Endocrine Abstracts

Introduction: Immunohistochemistry (IHC) of somatostatin receptor subtype 2 (SST₂) is a predictive factor for first-generation somatostatin receptor ligand (fg-SRL) response in acromegaly patients. A semi-quantitative immunoreactivity score, IRS, has been proposed as the gold-standard to evaluate SST₂ IHC expression. Recently, our group developed a quantitative method to determine SST₂ expression using an open-source digital image analysis (DIA). We aimed to validate the DIA on both SST₂ and SST₅ in a new cohort of GH-secreting pituitary tumors. We then correlated fg-SRL response with SST expression, evaluated with both IRS and DIA methods.

Material and Methods: SST₂ and SST₅ expression was assessed in paraffin-embedded tissues from 42 GH-secreting pituitary tumors, using both IRS and DIA. The DIA software calculates the staining intensity (intensity/area) and the percentage of positive cells (%PC-DIA) based on four representative images. The IRS was independently performed by two researchers. Correlations were performed evaluating the ‘total’ receptor expression (IRS vs intensity/area) and the %PC (%PC-IRS vs %PC-DIA). GH and IGF-1 data were collected at baseline after surgery and following 6-month fg-SRL treatment.

Results: Mean SST₂ IRS was 6.41±3.35, mean DIA intensity/area was 0.17±0.14, and mean %PC-DIA was 63.42±29.67%. As concerns SST₅, mean IRS was 4.66±3.26, mean DIA intensity/area was 0.07±0.09, and mean %PC-DIA was 38.27±38.84%. A good correlation was observed between DIA and IRS for ‘total’ receptor expression (rho=0.924 and rho=0.872, for SST₂ and SST₅, respectively), as well as for the %PC (rho=0.649 and rho=0.748, all P<0.0001). Twenty-four out of 42 patients (57%) were treated with fg-SRLs after surgery. A significant positive correlation was observed between GH decrease and SST₂ expression, quantified with DIA (intensity/area: rho=0.707, %PC: rho=0.625, P<0.005) and total IRS (rho=0.655, P=0.017), but not with the %PC-IRS (rho=0.356, P=0.123). Similarly, a significant positive correlation was observed between IGF-1 xULN decrease and SST₂ expression for both DIA parameters and total IRS (rho ranging from 0.517 to 0.617, P<0.05), but not for %PC-IRS (rho=0.235, P=0.334). No correlation was observed between GH, IGF-1 xULN decrease and SST₅ expression.

Conclusion: The DIA is a reliable quantification method to assess both SST₂ and SST₅ expression. SST₂, but not SST₅, expression correlated with GH and IGF-1 decrease following fg-SRL treatment. This correlation was statistically significant using both IRS and DIA when ‘total’ SST₂ expression was assessed. The %PC correlated with treatment response only when evaluated using the DIA, thus showing the superiority of this quantitative method.