Suboptimal Sensitivity and Specificity of PET and Other Gross Imaging Techniques in Assessing Lymph Node Metastasis

Abstract

In recent years, we have emphasized the limitations of positron emission tomography (PET) in imaging disorders that are beyond the capability of current instruments and available radiotracers [1–6]. These include imaging islets, bacteria, plaques and tangles, and atherosclerotic plaques. The spatial resolution of PET has substantially improved over the past decade, and therefore, high-quality images can be generated, particularly with specialized instruments for brain and animal studies. However, when the spatial resolution of PET instruments that are designed to image the entire body compared to that of dedicated brain scanners, it becomes quite evident that image quality deteriorates substantially due to large field of view and volume of the structures assessed. In other words, specially designed PET scanners for the brain (and possibly other small organs such as the breast) can provide images with spatial resolutions at around 4–5 mm that are similar to that of phantom studies (3–5mm) [7, 8]. Therefore, such instruments are capable of generating images with great details of regional function in various structures in the brain. Unfortunately, conventional body scanners suffer from significant deterioration of signals emitted and this adversely affects the spatial resolution of objects examined. In particular, this becomes a major source of error in obese patients where significant scattering of the gamma rays further degrades the image quality. Since most metastatic lymph nodes reside in the trunk in most common cancers, detection of spread in such structures is a major challenge and therefore unreliable for accurate staging. Recent advances in PET instrumentation including the introduction of time-of-flight PET machines have overcome this issue to some extent [9]. We should alsomention that metastatic lymph nodes are at times visualized on uncorrected images when the nodes are superficial in certain locations such as the groin and neck regions [10]. Additionally, small field of view PET imaging in the head and neck region can be employed to detect metastatic lymph nodes with higher sensitivity than that of the deeply seeded lymph nodes in the trunk. Overall, the spatial resolution of conventional PET body scanners is approximately 8–10 mm. Based on the assumption that the diameter of malignant cells averages around 15 μm, in order to detect metastasis with a diameter of 8–10 mm by PET, at least 150–300 million cells should be aggregated at the site. This calculation was based on dividing the volume of the lesion by the volume of an average cell, where all volumes were assumed to be spheres. For instance, in a lesion with a diameter of 8 mm (radius of 4 mm) and a malignant cell with a diameter of 15 μm (radius of 7.5 μm), the following formula was used to estimate the number of cells: