Physiological distribution of ¹⁸F-FDG in the spinal cord: A systematic review

Abstract

Context: The importance of physiologic distribution of ¹⁸F-FDG in the spinal cord.

Objective: The recognition of the physiologic distribution of ¹⁸F-FDG in the spinal cord is pivotal for accurate PET/CT imaging interpretation, especially in oncologic patients. Therefore, we performed a systematic review to investigate the normal distribution of ¹⁸F-FDG throughout the spinal cord.

Methods: Data sources: We carried out a comprehensive search of the literature on the physiologic patterns of ¹⁸F-FDG distribution in the spinal cord. PubMed and Scopus databases were searched using the following keywords: “spinal cord” AND “FDG”. Data extraction: Findings of the selected articles were described.

Results: Thirteen studies comprising 24,125 patients entered the systematic review. These investigations showed discrepancies in location, size, number, and intensity of ¹⁸F-FDG uptake throughout the spinal cord. However, cumulative results showed that ¹⁸F-FDG uptake was higher in the lower thoracic portion of spinal cord (T11-T12). Moreover, a decreasing trend in ¹⁸F-FDG uptake was observed from cervical to lumbar levels. Low maximal standardized uptake values, female sex, and higher body weight seem to be related to the physiological spinal cord ¹⁸F-FDG uptake.

Conclusions: On ¹⁸F-FDG PET/CT imaging, focal hypermetabolism of the spinal cord at the level of lower thoracic and lower cervical vertebrae should be considered physiological until proven otherwise.

Introduction

The application of Fluorine-18-Fluorodeoxyglucose Positron Emission Tomography (¹⁸F-FDG PET) has an incremental growth for staging, restaging as well as evaluating the response to treatment in oncologic patients.^1–4 This functional imaging modality provides information about the glucose metabolism of tumor lesions, while morphological imaging modalities such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) mostly show anatomic features.^1,2,5 Currently, hybrid imaging methods (i.e.: PET/CT and PET/MRI) allow to obtain metabolic and morphological information at the same time increasing the diagnostic accuracy in detecting tumor lesions. Nonetheless, nonspecific ¹⁸F-FDG uptake in body organs or in benign lesions can decrease the specificity of ¹⁸F-FDG PET.^1,2

Incidental increased ¹⁸F-FDG uptake in any organ may have clinical significance especially in oncologic patients.^6–11 This emphasizes the importance of recognition of the ¹⁸F-FDG normal distribution patterns from malignant metastatic lesions. The distinction of normal uptake patterns of ¹⁸F-FDG is pivotal for accurate PET/CT imaging interpretation.

¹⁸F-FDG uptake in the spinal cord seems to be not uncommon and it may be both physiological or expression of an underlying disease.³ Many reports of abnormal hypermetabolism in the spinal cord due to metastatic lesions or other kinds of myelopathies are available.^12–19 About 8–9% of central nervous system metastases occur in the spinal cord and lung and breast cancers are the most common primary malignancies with spinal cord metastases.^3,20–22 On the other hand, there are different patterns of the physiologic ¹⁸F-FDG uptake in the spinal cord (Figure 1).^4,23,24

Figure 1.

Focal physiologic hypermetabolism in the lower thoracic spinal cord segments in an oncologic patient who underwent ¹⁸F-FDG PET/CT scan in our center.

Given the importance of identifying the normal distribution of ¹⁸F-FDG uptake in the spinal cord for accurate PET/CT imaging interpretation, especially in oncologic patients, we aimed to perform a systematic review to add more evidence-based data in this setting.

Methods

We carried out a comprehensive search of the literature on the physiologic patterns of ¹⁸F-FDG uptake in the spinal cord. PubMed and Scopus databases were searched using the following keywords: “spinal cord” AND “FDG”. No date or language limit was used and the last search was performed on November 30, 2018. We also screened the references of the pertinent articles to find any other relevant study. We included all articles on physiologic uptake of ¹⁸F-FDG or incidental ¹⁸F-FDG avid lesion in the spinal cord.

We excluded the studies wherein the enrolled patients have any signs or symptoms suggesting spinal cord abnormality, history of previous spinal surgery or radiotherapy or spine diseases.

We included two groups of articles: those assessing the distribution of the ¹⁸F-FDG throughout the spinal cord and those determining the presence of incidental focal ¹⁸F-FDG uptake in the spinal cord. For each eligible study, information on study date, demographic data such as sample size, age and sex, as well as technical features such as blood sugar, indication, follow up, uptake period, ¹⁸F-FDG activity, PET/CT scanner type and PET mode were extracted. We also collected data regarding image analysis method, PET/CT findings, and SUVmax ranges.

Results

Thirteen studies encompassing 24,125 patients were included. PRISMA flowchart of the study can be found in Figure 2. Nine studies comprising 779 patients determined the distribution of the ¹⁸F-FDG in the spinal cord and the second group including four articles, comprising 23,346 patients, determined the presence of a focal accumulation of ¹⁸F-FDG throughout the spinal cord. In all studies, patients have a history of malignant disease, except for two which including 30 and 16 healthy subjects referred to 18F-FDG PET scan for cancer screening and unclear reasons, respectively.^25,26 As pointed above, the studied patients have neither a clinical spinal myelopathy nor a history of spine disease. They had no evidence of spinal disease or spinal tumoral involvement during clinical follow-up or in additional imaging (CT, MRI, ¹⁸F-FDG PET/CT).

Figure 2.

PRISMA flowchart of the current systematic review.

The details of the patient characteristics and technical features of each investigation related to the first and second groups are listed in the Tables 1 and 2, respectively.

Table 1. Characteristics of the included studies, which reported distribution of FDG in the spine.

Author, country	Sample size	Age (range)	Sex	Blood sugar	18F-FDG dose	Time to imaging from tracer injection	PET mode
Patel et al., USA	200	20–88	M:89 (44%) F:111 (56%)	<200 mg/dl	15mCi	60–90	3D
Jorgov et al., France	30	3–17	M:15 (50%) F:15 (50%)	N/A	3–5 MBq/kg	60	N/A
Taralli et al., Italy	167	3.9–18.9	M:97 (58%) F:70 (42%)	<160 mg/dl	81–336 MBq	50–70	3D
Chong et al., Korea	30	42–77	M:24 (80%) F:6 (20%)	<150 mg/dl	375 MBq	60	N/A
McCarville et al., USA	128	1–24	M:74 (58%) F:54 (42%)	<200 mg/dl	0.15 mCi/kg	60	2D
Nakamoto et al., USA	78	10–83	M:36 (46%) F:42 (54%)	<11.1 mmol/L	555–740 MBq	60	N/A
Bao et al., USA	92	55.5 ± 19	M:49(53%) F:43 (47%)	<300 mg/dl	N/A	N/A	2D
Kamoto et al., Japan	21	31–88	M:12 (57%) F:9 (43%)	N/A	244–488 MBq	40–80	2D
Uematsu et al., Japan	33	48–87	M:19 (57%) F:14 (43%)	<120 mg/dl	370 MBq	40	N/A

Table 2. Characteristics of the included studies, which reported focal uptake of the FDG in the spine.

Author, Country	Sample size	Age (range)	Sex	Blood sugar	18F-FDG dose	Time to imaging from tracer injection	PET mode
Lim et al., South korea	22,937	18–87	M:31 (45%) F:38 (55%)	<200	5 MBq/kg	60	2D
Amin et al., Egypt	101	21–84	M:49 (48.5%) F:52 (51.5%)	<150	280 ± 35 MBq	60	N/A
Bhatt et al., USA	200	18–80	M:118 (59%) F:82 (41%)	≤200	340.4–592.0 MBq	55–85	3D
Greenspan et al., USA	108	26–91	M:49 (45%) F:59 (55%)	60–198	296–610.5 MBq	60	N/A

Distribution of ¹⁸F-FDG uptake in the spinal cord

Investigation methods of this group of studies varied considerably. The results are summarized in Table 3.

Table 3. Results of the first group of the articles related to diffuse spinal 18F-FDG uptake.

Author, Country	Indication	Follow up	Image analysis	Results	Correlations (univariate analysis)	Correlations (multivariate analysis)
Patel et al., USA	Oncologic patients	By MRI	SUVmax of spinal cord at C2, C5, T6, T12, and L3 levels	Highest SUVmax at C2 (0.54–2.91, 1.76) level Lowest SUVmax at T6 (0.58–3.51, 1.37) level	– No association with sex, age, BMI & FBS  – Positive & strong correlation with corresponding vertebral marrow SUVmax	N/A
Jorgov et al., France	Oncologic pediatric patients (HL)	By PET/CT scan & clinical follow up (36.5 m)	Visual assessment of T12 spinal cord level	Non-focal FDG uptake at T12 level with mean SUVmax of 1.7 (more than liver FDG uptake)	N/A	N/A
Taralli et al., Italy	Oncologic pediatric patients	N/A	SUVmax of three parts of spinal cord (C1-7, T1-6, T7-L1)	Similar SUVmax at C1-7 & T1-6 levels (1.9 & 1.8, Res) Significantly lower SUVmax at T7-L1 level (1.6)	– Positive & statistically significant association with body weight, female sex and Hodgkin lymphoma.  – No association with season.	– Positive & statistically significant association with body weight and female sex  – No association with age, disease type and season
Chong et al., Korea	Cancer screening (Healthy)	By PET/CT scan (2.8 y)	SUVmax of each mid-vertebral level (C1 to L5)	Decreasing pattern from cervical to lumbar vertebrae with significant increase at the lower cervical (C4-6) and the lower thoracic (T12) levels	No association with sex or age	N/A
McCarville et al., USA	Oncologic pediatric patients	N/A	SUVmax of spinal cord at C2, C4, C7, T2, T4, T8, T11, T12, L1 and L4 levels	Highest SUVmax at C4, T11, and T12 levels (all three means = 1.6)	Positive & statistically significant association with age	N/A
Nakamoto et al., USA	Oncologic patients (non H&N cancer	N/A	mean SUV of the C1 spinal cord level	mean SUVs of 2.12 ± 0.48	No association with age, FBS & sex	N/A
Bao et al., USA	Oncologic patients (non-CNS cancer)	N/A	SUVmax of each mid-vertebral level (C1 to L3)	Decreasing pattern from cervical to lumbar vertebrae with significant increase at the T11-T12 vertebral levels and relative insignificant increase at C4	No association with age or sex	No association with age or sex
Kamoto et al., Japan	Oncologic patients	By MRI, CT & PET/CT scan	SUVmax of three parts of cervical spinal cord (C1-3, C4-5, C6-7)	Slight non-significant decreasing pattern from cranial to caudal parts (SUVmean of the upper, middle, and lower segments = 1.99, 1.93, and 1.91, res)	Negative association with age	N/A
Uematsu et al., Japan	Normal subjects (16) & oncologic patients (H&N cancer)(17)	N/A	SUVmax of whole cervical spinal cord	Non-focal FDG uptake in cervical spinal cord with a mean SUVmax of 2.09 ± 0.56	N/A	N/A

Chong et al. and Bao et al. assessed the maximal standardized uptake values (SUVmax) of all spinal segments in each mid-vertebral level from cervical to lumbar regions and reported spinal ¹⁸F-FDG uptake as a cord to background ratio.^24,25 They found a decreasing pattern in the cord to background ratio from cervical to lumbar levels of the spinal cord with a statistically significant increase at the lower thoracic and mid to lower cervical levels.^24,25

Kamoto et al. only evaluated cervical spinal cord levels;²⁷ they also found a decreasing pattern in ¹⁸F-FDG uptake from cranial to caudal parts of cervical levels of the spinal cord.²⁷ In another comparable study, the authors divided the total spinal cord segments (from C1 to L1) to 3 parts and assessed ¹⁸F-FDG uptake in each part separately.⁴ They showed similar ¹⁸F-FDG avidity in the cervical (C1-C7) and upper thoracic (T1-T6) parts and statistically significant lower uptake in the lower thoracic (T7-L1) parts.⁴ Another study by Uematsu et al. evaluated the whole cervical spinal cord as one region of interest (ROI).²⁶ They concluded that the cervical cord can be easily visualized due to high average SUVmax compared to the adjacent structures.

McCarville, Patel and their colleagues only evaluated ¹⁸F-FDG avidity in 10 and five levels from the entire of the spinal cord, respectively.^28,29 In the former study, the highest SUVmax values were in the mid-cervical and lower thoracic levels;²⁸ whereas, in the latter study, the highest SUVmax value was in upper cervical levels.²⁹ In addition, two studies only determined the SUVmax values of one segment of the cord, including T12 and C1.^2,30

Focal ¹⁸F-FDG uptake in the spinal cord

In this second group of articles, the presence of focal uptake of ¹⁸F-FDG was determined. The details regarding the methods of image analysis and results are presented in Table 4.

Table 4. Results of the second group of the articles related to focal spinal 18F-FDG uptake.

Author, Country	Indication	Follow up	Image analysis	Focal uptake	Max cord uptake level	SUVmax range (mean ± SD)	Correlations (univariate analysis)	Correlations (multivariate analysis)
Lim et al., south Korea	Oncologic patients	By FDG-PET/CT, MRI, CT or clinical follow up (20 m)	SUVmax of focal FDG uptake	69 (0.3%)	T12 (60.3%) L1 (25.6%) T11 (7.7%)	1.4–3.9 (2.5 ± 0.5)	N/A	N/A
Amin et al., Egypt	Oncologic patients	By MRI & FDG-PET/CT	SUVmax of each mid-vertebral level (C1-L3), with a SUVmax cut off ≥ 1	49 (48.5%)	T11-T12 (73.4%) C1-C7 (49%)	(2.1 ± 0.1)	– No association with age, sex, BMI, mobility status, cancer type or specific metastatic site and therapeutic modality  – More observed in winter	N/A
Bhatt et al., USA	Oncologic patients	N/A	SUVmax of C3-C5, & T11-T12 levels	45 (22.5%)	T11-T12	(2.53 ± 0.63)	– Association with weight, F-18 FDG dose, Caucasian race as well as liver, bowel & C3-C5 spine SUVmax.  – No association with ambulatory status	N/A
Greenspan et al., USA	Oncologic patients	N/A	SUVmax of focal FDG uptake	17 (15.7%)	T11–L1	1.6–3.7 (2.43)	– Association with female sex, white race, age <61.9 years & FBS <108 mg/dL  – No association with diabetes, cancer type, weight, F-18 FDG dose, or scan delay	Statistically significant association with female sex & lower FBS

Lim et al. and Greenspan et al. used similar analysis methods. They measured the SUVmax value of the hypermetabolic foci throughout the cord.^6,23 They reported the prevalence of focal ¹⁸F-FDG uptakes as 0.3% (with mean SUVmax of 2.5) and 15.7% (with mean SUVmax of 2.43), respectively.^6,18

Amin et al. determined SUVmax of each mid-vertebral level from C1 to L3 vertebrae and used SUVmax ≥1 as cut off value for ¹⁸F-FDG focal uptake.⁵ They reported ¹⁸F-FDG focal uptakes in 48.5% of patients (with mean SUVmax of 2.1 ± 0.1).⁵

Bhatt et al. explored SUVmax of C3-C5 and T11-T12 levels; they reported focal ¹⁸F-FDG avidity in 22.5% of their patients (mean SUVmax of 2.53 ± 0.63).³ As a common result from these four investigations, the hypermetabolic foci observed between the T11 to L1 cord levels were common more than the other spinal segments.

Factors associated with ¹⁸F-FDG uptake in the spinal cord

Several studies evaluated the association between the ¹⁸F-FDG uptake and multiple physiologic and pathologic factors of the patients. These data are summarized in Table 5. Several factors were evaluated including age, sex, weight, ¹⁸F-FDG dose, physical activity, season, etc.

Table 5. The correlation of 18F-FDG uptake with multiple physiologic and pathologic factors of the patients.

Author, Country	Patient number	Age	Sex	Race	Weight/ BMI	FBS	Other organs 18F-FDG uptake	Cancer type	Mobility status	Season	Other
Patel et al., USA	200	No association	No association	N/A	No association	No association	Positive & strong correlation with corresponding vertebral marrow SUVmax	N/A	N/A	N/A	N/A
Jorgov et al., France	30	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A
Taralli et al., Italy	167	N/A	Positive & statistically significant association with female sex	N/A	Positive & statistically significant association with body weight	N/A	N/A	Positive & statistically significant association with Hodgkin lymphoma	N/A	No association	N/A
Chong et al., Korea	30	No association	No association	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A

Factors associated with physiological spinal cord ¹⁸F-FDG uptake were numerous. Age of the patients showed conflicting results: one study found a negative correlation,²⁷ two studies showed a positive correlation^23,28 and five studies did not find any relation with the age.^2,5,24,25,29 One of the studies with the positive correlation was performed in pediatric patients,²⁸ and the other one only stated association with age lower than 61.5 years.²³

Positive association with female sex was shown in two studies,^4,23 while five investigations did not find any correlation.^2,5,24,25,29

Five studies evaluated the association between ¹⁸F-FDG uptake and body weight; the positive correlation was found in two studies,^3,4 whereas three investigations had not detected any association.^5,23,29

Physical activity of the patients’ physical activity does not seem to be related to the spinal cord ¹⁸F-FDG uptake as reported by two included studies.^3,5 Finally, two studies evaluated the correlation of ¹⁸F-FDG uptake with season; one found positive⁵ and the other found no association.⁴

Discussion

The frequency of spinal cord metastasis in oncologic patients’ autopsy series was reported to be 0.9–2.1%.^18,20–22 Even though the spinal cord metastases are rare, it can cause a serious neurologic deficit and short survival time, and timely diagnosis is imperative for proper managing.^18,20

It should be noted that increased ¹⁸F-FDG uptake throughout the spinal cord of oncologic patients, whether focal or diffuse, is not infrequent but such uptake must be considered as physiologic to avoid misdiagnosing as metastatic involvement.

According to the performed studies on diffuse ¹⁸F-FDG uptake throughout the spinal cord, most studies reported a decreasing uptake from cervical to lumbar spinal cord levels with two intermittent peaks at the lower cervical (C4-6) and the lower thoracic (T12) levels.^24,25,27 In accordance with these findings, studies evaluating the focal uptake in the cord reported that the most common site of focal hypermetabolism is in the lower thoracic level.^3,5,6,23 One study reported the lower cervical level as the second most common site.⁶

Several mechanisms have been proposed as the cause of physiological ¹⁸F-FDG uptake variations in the spinal cord including:

• Hypermetabolism of the cervical enlargement in spinal cord segments C6-C8 (correspond to the C3-T2 vertebral body levels) and the lumbar enlargement in spinal cord segments L4-S1 (correspond to the T9-T12 vertebral body levels), which are responsible for supplying the neuronal activity of the upper and lower limbs.^3,23–25

• White matter does not accumulate ¹⁸F-FDG in comparison with the gray matter. Therefore, it can be concluded that abundant amount of gray matter in the spinal cord enlargements at the C4-C8 and L4-S1 vertebral levels, involving in neuronal transmission for upper and lower extremities, could be responsible for the relative hyperactivity.^3,6,24,25

• Some investigators brought up a hypothesis that the sensory stimuli (e.g. tactile or thermal) from the extremities to the spinal cord can be the reason of regional hypermetabolism.^24,25

• Inadequate clearance of ¹⁸F-FDG from the Adamkiewicz artery (largest medullary segmental artery originated from aorta around the T9 to T11 thoracic vertebrae on the left side) is another explained mechanism.^3,31

• In a study performed on rats with arthritis, acute or chronic arthritis in the extremities can cause a significant increase in metabolic activity of the relevant cord levels.^32,33 This can also be true in human being although not specifically investigated.^3,24

• Technical limitations like partial volume effect are another proposed mechanism that may be a cause of underestimation of SUVmax in spinal segments except for cervical and lumbar enlargements.²⁴

Several factors can be helpful in differentiation between physiologic and malignant spinal cord ¹⁸F-FDG uptake including SUVmax, and location in the spinal cord.

The highest reported mean SUVmax in the physiological focal ¹⁸F-FDG uptake was 2.53 ± 0.63, whilst the reported mean SUVmax for spinal metastatic lesions is higher. Flanagan et al. reported that the mean SUV max of the neoplastic spinal cord lesions was 3.3 with the range of 1.8–14.8. This was higher than the inflammatory myelopathy lesions though (1.9 with the range of 1–4.5).¹⁶ Another study reported the mean SUVmax of 6.7 g/mL (with a range of 3.3–9.9 g/mL) in spinal cord metastases.¹⁸ Gilardi et al. also reported the SUV max of 4.1 g/mL and 4.7 g/mL for two metastatic lesions in the spinal cord due to breast cancer.³⁴

The location of the ¹⁸F-FDG avid areas in the spinal cord is another important factor to be considered. Several studies reported that intramedullary spinal cord metastasis can occur at every segment of the cord but the cervical spine seems to be the most common region.^21,22 This predilection can in part be explained by greater bulk of tissue and rich vascularity of the cervical spine.^21,22 On the other hand, the most common sites of physiological spinal hypermetabolism are lower thoracic (including the first lumbar segment) and lower cervical regions.

In summary, on ¹⁸F-FDG-PET/CT imaging incidental focal hypermetabolism of the spinal cord at the level of lower thoracic vertebrae as the most common site and lower cervical vertebrae as the second common site should be considered physiological until proven otherwise. Low SUVmax values (<4), female sex, and higher body weight seems to be associated with the physiological spinal cord ¹⁸F-FDG uptake.

More studies are needed to clarify the exact factors which are related to the physiological ¹⁸F-FDG uptake and its pattern.

Disclaimer statements

Contributors None.

Funding This work was not supported by any organization.

Conflicts of interest All the authors declare that they have no conflict of interest.