Gastroesophageal junction-to-incisura radiographic index as a predictor of GERDQ-defined reflux-related symptoms three years after laparoscopic sleeve gastrectomy: a diagnostic test accuracy study

Anatomic changes induced by laparoscopic sleeve gastrectomy (LSG) have been implicated in gastroesophageal reflux disease (GERD) and reflux-related symptoms following surgery. We evaluated the diagnostic accuracy of an imaging-derived metric for identifying GERDQ-defined reflux-related symptoms at three years after LSG. This retrospective, single-center diagnostic test accuracy study included 59 patients who underwent LSG between 2015 and 2018. Based on a simplified Bernoulli-inspired model, the index test was defined as the fourth-power ratio of radii at the gastroesophageal junction and incisura, I=(r_GE/r_incisura)^4. It was measured on contrast swallow studies obtained within two weeks after surgery by an investigator blinded to the three-year GERDQ outcome. A GERDQ score ≥ 8 at three years was used as the symptom-based reference standard; no routine postoperative endoscopy or ambulatory pH-impedance monitoring was performed. A candidate threshold was selected in the same cohort using the Youden index. Mean age was 36 years, 41 of 59 patients (69.5%) were female, and mean body mass index was 40.3 kg/m². The area under the receiver operating characteristic curve was 0.82 (95% CI 0.72–0.93). Thirty-five patients had an index < 7.25 and 24 had an index ≥ 7.25; mean (SD) GERDQ scores were 8.1 (1.7) and 6.1 (0.6), respectively. The GERDQ threshold was met by 20 of 35 (57.1%) versus 0 of 24 (0%) patients, respectively. At this candidate cut-off, sensitivity was 100% (95% CI 83.2%–100%), specificity 61.5% (44.6%–76.6%), positive predictive value 57.1% (39.4%–73.7%), negative predictive value 100% (85.8%–100%), and accuracy 74.6% (61.6%–85.0%). In this small retrospective cohort, the radiographic index showed apparent discrimination of GERDQ-defined reflux-related symptom status. It was not validated against objective reflux testing, and selecting and evaluating the threshold in the same cohort may have produced optimistic performance estimates. Independent prospective validation and multi-reader reproducibility testing are required before clinical use.