Fluorodeoxyglucose (FDG) positron emission tomography (PET) is now widely used for cancer imaging purpose, notably for preoperative work-up. It aims at visualizing organs metabolism. In case of cancer, metabolism is, classically, increased and some hot spots are visible on PET images. Because of respiratory motion and because the liver is intrinsically FDG avid, some tumours (especially the smallest ones) can be occulted and missed by the clinician.

The investigators developed a respiratory-gated PET method in order to reduce the motion issue. This protocol has been validated on lung pathologies. The investigators designed a study to investigate its effect on liver cancer (primary or metastasis) to check if it allows the detection of a higher number of tumour lesions.

To that aim, patients who are planned to undergo a surgical intervention on the liver can be proposed to participate this study. After the standard PET acquisition (acquired in free-breathing), an additional 10 minutes respiratory-gated PET acquisition is performed without additional injection. After that, a breath-hold (~10s) CT is performed.

Introduction:

18-fluoro-2-deoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) became a major imaging modality for management of patients with suspected hepatic cancer notably because of its ability to detect distant metastases. Hepatic resection is known to be the only curative treatment in a subset of patients suffering liver neoplasm. Eligibility for hepatic surgery lays on conventional staging comprising contrast-enhanced CT, magnetic resonance imaging or ultrasonography of the liver. Recently, 18F-FGD PET/CT has been introduced as a staging modality in the preoperative work up.

However, many physiological issues could affect PET interpretation. Indeed, some lesions could be missed by the physician due to histology of the lesion (e.g. in case of moderate or well-differentiated hepatocellular carcinoma (HCC) or mucinous carcinoma). Unlike, local inflammations could be wrongly considered as malignant.

Another issue for liver examination with PET imaging is the respiratory motion as it produces blurring in reconstructed images. Therefore, some lesions may be missed or underestimated. This motion issue is well known for thoracic imaging and various methods were proposed to deal with it. We have developed a respiratory motion compensation method where gated PET events are selected on the base of a breath-hold CT (CT-based).

To our knowledge, no sensitivity study was performed to assess the usefulness of gated acquisitions in term of lesions detection at abdominal stage. W In this trial, we apply our gating method on the liver to compare lesion-per-lesion sensitivity of clinical (Ungated) and CT-based PET images.

PET/CT acquisitions:

All acquisitions (Ungated and CT-based) are performed on a whole-body PET/CT system.

• Whole-body PET/CT (Ungated session) The Ungated acquisition consists in a whole-body, free-breathing CT (110 kV; 85 mAs; pitch: 1) followed by standard multistep PET (3 minutes per step), as used in routine clinical practice in the department.
• Respiratory-gated PET/CT (CT-based session) The CT-based method consists in an additional single-step, 10-minute List Mode respiratory gated PET acquisition followed by an end-expiration breath-hold CT (110 kV; 50 mAs; pitch: 2) added to the end of the clinical protocol, with continuous respiratory signal recording during these examinations.

To summarize respiratory-gated PET processing, the breath-hold CT sequence is visible on the respiratory signal as a plateau. A selection range is placed around this plateau to select only the PET events which correspond to the same position as that of tissues at the time of breath-hold CT.

Image Reconstruction:

After compensation for random coincidences, all 3-dimensional (3D) sinograms are FORE-rebinned into two dimensions and scatter-corrected. The attenuation coefficients at 511 keV are calculated from the CT acquisition, in order to correct for tissue self-attenuation. Ungated volumes are corrected with the free-breathing whole-body CT scan and CT-based volumes are corrected with the end-expiration breath-hold CT. All PET volumes are reconstructed using AWOSEM with the following parameters: 4 iterations, 8 ordered subsets in 168 x 168 x 81 matrices (4.06 mm x 4.06 mm x 2 mm). Finally, a 3D isotropic Gaussian post filter with a full width at half maximum of 5 mm was applied.

Image analysis:

Each PET examination is blindly and independently analysed by experienced nuclear medicine physicians, i.e. neither the type of image nor the patients' information are known. Each clinician has to report the number of lesions detected in both Ungated and CT-based PET datasets. Should it be the case, for each lesion, they are asked to precise its location according to the Couinaud segmental classification and its maximum standardized uptake value (SUVmax).

Surgical procedure and histopathological analysis:

During surgery, intraoperative ultrasound are performed in order to detect and localize all liver lesions. Surgery is aimed at obtaining disease-free resection margins. The type of liver resection (hepatectomy or wedge(s)) is at the surgeon discretion as well as the use of radiofrequency tumour ablation. When available, surgical specimen are analysed in the department of histopathological analysis of our institution to determine the type of cancer and the location of the resected lesions.

Inclusion Criteria:

• patient planed to undergo any liver surgical intervention
• age : over 18
• patients gave their written informed consent

Exclusion Criteria:

• pregnancy
• liver surgical intervention cancel upon surgical or medical decision

Patients presenting liver lesions planned for hepatic surgery

1 year