¹³N-ammonia positron emission tomography/magnetic resonance imaging in patients with ischemia with non-obstructive coronary arteries: focus on feature tracking and T1 mapping

Introduction

Epicardial coronary artery disease (CAD) has traditionally been viewed as the primary etiology of myocardial ischemia. Nevertheless, more than 50% of patients with stable angina who undergo elective invasive coronary angiography (ICA) show no evidence of obstructive CAD, and this condition is more prevalent among women than in men (1). This group of patients with symptoms suggestive of ischemia with non-obstructive coronary arteries (INOCA), once considered a benign condition, has a high morbidity and an elevated risk of adverse cardiovascular outcomes, including heart failure with preserved ejection fraction (HFpEF), myocardial infarction (MI), and cardiovascular death (2,3). Although the INOCA population is heterogeneous with diverse underlying etiological factors, up to 50–65% of these individuals are diagnosed with coronary microvascular dysfunction (CMVD) (4). The development of CMVD is linked to a variety of pathological mechanisms, including impaired vasodilation of the coronary microvasculature, which results in myocardial ischemia from an inadequate increase in coronary flow from rest to stress (5).

To date, after nearly three decades of clinical and research experience, positron emission tomography (PET) has emerged as one of the most well-validated and accurate non-invasive modalities for the assessment of coronary vascular function by quantifying myocardial flow reserve (MFR) (6). Furthermore, PET provides the most robust data on prognosis in patients with CMVD (7). Nevertheless, the broad clinical adoption of PET imaging remains constrained by its high costs, which stem from the requirement for on-site or proximal cyclotrons as well as expensive generators. Furthermore, PET entails ionizing radiation exposure for patients, rendering it ill-suited for routine serial follow-up.

In recent years, technological innovations in cardiac magnetic resonance (CMR), including native myocardial T1 mapping and CMR feature tracking (CMR-FT), have become valuable imaging tools for the early identification of subtle structural and functional abnormalities across a wide range of cardiac disorders (8). Notably, T1 mapping enables noninvasive evaluation of diffuse myocardial fibrosis and may offer novel insights into myocardial tissue characteristics (9). Similarly, CMR-FT can detect abnormalities in left ventricular (LV) deformation even when systolic functional parameters are normal (10).

However, to date, there remains a paucity of data regarding abnormalities assessed by T1 mapping and CMR-FT in patients with INOCA. Accordingly, this study aimed to determine whether patients with INOCA exhibit subclinical tissue alterations identified by T1 mapping and LV wall motion abnormalities evaluated by CMR-FT, using PET-derived MFR as the reference standard, compared with a cohort of matched healthy controls (HC). We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0509/rc).

Methods

Study population

Baseline data, including demographic characteristics, clinical history, cardiovascular risk factors, and laboratory parameters, were obtained from the Hospital Information System and Laboratory Information Management System. Each of the medical records was independently reviewed by two physicians (R.W. and X.M.) to ensure accuracy and consistency.

INOCA cases

This study prospectively enrolled consecutive participants aged 18–75 years at The First Affiliated Hospital of USTC between August 2024 and February 2026. The inclusion criteria for study enrollment were as follows: (I) clinical symptoms suggestive of myocardial ischemia, including chest pain and/or dyspnea; and (II) no obstructive coronary artery stenosis, defined as <50% luminal narrowing in any epicardial coronary vessel. The exclusion criteria were as follows: (I) coronary artery stenosis ≥50%, determined by ICA or coronary computed tomography angiography (CCTA); (II) known CAD, prior coronary artery bypass grafting (CABG), previous MI, cardiomyopathy, acute coronary syndrome, significant valvular heart disease, pregnancy, or breastfeeding; (III) any type of arrhythmia; and (IV) contraindications to regadenoson [including severe asthma, chronic obstructive pulmonary disease (COPD), or systolic blood pressure (SBP) <90 mmHg] or to CMR imaging [metallic implants or estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m²].

Finally, 44 patients diagnosed with INOCA were enrolled in this study (Figure 1). Based on the PET-derived MFR findings, these patients were further classified into two subgroups, which were as follows: those with INOCA and CMVD (n=28) and those with INOCA but without CMVD (n=16). The diagnostic cutpoint for CMVD was established at a PET-derived MFR value below 2.0.

Figure 1 Flowchart of selection based on the inclusion and exclusion criteria. INOCA, ischemia with non-obstructive coronary disease; MRI, magnetic resonance imaging; PET, positron emission tomography.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of The First Affiliated Hospital of University of Science and Technology of China (No. 2024KYER-232), and written informed consent was provided by all participants.

HC

The HC group (n=22) consisted of healthy cases who were age- and body mass index (BMI)-matched with the INOCA cohort and met the required criteria to serve as reference controls for the imaging study: (I) underwent PET/magnetic resonance imaging (MRI) but did not present with a previous history of cardiac disease or cardiac symptoms or impaired MFR and inadequate image quality; and (II) did not meet any of the exclusion criteria described above.

PET/MRI protocol

All participants underwent an overnight fast of at least 8 hours, with free access to water only. PET/MRI was performed using a hybrid equipment (Biograph mMR; Siemens Healthcare, Erlangen, Germany) with 3-Tesla (3T) MRI. All participants underwent a standardized PET/MRI protocol including cardiac cine imaging, stress/rest ¹³N-ammonia perfusion imaging, and native/post-contrast myocardial T1 mapping. The whole PET/MRI scanning protocol is presented in Figure 2. All participants were also instructed to abstain from caffeine and medications that might affect myocardial perfusion prior to the scheduled examination, as detailed previously. MR-based attenuation correction (MRAC) maps were generated using a four-class segmentation technique (11).

Figure 2 Protocol for PET/MRI imaging. 3D, three-dimensional; Gd, gadolinium; IV, intravenous; MRI, magnetic resonance imaging; PET, positron emission tomography.

Cardiac cine imaging and T1 mapping

All scans were acquired at end-expiration during breath-holding. All cardiac cine imaging acquisitions were performed at rest only. Briefly, after acquiring localizer images, cine imaging was performed using an electrocardiogram (ECG)-gated balanced steady-state free precession (b-SSFP) sequence, with a short-axis stack covering the entire LV and long-axis images (2-, 3-, and 4-chamber views) (12).

Native T1 mapping was performed using the shortened MOLLI (shMOLLI) prototype sequence with a 5[1]1[1] protocol (3 Look-Locker cycles over 9 heartbeats) before intravenous (IV) bolus injection of gadolinium-based contrast agent [gadopentetate dimeglumine (Gd-DTPA)], as described previously (13). The injection protocol was 0.15 mmol/kg of Gd-DTPA (Beilu Pharmaceutical, Beijing, China), followed by a 15-mL saline flush. Post-contrast T1 mapping images were acquired 15–20 minutes after gadolinium injection, during breath-hold at end-diastole. Three T1 map slices were obtained to cover the entire left ventricle. Careful optimization of ECG gating and breath-holding was implemented. During scanning, raw images and T1 error maps were inspected for possible image artifacts, which permitted the immediate repetition of any low-quality measurements (14). Manual epicardial and endocardial contours were delineated with a 10% margin offset to avoid partial volume effects. Global mean T1 values and segmental T1 measurements were then derived from the short-axis images (15). The acquisition parameters are described in Table S1.

PET acquisition

For rest imaging, immediately after IV administration of 370 MBq of ¹³N-ammonia, ECG-gated acquisition was performed for 10 minutes with 16 frames per cardiac cycle using parallel list-mode acquisition. Between rest and stress imaging, a one-hour interval was allowed to permit excretion of gadolinium-based contrast agent and ¹³N-ammonia. To ensure sufficient radiotracer clearance prior to subsequent image acquisition, patients were allowed to step off the scanner table and rest safely throughout this interval.

Stress perfusion imaging used the same parameters as rest imaging. Regadenoson (400 µg) was administered as a single IV bolus over 10 seconds, followed by a 5-mL saline flush immediately prior to PET scanning to ensure maximal coronary vasodilation. Three minutes after the administration of regadenoson, 370 MBq of ¹³N-ammonia was injected intravenously as a bolus. To enable photon attenuation correction, a two-point Dixon sequence was obtained under breath-hold conditions at rest and throughout hyperemic PET data acquisition. Dynamic PET images were reconstructed via three-dimensional ordered-subset expectation-maximization (3D-OSEM) with parameters set to 6 iterations, 21 subsets, and a 5-mm post-reconstruction filter. The dynamic framing was defined as 12×10 s, 2×30 s, 4×60 s, and 1×180 s.

To minimize potential interference from MR-related breath-holding maneuvers on PET quantification, all breath-hold MR sequences were deliberately performed during the later phase of the 10-minute list-mode PET acquisition (approximately 7–10 minutes after ¹³N-ammonia injection). By this time point, the tracer has completed the critical initial phase of blood pool clearance and myocardial uptake, and myocardial activity kinetics have stabilized, reducing the impact of respiratory maneuvers on perfusion quantification. This timing approach is consistent with the protocol design described by Ikeda et al. (16) for simultaneous cardiac PET/MRI. All breath-hold sequences were kept very short: the cine sequence required 8–10 breath-holds of 6–8 heartbeats each (approximately 5–7 seconds per breath-hold), and the post-contrast T1 mapping sequence required 3 breath-holds of 9 heartbeats each (approximately 7–9 seconds per breath-hold). The total cumulative breath-hold time during the 10-minute rest PET acquisition was less than 90 seconds. PET data were acquired in 3D list mode, allowing exclusion of any frames with severe motion artifacts during post-processing. All PET images were systematically reviewed by two experienced nuclear medicine physicians for respiratory motion artifacts, and no significant artifacts were observed that could affect myocardial blood flow quantification.

Image analysis and post-processing

Prior to quantitative analysis, all images were systematically assessed for quality according to predefined criteria. Studies were excluded from further analysis if core imaging sequences were of insufficient quality to allow reliable quantitative measurements.

CMR data analysis

All CMR data were imported into a dedicated post-processing workstation (Circle CVI42 5.11.2; Circle Cardiovascular Imaging, Calgary, AB, Canada) and analyzed in duplicate by two specially trained and experienced CMR physicians blinded to both clinical data and PET findings. Another senior observer with over 5 years of expertise in CMR acquisition and interpretation performed repeated analyses to assess intra- and interobserver variability. Any discrepancies were resolved by consensus following thorough discussion and joint review.

CMR images were considered inadequate if they showed severe respiratory motion artifacts, arrhythmia-related cine distortion, or significant signal inhomogeneity on T1 mapping sequences. According to the American Heart Association (AHA) 17-segment model (excluding the apical cap), native T1 values, post-contrast T1 values, and extracellular volume (ECV) were quantified in the 16 myocardial segments by dedicated post-processing software. Global values for each parameter were calculated using conservative epi- and endocardial contouring, defined as the mean value obtained from 16 segments per patient. Additionally, the cardiac function parameters, including LV ejection fraction (LVEF), end-diastolic volume index (LVEDVI), end-systolic volume index (LVESVI), stroke volume index (LVSVI), mass index (LVMI), cardiac output (LVCO), cardiac index (LVCI), global function index (LVGFI), and remodeling index (LVRI) were calculated from the cine images (17). Venous blood samples were collected from all participants within 24 hours of the PET/MRI examination for the measurement of hematocrit (HCT). ECV was calculated using the following formula: ECV = (1 − HCT) × [Δ(1/T1) myocardial/Δ(1/T1) blood pool] (18).

To comprehensively assess LV global strain parameters, including global radial strain (GRS), global circumferential strain (GCS), and global longitudinal strain (GLS), manual contour tracing was performed across all layers of short-axis cine slices and a single layer of 4- and 2-chamber long-axis cine slices using the CMR-FT module. Three-dimensional (3D) myocardial strain analysis was performed.

For both T1 mapping and CMR-FT analyses, the LV was sections into 16 segments following the AHA recommendations. The apical cap was excluded from analysis, since the analysis was performed on short-axis views. Additionally, slices demonstrating a visible outflow tract were also excluded.

PET data analysis

PET images were considered inadequate if they exhibited major attenuation correction errors, radiotracer extravasation at the injection site, severe respiratory motion artifacts, or incomplete dynamic acquisition. All PET images were evaluated by two nuclear medicine physicians with at least 5 years of experience with PET, using dedicated software (Syngo.via, Siemens, Germany). These physicians were completely blinded to all CMR structural and T1 mapping results throughout the entire analysis process. Myocardial blood flow (MBF) for each vascular territory, both at rest and under stress, was calculated as the average perfusion of the relevant segments. MFR was calculated as stress MBF divided by rest MBF.

The two readers were provided with polar maps and 16 slices covering the short-axis, vertical long-axis, and horizontal long-axis planes for analysis. Image interpretation was conducted using a 17-segment model and a semi-quantitative scoring system (0= normal, 1= mildly abnormal, 2= moderately abnormal, 3= severely abnormal, 4= complete defect) to evaluate perfusion severity. The summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) (SDS = SSS − SRS) were then calculated (19).

Statistical analysis

The normality of continuous variables was tested using the Shapiro-Wilk test. Continuous variables were reported as mean ± standard deviation (SD) for parametric variables and as median with interquartile range (IQR) for nonparametric variables. Categorical variables were reported as frequencies and corresponding percentages. Comparisons of continuous variables across the three groups were performed using the Kruskal-Wallis H test for nonparametric variables and analysis of variance (ANOVA) for parametric variables, whereas categorical variables were compared using the chi-square test. Bonferroni correction was applied to post hoc analysis of Mann-Whitney U test P values. The relationship was examined using Spearman’s correlation coefficient (r). Intraclass correlation coefficients (ICC) were used to assess both intra- and inter-observer variability of the measurements. All statistical analyses were performed using the software SPSS 26.0 (IBM Corp., Armonk, NY, USA). A two-tailed P value below 0.05 was considered statistically significant.

Results

A total of 60 consecutive patients with INOCA were initially screened for study eligibility. Based on the predefined exclusion criteria, 11 patients were excluded, leaving 49 patients for enrollment. Additionally, five patients were excluded due to suboptimal image quality. Of these, three had inadequate CMR image quality (two with severe respiratory motion artifacts affecting cine and T1 mapping sequences, one with failed gadolinium contrast injection), and two had poor PET perfusion image quality due to radiotracer extravasation. The final study cohort comprised 44 patients (median age, 55 years; men, 45%).

Baseline characteristics

Table 1 presents the baseline demographic and clinical characteristics of the final study participants. No significant differences in age, gender, heart rate, BMI, body surface area (BSA), or laboratory data were observed among the three groups (all P>0.05). Resting SBP was significantly higher in INOCA patients with CMVD than those in HC (P=0.002) and INOCA patients without CMVD (P=0.042), whereas the latter two groups did not show significant differences (P>0.05). INOCA patients with CMVD also had a significantly higher resting diastolic blood pressure (DBP) than INOCA patients without CMVD (P=0.018). Additionally, there were no significant differences in cardiovascular risk factors or medication use between INOCA patients with and without CMVD (all P>0.05).

Table 2 summarizes PET/MRI parameters in patients with INOCA and HC. All patients with INOCA presented with symptoms suggestive of myocardial ischemia. For PET functional indices, the SSS, SRS, and SDS were significantly lower in the HC group than those in the INOCA patients with or without CMVD (all P<0.05). Resting MBF levels were significantly higher in INOCA patients with CMVD than those in the INOCA patients without CMVD and those in the HC [1.04 (0.85, 1.40) vs. 0.83 (0.73, 0.97) vs. 0.80 (0.53, 0.94) mL/g/min, respectively; P<0.001]. In addition, global MFR values were significantly lower in INOCA patients with CMVD than those in the INOCA patients without CMVD and those in the HC [1.86 (1.37, 2.00) vs. 2.71 (2.42, 2.82) vs. 3.10 (2.56, 3.56), respectively; P<0.001]. Please refer to Table S2 in the Supplemental Material for absolute P value for each result.

Correlation analysis among GLS, native T1 and MFR

The correlations between CMR-derived parameters and MFR were further investigated. As shown in Figure 3A,3B, the GLS and native T1 values each demonstrated a significant negative correlation with MFR (r=−0.638, P<0.001; r=−0.739, P<0.001, respectively), highlighting their potential value in evaluating global cardiac function. In addition, a robust positive correlation was identified between the native T1 values and GLS (r=0.583, P<0.001) (Figure 3C). ECV was also positively correlated with GLS (r=0.583, P=0.004) (Figure 3D). No significant correlation was observed between LVEF and GLS across the entire study cohort, as presented in Figure S1.

Figure 3 Scatter plots of global MFR versus native T1 (A), global MFR versus GLS (B), native T1 versus GLS (C), and ECV versus GLS (D) of all participants. ECV, extracellular volume; GLS, global longitudinal strain; MFR, myocardial flow reserve.

Comparison of LV function, global strain, and T1 mapping among three groups

For CMR-derived LV global strain parameters, the absolute values of GLS demonstrated a pronounced decline moving from HC to INOCA patients with MFR >2, and further deteriorating in those with MFR <2 (−19.21%±1.84% vs. −17.83%±0.92% vs. −16.25%±1.59%; P<0.001). For GCS, there were no significant differences among the three groups (P=0.481), whereas the INOCA patients with MFR >2 had a higher GRS than the INOCA patient MFR <2 (P=0.043). Conversely, the LV function parameters were not observed this significant decline across the three groups (all P>0.05).

Native T1 values exhibited a gradually increasing trend across the HC, INOCA with MFR >2, and INOCA with MFR <2 groups (1,196±19 vs. 1,213±8 vs. 1,244±20 ms; P<0.001), whereas no such trend was observed for post-contrast T1 or ECV values (Figure 4).

Figure 4 Comparison of PET/MRI parameters among INOCA patients with MFR <2, INOCA with MFR ≥2, or the healthy controls. The three points on the line segment represent the median and the 95% confidence interval. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. ECV, extracellular volume; GCS, global circumferential strain; GLS, global longitudinal strain; GRS, global radial strain; INOCA, ischemia with non-obstructive coronary disease; LV, left ventricular; MBF, myocardial blood flow; MFR, myocardial flow reserve; MRI, magnetic resonance imaging; ns, not significant; PET, positron emission tomography.

Intra- and inter-observer variability

After the 2-week washout period, the intra- and inter-observer variability of CMR-derived parameters was assessed in all participants by two physicians independently. CMR-FT and T1 measurements were shown to be highly reproducible with respect to the intra- and inter-observer variability (Table 3). Two representative cases are shown in Figures 5,6.

Figure 5 Case of a patient with INOCA. A 56-year-old female patient with INOCA. (A-C) The endo- and epicardial contours of the left ventricles are delineated on standard cardiac short-axis, two-chamber, and four-chamber planes in end diastole. (D-F) Global strains and bullseye maps illustrating GRS, GCS, and GLS. (G-I) T1 mapping analysis including native T1, post-contrast T1, and ECV. (J-L) Perfusion maps showed stress MBF, rest MBF, and MFR. ECV, extracellular volume; GCS, global circumferential strain; GLS, global longitudinal strain; GRS, global radial strain; INOCA, ischemia with non-obstructive coronary disease; LAD, left anterior descending artery; LCX, left circumflex artery; MBF, myocardial blood flow; MFR, myocardial flow reserve; RCA, right coronary artery.

Figure 6 Case of a healthy control. A 37-year-old male volunteer. (A-C) The endo- and epicardial contours of the left ventricles are delineated on standard cardiac short-axis, two-chamber, and four-chamber planes in end diastole. (D-F) Global strains and bullseye maps illustrating normal GRS, GCS, and GLS. (G-I) T1 mapping analysis showing normal native T1, post-contrast T1, and ECV. (J-L) Perfusion maps showed stress MBF, rest MBF, and MFR. ECV, extracellular volume; GCS, global circumferential strain; GLS, global longitudinal strain; GRS, global radial strain; LAD, left anterior descending artery; LCX, left circumflex artery; MBF, myocardial blood flow; MFR, myocardial flow reserve; RCA, right coronary artery.

Discussion

The present study using an integrated PET/MRI scanner applied T1 mapping and CMR-FT techniques in a distinct population of patients with signs and symptoms of INOCA in comparison to matched reference controls. Native T1 values exhibited a progressive elevation across the reference control groups, INOCA patients without and with CMVD, whereas the absolute values of GLS showed a gradual reduction in this sequential cohort. Furthermore, a significant positive association was observed between elevated native T1 and GLS. These findings suggest that incorporating CMR into routine clinical assessment may optimize patient monitoring and treatment strategies, ultimately leading to better cardiovascular outcomes for these cohorts. The absolute native T1 values reported here are scanner-, sequence-, and institution-specific, consistent with well-recognized inter-laboratory variability. Our measurements were obtained on a 3T Siemens Biograph mMR system using the shMOLLI sequence, which typically yields higher values than 1.5T scanners or other T1 mapping techniques, explaining differences from prior studies. This study focused on relative differences between groups and correlations with MFR, not universal clinical cutoffs. All analyses used identical protocols within the same cohort, ensuring the internal validity of our findings.

Recently, the development and validation of a novel native T1 mapping sequence (shMOLLI) have enabled high-resolution myocardial T1 mapping during a single breath-hold (20). Building on this, the current study, conducted at 3T to address the limited sensitivity of prior 1.5T approaches, showed excellent agreement with PET (r=−0.739, P<0.001). Recent research has shown that the mean CMVD area obtained from native T1 maps does not differ significantly from that measured on late gadolinium enhancement (LGE) imaging, with high reproducibility between these two imaging types (21). Notably, this study did not compare T1 mapping with PET, the established gold standard, but only with CMVD areas identified by the LGE imaging.

Consistent with our observation that native T1 is elevated compared with HC, Shaw et al. (22) showed that native T1 was significantly higher in INOCA patients than that in the reference group (1,040.1±29.3 vs. 1,003.8±18.5 ms, P<0.001), and there was also a significant inverse association between elevated native T1 and impaired MFR (r=−0.715, P<0.001). Moreover, Shin et al. compared native T1 with LGE imaging to detect microvascular obstruction (MVO) in 20 acute MI patients, thus demonstrating that native T1 was a useful tool for MVO detection (21). In contrast, no statistical relationship between myocardial perfusion reserve and native T1 was observed in the iPOWER study (R²=0.004, P=0.64), which included a distinct patient population and used a single-vendor MRI protocol (23). Beyond methodological differences, this lack of association may be due to significant differences in patient characteristics and risk factors compared with our cohort, such as higher SBP (147 vs. 135 mmHg) and higher smoking prevalence (63% vs. 17%) in the iPOWER participants.

The multifactorial nature of native T1 as an imaging marker, along with the small sample size of this study, does not permit full elucidation of the mechanisms underlying the observed increase in native T1 among INOCA patients. However, measurements of post-contrast T1 and ECV values revealed that in patients with INOCA, those complicated with CMVD exhibit significantly elevated native myocardial T1 values, whereas their ECV showed no statistically significant elevation. Our PET perfusion data provide important context for this observation. The SSS and SRS were significantly higher in INOCA patients with CMVD compared with both HC and INOCA patients without CMVD. However, the absolute values of these scores remained relatively low across all groups, indicating only mild and focal myocardial perfusion abnormalities rather than severe or extensive ischemia. This pattern of mild, intermittent ischemia is characteristic of early-stage CMVD and is more likely to cause transient intracellular myocardial edema rather than persistent extracellular edema or established diffuse fibrosis. Native T1 mapping is highly sensitive to changes in intracellular water content and cellular membrane integrity, which occur early in ischemic injury. In contrast, ECV primarily reflects alterations in the extracellular compartment, which become more prominent only with more severe or chronic myocardial damage. This differential sensitivity to intracellular vs. extracellular changes explains the observed dissociation between native T1 and ECV values in our cohort. These findings suggest that myocardial tissue alterations in CMVD may arise primarily from pathophysiological mechanisms including intracellular myocardial edema and microcirculatory dysfunction, rather than manifesting as diffuse myocardial fibrosis. This aligns with the well-established clinical role of ECV as a key surrogate marker for diffuse interstitial fibrosis (24). To rule out technical factors that might affect ECV measurements, we implemented strict quality control protocols. All HCT samples were collected within 24 hours before the PET/MRI examination, closely matching the timing of contrast agent administration. The shMOLLI sequence included automatic B0 and B1 field homogeneity correction before each scan, and local shimming was performed for every cardiac slice to minimize magnetic field inhomogeneities. All T1 map images were independently reviewed by two experienced radiologists, and any images with artifacts were excluded. Our ECV measurements also showed excellent intra- and inter-observer reproducibility, confirming the reliability of our results. However, correlation analysis found that GLS was correlated negatively with global MFR, but positively with ECV. Consistent with this finding, ECV was identified as an independent predictor of GLS. Similarly, Samuel et al. demonstrated that ECV was significantly positively correlated with early diastolic longitudinal strain rate (r=0.26, P=0.041) in female patients with INOCA (25). Elevated ECV might therefore help in identifying patients with poorer prognosis who might otherwise go undetected by conventional LGE techniques, and correlates with impaired myocardial strain function. In patients with INOCA, a relatively early-stage clinical population, ECV is demonstrated to affect GLS, which in turn reflects MFR.

Further supporting this interpretation, the magnitude of native T1 elevation observed in our INOCA patients with impaired MFR is consistent with the difference in resting native T1 reported by Liu et al. (26) between obstructive CAD patients and controls, which was also explained by microcirculatory autoregulation (27). Future studies using multi-parametric tissue characterization and T1 mapping with heart-rate-independent techniques (28) may offer deeper insights into the mechanisms underlying elevated native T1 and its association with impaired MFR in this population.

In the present study, the absolute values of GLS were observed to exhibit a progressive reduction when comparing the HC group, INOCA patients without CMVD, and INOCA patients with CMVD (−19.21%±1.84% vs. −17.83%±0.92% vs. −16.25%±1.59%; P<0.001) and a statistically significant correlation (r=−0.638, P<0.001) was also observed between GLS and MFR. LV global strain is derived from tracking myocardial motion between the epicardial and endocardial border. It can sensitively detect early LV motion abnormalities (29). The absolute values of GLS were significantly reduced from the HC to INOCA patients without CMVD, and further decreased in those with CMVD, whereas the mean LVEF of all enrolled INOCA patients remained above 50% in the present study. It was indicated that the damage associated with INOCA and CMVD had already been present even in the presence of relatively normal LVEF. An interesting study by Sucato et al. found that GLS was significantly lower in CMVD patients with HFpEF than in the control group (−19.64%±1.91%, P=0.028) (30). Similarly, Yu et al. (31) reported that among 42 CMVD patients and 30 HC, GLS assessed by 3D speckle-tracking imaging (3D-STI) was significantly reduced in the CMVD group (−17.63%±1.50% vs. −20.09%±1.54%, P<0.001). The dysfunction of the microcirculation and the consequent alteration of the MBF probably result in a reduction in myocardial performance mainly in the subendocardial layers. The strain imaging can assess the deformation of longitudinal muscle fibers, which are responsible for 70% of the LV longitudinal systolic function. As a result of the poor hypoxia tolerance of the endocardium, if the coronary artery blood supply is decreased, the endocardial myocardium would be involved, which is associated with longitudinal strain and mainly responsible for the observed significant changes (32). This is an advantage over conventional echocardiographic parameters, which mainly evaluate radial myocardial mechanics but poorly detect the deformation of obliquely arranged subendocardial muscle fibers (33). In this regard, GLS appears to be more sensitive than LVEF in the detection of early systolic function impairments in this patient group. This finding was consistent with the findings of the present study.

Study limitations

The present study has several limitations. First, the native T1 values as well as CMR-FT determined in this study fall within the range of previously reported normal values at 3T (34), and the study was conducted at only a single academic medical center, which may affect generalizability. Second, Gd-DTPA (regionally approved per institutional protocols) may raise safety concerns in regions preferring macrocyclic agents for lower gadolinium retention. Third, the present study did not perform LGE imaging acquisitions in the study population. This was a deliberate protocol design decision focused on minimizing total scan duration and prioritizing non-contrast tissue characterization with native T1 mapping. We acknowledge that the absence of LGE is an important limitation, as LGE would have allowed us to exclude occult MI and detect focal myocardial fibrosis, which could have contributed to the observed abnormalities in native T1 and strain parameters. Fourth, this study used non-invasive PET-derived MFR to define CMVD instead of invasive functional testing including coronary flow reserve (CFR) and index of microcirculatory resistance (IMR) (35). We acknowledge that invasive coronary function testing is the gold standard for CMVD diagnosis as recommended in recent clinical guidelines. However, invasive functional testing was not performed in this study for two main reasons. First, most patients declined additional invasive procedures after diagnostic coronary angiography due to concerns about procedural risks and discomfort. Second, this study was specifically designed as a non-invasive imaging investigation to evaluate the performance of integrated PET/MRI parameters in INOCA patients. PET-derived MFR shows good correlation with invasive CFR measurements and has been widely used in clinical research. Nevertheless, the lack of invasive functional confirmation remains an important limitation of our work. Additionally, acetylcholine provocation testing was not performed in any enrolled patient for the diagnosis of coronary microvascular spasm. This prevented clear distinction between INOCA endotypes. However, our primary endpoint of global MFR provides an integrated measure of overall coronary microvascular function, so the observed associations between CMR parameters and impaired MFR remain valid. Furthermore, we dichotomized our INOCA cohort using an MFR threshold of 2.0, the most widely used cut-off for impaired coronary microvascular function. We recognize that this is an empirically derived value, and optimal thresholds may vary across populations. Notably, no true universal reference standard exists for CMVD. Even invasive tests, although considered the gold standard, have limitations such as procedural risks and operator dependence, a challenge shared by all CMVD research. Additionally, sex-specific effects on T1 values or CMVD mechanisms were not explored, despite the higher prevalence of INOCA among women. Fifth, this study did not perform cardiac MRI perfusion imaging. MRI perfusion has higher spatial resolution than does PET and could provide more detailed segmental information on myocardial perfusion. However, adding MRI perfusion sequences for both rest and stress would have significantly increased the total scan duration. This might have reduced patient compliance and increased the risk of motion artifacts affecting other core imaging sequences. The lack of MRI perfusion data prevents direct head-to-head comparison between PET and MRI perfusion measurements in this cohort. Future studies combining both modalities may offer a more comprehensive assessment of myocardial perfusion in patients with INOCA. Sixth, cardiac cine and post-contrast T1 mapping sequences were acquired simultaneously with rest PET data. Although we implemented measures to minimize the impact of MR-related breath-holding on PET quantification and observed no significant motion artifacts in the final images, this simultaneous acquisition approach remains a potential methodological limitation that could introduce minor variability in myocardial blood flow measurements. Seventh, synchronized blood pressure measurements were not obtained during PET acquisition, preventing calculation of rate-pressure product-corrected myocardial blood flow. Resting MBF is influenced by myocardial oxygen demand, which depends in part on heart rate and SBP. The absence of RPP correction may limit interpretation of absolute resting MBF values. However, our primary endpoint of MFR is a ratio of stress to rest MBF. This ratio reduces the impact of individual hemodynamic variations, so the main study conclusions remain valid.

Conclusions

Using an integrated 3T PET/MRI system, we found that native T1 and GLS show significant correlations with impaired MFR in patients with INOCA. These two parameters are consistent with subclinical LV systolic dysfunction and reduced LV contractile reserve, which are potentially driven by underlying myocardial ischemia. Further studies are warranted to elucidate the mechanisms underlying these association.

Acknowledgments

The authors thank the staff of the Department of Nuclear Medicine, The First Affiliated Hospital of the University of Science and Technology of China for their support of this research.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0509/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0509/dss

Funding: This work was supported by USTC Research Funds of the Double First-Class Initiative (No. YD9110002055 to X.W.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0509/coif). The authors have no conflicts of interest to declare.

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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the ethics committee of The First Affiliated Hospital of University of Science and Technology of China (No. 2024KYER-232), and written informed consent was obtained from all participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Berry C, Camici PG, Crea F, George M, Kaski JC, Ong P, Pepine CJ, Pompa A, Sechtem U, Shimokawa H, Zeitz C, Escaned J, van de Hoef TP, Beltrame JF, Merz CNB. Clinical standards in angina and non-obstructive coronary arteries: A clinician and patient consensus statement. Int J Cardiol 2025;429:133162. [Crossref] [PubMed]
2. Odanović N, Schwann AN, Zhang Z, Kapadia SS, Kunnirickal SJ, Parise H, Tirziu D, Ilic I, Lansky AJ, Pietras CG, Shah SM. Long-term outcomes of ischaemia with no obstructive coronary artery disease (INOCA): a systematic review and meta-analysis. Open Heart 2024;11:e002852. [Crossref] [PubMed]
3. Lee SH, Shin D, Lee JM, van de Hoef TP, Hong D, Choi KH, et al. Clinical Relevance of Ischemia with Nonobstructive Coronary Arteries According to Coronary Microvascular Dysfunction. J Am Heart Assoc 2022;11:e025171. [Crossref] [PubMed]
4. Marinescu MA, Löffler AI, Ouellette M, Smith L, Kramer CM, Bourque JM. Coronary microvascular dysfunction, microvascular angina, and treatment strategies. JACC Cardiovasc Imaging 2015;8:210-20. [Crossref] [PubMed]
5. Collet C, Sakai K, Mizukami T, Ohashi H, Bouisset F, Caglioni S, et al. Vascular Remodeling in Coronary Microvascular Dysfunction. JACC Cardiovasc Imaging 2024;17:1463-76. [Crossref] [PubMed]
6. Zhang X, Li H, Wu P, Yuan L, Wu Z, Liu H, Li S. The diagnosis and prognosis of coronary microvascular disease using PET/CT. Clin Hemorheol Microcirc 2022;80:153-66. [Crossref] [PubMed]
7. Bhandiwad AR, Valenta I, Jain S, Schindler TH. PET-determined prevalence of coronary microvascular dysfunction and different types in a cardio-metabolic risk population. Int J Cardiol Heart Vasc 2023;46:101206. [Crossref] [PubMed]
8. Wang X, Pu J. Recent Advances in Cardiac Magnetic Resonance for Imaging of Acute Myocardial Infarction. Small Methods 2024;8:e2301170. [Crossref] [PubMed]
9. Aquaro GD, Monastero S, Todiere G, Barison A, De Gori C, Grigoratos C, Parisella ML, Faggioni L, Cioni D, Lencioni R, Neri E. Diagnostic Role of Native T1 Mapping Compared to Conventional Magnetic Resonance Techniques in Cardiac Disease in a Real-Life Cohort. Diagnostics (Basel) 2023;13:2461. [Crossref] [PubMed]
10. Saito N, Kato S, Azuma M, Horita N, Utsunomiya D. Prognostic impact of MRI-derived feature tracking myocardial strain in patients with non-ischaemic dilated cardiomyopathy: a systematic review and meta-analysis. Clin Radiol 2024;79:e702-14. [Crossref] [PubMed]
11. Lassen ML, Rasul S, Beitzke D, Stelzmüller ME, Cal-Gonzalez J, Hacker M, Beyer T. Assessment of attenuation correction for myocardial PET imaging using combined PET/MRI. J Nucl Cardiol 2019;26:1107-18. [Crossref] [PubMed]
12. Lassen ML, Beyer T, Berger A, Beitzke D, Rasul S, Büther F, Hacker M, Cal-González J. Data-driven, projection-based respiratory motion compensation of PET data for cardiac PET/CT and PET/MR imaging. J Nucl Cardiol 2020;27:2216-30. [Crossref] [PubMed]
13. Piechnik SK, Ferreira VM, Dall'Armellina E, Cochlin LE, Greiser A, Neubauer S, Robson MD. Shortened Modified Look-Locker Inversion recovery (ShMOLLI) for clinical myocardial T1-mapping at 1.5 and 3 T within a 9 heartbeat breathhold. J Cardiovasc Magn Reson 2010;12:69. [Crossref] [PubMed]
14. Zhang H, Yang W, Zhou D, Wang Y, Zhu L, Jiang M, Zhang Q, Sirajuddin A, Arai AE, Zhao S, Lu M. Prognostic Value of Myocardial T1 Mapping and Extracellular Volume Fraction in Heart Failure: A Meta-Analysis. JACC Cardiovasc Imaging 2025;18:1203-17. [Crossref] [PubMed]
15. Guglielmo M, Fusini L, Muscogiuri G, Baessato F, Loffreno A, Cavaliere A, et al. T1 mapping and cardiac magnetic resonance feature tracking in mitral valve prolapse. Eur Radiol 2021;31:1100-9. [Crossref] [PubMed]
16. Ikeda A, Fukushima K, Katahira M, Ukon N, Yamakuni R, Akama J, Shimizu T, Kiko T, Ishii S, Oikawa M, Ito H, Takeishi Y. Simultaneous measurement of coronary sinus flow and myocardial flow reserve using hybrid (13)N-ammonia PETMR in patients with ischemic heart disease. Int J Cardiovasc Imaging 2025;41:2373-81. [Crossref] [PubMed]
17. Schulz-Menger J, Bluemke DA, Bremerich J, Flamm SD, Fogel MA, Friedrich MG, Kim RJ, von Knobelsdorff-Brenkenhoff F, Kramer CM, Pennell DJ, Plein S, Nagel E. Standardized image interpretation and post-processing in cardiovascular magnetic resonance - 2020 update : Society for Cardiovascular Magnetic Resonance (SCMR): Board of Trustees Task Force on Standardized Post-Processing. J Cardiovasc Magn Reson 2020;22:19. [Crossref] [PubMed]
18. Kato S, Misumi Y, Horita N, Yamamoto K, Utsunomiya D. Clinical Utility of Computed Tomography-Derived Myocardial Extracellular Volume Fraction: A Systematic Review and Meta-Analysis. JACC Cardiovasc Imaging 2024;17:516-28. [Crossref] [PubMed]
19. Angelidis G, Giannakou S, Valotassiou V, Tsougos I, Tzavara C, Psimadas D, Theodorou E, Ziaka A, Ziangas C, Skoularigis J, Triposkiadis F, Georgoulias P. Long-Term Prognostic Value of Automated Measurements in Nuclear Cardiology: Comparisons with Expert Scoring. Medicina (Kaunas) 2023;59:1738. [Crossref] [PubMed]
20. Dall'Armellina E, Piechnik SK, Ferreira VM, Si QL, Robson MD, Francis JM, Cuculi F, Kharbanda RK, Banning AP, Choudhury RP, Karamitsos TD, Neubauer S. Cardiovascular magnetic resonance by non contrast T1-mapping allows assessment of severity of injury in acute myocardial infarction. J Cardiovasc Magn Reson 2012;14:15. [Crossref] [PubMed]
21. Shin JM, Choi EY, Park CH, Han K, Kim TH. Quantitative T1 Mapping for Detecting Microvascular Obstruction in Reperfused Acute Myocardial Infarction: Comparison with Late Gadolinium Enhancement Imaging. Korean J Radiol 2020;21:978-86. [Crossref] [PubMed]
22. Shaw JL, Nelson MD, Wei J, Motwani M, Landes S, Mehta PK, Thomson LEJ, Berman DS, Li D, Bairey Merz CN, Sharif B. Inverse association of MRI-derived native myocardial T1 and perfusion reserve index in women with evidence of ischemia and no obstructive CAD: A pilot study. Int J Cardiol 2018;270:48-53. [Crossref] [PubMed]
23. Mygind ND, Michelsen MM, Pena A, Qayyum AA, Frestad D, Christensen TE, Ghotbi AA, Dose N, Faber R, Vejlstrup N, Hasbak P, Kjaer A, Prescott E, Kastrup J. Coronary microvascular function and myocardial fibrosis in women with angina pectoris and no obstructive coronary artery disease: the iPOWER study. J Cardiovasc Magn Reson 2016;18:76. [Crossref] [PubMed]
24. Clerc OF, Cuddy SAM, Jerosch-Herold M, Benz DC, Katznelson E, Canseco Neri J, Taylor A, Kijewski MF, Bianchi G, Ruberg FL, Di Carli MF, Liao R, Kwong RY, Falk RH, Dorbala S. Myocardial Characteristics, Cardiac Structure, and Cardiac Function in Systemic Light-Chain Amyloidosis. JACC Cardiovasc Imaging 2024;17:1271-86. [Crossref] [PubMed]
25. Samuel TJ, Wei J, Sharif B, Tamarappoo BK, Pattisapu V, Maughan J, Cipher DJ, Suppogu N, Aldiwani H, Thomson LEJ, Shufelt C, Berman DS, Li D, Bairey Merz CN, Nelson MD. Diastolic dysfunction in women with ischemia and no obstructive coronary artery disease: Mechanistic insight from magnetic resonance imaging. Int J Cardiol 2021;331:1-7. [Crossref] [PubMed]
26. Liu A, Wijesurendra RS, Francis JM, Robson MD, Neubauer S, Piechnik SK, Ferreira VM. Adenosine Stress and Rest T1 Mapping Can Differentiate Between Ischemic, Infarcted, Remote, and Normal Myocardium Without the Need for Gadolinium Contrast Agents. JACC Cardiovasc Imaging 2016;9:27-36. [Crossref] [PubMed]
27. Nagel E, Puntmann VO. Is Myocardial Native T1 the One Answer for All? JACC Cardiovasc Imaging 2016;9:37-9. [Crossref] [PubMed]
28. Piechnik SK, Neubauer S, Ferreira VM. State-of-the-art review: stress T1 mapping-technical considerations, pitfalls and emerging clinical applications. MAGMA 2018;31:131-41. [Crossref] [PubMed]
29. Zhou P, Dong Z, Hu X, Yang S, Wang J, Ma X, Tang Y, Xu J, Wei Z, Jia X, Chen X, Liu Y, Xiang X, Huang J, Zhao S. Incremental value of multiparametric cardiac magnetic resonance imaging for non-invasive identification of significant acute cardiac allograft rejection: a prospective and biopsy-proven study. Eur Heart J Cardiovasc Imaging 2025;26:886-94. [Crossref] [PubMed]
30. Sucato V, Galassi AR, Novo S, Saladino A, Evola S, Novo G. Correlation between longitudinal strain analysis and coronary microvascular dysfunction in patients with heart failure with preserved ejection fraction. Microcirculation 2020;27:e12605. [Crossref] [PubMed]
31. Yu Z, Pan H, Cheng Z, Lu K, Hu H. Evaluation of Left Ventricular Systolic Function in Patients with Coronary Microvascular Dysfunction by Three-Dimensional Speckle-Tracking Imaging. Braz J Cardiovasc Surg 2022;37:321-7. [Crossref] [PubMed]
32. Edvardsen T, Helle-Valle T, Smiseth OA. Systolic dysfunction in heart failure with normal ejection fraction: speckle-tracking echocardiography. Prog Cardiovasc Dis 2006;49:207-14. [Crossref] [PubMed]
33. Cadeddu C, Nocco S, Deidda M, Pau F, Colonna P, Mercuro G. Altered transmural contractility in postmenopausal women affected by cardiac syndrome X. J Am Soc Echocardiogr 2014;27:208-14. [Crossref] [PubMed]
34. Teixeira T, Hafyane T, Stikov N, Akdeniz C, Greiser A, Friedrich MG. Comparison of different cardiovascular magnetic resonance sequences for native myocardial T1 mapping at 3T. J Cardiovasc Magn Reson 2016;18:65. [Crossref] [PubMed]
35. Ang DTY, Berry C. What an Interventionalist Needs to Know About INOCA. Interv Cardiol 2021;16:e32. [Crossref] [PubMed]