Recent studies estimate that the prevalence of pituitary adenomas is approximately 1/1500 persons. Pituitary tumours are usually considered as benign. However, local invasion is reported in 35-40% of pituitary adenomas; resistance to medical treatment or recurrence leading to multimodal therapy is reported in about 15% of cases. These tumours are considered as aggressive pituitary tumours and present a distinct biological and clinical entity with continued growth despite multimodal therapy, including surgery and radiotherapy (McCormack et al., 2011). Whilst these tumours have malignant potential, the term of pituitary carcinoma is strictly reserved for those rare tumours (0.2%) with demonstrated craniospinal or systemic metastases (Heaney, 2011).

Pituitary aggressive and malignant tumours are very difficult to control and ultimately prove to be lethal. It was suggested that early aggressive treatments (chemotherapy, radiotherapy) may control progression and occurrence of metastases. However, these therapeutic options are associated with important side effects limiting their use and the prediction of pituitary tumor behaviour remains a challenge. At the diagnosis, clinical signs are not specific and the results concerning proliferative factors (Ki-67 and P53), putative oncogenes (PTTG) conflict from one series to another.

In a case-control retrospective study of a cohort of 410 patients (HYPOPRONOS), we validated a prognostic pathological classification based on histological and radiological data (J. Trouillas 2012 in preparation). Tumours were classified into 3 grades: grade 1= non-invasive tumour, grade 2= invasive tumour and grade 3 = aggressive-invasive tumor with the combination of radiological signs of invasion and 2 of 3 signs of increased proliferation (Ki-67 index>3%, number of mitoses>2 per 10 fields at 400X, P53 nuclear detection).

It is now widely accepted that cancer is a clonal disease, which arises from a single normal cell and progresses thanks to the accumulation of DNA alterations (Sanson et al., 2011). To identify the role of these DNA alterations, we conducted array CGH analysis limited to 13 prolactin pituitary tumours, from frozen fragments, and identified allelic loss of chromosome 11 associated with aggressiveness and malignancy (Wierinckx et al., 2011).

To confirm these encouraging results we propose to conduct a study on a large series of tumours, fixed and embed, and to be correlated the results to clinical data.

Inclusion Criteria:

• Only patient with complete clinical, radiological and hormonal data available during yearly follow-up will be included.
• Preoperative MRI will be used to classify the tumour as invasive, and postoperative MRI will be collected to confirm recurrence or progression of the tumour.
• Presence of tumour fragments fixed in Holland-Bouin's fluid or Neutral Buffered Formalin fixative available for aCGH analysis.

Exclusion Criteria:

• Patient who underwent systematic post-operative radiotherapy.
• Patient presenting Multiple Endocrine Neoplasia type 1 (MEN1) or aryl hydrocarbon receptor interacting protein (AIP) mutation since mechanism of tumorigenesis are different to sporadic pituitary tumours.

Patients presenting a pituitary tumour, including PRL, GH, ACTH and LH/FSH, operated on by transsphenoidal route between 1990 and 2008 with at least 5 years of follow-up