A Modified Approach to Measuring Femoro-Epiphyseal Acetabular Roof (FEAR) Index Has Better Intraobserver and Interobserver Reliability Compared With the Original FEAR Index

Abstract

PurposeTo propose a modified approach to measuring femoro-epiphyseal acetabular roof (FEAR) index while still abiding by its definition and biomechanical basis, and to compare the intra- and inter-observer reliabilities of the original and the modified FEAR index. To propose a classification for medial sourcil edges. MethodsWe retrospectively reviewed a consecutive series of patients treated with periacetabular osteotomy and/or hip arthroscopy at a single institute. Patients with unilateral or bilateral symptomatic borderline hip(s) were included. Hips with remarkable osteoarthritis, deformities, history of previous surgery, or without symptoms were excluded. A modified FEAR index was defined using a best-fit circle to determine the sourcil line and two ancillary lines connecting femoral head and sourcil edges to determine epiphyseal line. Lateral center-edge angle, Sharp’s angle, Tonnis angle on all hips, as well as FEAR index with original and modified approaches were measured. Intra- and inter-observer reliability were calculated as intraclass correlation coefficients (ICC) for FEAR index with both approaches and other alignments. A classification was proposed to categorize medial sourcil edges. ICC for the two approaches across different sourcil groups were also calculated. ResultsAfter reviewing 411 patients, 49 were finally included. Thirty-two patients (40 hips) were identified as having borderline dysplasia defined by an LCEA of 18 to 25 degrees. Intra-observer ICC for the modified method were good to excellent for borderline hips; poor to excellent for DDH; moderate to excellent for normal hips. As for inter-observer reliability, modified approach outperformed original approach with moderate to good inter-observer reliability (DDH group, ICC=0.650; borderline dysplasia group, ICC=0.813; normal hip group, ICC=0.709). The medial sourcil edge were classified to 3 groups upon its morphology. Type II (39.0%) and III (43.9%) sourcils were the dominant patterns. The sourcil classification had substantial intra-observer agreement (observer 4, kappa=0.68; observer 1, kappa=0.799) and moderate inter-observer agreement (kappa=0.465). Modified approach to FEAR index possessed greater inter-observer reliability in all medial sourcil edge patterns. ConclusionThe modified FEAR index has better intra- and inter-observer reliability compared with the original approach in all hip groups and sourcil groups. Type II and III sourcils accounts for the majority to which the modified approach is better. Level of EvidenceLevel II, Development of diagnostic criteria (consecutive patients with consistently applied reference standard and blinding)