A predictive algorithm for evaluating elevated serum prolactin in patients with a sellar mass.

Area of Science:

• Endocrinology
• Neurosurgery
• Oncology

Background:

• Hyperprolactinemia presents diagnostic challenges, stemming from either prolactinoma or pituitary stalk compression by a sellar mass.
• Accurate diagnosis is crucial for determining appropriate treatment, such as dopamine agonist therapy or surgical resection.

Purpose of the Study:

• To develop a predictive algorithm for diagnosing prolactinoma in patients with elevated serum prolactin and a sellar mass.
• To differentiate prolactinoma from other causes of hyperprolactinemia to guide clinical management.

Main Methods:

• A case-control study analyzed 177 patients with prolactinomas and 87 controls with non-endocrine secreting sellar masses.
• Classification and regression tree (CART) analysis identified significant predictive variables including demographic data, symptoms, lab values, and imaging findings.
• A subset with mildly elevated prolactin (25-125 ng/ml) was analyzed separately.

Main Results:

• Prolactinomas showed significantly higher mean preoperative prolactin levels (858 ng/ml) compared to controls (17.57 ng/ml).
• CART analysis identified preoperative prolactin (>41.5 ng/ml), age (<40.5 years), and size (<17 mm) as key predictors of prolactinoma, with a 7.9% misclassification rate.
• In patients with mildly elevated prolactin, tumor size (<2.5 cm) and prolactin level (>40 ng/ml) predicted prolactinoma in 98.6% of cases.

Conclusions:

• A simple predictive model incorporating prolactin levels, age, and tumor size effectively aids in diagnosing prolactinoma.
• This algorithm assists in identifying patients who may benefit from surgical intervention.
• The findings support a more streamlined diagnostic approach for hyperprolactinemia in the presence of sellar masses.