An interesting study from Italy looking at an anterior skull base classification that may be useful in predicting risk of intraoperative CSF leak during FESS surgery. Traditionally we have used the Keros classification system, developed in 1962 to categorise olfactory cleft depth of the cribriform plate as an indicator of risk for skull base penetration during FESS surgery. This technique identifies the lateral lamella of the cribriform plate as a likely area of risk and that the angle formed with the horizontal may predict risk of injury. This has been termed the Gera classification. Class 1 angle >80 degrees is low risk, Class 2 angle 45-80 degrees, medium risk and Class 3 angle <45 degrees, high risk. A retrospective analysis of CT scans of patients with an iatrogenic CSF leak was done: 6000 FESS patients from four centres over four years had 24 CSF leaks due to penetration of the lateral lamella of the cribriform plate. The 24 scans were selected and compared to 100 randomly selected from the 6000 patients. CT scans without contrast at 1.5mm axial sections had been performed and any which were rotated by more than five degrees were excluded. The scans were reconstructed and measurements taken to determine the Keros and Gera classifications. Statistically there was no difference between the two groups of patients in terms of age or sex. Right and left iatrogenic injuries occurred equally and there were no bilateral cases. In the CSF leak group, the mean depth of the cribriform plate was 4.3+/-1.7mm on the leak side and 4.5+/-1.6mm on the contralateral side which was not significantly different. The most common classification was Keros 2. In the control group, the mean depth was 5.2+/-1.7mm on the right and 5.5+/-1.6mm on the left, again no difference. The most common classification was again Keros 2. There was no significant difference between the measured depths or the Keros classification between the CSF leak group and the control group. In the CSF leak group, the mean angle between the lateral lamella of the cribriform plate and the horizontal was 41.2+/-10.3 degrees versus 50.1+/-10.4 on the contralateral side, which was a significant difference. Using Gera, the most common classification on the leak side, was Class 3, 19 patients (79%). On the contralateral side it was Class 2, 14 patients (58%) - a significant difference. In the control group, the mean angle on the right was 71.7+/-12.4mm and on the left 71.1+/-11.2 degrees; not significantly different. Significant differences between the two groups occurred at the angle of the LLCP and also the Gera classification. The sensitivity, specificity and positive and negative predictive values in this group were better with the Gera classification rather than Keros. This seems a good and straightforward adjunct to the preoperative CT scan assessment, I’ll certainly be adding the angle of the lateral lamella on coronal imaging to my checklist.