Clinical evaluation of (18)F-fludeoxyglucose positron emission tomography/CT using point spread function reconstruction for nodal staging of colorectal cancer

Abstract

Objective: For patients with colorectal cancer, lymph node metastasis is a very important factor for prognostic and treatment determinations. Fluorine-18 fludeoxyglucose positron emission tomography/CT ((18)F-FDG-PET/CT) is among the useful tools for detecting lymph node metastasis. Recently, a new (18)F-FDG-PET/CT reconstruction technique for improving spatial resolution and signal-to-noise ratios, point spread function (PSF), has become available. We assessed the effect of PSF reconstruction on standardized uptake values and its diagnostic accuracy for lymph node staging in patients with colorectal cancer.

Methods: We retrospectively analysed records from patients with colorectal cancer who underwent (18)F-FDG-PET/CT for pre-operative staging. All positron emission tomography CT (PET/CT) examinations were reconstructed using ordered subset expectation maximization (OSEM) and OSEM + PSF. We compared sensitivities, specificities, positive-predictive values (PPVs), negative-predictive values (NPVs) and accuracies of conventional PET/CT (reconstructed with OSEM) and PSF-PET/CT (reconstructed with OSEM + PSF) for identifying lymph node metastases. We also analysed the diagnostic confidence level on a 5-point scale.

Results: With conventional PET/CT, the sensitivity, specificity, PPV, NPV and accuracy were 53.1%, 99.1%, 94.4%, 88.3% and 89.1%, respectively. With PSF PET/CT, the corresponding values were 65.6%, 99.1%, 95.4%, 91.2% and 91.8%, respectively. Conventional PET/CT and PSF PET/CT did not differ significantly in terms of N-stage definition (p = 0.125). However, the diagnostic confidence level of PSF PET/CT was significantly higher than that of conventional PET/CT (p < 0.01).

Conclusion: PSF reconstruction might slightly increase sensitivity without impairing specificity. Moreover, this technique is expected to facilitate more confident radiological decisions when compared with conventional PET/CT. Advance in knowledge: This study demonstrates the clinical effectiveness of PSF PET/CT for lymph node staging in colorectal cancer.

Figure 1.

Relationship between lymph node size and the point spread function (PSF) to conventional fluorine-18 fludeoxyglucose positron emission tomography/CT maximum standard uptake value (SUV_max) ratio. The regression analysis showed a weak statistical correlation between the lymph node size and the PSF to conventional SUV_max ratio.

Figure 2.

A 63-year-old male with rectal carcinoma. Conventional positron emission tomography (PET) (a), conventional PET/CT fusion (b), point spread function PET (PSF PET) (c) and PSF PET/CT fusion (d) are demonstrated. PSF PET/CT images were sharper and clearer than conventional PET/CT images. The confidence levels in the proximal lymph node (arrows), assessed by two radiologists, were two on conventional images and three on PSF images. Therefore, lymph node metastasis could be diagnosed from PSF images. ROI, region of interest.