Development of a cost-effective diagnostic algorithm incorporating transcription factor immunohistochemistry in the evaluation of pituitary tumours

Abstract

Purpose: To determine the utility of the 2022 WHO Classification of pituitary tumours in routine clinical practice and to develop an optimal diagnostic algorithm for evaluation of tumour type in a real-world setting. Methods: Retrospective evaluation of pituitary tumour immunohistochemistry (IHC), operatively managed at St Vincent’s Hospital Sydney, between 2019 and 2021. Routine IHC comprised evaluation of transcription factors [steroidogenic factor 1 (SF1), T-box transcription factor 19 (TPIT) and pituitary-specific positive transcription factor (PIT1)] and anterior pituitary hormones. Three tiered algorithms were tested, in which hormone IHC was performed selectively based on the initial transcription factor results. These were applied retrospectively and compared with current practice ‘gold standard’ comprising all transcription factor and hormone IHC. Diagnostic accuracy and cost were evaluated for each. Results: There were 113 tumours included in the analysis. All three algorithms resulted in 100% concordance with the ‘gold standard’ in the characterisation of tumour lineage. While all three were associated with relative cost reduction, Algorithm #3, which omitted hormone IHC in the setting of positive SF1 or TPIT and performed IHC for growth hormone, prolactin and thyroid stimulating hormone only in the setting of PIT1 positivity, was the most cost-efficient. Additionally, there were 12/113 tumours with no distinct cell lineage. Conclusion: A diagnostic algorithm omitting hormone IHC except in cases of PIT1 positivity is an accurate and cost-effective approach to diagnose the type of pituitary tumour. A significant subgroup of pituitary tumours with no distinct cell lineage, frequently plurihormonal, remains difficult to classify with the new WHO criteria and requires further evaluation.