Introduction

Brown adipose tissue (BAT) has the property of being metabolically activated to take up ¹⁸F-fluorodeoxglucose (FDG) [1], which can be visualised on positron emission tomography/computed tomography (PET/CT) using FDG (Fig. 1). Several studies have used FDG PET/CT to evaluate the different factors influencing BAT activation. BAT activation is more prominent in paediatric patients than in adult patients [2, 3].

View Full Size

PET/CT fused images (1), CT images (2) and PET MIP image (3) of a 9-year-old paediatric patient with osteosarcoma. Visualisation of FDG uptake in the primary tumour of the right humerus (a), metastatic lymph node in the right axillary region (b) and brown adipose tissue (BAT) in the left axillary region (c)

Citation: Imaging 2024; 10.1556/1647.2024.00262

• Download Figure
• Download figure as PowerPoint slide

Cold external temperatures have been observed to increase the possibility of BAT activation [2, 4, 5]. A sex difference was also observed in BAT activation by Cohade et al. [2] (females had a higher incidence compared to males (10.5% vs. 2.9%); however, in this study, this difference was evaluated in a mixed group of both paediatric and adult patients. The sex disparity in BAT activation seen in adults wasn't evident in studies involving only children or adolescents [6, 7].

BAT activation may also depend on the histological type of the malignancy. In a study on adult lymphoma patients, Brendle et al. in 2020 [8] found that Hodgkin lymphoma (HL) had the highest frequency of BAT visualisation (36/206 = 17%) among all types of examined lymphomas. However, no significant association was found between BAT activation and metabolic activity of the lymphoma tissue.

We previously investigated the prevalence of BAT activity in a paediatric population [7]. We analysed the FDG PET/CT images of 135 paediatric patients with HL collected at the time of diagnosis. BAT visualisation was identified in 25 of 135 cases (19%) on FDG PET/CT scans. In this study, age group; external temperature; maximal standardised uptake value (SUVmax) of the liver, spleen, spinal cord, and bone marrow; and maximal SUVmax of HL lesions were not significantly associated with BAT activation. Low median values of total metabolic tumour volume (TMTV) and total lesion glycolysis (TLG) were observed in patients with BAT visualisation, although the difference was not statistically significant (Mann-Whitney U test results showed TMTV: P = 0.07; TLG: P = 0.06).

We further investigated the frequency of BAT activation in the paediatric population. To achieve this, we analysed FDG PET/CT images of paediatric patients diagnosed with HL or sarcoma subtypes.

Materials and methods

FDG PET/CT reading

The FDG PET/CT images of the participants were analysed by a nuclear medicine physician experienced in FDG PET/CT evaluation. Activated BAT was identified by high FDG uptake in the adipose tissue, characterised by negative Hounsfield units on CT scans, whereas the surrounding soft tissue showed a lower intensity of FDG uptake (Fig. 2).

View Full Size

FDG uptake according to BAT on PET/CT images of a 16-year-old paediatric patient with HL. High FDG uptakes in adipose tissue corresponding to negative Hounsfield units (HU) on CT (sign of activated BAT) on both sides in the supraclavicular regions (a). Fused FDG PET/CT images show FDG accumulations according to soft tissue masses on the left side of the neck (part of HL) (b)

Citation: Imaging 2024; 10.1556/1647.2024.00262

• Download Figure
• Download figure as PowerPoint slide

BAT activation was graded as a binary outcome (yes/no) for each patient without scaling the extent and intensity of the hyperfunctional foci observed on FDG PET/CT.

The investigator measured the FDG uptake of the tumour lesions and the following variables [9]: TMTV and TLG.

A volume of interest was set around each malignant lesion (lymph node or organ). TMTV was determined by summing the volume of all such lesions (tumour, lymph node, organ involvement) with a threshold of SUV ≥ 2.5. Bone marrow and spleen involvement were included in the volume measurement only if focal well-determined FDG uptake was present. TLG was calculated as the sum of the products of the metabolic volume by the SUVmean of each malignant lesion. Segmentations were performed using the Mediso Interview software.

Results

The frequency of BAT visualisation on bPET was 22/92 = 23.9% (CI: 15.2%–32.6%). A total of 12 out of 63 HL patients showed signs of BAT activation (12/63 = 19%), and 10 out of 29 sarcoma patients had signs of BAT activation (10/29 = 34.4%). Fischer's exact test performed to determine the prevalence of BAT activation in patients with HL compared with that in patients with sarcoma revealed no significant difference (P = 0.12).

To investigate potential factors influencing BAT activation, we conducted separate analyses on the entire patient cohort, the subgroup diagnosed with Hodgkin lymphoma, and the subgroup diagnosed with sarcoma. These analyses evaluated the influence of age, sex, external temperature on the day of the PET/CT examination, TMTV and TLG of the malignant tissue.

The Chi-2 test showed no significant correlation with season (whole group P = 0.18; HL subgroup P = 0.6; sarcoma subgroup P = 0.09), and there was no significant association between external daily temperature and BAT activation on FDG PET/CT.

For the TMTV and TLG of the malignant tissue, the Mann-Whitney U test showed no significant association with BAT activation in the group that included paediatric patients with HL and sarcoma (Table 1). However, by analysing the subgroup of patients with Hodgkin disease, we found a trend towards lower TMTV (P = 0.06) and TLG (P = 0.09) (Table 2). Regarding the association between BAT activation and TMTV, TLG values there was a difference between the HL and sarcoma subgroups (Table 2). In the HL subgroup, patients with BAT activation showed lower TMTV, TLG values (TMTV mean: 166.58 cm³; TLG mean: 760.24 g) than HL cases without BAT activation (TMTV mean: 384.32 cm³; TLG mean: 1765.08 g). In the sarcoma subgroup, patients with BAT activation showed higher TMTV, TLG values (TMTV mean: 263.89 cm³; TLG mean: 1352.87 g) than patients with sarcoma without BAT activation (TMTV mean: 180.19 cm³; TLG mean: 799.23 g). However, the Mann-Whitney U test showed no significant association between TMTV, TLG and BAT activation in the sarcoma subgroup. This could be attributed to the limited number of patients with sarcoma.

In the logistic regression, the aforementioned trends were not observed for TMTV and TLG in any group; however, a significant correlation between patient age and BAT activation was found in the HL subgroup (P = 0.04), and in the case of the whole group a P value near to the value considered as significant (P = 0.06) was observed. This correlation with age was not observed in the subgroup of patients with sarcoma (P = 0.72).

The logistic regression model showed a significant inverse relationship between the winter season and autumn reference season for FDG uptake in BAT in the whole group (Log. Odds −2.13; P = 0.02). A similar, but not significant trend was observed in patients with HL (log. Odds −2.45; P = 0.08). No similar relationship was observed in the subgroup of patients with sarcoma (probably because of the smaller sample size).

When analyzing the seasonal distribution of BAT activation on PET/CT scans, we observed that 45% (10 out of 22) of the PET/CT examinations demonstrating BAT activation were performed during the autumn season. Interestingly, 38% (10/26) from all PET/CT scans conducted in the autumn period showed signs of BAT activation (Fig. 3).

View Full Size

BAT activity as a function of season (mean, mean ± sd, min and max values). FDG PET/CT examinations performed at diagnosis showed the highest rate of BAT activation in autumn

Citation: Imaging 2024; 10.1556/1647.2024.00262

• Download Figure
• Download figure as PowerPoint slide

The prevalence of BAT visualisation did not differ between boys and girls, either in the whole group (10 out of 48 boys (10/48 = 20%); 12 of 44 girls (12/44 = 27%); P = 0,63) or in the subgroups of HL (4 out of 31 boys (4/31 = 13%); 8 out of 32 girls (8/32 = 25%); P = 0.36) or in the subgroups of sarcoma disease (6 out of 17 boys (6/17 = 35%); 4 out of 12 girls (4/12 = 33%); P = 1).

Discussion

Our study identified BAT visualization in 23.9% of the patient cohort. This finding is consistent with the results reported by Cohade et al. [2], where activated BAT was detected in 10 out of 42 patients (23.8%) who were 18 years of age or younger. This prevalence is similar to that of BAT activation in children and adolescents reported by Yeung et al. [3] (4/26 = 15%) and that reported in a previous study involving 135 paediatric patients with HL (25/135 = 19%) [7]. In a pilot study published in 2016 [10] involving 30 paediatric patients with HL, which aimed to evaluate the potential importance of non-disease-specific FDG uptake in HL activity and prognosis, activation of BAT at diagnosis was found in only two patients (2/30 = 7%). This result is consistent with that of Gilsanz et al. [4], who focused on FDG PET/CT examinations of paediatric patients with HL (n = 21) or non-HL (n = 10) and compared the prevalence of metabolically active BAT at the time of diagnosis and when there was no evidence of disease after effective treatment in the same patient population. The results showed that the overall prevalence of BAT visualisation on FDG PET/CT at the time of diagnosis was 3/31 = 10%.

Yeung et al. [3] investigated the frequency of BAT activation in the area of the neck and found a significantly higher incidence in the paediatric population (4/26 = 15%) compared to the incidence in the adult population (16/837 = 1.9%). A higher prevalence was reported by Drubach et al. [6], who evaluated the FDG PET/CT images of 172 patients aged 5–21 years. Contrary to the results of other studies in adults, they found a high prevalence of BAT visualisation (76/172 = 44.2%), but no significant difference was found between boys (42/97 = 43.3%) and girls (34/75 = 45.3%). Gilsanz et al. [4] also reported a higher frequency of BAT activation in treated paediatric patients with HL (18/21 = 86%) and in a mixed group with treated paediatric patients with HL or non-HL (24/31 = 77%). The inclusion of post-treatment FDG PET/CT examinations may explain this higher prevalence in both of these studies [4, 6].

The Mann-Whitney U test demonstrated a trend towards statistical significance in the correlation between lower values of TMTV and TLG of the tumour mass and the appearance of FDG uptake by BAT on PET/CT only in the subgroup of patients with HL. Compared with the results of our study from 2022 in a paediatric patient group involving only HL patients [7], we found that P-values for TMTV and TLG were P = 0.07 and 0.06, respectively, and the results of the current study showed P-values of the Mann-Whitney U test of P = 0.06 and 0.09, respectively, for the subgroup of HL patients. This finding for TMTV and TLG agrees with the results of Gilsanz et al. [4], suggesting that BAT visualisation could be associated with a non-detectable or lower HL tumour mass. However it should be noted that TMTV and TLG values are not completely independent.

Using similar univariate methods, we were able to replicate the results of our previous study regarding TMTV and TLG values in patients with HL [7]; however, these associations could not be demonstrated in a multivariate setting in a regression context. This difference is due to the ability of the regression model to not only assess the individual effects of predictors, but also to control for potential confounding effects between variables and to measure the complex interaction of all predictors in the model, thus providing a holistic picture of the underlying relationships between variables.

The Mann-Whitney U test showed no statistically significant trend for TMTV or TLG values in the whole group that included patients with HL and sarcoma. Also, no statistical significance was found in the subgroup of sarcoma patients, but a difference in the association between BAT activation and TMTV, TLG values was observed between the subgroup of HL patients and the subgroup of sarcoma patients. However, for the HL subgroup, BAT activation appears to be associated with lower TMTV and TLG median or mean values and in the sarcoma subgroup, BAT activation was associated with higher TMTV and TLG median or mean values compared to cases without BAT activation from the corresponding subgroup, statistical calculations did not confirm this point. Our findings for the sarcoma subgroup may be limited by the relatively small sample size.

The present study showed no statistically significant correlation between TMTV, TLG values of sarcoma mass and FDG uptake by BAT on PET/CT. The results may not be sufficiently powered due to the low number of sarcoma patients. This hypothesis should be evaluated in a larger group of sarcoma patients. In terms of BAT activation, the overall results of the current study suggest that there may be a difference between more focalised cancer tissues, such as sarcoma types and malignancies with more spreaded tumour cells, such as HL.

We found no difference in BAT activation frequency between male and female patients, supporting the previous observation that there is no significant difference between the sexes in the paediatric population. However, it stands in contrast to reports from studies involving adults, such as the one conducted by Cohade et al. where a sex difference was observed. [2]. In their study, women had a higher incidence (53/504 = 10.5%) than men (15/513 = 2.9%).

Analysing the frequency of BAT activation on PET/CT by season, we found that 45% of the PET/CT examinations that showed signs of BAT activation were performed in autumn, and the numbers of PET/CT examinations with signs of BAT activation in autumn accounted for 38%. However, the logistic regression model showed a significant inverse relationship between the winter and reference autumn seasons for FDG uptake by BAT in the entire group (log. Odds −2.13; P = 0.02). There was only a P value near to the value considered as significant in the subgroup of patients with HL, probably because of the smaller sample size (log. odds ratio −2.45; P = 0.08) and the subgroup of patients with sarcoma did not show a similar relationship.

Our results do not correlate with observations from studies such as the one by Cohade et al. [2], in which the incidence of activated BAT was higher in winter, in the period from January to March (38/278 = 13.7%) than in the rest of the year (30/739 = 4.1%), or with the results of Gilsanz et al. [4], in which BAT was found to have a higher incidence in winter months (November, December, January) than the rest of the year (22% vs. 4.5%). Seki et al. [5] demonstrated that cold-induced BAT activation substantially decreases blood glucose and impedes the glycolysis-based metabolism in cancer cells. This finding may even explain the results of our previous study and those of the current one that show a higher frequency of BAT activation in HL cases with low TMTV, TLG.

The temperature in the injection room of the nuclear medicine department could be an important consideration in terms of the effect of the external temperature on BAT activation. In a study by Zukotynski et al. [11], 103 patients were warmed up to 24°C prior to FDG PET/CT scanning. They compared the uptake of FDG by BAT in this group and a control group of 99 patients who underwent FDG PET/CT when the injection room temperature was as low as 21 °C. After patient warm-up, FDG uptake by BAT occurred in 9% of the PET/CT examinations compared to 27% of PET/CT examinations in the control group (P << 0.01). They concluded that by keeping the room temperature constant at 24 °C for 30 min before and 1 h after intravenous tracer administration significantly decreased FDG uptake by BAT in children. This effect was the greatest in winter and summer. In 2023, Pötzsch et al. [12] reported that in their study involving 528 FDG PET/CT scans from 241 children or adolescents with lymphoma, age, sex, and body mass index had no clear impact on the visualisation of BAT activation. However, a logistic regression model showed that the frequency of BAT activation primarily depends on the outside temperature (P = 0.005) and can be effectively reduced by warming up (P = 0.004), administering nonselective beta-blockers, or a combination of both. They also concluded that the effect of warm-up decreased with increasing outside temperature.

As in our study, the temperature in the injection room of the PET/CT centres was not recorded, and its effect on BAT visualisation could not be controlled; however, heated rooms in winter may influence the appearance of BAT activation.

Pre-medications applied for PET/CT examinations in children can be also an important factor influencing the FDG uptake by BAT. Sedatives (nalbuphine, midazolam, fentanyl) is widely used in paediatric imaging as an anxiolityc [13]. FDG uptake in BAT can be reduced by such pre-medications, and some other drugs are used intentionally to reduce the high FDG uptake in BAT on PET images (propranolol or reserpine) [14]. In the current study no such pre-medication was applied, thus we can not report on this possible influencing factor, but it is necessary to mention its possible effects on interpretation of BAT activation.

The univariate analysis showed no difference by age; however, the logistic regression analysis suggested that there may be a correlation with the age of the patients in this paediatric population (age range 3–17 years). We found a significant correlation between patient age and BAT activation in the HL subgroup (P = 0.04), and no statistical significant but a P value near to value considered as significant in the entire group (P = 0.06). Gilsanz et al. [15] reported that the appearance and amount of brown fat increases during puberty. They examined the PET/CT scans of patients aged 4–19.9 years who had been previously treated for paediatric malignancy, but were disease-free at the time of examination. This study included 73 patients with heterogeneous diseases (lymphoma and sarcoma as well). In this study, only the last follow-up examinations of disease-free patients with normal PET/CT findings were analysed. They concluded that metabolic and hormonal events associated with the attainment of sexual maturity were likely responsible for the rapid increase in brown fat observed during puberty.

Conclusions

The results of the current study confirm previously published evidence from heterogeneous conditions that the epidemiology of BAT activation differs between paediatric patients and adults. In particular, we found no difference between boys and girls regarding BAT visualisation. In this study, the results suggest that BAT visualisation may be associated with a lower HL tumour mass, but this hypothesis could not be confirmed when combinations of several quantitative variables were tested simultaneously. In terms of the connection between season or outside temperature and BAT activation, the present study also highlighted the importance of influencing factors such as the temperature in injection rooms in PET/CT laboratories. Multivariate analysis raised the possibility of a correlation between BAT appearance and age in this paediatric population of patients with HL, which may influence the results in the whole group.

Authors' contribution

LJ – Concept drafting, Data management, Statistical analysis, Editing of the manuscript, Preparation of the figures.

BM – Data management, Editing the manuscript, Proofreading, Preparation of the figures.

MP – Statistical analysis, Editing of the manuscript, Proofreading, Preparation of the figures.

SC – Data management, Editing of the manuscript, Proofreading.

TG – Concept Drafting, Editing the manuscript, Critical revision.

All authors reviewed the final version of the manuscript and agreed to submit it for publication.

Funding sources

No financial support was received for this study.

Conflict of interests

The authors have no conflict of interest to disclose.

Ethical statement

As this was a retrospective analysis, the approval of the institutional ethics committee has been waived. The studies reported in the manuscript were conducted with a clear and valid clinical indication (staging) according to standards and patients gave written informed consent of performing the scan.