Comparison of ⁶⁸Ga-DOTATOC and ¹⁸F-FDG Thoracic Lymph Node and Pulmonary Lesion Uptake Using PET/CT in Postprimary Tuberculosis

ABSTRACT.

Tuberculosis (TB) remains one of the world’s leading infectious cause of morbidity and mortality. Positron emission tomography (PET) associated with computed tomography (CT) allows a structural and metabolic evaluation of TB lesions, being an excellent noninvasive alternative for understanding its pathogenesis. DOTATOC labeled with gallium-68 (⁶⁸Ga-DOTATOC) can bind to somatostatin receptors present in activated macrophages and lymphocytes, cells with a fundamental role in TB pathogenesis. We describe ⁶⁸Ga-DOTATOC uptake distribution and patterns in thoracic lymph nodes (LN) and pulmonary lesions (PL) in immunocompetent patients with active postprimary TB, analyze the relative LN/PL uptake, and compare this two tracer’s uptake. High uptake of both radiotracers in PL and LN was demonstrated, with higher LN/PL ratio on ⁶⁸Ga-DOTATOC (P < 0.05). Considering that LN in immunocompetent patients are poorly studied, ⁶⁸Ga-DOTATOC can contribute to the understanding of the complex immunopathogenesis of TB.

INTRODUCTION

Tuberculosis (TB) is the leading cause of infectious disease–related death, accounting for about 1.5 million deaths annually.¹ Postprimary tuberculosis (PPT) that occurs in immunocompetent adults differs from primary TB in terms of host susceptibility, age distribution, clinical presentation, and possible complications. PPT, caused by reinfection or dormant bacilli reactivation, accounts for 80% of all clinical cases and nearly 100% of TB transmission.² In immunocompetent patients, PPT affects almost exclusively the lungs and intrathoracic lymph node (LN) enlargement is rarely demonstrated in the computed tomography (CT).^2,3

Chest X-rays and CT are useful in the morphological evaluation of TB-associated changes. Positron emission tomography (PET)/CT enables assessment of the metabolic characteristics of active PPT and primary TB.⁴ Fluorine-18-labeled fluorodeoxyglucose (¹⁸F-FDG) is used commonly in experimental and clinical PET/CT studies of TB, being the most validated tracer.⁴ However, it lacks cellular type specificity,⁴ and LN involvement in TB (especially PPT) infection is understudied.

The use of somatostatin analogs targeting surface somatostatin receptors (SSTR) not only in neuroendocrine tumors but in chronic inflammatory disease contexts has increased.^5,6 Gallium-68-labeled DOTATOC (⁶⁸Ga-DOTATOC) binds to receptors that can be overexpressed in activated inflammatory cells, particularly macrophages and lymphocytes,⁵ which play central roles in TB infection.^7,8 The aims of this study were to describe ⁶⁸Ga-DOTATOC uptake distribution and patterns in thoracic LNs and pulmonary lesions (PL) in immunocompetent patients with active PPT, analyze the relative LN/PL uptake, and compare ⁶⁸Ga-DOTATOC and ¹⁸F-FDG uptake. We hypothesize that ⁶⁸Ga-DOTATOC may have high PL and LN uptake, as in ¹⁸F-FDG, but possible differences in uptake values may correlate to their specific targets and, thus, ⁶⁸Ga-DOTATOC tracer may represent a new alternative in noninvasive study of TB pathogenesis.

MATERIALS AND METHODS

Patients.

Our institutional ethics committee approved this prospective observational study conducted between June of 2018 and April of 2019, and all patients provided written informed consent. Inclusion criteria were active PPT in immunocompetent adults with < 30 days of TB treatment at the time of the second PET/CT acquisition. Active PPT was diagnosed by symptoms; suggestive chest X-ray findings; and Mycobacterium tuberculosis positivity on bacilloscopy, sputum culture, or Xpert MTB RIF. Exclusion criteria were active/recent inflammatory/tumoral disease, pregnancy/lactation, multidrug-resistant TB infection, and HIV positivity or immunosuppression. A total of 31 patients were screened by consecutive sampling, 8 agreed to participate and 1 was excluded after missing the scheduled dates for the exams.

Imaging.

All patients underwent first ¹⁸F-FDG PET/CT (after a ≥ 6 hours fast) followed by ⁶⁸Ga-DOTATOC PET/CT in a dedicated scanner (Siemens Biograph^®; Siemens, Erlangen, Germany) at least 6 days later. Neck base–upper abdomen scans were acquired ∼60 minutes after the intravenous injection of 0.12 mCi/kg ¹⁸F-FDG or 0.05 mCi/kg ⁶⁸Ga-DOTATOC. Thoracic breath-hold CT was performed for lung parenchymal analysis. Each scan took approximately 10 minutes.

Image analysis.

Two nuclear medicine specialists (SAA, PHRC) and two chest radiologists (RSR, MMB) analyzed the PET and CT images using a DICOM viewer (Osirix Imaging Software, Geneva, Switzerland). LNs with short axes ≤ 10 mm were deemed normal.⁹ For the metabolic activity quantification from ¹⁸F-FDG studies, up to three LNs and PLs each with the highest maximum standardized uptake (SUVmax) values were chosen. To compare tracer uptake, single-slice regions of interest were copied to corresponding lesions on the ⁶⁸Ga-DOTATOC scans (Figure 1). To minimize measurement variability,^10–12 SUVmax values were divided by an internal reference tissue (mediastinal blood pool SUVmean with a 1.0 cm² ROI) value, yielding cSUVmax_FDG and cSUVmax_DOTATOC values. To assess relative uptake, we calculated SUVmax and cSUVmax LN/PL ratios.

Figure 1.

Patient 6 with active postprimary tuberculosis. Fluorine-18–labeled fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) (A and B) images show remarkable uptake in pulmonary lesions and only mild in lymph nodes. However, corresponding gallium-68_labeled DOTATOC (⁶⁸Ga-DOTATOC) (C and D) scans reveal mild uptake in pulmonary lesions and greater relative uptake in lymph nodes, especially in subcarinal level (arrow). Vertical color bars to the left side of images (A) and (C) and color bars to the right side of images (B) and (D) in standardized uptake value. Color bar to the right side of images (A) and (C) in Hounsfield units. This figure appears in color at www.ajtmh.org.

Statistical analysis.

Statistical analysis was performed using the R application (The R Foundation for Statistical Computing, Wien, Austria). Two-way analysis of variance (ANOVA) for repeated measures were performed for SUVmax and cSUVmax using a mixed model using random intercept. The calculations used both variables on a logarithmic scale to achieve normality, which made possible the use of a parametric model. The post-hoc analysis was performed using the Tukey test. The Shapiro–Wilk test and the graphical evaluation using the qq-norm graph were used to assess the normal distribution of variables. Data are presented as medians with interquartile ranges or means with SDs, according to distribution of variables. SUVmax and cSUVmax LN/PL ratios were compared between studies using the paired samples Wilcoxon test. Spearman coefficients of correlation between LN size and SUVmax or cSUVmax measured after ⁶⁸Ga-DOTATOC and ¹⁸F-FDG administration were determined. P values < 0.05 were deemed significant. The main analyzes were descriptive and, therefore, less dependent on the sample size.

RESULTS

Seven patients (five males, two females; mean age 29 [19–39] years) with culture-proven active PPT were included (Table 1). The TB treatment duration for ¹⁸F-FDG and ⁶⁸Ga-DOTATOC studies were 6 (3–9; median 7) and 15 (6–17; median 7) days, respectively. Mean uptake time for ¹⁸F-FDG studies was 63 minutes (range 55–68 minutes) and for ⁶⁸Ga-DOTATOC was 64 minutes (range 58–83 minutes). There was no statistically significant difference between uptake times (P = 0.733).

Table 1.

Patient characteristics and their diagnostic exams results

	Age	Sex	Interval between start of treatment and 18F-FDG study (days)	Interval between studies (days)	Smear	Culture	RMT
Patient 1	24	M	3	7	+	Positive	Detectable
Patient 2	39	F	9	6	–	Positive	Detectable
Patient 3	36	M	3	7	++	Positive	Detectable
Patient 4	20	M	7	14	+	Positive	Detectable
Patient 5	24	M	3	7	++	Positive	NP
Patient 6	24	F	9	7	++	Positive	NP
Patient 7	34	M	9	17	–	Positive	Detectable

¹⁸F-FDG = fluorine-18-labeled fluorodeoxyglucose; NP = not performed; RMT = rapid molecular test.

We analyzed and compared ¹⁸F-FDG and ⁶⁸Ga-DOTATOC uptake in 18 PLs (mean 2.57 lesions by patient). The metabolic activity after the injection of ¹⁸F-FDG was quantified and compared with the SUVmax of the same region after the injection of ⁶⁸Ga-DOTATOC, before and after the correction for the blood pool (Table 2). The SUVmax analysis of the mean SUVmax after injection of ¹⁸F-FDG from each patient’s lung lesions revealed values between 4.09 and 9.93 with a median of 7.09 (IQ = 4.46–9.72) before correction and between 3.08 and 7.50 with a median of 5.56 (IQ = 3.31–7.81) after correcting by the SUVmax of the blood pool. The SUVmax evaluation obtained from the average of the same regions after the administration of ⁶⁸Ga-DOTATOC resulted in values between 1.63 and 4.06 with a median of 2.09 (IQ = 0.90–3.28) before correcting and between 3.54 and 7.84 with a median of 4.36 (IQ = 3.42–5.59) after correcting by the SUVmax of the blood pool.

Table 2.

Mean SUVmax and cSUVmax of each patient obtained from lymph nodes after administration of ¹⁸F-FDG and ⁶⁸Ga-DOTATOC

	18F-FDG		68Ga-DOTATOC
	SUVmax	cSUVmax	SUVmax	cSUVmax
Patient 1	2.42 (0.45)	1.51 (0.28)	2.08 (0.31)	3.64 (0.55)
Patient 2	12.75 (1.56)	8.55 (1.05)	1.35 (0.30)	2.93 (0.65)
Patient 3	4.50 (0.38)	3.33 (0.28)	2.83 (0.57)	5.24 (1.06)
Patient 4	4.03 (0.44)	2.62 (0.29)	2.55 (0.72)	6.23 (1.75)
Patient 5	2.22 (0.15)	3.04 (0.20)	2.66 (0.13)	4.67 (0.24)
Patient 6	3.07 (0.23)	2.96 (0.23)	2.37 (0.30)	3.88 (0.50)
Patient 7	3.70 (0.38)	2.27 (0.24)	3.00 (0.44)	6.98 (1.03)

¹⁸F-FDG = fluorine-18-labeled fluorodeoxyglucose; ⁶⁸Ga-DOTATOC = gallium-68-labeled DOTATOC; cSUVmax = SUVmax divided by the SUVmax of the vascular pool; SUVmax = standardized maximum uptake value. Values are presented as mean (SD).

Likewise, LNs also had their SUVmax analyzed both in studies with ¹⁸F-FDG and with ⁶⁸Ga-DOTATOC and tabulated before and after correction by SUVmax of the blood pool (Table 3). The median LN short-axis length was 1.0 (0.60–1.40) cm; 13 (62%) LNs were normally sized. The SUVmax analysis of the mean SUVmax from each patient’s LNs after administration of ¹⁸F-FDG revealed values between 2.22 and 12.75 with a median of 3.55 (IQ = 2.01–5.09) before correction and between 1.51 and 8.55 with a median of 2.87 (IQ = 1.94–3.80) after correcting by the SUVmax of the blood pool. The SUVmax evaluation obtained from the average of the same LNs after the administration of ⁶⁸Ga-DOTATOC resulted in values between 1.35 and 3.00 with a median of 2.49 (IQ = 1.83–3.15) before correction and between 2.93 and 6.98 with a median of 4.40 (IQ = 3.79–5.79) after correcting by the SUVmax blood pool. There was no correlation between LN size and uptake in ¹⁸F-FDG (r = 0.2; P = 0.39) or ⁶⁸Ga-DOTATOC (r = −0.07; P = 0.76).

Table 3.

Mean SUVmax and cSUVmax of each patient obtained from lung lesions after administration of ¹⁸F-FDG and ⁶⁸Ga-DOTATOC

	18F-FDG		68Ga-DOTATOC
	SUVmax	cSUVmax	SUVmax	cSUVmax
Patient 1	8.16 (0.91)	5.10 (0.57)	2.31 (0.47)	4.05 (0.82)
Patient 2	9.93 (4.83)	6.66 (3.24)	1.63 (0.57)	3.54 (1.24)
Patient 3	7.25 (1.36)	5.37 (1.00)	2.26 (0.65)	4.19 (1.20)
Patient 4	4.71 (1.97)	3.08 (1.29)	2.06 (0.59)	5.01 (1.43)
Patient 5	4.09 (0.57)	5.60 (0.78)	4.06 (0.24)	7.12 (0.42)
Patient 6	7.80 (1.02)	7.50 (0.98)	2.78 (0.70)	4.56 (1.14)
Patient 7	8.06 (5.79)	4.94 (3.55)	3.37 (3.32)	7.84 (7.72)

¹⁸F-FDG = fluorine-18-labeled fluorodeoxyglucose; ⁶⁸Ga-DOTATOC = gallium-68-labeled DOTATOC; cSUVmax = SUVmax divided by the SUVmax of the vascular pool; SUVmax = standardized maximum uptake value. Values are presented as mean (SD).

We verified from the Shapiro–Wilk test that the variables logarithm of SUVmax and logarithm of cSUVmax follow a normal distribution (P = 0.1070, P = 0.9380). Such evidences are confirmed by the qq-norm graphs analysis, where quantiles of each variable are approximately aligned with the theoretical quantiles of the corresponding normal distribution. This supports the parametric analysis performed. The ANOVA tests to assess whether differences between LN and PL and the radiotracers were statistically significant for logarithm of cSUVmax (P = 0.0231, P = 0.0125) but not for logarithm of SUVmax (P = 0.0741, P = 0.057). The subgroup post-hoc analysis revealed a significant difference between LN and PL for ¹⁸F-FDG (P = 0.0416) and between both radiotracers in LN (P = 0.0211). The LN/PL ratio for cSUVmax was significantly higher for ⁶⁸Ga-DOTATOC than for ¹⁸F-FDG (P = 0.0156; Figure 2).

Figure 2.

Higher mean cSUVmax (standardized maximum uptake divided by that of the mediastinal blood pool) lymph node/pulmonary lesion (LN/PL) ratio for gallium-68_labeled DOTATOC (⁶⁸Ga-DOTATOC) than fluorine-18–labeled fluorodeoxyglucose (¹⁸F-FDG) (P = 0.0156). This figure appears in color at www.ajtmh.org.

DISCUSSION

We observed the high uptake of LN, including those normally sized, and PL with both radiotracers—all LN and PL analyzed with ¹⁸F-FDG had significant ⁶⁸Ga-DOTATOC with uptake. Few studies^13,14 with ⁶⁸Ga-DOTATOC in inflammatory diseases sought a direct comparison with ¹⁸F-FDG, but the findings of high uptake in areas with inflammatory activity are in agreement with other studies performed in atherosclerosis and sarcoidosis.^5,6 In our study, the higher SUVmax values found in ⁶⁸Ga-DOTATOC compared with previous studies⁶ may be related to the correction using an internal reference tissue to minimize measurement variability.^10–12,15

In our study, LN size was not related to tracer uptake intensity and the LN/PL uptake ratio was greater for ⁶⁸Ga-DOTATOC than for ¹⁸F-FDG. Soussan et al.¹⁶ found LN with an average size of 15 mm, above the 10 mm found in 62% of our patients, which suggests cell activity in postprimary tuberculosis regardless of adenomegaly. ¹⁸F-FDG PET/CT has revealed increased glycolytic activity in normally sized LNs in patients with latent TB, indicating that cellular activation can occur without active disease.¹⁷ Moreover, latent TB reactivation can start in the LNs and be predicted by ¹⁸F-FDG PET/CT in immunocompromised patients.¹⁸

SUV values from ⁶⁸Ga-DOTATOC have been shown to correlate with the immunohistopathological expression of SSTR, especially type 2, which is expressed on cells that play major roles in TB pathogenesis (i.e., macrophages, epithelioid cells, giant cells, and lymphocytes).⁵

Somatostatin target tissues, including lymphoid tissues, express multiple SSTR, and somatostatin analog distribution and quantification can be evaluated accurately.⁵ In TB, the LNs initiate and shape adaptive immune responses, and serve as niches for M. tuberculosis growth and persistence.¹⁸ In a ⁶⁸Ga-DOTANOC and ¹⁸F-FDG PET/magnetic resonance imaging study, Naftalin et al.⁶ found LN uptake in only four of eight patients with active TB, visible with both tracers. The lesion detection rate in our study was higher, with at least three LNs per patient taking up ⁶⁸Ga-DOTATOC, although the two tracers target similar ranges of somatostatin subtype receptors. The literature on the pathogenesis of PPT is scarcer than on latent infection and primary pulmonary TB, mainly due to the difficulty in reproducing human models in animals.² In addition, the information is restricted to immunological and cellular changes that occur in lung parenchyma, without assessing the repercussions on thoracic LN.² Thus, the greater macrophages and lymphocytes specificity of ⁶⁸Ga-DOTATOC¹⁹ compared with ¹⁸F-FDG,²⁰ and the lower ¹⁸F-FDG uptake in LN and similar ⁶⁸Ga-DOTATOC PL uptake, led us to speculate that the high ¹⁸F-FDG cSUVmax in LN represent the migration of activated defense cells without local inflammatory activity. Another factor that could, at least in part, account for the discordance in LN/PL uptake ratios (FDG versus DOTATOC) is the severe hypoxia in TB lesions, as demonstrated by Belton et al. using fluorine-18 fluoromisonidazole ([¹⁸F]FMISO).^21,22

Limitations of this study include the small sample, although our intent was to perform descriptive analysis. Additionally, performance of the PET exams on different days may have influenced comparison between the two radiotracers. However, given the slow treatment response of TB,⁶ the short intervals between exams probably did not significantly affect the results.

To our knowledge, this study was the first to describe ⁶⁸Ga-DOTATOC uptake pattern of LN and PL in PPT and compare it with the widely validated uptake of ¹⁸F-FDG. Additionally, we observed ¹⁸F-FDG and ⁶⁸Ga-DOTATOC uptake in (even normally sized) LNs in all patients with active PPT, with higher LN/PL ratios on ⁶⁸Ga-DOTATOC, which raised hypothesis related to PPT parenchymal-LN cell migration despite normal LN size. No lesion had uptake restricted to one of the radiotracers and, therefore, although ⁶⁸Ga-DOTATOC can detect TB PL and LN, it does not add diagnostic benefit but should prompt TB as a differential diagnosis in the follow-up use of SSTR radiotracers for neuroendocrine tumors. Considering that normal-sized LN in immunocompetent patients with TB are poorly studied, ⁶⁸Ga-DOTATOC may contribute to in vivo TB pathogenesis studies. New studies, however, with immunocytological evaluation are needed to confirm our results and their relevance.