Comparison of static F-FDG-PET/CT (SUV, SUR) and dynamic F...

1

Comparison of static 18F-FDG-PET/CT (SUV, SUR) and dynamic 18F-FDG-PET/CT (Ki) for quantification of pulmonary inflammation in acute lung injury

Anja Braune1,2, Frank Hofheinz3, Thomas Bluth1, Thomas Kiss1, Jakob Wittenstein1, Martin Scharffenberg1, Jörg Kotzerke2, Marcelo Gama de Abreu1

1Pulmonary Engineering Group, Department of Anesthesiology and Intensive Care Medicine, University Hospital Carl Gustav Carus at the Technische Universität Dresden, Dresden, Germany

2Department of Nuclear Medicine, University Hospital Carl Gustav Carus at the Technische Universität Dresden, Dresden, Germany

3Helmholtz-Zentrum Dresden-Rossendorf, PET Center, Institute of Radiopharmaceutical Cancer Research, Dresden, Germany

Corresponding Author / Reprint request: Anja Braune Department of Nuclear Medicine University Hospital Carl Gustav Carus Fetscherstr. 74 D-01307 Dresden Phone: +49 (0)351-458 12128 Fax: +49 (0)351-458 5310 E-mail: [email protected]

Short running title: 18F-FDG-PET/CT imaging in acute lung injury

Words: 4993, 345 (abstract)

Journal of Nuclear Medicine, published on May 3, 2019 as doi:10.2967/jnumed.119.226597by on January 25, 2020. For personal use only. jnm.snmjournals.org Downloaded from

http://jnm.snmjournals.org/

2

ABSTRACT

Positron Emission Tomography (PET) imaging with 18F-FDG followed by mathematical modelling

of the pulmonary uptake rate (Ki) is the gold standard for assessment of pulmonary inflammation

in experimental studies of acute respiratory distress syndrome (ARDS). However, dynamic PET

requires long imaging and only allows the assessment of one cranio-caudal field of view

(~15 cm). We investigated whether static 18F-FDG-PET/CT and analysis of standard uptake

values (SUV) or standard uptake ratios (SURstat, uptake time corrected ratio of 18F-FDG-

concentration in lung tissue and blood plasma) might be an alternative to dynamic 18F-FDG-

PET/CT and Patlak analysis for quantification of pulmonary inflammation in experimental ARDS.

METHODS: ARDS was induced by saline lung lavage followed by injurious mechanical

ventilation in fourteen anesthetized pigs (29.5-40.0 kg). PET/CT imaging sequences were

acquired before and after 24 h of mechanical ventilation. Ki and the apparent volume of

distribution (Vdist) were calculated from dynamic 18F-FDG-PET/CT scans using the Patlak

analysis. Static 18F-FDG-PET/CT scans were obtained immediately after dynamic PET/CT and

used for calculations of SUV and SURstat. Mean Ki values of the whole imaged field of view and

of five ventro-dorsal lung regions were compared with corresponding SUV and SURstat values,

respectively, by means of linear regression and concordance analysis. The variability of the 18F-

FDG concentration in blood plasma (arterial input function) was analyzed.

RESULTS: Both for the whole imaged field of view and ventro-dorsal subregions, Ki were

linearly correlated with SURstat (r2 ≥ 0.84), while Ki-SUV correlations were worse (r2 ≤ 0.75). The

arterial input function exhibited an essentially invariant shape across all animals and time points

and can be described by an inverse power law. Compared to Ki, SURstat and SUV tracked the

same direction of change in regional lung inflammation in 98.6 % and 84.3 % of measurements,

respectively.

CONCLUSION: The Ki-SURstat correlation was considerably stronger than the Ki-SUV

correlation. The good Ki-SURstat correlation suggests that static 18F-FDG-PET/CT and SURstat

by on January 25, 2020. For personal use only. jnm.snmjournals.org Downloaded from


3

analysis provides an alternative to dynamic 18F-FDG-PET/CT and Patlak analysis, allowing the

assessment of inflammation of whole lungs, repeated measurements within the period of 18F-

FDG decay, and faster data acquisition.

KEY WORDS: pulmonary inflammation, positron emission tomography, 18F-FDG, tumor-to-blood

standard uptake ratio, standard uptake value



4

INTRODUCTION

The acute respiratory distress syndrome (ARDS) is an inflammatory condition of the lung

and associated with high morbidity and mortality (1). The non-invasive and in vivo measurement

of the degree and distribution of pulmonary inflammation can improve the understanding of this

syndrome and the impact of mechanical ventilation. Positron emission tomography / computed

tomography (PET/CT) imaging of the uptake rate of 18F-FDG is a valuable method to determine

the pulmonary inflammatory response in ARDS. 18F-FDG-PET/CT measurements are based on

the fact that pulmonary inflammation is associated with regionally increased accumulation of

inflammatory cells, especially neutrophils, which have higher glucose metabolism compared to

other pulmonary cells (2,3). The more pronounced regional uptake of 18F-FDG and the

associated higher radioactivity originating from a local inflamed region can therefore be used to

assess the degree and the distribution of lung inflammation in ARDS.

Dynamic 18F-FDG-PET/CT acquires time-activity data over a long period after 18F-FDG

injection (typically over 60 to 75 min). Upon mathematical modeling, they allow the calculation of

dynamic indices describing the uptake rate of 18F-FDG (Ki). Such models take the transportation

rates between blood and tissue compartments into account. However, dynamic PET requires

long image acquisition and only allows the assessment of one cranio-caudal field of view (FoV),

which usually captures approximately 15 cm and thus not the whole lung. The captured lung

region has to be defined beforehand when degree and distribution of pulmonary inflammation is

not yet known.

Static PET scanning of the decay rate of 18F-FDG allows fast image acquisition and can

cover an unlimited FoV, enabling the acquisition of the whole lung. The standard uptake value

(SUV) is a simple and widely used parameter for quantification of static PET scans, which

represents the mean activity concentration within a region of interest normalized to the injected

dose and body weight. However, SUV values strongly depend on the 18F-FDG uptake of other

organs and tissue, affecting the amount of 18F-FDG in blood plasma available for the uptake by



5

lung tissue. This is of particular importance in the lung due to its much lower 18F-FDG uptake

compared to other organs such as kidney, heart or brain (4) and has potentially caused weak

correlation between Ki and SUV in dogs suffering from lung injury (2), patients with liver

metastases (5) and patients with non-small cell lung cancer (6). Van den Hof et al. introduced

the standard uptake ratio, which is defined as tissue SUV normalized to the 18F-FDG

concentration in blood plasma available for influx into the tissue (7). Therefore, the SUR value

takes the 18F-FDG uptake of other bodily tissue and organs into account. In comparison to the

dynamic index Ki reflecting the variation of the 18F-FDG uptake rate over time, static indices such

as SUR reveal the amount of 18F-FDG within a region of interest at the time point of a static

PET/CT scan.

In this study we investigated whether SUR or SUR values derived from static PET

scanning can be used as alternative to dynamic PET for the quantification of regional lung

inflammation in experimental ARDS.



6

MATERIALS AND METHODS

Experimental Protocol

The Institutional Animal Care and Welfare Committee and the Government of the State of

Saxony, Germany, approved all animal procedures in accordance to federal law (AZ 24-

9168.11-1/2013-53). The time course of interventions is shown in Figure 1. Briefly, after

premedication (1 mg/kg midazolam, 10 mg/kg ketamine, 0.05 mg/kg atropine), 14 juvenile pigs

(29.5 - 40.0 kg) were intravenously anesthetized (5 - 15 mg/kg ketamine, 0.3 – 1 mg/kg

midazolam, both as bolus), paralyzed (3 mg/kg atracurium), oro-tracheally intubated and

mechanically ventilated (Evita XL, Dräger Medical AG, Lübeck, Germany) in supine position.

Lungs were ventilated in volume controlled mode using the following settings: fraction of inspired

oxygen: 1.0; tidal volume (VT): 6 mL/kg; positive end-expiratory pressure: 10 cmH20; inspiratory

to expiratory ratio: 1:1; constant airway flow: 35 L/min, and respiratory rate adjusted to achieve

an arterial partial pressure of carbon dioxide between 35 and 45 mmHg. During preparation, a

crystalloid solution (E153, Serumwerk Bernburg AG, Bernburg, Germany) was infused

intravenously at a rate of 10 mL/kg/h via a peripheral vein. An 8.5 French sheath was inserted in

the right internal carotid artery and a 7.5 French pulmonary artery catheter was advanced

through another sheath placed in the right external jugular vein. The lungs were recruited with

continuous positive airway pressure of 30 cmH2O for 30 s followed by 15 min of stabilization.

Experimental ARDS was induced using a double hit model consisting of surfactant depletion

(eight repetitive isotonic saline lung lavages alternating in prone and supine position) followed by

injurious mechanical ventilation with high VT (20 mL/kg) until Horovitz index < 100 mmHg for at

least 30 min. After acquisition of baseline PET/CT imaging data, animals were randomly

assigned to mechanical ventilation with either variable volume controlled ventilation with a mean

VT of 6 mL/kg and coefficient of variation in VT of 30 % (n = 7) or volume controlled ventilation

with non-variable VT (n = 7). Further mechanical ventilation settings were: fraction of inspired

oxygen titrated according to the low positive end-expiratory pressure table of the ARDS network,



7

inspiration-expiration-ratio: 1:1; respiratory rate adjusted to arterial pH > 7.30, maximal plateau

pressure: 30 cmH2O and 45 cmH2O for variable and non-variable ventilation, respectively, mean

plateau pressure: 30 cmH2O in the variable ventilation mode.

After randomization, the crystalloid solution infusion rate was changed to 4 mL/kg/h to

maintain intravascular volume. Colloid solution (6 % hydroxyenthyl starch, Fresenius Kabi

Deutschland GmbH, Bad Homburg, Germany) was administered as necessary to keep the

hemoglobin concentration in the blood approximately constant. After 24 h of mechanical

ventilation in variable or non-variable ventilation mode, PET/CT imaging was repeated.

Respiratory mechanics, gas exchange and hemodynamics were assessed before and after

induction of ARDS (BL 1, injury), before start of mechanical ventilation (BL 2) and in 6 h intervals

thereafter (Time 1 to 4). At the end of the experiments, animals were killed by intravenous

injections of thiopental (2 g), followed by potassium chloride (1 M, 50 mL).

Lung Imaging Protocol and Image Processing

After induction of lung injury and before start of 24 h of mechanical ventilation, as well as

after 24 h of mechanical ventilation, imaging data were acquired according to the imaging

protocol illustrated in Figure 1. Briefly, low dose helical CT scans of the thorax were obtained

during mechanical ventilation and used for attenuation correction of the following PET images

(attenuation correction CT scans - ACCT) (Biograph16 Hirez PET/CT, Siemens, Knoxville, TN,

USA). 18F-FDG (198.6 ± 42.3 MBq) was injected intravenously over 60 s. Starting at the

beginning of 18F-FDG infusion, sequential PET frames (6 × 30″, 7 × 60″, 15 × 120″, 1 × 300″, 3 ×

600″) were acquired over 75 min. The 15 cm cranio-caudal FoV of the dynamic PET scans was

set above the diaphragmatic dome to reduce artifacts due to motion of the diaphragm.

Pulmonary arterial blood was sampled during the time course of the dynamic PET scans (12 ×

15″, 4 × 30″, 5 × 60″, 11 × 300″, and 75′). The concentration of 18F-FDG in 1 mL blood plasma

was measured in a gamma counter cross-calibrated with the PET scanner. Immediately after



8

dynamic PET and 77-81 min after 18F-FDG injection, static 18F-FDG-PET/CT scans were

obtained in three bed positions assessing the whole lung.

ACCT scans were reconstructed with 2.0 mm slice thickness, yielding matrices with 512

× 512 pixels (1.37 × 1.37 mm2). Static and dynamic PET scans were reconstructed with 2.0 mm

slice thickness, yielding matrices with 168 × 168 pixel (2.03 × 2.03 mm2). The reconstruction

was carried out iteratively (ordered subset expectation maximization, six iterations, four subsets,

post-filtering Gauss 5 mm) with correction for scatter and attenuation.

Analysis of Blood Plasma Samples

For each animal and imaging sequence, the activity measurements of 18F-FDG in blood

plasma were interpolated to the mean frame time points of the dynamic PET scans, giving a

subject-specific arterial input function (Cp(t)). For each animal and time point, an inverse power

law (Eq. 1) was fitted to the input function using the data at t ≥ 10 min after 18F-FDG injection:

C t A ∗ t (Eq. 1)

The resulting input function was extrapolated to the time p.i. of the respective static PET

scan. The extrapolated Cp values were used to compute SUR of the static PET scans (see

below), for which no blood samples were available.

To validate wether equation 1 can be used to adequately describe the input function, the

time course of Cp was normalized to its mean value over the period of the dynamic PET scan to

account for differing amount of injected 18F-FDG and for differing body weight (blood volume).

The time-averaged Cp was compared between imaging sequences and animals by graphical

illustration and by fitting equation 1 to the time-averaged Cp data at t > 3 min and t ≥ 10 min.



9

Image Analysis

ACCT scans and static PET scans were coregistered to the dynamic PET scans.

Segmentation was performed on coregistered ACCT scans to define regions of interest (ROI),

from which major airways and vessels were excluded. The acquired 15 cm cranio-caudal lung

fields of view of the dynamic PET scans were divided in 5 iso-gravimetric subregions reaching

from ventral to dorsal. The ROIs were applied to dynamic and static PET scans and were used

to compute the corresponding concentration of 18F-FDG in lung tissue (CPET).

18F-FDG uptake rates (Ki) and the apparent distribution volume of 18F-FDG in blood

plasma as a fraction of tissue volume (Vdist) were derived from the Patlak graphical analysis of

the dynamic PET frames acquired 10 - 75 min after 18F-FDG injection using the following

equation:

K ∗ θ t V . with θ t∗

(Eq. 2)

where Θ(t) is the so-called Patlak time and is the integration variable. Ki and Vdist were

averaged for each ROI as well as for the whole FoV.

SUV was calculated from static PET scans as:

SUV T /

(Eq. 3)

SUR was computed for the data of the dynamic PET scans acquired 40 min to 75 min p.i.

(SURdyn) and static PET scans (SURstat) as the uptake time corrected ratio of tissue

concentration and blood concentration as described in (8):

SUR T ∗ (Eq. 4)

where T is the actual scan time p.i. and T0 is the chosen standard scan time to which SUR

values are normalized. By definition the uptake time of SURdyn is the same for all measurements

and, therefore, T0 = T was chosen (i.e. no scan time correction). The mean frame time point of

the static PET scans ranged from 80.0 min to 83.4 min. The mean value of the scan times was

chosen as reference time T0 = 81.0 min p.i.. Mean SUV and SUR values were calculated for the



10

same ROIs as used for the Patlak analysis, thus covering only the 15 cm cranio-caudal FoV, and

for the whole FoV.

For PET measurements before and after 24 h of mechanical ventilation, respectively,

correlations of Ki vs. SUV, Ki vs. SURstat and Ki vs. SURdyn were investigated by means of linear

regression and comparison of the coefficients of determination (r2) for regional values and the

whole FoV.

The ability of static PET scanning and SUV and SURstat analysis, respectively, to track

the direction of change in regional lung inflammation induced by 24 h of mechanical ventilation

and determined by dynamic PET scanning and Patlak Ki analysis was assessed by concordance

analysis and calculation of Cohen’s Kappa.

Statistics

A sample size calculation was not performed. Data are presented as mean and standard

deviation (SD) if not stated otherwise. Wilcoxon tests were used for comparisons between

measurement time points. For the analysis of hemodynamics, gas exchange and lung

mechanics, differences between and within groups (Group Effect, Time*Group Effect) were

tested with general linear model statistics. Differences between groups at time point Injury and

Time 4, respectively, were tested with Mann-Whitney-U tests. Significance was accepted at

p<0.05. Statistical analysis was performed with SPSS (version 23, SPSS, Chicago, IL).



11

RESULTS

Hemodynamics, gas exchange and lung mechanics data are shown in Supplemental

Table 1 of the Appendix. There was no group effect or time-Group-effect for any of the variables.

Variables regarding hemodynamics and gas exchange were comparable between groups at the

imaging time points Injury and Time4, respectively (Supplemental Table 1).

Maps of pulmonary inflammation of one representative animal obtained by static and

dynamic 18F-FDG-PET/CT scanning before and after 24 hare shown in Figure 2. The 15 cm

cranio-caudal FoV assessed by dynamic PET scanning covered 76.43 ± 9.61 % of the volume of

the whole lung and 65.58 ± 5.96 % of the cranio-caudal lung expansion (see Fig. 2 for

illustration). In comparison, static PET scans covered the whole lung. Static PET scans were

acquired 81.0 ± 0.81 min after injection of 18F-FDG, and 11.0 ± 0.81 min after the mean time

point of the last dynamic PET frame. Dynamic PET scans were acquired over 75 min, while

acquisition of static PET scans lasted 9 min.

The period of 24 h of mechanical ventilation was associated with a 127.2 ± 79.4 %, 63.2

± 68.0 %, and 99.2 ± 76.5 % increase in regional Ki, SUV and SURstat, respectively (Figs. 3 and

4).

Before the 24 h ventilation period, the linear correlation between Ki and SUV was weak,

both for the whole FoV (r2 = 0.08) and 5 ventro-dorsal ROIs (Fig. 3, r2 = 0.12). Ki-SUV correlation

was stronger after 24 h of mechanical ventilation (whole FoV: r2 = 0.73; 5 ventro-dorsal ROIs: r2

= 0.75, Fig. 3). Linear correlation between Ki and SUV was worse than Ki-SURstat correlation

(compare Fig. 3 and 4). Before and after 24 h of mechanical ventilation, Ki and SURstat were

correlated, both for the whole FoV (Supplemental Fig. 2, r2 = 0.94 and 0.97, respectively) and for

5 ventro-dorsal ROIs (Fig. 4, r2 = 0.84 and 0.97, respectively). The correlation between Ki and

SURstat was higher after 24 h of mechanical ventilation, when inflammation increased

substantially (Fig. 4).



12

Static PET scanning and SURstat and SUV analysis, respectively, was able to predict the

direction of change in regional lung inflammation, as determined by dynamic PET and Ki

analysis, in 98.6 % and 84.3 % of measurements (Fig. 5). The smallest change in Ki that was

still detected as increase in SURstat was 0.0006 mL/mL/min (Fig. 5). The agreement between

changes in Ki and SURstat induced by 24 h of mechanical ventilation was good (Cohen’s Kappa:

0.66), while there was no agreement between changes in Ki and SUV (Cohen’s Kappa: -0.027).

Figure 6 A shows the time course of the time-averaged Cp of all animals and both

imaging sequences. The very small standard deviation of CP at mean frame time points beyond

t > 3 min after 18F-FDG injection and the excellent agreement of the interpolation function with

time-averaged Cp values (r2 = 0.99) illustrates the very small inter- and intra-subject variability of

the mean normalized activity of 18F-FDG in blood plasma, despite much higher inflammatory

values after a 24 h period of mechanical ventilation. The low variability is especially true at late

time points, which are the relevant ones for the correlation Ki vs. SURstat. As a consequence, Θ

featured a relatively small inter-study variability and was linearly correlated with time for each

animal and measurement time point (Fig. 6 B). The low relative standard deviations of Θ at later

time points (e.g. at t = 70 min: 8.7 %) indicates the small inter- and intra-subject variability of the

Patlak timeVdist was lowest in ventral regions and increased along the gravitational gradient at

both imaging time points (Fig. 7). The 24 h period of mechanical ventilation was associated with

a 23.2 ± 25.0 % increase in mean regional Vdist.

Ki and SURdyn were strongly correlated (before and after 24 h of mechanical ventilation:

r2 = 0.78 and r2 = 0.97, Supplemental Fig. 3). While for higher inflammatory values obtained after

24 h of mechanical ventilation the Ki-SURdyn correlation was similar to the Ki-SURstat correlation,

the Ki-SURdyn correlation was worse than the Ki-SURstat correlation before 24 h of mechanical

ventilation (lower r2).



13

DISCUSSION

The main results of this study were that, in an experimental model of ARDS: 1) the

agreement between Ki and SURstat was stronger than the agreement between Ki and SUV at

different Ki levels; 2) Ki and SURstat were strongly correlated at different levels of lung

inflammation; 3) the arterial input function exhibited an essentially invariant shape across all

animals and time points; 4) compared to Ki, SURstat and SUV tracked the same direction of

change in regional lung inflammation in 98.6 % and 84.3 % of measurements, respectively.

The SUV is a widely used parameter for quantification of the uptake of 18F-FDG.

However, the SUV is dependent on factors such as body mass, injected dose of 18F-FDG and

other confounding factors (9,10). We found a weak correlation between Ki and SUV, despite

similar weight of the pigs (35.3 ± 3.6 kg), small variation of the injected dose of 18F-FDG (198.6 ±

42.3 MBq) and acquisition of the static PET/CT scans at a similar time after 18F-FDG injection

(78.4 ± 0.9 min). The Ki-SUV-correlation was especially weak at low levels of lung inflammation

where variations in body mass and injected 18F-FDG dose have a comparatively high impact on

SUV.

The linear correlation between Ki and SUR was much stronger than the Ki-SUV

correlation at different levels of lung inflammation. Similarly, Chan et al. showed in an

experimental ARDS study in dogs that Ki, determined from compartment modeling of dynamic

18F-FDG-data, strongly correlated with tissue-to-plasma activity ratios (calculated from the last

frame of a dynamic 18F-FDG scan), while Ki-SUV correlation was weak (2).

A prerequisite for the good correlation between Ki and SUR is the shape invariance of

the arterial input function of 18F-FDG. This shape invariance across different subjects and time

points has been shown in patients with liver metastasis (7) and colon cancer metastatic to the

liver (11). Van den Hoff et al. showed that, in case the arterial input function can additionally be

described by an inverse power law, the shape invariance translates into: same exponent b but

different scale factor A in Eq. 1 (8). As a direct consequence, the Patlak time Θ(t) does not



14

depend on the individual arterial input function but is rather proportional to real time t. Therefore,

the Patlak time is comparable between subjects at any time point after an initial period of about

3 min (8). Van den Hoff showed that these theoretical consideration are approximately fulfilled in

measurements obtained from patients with liver metastases (8). The investigation of the arterial

input function performed in this study demonstrates that both the shape invariance of the 18F-

FDG input function across different animals and imaging time points and the description of the

arterial 18F-FDG-time-activity curve by an inverse power law is also valid in pigs with ARDS.

However, the exponent of the power law seems to be notably larger in pigs than in humans (0.52

compared to 0.31). In general, a shape invariant arterial input function (as indicated by the

constant exponent) is very likely a result of a constant systemic glucose metabolism (12).

Therefore, the differing shape of the arterial input function between pigs and humans reflect

different systemic metabolism. However, further investigations are necessary to confirm this

hypothesis.

The slightly worse correlation of Patlak Ki and SUR directly after induction of lung injury

compared to imaging data obtained after 24 h mechanical ventilation might be caused by the

lower inflammatory values and the resulting higher contribution of the variability of Vdist. The

rather high intra-subject variability of Vdist might be explained by a substantial increase in lung

perfusion from ventral to dorsal regions in supine positioned animals (13-15), potentially

increasing the fractional blood volume.

The 24 h period of mechanical ventilation and the associated ventilator induced lung

injury was associated with a 127.2 ± 79.4 % increase in regional Ki, while regional Vdist did

increase by 23.2 ± 25.0 %. Therefore, the contribution of the variability of regional Vdist was much

lower after the 24 h period of mechanical ventilation. This might, at least partly, explain the better

correlation between Ki and SUR at higher levels of pulmonary inflammation obtained after the

24 h period of mechanical ventilation.



15

The comparison of Patlak Ki and SUR was performed using two different static PET

images for SUR computation: the static PET scan measured after dynamic data acquisition

(covering the whole lung, giving SURstat) and one image generated from the last frames of the

dynamic PET scan (giving SURdyn). The latter was analyzed, since the analysis of a static PET

scan has two artificial sources of errors. First, there were no blood samples taken at the time

point of the static PET scan. Instead, the measured arterial input function was extrapolated to

this time point. Second, due to the different FoV, the image data had to be coregistrated to the

corresponding dynamic PET scan. Both aspects introduce an additional small error, which would

not be present in a study designed accordingly. Therefore, the accuracy, which can be expected

when Ki is replaced by SUR, is given by the results for the generated static image (SURdyn) not

by the measured static image (SURstat).

A limitation of the current study is the above mentioned lack of blood samples at the time

point of the static whole lung PET scan. A second limitation is the low number of investigated

subject. Further investigations with larger sample size have to be performed before these results

can be transferred to patient investigations.

CONCLUSION

In this model of experimental ARDS, the SUR analysis provided an alternative to

dynamic PET scanning and Patlak modeling of the uptake rate of 18F-FDG, allowing assessment

of inflammation of whole lungs, repeated measurements within the period of the 18F-FDG decay,

and faster data acquisition.



16

DISCLOSURE

No potential conflicts of interest relevant to this article exist.

ACKNOWLEDGMENTS

We thank Susanne Henninger Abreu, Gabriele Kotzerke, Kathrin Rosenow, and Michael

Andreeff for their valuable support during the experiments.



17

KEY POINTS

QUESTION: Can static 18F-FDG-PET/CT and analysis of standard uptake values (SUV) or

standard uptake ratios (SURstat) be used as alternative to Patlak- Ki values derived from dynamic

18F-FDG-PET/CT for quantification of regional lung inflammation in experimental acute

respiratory distress syndrome in pigs?

PERTINENT FINDINGS: An experimental study in fourteen anesthetized pigs suffering from

acute respiratory distress syndrome revealed a weak and a strong linear correlation between

SURstat and Ki at two separate imaging time points. The good SURstat-Ki–correlation can be

explained by the shape invariance of the arterial input function of 18F-FDG across all animals

and time points.

IMPLICATIONS FOR PATIENT CARE: The findings suggest that SURstat derived from static 18F-

FDG-PET/CT provides an alternative to dynamic 18F-FDG-PET/CT and Patlak-Ki analysis,

allowing the assessment of inflammation of whole lungs, repeated measurements within the

period of 18F-FDG decay, and faster data acquisition.



18

REFERENCES

1. Bellani G, Laffey JG, Pham T, et al. Epidemiology, Patterns of Care, and Mortality for Patients With Acute Respiratory Distress Syndrome in Intensive Care Units in 50 Countries. JAMA. 2016;315:788-800.

2. Chen DL, Mintun MA, Schuster DP. Comparison of methods to quantitate 18F-FDG uptake with PET during experimental acute lung injury. J Nucl Med. 2004;45:1583-1590.

3. Musch G, Venegas JG, Bellani G, et al. Regional gas exchange and cellular metabolic activity in ventilator-induced lung injury. Anesthesiology. 2007;106:723-735.

4. Paquet N, Albert A, Foidart J, Hustinx R. Within-patient variability of (18)F-FDG: standardized uptake values in normal tissues. J Nucl Med. 2004;45:784-788.

5. van den Hoff J, Oehme L, Schramm G, et al. The PET-derived tumor-to-blood standard uptake ratio (SUR) is superior to tumor SUV as a surrogate parameter of the metabolic rate of FDG. EJNMMI Res. 2013;3:77.

6. Hofheinz F, Hoff J, Steffen IG, et al. Comparative evaluation of SUV, tumor-to-blood standard uptake ratio (SUR), and dual time point measurements for assessment of the metabolic uptake rate in FDG PET. EJNMMI Res. 2016;6:53.

7. van den Hoff J, Oehme L, Schramm G, et al. The PET-derived tumor-to-blood standard uptake ratio (SUR) is superior to tumor SUV as a surrogate parameter of the metabolic rate of FDG. EJNMMI research. 2013;3:77.

8. van den Hoff J, Lougovski A, Schramm G, et al. Correction of scan time dependence of standard uptake values in oncological PET. EJNMMI research. 2014;4:18.

9. Carlier T, Bailly C. State-Of-The-Art and Recent Advances in Quantification for Therapeutic Follow-Up in Oncology Using PET. Front Med (Lausanne). 2015;2:18.

10. Chen DL, Cheriyan J, Chilvers ER, et al. Quantification of Lung PET Images: Challenges and Opportunities. J Nucl Med. 2017;58:201-207.

11. Graham MM, Peterson LM, Hayward RM. Comparison of simplified quantitative analyses of FDG uptake. Nuclear Medicine and Biology. 2000;27:647-655.

12. Paquet N, Albert A, Foidart J, Hustinx R. Within-patient variability of (18)F-FDG: standardized uptake values in normal tissues. Journal of Nuclear Medicine: Official Publication, Society of Nuclear Medicine. 2004;45:784-788.



19

13. Braune A, Scharffenberg M, Naumann A, Bluth T, de Abreu MG, Kotzerke J. Comparison of 68Ga- and fluorescence-labeled microspheres for measurement of relative pulmonary perfusion in anesthetized pigs. Nuklearmedizin Nuclear Medicine. 2018;57:100-107.

14. Glenny RW, Lamm WJ, Albert RK, Robertson HT. Gravity is a minor determinant of pulmonary blood flow distribution. Journal of Applied Physiology (Bethesda, Md: 1985). 1991;71:620-629.

15. Walther SM, Domino KB, Glenny RW, Hlastala MP. Pulmonary blood flow distribution in sheep: effects of anesthesia, mechanical ventilation, and change in posture. Anesthesiology. 1997;87:335-342.



20

FIGURES WITH LEGENDS

Premedication, surgical preparation

Euthanasia, organ harvest

Induction of ARDS

8 h

Randomization

Start of therapy

2 h 22 h 2 h

PET/CT

ACCT Dynamic PET Static PET1 min 75 min 9 min

18F-FDG injection (190 ± 23 MBq)

PET/CT

ACCT Dynamic PET Static PET1 min 75 min 9 min

18F-FDG injection (206 ± 23 MBq)

Figure 1: Time course of interventions.

ACCT, CT-based attenuation correction; PET, positron emission tomography; CT, computed

tomography



21

FoV of dynamic PET scan

Missing part in dynamic PET scan

0 2.7 0 2.3 0 2.3

Dynamic PET: Ki (102 mL/min/mL) Static PET: SUV Static PET: SUR3D

lu

ng

co

nto

ur

Bef

ore

24 h

ven

tila

tio

nA

fter

24

h v

enti

alti

on

Figure 2: 3D lung contour (upper line) and 2D transversal slices of one representative animal

obtained before (middle) and after 24 h of mechanical ventilation (lower line). 18F-FDG uptake

rate (Ki, left column) was derived by dynamic PET/CT, while SUV (central column) and SURstat

(right column) were derived from static PET/CT. The location of the 2D transversal slices is

shown in the 3D contours. Caudal lung regions that are not assessed by the dynamic PET scan

due to its limited field of view are highlighted in the 3D contours.



22

Figure 3: Linear correlation between Ki and SUV obtained from PET/CT imaging data acquired

before (left) and after 24 h of mechanical ventilation (right) and divided in 5 iso-gravimetric

ventro(gravitational non-dependent)-dorsal (gravitational dependent) regions. Red lines

represent linear regression lines. Note the differing axis scales, for which slope and intercept are

specified.



23

Figure 4: Linear correlation between Ki and SURstat obtained from PET/CT imaging data

acquired before (left) and after 24 h of mechanical ventilation (right) and divided in 5 iso-

gravimetric ventro(gravitational non-dependent)-dorsal (gravitational dependent) regions. Red

lines represent linear regression lines, for which slope and intercept are specified. Note the

differing axis scales.



24

Figure 5: Linear correlation between changes in regional pulmonary uptake rates of 18F-FDG

(∆Ki) and SURstat (∆SURstat, left), and linear correlation between changes in regional Ki (∆Ki) and

SUV (∆SUV, right), induced by 24 h of mechanical ventilation in 14 animals and divided in 5 iso-

gravimetric ventro(gravitational non-dependent)-dorsal (gravitational dependent) regions. Red

lines represent linear regression lines, for which slope and intercept are specified.



25

Figure 6: Arterial input function of 18F-FDG (A) and Patlak time Θ (B) at mean frame time points

of the dynamic PET scans. Courses are shown for 14 animals before (blue) and after 24 h of

mechanical ventilation (black). Red lines and error bars represent averages and standard

deviations of each frame. In A, interpolation was performed using the following power function:

A ∗ t .



26

Figure 7: Regional volume of distribution (Vdist) of 14 animals before (left) and after (right) the

24 h period of mechanical ventilation.



Supplemental Digital Content

Comparison of static 18F-FDG-PET/CT (SUV, SUR) and dynamic 18F-FDG-PET/CT (Ki) for quantification of pulmonary inflammation in acute lung injury

Anja Braune1,2, Frank Hofheinz3, Thomas Bluth1, Thomas Kiss1, Jakob Wittenstein1, Martin Scharffenberg1, Jörg Kotzerke2, Marcelo Gama de Abreu1

1Pulmonary Engineering Group, Department of Anesthesiology and Intensive Care Medicine, University Hospital Carl Gustav Carus at the Technische Universität Dresden, Dresden, Germany

2Department of Nuclear Medicine, University Hospital Carl Gustav Carus at the Technische Universität Dresden, Dresden, Germany

3Helmholtz-Zentrum Dresden-Rossendorf, PET Center, Institute of Radiopharmaceutical Cancer Research, Dresden, Germany

Corresponding Author / Reprint request: Anja Braune Department of Nuclear Medicine University Hospital Carl Gustav Carus at the Technische Universität Dresden Fetscherstr. 74 D-01307 Dresden Phone: +49 (0)351-458 12128 Fax: +49 (0)351-458 5310 E-mail: [email protected]

Short running title: 18F-FDG-PET/CT imaging in acute lung injury



Figure SDC1: Patlak plot of imaging data obtained from one representative animal before and

after 24 h of mechanical ventilation and divided in 5 iso-gravimetric ventro-dorsal lung regions.

Ventral (gravitational non-dependent), mid-ventral, middle, mid-dorsal, and dorsal (gravitational

dependent) subregions are shown in orange, red, green, blue, and black, respectively. The solid

and dashed lines represent the linear regression lines of the data obtained before and after 24 h

of mechanical ventilation, respectively. Linear regression was performed for data obtained ≥

10 min after injection of 18F-FDG (frame 14 to 32). The slope of the linear regression line

represents the 18F-FDG net uptake rate (Ki) while the ordinate-intercept of the prolonged linear

regression line corresponds to the apparent distribution volume (Vdist).



Figure SDC2: Linear correlation between Ki and SURstat obtained from PET/CT imaging data of 14 animals acquired before (black markers) and after 24 h mechanical ventilation (gray markers) and averaged over the whole field of view (15 cm craniao-caudal field of view). SURstat were obtained from the static PET/CT scans and analysis of the same 15 cm cranio-caudal field of view as used for the Patlak analysis. The red solid and dashed lines represent the linear regression lines of the data obtained before and after 24 h of mechanical ventilation, respectively.



Figure SDC3: Linear correlation between Ki and SURdyn obtained from PET/CT imaging data of 14 animals acquired before (left) and after 24 h mechanical ventilation (right) and divided in 5 iso-gravimetric ventro-dorsal regions. Pulmonary uptake rates of 18F-FDG (Ki) were derived by dynamic PET scanning followed by Patlak analysis. SURdyn data were obtained from the dynamic PET scan and analysis of frame 29 – 32 acquired 40 min to 75 min post injection of 18F-FDG and analysis of the same 15 cm cranio-caudal field of view as used for the Patlak analysis. Ventral (gravitational non-dependent), mid-ventral, middle, mid-dorsal, and dorsal (gravitational dependent) subregions are shown in orange, red, green, blue, and black, respectively. The red lines represent the linear regression lines. Note the differing axis scales.



Table 1: Hemodynamics, gas exchange and lung mechanics data.

Group BL 1 Injury BL 2 Time 1 Time 2 Time 3 Time 4

Group Effect Time*Group Effect

Hemodynamics

CO (l/min) nVCV 3.9 ± 0.8 6.7 ± 1.9 4.4 ± 0.3 6.3 ± 2.3 6.8 ± 2.0 6.9 ± 1.9 6.74 ± 1.70

n.s. n.s.

VCV 3.4 ± 0.6 5.6 ± 2.3 4.7 ± 1.5 5.5 ± 2.1 5.3 ± 0.5 6.4 ± 1.6 5.90 ± 0.76

n.s. n.s.

HF (min-1)

nVCV 105 ± 18 113 ± 27 100 ± 13 107 ± 20 109 ± 12 113 ± 8 107 ± 12 n.s. n.s. VCV 97 ± 13 107 ± 34 107 ± 22 106 ± 14 100 ± 14 111 ± 9 108 ± 12

n.s. n.s.

MAP (mmHg)

nVCV 64.7 ± 6.4 81.0 ± 8.7 77.6 ± 9.7 70.3 ± 10.3 70.4 ± 7.4 71.6 ± 9.5 71.1 ± 6.9 n.s. n.s. VCV 73.9 ± 14.5 80.1 ± 9.5 77.9 ± 13.3 72.3 ± 16.4 67.4 ± 9.4 68.9 ± 8.2 71.4 ± 8.6

n.s. n.s.

MPAP (mmHg)

nVCV 18.1 ± 3.6 31.7 ± 4.2 27.1 ± 3.6 26.6 ± 2.1 26.4 ± 4.9 26.4 ± 3.5 24.7 ± 4.2 n.s. n.s. VCV 18.9 ± 2.7 31.4 ± 4.5 31.7 ± 7.0 28.3 ± 5.2 28.7 ± 4.4 28.0 ± 2.0 27.3 ± 3.2

n.s. n.s.

Hct nVCV 0.27 ± 0.02 0.29 ± 0.05 0.29 ± 0.02 0.26 ± 0.03 0.25 ± 0.03 0.25 ± 0.04 0.25 ± 0.04 n.s. n.s. VCV 0.26 ± 0.03 0.27 ± 0.04 0.28 ± 0.04 0.26 ± 0.04 0.24 ± 0.03 0.24 ± 0.02 0.24 ± 0.03

n.s. n.s. Gas Exchange

PaO2 nVCV 600.6 ± 61.6 69.14 ± 16.22 86.86 ± 13.06 83.57 ± 17.82 82.29 ± 16.12 83.29 ± 14.04 87.86 ± 18.28 n.s. n.s. VCV 599.7 ± 60.2 64.43 ± 14.79 80.86 ± 6.26 75.71 ± 11.09 74.00 ± 9.83 77.57 ± 10.75 74.57 ± 8.98

n.s. n.s.

PaCO2 nVCV 47.7 ± 6.4 89.17 ± 10.42 87.71 ± 18.03 83.91 ± 11.24 81.63 ± 7.64 91.54 ± 10.41 95.74 ± 15.55 n.s. n.s. VCV 50.9 ± 5.8 88.49 ± 27.48 88.99 ± 19.89 80.41 ± 15.14 80.41 ± 15.62 86.73 ± 19.34 84.20 ± 10.36

n.s. n.s.

PaO2/ FiO2

nVCV 600.6 ± 61.6 69.1 ± 16.2 202.6 ± 81.7 214.2 ± 79.4 220.0 ± 222.6 ± 74.4 228.3 ± 85.6

n.s. n.s. VCV 599.7 ± 60.2 64.4 ± 14.8 158.2 ± 42.0 167.6 ± 32.5 189.4 ± 198.4 ± 47.7 190.4 ± 41.8

n.s. n.s.

pH nVCV 7.4 ± 0.0 7.23 ± 0.06 7.26 ± 0.08 7.30 ± 0.06 7.33 ± 0.06 7.32 ± 0.06 7.30 ± 0.07 n.s. n.s. VCV 7.4 ± 0.0 7.25 ± 0.11 7.22 ± 0.08 7.30 ± 0.05 7.32 ± 0.04 7.34 ± 0.05 7.35 ± 0.05

n.s. n.s.

Tempe-rature

nVCV 37.8 ± 0.9 37.99 ± 0.68 38.63 ± 1.19 38.89 ± 0.62 38.23 ± 0.56 38.66 ± 0.34 38.81 ± 0.40 n.s. n.s. VCV 37.6 ± 0.6 37.71 ± 0.83 38.13 ± 1.02 38.39 ± 0.82 37.93 ± 0.45 38.43 ± 0.37 38.41 ± 0.34

n.s. n.s.



Group BL 1 Injury BL 2 Time 1 Time 2 Time 3 Time 4

Group effect Time* Group effect

Lung Mechanics VT (mL/kg) nVCV 6.4 ± 0.1 6.4 ± 0.1 6.5 ± 0.0 6.2 ± 0.5 6.2 ± 0.5 6.1 ± 0.6 6.0 ± 0.5 n.s. n.s.

VCV 6.5 ± 0.2 6.6 ± 0.2 6.8 ± 0.9 6.4 ± 0.4 6.4 ± 0.5 6.5 ± 0.4 6.6 ± 0.2 n.s. p=0.009 RR (min-1)

nVCV 33.6 ± 2.5 33.6 ± 2.5 35.1 0.1 28.3 ± 7.2 27.5 ± 8.1 25.3 ± 8.6 26.0 ± 9.3 n.s. n.s. VCV 33.6 ± 2.5 33.6 ± 2.5 35.1 0.0 32.9 ± 2.7 29.3 ± 6.1 27.9 ± 5.7 26.4 ± 5.6

n.s. n.s. MV (l/min)

nVCV 7.9 ± 0.6 7.9 ± 0.6 8.3 0.7 6.2 ± 1.4 6.0 ± 1.6 5.4 ± 1.5 5.5 ± 1.5 n.s. n.s. VCV 7.6 ± 0.7 7.6 ± 0.6 8.2 0.9 7.3 ± 0.6 6.4 ± 0.8 6.1 ± 1.4 5.8 ± 1.5

n.s. n.s. RRS (cmH2O l-1/s)

nVCV 7.3 ± 0.6 10.7 ± 2.0 7.4 0.3 8.2 ± 0.6 9.1 ± 1.1 9.8 ± 2.1 10.7 ± 4.2 n.s. n.s. VCV 7.6 ± 1.1 10.0 ± 1.6 8.5 ± 1.9 7.9 ± 0.6 8.4 ± 2.1 9.4 ± 2.2 9.6 ± 2.4

n.s. n.s. ERS (cmH2O l-1)

nVCV 24.1 ± 2.7 81.2 ± 7.0 69.2 12.4 74.6 ± 21.6 74.1 ± 23.7 71.3 ± 22.6 70.1 ± 23.2 n.s. n.s. VCV 23.6 ± 4.3 67.7 9.8 69.1 ± 8.8 79.3 ± 13.5 78.0 ± 14.4 74.6 ± 11.1 71.0 ± 10.2

p=0.018 p=0.805 Pmax (cmH2O)

nVCV 21.0 ± 0.7 34.4 ± 2.4 27.6 ± 4.2 27.1 ± 5.2 27.6 ± 6 26.8 ± 5 27.2 ± 4.7 n.s. n.s. VCV 20.9 ± 0.7 31.1 ± 2.6 30.5 ± 3.6 29.1 ± 3.9 28.1 ± 2.3 28 ± 2.5 26.6 ± 2.4

p=0.048 n.s. Pmean (cmH2O)

nVCV 14.0 ± 0.2 19.2 ± 0.8 15.5 ± 3.2 14.2 ± 2.7 14.4 ± 3.5 13.8 ± 3.1 14.0 ± 3.2 n.s. n.s. VCV 14.0 ± 0.3 17.9 ± 0.9 17.6 ± 3.1 15.9 ± 2.8 15.3 ± 2.2 15.2 ± 2.1 13.9 ± 1.7

p=0.026 n.s. Pplat (cmH2O)

nVCV 17,4 0.7 30.8 2.3 25.6 4.3 24.6 5.2 24.7 7.0 23.5 6.1 23.4 6.2 n.s. n.s. VCV 17,3 0.6 27.0 2.3 27.8 4.8 27.1 4.2 25.7 3.4 25.2 3.1 23.6 3.0

p=0.018 n.s. PEEP (cmH2O)

nVCV 10.0 ± 0.0 9.8 ± 0.2 7.7 ± 2.9 6.2 ± 1.5 6.3 ± 62 5.8 ± 1.9 6.2 ± 2.0 n.s. n.s. VCV 10.0 ± 0.0 9.8 ± 0.2 9.7 ± 2.8 7.6 ± 2.0 6.9 ± 1.9 6.7 ± 1.9 5.6 1.5

n.s. n.s. Values are given as mean and standard deviation. Differences between and within groups were tested with general linear model statistics with BL 2 as covariate. Differences between groups at time point Injury and Time 4, respectively, were tested with Mann-Whitney-U tests and are specified in the respective columns. Statistical significance was accepted at p<0.05. BL1/2, Baseline 1/2; nVCV, volume controlled ventilation with variable tidal volumes; VCV, volume controlled ventilation with non-variable tidal volume; CO, cardiac output; HR, heart rate; MAP, mean arterial blood pressure; MPAP, mean pulmonary arterial blood pressure; Hct, hematocrit; PaO2, arterial partial pressure of oxygen; FiO2, fraction of inspired oxygen; PaCO2, arterial partial pressure of carbon dioxide; VT, tidal volume; RR, respiratory rate; MV, minute ventilation; RRS, resistance of the respiratoy system; Pmax, maximal airway pressure; Pmean, mean airway pressure; PEEP, positive end-expiratory pressure; n.s., no significance.



Doi: 10.2967/jnumed.119.226597Published online: May 3, 2019.J Nucl Med. Marcelo Gama de AbreuAnja Braune, Frank Hofheinz, Thomas Bluth, Thomas Kiss, Jakob Wittenstein, Martin Scharffenberg, Joerg Kotzerke and for quantification of pulmonary inflammation in acute lung injury

F-FDG-PET/CT (Ki)18F-FDG-PET/CT (SUV, SUR) and dynamic 18Comparison of static

http://jnm.snmjournals.org/content/early/2019/05/03/jnumed.119.226597This article and updated information are available at:

http://jnm.snmjournals.org/site/subscriptions/online.xhtml

Information about subscriptions to JNM can be found at:

http://jnm.snmjournals.org/site/misc/permission.xhtmlInformation about reproducing figures, tables, or other portions of this article can be found online at:

and the final, published version.proofreading, and author review. This process may lead to differences between the accepted version of the manuscript

ahead of print area, they will be prepared for print and online publication, which includes copyediting, typesetting,JNMcopyedited, nor have they appeared in a print or online issue of the journal. Once the accepted manuscripts appear in the

. They have not beenJNM ahead of print articles have been peer reviewed and accepted for publication in JNM

(Print ISSN: 0161-5505, Online ISSN: 2159-662X)1850 Samuel Morse Drive, Reston, VA 20190.SNMMI | Society of Nuclear Medicine and Molecular Imaging

is published monthly.The Journal of Nuclear Medicine

© Copyright 2019 SNMMI; all rights reserved.


http://jnm.snmjournals.org/content/early/2019/05/03/jnumed.119.226597

http://jnm.snmjournals.org/site/misc/permission.xhtml

http://jnm.snmjournals.org/site/subscriptions/online.xhtml


Date post:	03-Jan-2020
Category:	Documents
Upload:	others
View:	15 times
Download:	0 times

Comparison of static F-FDG-PET/CT (SUV, SUR) and dynamic F...

Documents