Impact of probe choice on diagnostic performance of vibration controlled transient elastography

Background and aims Non-invasive assessment of fibrosis and steatosis is widely used. Especially vibration controlled transient elastography (VCTE) with liver stiffness measurement (LSM) and controlled attenuation parameter (CAP) has been evaluated in many histology-controlled studies. International guidelines recommend the M probe for leaner and the XL probe for obese patients. The probe choice is based on the patient’s anthropometry and can be made either user-dependent or based on an automated probe selection tool (AUTO). The a im of this study was to evaluate AUTO in comparison to skin-to-liver-capsule distance (SLCD), focussing on the effect of whether the M or XL probe is opted for first. Methods VCTE (FibroScan 530, Echosens, France) was performed in a prospective cohort with chronic liver disease. All examinations were performed with M and XL probe. The choice of the first probe was randomized and the resulting probe recommendation (AUTO) was recorded. In addition, the SLCD at the LSM site was determined. A SLCD >25 mm indicated correct use of the XL probe. Results 200 patients (58 % female, age 56 ± 14 years) were included. 34 %, 33 %, 33 % were normal weight, overweight and obese, respectively, and 37 % had NAFLD/NASH. Three patients had an invalid VCTE measurement (SLCD). Using the SCLD criterion, 163 and 37 patients required use of M and XL probe, respectively. AUTO resulted in a “M probe recommendation” of 184 and 133, depending on the initial user-dependent probe choice (M/XL), see figure. Three to five cases were indeterminate according to AUTO. Compared to SLCD, there was a discrepancy in 23 patients using the M probe and 34 using the XL probe, p = 0.12. CAP disagreed with the SLCD-based value by > 10 dB/m in 10 patients when beginning with the M probe, but in 30 patients when beginning with the XL probe, p < 0.001. A different LSM risk category (> 7.9 or > 9.6 kPa) was only found in 1 to 3 patients. Conclusions The correct choice of the VCTE probe is user-dependent even if automated probe selection is applied. The arbitrary choice of which probe to try out first can affect the “correct” final result, especially for CAP. An improved standardized probe selection algorithm should be established for further studies and clinical application. Publication History Article published online: 08 September 2020 © Georg Thieme Verlag KG Stuttgart · New York