Utility of Pelvis CT to Assess Occult Intertrochanteric Extension of Greater Trochanteric Fractures.

Objectives: To determine whether computed tomography (CT) differences of bone density between the injured and noninjured femora in patients with greater trochanteric fractures can be used to identify intertrochanteric extension (ITE).

Methods:

Design: Retrospective cohort series.

Setting: Multihospital academic institution.

Patient selection criteria: Included were patients for a 7-year period (January 2014-December 2021) with greater trochanteric fractures (OTA/AO 31A1.1) without evident ITE on CT that also underwent pelvis magnetic resonance imaging (MRI) to assess for occult ITE.

Outcome measures and comparisons: The primary outcome measures were CT findings of intertrochanteric (IT) curvilinear density (CL) and subtrochanteric bone density (ST) in the injured femur in patients with greater trochanteric fractures. CT findings (CL and ST) were compared with the patient's MRI, which was the reference standard for occult IT fractures. The MRI determined presence of occult intertrochanteric extension and, if present, MRI determined fracture extension into the IT region were categorized as (1) less than 50% or (2) 50% or greater. Descriptive statistics, sensitivity, specificity, and inter-rater reliability were calculated assessing the presence of the CT findings of CL and ST, compared with reference standard MRI ITE. Sensitivity and specificity for CL and ST were calculated for (1) any degree of MRI ITE (<50% and ≥50%) and (2) only MRI ITE 50% or greater.

Results: Eighty-one patients (54 women, 27 men, mean age 82 years, range 54-102) were included. Fourteen (17%) patients had no MRI ITE, 11 (14%) patients had <50% MRI ITE, and 56 (69%) patients had ≥50% MRI ITE. The presence of CL on CT corresponded to any MRI ITE (<50% and ≥50%) with sensitivity of 55.2%, specificity 100%, positive predictive value (PPV) 100%, and negative predictive value (NPV) 31.8%. In patients with MRI ITE 50% or greater only and CL presence, specificity was 92%, and sensitivity was 62.5%. Presence of ST on CT was associated with any MRI ITE with sensitivity of 34.3%, specificity 100%, PPV 100%, and NPV 24.1%. Patients with MRI ITE 50% or greater and ST presence, specificity was 96% and sensitivity was 39%.

Conclusions: In patients with apparent isolated greater trochanteric fractures, the presence of curvilinear IT and subtrochanteric density in the medullary bone in the injured femur on pelvis CT was highly predictive of ITE. Patients with these CT findings in the injured femur on pelvis CT can be assumed to have ITE and treated accordingly, obviating the need for MRI. The absence of the curvilinear IT and subtrochanteric densities did not rule out possible ITE, and MRI can be further considered in this population.

Level of evidence: Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.