Distinguishing Of Prolactinomas From Other Hyperprolactinemic Pituitary Adenomas Before Treatment

Object: It is significant but sometimes difficult to distinguish between prolactinomas and non-functioning adenomas before treatment, especially in patients with a pituitary macroadenoma. Therefore, the main objective of this study was to carry out two new indicators: PRL/MD and PRL/V, to compare the performances of PRL, PRL/MD and PRL/V for distinguishing between non-functioning pituitary macroadenomas(NFPMAs) and macroprolactinomas(MPs) before treatment and to assess if the two indicators could be proposed as useful diagnostic tools.Methods: 178 patients presenting with pituitary macroadenoma and hyperprolactinemia to the Department of Neurosurgery in Fuzhou General Hospital between December 2008 and December 2014 were studied. Patients were divided into two groups according to immunohistochemical results:(A) macroprolactinoma(n=86); and(B) other hyperprolactinemic macroadenomas(n=92). Performances of PRL, PRL/MD and PRL/V were compared by two comparative studies. The first comparative study included all cases with no PRL limit, and the second comparative study included the cases with PRL levels below 250 μg/L(a grey area associated with diagnostic uncertainty).Results: In the first comparative study, the analysis included a total of 178 patients presenting with pituitary macroadenoma and hyperprolactinemia. The best cutoff point of PRL in the distinguishing of macroprolactinomas from other hyperprolactinemic pituitary macroadenomas showed by ROC curve was 120.37μg/L, while the SE, SP, PPV, NVP and Youden index were 0.826, 0.913, 0.899, 0.848 and 0.793 individually; the best cutoff point of PRL/MD was 5.91μg/(L·mm), while the SE, SP, PPV, NVP and Youden index were 0.872, 0.93, 0.926, 0.887 and 0.807 individually; and the best cutoff point of PRL/V was 53.94 μg/(L·cm3), while the SE, SP, PPV, NVP and Youden index were 0.953, 0.946, 0.943, 0.956 and 0.899individually. Diagnostic accuracies of PRL, PRL/MD and PRL/V were all satisfactory, and the results had no statistical difference. In the second comparative study, the analysis included a total of 118 pituitary macroadenoma patients with the PRL level between the upper limit of normal and 250 μg/L. The best cutoff point of PRL showed by ROC curve was 55.65μg/L, while the SE, SP, PPV, NVP and Youden index were 0.800, 0.716, 0.857, 0.933 and 0.516 individually; the best cutoff point of PRL/MD was 4.03μg/(L·mm), while the SE, SP, PPV, NVP and Youden index were 0.800, 0.898, 0.727, 0.929 and 0.698 individually; and the best cutoff point of PRL/V was 54.00μg/(L·cm3), while the SE, SP, PPV, NVP and Youden index were 0.900, 0.966, 0.900, 0.966 and 0.866 individually. Diagnostic accuracy of PRL/MD tended to be higher than it of PRL, although the results didn’t have statistical difference(p = 0.097), the diagnostic accuracy of PRL/V, however, was significantly higher than it of PRL(p=0.028).Conclusion: We suggested that PRL/MD and PRL/V were both helpful indicators in the distinguishing of macroprolactinomas from other hyperprolactinemic pituitary macroadenomas before treatment. All of PRL, PRL/MD and PRL/V were accurate indicators. However, PRL/V was more accurate than PRL, especially in the hyperprolactinemic pituitary macroadenoma patients with a PRL level between the upper limit of normal and 250 μg/L. 54.00 μg/(L·cm3) was the best cutoff point of PRL/V.