Background
The metabolic characterization of undetermined solid pulmonary nodules detected at computed tomography (CT) is one of the first indications for
18 F-fluoro-deoxy-glucose (
18F-FDG) positron emission tomography (PET) [
1]. In this setting, integrated
18 F-FDG PET-CT provides higher values of sensitivity, specificity, and accuracy (97%, 85%, 93%, respectively) when compared with
18 F-FDG PET alone [
2]. The diagnostic performance of
18 F-FDG is similar for nodules measuring at least 1 cm and for larger masses, but few data exist on its diagnostic performance for nodules smaller than 1 cm [
3]. The few false-negative
18 F-FDG results can be related to ‘metabolic’ causes such as low
18 F-FDG avidity, or to ‘technical’ aspects, commonly related to the nodule size [
2]. Likewise, false-positive
18 F-FDG results can be related to ‘metabolic’ causes such as inflammatory processes, or to ‘technical’ aspects such as micro-embolisms provoked during the tracer injection.
Herein we report on five cases of focal lung 18 F-FDG uptake in pulmonary nodules smaller than 1 cm, and we briefly discuss the most common causes of 18 F-FDG false-positive and false-negative results in the pulmonary parenchyma.
Discussion
Contrast-enhanced CT is the best diagnostic technique to detect pulmonary nodules providing anatomic and morphologic information [
2]. One of the well-established indications for
18 F-FDG PET-CT is the metabolic characterization of undetermined pulmonary nodules either in oncologic patients or in patients without any known malignancy but with a high risk of lung cancer [
4]. The complementary roles of anatomic and metabolic imaging improve the diagnostic accuracy in the diagnosis and characterization of pulmonary nodules [
2]. The risk of malignancy in pulmonary nodules, especially in those smaller than 1 cm, depends on several factors related to the patient, such as a clinical history of previous malignancy, age, smoking history, and on radiological criteria such as nodule size, margins, and density [
5‐
7].
The diagnostic performance of
18 F-FDG is similar for nodules measuring at least 1 cm and for larger masses with a sensitivity of 96% and specificity of 73% [
3]. Unfortunately, few data exist on the diagnostic performance of
18 F-FDG for nodules smaller than 1 cm and the relationship with the risk of malignancy. In the oncologic population, Reinhardt
et al. [
8] report a PET sensitivity of 40% and 78% in pulmonary metastases ranging from 5–7 mm and 8–10 mm, respectively. In both the oncologic and non-oncologic populations, Herder
et al.[
9] report a PET sensitivity of 93% and specificity of 77% in undetermined pulmonary nodules ≤10 mm; in the non-oncologic high risk population, Kim
et al. [
2] report a sensitivity of 50% in nodules ranging from 7–10 mm and Divisi
et al. [
10] report a sensitivity of 95% in solitary lung nodules between 5 and 9.9 mm.
18 F-FDG PET sensitivity for characterizing pulmonary nodules as probably malignant can be compromised by ‘metabolic’ causes and ‘technical’ aspects, which lead to false-negative results.
The metabolic causes which can reduce sensitivity are: (1) a high blood glucose level, because of competitive reaction [
11]; (2)
18 F-FDG avidity related to the histological pattern: low in some tumor types and in slow-growing tumors (bronchiolo-alveolar carcinoma, carcinoid, metastasis of clear-cell renal-cell carcinoma, etc.) [
12]; (3) the
18 F-FDG uptake related to the degree of cell differentiation: lower in well-differentiated cells than in moderately differentiated ones [
13]; (4) the number of viable malignant cells. Fischer
et al. [
14] demonstrated ‘in vitro’ that the theoretical detection limit of
18 F-FDG is in the magnitude of 10
5 to 10
6 malignant cells, depending on the glucose turnover of the specific cancer. Recently, Wahl
et al. [
15] reported ‘in vivo’ that the limit of
18 F-FDG PET for detecting cancers is generally in the magnitude of 10
8-10
9 cells, which translates into a tumor size between 0.4 and 1 cm in diameter.
The technical aspects that can determine an underestimation of the true
18 F-FDG activity, especially in nodules smaller than 1 cm, are: the respiratory motion because of the displacement caused by shallow breathing, particularly in nodules located in the periphery and in the base of the lungs; the partial volume effect because nodules smaller than the resolution of the PET scanners (ranging from 6 to 10 mm in clinical applications) are not, or only faintly visualized [
8,
16,
17]. Likewise,
18 F-FDG PET specificity for characterizing pulmonary nodules as probably malignant can be compromised by ‘metabolic’ causes and ‘technical’ aspects, which lead to false-positive results.
The metabolic causes are: benign neoplasms (sclerosis hemangioma, leiomyoma, and so on), infection (tuberculosis, sarcoidosis, and so on) or inflammation (acute inflammation associated with bronchiectasis or thromboembolic disease, and so on) [
12,
18‐
20]. To our knowledge, a pulmonary micro-embolism provoked during
18 F-FDG injection can be considered the only reason ‘technically’ responsible for a false-positive result. The vascular endothelium can be damaged by several factors, such as a para-venous or ‘in bolo’ injection, producing micro-emboli. The cellular activation process at the site of pulmonary micro-emboli requires energy with a consequent increased glucose uptake [
21,
22]. From a ‘scintigraphic’ point of view, this is evident as a focal
18 F-FDG lung uptake but, thanks to CT integrated with the PET scanner, its artifactual nature should be suspected in the absence of corresponding abnormalities at the co-registered CT [
23].
Our first case showed focal
18 F-FDG activity in the lung parenchyma in the absence of any detectable abnormality, even at co-registered unenhanced CT. We could exclude an artifact caused by the micro-embolism provoked during injection because: we routinely inject the radiotracers through a venous cannula avoiding a para-venous injection and repeated blood aspirations; no bolus injection was performed and no vascular activity due to a para-venous injection in the arm was evident in the images. The focus of
18 F-FDG activity projecting onto the subpleural parenchyma was considered as non-specific, also because at contrast-enhanced CT performed only ten days before, no other morphological abnormalities were detectable except the known mass. It is very interesting to note that the
18 F-FDG activity was already evident 7 months before the CT appearance of a 1 cm nodule, therefore indicating its malignant nature. The absence of a detectable nodule at the first CT examination suggests that the area of focal
18 F-FDG uptake contained a very small number of tumor cells, not enough for anatomical detection, and therefore, it would be reasonably to say around the lowest
18 F-FDG detection limit, as reported [
15]. This supports the well-known concept that the metabolic/functional alterations may precede the morphologic ones, and PET can sometimes detect early changes not, or only minimally revealed by morphological imaging [
24].
Also in case 2 we could exclude an artifact due to a micro-embolism by applying the same considerations previously described. Differently from case 1, taking into account the very high oncologic risk of this patient, the PET finding was considered highly suspicious for malignancy, even in the absence of any clear morphologic abnormalities at co-registered unenhanced CT. No clear evidence of a nodule at co-registered unenhanced CT could be explained by some technical acquisition reasons: slice thickness (5 mm), free breathing and ‘low dose’ setting (tube current of about 40 mA/s). These acquisition parameters could limit the detection of small pulmonary nodules at co-registered unenhanced PET-CT, especially in a juxtavascular location. A volumetric high-resolution CT was suggested to overcome these limits; with this technique, slice thickness is 1 mm or less, in a single, breath-held inspiration, with a full radiation dose, allowing the identification of small pulmonary nodules. However, a delay of 3 months was suggested to identify dimensional growth, as this is the only reliable criterion in favor of malignancy; the other criteria commonly used to define malignancy of a nodule (dimensions, site, margins) are less reliable in small nodules [
25‐
27]. Also in this case, PET was able to detect malignant findings earlier than morphological imaging. No clear detection of the nodule at co-registered unenhanced CT and also at the first high-resolution CT suggests that the number of tumor cells showing
18 F-FDG uptake was very low, as previously reported for case 1 [
15]. In addition, very early
18 F-FDG uptake could be explained by the high mitotic index and by the presence of highly aggressive and undifferentiated cells, as proved by histopathology.
In case 3, PET showed an intense focal
18 F-FDG uptake in the right lung, located between two vessels, where no abnormalities were recognized either at the staging contrast-enhanced CT, or at CT integrated with PET. Also in this case we could exclude the ‘technical’ origin of a micro-embolism. So, guided by focal
18 F-FDG uptake, we carefully looked for any alteration in lung parenchyma at co-registered unenhanced CT as well as at contrast-enhanced CT. Eventually, a small juxtavascular nodule corresponding to the focal uptake was recognized, and its neoplastic nature was confirmed by the subsequent response to radiation therapy, even if a disease progression with multiple new pulmonary nodules and neoplastic abdominal disease were found. Possible causes of failed identification of pulmonary nodules at CT are a central location, either within the bronchi (as in case 2) or adjacent to vessels, as in this case; other possible causes are a small size, faint attenuation, lower lobe location or location adjacent to other parenchymal abnormalities such as inflammatory lesions [
28]. This case underlines the importance of looking at the pulmonary parenchyma with high accuracy, especially in high-risk oncological patients, searching for any abnormality and eventually using PET findings as a guide because PET could sometimes guide the identification of alterations missed at morphologic imaging.
In case 4, CT revealed a pulmonary nodule not present in the previous CT scan, difficult to be characterized because of its small size and the presence of the atelectatic band. PET performed only 1 month later demonstrated that the nodule had metabolic activity, strongly suggesting its neoplastic nature. Unfortunately, the patient died for cancer progression and no proof of its true nature was available. Multidetector CT can detect nodules as small as 1 or 2 mm but without being able to characterize them. Nevertheless, in this case, the pulmonary nodule was characterized by PET even though located at the base of right lung, in the context of the atelectatic band and only 8 mm in size.
In case 5, PET was able to characterize a pulmonary micronodule of 4 mm, impossible to characterize at morphologic imaging. Despite the very small size, the
18 F-FDG uptake can be explained by ‘metabolic’ reasons such as the aggressive histotype of the primary thyroid cancer (Hurthle cell carcinoma), and the de-differentiation of the tumor cells, rather than their number. It has been widely demonstrated that secondary lesions of thyroid carcinomas with low iodine avidity tend to have higher glucose metabolism and are more likely to be positive at
18 F-FDG PET [
29].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MLC and ST drafted the manuscript, contributed to conception, manuscript and design preparation, and conducted a literature search. MLC, ST, FM, VR, and LB performed and analyzed PET and CT images. LL, GT, and FM contributed to acquisition of data and obtained images used in the manuscript. VR participated in the design of the study. LB and AG critically revised the manuscript. All authors read and approved the final manuscript.