Swallowing MRI—a reliable method for the evaluation of the postoperative gastroesophageal situs after Nissen fundoplication

Since its introduction in 1991, laparoscopic Nissen fundoplication has become the most common procedure in antireflux surgery [1] and long-term follow-up studies of 5 years and longer have shown a satisfactory rate of 93% [2]. However, 2–17% of these patients report new or recurrent postoperative symptoms, such as dysphagia, heartburn, and regurgitation [3]. The symptoms should resolve within 6 weeks; otherwise, a postsurgical failure as a possible reason for the symptoms has to be considered.

Postsurgical failure and its definition have been extensively discussed in the literature [4]. In some patients, failure means resumption of drug therapy. Others may benefit only from surgical revision [2], because the majority of cases are subject to morphological changes either attributable to the fundoplication wrap or the hiatal closure [5]. While different diagnostic follow-up options, such as manometry studies and endoscopy, exist, dynamic MR imaging was recently introduced as a promising non-invasive method for evaluating the gastroesophageal junction after Nissen fundoplication [6]. This method also may serve as an accessory method for the evaluation of esophageal motility [7, 8].

All patients were able to complete the MR swallowing exam. The mean age at first presentation was 52.3 years (range, 26–80 years). The patients had undergone surgery from 6 weeks to 3 years before a swallowing MRI exam (mean, 2.6 years). The leading symptoms included dysphagia (n = 33), heartburn (n = 34), recurrent regurgitation (n = 17), and feeling of chest or upper abdominal pressure (n = 12).

No complications occurred during the examinations. Identification of all anatomic landmarks, such as the course of the esophagus, the esophagogastric junction, and the stomach, as well as the fundoplication region, was possible in all patients. The average examination time, including patient positioning on the MR table, was 32 ± 7 min.

This study investigated the diagnostic performance and interrater agreement displayed by kappa statistics for MRI of the postoperative GEJ after laparoscopic Nissen fundoplication. MRI yielded good agreement with surgery and endoscopy for key morphologic findings that were causative for postoperative complaints. Interrater analysis showed moderate-to-substantial reliability. Furthermore, this study revealed that even less experienced and non-imaging-specialized readers achieve good detection of failure mechanisms. MR swallowing exams, therefore, provide a reliable non-invasive method for the assessment of complaints in patients after surgical treatment of GERD.

GERD is a mechanical disorder caused by a dysfunction of the lower esophageal sphincter (LES), failed esophageal peristalsis, or gastric emptying disorder. It is, therefore, a failure of the antireflux barrier that exposes the esophagus to gastric contents, with a broad spectrum of symptoms and complications that range from heartburn to malignant disease [9, 10]. Treatment options in long-term follow-up studies for up to 10.6 years [11‐13] range from gastric acid suppression through proton pump inhibitors (PPI) to antireflux surgery. A recent study by Galmiche et al [14] reported that, with contemporary antireflux therapy for GERD, either by drug-induced acid suppression or by LARS, most patients remain in remission at 5 years. According to the guidelines of the Society of American Gastrointestinal and Endoscopic Surgeons, surgery should be offered to patients that (a) have failed medical treatment, (b) opt for surgery due to quality-of-life reasons, (c) have complications (e.g., Barrett’s esophagus), or (d) suffer from extra-esophageal manifestations of GERD (e.g., asthma) [15].

Although high satisfactory rates have been reported [2] in the literature, in up to 21% of the patients, fundoplication fails and recurrent symptoms, such as heartburn, dysphagia, or bloating, occur [3]. Those patients are subject to detailed multimodal examinations to elucidate the underlying cause. Furthermore, it has been shown that surgical success diminishes with each reoperation [4]. Compared with primary repair, revision surgery is associated with higher conversion rates to open surgery, longer operating times, and higher complication rates [15]. Thus, finding the reason for treatment failure is indispensable in the decision to reoperate or not and the planning of further treatment. This study supports the preliminary results of Kulinna-Cosentini et al [6], who showed that an MR swallowing exam offers a non-invasive method by which to visualize the postsurgical gastroesophageal situs and provides crucial anatomical information to the surgeon when considering a reoperation.

In comparison to fluoroscopy where the wrap can only be identified indirectly, MR imaging is the first radiological method that provides a direct view of the wrap itself. In the aforementioned preliminary study [6] with a small patient population, MR imaging yielded better results than endoscopy, which failed to diagnose the location of the wrap or even its disruption: Of the nine cases, five were misdiagnosed as recurrent hiatal hernia or newly acquired paraesophageal hernia, and, in four cases, there was a false diagnosis of reflux esophagitis. In the literature, endoscopy also shows a poor objective correlation with patient-reported symptoms, especially if the recurrent symptom is heartburn. Lord et al reported that no endoscopic abnormalities could be found in 62% of all symptomatic patients [16].

Although this study showed better results of MR imaging, imaging should not replace endoscopy. These methods are complementary in diagnosing the postoperative wrap situation.

As recurrent hernia is known to be the major cause of reoperation [4], it is not surprising that, in this study, 82.9% of the patients in the revision surgery group also had an hernia. An MR swallowing exam in consensus reading was correctly positive for hernia in 82.8% of cases, with five cases missed (17.2%). However, only 10.1% were missed in the non-surgical group. This might jeopardize surgery as the standard of reference, as results have been published, which indicate that high-resolution manometry is a reliable method for the detection of hernia. It has been hypothesized [17] that false-negative results on HRM may occur due to insufflation of gas during laparoscopy, thus exaggerating the number of hernias found at surgery.

For slipping and hernia, consensus terminology is important. According to the Hinder classification, there are four types of failure: disruption of the wrap (Hinder type I), stomach slippage above the diaphragm (Hinder type II), slipped Nissen (Hinder type III; the wrap slips down the stomach resulting in a telescope phenomenon with the gastric pouch above the level of the wrap), and transdiaphragmatic wrap herniation (Hinder type IV) [18]. In our institution, slipping is diagnosed if the stomach slips proximally through an intact fundoplication wrap (Hinder type III). However, recurrent hernia is diagnosed if there is displacement of gastric components above the level of the hiatus without regard to the status of the wrap (there can be both slipping and hernia). However, we are aware that various definitions of a “slipped Nissen” exist, including the transthoracic migration of the wrap [19], which is actually consistent with a Hinder type IV failure.

Slipping was found in a total of 11 subjects (seven cases in the surgery group) and correctly diagnosed in 10/11 patients (90.1%) by consensus. However, there was poor agreement, as no cases were detected correctly by all of the readers.

One explanation might be that a tube-like, malformed part of the gastric fundus was misinterpreted as the distal esophagus. Mostly, the so-called telescope phenomenon, or slipping, develops gradually if the cuff has been located too low and the intraabdominal part of the esophagus is not well mobilized during the first operation [20].

Wrap disruption has been reported in the literature in up to 12.8% of cases that undergo revision surgery [21]. In our study, reoperation revealed disruption in 28.6% of the cases. A correct diagnosis of wrap disruption was achieved by the more experienced reader (R2) of the radiologists and by R4 of the surgeons (93.3 and 86.7%, respectively), whereas R1 and R3 achieved a correct diagnosis in 73.4%.

A certain level of experience in reading MR swallowing seems to be important in the assessment of wrap disruption. Often, it can be difficult to differentiate between complete disruption and partial disruption, but clinical consequences in symptomatic patients are the main reason for reoperation. The LARS procedure performed in this study was Nissen fundoplication, where the gastric fundus is wrapped in a 360° fashion around the GEJ [22], resulting in a typical pseudo-ring-like appearance on MRI [6]. However, several variations of fundoplication exist (e.g., Toupet 270°) [23], which change the “normal” wrap presentation on MRI. Thus, it is crucial that the radiologist be familiar with the type of LARS performed to prevent reporting of a false-positive wrap disruption.

With a restored reflux barrier after fundoplication, dysphagia occurs in 3 to 17% [15]. Twenty-one of our patients (26.6%) were diagnosed with esophageal motility disorder on HRM, and an MR swallowing exam in consensus reading was able to correctly identify approximately half the cases (52.4%). However, there was only fair interrater agreement among the four readers, with overall k values of 0.200. HRM remains the absolute gold standard for motility disorders, but an MR swallowing exam may be able to identify causative morphologic changes, such as a too-tight wrap [6].

It is known that the main limitation of an MR swallowing exam is its low temporal resolution of two pictures/second, in contrast to videofluoroscopy with 25 frames/second [7, 24]. In accordance with the previous findings of Kulinna-Cosentini et al, where the sagittal oblique plane demonstrated the best visualization of the esophagus with a single-slice mean length of 16 cm on an MR swallowing exam [25], this also applied in our study.

This study has some limitations: Our main objective was to determine the interrater reliability. Since all diagnostic criteria were dichotomous variables (yes/no), agreement by chance in some cases is very likely. Otherwise, kappa statistics reflect chance-corrected agreement, which, on the other hand, may be excessive in dichotomous variables, resulting in too-low values. Second, in order to facilitate the reading, we did not make a distinction between partial or complete wrap disruption. However, it is very likely that symptomatic patients with suspected disruption—both partial and complete—will undergo revision surgery. Furthermore, it should be noted that swallowing a bolus of buttermilk consistency in the supine position does not simulate daily food intake and may cause false-positive findings in motility assessment. Moreover, it has been shown that including multiple water swallows and a solid meal test increase the diagnostic yield for HRM [26].

MR swallowing exams readily depict the major failure mechanisms of LARS and provide good reliability, even in non-experienced readers. It will be of great benefit to surgeons in considering and planning a reoperation and should be included in the preoperative workup for revision surgery after fundoplication.