Clinical Significance of MRI and Pathological Features of Giant Cell Tumor of Bone Boundary.

OBJECTIVE: To find new clues to reduce postoperative recurrence after intralesional curettage by studying MRI and pathological features of giant tumor of bone (GCTB) boundaries. METHODS: A retrospective study was performed in the departments of orthopaedic surgery and medical imaging at our hospitals from January 2006 to August 2016. A total of 16 GCTB patients confirmed by pathology were asked to participate in the present study. The age range was from 18 to 44 years (9 women and 7 men). All patients underwent MRI examination. All patients underwent en bloc resection and complete postoperative tumor segments were obtained. Five specimens were obtained randomly at the place of the segments where the GCTB boundary showed different types on MRI. Ordinary HE staining was used for all specimens and we measured the depth of local tumor cell infiltration (240 measurements). Results were expressed as means ± standard deviation. Statistical analyses were carried out with one-way ANOVA and the Student-Newman-Keuls test. P < 0.05 was considered statistically significant. The kappa test was used to analyze the degree of agreement of observers. RESULTS: A total of 16 patients (median age 30.56 years; range, 18-44 years) with GCTB (the number of distal femurs and proximal tibias was 9 and 7, respectively) were tested. The boundaries of all GCTB cases were composed of clear boundary, relatively clear boundary, and blurred boundary in different proportions on MRI. Based on continuous observation of all MRI, all boundaries were incomplete. The kappa value between two radiologists and two pathologists was 0.91 and 0.88, respectively. The average depth of local tumor cell infiltration in the clear boundary, relatively clear boundary, and blurred boundary groups was 0.42 ± 0.11 mm, 2.85 ± 0.21 mm, and 4.83 ± 0.12 mm, respectively. There was statistical difference among the three groups (F = 17.62, P < 0.05). There was also statistical difference between each of the two groups (q-value was 8.95, 14.28, and 5.21, respectively, P < 0.05). The depth of local tumor cell infiltration with blurred boundaries on MRI was the largest and the depth with clear boundaries was the smallest. CONCLUSION: The intralesional curettage boundaries need to be expanded on the basis of different types of boundaries provided by MRI.