evaluation and diagnostic imaging. Rectal MRI is an established examination in both baseline and post-total neoadjuvant treatment (TNT) evaluation in rectal cancer patients [2,3]. Within post-TNT rectal MRI, diffusion-weighted imaging (DWI) sequences have shown added value, particularly to determine the presence of viable tumor [4–7]. As such, the latest recommendations from the 2016 European Society of Gastrointestinal and Abdominal radiology (ESGAR) consensus meeting included the routine use of DWI sequences in rectal cancer imaging, particularly for restaging to evaluate response to TNT [8]. However, clinically utilized DWI sequences in rectal MRI are prone to susceptibility artifacts, including gas-related artifacts due to a gas-filled bowel [9].

Magnetic susceptibility artifacts occur at the interface of two matters secondary to differences in the level of magnetization, leading to image ghosting and distortion. These artifacts are reportedly exaggerated at higher field strengths [2,10–12]. In the case of a gas-filled bowel, when the rectum is evaluated, gas-induced susceptibility artifacts can markedly compromise the diagnostic image quality, sometimes even rendering the images non-diagnostic leading to decreased accuracy in interpretation.

Decreasing gas-induced susceptibility artifacts is essential for accurate imaging interpretation. Attempts to minimize gas-related artifacts have largely relied on manipulating DWI parameters [13,14]. Recently, several institutions have implemented different preparatory approaches, including the use of microenema. The use of microenema prior to rectal MRI post TNT demonstrated significant reduction in the incidence and severity of gas-related artifacts when evaluating artifacts performed on 1.5T systems using a b-value of 1000 s/mm² [15].

The purpose of our study was to determine whether administration of microenema prior to rectal MRI decreases the level of gas-related artifacts on DWI sequences in various imaging scenarios when the rectum is the organ of investigation. We hypothesized that there would be fewer and less severe artifacts when a microenema is used in both baseline and post-neoadjuvant treatment imaging under varying b-values and magnetic field strengths.

Materials and methods

Study population

This retrospective, single-institution study was approved by the Institutional Review Board with a waiver of written informed consent and was compliant with the Health Insurance Portability and Accountability Act. We retrospectively examined our database for consecutive patients who underwent rectal MRI after the introduction of microenema to our routine rectal MRI protocol on January 30, 2019, and identified all scans performed until January 30, 2020. The inclusion criterion for this study was MRI scan performed for a diagnosis of rectal cancer at baseline (i.e. baseline MRI) or within one year of completion of TNT (i.e. post-TNT MRI). The exclusion criteria were lack of information whether microenema was used or not, duplicate studies, same-day enema for a different procedure, post-surgical studies, non-rectal cancer patients, recurrences, patients with stents in situ, post-induction-chemotherapy scans, and post-TNT scans performed one year after completion of chemoradiotherapy. The study selection process is outlined in Figure 1.

Fig 1.

Flowchart of included MRI rectal scans

Rectal MRI protocol

All scans were acquired using different GE Healthcare System platforms (Discovery 2MR750, Optima MR450w, Signa EXCITE, and Signa PET MRI; Waukesha, WI) with either 1.5 Tesla (1.5T) or 3 Tesla (3T) magnet strength and a pelvic phased-array coil. Our routine rectal MRI protocol included a straight axial T2-weighted sequence and two separate DWI sequences in addition to other T2-weightd multiplanar sequences. Axial DWI included a large field of view (FOV) axial sequence (24 cm) with a b-value of 800 s/mm² and matched the straight axial T2-weighted sequence slice locations from the iliac crest to anal verge. FOV optimized and constrained undistorted single-shot (FOCUS) DWI included a smaller FOV axial sequence (varied based on patient size) with a b-value of 1500 s/mm² and matched to the straight axial T2-weighted sequence slice locations (16 slices only) covering the area of the tumor. Table 1 details the the routine acquisition protocol for T2-weighted axial and DWI sequences for both 1.5T and 3T magnet strength scanners. Table 1 shows our acquisition parameters for the axial T2 and DWI sequences used for scoring in our study for both 1.5T and 3T magnet strength scanners. It does not include our complete imaging protocol.

Table 1.

MRI protocol for the axial T2 and DWI sequences used for scoring in our study.

	1.5T magnet strength scanner			3T magnet strength scanner
Series Descriptor	Axial T2	Axial DWI b 800	Focus DWI b 1500	Axial T2	Axial DWI b 800	Focus DWI b 1500
Generic sequence name	FSET2	2D	Focus Diffusion	FSET2	2D	Focus Diffusion
Plane	Axial	Axial	Axial	Axial	Axial	Axial
Field of View (cm)	20–24	24	24	20–24	24	20–24
Slice Thickness (mm)	5	5	5	5	5	5
Gap (mm)	1	1	1	1	1	1
b-value		800	1500		800	1500
Saturation Pulse	S/I/A	N/A	N/A	S/I/A	N/A	N/A
TE1/TE2	110	Min	Min	110	Min	Min
TR	4000–6000	Auto-TR (4000–10000)	Auto-TR (4000–10000)	4000–6000	Auto-TR (4000–10000)	Auto-TR (4000–10000)
Flip Angle	90	N/A	N/A	90	N/A	N/A
Bandwidth (kHz)	32	N/A	N/A	41	N/A	N/A
ETL	24	N/A	N/A	24	N/A	N/A
NEX	3	16	16	3	16	16
Frequency Steps	320	128	128	320	128	140
Phase Encoding Steps	224	128	128	224	128	70
Frequency Direction	A/P	R/L	R/L	A/P	R/L	R/L

TE, echo time; TR, repetition time; ETL, echo train length; NEX, number of excitations; FSE, fast spin echo; A/P, Anterior/Posterior; R/L, Right/Left.

The perscribed microenema was a hyperosmolar docusate sodium laxative in 5 ml liquid (Enemeez®). The onset of action is reported to be within 2–15 mins of application and was self-administered by the patient immediately before the imaging acquisition. Microenema was not administered in patients who had colostomy, known anal stricture, rectal abscess or fistula, anal or rectal pain, current feeling of malaise, nausea, or abdominal pain, or if patients expressed personal preference of not having the microenema.

MRI image assessment

We evaluated the presence and severity of gas-related artifacts (i.e. artifact level) within the rectum in all DWI studies with and without microenema as well as the presence and amount of rectal gas (i.e. gas level) in all axial T2-weighted studies.

DWI studies were scored by two readers (VSJ and SJT) with experience in oncologic imaging blinded to clinical information, to each other’s scores, as well as to whether microenema was or was not administrated prior to the scan. Artifact level was scored based on a 5-point scale as: 1 = no artifact, complete visualization of the rectal wall and lumen; 2 = mild artifact (no effect on diagnostic image quality, < 25% rectal wall distortion), 3 = moderate artifact (some effect on diagnostic image quality, 25–50% rectal wall distortion), 4 = severe artifact (considerable effect on diagnostic image quality but still interpretable, 50–75% rectal wall distortion), and 5 = extensive artifact (non-diagnostic images, > 75% rectal wall distortion) [16,17] (Figure 2). The initial 20 studies were jointly read by both readers in order to normalize the subsequent scoring and the remaining studies were read independently. There was a one-week period between the scoring of axial DWI and FOCUS DWI studies to reduce recall bias. Axial T2-weighted studies were similarly assesed for gas level based on a 5-point scale as: 1 = no gas, 2 = minimal gas, 3 = small amount of gas, 4 = moderate amount of gas, and 5 = large amount of gas [16,17] (Figure 3).

Fig 2.

Representative axial slices depicting the presence and severity of artifacts within the large field of view axial DWI sequences with b = 800. The scoring system was based on 5-point scale as follows: Score 1 = No artifact (a). Score 2 = Mild artifact (b) (arrow). Score 3 = Moderate artifact (c). Score 4 = Severe artifact (d). Score 5 = Extensive artifact (e)

Fig 3.

Representative T2 axial slices depicting the presence and amount of gas within axial T2 sequences. The scoring system was based on a 5-point scale as follows: 1 = No gas (a). 2 = Minimal amount of gas (b). 3 = Small amount of gas (c). 4 = Moderate amount of gas (d). 5 = Large amount of gas (e)

Statistical analysis

Fisher’s exact test was used to examine the associations between microenema use and patient characteristics, artifact level (no/mild vs moderate/severe/extensive), and gas level (no/minimal/small vs moderate/large). The Rao–Scott Chi-squared test was used to examine the overall association between microenema use and artifact level for each reader separately. Some patients had multiple studies included; we analyzed these studies independently considering that the artifact level is primarily gas-level related rather than patient-level related.

Subsequently, two sets of multivariable regression models were performed for each reader to examine the effect of microenema use under specified combinations of DWI view (axial DWI and FOCUS DWI views), scan timing (baseline and post-TNT groups), field strength (1.5 and 3T), and gas level.

Model I: To examine the effects of microenema use on artifact level in subsets of baseline and post TNT, and within each timing group, on axial DWI and on focus DWI views, and further under 1.5 and 3T field strength, we adopted general estimating equations (GEE) regressions with interactions terms, a logit link function and an exchangeable correlation structure. No model selection was performed. Adjustment for multiple testing was applied on confidence interval (CI) estimations and p-values using the simultaneous inference approach [18] on multiple contrasts.

Model II: To test whether the effect of microenema use on artifact level is different under different gas levels in all scans, GEE with a logit link function and an exchangeable correlation structure was used. Adjustment for multiple testing was also applied.

To assess inter-reader agreement on artifact level and gas level, Cohen’s kappa (κ) was estimated with bootstrapping 95% CIs. κ values were interpreted as follows: 0.00–0.20 = slight agreement, 0.21–0.40 = fair agreement, 0.41–0.60 = moderate agreement, 0.61–0.80 = substantial agreement, and 0.81–1.00 = almost perfect agreement.

All statistical analyses were performed in software packages R version 3.6.2 (The R Foundation for Statistical Computing).

Results

The study group consisted of 394 patients (236 males and 158 females) with a mean age of 59 years (ranging from 19 to 96 years) and a total of 492 MRI scans. A lower proportion of women had scans with microenema compared to men (P = 0.008). There was no significant difference in the number of scans with microenema and without microenema based on scan timing (baseline vs. post-TNT) and magnet field strength (1.5T vs. 3T). A summary of scan characteristics is presented in Table 2.

Table 2.

Summary of patient characteristics, scan artifact levels, and associations with microenema (ME) use.

	Overall1	ME Used1	ME not used1	p-value2
Overall	492	196	296
Age at diagnosis	59 (19, 96)	58 (22, 88)	60 (19, 96)	0.046
Sex				0.008
Male	301 (61%)	134 (68%)	167 (56%)
Female	191 (39%)	62 (32%)	129 (44%)
Scan timing				0.6
Baseline	183 (37%)	76 (39%)	107 (36%)
Post CRT	309 (63%)	120 (61%)	189 (64%)
Field strength				0.7
T1.5	113 (23%)	43 (22%)	70 (24%)
T3	379 (77%)	153 (78%)	226 (76%)

¹Statistics presented: median (minimum, maximum); n (%)

²Statistical tests performed: Wilcoxon rank-sum test; Fisher’s exact test

Both readers reported a significant decrease in artifact level in studies with microenema compared to studies without microenema in the entire cohort (P < 0.001) as well as within axial DWI views (P < 0.001) and FOCUS DWI views (P ≤ 0.003).

Overall, for post-TNT studies, microenema use showed a significant decrease in artifact level. Post-TNT studies performed with axial DWI views showed consistently lower odd ratios (ORs) of moderate to extensive artifacts with microenema use. For post-TNT FOCUS DWI views, one reader showed significant decrease in the level of artifacts with microenma use. When stratified according to field strength, among the post-TNT studies performed with a 3T field strength, both readers reported a significant decrease in artifact level with microenema use; however, among the post-TNT studies performed using a 1.5T scanner, only one reader reported a significant decrease in artifact level with microenema use for both axial and FOCUS DWI views (Figure 4).

Fig 4.

Forest plot on odds ratio estimates of artifact levels for studies performed with microenema (ME) vs without microenema (ME vs No ME) in all post-TNT studies for reader 1 (R1) and reader 2 (R2), as well as stratified first by field of view (axial DWI and FOCUS DWI) followed by magnet strength (1.5T and 3T). Except for studies performed on FOCUS DWI and 1.5 magnet strength for one reader, there was a significant decrease in artifact level for all post-TNT studies.

On the other hand, for baseline studies, microenema use did not show any significant improvement in artifact level, including when further stratified acording to DWI view or field strength (Figure 5).

Fig 5.

Forest plot on odds ratio estimates of artifact level for studies performed with microenema (ME) vs without microenema (ME vs No ME) in all baseline studies for reader 1 (R1) and reader 2 (R2), as well as stratified first by field of view (axial DWI and FOCUS DWI) followed by magnet strength (1.5T and 3T). There was no significant decrease in artifact level with microenema use in any baseline group.

Between the axial DWI and FOCUS views, there was no significant difference in artifact level either overall or when categorized by field strength (P > 0.1 for both readers in all groups).

Regarding microenema use and gas level, studies with microenema showed a signficantly decreased gas level (P < 0.001 for both readers). However, if the gas level was moderate to large, there was no significant difference in artifact level whether or not microenema was used (OR = 0.79 for both readers; 95% CI: 0.45–1.4 and 0.41–1.52 for R1 and R2, respectively). Similar results were seen in studies with no or small amount of gas (OR = 0.61 and 0.55, 95% CI 0.33–1.13 and 0.3–1.02 for R1 and R2, respectively).

Inter-reader agreement on artifact level and gas level assessments ranged from slight to substantial between the two readers (κ = 0.273–0.685) for overall scans and in subgroups of scan timing, scan type, field strength, and gas level (Figure 6).

Fig 6.

Cohen’s κ along with bootstrapping confidence intervals for inter-reader agreement on artifact levels for all studies performed with microenema vs without microenema as well as in all subgroups and for rectal gas.

Discussion

In this study, we investigated whether microenema use prior to rectal MRI reduces gas-related artifacts on diffusion-weighted images. Our results showed that, overall, when microenema is administered just before the acquisition of MRI rectal studies, significantly lower levels of artifacts are present on diffusion-weighted images compared to studies performed without microenema. This remained true when results were stratified for the axial DWI views and FOCUS DWI views. When comparing the effect of microenema on artifact level between post-TNT studies and baseline studies, post-TNT studies (both axial DWI and FOCUS DWI) performed with a 3T magnet showed a significant decrease in artifact level with microenema use.

DWI sequences are an integral part of the restaging MRI post TNT as they aid in lesion detection and characterization [18,19]. Previously, Van Griethuysen et al. studied the effect of administering microenema to improve the image quality of DWI of the rectum at 1.5T after long course chemoradiotherapy [15]. The study reported an excellent interobserver agreement and showed a significant decrease in the incidence and severity of the susceptibility artifact in the group with microenema. Our study featured a larger sample size and all studies were scored by both readers. In addition, the majority of our scans were performed on 3T scanners and we analyzed post-TNT scans using two FOVs in DWI. The added value in our study is that we showed that microenema use can significantly improve the quality of post-TNT rectal MRI performed at 3T magnet strength in both axial DWI and FOCUS DWI sequences.In addition to post-TNT studies, we also included baseline MRI studies for evaluation. Although the recent literature suggests that DWI sequences can help in tumor delineation and also provide prognostic information [20], the routine use of DWI in the baseline evaluation of rectal MRI is still debated and performed based on each institution’s preference. Our study did not show any significant decrease in the level of gas-related artifacts in baseline studies in general or when stratified according to FOV or magnet field strength when microenema was administered. This lack of significance may in part be due to the small number of baseline studies compared to the post-TNT studies, especially when further stratified into various subgroups (field strength, DWI view).

At our institution, we routinely perform two DWI sequences in the straight axial plane, one with a large FOV with the highest b-value of 800 s/mm² and a second with a small FOV with the highest b-value of 1500 s/mm² centered over the tumor. The effect of microenema use under a high b-value of 1500 s/mm² has not been studied. Thus, our study newly shows that microenema use can significantly decrease the level of artifacts on a sequence with a high b-value of 1500 s/mm, although it did not demonstrate that the level of improvement was better with a higher or lower b-value.

In our study, microenema use showed a significant decrease in the amount of rectal gas. An important consideration here is that the microenema was administered just before MRI aquisition. Rectal gas may accumulate if there is a considerable time lapse between emptying the bowel and MRI acquisition, although this time interval may vary between individuals. When there was a moderate to large amount of gas in the rectum, we did not detect any significant improvement in the level of artifacts despite microenema use. This suggests that rectal gas plays an important role in the development of higher levels of artifacts and we postulate that improvements in image quality in studies with microenema is secondary to emptying bowel gas immediately before the scan. This theory might be applied to baseline studies where the gaseous dilatation of bowel may be limited by the presence of significant intraluminal tumor causing luminal rigidity, thereby questioning the value of microenema in these cases. However, further studies will be required to prove this concept.

One interesting observation is that there was a significant sex imbalance in our cohort, with a higher proportion of male patients undergoing microenema administration compared to female patients. We did not specifically study the effect of microenema based on the anatomical differences between male and female patients as it was beyond the scope of this study. However, this could certainly be investigated in furture studies. The exact cause for not administering microenema was not recorded consistently in our database; however, the most common cause was felt to be patient preference. This may in part explain the greater proportion of females undergoing studies without microenema and strengthens the argument for patient as well as clinician education on the usefulness of microenema in this clinical scenario.

Our study has certain limitations. First, we were limited by an imbalanced cohort, with the majority of patients undergoing studies on a 3T scanner compared to 1.5T which probably affected lack of statistical significance in the 1.5T group. However, susceptibility artifacts are exaggerated at 3T [10], and thus our results in the 3T group are particularly relevant. Second, we did not separate our analyses based on the initial staging of the disease or whether residual disease was present on the post-treatment scans. However, whether the presence of significant intraluminal tumor volume impacts the value of microenema would require further research and was beyond the scope of current study. Third, image quality was assessed based on the subjective assessment of individual readers. We are aware that such subjective judgment has the potential to yield inconsistency between observers, and we may have introduced observer bias. To alleviate this limitation, we used a 5-point scale and implemented joint training by the two readers at the commencement of study. Fourth, patients were scanned on different scanners and we did not evaluate the technical acquisition parameters. However, the goal of this study was not to optimize the technical parameters of DWI. Finally, this was a retrospective analysis with its inherent limitations. However, the acquisition of rectal MRI scans and the administration of microenema in our institute are very much standardized.

In conclusion, our study shows microenema use in bowel preparation immediately before acquisition of rectal MRI significantly reduces susceptibility artifacts in both axial DWI and FOCUS DWI sequences. The effect is most consistently seen on post-TNT studies performed with a 3T magnet strength. The presence of rectal gas has a significant impact on the severity of artifacts and is appreciably reduced by the administration of microenema just before the acquisition of scan.