A retrospective cohort study assessing geometrical parameters related to visual deterioration in pituitary macroadenoma.

PURPOSE: We aim to identify specific parameters that will allow identification of those patients at greater risk of vision loss, specifically variables such as geometry, volume and size of the pituitary tumour.

METHODOLOGY: Volume measurements were made with both manual (using the formula width × height × length × 0.5) and video size, sub-voxel interpolation using Brainlab software. The shape of the tumour was divided into four categories, ovoid, dumbbell, pyramidal and other. Chiasmal compression was divided into central and lateral compression and the chiasm location was classified based on pre/post and normal fixed chiasm in relation to the tuberculum sella. We used a visual impairment score ranging from (0-100), which was then further subdivided in to four grades: Grade I - 0-25 (mild); Grade II - 26-50 (moderate); Grade III - 51-75 (severe); Grade IV - 76-100 (subtotal or complete).

RESULTS

Univariate analysis of variables related to vision loss found tumour volume, suprasellar growth and lateral chiasmal compression related to pre-operative vision loss. On multivariate regression analysis tumour volume and lateral chiasmal compression remained significant (p < 0.008). In relation to post-operative vision, univariate analysis indicated that variables such as giant macroadenoma, pre-operative visual impairment score and suprasellar growth were significant. Multivariate analysis showed pre-operative visual impairment and suprasellar growth as the only statistically significant variables.

CONCLUSION

Our ROC analysis suggests tumours with suprasellar extension 16 mm (sensitivity 85 %, specificity 67 %) in sagittal plane are more likely to result in moderate to severe grade vision loss (VIS grade ≥ 2).