In this study, we observed discrepancies in the JIC classification between radiographic and MRI evaluations in 14.9% of ONFH cases. Cases showing discrepancies in type classification demonstrated a more favourable prognosis for femoral head collapse. It was found that the extent of the AP necrotic regions was associated with these discrepancies, and the cutoff value for the AP necrotic regions that aided the identification of discrepancies was 66.9% (Fig. 5).

Osteonecrosis of femoral head of a 28-year-old man with discrepancy in the Japanese Investigation Committee type classification between radiography and magnetic resonance imaging (MRI). (A) The central coronal section of the femoral head on MRI T1-weighted image, it is classified as type C1. (B) The anteroposterior (AP) radiographic view of the femoral head, it is classified as type C2. (C) The lateral radiographic view. The extent of necrotic area is located anteriorly, and the extent of the AP necrotic region is 60%. (D) AP radiograph obtained 5 years after the first visit shows no collapse progression of the right femoral head

The JIC classification system proposed in Japan, is now used worldwide to classify ONFH. Takashima et al. reported that compared to the Kerboul and Steinberg classifications, the JIC classification has smaller interobserver variability and is a highly versatile and effective classification system [9]. Although the evaluation method involves the use of either AP radiographs of the hip joint or MRI coronal sections at the central level of the femoral head, there is no defined priority between these imaging modalities. Depending on the localization of necrosis, discrepancies may arise in the decision of type classification between radiography and MRI, potentially leading to misjudgment of the collapse risk. The results of this study showed discrepancies in the JIC-type classification between radiography and MRI in 28 (14.9%) of 188 joints. Although type classification based on plain radiography is a highly versatile approach, surgeons should be aware of the possibility of discrepancies with the MRI-based classifications.

Most large-scale studies on survival by type in the JIC classification have been based on MRI. Asada et al. reported collapse rates at 36 months for each JIC type as follows: type A, 0%; type B, 10.8%; type C1, 48.5%; and type C2, 76.7% [17]. Similarly, Kuroda et al. reported five year collapse rates for types A (0%), B (7.9%), C1 (36.6%), and C2 (84.8%) [18]. Our results showed that patients who showed discrepancies between radiography and MRI exhibited significantly better survival rates, with a collapse of > 3 mm as the endpoint, than those without discrepancies. Therefore, for cases showing discrepancies, MRI-based classification may be more appropriate than radiography-based classification for evaluating survival rate.

Necrotic lesions of ONFH are generally located anteriorly [12]. Furthermore, the extent of AP necrosis in ONFH has been shown to significantly affect prognosis. Osawa et al. reported that the extent of AP necrotic regions in the femoral head was associated with the cessation of femoral head collapse [19]. They found that the prognosis was significantly better with collapse progression as the endpoint when the AP necrotic region on MRI was < 62.1%. Ikemura et al. investigated the posterior boundary of necrotic lesions in ONFH and reported that cases with progressive collapse had significantly larger necrotic angles in the axial and oblique axial slices on MRI than cases with non-progressive collapse [20]. In our study, we found that the cutoff value for the AP necrotic region on lateral hip radiographs for predicting discrepancies in type classification between radiographs and MRI was 66.9%. Cases with AP necrotic regions < 66.9% on lateral radiographs may be overestimated in the type classification and should be further evaluated using MRI.

This study has several limitations. First, this was a retrospective study. A prospective investigation is necessary to validate our findings regarding the progression of collapse. Second, our study was limited to type C radiographs and may not be applicable to other types. Furthermore, the survival rates obtained in this study did not consider risk factors for collapse, such as necrotic size or volume. Additionally, the cutoff value obtained in this study was used to predict discrepancies in type classification between radiography and MRI, but it did not directly predict prognosis. Nevertheless, this study had several strengths. MRI cannot be performed in all cases in clinical practice. The threshold of the AP necrotic region in the lateral hip radiograph identified in this study can be used to identify cases that should be evaluated using MRI. Therefore, the results of this study may help determine treatment strategies for ONFH.

In conclusion, there were discrepancies in the JIC-type classification between radiography and MRI in 14.9% of ONFH type C cases. Patients who had with discrepancies demonstrated significantly better survival rates with femoral head collapse > 3 mm as the endpoint than those without discrepancies. The cutoff value for the AP necrotic region on lateral hip radiographs that aided the identification of discrepancies in type classification between radiography and MRI was 66.9%. These results may be useful in determining a treatment plan for ONFH.