Introduction

The kidneys are complex metabolic organs that perform osmoregulatory and key endocrine functions [1]. They effectively filter, remove, and reabsorb both essential and non-essential solutes from plasma, while contributing to the regulation of the acid–base balance in the body [2, 3]. They also play important roles in the metabolism of carbohydrates, proteins, and other nutrients [4]. Given their functional and metabolic roles, diseases affecting kidneys such as acute or chronic renal failure usually result in complex deleterious alterations in the overall physiological and metabolic processes in the body [5].

The kidneys utilize several different substrates such as lactate, free fatty acids, glutamine, 3-hydroxybutyrate, glycerol and citrate for the functional activities [6]. The preferred substrates for cellular respiration in the renal cortex include both short- and long-chain fatty acids and some amino acids [7]. In the outer medulla, lactate appears to be the preferred substrate over glucose [7]. And in the inner medulla, oxygen consumption is much lower than in the cortex, while glucose is the preferred metabolic substrate [6].

During the postabsorptive period, the kidneys account for 10% of the whole-body glucose utilization. Normally, approximately 180 g of glucose are filtered every 24 h and almost all of it are reabsorbed by means of the sodium/glucose cotransporter 2 (SGLT2) system, which is expressed in the early proximal tubules of the nephron [2, 3]. Most of the remaining glucose absorption is by sodium/glucose cotransporter 1(SGLT1) located at a more distal sections of the proximal tubules [8]. Importantly, the reabsorption of sodium ions is mainly due to the extremely high oxidative metabolism of the renal system [9].

Metabolic disease/disorder is characterized by the disruption in normal metabolism which affects the performance of critical biochemical reactions involving the processing or transport of proteins (amino acids), carbohydrates (glucose), and lipids (fatty acids) [10]. Moreover, metabolic disease is associated with oxidative stress and mitochondrial dysfunction [10]. Previous studies have reported that individuals with metabolic abnormalities are at an increased risk of renal dysfunction, including chronic kidney disease which may lead to kidney failure and end-stage renal disease [11,12,13,14].

Positron emission tomography (PET) is a nuclear medicine-based imaging technique that can functionally evaluate metabolic processes within the body by the use of non-nephrotoxic tracer compounds [15, 16]. Its application has been extensive in the study of a variety of organs, and it is currently considered the gold standard for measuring tissue metabolic rates in vivo. PET data can be analyzed either semiquantitatively (through standardized uptake rate-SUV) or quantitatively (through graphical analysis, fractional uptake rate, compartmental models). The latter requiring determination of plasma input function, thus tracking the tracer uptake in the tissue of interest in respect to its availability in the circulation [17].

While PET applications in studies of brain [18], heart [19], liver [20], skeletal muscle [21] and adipose tissue [22, 23] are well established, there have been challenges in its implementation for the study of the kidneys. We believe, however, that renal PET has begun to demonstrate its unique potential in providing useful information in a relatively noninvasive manner. The present systematic review, therefore, aimed to identify and summarize all available studies that have evaluated kidney physiology (i.e., renal perfusion, glomerular filtration rate, glucose, and fatty acid metabolism) and pathophysiology (i.e., renal failure, acute rejection following kidney transplantation, renal cysts) though PET imaging in non-oncological patients with metabolic disorders.

Methods

Eligibility criteria

Peer reviewed publications reporting on the utilization of PET imaging for the evaluation of renal (patho) physiology in humans (namely renal perfusion, glucose, and fatty acid metabolism) from 1980 to 2021 (based on the time span of the technique’s existence) were included in this review. Editorials, letters, commentaries, perspectives, and conference abstracts were excluded together with PET studies involving oncological diseases with renal involvement. Only articles written in the English language were considered.

Information sources and search strategy

  1. 1.

    Words that relate to the system: renal, kidney, nephrology.

  2. 2.

    Words that relate to the technique: positron emission tomography.

  3. 3.

    Words that relate to the studied function: renal metabolism in metabolic disorders.

We searched the MEDLINE (through PubMed), Science Direct and Cochrane databases. These databases were selected, because they provide extensive international journals and are regularly updated with relevant resources covering medicine, health sciences and related topics. Reference lists of relevant articles were hand-searched to identify additional articles. Relevant articles were then exported to and managed in the RefWorks Citation Manager®.

Study selection

After the deletion of duplicates, all titles and abstracts were duly reviewed and screened against the eligibility criteria. All ineligible articles were removed. Full texts of all relevant articles were then retrieved.

Data items and collection process

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) statement [24]. Data were extracted into a standardized data collection form. Extracted data included: first author, publication year, demographics, study design, limitations, major findings, conclusion, and remarks. Each study was read and reread to identify core methodological aspects, perform data reduction and when possible, some comparisons. Two reviewers independently abstracted the contents of each included study. The risk of bias was appraised using Joanna Briggs Institute Critical Appraisal Checklist for Qualitative Research [25]. Discrepancies and disagreements were resolved through a third independent researcher.

Results

A total of 2502 articles were identified. After removal of duplicates and screening based on titles and abstracts, 73 articles underwent a full-text review out of which thirty met the inclusion criteria (Fig. 1, Table 1). Designs included prospective (73.33%, n = 22), retrospective (16.12%, n = 5), cross-sectional (6.45%, n = 2), and observational studies (3.22%, n = 1) studies. Most of the studies used 15O-labeled water (n = 14) or 18F-fluorodeoxyglucose (18F-FDG, n = 8) as the PET radiotracer. Tracers that were use in at least on study included 15O, 11C-acetate 18FFTHA, RadioCF, 82Rb, 11C-PABA, F-NaF, 18F-FDS, 68Ga-pentixafor, 82Rb, 13N-ammonia, and 68Ga-EDTA (Table 1).

Fig. 1
figure 1

Study flow chart

Full size image
Table 1 Description of the utilization of positron emission tomography in renal system
Full size table

Renal perfusion

A radiopharmaceutical is suitable for renal perfusion quantification if it shows high perfusion-rate-independent extraction from arterial blood in its first-pass transit through the renal capillaries [26]. The most commonly used PET tracer for perfusion measurement has been oxygen-15 radiolabeled water (H215O, radiowater) n = 15 [16, 27,28,29,30,31,32,33,34,35,36,37,38,39,40]. One study also reported that carbon-15 labeled acetate (11C-acetate) can be used for the same purpose [37]. A one-compartment model was used for the quantification of these tracers. Other tracers, which are not entirely freely diffusible, have also been used for perfusion quantification through implementation of the two-compartment Patlak model namely, Nitrogen-13 labeled ammonia (13N-ammonia) n = 1 [28] and rubidium-82 chloride (82Rb) n = 2 (which behaves like a K+ analogue and is very commonly used to measure myocardial perfusion) [41, 42].

Energy substrate metabolism

In a recent study that aimed to automatically quantify tracer uptake during euglycemic hyperinsulinemic clamp conditions, volume estimation by 18FDG was comparatively inaccurate in the kidneys [43]. This was largely due to spillover from urine. Furthermore, 14(R, S)-[18F] fluoro-6-thia-heptadecanoate (18F-FTHA) has been used to estimate cortical and medullar fatty acid uptake (n = 1) [38], while oxygen consumption has been quantified with 11C-acetate n = 1 [37].

Inflammation

18FDG has been shown to be useful in the noninvasive diagnostic workup in suspected allograft rejection n = 2 [44, 45]. In a small proof-of-concept study that included only two human subjects, Werner et al. suggested 2-deoxy-2-[18F] fluorosorbitol (18F-FDS) as an alternative to 18F-FDG in pediatric patients due to a lower radiation exposure [46]. In one study, gallium labelled 68Ga-pentixafor has been proposed as useful for detecting leukocytes in allografts when PET is combined with magnetic resonance imaging (MRI) [47].

Kidney function

In a study by Geist et al. [48], both the cortex and total kidney 18FDG glomerular filtration rate obtained from regression analysis of Patlak plots were shown to have strong positive correlations with the reference value for estimated glomerular filtration rate (eGFR) [48]. N = 2 studies have quantified renal clearance rate of 18FDG [49, 50]. One study calculated kidney function parameters from 18FDG standardized uptake values and time activity curves [51]. Parallelly, Minamimoto et al. [52] showed that mean 18FDG SUV in the kidney is higher in healthy volunteers than in patients with suspected renal failure [52]. GFR has been quantitated in n = 1 study with 68Ga-EDTA [53], while 11C-PABA has been shown to be safe in humans and could be used for functional renal imaging [54]. In a study by Baldelomar et al. [55], a mouse model was used to map nephron mass as proof of concept in a donated human kidney in vivo with RadioCF [55]. Higher renal artery F–Na–F uptake has also been associated with higher cardiovascular risk and lower GFR [56].

Discussion

Our systematic review assessed the application of PET imaging in the evaluation of renal metabolism in metabolic disorders based on 30 studies. The following key finding emerged. First, the largest proportion of studies reporting on kidney metabolism with PET were prospective reports. Second, 15O-labeled water and 18F-fluorodeoxyglucose were the most commonly used radiotracers for kidney metabolism. Third, a range of clinical features such as renal inflammation, glomerular filtration rate, and renal pathologies such as acute and chronic kidney diseases were studied with PET. Our findings reinforced the fact that PET imaging holds the potential for the assessment of renal function, and for the diagnosis of renal pathologies.

Renal physiology

Many tracers are cleared through the kidneys, which is why uptake appears high. This is the case with the widely used tracer 18F-FDG. Several experimental tracers, most of which likely have no or very little true uptake in the kidneys, also show high renal uptake: 68Ga-NODAGA-exendin-4 and 18F-FP-(+)-DTBZ, used for β-cell imaging [57, 58], 18F-Fluoro-L-DOPA, studied in infants with congenital hyperinsulinism [59], 11C-PK11195, used for assessing peripheral benzodiazepine receptor activity [60, 61], 18F-AH111585, a peptide with a high affinity for the αvβ3 integrin receptor involved in angiogenesis [62], 1C-PE2I, a cocaine analog tracer [63], 11C-PIB and 18F-FPYBF-2, which are amyloid imaging agents [64, 65], 11C-nicotine, used to detect nicotine receptors [59], 11C-metformin, used for studying the biodistribution and excretion of metformin [66], 11C-MP4B, a tracer for the assessment of butyrylcholinesterase activity [67], 86Y-SMT487, a radiolabeled somatostatin analogue [68] and 68Ga-NOTA-UBI, a potential tracer for diagnosing infectious processes [69].

Renal perfusion

Clearance-based methods used to estimate effective renal plasma flow (ERPF) are time consuming and stressful to the patient [48]. PET provides a noninvasive method for the quantification of renal perfusion [70]. Alternatively, current modalities such as magnetic resonance imaging (MRI) arterial spin labelling (ASL) technique utilizes magnetic labelling and endogenous tracer of water in arterial blood to generate maps of absolute regional perfusion without the use of exogenous contrast [71].

15O-H2O, also called “radiowater”, is the most commonly used tracer for measuring renal perfusion [33]. Damkjær et al. showed that even medullar blood flow can be separately assessed with radiowater [34]. Several studies investigating physiological mechanisms have been conducted with radiowater. Modest saline loading, which leads to a natriuretic response, did not always induce changes in renal perfusion in seven healthy subjects [34]. When the ACE inhibitor quinaprilat was given as an injection to n = 8 patients with hypertension and moderate chronic heart failure, renal blood flow was increased [32]. Enalapril also increases renal perfusion [33]. Middlekauff et al. observed decreased renal perfusion due to vasoconstriction in n = 29 healthy subjects in response to a static handgrip exercise [30]. The magnitude and duration of this effect were exaggerated in n = 39 heart failure patients, compared to n = 38 controls [31]. Middlekauff et al. [29] also investigated pulmonary baroreflex in n = 8 heart failure patients and noted that reflex control of circulation in the forearm but not kidneys was abnormal compared to n = 10 controls [29]. Radiowater was used to measure regional hemodynamics and metabolism in obese subjects and non-obese controls to determine the effects of major weight loss before and after bariatric surgery [38]. Cortical perfusion was found to be significantly higher as compared with medullary perfusion in both groups. Even though cortical and medullary blood perfusion rates (ml/100 g/min) were not different between the two groups, total renal blood flow was higher in the obese individuals, and it was significantly decreased following weight loss [38].

Three other PET tracers have also been shown to be usable in humans for the assessment of renal blood flow. Normand et al. validated a modeling method for [11C] acetate to measure perfusion, using radiowater as gold standard [37]. The benefit of the tracer is that it can also be used to measure oxidative metabolism and can thus provide information on two different functions [37]. 13N-ammonia has also been validated for measuring perfusion [28], but has since not been utilized in kidney studies. It is instead used in myocardial perfusion and blood flow studies [72]. 82Rb is also most commonly used to assess myocardial blood flow in patients suspected of ischemic heart disease, but 82Rb is also appropriate for modelling renal blood flow using dynamic PET methods, since the method has shown high image quality [41]. In a study assessing myocardial perfusion, regadenoson was shown to also increase kidney perfusion [42]. Another potential tracer for evaluating renal perfusion would be 62Cu-ETS [35], but so far validation results in humans comparing the tracer with radiowater, have only been presented as a meeting report [40].

Renal metabolism

Renal substrate metabolism is closely tied to renal hemodynamics; the well-perfused renal cortex is rich in mitochondria and depends predominantly on oxidative metabolism—mainly of fatty acids, but also of ketone bodies, and lactate, whereas the renal medulla, which operates in relative hypoxia, relies predominantly on glycolysis [73]. The proximal tubules are also equipped with the enzymes responsible for gluconeogenesis (glucose 6-phosphatase, fructose 1,6 diphosphatase, and phosphoenolpyruvate carboxykinase), making the kidney one of the organs which contribute to endogenous glucose production after the liver, and also to a lesser extent the intestines [74].

PET provides a method for investigating kidney energy metabolism in humans in vivo, yet very few studies have addressed these questions so far. 11C-acetate can be used for direct estimates of renal oxygen consumption [37]. It has recently been shown that renal FFA (i.e., cortical, and medullary) uptake using FTHA is increased in morbidly obese subjects compared to lean individuals, and that 6 months following bariatric surgery, renal FFA uptake remains high, because of the ongoing catabolic state [38]. 18F-FDG is a widely available tracer, which can be used for studying metabolic shifts in glucose utilization [75]. To the best of our knowledge thus far, there has been no study evaluating renal glucose uptake using 18FDG-PET. 18FDG is excreted into urine and thus addressing renal metabolism with 18F-FDG with current modeling methods is considered thus far to be heavily biased [76]. If the challenges with modeling kidney tracer uptake could be overcome, 18F-FDG could be used for studying metabolic shifts in kidneys as well. Since kidney energy metabolism is suggested to take part in the development of pathologies such as renal hypertension [74], PET has unused potential in investigating the development of these diseases in humans.

Renal function

Glomerular filtration rate (GFR) is defined as the rate of plasma flow through the glomerulus into the urinary space of the Bowman’s capsule and is the most effective index and key indicator for renal function [77]. GFR is of high clinical significance especially in the detection, treatment, and prevention of kidney disease, but it cannot discern between one- or two-sided defects in filtration rate.

The most commonly used PET radio tracer for examinations of renal functions is 18FDG. Pharmacokinetically, 18FDG enters the kidney through renal arteries, is filtered in the glomeruli, then partially reabsorbed in the proximal tubule, and finally excreted in the urine [78, 79]. As partly mentioned in the results, early 18FDG clearance (from injection) may also be used to estimate glomerular filtration rate (GFR) [48]. Both the cortex and total kidney 18FDG glomerular filtration rate [GFR(FDG)] were shown to have strong positive correlations with the reference value for GFR [GFR(ref)] [48]. The GFR(FDG) was obtained from a regression analysis of Patlak plots [48], whereas the GFR(ref) was estimated from creatinine values using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [80]. Based on this, it was estimated that GFR and EPRF can be assessed with 18FDG-PET in healthy kidneys [48]. FDG renal clearance has since been utilized to also estimate urinary glucose loss [81]. Rasul et al. have shown that mean transit time of 18FDG decreased with 2 weeks of SGLT2i treatment in T2D patients, shifting the transit time towards that of healthy controls [50].

Renal failure is characterized by the significant loss of kidney function with a GFR of less than 15 ml/min/1.73m2 [82]. Available treatment options for end-stage renal failure include kidney transplantation, hemodialysis, and peritoneal dialysis [83]. Although 18FDG-PET/CT can differentiate kidneys with end-stage renal disease from healthy kidneys, this by itself would be of limited clinical utility for patients who are already dependent on renal replacement therapy. Research shows that in patients with normal functioning kidneys, roughly 10% of the injected 18FDG is excreted in the urine in about 70 min after the tracer injection [84]. However, in individuals with chronic kidney failure (CKF) who are on hemodialysis, urinary 18FDG excretion is impaired and hence these subjects experience constant volume overload [85]. The 18FDG tracer is distributed more in regions such as brain, and less in the cardiovascular system compared to control subjects without CKD [52].

Assessment of GFR can also be done using 68Ga-EDTA [53] and 68Ga [15]. This agrees with Hofman et al. [86] who reported that 68Ga-GFR agreed well with 51Cr-GFR for estimation of GFR. It was also stated that PET dynamic imaging offers a method to estimate GFR with an added advantage of enabling renal imaging in a single study [86]. There was good agreement between 68Ga and 51Cr- EDTA GFR despite injection of both radiotracers simultaneously and the use of the same samples for positron and gamma counting [86]. Hofman et al. [86] showed that 51Cr- EDTA overestimates a GFR by about 10%. Despite this occurrence, Goethals et al. [87] stated that 51Cr-EDTA is accepted as the most reliable agent to study whole kidney GFR. The ability of PET to quantify activity in three dimensions makes it ideal for applications such as GFR estimation [86].

Renal vasculature

Renal artery stenosis is a form of atherosclerotic vascular disease and can result in kidney insufficiency and renal hypertension [36]. Few studies have addressed effects of renal vasculature with PET in humans. Higher renal artery wall F-NaF activity was associated with CV risk and inversely associated with GFR in a group with high CV risk [56]. In an observational study by Koivuviita et al. [36], it was shown that angiographic severity of renal artery stenosis did not determine the perfusion response to revascularization. They hypothesized that it might be due to concomitant microvascular disease [36].

PET as a diagnostic tool

We have previously discussed the impact of tracer secretion in renal insufficiency and have identified how FDG secretion is impaired. This could be of concern, since FDG is a commonly used tracer in oncology and in the diagnostic work up of fever of unknown origin, and renal failure is very common in a clinical setting. Although compromised kidneys tend to alter the biodistribution of drugs cleared by the renal systems, the presentation of renal diseases do not necessarily impact FDG uptake values in other organs and, therefore, tracer dose adjustment is not required in such subjects [83].

Two studies have assessed 18FDG PET/CT in the diagnostic work up of fever of unknown origin in end-stage renal disease patients, treated with dialysis. In a retrospective study on n = 22 patients, 15 patients showed metabolically active lesions and the scans lead to a change in treatment in 21 patients [88]. In a retrospective study on n = 46, 22 out of 46 scans identified the cause and a higher C-reactive protein level was predictive of a positive outcome [89]. The clinical utility was comparable to unselected patient populations. In a case study, including seven PET scans on three patients with autosomal dominant polycystic kidney disease, 18FDG PET could be used to diagnose or rule out cyst infection [90]. High 18FDG uptake can also be caused by urinary calculi [91].

Slowly progressing chronic kidney disease is common in kidney transplants, but PET has not been widely utilized for assessing this pathology. A pilot study showed that PET could measure kidney transplant perfusion and that microvascular dysfunction could be detected in transplants [39]. Larger studies are needed on transplant patients and patients with advanced CKD before the method can be implemented in clinical use.

Acute kidney injury caused by acute rejection (AR) is also common in renal transplants. Current immunosuppressive treatments can efficiently treat AR, and thus the timely diagnosis of AR is fundamental. In acute rejection, there is accumulation of activated leucocytes in the transplant, which are characterized by an increased avidity for glucose (and 18FDG) [92]. Diagnosis of AR is based on transplant needle biopsy, which is an invasive procedure. PET could, therefore, provide a noninvasive means for directing further diagnostic testing.

In a preclinical study in rats, Reuter and colleagues demonstrated, that 18FDG-PET uptake was increased in transplants which were undergoing AR [93]. This finding was later confirmed in a study in n = 31 humans [45]. However, in this study, even though the sensitivity of 18FDG-PET was 100%, the specificity was only 50% [45]. Jadoul et al. [44] later showed in a prospective n = 95 study that 18FDG is usable for disproving acute rejection since quantification was repeatable and reproducible [44]. Parameters obtained from 18FDG PET/MRI have also shown a possible predictive feature for renal recovery in solid organ kidney transplantation patients undergoing acute kidney injury [51]. 68Ga-pentixafor has been investigated in 13 patients with renal allografts and complicated urinary tract infection, to detect leukocytes in grafts [47].

A cross-sectional diagnostic test study by Aaltonen et al. [94], where 26 dialysis patients suspected for renal osteodystrophy were studied with 18F-NaF, showed that PET could provide a noninvasive diagnostic tool in patients that usually undergo bone biopsy for diagnosis [94].

Strengths and limitations

This review has followed the recommendations for the systematic search of articles and the PRSIMA statement, which is internationally recognized. The limitation of this study is that only single studies have been conducted for several PET tracers, such as 11C-acetate, 18FFTHA, 18F-NaF, RadioCF, 11C-PABA, 13N-ammonia 18F-FDS, 68Ga-pentixafor and 68Ga-EDTA, which importantly reduced the analytical and comparative possibilities of our report. Nevertheless, we have maintained the descriptive axis of a systematic review to inform about this publication pitfalls and current knowledge gaps.

Conclusion

PET imaging provides an effective modality for evaluating a range of metabolic functions including glucose and fatty acid uptake, oxygen consumption and renal perfusion. Patients may benefit from undergoing PET imaging of the renal system for a variety of purposes, including the assessment of renal function and metabolism, disease diagnosis and prognosis. However, several useful tracers, however, have only been marginally utilized in clinical studies. PET imaging is an effectively growing, and currently underused modality for the noninvasive evaluation of renal (patho)physiology and its expansion could benefit non-oncological patients who suffer from metabolic diseases.