Quantifying the subjective: length-to-height ratio characterizes achalasia esophageal tortuosity.

Abstract

Esophageal morphology in achalasia is thought to affect outcomes, with "end-stage" sigmoidal morphology faring poorly; however, evaluation of morphology's role in outcomes has been limited by lack of objective characterization. Hence, the goals of this study were twofold: characterize the variability of timed barium esophagram (TBE) interpretation and evaluate an objective classification of TBE tortuosity: length-to-height ratio (LHR). We hypothesized that the esophagus must elongate to become sigmoidal such that sigmoidal morphology would demonstrate a larger LHR. Ninety pre-operative TBEs were selected from an institutional database. Esophageal morphology was categorized as straight, intermediate, or sigmoidal. Esophageal length was measured by a mid-lumen line from the aortic knob to the esophagogastric junction on TBE; height was measured vertically from the aortic knob to the level of the esophagogastric junction. The length divided by the height generated the LHR. Descriptive statistics and frequency of expert agreement were calculated. Median LHR was compared between consensus morphologies. A receiver operating characteristic (ROC) determined the optimal LHR for sigmoidal vs non-sigmoidal characterization. From a total of 90 pre-operative TBEs, expert consensus morphology was reached in 56 (62.2%) cases. Pairs of experts agreed on morphology in 62-74% of TBEs, with all three experts agreeing on 46.7-48.9% of cases. Median LHR between expert consensus morphologies was 1.03, 1.09, and 1.24 for straight, intermediate, and sigmoidal morphologies, respectively (p < 0.001). ROC demonstrated that an LHR cutoff of 1.13 was 100% sensitive and 95% specific (AUC 0.99) for ruling out sigmoidal morphology. These findings confirm our anecdotal experience that subjective morphology interpretation is variable, even between experts at a high-volume center. LHR provides an objective method for classification, allowing us to overcome the limitations of inter-observer variability, thus paving the way for future study of the role of morphology in achalasia outcomes.