Gastro-intestinal graft versus host disease (GVHD) is a major source of morbidity and mortality amongst allogenic hematopoietic stem cell transplantation (Allo-HSCT). The diagnosis is based on histological findings that involve colonic biopsies with a risk of bleeding, especially in case of thrombocytopenia. Moreover the diagnosis is frequently made at a clinical stage of the disease, after the appearance of gastro-intestinal symptoms. Endo-microscopy is a novel endoscopic technique that allow "optical biopsies" during a conventional endoscopy and has proved its efficiency in several indications. In a pilot study the investigators showed that it had good sensibility and sensitivity compared to histology as gold standard. Therefore this study aim to identify endo-microscopic criteria allowing the early diagnosis of GHVD before its clinical manifestations.

The investigators propose a mini-invasive alternative technique to perform biopsies in the terminal part of the colon with rectosigmoidoscopy alone. Confocal laser endomicroscopy, either using an endoscope-based or a probe based technology , aims to decrease the number of standard biopsies and their associated risks, by providing real-time in situ microscopy. In a previous pilot study performed in 15 patients at the University hospital of Nantes, the investigators suggested that combining probe-based confocal laser endomicroscopy (pCLE) and wireless capsule endoscopy (WCE) could detect early mucosal abnormalities suggestive of GVHD and thus be part of a mini-invasive algorithm for early detection of GI-GVHD. However, in this study, the investigators showed that pCLE was a sensitive method for the early diagnosis of GVH with a sensibility of 85% and a specificity of 71.5%. These interesting results are however to put in question due to several reasons. First of all the pCLE criteria used to define a GVHD in this study were established with data stemming of literature in other validated indications such as the MIAMI score for the diagnosis of malignant biliary structures and therefore can turn out not to be as relevant as expected. Secondly the interpretations of the images obtained with pCLE were only interpreted by a single endoscopist who by the way wasn't blinded from the macroscopic appearance of eventual lesions during rectosigmoidoscopy resulting in a bias. Thirdly the number of patient as in a pilot study was probably not sufficient to evaluate the real sensibility and sensitivity of those new indication and technique for the early diagnosis of GVH. All together these observations lead us to the necessity of developing the promising results obtained previously and allow their realization by the creation of an innovating pCLE score : the E-mage score. The different criteria of the score will be established after the one defined in the pilot study reviewed by an expert committee combining anatomo-pathologists and endoscopists to be as objective as possible with quantitative and semi-quantitative assessments focusing particularly on crypt architecture, morphologic and dynamic criteria and fluorescein leakage with a cut off standing as no GVHD, likely GVHD or certain GVHD.

Inclusion Criteria:

• Patient at risk for GVHD, i.e. undergoing HSCT for hematologic disorder
• Patient receiving a reduce intensity conditioning
• Aged 18 years or over
• Underlying hematologic disorder in complete remission
• Willing to participate after informed consent signed

Exclusion Criteria:

• Patient receiving a myeloablative conditioning regimen
• Patients receiving allo-HSCT from an haploidentical donor or an umbilical cord blood
• Patients receiving a second or beyond allo-HSCT
• Patient with the underlying hematologic disorder in relapse
• Patients having developed a GVHD before inclusion in the study
• Patients having developed engraftment syndrome
• Patients receiving corticosteroid upper 0.5 mg/kg/d
• Patients with ongoing uncontrolled medical condition
• Prior history of allergy to fluorescein
• Renal dysfunction
• Suspected or documented bowel obstruction
• Known inflammatory bowel disease
• History of major abdominal intervention
• No contraception
• Breast-feeding or pregnancy