Radiological evolution of autograft fat used for skull base reconstruction after transsphenoidal surgery for pituitary adenomas.

Cerebro-spinal fluid leak after transsphenoidal surgery for pituitary adenomas may be prevented by skull base reconstruction with fat autograft. However, graft changes may interfere with the interpretation of postoperative images. Our aim is to describe the radiological evolution of the fat autograft. A retrospective analysis was performed, including patients undergoing a transsphenoidal surgery for pituitary adenomas with a fat autograft for skull base reconstruction. Clinical and radiological data were collected, with assessment of fat autograft and extent of resection. Statistical analysis was performed using Kruskal-Wallis and Wilcoxon signed-rank test while Spearman's Rho was used to analyze the relationship between variables. Seventy-two patients were included. Macroadenomas were diagnosed in 62 cases (86.1%) and in 21 cases an invasion of the cavernous sinus was described (29%). Gross total resection was achieved in 84.7% of cases. The volume of the fat graft significantly decreased between 3 months and 1 year after surgery (p = 0.01) and between 1 year and the last follow-up (mean 4.63 years, p < 0.01). Fat signal ratio significantly diminished between 3 months and 1 year in unenhanced and enhanced T1-weighted sequences (p = 0.04 and p = 0.02 respectively). Volume reduction was related to the decrease in signal ratio in unenhanced T1 sequences (p = 0.008). Fat resorbs with time: almost 50% of the fat volume is lost during the first year after surgery and 60% is resorbed at 4.6 years. T1-signal, before and after gadolinium injection, also decreases during the first year, probably because of the progressive fibrosis of the graft. This information will contribute to the interpretation of postoperative images