Evaluation of skeletal muscle microcirculation via magnetic resonance intravoxel incoherent motion perfusion imaging in an acute ischemia rabbit model

Abstract

Background

Acute limb ischemia (ALI) necessitates prompt intervention to prevent severe complications such as amputation. Current clinical assessments lack reliable quantitative methods for gauging skeletal muscle ischemia severity. Intravoxel incoherent motion (IVIM) perfusion imaging is a noninvasive approach for quantifying microvascular perfusion. We aimed to assess microcirculation alterations in rabbit vastus lateralis muscle following acute ischemia using IVIM.

Methods

Acute ischemia models were established in 25 New Zealand white rabbits through arterial ligation of their right hind legs. Magnetic resonance imaging (MRI) examinations of the vastus lateralis muscle were conducted hourly postsurgery for a duration of 7 hours. The scan sequences included IVIM, adenosine triphosphate (APT), T1-weighted imaging (T1WI), T2-weighted imaging with fat suppression (T2WI-FS), T2 mapping, and diffusion-weighted imaging (DWI). The correlations between MRI results, ischemic time, and pathological changes were analyzed.

Results

The perfusion fraction (f) of ischemic muscle significantly decreased from 6.19%±1.13% at 1 hour to 2.50%±0.64% at 7 hours (control: 7.19%±1.03%), representing a strong negative correlation with ischemic duration (r=−0.790). The true diffusion coefficient (D) remained relatively stable [(1.41–1.46)×10⁻³ mm²/s] but was slightly elevated compared to controls. The pseudo-diffusion coefficient (D*) showed a sharp increase at 5 hours [(74.01±5.79)×10⁻³ mm²/s]; control: [(61.28±9.31)×10⁻³ mm²/s], followed by a drop at 6 hours [(59.44±15.77)×10⁻³ mm²/s], suggesting sudden structural changes, which were confirmed by histopathology. T2WI-FS and DWI showed increased signal intensity in ischemic muscle, with the T2 relaxation times being significantly elevated (P<0.001) and positively correlated with ischemic duration (r=0.807). Apparent diffusion coefficient (ADC) values also increased with time (r=0.623). The amide proton transfer effect was enhanced in ischemic skeletal muscle throughout the 2–7-hour post-ischemic period (ischemic: 2.26%±0.39% at the 7th hour vs. control: 1.77%±0.33%; P<0.05).

Conclusions

MRI effectively visualizes and detects skeletal muscle ischemia, with IVIM-derived f values providing a quantitative measure of microcirculatory impairment. D* potentially serves as a biomarker for identifying irreversible muscle fiber damage and may be a valuable tool for the quantitative assessment of ischemic injury to skeletal muscle.

Introduction

Acute limb ischemia (ALI) is a critical condition that may result in severe complications, including muscle infarction and functional impairment. Studies indicate that up to 25% of amputations are attributable to skeletal muscle ischemia (1). Evidence demonstrates that prompt intervention to restore blood flow in patients with ALI, regardless of severity, significantly improves prognosis and reduces the associated complications (2). Consequently, accurate assessment of skeletal muscle ischemia severity is essential for formulating an effective treatment strategy for ALI.

Currently, the clinical evaluation of skeletal muscle ischemia primarily relies on two approaches: estimation of ischemia duration based on clinical manifestations and imaging-based assessments. The former depends heavily on subjective clinical judgment and is often unreliable in cases of delayed ischemia. Imaging techniques, such as Doppler ultrasound, computed tomography angiography (CTA), and magnetic resonance angiography (MRA), allow for the determination of vascular obstruction sites, the extent of occlusion, and the status of collateral circulation (3-8). However, these modalities are limited in their ability to quantitatively assess the severity of ischemic damage to skeletal muscle tissue. Although digital subtraction angiography (DSA) provides enhanced visualization of vascular architecture due to its high-resolution imaging, its invasive nature and associated risk of complications limit its clinical applicability to cases even for which it is clearly indicated (9). There thus exists an urgent need to develop quantitative and noninvasive assessment tools that can accurately monitor the dynamic changes associated with acute skeletal muscle ischemia in both clinical practice and research settings.

In recent years, advancements in magnetic resonance imaging (MRI) techniques have progressed rapidly, allowing them to be applied in the assessment of blood perfusion. Intravoxel incoherent motion (IVIM) perfusion imaging, an MRI-based technology, extracts microvascular perfusion information from diffusion-weighted imaging (DWI) through use of multiple b-values, thereby facilitating the evaluation of capillary blood flow perfusion without the need for exogenous contrast agents (10,11). This technique yields three key parameters: the molecular diffusion coefficient (D), the pseudo-diffusion coefficient (D*), and the perfusion fraction (f), which reflect the tissue diffusion coefficient, capillary blood flow velocity, and capillary blood flow fraction, respectively. Studies have demonstrated that physiological muscle perfusion in the forearm, both at rest and postexercise, can be quantitatively assessed via analysis of perfusion-dependent parameters such as D* and f D* (12). Moreover, D and D* values in forearm skeletal muscles increase following exercise. IVIM parameters have also been employed to evaluate blood perfusion in the lower limbs (13), forearms (12,14), and masticatory muscles (15), with findings indicating that IVIM parameters can effectively reflect muscle blood flow perfusion under various conditions.

Additionally, magnetic resonance amide proton transfer (APT) imaging has been applied to detect changes in abnormal proteins and polypeptides in tumors and other pathological conditions through measurement of the concentration of exchangeable amide protons (16-18). This technique can also generate contrast based on endogenous pH variations (19). Research has shown that APT imaging can identify ischemic lesions (18), particularly in the early stages of stroke, when impaired cerebral blood flow results in reduced APT effects in ischemic tissues relative to healthy tissues (20,21). However, the application of APT imaging for assessing the extent of skeletal muscle ischemia has not yet been intensively investigated.

Based on the above-mentioned advancements in MRI-based blood perfusion assessment, this study aimed to employ IVIM and APT imaging techniques to quantitatively evaluate the severity of acute skeletal muscle ischemia and to characterize dynamic changes in relevant parameters during the ischemic process. Furthermore, MRI techniques such as DWI and T2-weighted imaging (T2WI), in conjunction with histopathological analysis, were employed to validate the imaging findings. Time-series imaging of an in vivo rabbit model of femoral artery ischemia demonstrated that IVIM imaging can effectively evaluate the progression and severity of early skeletal muscle ischemia. The results of this study support the viability of a novel and viable approach for the quantitative imaging assessment of acute skeletal muscle ischemia. We present this article in accordance with the ARRIVE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-711/rc).

Methods

Rabbit hindlimb acute ischemia model

All experiments in this study were performed under a project license (No. 2023-0445) granted by the Animal Welfare and Ethics Committee of Shanghai Sixth People’s Hospital and in compliance with institutional guidelines for the care and use of animals.

This study included 25 adult New Zealand White rabbits (14 males and 11 females) with an average weight of 2.5 kg. Prior to the experiment, each rabbit was weighed and anesthetized via marginal ear vein injection with 1.5% sodium pentobarbital (1 mL/kg). The lower abdomen was then prepared as the surgical site. Following hair removal and rigorous disinfection, a midline laparotomy was performed to expose the lower abdominal aorta, bilateral common iliac arteries, and their proximal branches. During the surgery, 3-0 nonabsorbable sutures were used to ligate and sever the right (left) common iliac artery and its proximal internal iliac artery, as well as the proximal internal iliac artery on the opposite side. Throughout the procedure, close monitoring was conducted to ensure that no significant bleeding occurred in the surgical area. The abdominal wall incision was then meticulously closed and sutured. In order to further restrict collateral circulation in the tail arteries, we applied binding straps to securely and carefully tie the tails of all rabbits, a measure intended to enhance the ischemic state of the experimental model. One rabbit was excluded from the study due to an anesthetic complication. Subsequently, 20 rabbits underwent MRI scanning for evaluation. Within this cohort, two rabbits were randomly selected for DSA to confirm segmental occlusion of the ipsilateral iliac vessels. Additionally, muscle biopsies were obtained from the lateral femoral muscles of two other randomly selected rabbits with confirmed ischemia at each time point for histological analysis.

MRI scanning protocol

In this study, both hind limbs of each rabbit underwent MRI scanning with a 3.0-T MR Elition scanner (Philips Healthcare, Best, the Netherlands) equipped with a dedicated cranial coil. To monitor the progression of muscle ischemia following arterial ligation, MRI scans were performed at 1-hour intervals for 7 hours after ligation. The MRI protocol included standard T1WI, T2WI with fat suppression (T2WI-FS), amide proton transfer weighted imaging (APTw), a DWI sequence (b=50 and 800 s/mm²), an IVIM sequence (b=0, 5, 10, 20, 50, 80, 100, 150, 200, 500, 800, and 1,000 s/mm²), and a T2 mapping sequence based on a multiecho turbo spin echo (TSE) technique. The main parameters for APTw were as follows: radiofrequency (RF) saturation duration, 2 s; RF saturation power, 2 µT; repetition time/echo time (TR/TE), 9 ms/14,000 ms; flip angle, 90°; slices, 6; total of number of off-resonance frequency points acquired, 9 (at ±2.7 ppm, ±3.5 ppm, ±4.3 ppm, and −1,560 ppm, with 3 images acquired at 3.5 ppm with different echo shift times for B0 field correction); and acquisition time, 4 min 12 s. At 1 hour after ligation, T1-weighted dynamic contrast-enhanced (DCE) imaging was acquired to assess immediate hemodynamic changes and confirm ischemia in the affected limb. The details of the scanning parameters are provided in Table 1.

Table 1. MRI scanning parameters.

Parameter	T1WI	T2WI-FS	T2 map	ADC	IVIM	APT
TR (ms)	540	2,370	2,000	3,000	3,000	14,000
TE (ms)	14	60	13/26/39/52/65/78	64	68	9
Field of view (cm × cm)	12×12	12×12	12×12	12×12	12×12	12×12
Matrix	256×256	256×256	256×256	256×256	256×256	256×256
NEX	4	4	4	4	4	4
b value (s/mm2)				800	0, 5, 10, 20, 50, 80, 100, 150, 200, 500, 800, 1,000
Slice thickness/gap (mm)	3/0	3/0	3/0	3/0	3/0	3/0

ADC, apparent diffusion coefficient; APT, adenosine triphosphate; IVIM, intravoxel incoherent motion; MRI, magnetic resonance imaging; NEX, number of excitations; T2WI-FS, T2-weighted imaging with fat suppression; TE, echo time; TR, repetition time.

Region of interest

Quantitative analysis of signal intensity changes in the relevant MRI sequences was performed using manually defined regions of interest (ROIs). ROIs were meticulously delineated on the axial slice exhibiting the most pronounced ischemic changes within the vastus lateralis muscle. Three ROIs, each measuring 10–15 mm², were carefully drawn, ensuring alignment with the corresponding slice exhibiting the most prominent ischemic manifestation on the T2WI-FS sequence. To minimize confounding factors, ROIs were strategically positioned to avoid nontarget structures such as the femur, major blood vessels, tendons, and fascia. For comparison, three identical ROIs were symmetrically placed in the mirroring muscle regions of the unaffected contralateral limb. T2 relaxation times, apparent diffusion coefficient (ADC) values, APT parameters, and IVIM parameters were recorded for both the ischemic and control sides at hourly intervals from the first to the seventh hour after ischemia. T2 relaxation times, ADC values, and APT parameters were measured directly on a IntelliSpace Portal version 9.0.4.31010 workstation (Philips Healthcare). IVIM sequences underwent postprocessing, and IVIM-derived parameters (f, D, and D*) were quantified via MRIcroGL and MITK software.

The IVIM-derived parameters, including f, D, and D*, were quantified via MITK Diffusion software (22). The IVIM data were modeled with the following biexponential form: $S\left(b\right)=S_0\left(1-f\right)e^{-bD}+f{e}^{-bD\ast }$, where S(b) is the signal intensity with the b value of b, and $S_0$ is the signal value without diffusion weighting (i.e., b=0 s/mm²) (10). The f represents the volume fraction of the microcirculation, D is the true diffusion coefficient of tissue water molecules, and D* is the pseudo-diffusion coefficient of water molecules in the microcirculation blood. The IVIM was analyzed via segmented fitting (23,24), with a threshold b value of 200 s/mm². The estimation of D was obtained through least-squares linear fitting of the image intensity to a monoexponential ADC model with b values greater than or equal to a threshold b value of 200 s/mm². The fitted curve was then extrapolated to obtain an intercept at b=0. The ratio between this intercept and the $S_0$ provided an estimate of f. Finally, the obtained D and f were substituted into the biexponential form and were nonlinearly least-square fitted against all b values to estimate D*.

Histopathology

To assess the histopathological evolution of skeletal muscle ischemia, two rabbits (randomly selected) underwent serial biopsies of the lateral femoral muscle. In the first hour, based on preliminary experimental results, we selected the site on the vastus lateralis muscle most likely to exhibit ischemia for biopsy. From the second to the seventh hour, we conducted an MR-T2WI scan each hour to identify ischemia and determine the most prominent level of localization. Subsequently, a 10 mm × 5 mm × 5 mm tissue sample was harvested from the corresponding site on the ischemic lateral femoral muscle. Samples were immediately immersed in formalin fixative for 24 hours, followed by hematoxylin and eosin (HE) staining and pathological sectioning. A light microscope was employed to evaluate the following characteristics of ischemic changes: muscle fiber morphology; skeletal muscle cell membrane and nuclei integrity; cellular rupture; interstitial edema severity; and the presence, distribution, and extent of inflammatory cell infiltration.

Statistical analyses

Statistical analyses were performed with R version 4.1.3 (The R Foundation for Statistical Computing). All parameter values are presented as the mean ± standard deviation. Repeated measures analysis of variance (ANOVA) was employed to compare parameter values between the ischemic and control sides at each time point (1–7 hours) following skeletal muscle ischemia. Paired t-tests were used for within-group comparisons across time points. Statistical significance was defined as P<0.05. Correlation analyses were conducted to assess the relationship between each parameter and ischemic duration, with correlation coefficients calculated accordingly.

Results

An animal model of acute lower limb ischemia was established through ligation of the right lower limb artery in live rabbits. To confirm the successful construction of the ischemia model, DSA was performed. As shown in Figure 1, DSA imaging revealed that the blood vessels in the non-ligated limb were fully patent, whereas arterial blood flow in the ligated limb was completely interrupted, indicating the successful establishment of the acute lower limb ischemia model. Through this approach, acute lower limb ischemia models were successfully established in 20 rabbits.

Figure 1.

Following the establishment of a skeletal muscle ischemia model in one lower limb of a randomly selected experimental rabbit, DSA was performed for verification. After catheterization of the left femoral artery and administration of contrast medium, the right common iliac artery, along with the internal and external iliac arteries, was not visualized (arrow), whereas the left common iliac artery and its internal and external branches were clearly delineated. DSA, digital subtraction angiography.

Table 2 presents the hourly changes in IVIM, APTw, ADC, and T2WI parameter values for the vastus lateralis muscle on both the ischemic and control sides following induced ischemia.

Table 2. Comparison of parameters measured between the ischemic side of the vastus lateralis muscle and the control side, including IVIM, APT, ADC, and T2 map, with the prolongation of ischemic time.

Quantitative parameter	1 h	2 h	3 h	4 h	5 h	6 h	7 h
f (%)
Ischemia†	6.19±1.13‡	5.60±0.78‡	3.46±0.58‡	3.18±0.69‡	2.65±0.59‡	2.96±0.66‡	2.50±0.64‡
Control	7.19±1.03	7.21±0.94	7.19±1.21	7.05±1.27	7.23±1.04	7.28±2.30	7.02±1.12
D (10−3 mm2/s)
Ischemic†	1.41±0.06‡	1.42±0.08‡	1.41±0.08‡	1.39±0.11‡	1.34±0.07‡	1.38±0.11	1.46±0.08‡
Control	1.2±0.04	1.25±0.09	1.25±0.13	1.29±0.09	1.23±0.05	1.26±0.09	1.25±0.09
D* (10−3 mm2/s)
Ischemia†	59.42±7.67	59.62±9.02	63.28±7.90	62.24±5.20	74.01±5.79‡	59.44±15.77	62.69±9.55
Control	62.65±7.67	60.88±5.53	62.42±4.13	62.58±7.83	61.28±9.31	60.92±4.94	62.31±4.39
APT (%)
Ischemia†	1.78±0.36	2.43±1.09‡	2.05±0.39‡	2.06±0.35‡	2.38±0.64‡	2.13±0.38‡	2.26±0.39‡
Control	1.75±0.53	1.88±0.59	1.75±0.42	1.75±0.33	1.99±0.42	1.68±0.31	1.77±0.33
ADC (10−3 mm2/s)
Ischemia†	1.32±0.06‡	1.36±0.07‡	1.40±0.06‡	1.43±0.09‡	1.48±0.10‡	1.51±0.10‡	1.50±0.09‡
Control	1.23±0.04	1.24±0.06	1.25±0.05	1.25±0.05	1.26±0.04	1.27±0.05	1.25±0.07
T2 value (ms)
Ischemia†	36.96±2.94‡	41.34±3.43‡	45.42±3.79‡	49.03±4.08‡	52.01±5.63‡	55.02±6.66‡	56.66±6.49‡
Control	33.42±0.92	33.57±1.27	33.58±0.96	33.85±0.87	34.04±0.96	33.97±0.55	33.98±1.02

All parameter values are presented as the mean ± standard deviation. ^†, statistical difference between ischemia groups. ^‡, statistical difference between the ischemia group and control group. ADC, apparent diffusion coefficient; APT, amide proton transfer; D, molecular diffusion coefficient; D*, pseudo-diffusion coefficient; f, perfusion fraction; IVIM, intravoxel incoherent motion.

To investigate the changes in the vastus lateralis muscle over a period of 1 to 7 hours after ischemia, multiple MRI sequences were applied to the rabbit models. First, conventional T2WI-FS imaging was performed. The results demonstrated a slight increase in signal intensity in the vastus lateralis muscle on the ischemic side as compared to the control side at 1 hour postischemia (Figure 2). Over time, the ischemic vastus lateralis muscle consistently exhibited high signal intensity, which progressively increased throughout the observation period. In the quantitative analysis, the T2 value of skeletal muscle on the ischemic side increased gradually over 7 hours, while the T2 value of skeletal muscle on the control side remained stable.

Figure 2.

Representative T2WI-FS images of the right vastus lateralis muscle of an experimental rabbit were obtained at different times after arterial ligation (1–7 hours) through the approximate central plane of the ischemic tissue. Ischemic muscle tissue showed a high signal (arrow) on T2WI-FS at 2 hours, and the ischemic high signal (arrow) became clearer with time. Moreover, the T2 value of ischemic muscle tissue continued to increase with time. T2WI-FS, T2-weighted imaging with fat suppression.

DWI results were consistent with those of the T2WI-FS sequence, showing a gradual increase in signal intensity in ischemic skeletal muscle over time (Figure 3). The ADC values of the ischemic muscle increased from (1.32±0.06)×10⁻³ mm²/s at 1 hour to (1.50±0.09)×10⁻³ mm²/s at 7 hours, whereas the ADC values of the skeletal muscle on the control side remained relatively stable, with a range of (1.23–1.25)×10⁻³ mm²/s during the same period. These findings indicate that both T2WI-FS and DWI sequences can effectively detect progressive changes in ischemic skeletal muscle over time.

Figure 3.

Representative DWI images of the right vastus lateralis muscle of experimental rabbits obtained at different times after arterial ligation (1–7 hours) showed that the signal of ischemic muscle tissue (arrow) increased slightly at 2–4 hours and showed clear high signal at 5 hours. Moreover, the ADC of ischemic tissue continued to increase with time. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging.

The APTw sequence was subsequently used to scan the rabbit vastus lateralis muscle. The results demonstrated that the APT signals in the ischemic vastus lateralis muscle were isointense. Quantitative analysis indicated that the APT value of skeletal muscle on the ischemic side was significantly higher than that on the control side as early as 2 hours post-ischemia, with this elevation sustained throughout the 2–7-hour observation period (ischemic: 2.26%±0.39% vs. control: 1.77%±0.33% at the 7th hour; P<0.05) (Figure 4).

Figure 4.

Representative APT images of the right vastus lateralis muscle of experimental rabbits obtained at different times after arterial ligation (1–7 hours). The APT images of ischemic muscle tissue (shown by dotted circles) could not identify ischemic focus at 1–7 hours; quantitative statistics indicated that the APT value increased at 2–7 hours. APT, amide proton transfer.

Following IVIM imaging of ischemic skeletal muscle, changes in f, D, and D* were analyzed (Figure 5). Compared with that of the control side, the f of ischemic skeletal muscle was significantly reduced, with f=6.19%±1.13% at the first hour (control side: f=7.19%±1.03%), and gradually decreased over time. At the seventh hour, there was a significant negative correlation between f (2.50%±0.64%) and the duration of ischemia (r=−0.790). Quantitative analysis further indicated that the D value remained relatively stable from 1 to 7 hours on the ischemic side [first hour: (1.41±0.06)×10⁻³ mm²/s; seventh hour: (1.46±0.08)×10⁻³ mm²/s], with a slight increase compared with that of the control group. D* increased significantly at the fifth hour [D*=(74.01±5.79)×10⁻³ mm²/s; control side: D*=(61.28±9.31)×10⁻³ mm²/s] and then decreased at the sixth hour [D*=(59.44±15.77)×10⁻³ mm²/s].

Figure 5.

Diagram of the statistical trend of the IVIM parameters for the ischemic side vastus of the lateralis muscle and the control side at 1–7 hours after ischemia. The perfusion fraction (f) gradually decreased with the prolongation of ischemia time and was negatively correlated with the ischemia time. D* increased at 5 hours. D, molecular diffusion coefficient; D*, pseudo-diffusion coefficient; f, perfusion fraction; IVIM, intravoxel incoherent motion.

These findings suggest that skeletal muscle ischemia is associated with a time-dependent decline in f. However, the sudden increase in the D* observed at 5–6 hours may reflect qualitative changes in skeletal muscle tissue, potentially indicating the onset of irreversible ischemic damage during this period.

To further validate the abovementioned findings, histological staining of ischemic skeletal muscle tissue was performed at different time points (Figure 6). Cross-sectional analysis revealed slight edema in the muscle fibers of the vastus lateralis muscle 1–2 hours after ischemia. At 3–4 hours, muscle fiber swelling persisted, and by 5–6 hours, significant swelling was observed, along with rupture of the cell membranes. At 7 hours, the muscle fibers exhibited pronounced swelling, loss of cell membrane integrity, and the presence of a small number of inflammatory cells surrounding the fibers. Longitudinal section analysis indicated the appearance of interstitial gaps between muscle bundles and muscle cells 1–2 hours after ischemia. These gaps narrowed hours 3–4, while the partial rupture of muscle fibers was observed at 5–6 hours. By hour 7, the extent of muscle fiber rupture had further increased.

Figure 6.

Representative pathological sections (hematoxylin and eosin staining; ×20) from 1 to 7 hours after ischemia of the vastus lateralis muscle in experimental rabbits. (A) Transverse section. (B) Longitudinal section. At 2 hours, the muscle fibers were slightly endemic, and there were gaps between the muscle bundles and muscle cells. At 4 hours, the degree of muscle fiber swelling increased compared with that previously, and the space around the muscle fibers narrowed. At 6 hours, the muscle fibers were swollen and the cell membrane structure was unclear. The box shows the intracellular fibrous degeneration. At 7 hours, the muscle fibers were swollen and the structure was unclear. The transverse section showed partial cell membrane rupture (arrow), and the longitudinal section showed muscle fiber rupture (arrow).

The histological findings suggest that significant structural changes, including cell membrane rupture and muscle fiber damage, begin to occur in ischemic skeletal muscle 5–6 hours after ischemia. These observations align with the trend of D* changes detected through magnetic resonance IVIM imaging, further indicating the presence of a relationship between ischemia-induced tissue damage and imaging parameters.

Discussion

In this study, magnetic resonance IVIM imaging was employed to quantitatively analyze changes in microcirculation perfusion in skeletal muscles following acute ischemia, which indicated a significant negative correlation between f and ischemic duration (r=−0.790). Although the measured f value represents the f derived from the IVIM model, its interpretation in the context of prolonged ischemia is complex and influenced by both physiological changes and alterations in tissue relaxation properties. The observed decrease in the measured f value with prolonged ischemia was closely aligned with pathological observations, in which with prolonged ischemia, muscle fibers progressively developed edema, intercellular spaces narrowed, and cell membrane structures became blurred or ruptured. These pathological changes indicate severe tissue damage and are associated with reduced capillary blood volume.

The gradual decline in the measured f value during ischemia may be attributed to several factors: (I) arterial ligation reducing capillary blood flow, thus leading to a decrease in water molecules; (II) unobstructed venous return enabling residual blood to drain from the veins and further depleting water molecules within the capillaries; and (III) ischemia impairing the function of the cellular sodium-potassium pump, causing cell edema, which compresses the capillary network and further reduces water molecules in the capillaries. These mechanisms are consistent with the findings reported in previous studies (25,26). Furthermore, the pathological changes, particularly the development of tissue edema, led to an increase in tissue T2 relaxation time over the course of ischemia. Two recent studies by Ma and Wang (27) and Wang (28) have indicate that an increase in T2 values can affect the IVIM signal attenuation characteristics and influence the accuracy of the standard biexponential fitting model, potentially leading to a decrease in the true f value. Therefore, the observed decrease in the measured f value likely reflects the combined effect of reduced capillary blood volume or flow and the influence of increasing tissue T2 on the IVIM measurement.

Suo et al. reported that the IVIM-derived f value increased following cuff-induced ischemia in lower limb skeletal muscles (29), which is in contrast with our results, likely due to differences in the experimental design. In cuff-induced ischemia, venous return is obstructed, leading to an accumulation of water molecules in the capillary network and a relative increase in the f value. However, our experimental design involved arterial ligation without the impairment of venous return, allowing for water molecules to drain from the capillary network; therefore, the changes reflected the microcirculation perfusion under conditions in which arterial blood supply was restricted but venous return unaffected. It is important to note that while collateral circulation may develop over longer periods, the severe acute arterial ligation model employed in our study primarily led to a significant reduction in overall blood flow in the studied timeframe, which was reflected in the trend of the measured f value, with some influence from other factors mentioned above.

In a recent review (30), the mean f value of skeletal muscle in healthy human volunteers at rest is 11.1%±6.7%, with no significant differences between the paraspinal, thigh, and calf muscles. To date, no other animal studies have reported f values in limb skeletal muscles. In our study, the average f value in the ischemic control group was 7.2%. It is widely acknowledged that IVIM parameter estimation is highly sensitive to acquisition settings, particularly the b value distribution and TE. As demonstrated in recent research (31), an extended TE can lead to systematic overestimation of the f value. Although this study focused on hepatic tissue, the underlying principles are broadly applicable and provide valuable insight into the influence of TE on IVIM parameter estimation across different tissue types. The primary objective of our study was to the evaluate relative changes in the f in response to ischemic insult, and we found a significant reduction in f values during the ischemic period. We believe that these changes constitute meaningful information regarding muscle microvascular function and the degree of ischemia. Moreover, these relative measurements are less susceptible to the variability introduced by sequence-specific parameters as compared to absolute baseline values.

Statistically significant differences in D* values were observed exclusively 5 hours after ischemia (P=0.002), with no significant alterations being detected at other time points. The transient increase in D* values at the fifth hour and the subsequent decrease at the sixth hour are an unusual finding. The D* parameter id the perfusion-related diffusion coefficient and reflects the flow velocity of water molecules within the capillary network (11). These findings indicate the absence of substantial changes in microcirculatory perfusion during the initial phases of ischemia (1 to 7 hours), aligning with the anticipated lack of novel blood flow perfusion in capillaries during this period. However, the underlying mechanism for the increase in the D* value at the fifth hour remains unclear.

D* value measurement, especially in the case of low perfusion and certain b-value selection, is challenging and susceptible to large variability. As described in the literature (32), D* and f are generally considered to be positively correlated. This positive correlation means that in the case of ischemia, when f is significantly reduced, the precise estimation and interpretation of D* becomes less reliable. Furthermore, the considerably large standard deviation associated with D* values may be attributed to data instability and susceptibility to signal-to-noise ratio variations, potentially compromising measurement accuracy (32,33). Of note, the observations around 5 hours coincide with our pathological findings, which included degeneration of ischemic skeletal muscle fibers characterized by blurred and progressive rupture of the cell membrane. Therefore, despite the quantitative uncertainty in the D* values at this specific time point, the observed temporal correlation between the fluctuations in IVIM parameters and the onset of cell degradation may indicate that dynamic qualitative changes are occurring in skeletal muscle tissue during this period, which could be associated with the transition to irreversible ischemic damage. However, definitive conclusions based solely on fluctuations in D* values would be limited, as they depend on the reliability of the measurements.

In our study, the D value in the experimental group was significantly different from that of control group at the majority of the time points. This may be attributed to the restricted diffusion of water molecules, potentially induced by muscle edema and cellular membrane compromise resulting from ischemia. In contrast to the findings of Suo et al., who reported a decrease in D values under ischemic conditions, which was interpreted as a reduction in the diffusion rate of water molecules (29), we observed a marginal increase in D values. The exact reason for this discrepancy remains to be fully elucidated and is likely multifactorial. Furthermore, it is important to consider potential interactions and constraints between IVIM parameters during the fitting process. As noted in the literature, including the work by Wang (34), there can be an observed inverse relationship or coupling between f and D, where a decrease in f may facilitate or be associated with an increase in the measured D value (or vice versa). Given that skeletal muscle ischemia profoundly reduces blood flow and thus leads to a significant decrease in f values, this potential interdependence between f and D could contribute to or influence the observed marginal elevation in D values, in addition to the biological factors. The D value represents the diffusion capacity of water molecules within the extravascular space and is modulated by various factors, including muscle cell dimensions and intracellular water content (35,36). Under ischemic conditions, a decline in muscle temperature and the development of cellular edema may lead to the compression of the extracellular space, typically resulting in a decrease in D values.

In contrast to acute brain ischemia, where ADC typically decreases due to restricted diffusion from cytotoxic edema, ischemic injury in skeletal muscle often leads to cellular dysfunction and swelling (edema), which can result in increased water diffusivity. In skeletal muscle, the diffusion of water molecules primarily reflects the intracellular space (35,37), and an increase in D and ADC values is commonly observed in conjunction with muscle cell swelling (38-41). The marginal elevation in D values observed in our study may indicate more intricate cellular and tissue responses during acute ischemia, potentially encompassing factors such as the redistribution of water between intracellular and extracellular compartments. We speculate that the dynamic alterations in water molecule diffusion during ischemia induced by arterial ligation may be more complex than initially anticipated.

Zhang et al. investigated skeletal muscle changes associated with acute ischemia-induced rhabdomyolysis in a rabbit model using T2 mapping, ADC, and diffusion tensor imaging (DTI) (42). They found a positive linear correlation between T2 and ADC values and ischemic duration, which is consistent with the increase in ADC values observed in our study’s ischemic group. Notably, we observed relatively higher T2 and ADC values compared those reported by Zhang et al., potentially due to differences in MRI scanners and acquisition parameters. These findings further suggest that the observed increase in ADC is a characteristic response of skeletal muscle to ischemia, reflecting the underlying edema and altered water dynamics.

It is important to note that that the measurement of ADC values may be affected by the T2 relaxation time. Recent research (28) suggests that T2 can be divided into short T2 time bands (<60 ms), medium T2 time bands (60–80 ms), and long T2 time bands (>80 ms); moreover, in the short T2 time bands, there is a negative correlation between T2 and ADC, while in the long T2 time bands, there is a positive correlation. In our study, the T2 value of skeletal muscle after ischemia gradually increased from 36.96 ms at the first hour to 56.66 ms at the seventh hour and was consistently in a short T2 time band (<60 ms). Based on the theoretical model described above, this may mean that the ADC value should show a downward trend. However, in our study, the ADC values showed an upward trend after ischemia, which is somewhat different from the T2 time bands predicted in the literature. A possible explanation for this is that under the pathophysiological conditions of skeletal muscle ischemia, tissue edema (including intracellular and extracellular edema) may cause complex changes in the diffusion environment of water molecules, thus affecting ADC measurements. For example, cell swelling may enhance intracellular water diffusion, which in turn leads to an increase in ADC values. This effect may partially defy the theoretical expectation of a negative correlation between T2 and ADC in a short T2 time band. In addition, Zhang et al.’s study also suggests that there is a trend of increasing ADC value after skeletal muscle ischemia (42), indicating that the dynamic changes of ADC may be affected by a combination of multiple factors, including, but not limited, to T2 relaxation time, tissue edema, and the redistribution of water molecular components.

Zhang et al. reported that fractional anisotropy (FA) values decreased over time following acute skeletal muscle ischemia (42). DTI specifically examines the directional diffusion of water molecules within muscle tissue, with FA quantifying this anisotropy and reflecting microstructural features such as myofiber size, aspect ratio, and sarcoplasmic reticulum organization (37,43-46). In contrast, we employed the IVIM sequence to investigate skeletal muscle ischemia, with derived IVIM parameters reflecting the degree of water diffusion within and surrounding microcirculation and indirectly indicating ischemia severity.

To our knowledge, this study is the first to report on the application of APTw imaging in the investigation of skeletal muscle ischemia. Consequently, direct comparisons with existing literature, particularly studies on cerebral ischemia in which APT effects have been well-documented, are challenging. Our findings indicate that the expected attenuation of the APT effect was not present in ischemic skeletal muscle, contrasting with the diminished APT effect observed in cerebral ischemic acidosis. This may be due to fundamental distinctions in metabolic pathways and the pH regulation between the two tissues under ischemic conditions. Skeletal muscle has a high glycolytic capacity and significant glycogen reserves. During ischemia, anaerobic glycolysis in skeletal muscle increases significantly, leading to lactate accumulation and a decrease in intracellular pH (47). However, skeletal muscle has considerable tolerance to lactate accumulation, likely due to its efficient adenosine triphosphate regeneration under anaerobic conditions (48). Furthermore, skeletal muscle possesses a strong metabolic buffering capacity, allowing it to effectively modulate the local acid-base status in the early stages of ischemia. Other investigations have demonstrated minimal early-phase decreases in intracellular pH during skeletal muscle ischemia. In a rabbit gastrocnemius complete tourniquet model, Hagberg et al. found that following 4 hours of ischemia, intracellular pH decreased from 7.00±0.03 to 6.60±0.05 (49). This limited pH reduction reflects the strong metabolic buffering capacity of skeletal muscle. Skeletal muscle can modulate local acid-base balance, preventing the inhibition of chemical exchanges via excessive acidification (50). Although pH is a critical determinant of the APT effect and the chemical exchange rate between amide protons and water is highly pH-sensitive, the limited intracellular pH decrease in the early stage of skeletal muscle ischemia, compared to the more pronounced acidosis typically observed in brain ischemia, may explain why the attenuation of the APT signal was not significant. Furthermore, ischemic skeletal muscle undergoes additional pathological alterations, including enhanced membrane permeability and intracellular edema (51,52), which may modify the local microenvironment and potentially influence the APT signal or its sensitivity to pH changes. We hypothesize that the observed enhancement of the APT effect in ischemic muscle results from the combined influence of the relatively buffered initial acidosis and ischemia-induced pathophysiological changes, including cellular edema and membrane dysfunction. The imaging parameters used in this study were originally optimized for tumor imaging, which may partially explain the unexpected APTw results observed in our skeletal muscle ischemia model. Despite these limitations, we believe that retaining this exploratory APTw analysis is valuable, as it represents the first attempt to apply this technique in this context and may provide preliminary insights for future studies. We acknowledge that these explanations are preliminary hypotheses and should be validated in further mechanistic studies, which will be an important direction for our future research.

In the experimental modeling process, the right (or left) common iliac artery, the internal iliac artery, and the contralateral internal iliac artery were ligated while the rabbit’s tail was secured to create a skeletal muscle ischemia model. Vascular DSA imaging, illustrated in Figure 1, confirmed the successful establishment of a lower limb vascular interruption ischemia model. However, due to the extensive vascular network in the lower limbs, this model did not induce ischemic damage in all muscles, as depicted in Figure 2. Complete occlusion of the lower limb vessels was not implemented because this approach consistently resulted in ischemic damage to the vastus lateralis muscle, which is more clinically relevant. Traumatic vascular disruption often leads to segmental occlusion or stenosis due to collateral circulation (53).

This study involved several limitations that should be acknowledged. First, although this study employed a rigorous design and standardized experimental procedures, the limited availability of resources constrained the sample size to 20 New Zealand white rabbits. This relatively small sample size may potentially limit the generalizability and statistical power of our findings. Future research could benefit from increased sample sizes, which would enhance the robustness and applicability of the conclusions. Second, the DWI protocol predominantly employed low b values (b <200 s/mm²) in the IVIM sequence. This methodological choice, while common, has been associated with an increase in the frequency of parameter estimation errors (32). To mitigate this potential source of inaccuracy, we implemented a rigorous ROI selection process. Specifically, ROIs were carefully delineated at the central level of the lateral femoral muscle, with a reasonable distance from the body surface, osseous structures, and major vascular elements being maintained. To enhance measurement reliability, a minimum of three independent measurements were obtained for each ROI. Third, our study did not adequately account for the potential confounding effects of surface temperature variations despite previous research indicating that such variations can significantly influence the measurement of ADC and D values (35). Future studies should incorporate temperature monitoring and control measures to mitigate this potential source of bias and enhance the reliability of the results.

Conclusions

The experimental results demonstrate that the blood f as measured by magnetic resonance IVIM scanning is negatively correlated with the duration of ischemia, supporting its potential as a quantitative indicator for evaluating skeletal muscle ischemic injury. Furthermore, the D* may be used as an indicator to assess whether ischemia causes irreversible damage to muscle fibers. These findings provide a foundation for the application of these for precise assessments of acute lower limb ischemic injury and guidance in developing corresponding clinical treatment strategies.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Experiments were performed under a project license (No. 2023-0445) granted by the Animal Welfare and Ethics Committee of Shanghai Sixth People’s Hospital and in compliance with institutional guidelines for the care and use of animals.

Data Sharing Statement

Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-711/dss

qims-15-09-8064-dss.pdf ^{(157.8KB, pdf)}

DOI: 10.21037/qims-2025-711