Introduction: The use of technetium 99m diethylenetriaminepentaacetic acid-galactosyl human serum albumin (99mTc-GSA) scintigraphy parameters, HH15 and LHL15, in assessing the future liver remnant function is not expedient because of their nonlinear behaviour against liver volume. Uptake rate constant for the binding of 99mTc-GSA to asialoglycoprotein receptors is probably more favourable, but the reported calculation methods are complex. We devised a simple method to calculate the uptake rate constant, KrGSA. Methods: Radioactivity counts for the entire liver and heart regions were extracted at 10, 20, and 30 min. Using whole liver and heart volumes measured from single-photon emission computed tomography images, free radioactivity corresponding to the liver blood pool was subtracted. The time activity curve was fitted to the equation L(t) = L(∞) × [1 − Exp (−kt)] using Microsoft Office Excel (add-in free programme Solver)®, where L(∞) is the count at plateau level and k denotes KrGSA. Results: KrGSA values accurately identified liver cirrhosis and were similar to the KICG. The areas under the curve for KrGSA and KICG in the receiver operating characteristic analysis were 0.808 and 0.795, respectively, and a good correlation was seen between KrGSA and KICG. Discussion/Conclusion: KrGSA can be utilized as an alternative to KICG in assessing the future liver remnant function.

© 2022 The Author(s). Published by S. Karger AG, Basel

Introduction

Technetium 99m diethylenetriaminepentaacetic acid-galactosyl human serum albumin (99mTc-GSA) scintigraphy is a well-established liver function test, in which the uptake of 99mTc-GSA reflects living hepatocytes [1, 2]. The most widely used indices are clearance index (HH15) and receptor index (LHL15). Conversely, the present gold standard of preoperative assessment of the liver function before hepatectomy is a fraction of plasma elimination rate constant of indocyanine green (ICG) clearance test that is expected in the future liver remnant (FLR): remKICG = KICG × FLR%, where KICG is the plasma elimination rate constant of ICG and FLR% is defined as the ratio of the FLR volume to the total liver volume multiplied by 100 [3]. This formula holds because the fraction of KICG value is linearly proportional to the FLR% [4]. Unlike KICG, HH15 and LHL15 are not suitable parameters for this purpose because their changing behaviour is not linearly proportional to the liver volume and therefore does not attract the attention of liver surgeons as much. On the other hand, the shortcoming of ICG clearance test is that it is not suitable in the presence of jaundice, which we often encounter in the patients with perihilar cholangiocarcinoma or liver cirrhosis. In addition, although not many, there are patients with constitutional ICG excretory defect. Conversely, 99mTc-GSA scintigraphy has a great advantage of being applicable even in patients with high bilirubinaemia or intrahepatic vascular shunts and those with the presence of a constitutional ICG excretory defect [5, 6]. Given that single-photon emission computed tomography (SPECT) during 99mTc-GSA scintigraphy has recently gained attention as a means of evaluating regional heterogeneity of liver function [7-9], the parameter development for 99mTc-GSA scintigraphy that can be expediently utilized as an alternative to KICG value is strongly anticipated.

Since the uptake of 99mTc-GSA at the hepatic asialoglycoprotein receptors, that is, GSA receptors, is a reverse reaction of 99mTc-GSA elimination from the blood. The most rational parameter that is comparable with KICG value seems to be the uptake rate constant of 99mTc-GSA. However, it is difficult to determine the uptake rate constant based on time activity curve (TAC) data alone in 99mTc-GSA scintigraphy because radioactivity at saturation (plateau) varies among individuals, depending on the number of GSA receptors, the quantity of GSA molecules administered, and the dose of radioactivity used. Further, given that scintigraphy uses a rapidly decaying isotope, it is extremely difficult to simultaneously standardize the GSA dose and the radioactivity on a patient-by-patient basis. To counter these problems, an attempt was made to use the degree of convexity of TAC as an index [10, 11]; however, it is still a fairly intuitive parameter that does not consider the radioactivity derived from receptor nonbinding 99mTc-GSA in the hepatic blood pool. A graphical approach to calculate the uptake rate constant has also been reported [12-14], and although it is theoretically appropriate, this calculation method is unlikely to be adopted in today’s digital era. What is needed now is to acquire a simple method for calculating an uptake rate constant of 99mTc-GSA and to determine whether the uptake rate constant is available for assessing liver function as an alter­native of KICG value. Therefore, we devised a method to calculate the uptake rate constant (KrGSA), without knowing the plateau value of TAC, according to a one-compartment model proposed by Shuke [15, 16], and verified the significance of KrGSA by comparing it with conventional parameters, including KICG value.

Materials and MethodsPatients

We retrospectively analyzed data from our patient cohort who were being evaluated for a potential liver resection at the Department of Gastroenterological Surgery, Akita University Hospital, between June 2010 and May 2020. We included only those who fulfilled the following criteria: (1) dynamic and SPECT data for 99mTc-GSA scintigraphy was available, (2) availability of concurrent ICG test data (measured at the serum bilirubin level <3.0 mg/dL, 0.974 ± 0.578 mg/dL, n = 89) and 99mTc-GSA scintigraphy data, and (3) no hepatic interventions, other than biliary drainage, before assessment. Thus, 89 patients were enrolled in this study. Patient demographics and modes of hepatectomies planned at the time of 99mTc-GSA scintigraphy examination are summarized in Table 1.

Table 1.

Hepatic function was comprehensively classified as normal, chronic hepatitis, or liver cirrhosis, based on clinical manifestation, hepatitis virus infection, and/or pathological background (in the patients who subsequently underwent surgery). For hepatic pathology, f1- and f2-grade fibrosis were defined as chronic hepatitis, while f3 and f4 were categorized as liver cirrhosis.

This study was conducted in accordance with the ethical standards of the Declaration of Helsinki. Informed consent was obtained from all patients before 99mTc-GSA scintigraphy, CT scanning, and the ICG test, separately. The study protocol was approved by the Ethics Committee of the Graduate School of Medicine, Akita University (approval code: 2078). According to the IRB instruction, the study protocol was informed on the web page of Akita University Hospital and obtained consent of patients in an opt-out manner.

99mTc-GSA Scintigraphy

A bolus of 1 mL of 99mTc-GSA (3 mg of GSA, 185 MBq or more, Nihon Medi-physics Co. Ltd., Nishinomiya, Japan) was intravenously injected. Dynamic scintigrams were obtained with a gamma camera (Symbia E®, Siemens Healthcare, Erlangen, Germany) equipped with a low-energy, high-resolution collimator. First, dynamic scintigraphy data during the first 30 min after bolus injection was acquired. Next, 360-degree SPECT images were acquired in 36 steps with a 5-degree rotation angle (20 s/step). The SPECT image (128 × 128 matrix) was reconstructed to yield volume data using the filtered back projection method and the Butterworth filter.

Calculation of the Uptake Rate Constant

In a one-compartment kinetic model, after a single bolus dose, the tracer reaches the liver through the plasma, is distributed in liver blood pool, and binds to the receptor. Among these processes, movement of the tracer between plasma and the liver blood pool is assumed to be in a steady state; however, that between the liver blood pool and the receptor is an irreversible unidirectional trap (Fig. 1). TAC from the liver ROI captures total hepatic radioactivity L(t) but is the sum of the tracer trapped by the receptor, L1(t) and that present in the liver blood pool, L2(t). These variables cannot be distinguished from each other on the TAC of the liver ROI. On the other hand, the L1(t) curve can be deemed to follow a first-order reaction because the binding process of the tracer (99mTc-GSA) to the receptor (asialoglycoprotein receptor) is irreversible, i.e., L1(t) = L1(∞) × [1 − Exp(−kt)]. Although the desired TAC is L1(t), subtraction of L2(t) from L(t) is not a simple mathematical operation because the L2(t) curve is dynamic and complex. However, we can confirm that the shape of the L2(t) curve is similar to the one shown in Figure 1. To arrive at the best estimate of L2(t), we calculated the volume of interest (VOI) for the liver and heart from the SPECT data (see below), and L2(t) was calculated as (radioactivity of the heart ROI at time t) × ([liver VOI volume] × 0.25 / [heart VOI volume]) because the liver blood pool is about 25% of the liver volume [17, 18] and the blood volume of the heart is almost equal to the heart volume.

Fig. 1.

Schematic representation of a one-compartment kinetic model. TACs are illustrated after bolus administration of the tracer, where three unknown parameters are GSA receptor uptake rate (KrGSA), total excretion rate (ke), liver blood pool (Vh). H(t) and L(t) indicate blood tracer concentration and tracer amount in the liver at time t, respectively. ke is assumed to be negligible during measurement within 30 min for 99mTc-GSA scintigraphy. Vh is estimated as 0.25 of liver volume. H(t) is calculated as whole heart tracer activity over whole heart volume.

The constant k denotes receptor binding rate of GSA (KrGSA) and was calculated by best fitting it to the equation L1(t) = L1(∞) × [1 − Exp(−kt)] using actual time activity of L1(t) obtained by dynamic scintigraphy and a free add-in mathematical optimization programme of Solver on Microsoft Office Excel® (Microsoft Inc., Redmond, WA, USA). Timepoints for measurement were set at 10, 20, and 30 min after injection of 99mTc-GSA because the decrease in radioactivity in the heart ROI achieved steady state and was almost negligible by 10 min. For the actual time radioactivity counts at ROIs, those from 9 to 10 min, 19–20 min, and 29–30 min after injection of 99mTc-GSA were used for the 10, 20, and 30 min timepoints, respectively (see online suppl. Material 1: Calculation of the Uptake Rate Constant, for detail; for all online suppl. material, see www.karger.com/doi/10.1159/000525892).

Calculation of Liver and Heart VOI

The liver and heart volumes were calculated from SPECT data using the outline extraction method (Fig. 2) [13]. As visualization of the liver in the SPECT image obtained after dynamic study was clear, the outline of the liver could be uniformly detected with a cut-off level of 20%. However, as the heart signal had significantly reduced by the time SPECT imaging could be acquired, it was difficult to extract the outline with a uniform cut-off level. Thus, cut-off levels were determined individually such that the best outline obtained was visually verified.

Fig. 2.

Outline extraction of liver and heart VOIs. The outline of the liver could be uniformly detected at a cut-off level of 20%. The outline of the heart could be extracted only using individually determined cut-off levels; the best outline obtained was visually checked. A, anterior; F, foot; H, head; L, left; P, posterior; R, right.

Curve Fitting Using Excel-Solver®

The equation, L1(t) = L1(∞) × [1 − Exp(−kt)] can be rearranged to give L1(∞) = L1(t) / [1 − Exp(−kt)]. As the saturated count, L1(∞) should be the same value for any (t, L1(t)). We generated this relationship for L1 (t) for t1–3 and for L1(∞) for t1–3 on an Excel sheet. For determining the k value that provides the best fit curve for L1 (t) = L1(∞) × [1 − Exp(−kt)] at the observed radioactivity requires us to find k such that L1(∞) for t1, L1(∞) for t2, and L1(∞) for t3 are all equal. Practically, we developed a programme to find a k value that will yield minimum difference between the L1(∞) that fulfils the curve passing the measured point for t1, t2, and t3 (see online suppl. Material 2: Curve fitting using Excel-Solver, for detail.) These operations were performed using the Excel-Solver® (Generalized Reduced Gradient Method) (Programmed Excel sheet is available in online suppl. Material 3: Excel sheet for KrGSA calculation.)

CT Volumtery

Dynamic contrast-enhanced CT scanning of the abdomen was performed on a 64-row scanner (Revolution CT® or Discovery CT 750 HD®, GE Healthcare Japan, Co. Ltd., Tokyo, Japan) at 0.625–1.25-mm intervals. Nonionic contrast media (650 mg iodine/kg) was intravenously injected in 30 s using an automatic power injector. Images were obtained in 4 phases, namely, at 10 (early arterial phase), 20 (late arterial phase), 45 (portal phase), and 165 (equilibrium phase) seconds. Volumetric analysis was performed using Digital Imaging and COmmunications in Medicine data for dynamic CT scanning and with the help of liver analysis application of 3D image analysis system SYNAPSE VINCENT® Ver.4.0 (Fujifilm Medical Co., Tokyo, Japan).

Statistical Analysis

All data are expressed as mean ± standard deviation (SD), unless otherwise mentioned. Statistical analyses were performed using SPSS, version 26, (SPSS Inc., Chicago, IL, USA). Correlation analysis was done using Pearson’s correlation coefficients, i.e., r and p values. Kolmogorov-Smirnov test and Levine’s test were used for testing normality and equality of variance, respectively. Kruskal-Wallis test with Bonferroni correction was applied for comparison of mean values. Receiver operating characteristic (ROC) analysis was used to test the ability of the two methods to accurately identify liver cirrhosis. The Youden index was utilized to estimate the best cut-off values, and p < 0.05 was considered statistically significant.

ResultsBasic Parameters for Calculating KrGSA

Liver volume was calculated to be 1,344 ± 289 mL and 1,226 ± 285 mL from 99mTc-GSA SPECT images and CT volumetry, respectively; the average difference between the two measurements was 207 mL. A good correlation was obtained between liver volumes calculated using 99mTc-GSA scintigraphy and CT volumetry (r = 0.669, n = 89, p < 0.001), indicating that liver volume obtained from SPECT images would permit subsequent calculation of KrGSA. Heart volume based on SPECT image data was calculated to be 639 ± 108 mL, and even though CT volumetry data was not available for comparison, this value is consistent with reported values of 408 mL/m2, given the average body surface area of the Japanese population [19]. Consequently, radioactivity counts due to liver blood pool constituted about 10% of the liver ROI counts.

KrGSA

KrGSA values ranged from 0.063 to 0.204 (0.132 ± 0.028, n = 89). First, we evaluated the ability of KrGSA to accurately identify three categories of liver damage, namely, normal, chronic hepatitis, and liver cirrhosis, and compared it to that with KICG (Fig. 3a). Both indices showed significantly lower values for liver cirrhosis, but there was no significant difference with respect to normal liver or chronic hepatitis. The ROC curves of KrGSA, KICG, HH15, and LHL15 for identifying liver cirrhosis are shown in Figure 3b, and the area under the curve (AUC) for these indices was 0.808, 0.795, 0.848, and 0.671, respectively. Although the discrimination abilities of KrGSA, KICG, and HH15 were equally good, that of LHL15 was disappointing. KrGSA was able to identify cirrhosis at a cut-off value of 0.117; this value was similar to that of KICG.

Fig. 3.

a KrGSA and KICG for the three categories of liver damage, as a box-and-whisker plot. × = mean. CH = chronic hepatitis; N = normal; Z = liver cirrhosis. KICG, indocyanine green elimination rate constant. b ROC curves for KrGSA, KICG, HH15, and LHL15, to distinguish liver cirrhosis. Cut-off values were determined from the largest Youden index. AUC, area under the curve; KICG, indocyanine green elimination rate constant.

Next, correlations between KrGSA and HH15, LHL15, KICG, or other conventional parameters of liver damage were assessed. Table 2 shows the correlation of KrGSA or KICG with conventional parameters of liver damage. KrGSA showed a strong correlation with LHL15 and HH15 (r = 0.65) but a little bit weaker correlation with KICG (r = 0.63). On the other hand, neither KrGSA nor KICG was correlated with platelet counts, serum albumin, cholinesterase, prothrombin time (INR), albumin-bilirubin score (ALBI score), or Model for End-Stage Liver Disease score (MELD-Na score). KrGSA did not necessarily agree with KICG; nevertheless, there was a nearly strong correlation. As depicted in Figure 4, although the AUC for HH15 in discriminating the liver damage was better than KrGSA, correlation with KICG was superior in KrGSA (r = 0.6276) than HH15 (r = −0.4692). LHL15 showed a better correlation with KICG than HH15. Conversely, the distribution range of LHL15 was significantly different from that of KICG and leads to a suggestion that LHL15 was not a good alternative to KICG to determine FLR function.

Table 2.

Correlation between KrGSA or KICG and other conventional liver damage parameters

Fig. 4.

Correlation with KICG in KrGSA, HH15 and LHL15. KICG, indocyanine green elimination rate constant.

remKrGSA versus remKICG

We have retrospectively calculated remKrGSA, defined as KrGSA × FLR%, according to the actual plan of hepatectomy modes at the time of 99mTc-GSA scintigraphy evaluation (Table 1). As illustrated in Figure 5, remKrGSA value showed a very strong correlation with remKICG value (r = 0.8949).

Fig. 5.

Correlation between remKrGSA and remKICG. remKrGSA = KrGSA × FLR%; remKICG = KICG × FLR%. KICG, indocyanine green elimination rate constant, FLR%, the ratio of the future liver remnant volume to the total liver volume multiplied by 100.

Discussion/Conclusion

We describe a new method for calculating the rate constant of the binding of the 99mTc-GSA to the GSA receptor. This method effectively incorporates TAC information obtained during dynamic studies to the extent that it is not too complex. This rate constant was named KrGSA based on liver uptake through receptor binding.

The range of KrGSA and KICG values in our cohort were similar. The minimum value was 0.063 because the cohort consisted of patients undergoing preoperative assessment for hepatectomy, and thus, the study population did not include patients with severe liver damage. Additional investigations among patients with severe liver cirrhosis are anticipated. Nevertheless, although there was no significant difference in values between normal liver and chronic hepatitis, KrGSA was significantly lower in liver cirrhosis and it could identify liver cirrhosis at a cut-off value of 0.117. This ability to identify liver damage was at the least not less than comparable that of KICG, which is thought to be the most reliable index, as shown by the AUC in ROC analysis. As expected, the correlation between KrGSA and LHL15, or HH15, remained only strong and did not reach the “very strong” level despite being obtained from identical measurements. This might arise from the fact that HH15 and LHL15 merely evaluate hepatic accumulation or blood retention of 99mTc-GSA at a given point in time (3 and 15 min after injection), and HH15 only reflects the radioactivity count in the heart pool. Limited use of TAC information while calculating LHL15 or HH15 seemed to have a substantial impact on the how TAC characteristics were utilized.

Two results are shown above: (1) there was a clear, positive correlation between KrGSA and KICG and (2) the measurement scale (range) of KrGSA value was equal to that of KICG. These findings support that the values of KrGSA and KICG are comparable with each other or can be used as an alternative. Unlike HH15 and LHL15, KrGSA is applicable for calculating remKICG – to be precise, it should be called remKrGSA. Conversely, the fact that correlation between KrGSA and KICG did not reach the “strong” level was not surprising because, theoretically, KrGSA and KICG represent different aspects of liver quality, and hence, cannot have perfect correlation. Specifically, while KrGSA is predominantly dependent on the number of asialoglycoprotein receptors on the hepatocyte membrane [20], KICG reflects not only membrane transport of the ICG molecule via OATP1B3 and NTCP but also hepatic bloodstream transport of ICG molecules to hepatic sinusoids [21, 22]. Therefore, KrGSA is less dependent on hepatic blood flow than KICG. Several reports have described the correlation coefficient between ICG test (ICGR15) and HH15 or LHL15 to be 0.477–0.574 and 0.52–0.717, respectively [23, 24]. We also report similar correlation coefficients of 0.469 for HH15 and 0.530 for LHL15. In contrast, the correlation between KrGSA and KICG was much better and may be due to a reduction in measurement errors when a theoretical curve was fit using sufficient information from TAC rather than HH15 or LHL15 values. Furthermore, a moderately suppressed but undoubtedly positive correlation between KrGSA and KICG (r = 0.6276) supports the notion that KrGSA can be an efficient and independent index of liver function. Despite such an independence of KrGSA from KICG, a very strong correlation between remKrGSA and remKICG (r = 0.8949) was observed in actual patients who are going to undergo hepatectomy. This finding appears to give support to the fact that KrGSA could be a useful alternative to KICG.

The limitations to this study are as follows. First, the patient cohort used in this study was biased towards a potential hepatectomy and those with greater liver damage, like decompensated liver cirrhosis, were not included. Second, subtracting radioactivity of the liver blood pool is not a direct method; rather it uses a factor of 0.25 to estimate the volume of the liver blood pool in the total liver volume. Nevertheless, this subtraction corrected radioactivity counts of the liver ROI by about 10% and appears to contribute to a more accurate TAC assessment than described elsewhere. Third, the present study could not directly demonstrate whether remKrGSA has sufficient sensitivity and specificity to predict posthepatectomy liver failure. However, this is because hepatectomies of a patient cohort carefully selected by remKICG in a single center did not experience sufficient posthepatectomy liver failure to evaluate them. Of course, in terms of patient safety, it is clinically good in a sense. Multicenter studies or big data analysis is anticipated. Fourth, this is a retrospective study from single institution, and prospective validation cohort of the calculation methods has not been attempted. Validation at other institutions is anticipated.

In conclusion, we have devised a method to accurately calculate the value of the uptake rate constant for the binding of the 99mTc-GSA to the GSA receptor using TAC information. This method does not require specially designed programmes, other than the Excel-Solver®. Along with SPECT and CT fusion images, the provision of KrGSA will contribute to expanding the playing field of 99mTc-GSA scintigraphy to provide safe hepatectomy. In particular, it will contribute to patients who cannot undergo ICG clearance test for some reason.

Acknowledgments

We express our sincere gratitude to Professor Emeritus Dr. Akira Nakamura, Department of Medical Information Science, Akita University School of Medicine, for valuable suggestions on Excel-Solver® and statistical analyses.

Statement of Ethics

This study was conducted in accordance with the ethical standards of the Declaration of Helsinki. This study protocol was reviewed and approved by the Ethics Committee of the Graduate School of Medicine, Akita University, approval number [2078]. All participants provided written informed consent for research use of their data obtained during disease treatment. The study protocol was made available on the web page of the Akita University Hospital and participants were provided the option to opt-out even after providing consent. No extra participant consent was required because of the retrospective study design.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

Dr. Yamamoto reports having received research funding from the Japan Society for the Promotion of Science (No. 20K09072). The other authors report no funding received in relation to this work.

Author Contributions

Concept/study design: Yuzo Yamamoto, Masatake Iida, and Yoshihiro Abukawa; acquisition of data: Yuzo Yamamoto, Masatake Iida, Kimihiko Sato, and Yoshihiro Abukawa; analysis and interpretation of data: Yuzo Yamamoto, Masatake Iida, Kimihiko Sato and Go Watanabe; drafting of manuscript: Yuzo Yamamoto; critical revision: Yasuhiko Nakagawa and Manabu Hashimoto.

Data Availability Statement

This article is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC). Usage and distribution for commercial purposes requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.