Abstract
OBJECTIVE: This cross-sectional study aimed to evaluate the correlation between the necrotic index and the modified necrotic index with the Harris hip score (HHS), as well as to assess their interobserver and intraobserver reliability in patients with osteonecrosis of the femoral head (ONFH). METHODS: Fifty-four adult patients (88 hips) presenting with atraumatic hip pain and radiographic evidence of ONFH (Ficat-Arlet grades 1 and 2) were included. Participants abstained from analgesics for one week prior to the study. All underwent MRI, and HHS was calculated by a resident uninvolved in the study. The necrotic index was determined using angles obtained from mid-coronal and mid-sagittal MRI slices, whereas the modified necrotic index was derived from angles measured in MRI slices where the necrotic lesion appeared at its largest size. Three senior consultants performed measurements on two separate occasions to assess reliability. RESULTS: Both indices showed significant negative correlations with HHS. The modified necrotic index demonstrated a strong correlation (r = -0.940, p < 0.001), whereas the original necrotic index exhibited a moderate correlation (r = -0.790, p < 0.001). The interclass correlation coefficient (ICC) for assessment of lesion size was significant for both indices (p < 0.05), but it was higher for the modified index (ICC = 0.81) compared to the original index (ICC = 0.69). Levene's test for homogeneity of variance indicated unequal variance across the two methods (p < 0.01). Measurement variance was smaller for the modified index than for the original index (p < 0.01), indicating greater precision. Both methods showed significant interobserver and intraobserver agreement (p < 0.05) Conclusion: Both the necrotic index and the modified necrotic index showed significant negative correlations with HHS. The modified necrotic index demonstrated a strong correlation, whereas the original necrotic index exhibited a moderate correlation. Both methods proved reliable for measuring lesion size, with the modified method showing greater precision. Additionally, both indices demonstrated significant interobserver and intraobserver reliability.