Elsevier

Photoacoustics 光声学

Volume 34, December 2023, 100569
第 34 卷, 2023 年 12 月, 100569
Photoacoustics

In vivo liver function reserve assessments in alcoholic liver disease by scalable photoacoustic imaging
通过可扩展光声成像对酒精性肝病中的体内肝功能储备评估

IF 7.1EI检索SCI升级版 医学1区SCI基础版 医学2区
https://doi.org/10.1016/j.pacs.2023.100569 Get rights and content 获取权利和内容
Under a Creative Commons license
在知识共享许可下
open access 开放获取

Abstract 摘要

We present a rapid and high-resolution photoacoustic imaging method for evaluating the liver function reserve (LFR). To validate its accuracy, we establish alcoholic liver disease (ALD) models and employ dual-wavelength spectral unmixing to assess oxygen metabolism. An empirical mathematical model fits the photoacoustic signals, obtaining liver metabolism curve and LFR parameters. Liver oxygen metabolism significantly drops in ALD with the emergence of abnormal hepatic lobular structure. ICG half-life remarkably extends from 241 to 568 s in ALD. A significant decline in LFR occurs in terminal region compared to central region, indicated by a 106.9 s delay in ICG half-life, likely due to hepatic artery and vein damage causing hypoxia and inadequate nutrition. Reduced glutathione repairs LFR with a 43% improvement by reducing alcohol-induced oxidative damage. Scalable photoacoustic imaging shows immense potential for assessing LFR in alcoholic-related diseases, providing assistance to early detection and management of liver disease.
我们提出了一种快速高分辨率的光声成像方法，用于评估肝功能储备（LFR）。为了验证其准确性，我们建立了酒精性肝病（ALD）模型，并采用双波长光谱解混技术评估氧代谢。一个经验数学模型拟合光声信号，获得肝脏代谢曲线和 LFR 参数。在 ALD 中，肝脏氧代谢显著下降，伴随异常肝小叶结构的出现。在 ALD 中，ICG 的半衰期显著延长，从 241 秒延长到 568 秒。与中央区域相比，终末区域的 LFR 显著下降，ICG 半衰期延迟 106.9 秒，这可能是由于肝动脉和静脉损伤导致缺氧和营养不足。还原型谷胱甘肽通过减少酒精诱导的氧化损伤，使 LFR 改善 43%。可扩展的光声成像在评估与酒精相关的疾病中的 LFR 方面显示出巨大潜力，为肝病的早期检测和管理提供了帮助。

Keywords 关键词

Scalable photoacoustic imaging
Dynamic contrast enhancement
Alcoholic liver disease
Liver function reserve

可扩展的光声成像
动态对比增强
酒精性肝病
肝功能储备

1. Introduction 1. 引言

Alcoholic liver disease (ALD) is a prevalent and serious health issue worldwide, resulting from the excessive consumption of alcohol [1], [2]. It poses a significant risk to individuals, leading to severe complications, ranging from steatosis to steatohepatitis [3], fibrosis [4], cirrhosis [5], and hepatocellular carcinoma [6]. Early detection and accurate assessment of ALD are crucial for effective management and improved patient outcomes.
酒精性肝病（ALD）是全球普遍存在的严重健康问题，源于过量饮酒。它对个人构成重大风险，导致严重并发症，从脂肪肝到脂肪性肝炎、纤维化、肝硬化和肝细胞癌。早期发现和准确评估 ALD 对于有效管理和改善患者预后至关重要。
In recent years, advancements in medical imaging techniques have opened new avenues for non-invasive assessment of liver diseases. Photoacoustic imaging (PAI) has emerged as a promising modality that combines the benefits of high-resolution imaging, deep tissue penetration [7], [8], and functional information [9]. PAI utilizes laser-induced photoacoustic signals to generate detailed images by detecting the acoustic waves produced from the absorption of pulsed laser light by tissue chromophores [10], [11]. PAI has demonstrated a variety of biological applications, such as detecting hemoglobin oxygen saturation (sO2) [12], [13], measuring blood flow velocity [14], [15], enabling whole-body imaging in small animals [16], and facilitating preclinical cancer diagnosis [17], [18], [19].
近年来，医学成像技术的进步为非侵入性评估肝脏疾病开辟了新的途径。光声成像（PAI）作为一种有前景的模式，结合了高分辨率成像、深层组织穿透和功能信息的优点。PAI 利用激光诱导的光声信号，通过检测组织色素对脉冲激光光的吸收所产生的声波来生成详细的图像。 PAI 已展示出多种生物应用，例如检测血红蛋白氧饱和度（sO₂）、测量血流速度、实现小动物的全身成像，以及促进临床前癌症诊断。
PAI holds great promise in ALD assessment, particularly in evaluating hepatic vascular structure and oxygen metabolism [20], [21], which are closely associated with liver function reserve (LFR) [16], [22]. The vascular network is vital for liver health, providing the necessary nutrients and oxygen [23], [24]. Disruption in vascular structure and oxygen metabolism indicate liver damage and impaired LFR [25]. LFR represents the metabolic capacity of liver and is a critical parameter for assessing disease severity and predicting patient outcomes [26]. Traditional static liver function tests and imaging modalities have limitations in comprehensive and real-time LFR evaluation. While hematological tests and clinical grading systems can identify liver lesions to some extent, they fall short in assessing LFR comprehensively [27], [28].
PAI 在 ALD 评估中具有很大潜力，特别是在评估肝脏血管结构和氧代谢方面，这与肝功能储备（LFR）密切相关。血管网络对肝脏健康至关重要，提供必要的营养和氧气。血管结构和氧代谢的破坏表明肝脏损伤和 LFR 受损。LFR 代表肝脏的代谢能力，是评估疾病严重性和预测患者结果的关键参数。传统的静态肝功能测试和影像学方法在全面和实时 LFR 评估方面存在局限性。 虽然血液学测试和临床分级系统在一定程度上可以识别肝脏病变，但它们在全面评估 LFR 方面仍然不足。
Imaging techniques like computed tomography and magnetic resonance imaging lack quantitative evaluation and real-time monitoring of LFR changes [29], [30].
影像技术如计算机断层扫描和磁共振成像缺乏对 LFR 变化的定量评估和实时监测。
The decline in indocyanine green (ICG) concentration reliably indicates liver function and is commonly employed in LFR assessment [10], [31]. ICG clearance is calculated by collecting blood samples are 2–4 times within 15 min after intravenous ICG infusion and measuring their absorbance in vitro [28]. However, this process is invasive, inconvenient, and lacks real-time LFR results. Current ALD diagnosis methods, such as liver biopsy and imaging techniques, have drawbacks including invasiveness [32], sampling errors, low specificity [33], and resolution limitations [34]. Consequently, there is an urgent need for a comprehensive and precise technology to assess the metabolic function and vascular structure, highlighting the importance of LFR assessment.
吲哚氰绿（ICG）浓度的下降可靠地指示肝功能，并常用于 LFR 评估[10]，[31]。ICG 清除率是通过在静脉注射 ICG 后 15 分钟内收集 2-4 次血样并测量其体外吸光度来计算的[28]。然而，这一过程具有侵入性、不便，并且缺乏实时 LFR 结果。目前的 ALD 诊断方法，如肝活检和影像学技术，存在侵入性[32]、取样误差、低特异性[33]和分辨率限制[34]等缺点。因此，迫切需要一种全面而精确的技术来评估代谢功能和血管结构，突显了 LFR 评估的重要性。
We aim to enhance the ability to assess the LFR in ALD patients by leveraging the advantages of PAI. To validate the accuracy and sensitivity of PAI in assessing LFR, we established an ALD mouse model at different stages. The high-resolution advantage of PAI allowed to visualize vascular structure differences in the liver at varying health levels and located lesions. Dual-wavelength spectral unmixing was employed to evaluate liver hemoglobin oxygen saturation (sO2), and an empirical mathematical model derived LFR-related parameters, including maximum peak time (tmax) and half-life (t1/2). Statistical analysis of these parameters facilitates a comprehensive comparison of the pathological features between normal liver and ALD liver, further enabling the evaluation of metabolic capacity in different liver sub-regions.
我们旨在通过利用 PAI 的优势来增强评估 ALD 患者 LFR 的能力。为了验证 PAI 在评估 LFR 方面的准确性和敏感性，我们在不同阶段建立了 ALD 小鼠模型。PAI 的高分辨率优势使得能够可视化肝脏在不同健康水平下的血管结构差异并定位病变。采用双波长光谱解混技术评估肝脏血红蛋白氧饱和度（sO2），并通过经验数学模型推导出与 LFR 相关的参数，包括最大峰值时间（tmax）和半衰期（t1/2）。对这些参数的统计分析有助于全面比较正常肝脏与 ALD 肝脏之间的病理特征，进一步使得能够评估不同肝脏亚区域的代谢能力。
In addition, reduced glutathione (GSH) is a potent antioxidant known for its important role in repairing liver function and alleviating alcohol-induced oxidative damage effects [6]. Combined GSH as a therapeutic intervention for ALD is expected to improve LFR and alleviate disease progression. PAI can potentially aid in monitoring the efficacy of GSH-based therapies by non-invasively assessing changes in the vascular structure and oxygen metabolism of the liver.
此外，减少型谷胱甘肽（GSH）是一种强效抗氧化剂，以其在修复肝功能和减轻酒精引起的氧化损伤方面的重要作用而闻名。将 GSH 作为 ALD 的治疗干预预计将改善 LFR 并减轻疾病进展。PAI 可能通过非侵入性评估肝脏血管结构和氧代谢的变化来帮助监测基于 GSH 的疗法的疗效。
We offer an innovative method for the application of scalable PAI to the assessment of LFR in ALD, which is expected to unravel the intricate relationship between vascular abnormalities, oxygen metabolism and LFR, thereby enhancing the understanding of the pathogenesis of ALD (Fig. 1). Further research and clinical application of PAI in the treatment of ALD have the potential to revolutionize the diagnosis and treatment of liver disease, ultimately improving patient outcomes and quality of life.
我们提供了一种创新的方法，将可扩展的 PAI 应用于 ALD 中 LFR 的评估，预计将揭示血管异常、氧代谢和 LFR 之间的复杂关系，从而增强对 ALD 发病机制的理解（图 1）。进一步研究和临床应用 PAI 在 ALD 治疗中的潜力有望彻底改变肝病的诊断和治疗，最终改善患者的预后和生活质量。
Fig. 1
  1. Download: Download high-res image (362KB)
    下载：下载高分辨率图像 (362KB)
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 1. Schematic diagram for establishment of alcoholic models and representative implementations of PAI.
图 1。酒精模型建立的示意图及 PAI 的代表性实现。

2. Material and methods 2. 材料和方法

2.1. Preparation of animals
2.1. 动物的准备

All animals were obtained from the Guangdong Experimental Animal Center. The experimental protocols were approved by the ethics committee of Guangdong provincial people′s hospital, and complied with all relevant ethical regulations. Male BALB/c-nude mice aged 6–8 weeks were housed in a 12-hour light/12-hour dark cycle with food and water. A total of 49 male mice were randomly divided into four groups: acute ALD (n = 14), chronic ALD (n = 7), GSH intervention (n = 7) and control (n = 21). Acute ALD group received alcohol (3 mg/kg body weight with 56% (v/v)) before imaging, while the chronic ALD group received 56% (v/v) alcohol (1 mg/kg body weight) daily for four weeks. The GSH group received 56% (v/v) alcohol (1 mg/kg body weight) intragastric administration, followed by an intravenous injection of reduced GSH 12 h later (5 mg/kg body weight), once a day for four weeks. The control group received a conventional diet.
所有动物均来自广东省实验动物中心。实验方案经过广东省人民医院伦理委员会批准，并遵循所有相关伦理规定。6-8 周龄的雄性 BALB/c 裸鼠在 12 小时光/12 小时暗的周期中饲养，提供食物和水。共 49 只雄性小鼠随机分为四组：急性 ALD 组（n=14）、慢性 ALD 组（n=7）、GSH 干预组（n=7）和对照组（n=21）。急性 ALD 组在成像前接受酒精（3 mg/kg 体重，56% (v/v)），而慢性 ALD 组每天接受 56% (v/v)酒精（1 mg/kg 体重）为期四周。GSH 组接受 56% (v/v)酒精（1 mg/kg 体重）胃内给药，然后在 12 小时后接受还原型 GSH 的静脉注射（5 mg/kg 体重），每天一次，持续四周。对照组接受常规饮食。

2.2. PAM experiment 2.2. PAM 实验

In the acute ALD mice liver structure imaging experiment, mice were anesthetized with 2% isoflurane. Abdominal incisions of 5 mm were made to expose the liver lobe. The functional information of the liver lobule was measured by dual wavelengths of 532 nm and 559 nm (high hemoglobin absorption), monitoring sO2 changes within 30 min. Scanning parameters were set as follows: imaging range 2 mm × 2 mm, scanning step 5 µm, and laser frequency 1 kHz. For the chronic ALD and the GSH groups, the liver lobes were exposed surgically as described above. The wavelength of 532 nm was selected to obtain the structural information of the liver lobule.
在急性 ALD 小鼠肝脏结构成像实验中，小鼠被麻醉于 2%异氟烷。进行了 5 毫米的腹部切口以暴露肝叶。通过 532 nm 和 559 nm 的双波长（高血红蛋白吸收）测量肝小叶的功能信息，监测 sO₂在 30 分钟内的变化。扫描参数设置如下：成像范围 2 mm × 2 mm，扫描步长 5 µm，激光频率 1 kHz。对于慢性 ALD 和 GSH 组，肝叶如上所述进行外科暴露。选择 532 nm 波长以获取肝小叶的结构信息。

2.3. PACT experiment 2.3. PACT 实验

Each mouse remained under 2% isoflurane anesthesia throughout the experiment, with the body fixed upright in the center of the tank. In the acute ALD group, images at wavelengths of 1064 nm and 780 nm were captured after alcohol administration to calculate sO2. In the chronic ALD and GSH groups, an indwelling needle was inserted into the mouse tail vein prior to imaging. An injection pump (D107886, KD Scientific Inc., USA) delivered 0.1 mL (3 mmol/kg body weight) of ICG solution at 0.2 mL/s. The laser wavelength was set at 780 nm, and a cross-section of mouse liver was scanned and imaged. Photoacoustic signals were continuously collected for 10 min. The control group was injected with the same volume of normal saline.
每只小鼠在实验过程中保持在 2%以下的异氟烷麻醉状态，身体固定在水槽的中心直立。在急性 ALD 组中，在酒精给药后捕获 1064 nm 和 780 nm 波长的图像以计算 sO2。在慢性 ALD 和 GSH 组中，在成像之前将一个留置针插入小鼠的尾静脉。注射泵（D107886，KD Scientific Inc.，美国）以 0.2 mL/s 的速度输送 0.1 mL（3 mmol/kg 体重）的ICG溶液。激光波长设置为 780 nm，并对小鼠肝脏的横截面进行扫描和成像。光声信号连续收集 10 分钟。对照组注射相同体积的生理盐水。

2.4. Photoacoustic imaging system
2.4. 光声成像系统

PAM (G2, InnoLaser) offers optical-to-acoustic resolution conversion with up to 5 µm optical resolution, supporting multi-wavelength laser scanning at 532 nm, 559 nm, and 750–840 nm. The acoustic signal generated by laser irradiation on the tissue was received by a 50 MHz ultrasonic sensor. The laser pulse repetition frequency is set to 10 kHz, and the X-axis B-scan rate is 25 Hz.
PAM (G2, InnoLaser) 提供高达 5 微米的光学分辨率转换，支持在 532 纳米、559 纳米和 750–840 纳米的多波长激光扫描。激光照射组织产生的声信号由 50 兆赫的超声传感器接收。激光脉冲重复频率设置为 10 千赫，X 轴 B 扫描速率为 25 赫。
PACT (SIP-PACT, Union Photoacoustic Technologies) can be used for organs cross-section imaging. It is equipped with a laser with a 20 Hz repetition rate, enabling imaging at wavelengths of 680–950 nm, 1064 nm, and 1190–2600 nm. A 512-element ring array ultrasonic transducer with a center frequency of 5.5 MHz is used to collect ultrasonic signals. The imaging resolution is 125 µm, and the imaging depth is approximately 5 cm.
PACT（SIP-PACT，联合光声技术）可用于器官横截面成像。它配备了一个重复频率为 20 Hz 的激光，能够在 680–950 nm、1064 nm 和 1190–2600 nm 的波长下进行成像。使用一个中心频率为 5.5 MHz 的 512 元环阵列超声换能器来收集超声信号。成像分辨率为 125 µm，成像深度约为 5 cm。

2.5. Signal and image processing
2.5. 信号与图像处理

MATLAB 9.8 software (R2020a, MathWorks) was used for image processing. Raw data was reconstructed by the dual acoustic back-projection algorithm to generate a PA image. Vascular features were extract from the PA images using the Hessian Filter Enhancement algorithm, providing more detailed information. The differential image was obtained by subtracting the baseline image from the post-injection. The portion of the image whose signal value was greater than the average pixel was chosen as the enhancement part and assigned color pixels. It was then overlaid on a 1064 nm background image.
MATLAB 9.8 软件（R2020a，MathWorks）用于图像处理。原始数据通过双声学反投影算法重建，以生成 PA 图像。血管特征通过 Hessian 滤波增强算法从 PA 图像中提取，提供了更详细的信息。通过从注射后的基线图像中减去基线图像获得差异图像。信号值大于平均像素的图像部分被选择为增强部分并分配颜色像素。然后将其叠加在 1064 nm 背景图像上。
To image the mouse liver, two different wavelengths of laser pulses (1064 nm and 780 nm) were used. The 1064 nm wavelength provides deep penetration, enabling clearer image of the liver structure. In the range of 600–800 nm, oxygenated hemoglobin (HbO2) and deoxygenated hemoglobin (Hb) have distinct absorption coefficients. The wavelength of 780 nm can achieve high contrast to distinguish the signals of oxyhemoglobin and deoxyhemoglobin. Photoacoustic images at 1064 nm and 780 nm were used for spectral unmixing to calculate sO2 [35].
为了成像小鼠肝脏，使用了两种不同波长的激光脉冲（1064 nm 和 780 nm）。1064 nm 波长提供了深层穿透，使肝脏结构的图像更加清晰。在 600–800 nm 范围内，氧合血红蛋白（HbO2）和去氧血红蛋白（Hb）具有不同的吸收系数。780 nm 的波长可以实现高对比度，以区分氧合血红蛋白和去氧血红蛋白的信号。使用 1064 nm 和 780 nm 的光声图像进行光谱解混，以计算 sO2[35]。
Binary conversion was applied to transform photoacoustic images into binary images represented by single-bit pixels (0 or 1). Intensity-based thresholding (0.3) was employed to identify and extract regions of interest. Abnormal blood vessels were recognized using a clustering algorithm based on vessel diameter characteristics [36].
二进制转换被应用于将光声图像转换为由单比特像素（0 或 1）表示的二进制图像。采用基于强度的阈值（0.3）来识别和提取感兴趣区域。使用基于血管直径特征的聚类算法识别异常血管。[36]

2.6. Empirical mathematical model
2.6. 实证数学模型

Origin 2016 was used for all data processing. The base value for monitoring liver ICG metabolism was obtained by averaging 300 image signal values before ICG injection. The image signal value was quantified within 10 min of ICG injection, and the metabolic curve was drawn after filtering and smoothing. Relative intensity (RI) of the ICG signal was calculated using the following formula [37]:(1)RI=[PAt−PA0]/[PA0]where PA(t) represents post-injection time-varying photoacoustic signal value, and PA(0) is the pre-injection basis value. The metabolic curve can be fitted according to the following formula [38]:(2)RI=0,0≤t<t0A·1−e−αt−t0·q·eβt−t0,t0≤twhere A is the maximum amplitude of RI, t is the time variable (s), α is the rate of drug uptake (s−1), β is the rate of drug metabolism (s−1), and q is the parameter associated with the slope of early uptake. Maximum peak time (tmax) of RI and half-life of metabolism (t1/2) were derived from the fitted curves.
Origin 2016 用于所有数据处理。监测肝脏 ICG 代谢的基值通过在 ICG 注射前平均 300 个图像信号值获得。图像信号值在 ICG 注射后 10 分钟内量化，经过滤波和平滑后绘制代谢曲线。ICG 信号的相对强度（R）使用以下公式计算： (1)RI=[PAt−PA0]/[PA0] ，其中 PA(t)表示注射后时间变化的光声信号值，PA(0)是注射前的基值。代谢曲线可以根据以下公式拟合： (2)RI=0,0≤t<t0A·1−e−αt−t0·q·eβt−t0,t0≤t ，其中 A 是 R 的最大幅度，t 是时间变量（秒），α是药物摄取速率（s−1），β是药物代谢速率（s−1），q 是与早期摄取斜率相关的参数。 最大峰值时间 (tmax) 和代谢的半衰期 (t1/2) 是从拟合曲线中得出的。

2.7. Statistical analysis
2.7. 统计分析

SPSS27 software (IBM, USA) was utilized for the statistical analysis. One-way ANOVA was used to compare the results between groups after checking the normality of sO2 data. A two-tailed P value less than 0.05 was indicative of a significant difference. Quantitative data of hepatic metabolic parameters were displayed as mean ± standard deviation.
使用 SPSS27 软件（IBM，美国）进行统计分析。在检查 sO2数据的正态性后，采用单因素方差分析比较组间结果。双尾 P 值小于 0.05 表示存在显著差异。肝脏代谢参数的定量数据以均值±标准差的形式显示。

2.8. Histopathology of liver
2.8. 肝脏的组织病理学

After paraffin embedding, the liver tissue was fixed with a 10% paraformaldehyde solution and sectioned. The tissue slices were stained by hematoxylin and eosin. Frozen liver slices were dried and incubated in 100% isopropanol before being stained with oil red O. Histopathological changes in the liver were examined under a microscope.
在石蜡包埋后，肝组织用 10%的多聚甲醛溶液固定并切片。组织切片用苏木精和伊红染色。冷冻肝切片被干燥并在 100%的异丙醇中孵育，然后用油红 O 染色。肝脏的组织病理变化在显微镜下进行检查。

3. Results 3. 结果

3.1. Blood oxygen levels in mice with acute ALD
3.1. 急性 ALD 小鼠的血氧水平

PAM and PACT monitored sO2 changes in the livers of mice after acute alcohol administration at multi-scale. PAM images (Fig. 2a) displayed the microstructure and function of the hepatic lobules, which are the primary liver unit. Alcohol gavage had no significant impact on the microstructure of hepatic lobules, but sO2 steadily declined (Before: 84.3 ± 2.1%, 30 min: 71.5 ± 5.0%, p < 0.05). In the control group, sO2 initially dropped and then increased (Before: 84.79 ± 2.86%, 30 min: 83.0 ± 2.4%, p < 0.05). Fig. 2b depicted the entire liver cross-section by PACT. In the acute ALD group, sO2 gradually decreased within 30 min after alcohol gavage (Before: 91.8 ± 5.0%, 30 min: 52.1 ± 3.4%, p < 0.01). In the control group, sO2 decreased slightly after 5 min of saline injection, then returned to normal (Before: 84.9 ± 0.2%, 30 min: 83.8 ± 1.4%, p < 0.01), consistent with PAM results.
PAM 和 PACT 监测了急性酒精给药后小鼠肝脏中的 sO2 变化。PAM 图像 (图 2a) 显示了肝小叶的微观结构和功能，肝小叶是肝脏的主要单位。酒精灌胃对肝小叶的微观结构没有显著影响，但 sO2 稳定下降（之前：84.3 ± 2.1%，30 分钟：71.5 ± 5.0%，p < 0.05）。在对照组中，sO2 最初下降然后上升（之前：84.79 ± 2.86%，30 分钟：83.0 ± 2.4%，p < 0.05）。图 2b 描绘了 PACT 的整个肝脏横截面。在急性 ALD 组中，sO2 在酒精灌胃后 30 分钟内逐渐减少（之前：91.8 ± 5.0%，30 分钟：52.1 ± 3.4%，p < 0.01）。在对照组中，sO2 在生理盐水注射后 5 分钟略有下降，然后恢复正常（之前：84.9 ± 0.2%，30 分钟：83.8 ± 1.4%，p < 0.01），与 PAM 结果一致。
Fig. 2
  1. Download: Download high-res image (2MB)
    下载：下载高分辨率图像 (2MB)
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 2. Microstructural and oxygen metabolism imaging of the liver in acute ALD model. (a) PAM was used to monitor sO2 images in the liver within 30 min in control and acute ALD groups. (b) PACT was used to monitor sO2 images in the liver within 30 min in control and acute ALD groups. Quantitative oximetry data from (c) PAM and (d) PACT. Scale bars: 500 µm (PAM), 5 mm (PACT). *, p < 0.05, **, p < 0.01.
图 2。急性 ALD 模型中肝脏的微观结构和氧代谢成像。(a) PAM 被用于监测对照组和急性 ALD 组中肝脏的 sO2 图像，时间为 30 分钟。(b) PACT 被用于监测对照组和急性 ALD 组中肝脏的 sO2 图像，时间为 30 分钟。来自 (c) PAM 和 (d) PACT 的定量 血氧测定 数据。比例尺：500 微米 (PAM)，5 毫米 (PACT)。*, p < 0.05，**，p < 0.01。

3.2. Evaluating the LFR in chronic ALD mice by PACT
3.2. 通过 PACT 评估慢性 ALD 小鼠中的 LFR

PACT enabled macroscopic liver imaging, visualizing the vascular network and hepatic cross-section tissues. Dynamic recordings captured vibrations of the portal vein, inferior vena cava, and abdominal aorta. ICG is an FDA-approved contrast agent, served as an optical indicator for clinical liver function testing. Fig. 3(a)–(c) illustrated the photoacoustic signals from liver slices recorded within 600 s after ICG injection for LFR assessment. The color signal indicated time-dependent ICG uptake in the liver. LFR curves for the three liver groups were shown in Fig. 3(d)–(f). Long-term alcohol consumption negatively affects liver blood vessels, further impairing liver function. Fig. 3(g) displayed prolonged tmax of ICG in the chronic ALD group compared to the control, while it was shorter following GSH administration (Control: 91.9 ± 3.0 s, Chronic ALD: 139.8 ± 12.4 s, GSH: 120.8 ± 13.4 s, p < 0.05). The rate of ICG clearance in the chronic ALD group was approximately twice as slow as the control group. Compared to the control group, the t1/2 in the GSH group was a slightly longer (Control: 241.0 ± 8.0 s, Chronic ALD: 568.2 ± 23.2 s, GSH: 356.1 ± 10.8 s, p < 0.01). These findings indicate that alcohol impaired liver LFR. GSH effectively repaired LFR (t1/2) with ∼43% improvement by reducing alcohol-induced oxidative damage. Interestingly, early intervention alleviated the negative effects of alcohol on liver function.
PACT 实现了宏观肝脏成像，能够可视化血管网络和肝脏横截面组织。动态记录捕捉了门静脉、下腔静脉和腹主动脉的振动。ICG 是一种 FDA 批准的对比剂，作为临床肝功能测试的光学指示剂。图 3(a)–(c)展示了在 ICG 注射后 600 秒内记录的肝脏切片的光声信号。颜色信号表明肝脏中 ICG 的时间依赖性摄取。三组肝脏的 LFR 曲线显示在图 3(d)–(f)中。长期饮酒对肝脏血管产生负面影响，进一步损害肝功能。图 3(g)显示慢性 ALD 组中 ICG 的 tmax延长，与对照组相比，而在 GSH 给药后则较短（对照组：91.9 ± 3。0 s，慢性 ALD：139.8 ± 12.4 s，GSH：120.8 ± 13.4 s，p < 0.05）。慢性 ALD 组的 ICG 清除率大约是对照组的两倍。与对照组相比，GSH 组的 t1/2稍长（对照组：241.0 ± 8.0 s，慢性 ALD：568.2 ± 23.2 s，GSH：356.1 ± 10.8 s，p < 0.01）。这些发现表明，酒精损害了肝脏 LFR。GSH 通过减少酒精诱导的氧化损伤有效修复了 LFR（t1/2），改善约 43%。有趣的是，早期干预减轻了酒精对肝功能的负面影响。
Fig. 3
  1. Download: Download high-res image (1MB)
    下载：下载高分辨率图像（1MB）
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 3. Time course of LFR in mice liver was measured by PACT. PACT images visualizing ICG clearance in (a) control, (b) chronic ALD, and (c) GSH groups. (d), (e), and (f) Liver metabolic curves in different groups. (g) Quantitative data of kinetic parameters tmax and t1/2. All scale bars are 5 mm. *, p < 0.05, **, p < 0.01.
图 3。小鼠肝脏中 LFR 的时间过程通过 PACT 测量。PACT 图像可视化了(a)对照组、(b)慢性 ALD 组和(c)GSH 组中的 ICG 清除情况。(d)、(e)和(f)不同组的肝脏代谢曲线。(g)动力学参数 tmax和 t1/2的定量数据。所有比例尺为 5 mm。*, p < 0.05, **, p < 0.01。

3.3. Changes in LFR based on different liver sub-areas
3.3. 基于不同肝脏亚区的 LFR 变化

Drug metabolism is a crucial liver function for eliminating foreign substances such as drugs, alcohol, and poisons. It was found that ICG was metabolized differently in the various areas of the whole liver. The results showed that t1/2 in the central region (yellow dashed box, Fig. 4(b)) was 208.5 ± 9.0 s, and 220.3 ± 8.9 s in the terminal region (red dashed box, Fig. 4(c)) in the control group. In the chronic ALD group, t1/2 in the central region (yellow dashed box, Fig. 4(e)) was 500.6 ± 21.4 s, and it was 619.3 ± 18.3 s in the terminal region (red dashed box, Fig. 4(f)). Analysis revealed a ∼400 s prolongation of t1/2 in the terminal region and ∼300 s in the central region in the chronic ALD group compared to the control. Long-term alcohol use affected liver metabolism more severely in the terminal region than in the central region. It is speculated that chronic ALD leads to hepatocyte necrosis, causing a disruption in the continuity of hepatic vascular structure. Moreover, the functions of the portal vein (providing nutrients) and hepatic artery (providing oxygen) were impaired. Fig. 4(a) and (d) depicted insufficient nourishment and oxygen in the tissues of the terminal liver region compared to the central region. Hence, the damage was more severe, resulting in a significant decline in LFR.
药物代谢是肝脏消除药物、酒精和毒素等外来物质的重要功能。研究发现，ICG 在整个肝脏的不同区域代谢方式不同。结果显示，控制组中央区域（黄色虚线框，图 4(b)）的 t1/2为 208.5 ± 9.0 s，而终末区域（红色虚线框，图 4(c)）为 220.3 ± 8.9 s。在慢性 ALD 组中，中央区域（黄色虚线框，图 4(e)）的 t1/2为 500.6 ± 21.4 s，而终末区域（红色虚线框，图 4(f)）为 619.3 ± 18.3 s。分析显示，与控制组相比，慢性 ALD 组终末区域的 t1/2延长了约 400 s，中央区域延长了约 300 s。长期饮酒对终末区域的肝脏代谢影响比中央区域更为严重。 据推测，慢性酒精性肝病导致肝细胞坏死，造成肝脏血管结构的连续性中断。此外，门静脉（提供营养）和肝动脉（提供氧气）的功能受到损害。图 4（a）和（d）显示了终末肝区的组织与中央区域相比，营养和氧气不足。因此，损伤更为严重，导致肝脏功能显著下降。
Fig. 4
  1. Download: Download high-res image (597KB)
    下载：下载高分辨率图像 (597KB)
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 4. Time course of LFR in mice liver sub-areas. Blood vessels transporting nutrients and oxygen in the liver in (a) control and (d) chronic ALD groups. Metabolic curves of the central region in (b) control and (c) chronic ALD groups. Metabolic curves of the terminal region in (e) control and (f) chronic ALD groups. All scale bars are 5 mm.
图 4。小鼠肝脏亚区 LFR 的时间过程。血管在(a)对照组和(d)慢性 ALD 组中运输肝脏的营养和氧气。中央区域的代谢曲线在(b)对照组和(c)慢性 ALD 组中。终末区域的代谢曲线在(e)对照组和(f)慢性 ALD 组中。所有比例尺为 5 毫米。

3.4. Microscopic imaging of liver structures in chronic ALD mice
3.4. 慢性酒精性肝病小鼠肝脏结构的显微成像

To examine micro-level structural abnormalities in the livers of chronic ALD mice, PAM scans were conducted. Fig. 5(a) revealed distinct liver lobule structure and well-organized blood vessels in healthy mice. In the chronic ALD group, blood vessels appeared blurry with numerous vascular nodules. Furthermore, liver lobes presented partial edema and fatty lesions. The GSH group showed mild vessel swelling and a few vascular nodules, primarily due to local inflammation caused by alcohol abuse. Fig. 5(b) binarized the microscopic images to highlight vascular continuity. The chronic ALD group showed a wide range of vascular signal loss, indicating a higher prevalence of steatosis around vessels. In comparison, the GSH group exhibited less damage. Fig. 5(c) utilized a vascular feature algorithm to identify the abnormal vascular region in red. The chronic ALD group displayed the largest abnormal vascular area.
为了检查慢性 ALD 小鼠肝脏的微观结构异常，进行了 PAM 扫描。图 5(a)显示健康小鼠的肝小叶结构清晰，血管组织良好。在慢性 ALD 组中，血管显得模糊，并伴有大量血管结节。此外，肝叶呈现部分水肿和脂肪病变。GSH 组显示轻微的血管肿胀和少量血管结节，主要是由于酒精滥用引起的局部炎症。图 5(b)对显微图像进行了二值化，以突出血管的连续性。慢性 ALD 组显示出广泛的血管信号丧失，表明血管周围的脂肪变性发生率更高。相比之下，GSH 组的损伤较少。图 5(c)利用血管特征算法识别异常血管区域，并用红色标出。慢性 ALD 组显示出最大的异常血管区域。
Fig. 5
  1. Download: Download high-res image (2MB)
    下载：下载高分辨率图像 (2MB)
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 5. Structural images of the liver were obtained by PAM at 532 nm. (a) Photoacoustic images. (b) The image after filtering and binarization of figure (a). (c) Abnormal blood vessel images by feature extraction. All scale bars are 500 µm.
图 5。肝脏的结构图像通过 532 nm 的光声显像获得。(a) 光声图像。(b) 图(a)经过滤波和二值化后的图像。(c) 通过特征提取得到的异常血管图像。所有比例尺为 500 µm。

3.5. Pathophysiological examination
3.5. 病理生理检查

The pathogenic structure of chronic ALD was showed in Fig. 6(a), highlighting noticeable pathological abnormalities compared to the control group. These abnormalities included disorganized hepatic cords, irregularly arranged hepatic plates, neutrophil infiltration, hepatocyte vacuolation, mallory body presence, and steatosis (shown in Ⅰ and II in the second row). Similarly, Fig. 6(b) showed a more severe fatty liver in the chronic ALD group.
慢性 ALD 的致病结构在图 6(a)中显示，与对照组相比，突出了明显的病理异常。这些异常包括肝索紊乱、肝板不规则排列、中性粒细胞浸润、肝细胞空泡化、马洛里小体的存在，以及脂肪变性（在第二行的Ⅰ和Ⅱ中显示）。同样，图 6(b)显示了慢性 ALD 组中更严重的脂肪肝。
Fig. 6
  1. Download: Download high-res image (2MB)
    下载：下载高分辨率图像 (2MB)
  2. Download: Download full-size image
    下载：下载完整尺寸的图像

Fig. 6. Pathological sections of mouse livers from three groups were stained with hematoxylin and eosin. (a) The sections were magnified at × 100 and × 400 in Row 1 and Row 2, respectively. The second row is an enlarged image of the area enclosed by the yellow box from the first row. (b) The oil-red O section was magnified at × 400. (c) Blood biochemical index. (d) Enzyme activity.
图 6。小鼠肝脏的病理切片来自三个组，使用苏木精和伊红染色。(a) 第一行和第二行的切片分别放大至×100 和×400。第二行是第一行中黄色框所包围区域的放大图。(b) 油红 O 切片放大至×400。(c) 血液生化指标。(d) 酶活性。

The changes in serum alanine transaminase (ALT), aspartate transaminase (AST), triglyceride (TG), and glutathione peroxidase (GSH-PX) can reflect the degree of ethanol-induced liver injury, as shown in Fig. 6(c). The ALT, AST, and TG levels in the chronic ALD group were 110%, 163%, and 73% higher than the control, whereas GSH-PX was 20% lower. After GSH intervention, serum ALT and AST levels in chronic ALD mice decreased to some extent. Acetaldehyde dehydrogenase (ADH) and alcohol dehydrogenase (ALDH) are key enzymes in ethanol metabolism. Compared to the control group, the ADH activity in chronic ALD group was noticeably reduced, although it increased after GSH intervention, as shown in Fig. 6(c). Liver ALDH activity in both the chronic ALD and GSH groups increased by 43% and 118%, respectively. These results indicated that the liver of chronically alcohol-consuming mice could better metabolize alcohol with the help of GSH.
血清中丙氨酸转氨酶（ALT）、天冬氨酸转氨酶（AST）、甘油三酯（TG）和谷胱甘肽过氧化物酶（GSH-PX）的变化可以反映乙醇诱导的肝损伤程度，如图 6(c)所示。慢性 ALD 组的 ALT、AST 和 TG 水平分别比对照组高 110%、163%和 73%，而 GSH-PX 则低 20%。在 GSH 干预后，慢性 ALD 小鼠的血清 ALT 和 AST 水平在一定程度上下降。乙醛脱氢酶（ADH）和酒精脱氢酶（ALDH）是乙醇代谢中的关键酶。与对照组相比，慢性 ALD 组的 ADH 活性明显降低，尽管在 GSH 干预后有所增加，如图 6(c)所示。慢性 ALD 组和 GSH 组的肝 ALDH 活性分别增加了 43%和 118%。 这些结果表明，长期饮酒的小鼠的肝脏在谷胱甘肽的帮助下能够更好地代谢酒精。

4. Discussion 4. 讨论

In this study, dynamic PAI was utilized to demonstrate the microscopic and macroscopic features of the early stages of ALD. Traditional ALD diagnostic methods, such as liver biopsy and various imaging techniques, suffer from invasiveness, sampling errors, low specificity, and limited resolution. Therefore, we combined PAM and PACT to provide a more comprehensive perspective for ALD research. Alcohol consumption altered liver vascular structure and functional parameters, including oxygen metabolism, demonstrated through PAM and PACT. Quantitative analysis of the light absorption intensity of HbO2 and Hb at different wavelengths allows accurate assessment the sO2 levels. Excessive alcohol intake rapidly decreases the sO2 levels in the liver, causing acute hypoxia and potential liver failure. Studies suggest alcohol hinders the absorption of oxygen by hemoglobin, causing red blood cell agglutination and impaired organ oxygen delivery [39].
在这项研究中，动态 PAI 被用来展示 ALD 早期阶段的微观和宏观特征。传统的 ALD 诊断方法，如肝活检和各种成像技术，存在侵入性、取样误差、特异性低和分辨率有限等问题。因此，我们结合了 PAM 和 PACT，为 ALD 研究提供了更全面的视角。酒精消费改变了肝脏血管结构和功能参数，包括通过 PAM 和 PACT 展示的氧代谢。对不同波长下 HbO2和 Hb 的光吸收强度的定量分析可以准确评估 sO2水平。过量饮酒迅速降低肝脏中的 sO2水平，导致急性缺氧和潜在的肝功能衰竭。研究表明，酒精阻碍了血红蛋白对氧的吸收，导致红细胞聚集和器官氧气输送受损。[39]
ICG concentration decline is widely used to reflect liver function. However, ICG clearance rate testing is invasive and not real-time. We present a non-invasive method based on rapid and dynamic PACT for accurate LFR assessment. Understanding liver metabolism is crucial for optimal medication therapy. ICG is a key dye for monitoring LFR and hemodynamics. Combining ICG with photoacoustic technology enables us to evaluate the severity of liver disease by measuring dye metabolism and clearance rate of the. We quantified the time course curve of ICG by measuring photoacoustic signal intensity changes. Results showed significant metabolic differences between healthy and sick livers, with prolonged tmax and t1/2 of ICG observed after alcohol injury. GSH intervention partially recovers liver metabolic capacity and reduces vascular steatosis severity. This non-invasive method, with fast, multi-wavelength, wide-field, and high-resolution capabilities, provides a new strategy for clinical LFR evaluation.
ICG 浓度下降被广泛用于反映肝功能。然而，ICG 清除率测试是侵入性的，并且不是实时的。我们提出了一种基于快速动态光声成像（PACT）的非侵入性方法，用于准确评估肝脏功能。理解肝脏代谢对优化药物治疗至关重要。ICG 是监测肝脏功能和血流动力学的关键染料。将 ICG 与光声技术结合，使我们能够通过测量染料代谢和清除率来评估肝病的严重程度。我们通过测量光声信号强度变化量化了 ICG 的时间过程曲线。结果显示健康肝脏与病肝之间存在显著的代谢差异，观察到酒精损伤后 ICG 的 tmax和 t1/2延长。GSH 干预部分恢复了肝脏的代谢能力，并减少了血管脂肪变性严重程度。这种具有快速、多波长、宽视野和高分辨率能力的非侵入性方法，为临床肝脏功能评估提供了一种新策略。
Alcohol primarily damages the liver terminal region, impacting the nutrients and oxygen transport during ALD [35]. The integrity and functionality of the liver blood vessels are also damaged. Moreover, since the terminal region is far from the center, these factors lead to insufficient nutrient supply (portal vein) and oxygen supply (hepatic artery) to the terminal region. PAI can zone the LFR by accurately evaluating the metabolic capacity of different sub-areas.
酒精主要损害肝脏末端区域，影响 ALD 期间的营养和氧气运输。肝脏血管的完整性和功能也受到损害。此外，由于末端区域远离中心，这些因素导致末端区域营养供应（门静脉）和氧气供应（肝动脉）不足。PAI 可以通过准确评估不同子区域的代谢能力来划分 LFR。
The liver converts alcohol into water and carbon dioxide [3], [23]. Prolonged alcohol consumption alters enzyme activity. According to studies, ethanol exposure generates free radicals and oxidants, increasing GSH utilization for detoxification [40]. GSH is crucial in mitigating alcohol-induced oxidative damage.
肝脏将酒精转化为水和二氧化碳。长期饮酒会改变酶活性。根据研究，乙醇暴露会产生自由基和氧化剂，增加谷胱甘肽的利用以进行解毒。谷胱甘肽在减轻酒精引起的氧化损伤中至关重要。
This study utilizes two imaging technologies: Photoacoustic microscopy excels in high-resolution imaging, revealing microvascular morphology and providing vital tissue functionality information, including oxygen saturation. Additionally, its precise localization capabilities enhance accuracy and visual effectiveness in detecting anomalous blood vessels. On the other hand, photoacoustic computed tomography demonstrates real-time dynamic visualizations of the liver noninvasively. It offers a powerful tool for diagnosing and monitoring liver diseases with broad clinical applications. By harnessing the synergies of these two imaging systems, this study maximizes the unique advantages of photoacoustic technology in liver research and medical applications.
本研究利用两种成像技术：光声显微镜在高分辨率成像方面表现出色，揭示微血管形态并提供重要的组织功能信息，包括氧饱和度。此外，其精确的定位能力增强了检测异常血管的准确性和视觉效果。另一方面，光声计算机断层扫描展示了肝脏的实时动态可视化，非侵入性地提供了强大的工具用于诊断和监测肝脏疾病，具有广泛的临床应用。通过利用这两种成像系统的协同作用，本研究最大限度地发挥了光声技术在肝脏研究和医疗应用中的独特优势。
In summary, recent advancements in PAI hold promise for assessing ALD by visualizing and quantifying liver vascular structure, oxygen saturation, and LFR. This non-invasive technique has significant clinical applications in ALD diagnosis and management. Ongoing research in PAI techniques have the potential to revolutionize ALD assessment, leading to improved patient outcomes.
总之，最近在光声成像（PAI）方面的进展为评估酒精性肝病（ALD）提供了希望，通过可视化和量化肝脏血管结构、氧饱和度和肝脏血流量（LFR）。这种非侵入性技术在酒精性肝病的诊断和管理中具有重要的临床应用。光声成像技术的持续研究有可能彻底改变酒精性肝病的评估，从而改善患者的治疗结果。

5. Conclusion 5. 结论

We introduced an innovative multi-scale PAI technique for precise measurement of LFR. Compared to traditional clinical imaging, PAI offers non-invasiveness and high-contrast visualization, potentially enhancing comprehensive assessment of hepatic diseases. However, challenges in imaging speed arise due to limitations in laser frequency and B-scan rates. PAM primarily targets surface microvasculature, yet deep vascular lesions are crucial for hepatic disease diagnosis. Our next step improvements involve enhancing imaging speed and depth while maintaining system stability through higher laser frequencies, and optimized scanning algorithms.
我们引入了一种创新的多尺度 PAI 技术，用于精确测量 LFR。与传统的临床成像相比，PAI 提供了非侵入性和高对比度的可视化，可能增强对肝脏疾病的综合评估。然而，由于激光频率和 B 扫描速率的限制，成像速度面临挑战。PAM 主要针对表面微血管，但深层血管病变对于肝脏疾病的诊断至关重要。我们下一步的改进涉及在保持系统稳定性的同时，通过更高的激光频率和优化的扫描算法来提高成像速度和深度。
In summary, our research provides opportunities for hepatic disease diagnosis and monitoring. We anticipate wide interest across research domains, advancing medical imaging and disease diagnosis research.
总之，我们的研究为肝脏疾病的诊断和监测提供了机会。我们预期在研究领域中会引起广泛的兴趣，推动医学影像和疾病诊断研究的发展。

Funding 资金

This investigation was supported by grants from National Natural Science Foundation of China (82372010&82302253) and GDPH Supporting Fund for Talent Program (KJ01200638).
本研究得到了中国国家自然科学基金（82372010 和 82302253）以及 GDPH 人才计划支持基金（KJ01200638）的资助。

Declaration of Competing Interest
利益冲突声明

The authors declare that there are no conflicts of interest.
作者声明不存在利益冲突。

Acknowledgments 致谢

T. Sun and J. Lv contributed equally to this work. All authors contributed to the study conception and design. All authors read and approved the final manuscript.

Data availability

For inquiries regarding the availability of the code, you are welcome to contact after the article publication, due to intellectual property considerations.

References

Cited by (25)

  • Exosomal ncRNAs in liquid biopsies for lung cancer

    2025, Clinica Chimica Acta

  • Reconstruction error based implicit regularization method and its engineering application to lung cancer diagnosis

    2025, Engineering Applications of Artificial Intelligence

  • DSEUNet: A lightweight UNet for dynamic space grouping enhancement for skin lesion segmentation

    2024, Expert Systems with Applications
    Citation Excerpt :

    Accurate diagnosis of various diseases is one of the critical demands nowadays due to potential benefits that can significantly enhance treatment results, decrease healthcare budgets, and recover patient health by empowering appropriate and besieged involvements (Chen et al., 2022; Sun et al., 2023).

  • Comprehensive insights into oral squamous cell carcinoma: Diagnosis, pathogenesis, and therapeutic advances

    2024, Pathology Research and Practice
    Citation Excerpt :

    There are a few ways to build theranostic nanoparticles: attachment of an anti-cancer regimen or nanoparticles that stimulate light, combining the nanoparticles with the image contrast enhancer, and finally loading the complex into one nanoparticle [272]. Pretargeting therapy is done to label and locate the target and collect information regarding the tumor using a few methods avidin-biotin interaction, Staudinger ligation, trans-cyclooctene/tetrazine bioorthogonal click reaction, copper-free azide-cyclooctyne click chemistry, and peptide nucleic acids interaction) and followed by second therapeutic components which selectively cross-link with the pre-target on the cell surface [273]. New advancements have been reported including superparamagnetic nanoparticles MRI contrast, carbon nanomaterials used for optical and photoacoustic imaging, stealth and polymer-based nanoparticles with photothermal properties, gene delivery using nanoparticle chauffeur, stimuli-responsive nanomaterials and multifunctional nanoparticles [274].

  • Direct Monitoring of Whole-Brain Electrodynamics via High-Spatiotemporal-Resolution Photoacoustics with Voltage-Sensitive Dye

    2024, Laser and Photonics Reviews

  • Advancements in tissue engineering for cardiovascular health: a biomedical engineering perspective

    2024, Frontiers in Bioengineering and Biotechnology
View all citing articles on Scopus
Tong Sun is a master's student at South China Normal University. Her research interests are in the development of photoacoustic imaging systems for medical applications.
Jing Lv is currently a postdoctoral fellow of Research Center of Medical Sciences of Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences. She got his Ph.D. degree from Xiamen University, China in 2022. Her main research interests are the structural and functional photoacoustic imaging application of cardiovascular and cerebrovascular disease models as well as contrast-enhanced quantitative, dynamic, metabolic imaging of livers and kidneys.
1
These authors contributed equally to this work.
© 2023 The Author(s). Published by Elsevier GmbH.