Dual perfusion imaging with integrated positron emission tomography-magnetic resonance imaging for the detection of coronary microvascular dysfunction in patients with ischemia with no obstructive coronary artery disease: a study on myocardial perfusion imaging and T1 mapping

Introduction

Ischemia with no obstructive coronary artery (INOCA) disease, a chronic disorder with a poor prognosis, represents a challenging health issue (1). Coronary microvascular dysfunction (CMVD) is a common cause of INOCA and is defined by impaired microvascular function and a suboptimal coronary vasodilatory response to exercise or pharmacological stress. It is associated with adverse outcomes comparable to those associated with major cardiac risks and imposes a significant economic burden (2,3). Therefore, accurate methods for the diagnosis of CMVD are needed for optimal patient management.

Historically, diagnosing CMVD required invasive coronary angiography combined with thermodilution or Doppler coronary flow assessment (4). However, after nearly three decades of clinical practice and research efforts, noninvasive quantification of myocardial flow reserve (MFR) with positron emission tomography (PET) has emerged as an alternative reference standard for the assessment of CMVD (5). However, the relatively high cost of PET imaging has limited its widespread clinical application due to the need for onsite or nearby cyclotrons or costly generators. Additionally, PET imaging exposes patients to ionizing radiation, which is undesirable for regular follow-up.

Recently, magnetic resonance imaging (MRI) has been well validated as a modality for evaluating CMVD in patients with INOCA, and its utilization for this purpose is recommended by the relevant guidelines (6). MRI has also been employed to accurately quantify the myocardial perfusion reserve index (MPRI) via first-pass perfusion techniques (7). The calculation of MPRI follows a method similar to that used in PET and has been validated against PET-based measurements (8). However, in addition to the nonlinear relationship between contrast enhancement and its concentration, the main disadvantages of MRI perfusion imaging include its relatively complex postacquisition processing and higher costs. Caution is also warranted in the administration of gadolinium-based agents because the clinical significance of their tissue deposition remains incompletely understood (9). Furthermore, reports suggest that gadolinium can be deposited in the central nervous system, particularly in cases of contrast-enhanced repeated MRI scans and cumulative contrast exposure (10).

T1 mapping without contrast agent administration, referred to as native T1 mapping, is an emerging MRI technique that has demonstrated considerable ability to characterize abnormalities related to acute and chronic myocardial injury, providing novel insights into myocardial tissue properties (11,12). However, despite a rapidly growing body of evidence supporting the clinical utility of this technique, only a few studies have reported the diagnostic value of native T1 mapping in detecting CMVD (13-15). Therefore, the aim of this study was to investigate the diagnostic performance of two MRI techniques, native T1 mapping and MPRI, for the detection of CMVD in patients with INOCA and to compare their diagnostic accuracy with that of the reference standard, PET. We present this article in accordance with the STARD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0413/rc).

Methods

Participants

Consecutive eligible patients with INOCA were prospectively enrolled between August 2024 and February 2026 at The First Affiliated Hospital of University of Science and Technology of China (USTC). The inclusion criteria were as follows: (I) symptoms of myocardial ischemia; and (II) documented nonobstructive coronary artery disease (defined as stenosis <50% on coronary angiography in any coronary artery). Meanwhile, the exclusion criteria were as follows: (I) history of coronary revascularization (including percutaneous coronary intervention or coronary artery bypass grafts); (II) acute coronary syndrome; (III) severe valvular heart disease; (IV) heart failure; (V) cardiomyopathy; (VI) renal dysfunction (estimated glomerular filtration rate <30 mL/min/1.73 m²); (VII) contraindications to MRI (e.g., claustrophobia, pacemaker, and brain aneurysm clips) or to regadenoson; and (VIII) second- or third-degree atrioventricular block, active asthma, seizures, uncontrolled hypertension, pregnancy, breastfeeding, use of caffeine, nicotine, or over-the-counter cold medicines within 12 hours of the PET/MRI examination. A detailed study enrollment flowchart is presented in Figure 1. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and was approved by the Ethics Committee of The First Affiliated Hospital of USTC (approval No. 2024KYER-232). Written informed consent was obtained from all participants.

Figure 1 Flowchart of participants selection based on the inclusion and exclusion criteria. INOCA, ischemia with no obstructive coronary artery disease; MRI, magnetic resonance imaging; PET, positron emission tomography.

Image acquisition

Cardiac PET/MRI protocol

All participants underwent an overnight fast of ≥8 hours but could intake water. Imaging was performed with an integrated whole-body 3-T PET/MRI system (Biograph mMR, Siemens Healthineers, Erlangen, Germany). Imaging protocols included cine imaging, native/post-contrast myocardial T1 mapping, and myocardial perfusion imaging. All participants underwent simultaneous ¹³N-ammonia PET and gadopentetate dimeglumine (Gd-DTPA) perfusion MRI at rest and during regadenoson-induced stress hyperemia. Between rest and stress imaging, a 1-hour interval was provided for clearance of the gadolinium-based contrast agent and ¹³N-ammonia. Patients were permitted to get off the scanner table and rest safely during this period to ensure adequate clearance before subsequent imaging. This protocol is summarized in Figure 2.

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

Cine imaging and T1 mapping

For image acquisition, patients were positioned in a supine position, and heart localization was performed with True Fast Imaging with Steady-State Precession (True FISP; Siemens Healthineers) with electrocardiographic (ECG) gating. Cine images were acquired with a balanced steady-state free precession (b-SSFP) sequence in the short-axis and 2-, 3-, and 4-chamber views of the left ventricle (LV) (16).

At rest, native T1 mapping was performed with the shortened modified Look-Locker inversion recovery (shMOLLI) prototype sequence with a 5[1]1[1] protocol (three Look-Locker cycles over nine heartbeats) before the administration of an intravenous bolus of Gd-DTPA (Beilu Pharmaceutical Co., Ltd., Beijing, China), as described previously (17). This sequence has been shown to exhibit minimal dependence on heart rate and to require a short breath-hold. Postcontrast T1 mapping images were acquired 15–20 minutes after gadolinium injection during a breath-hold at end-diastole.

Myocardial perfusion

For rest PET/MRI perfusion imaging, a 10-minute dynamic PET perfusion scan was initiated simultaneously with the administration of 370 MBq (10 mCi) of ¹³N-ammonia at a flow rate of 0.8 mL/s, which was immediately followed by a 20-mL saline flash (2 mL/s). In addition, MRI first-pass perfusion imaging was performed concurrently with the PET perfusion scan. A single bolus of Gd-DTPA contrast (0.05 mmol/kg of body weight) was administered with a power injector after the start of the PET scan at a flow rate of 5 mL/s, which was followed by a 25-mL saline flush at 5 mL/s through a 20-gauge intravenous cannula in the antecubital vein. Magnetic resonance (MR) perfusion imaging was performed with three short-axis slices, which covered 16 standard myocardial segments per heartbeat, excluding the apex. Seventy ECG-triggered dynamic MRI images per slice of three short-axis myocardial slices (basal, midventricular, and apical) were acquired during an end-expiratory breath-hold with a saturation-recovery turbo fast low-angle shot (TurboFLASH) sequence.

Stress PET/MRI perfusion images were acquired with the same parameters as those of rest imaging. Regadenoson was administered intravenously over 10 seconds, and a second equivalent dose of ¹³N-ammonia was administered 2 minutes after regadenoson administration to ensure maximal vasodilation. A two-point Dixon sequence was acquired during a breath-hold at rest and during stress hyperemic PET scans to correct for PET photon attenuation; this sequence enabled segmentation of fat and water tissues, lung, and air, which served as the basis for generating MRI-based attenuation maps. Dynamic PET images were reconstructed with the three-dimensional ordered-subset expectation-maximization (3D-OSEM) algorithm with six iterations and a 5-mm Gaussian postfilter.

All acquisition parameters are described in Table S1.

Image analysis

Two independent nuclear medicine physicians performed qualitative and quantitative analyses of all PET/MRI images. The physicians were blinded to patient identities, and the order of image presentation was randomized. The physicians reviewed concurrently the two image sets (MRI perfusion imaging and T1 maps) per patient. They were aware that two sets of images originated from the same patient but were unaware of the patients’ PET results. MRI perfusion images and T1 mapping images were analyzed over a period of approximately 30 minutes per patient via dedicated research software (CVI42 version 5.11.2, Circle CVI Inc., Calgary, Canada).

A region of interest (ROI) was manually placed in the central LV blood pool of the image series acquired for the purpose of arterial input function analysis. The papillary muscles were excluded by careful placement. The endocardial and epicardial contours were manually delineated on one phase of each MRI myocardial perfusion slice. Subsequently, these contours were automatically propagated to the other phases. LV blood pool and epicardial fat were carefully excluded during this process. MRI first-pass myocardial perfusion images were analyzed with same postprocessing software to quantify MPRI at the segmental level. For each segment, the MPRI was determined by calculating the ratio of the stress upslope to the remaining upslope of the time-intensity curve (TIC). The myocardial upslope at rest and stress was normalized to the corresponding upslope of the LV blood pool TIC to correct for variations in contrast agent delivery and systemic hemodynamics. The average of the segmental MPRI values was defined as the global MPRI for each patient.

For all participants, native T1, T2, and postcontrast T1 values were analyzed in three ventricular slices from the inline-generated T1 and T2 maps via CVI42 software, with blinding conducted for function data. For global native T1 measurements, endocardial and epicardial borders were carefully segmented to exclude partial volume effects from the LV blood pool and adjacent tissues or fat. Extracellular volume (ECV) fraction maps were generated by combining the shMOLLI images acquired before and 15 minutes after gadolinium administration, as described previously (18). ECV values were calculated with pre- and postcontrast T1 and blood pool values according to the following formula: ECV =(1−hematocrit) × [Δ(1/T1) myocardial/Δ(1/T1) blood flow] (19). Venous blood samples were obtained 24 hours before the PET/MRI examination, and the hematocrit levels of all participants were measured. Global T1, ECV, and T2 values were averaged over the short-axis slices of the whole LV myocardium.

All PET data were analyzed semiautomatically with commercially dedicated software (syngo.via, Siemens Healthineers), with MFR values for the global LV being generated. The LV was automatically outlined, with manual adjustments being applied as needed. Time-activity data from the LV blood pool and myocardial wall served as inputs for the DeGrado compartment model for ¹³N-ammonia imaging, which allowed for the quantification of resting and stress global myocardial perfusion in milliliters per gram per minute (mL/g/min). As previously established, an MFR <2.0 was considered pathologic for the diagnosis of CMVD (20).

LV end-diastolic volume (LVEDV), end-systolic volume (LVESV), stroke volume (LVSV), LV mass (LVM), and ejection fraction (LVEF) were calculated from the cine images. The papillary muscles and trabeculae were included in the LV ventricular volume analyses but excluded from the LV myocardial mass analysis. Subsequently, to facilitate body surface area-adjusted comparisons, these parameters were normalized via the Mosteller formula (21) to derive the. indexed values for LVEDV (LVEDVI), LVESV (LVESVI), LVSV (LVSVI), and LVM (LVMI). Moreover, the LV global function index (LVGFI) was calculated with a previously established formula (22). The LV remodeling index (LVRI) was calculated as follows: the ratio of LVM to LVEDV. The standard 16-segment American Heart Association model was used for myocardial segmentation.

Statistical analysis

Data are reported as the mean ± standard deviation for all the parametric variables and as the median with interquartile range (IQR) for nonparametric variables, with normality determined via Shapiro-Wilk tests. The differences between the groups were evaluated with unpaired t-tests (for continuous normally distributed data) or with chi-squared tests (for categorical variables). The bivariate correlations were assessed via the Pearson correlation coefficient (r). Cohen’s kappa was used to assess interrater agreement for the categorical data, while the McNemar test examined the differences in the paired dichotomous outcomes under related conditions. This kappa value was interpreted based on a classification provided by Landis and Koch (23) as follows: 0.0, poor; 0.0–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1.00, excellent. Intraclass correlation coefficients (ICCs) were used to assess both the intra- and interobserver variability of the measurements. Univariate linear regression analysis was conducted to identify potential predictors of impaired MFR. Variables with a P<0.1 in univariate regression were entered into a multivariate linear regression model. In addition, variance inflation factor (VIF) analysis was performed to check for multicollinearity, and variables with VIF values exceeding 10 were excluded. The remaining variables were then entered into multivariate models to determine the independent predictors of MFR. Diagnostic accuracy was determined according to standard diagnostic metrics, including sensitivity and specificity. Diagnostic performance was evaluated according to the receiver operating characteristic (ROC) curve. The Youden index was applied to identify the optimal cutoff values, and differences between two indices were evaluated via DeLong ROC comparison analysis. To assess threshold stability and generalizability, 1,000 bootstrap resampling was performed for ROC analysis of native T1 and MPRI.

Statistical analyses were conducted with GraphPad Prism v. 9.0.0 (Dotmatics, Boston, NY, USA) and IBM v. 26.0 (IBM Corp., Armonk, NY, USA). All statistical tests were two-tailed, with P <0.05 denoting statistical significance.

Results

Clinical characteristics

The demographic, clinical, and biochemical data of the 66 participants are summarized in Table 1 (28 had CMVD, while 38 had preserved MFR and were placed in the reference group). In terms of conventional cardiovascular risk factors, 22.7% of the participants were diagnosed with diabetes mellitus, 30.3% with hypertension, and 25.8% with dyslipidemia. In unadjusted analyses, resting systolic blood pressure and triglycerides showed nominal between-group differences (P=0.045 and P=0.038, respectively). However, after the Benjamini-Hochberg false-discovery rate adjustment was applied for 49 multiple comparisons, these findings were no longer statistically significant, confirming they are likely chance findings due to the large number of tests performed (Table S2).

PET/MRI measures

The PET/MRI parameters in patients with INOCA are presented in Table 2. No statistically significant differences in LV function parameters derived from MRI were observed between patients with and without CMVD (all P values >0.05). MRI indicated normal LV global systolic function in all 66 participants, with a mean LVEF of 62.0%±6.9%.

As expected, patients with CMVD exhibited a higher median resting myocardial blood flow (MBF; 1.04 mL/g/min, IQR 0.85–1.40 mL/g/min) and therefore a lower median MFR (1.86, IQR 1.37–2.00) compared with reference patients (MBF: median 0.82 mL/g/min, IQR 0.63–0.94 mL/g/min, P<0.001; MFR: median 2.75, IQR 2.52–3.35; P<0.001). Furthermore, native T1 values were significantly different between the CMVD and reference groups (1,239±23 vs. 1,205±17 ms; P<0.001) as were the MPRI values (1.66±0.36 vs. 2.15±0.38; P<0.001), whereas no such differences were observed for postcontrast T1 (P=0.475) or ECV (P=0.357) (Figure 3). Consistent findings were observed in the segmental analysis across basal, midventricular, and apical short-axis slices, the details of which are presented in Figures S1,S2.

Figure 3 Comparison of (A) native T1 and (B) MPRI between the two groups. MFR <2 was defined as positive for CMVD, and MFR >2 as negative. ****, P<0.0001. CMVD, coronary microvascular dysfunction; MPRI, myocardial perfusion reserve index.

Association between MRI parameters and MFR

The bivariate correlations between MRI-derived parameters and impaired MFR were further assessed. At the per-patient level, an impaired MFR was significantly correlated with MPRI (r=0.577; P<0.001) and native T1 (r=−0.513; P<0.001) (Figure 4A,4B). Furthermore, a moderate negative correlation was observed between MPRI and native T1 values (r=−0.538; P<0.001) (Figure 4C). These findings were corroborated by the segmental analysis (Figure S3).

Figure 4 Scatter plots from all participants of (A) global MFR vs. native T1, (B) global MFR vs. MPRI, and (C) native T1 vs. MPRI. MFR, myocardial flow reserve; MPRI, myocardial perfusion reserve index.

The linear regression models for the determinants of MFR are summarized in Table 3. Body mass index (BMI), LV cardiac output, MPRI, and native T1 were all significantly correlated with MFR in the univariate analyses (all P values <0.05), and these parameters were therefore all included in the multivariate model. Multivariate linear regression identified the independent predictors of MFR to be BMI [β=−0.065, 95% confidence interval (CI): −0.128 to −0.003; P=0.040], MPRI (β=0.764, 95% CI: 0.292 to 1.237; P=0.002), and native T1 values (β=−0.010, 95% CI: −0.018 to −0.002; P=0.015).

ROC analyses

Figure 5 depicts the ROC curves for detecting CMVD as defined by ¹³N-ammonia PET. The MPRI threshold of 1.8 (unitless) showed excellent specificity in detecting CMVD, with a sensitivity of 75.0% and a specificity of 81.6%. The native T1 threshold of 1,225 ms showed excellent specificity in detecting CMVD, with a sensitivity of 75.0% and a specificity of 100%. The area under the curve (AUC) calculated for the native T1 of 0.905 was not significantly higher than that for MPRI of 0.829 (P=0.194). In the segmented analysis, the AUC of native T1 (AUC =0.768) was also not significantly different from that of MPRI (AUC =0.773; P=0.923) (Figure S4). Thus, the higher diagnostic accuracy of native T1 compared to that of MPRI in the detection of CMVD did not translate into a significant discriminatory value in terms of the AUC (Table 4). Bootstrap resampling confirmed the stability of the diagnostic thresholds and performance for native T1 and MPRI (Table S3).

Figure 5 Diagnostic performance of MPRI and native T1 in the identification of coronary microvascular dysfunction. AUC, area under the curve; MPRI, myocardial perfusion reserve index.

Comparisons of diagnostic performance relative to reference standard diagnoses

The diagnostic results of native T1 and MPRI are shown in Table 5. Table 6 presents the agreement of the two MRI-derived parameters with the reference standard diagnoses for CMVD detection. The analysis revealed moderate-to-substantial agreement for MPRI (κ=0.566, moderate agreement) and native T1 (κ=0.776, substantial agreement), with statistically significant differences (both P values <0.05). Figure 6 presents a case example of concordance between MRI findings and ¹³N-ammonia PET findings in a patient with CMVD.

Figure 6 A representative case of a symptomatic 59-year-old woman in the CMVD group. (A) MRI first-pass perfusion images (left) at rest and stress and MPRI (right) values in each segment. (B) Native T1 maps (left) and a polar map (right) demonstrating T1 values in each segment. (C) ¹³N-ammonia PET perfusion imaging (left) at rest and stress and MFR (right) values in each segment. CMVD, coronary microvascular dysfunction; MFR, myocardial flow reserve; MPRI, myocardial perfusion reserve index; MRI, magnetic resonance imaging; PET, positron emission tomography.

As shown in Table 7, the reproducibility of the key MRI parameters was statistically excellent for both interobserver and intraobserver assessments (all ICCs >0.7).

Discussion

This study used an integrated PET/MRI scanner and employed native T1 and MPRI in patients with signs and symptoms of INOCA (I) to evaluate whether native T1 is abnormally elevated in patients with CMVD; and (II) to separately examine the association between elevated native T1 values and impaired PET-derived MFR and that between impaired MFR and impaired MPRI values (a surrogate measure of vasodilator-induced ischemia). It was found that the native T1 values in the CMVD group were elevated as compared to those of the reference group. Furthermore, significantly lower MPRI values were observed in the CMVD group than in the reference groups, and a statistically significant inverse association was observed between elevated native T1 values and impaired MFR. Meanwhile, the native T1 performed similarly to MPRI in detecting abnormal MFR as defined by ¹³N-ammonia PET. Additionally, the interrater agreement for MPRI and native T1 measurements was moderate to substantial, indicating that radiologists could consistently and reliably interpret these images. Based on these results, we propose that native T1 mapping could serve as an initial imaging approach for the evaluation of CMVD in the patients with INOCA.

The high-risk cardiac factors for the patient group in our study, with a notably high prevalence of hypertension (30.3%), are consistent with previous large-sample studies of patients with INOCA (24) and, along with aging, are recognized as common risk factors driving the pathophysiology in this population (25,26). Notably, research has shown that hypertension is associated with diffuse remodeling of the intramyocardial microvasculature (27), a process that can occur in patients with hypertension and microvascular angina even without LV hypertrophy (28). Importantly, CMVD acts as an early precursor to adverse cardiovascular events. Studies have demonstrated that this dysfunction is associated with a 2.5% annual rate of major adverse events, including cardiovascular mortality, nonfatal myocardial infarction, nonfatal stroke, and heart failure, even among the individuals without epicardial coronary artery stenosis (29). Thus, the early detection of CMVD may be help facilitate prognostic assessment and patient stratification to optimize therapy (30).

Native T1 mapping quantifies myocardial tissue properties without contrast, allowing for the early detection of microvascular injury (31). Parametric mapping may represent the first sign of myocardial injury and often precedes symptoms, changes in ejection fraction, and irreversible myocardial remodeling (12). Among these, native T1 mapping quantifies myocardial intracellular and extracellular compartments without the need for intravenous contrast agents, making this examination accessible and safe for patients with chronic kidney disease (13). Native T1 mapping quantifies the myocardial longitudinal relaxation time. In the native myocardium, native T1 prolongation occurs in response to an increased free water content or interstitial accumulation of macromolecules (e.g., diffuse fibrosis and amyloid deposition). Conversely, native T1 shortening results from intracellular or extracellular iron accumulation (e.g., primary or secondary hemochromatosis) or intracellular lipid deposition (e.g., Anderson-Fabry disease).

Consistent with findings of elevated native T1 values in the CMVD compared with those of the reference group (1,239±23 vs. 1,205±17 ms; P<0.001), Shaw et al. reported that the native T1 value was significantly elevated in patients with INOCA compared with controls (1,040.1±29.3 vs. 1,003.8±18.5 ms; P<0.001) (13); they further found a significant inverse association between increased native T1 values and a reduced MPRI (r=−0.481; P=0.004), which was considered a surrogate measure of ischemia severity in this cohort (32). Moreover, Shin et al. compared native T1 mapping with late gadolinium enhancement (LGE) imaging for detecting microvascular obstruction (MVO) in 20 patients with acute myocardial infarction, reporting that native T1 mapping was a useful tool for MVO detection (11). Notably, they did not compare native T1 mapping to ¹³N-ammonia PET but only to the CMVD areas identified by LGE imaging. Additionally, the iPOWER study, which used a different patient population and a single-vendor MRI protocol, failed to observe a statistical relationship between the myocardial perfusion reserve and native T1 values (r²=0.004; P=0.64) (33). Beyond methodological discrepancies, the lack of association may be explained by significant differences in the patient characteristics and risk factors between the iPOWER participants and our cohort, such as the higher systolic blood pressure (147 vs. 135 mmHg) and higher smoking prevalence (63% vs. 21.2%).

In this study, the differences in the kappa coefficients for the two imaging parameters were significant, with the value for the native T1 (77.6%) being the highest, followed by those for MPRI (56.6%). Kappa coefficients represent the agreement between classifications, with ranges of 0.0–0.20, 0.21–0.40, 0.41–0.60, and 0.61–0.80 indicating slight, fair, moderate, and substantial agreement, respectively. Therefore, the agreement for the native T1 was superior to that of the MPRI. These results were consistent with those from our ROC analyses. Although the differences were not significant, the AUC for native T1 was 0.905, which was higher than that for the MPRI (0.829). Thus, the consistency of native T1 results was optimal compared to that of the MPRI.

The multifactorial nature of native T1 as a noninvasive imaging marker and the small sample size in this study do not allow for fully elucidation of the mechanisms underlying the observed increase in native T1 values in patients with INOCA. However, measurements of postcontrast T1 and ECV values revealed that in patients with INOCA, those with CMVD exhibited significantly elevated native T1 values, while their ECV had no statistically significant increase. This discrepancy is attributed to the distinct pathophysiological information reflected by native T1 and ECV. Native T1 mapping is sensitive to subtle changes in myocardial free water content. Early microcirculatory disturbance in CMVD can lead to elevated myocardial water and prolonged native T1 values. ECV is employed to specifically assess the degree of diffuse myocardial fibrosis. The unchanged ECV indicates that significant interstitial fibrosis was not detected in the study population. In addition, the divergent statistical results can be further explained by the difference in measurement variability between the two parameters. Native T1 mapping is acquired via a single standardized measurement with excellent intra- and interobserver reproducibility, as confirmed by the high ICCs in our study. The narrow standard deviation of native T1 values reflects low within-group variability, and thus the observed between-group difference is unlikely to be due to measurement error. In contrast, ECV calculation requires multiple independent measurements, including native and postcontrast T1 mapping and hematocrit testing, which introduce additional sources of random error. The higher within-group variability of ECV resulted in a lack of statistical significance for the between-group difference, despite a relative change comparable to that of native T1. This finding suggests that myocardial tissue changes associated with CMVD may be primarily related to key pathophysiological processes such as myocardial edema and microcirculatory disturbances and not to diffuse myocardial fibrosis, which is consistent with ECV being clinically significant as an important surrogate marker for diffuse fibrosis (34). This assumption is further supported by the similarity between our native T1 value differences (CMVD patients vs. controls) and those reported by Liu et al. (35); in their study, resting native T1 values in patients with obstructive CAD were compared with those of controls at 1.5 T, and the differences were attributed to microcirculatory autoregulation. Future investigations incorporating multiparametric tissue characterization, including native T1 mapping under heart-rate-independent techniques (12), may help to clarify the underlying mechanisms contributing to the increased native T1 values and their negative association with MFR in this population.

The primary challenge in MRI mapping development is the absence of standardized protocols for T1 and T2 mapping reference sequences, which has led to significant vendor-specific variability. Furthermore, establishing local normal ranges in line with the current consensus is resource-intensive for MRI centers. It should be noted that significantly altering b-SSFP readout parameters may introduce bias into T1 measurements due to a dependence on T2 relaxation, resonance frequency offsets, or magnetization transfer effects. Caution is therefore warranted in the comparison of T1 values obtained with different techniques or parameter sets. Additionally, implementing these techniques into routine clinical practice requires specialized training for healthcare providers and the validation of protocols across multiple institutions (31).

Study limitations

Several limitations should be noted in relation to the interpretation of our findings. To begin, the use of Gd-DTPA as a contrast agent, while regionally approved and compliant with institutional protocols, may raise safety concerns in certain regions where macrocyclic agents are preferred due to lower gadolinium retention risks. Additionally, the small sample size limits the statistical power for subgroup analyses and robust cutoff validation, and an a priori sample size calculation was not performed for this comparison. Furthermore, the diagnostic cutoff values for native T1 mapping and MPRI were derived under specific conditions: 3-T MRI and regadenoson-induced stress. These thresholds may not generalize to 1.5-T systems or adenosine-induced stress protocols, underscoring the need for site-specific validation. Moreover, it should be noted that our study evaluated only one type of INOCA, thereby limiting the generalizability of our findings to other INOCA phenotypes. Another concern is that our imaging protocol used a clinically available perfusion MR technique, which, similarly to previous studies (13,36), restricted our analysis to the semiquantitative assessment of perfusion reserve (i.e., MPRI). Future studies that incorporate absolute quantification of MBF at rest and hyperemic peak may further clarify the mechanisms underlying the impaired perfusion reserve and its association with elevated native T1. Finally, INOCA is more prevalent in women, and sex differences may influence native T1 values or CMVD-related mechanisms. However, our study did not systematically investigate the sex-specific effects on myocardial T1 mapping outcomes.

Conclusions

Our findings demonstrate a significant inverse association between elevated native T1 values and PET-derived MFR. Furthermore, the native T1 mapping offered consistent and complementary diagnostic value with MPRI in detecting CMVD as defined by ¹³N-ammonia PET. However, the underlying mechanisms of this association remain to be further investigated. If validated in larger studies, this inverse relationship may help establish native T1 mapping as a gadolinium-free surrogate marker of CMVD severity in patients with INOCA, allowing for serial imaging-based therapy evaluations that are free of gadolinium-based contrast agents.

Acknowledgments

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

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0413/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0413/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-2026-1-0413/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. Jenkins K, Pompei G, Ganzorig N, Brown S, Beltrame J, Kunadian V. Vasospastic angina: a review on diagnostic approach and management. Ther Adv Cardiovasc Dis 2024;18:17539447241230400. [Crossref] [PubMed]
2. Chen C, Wei J, AlBadri A, Zarrini P, Bairey Merz CN. Coronary Microvascular Dysfunction - Epidemiology, Pathogenesis, Prognosis, Diagnosis, Risk Factors and Therapy. Circ J 2016;81:3-11. [Crossref] [PubMed]
3. Schumann CL, Mathew RC, Dean JL, Yang Y, Balfour PC Jr, Shaw PW, Robinson AA, Salerno M, Kramer CM, Bourque JM. Functional and Economic Impact of INOCA and Influence of Coronary Microvascular Dysfunction. JACC Cardiovasc Imaging 2021;14:1369-79. [Crossref] [PubMed]
4. Thakker RA, Rodriguez Lozano J, Rodriguez Lozano P, Motiwala A, Rangasetty U, Khalife W, Chatila K. Coronary Microvascular Disease. Cardiol Ther 2022;11:23-31. [Crossref] [PubMed]
5. Al-Mallah MH, Nayfeh M, Alrifai M. The role of cardiac PET in diagnosis and prognosis of patients with ischemia with no obstructive coronary arteries (INOCA). Am Heart J Plus 2024;43:100399. [Crossref] [PubMed]
6. Vrints C, Andreotti F, Koskinas KC, Rossello X, Adamo M, Ainslie J, et al. 2024 ESC Guidelines for the management of chronic coronary syndromes. Eur Heart J 2024;45:3415-537. [Crossref] [PubMed]
7. Nelson MD, Gomez-Arnold JM, Wei J, Lauzon M, Zamani SK, Maughan J, Obrutu O, Shufelt C, Handberg E, Pepine C, Bairey Merz CN. Contributors to high left ventricular ejection fraction in women with ischemia and no obstructive coronary artery disease: Results from the Women's Ischemia Syndrome Evaluation-Coronary Vascular Dysfunction (WISE-CVD) Study. Am Heart J 2024;278:41-7. [Crossref] [PubMed]
8. Kotecha T, Martinez-Naharro A, Boldrini M, Knight D, Hawkins P, Kalra S, Patel D, Coghlan G, Moon J, Plein S, Lockie T, Rakhit R, Patel N, Xue H, Kellman P, Fontana M. Automated Pixel-Wise Quantitative Myocardial Perfusion Mapping by CMR to Detect Obstructive Coronary Artery Disease and Coronary Microvascular Dysfunction: Validation Against Invasive Coronary Physiology. JACC Cardiovasc Imaging 2019;12:1958-69. [Crossref] [PubMed]
9. Bonner BP, Yurista SR, Coll-Font J, Chen S, Eder RA, Foster AN, Nguyen KD, Caravan P, Gale EM, Nguyen C. Contrast-Enhanced Cardiac Magnetic Resonance Imaging With a Manganese-Based Alternative to Gadolinium for Tissue Characterization of Acute Myocardial Infarction. J Am Heart Assoc 2023;12:e026923. [Crossref] [PubMed]
10. Vymazal J, Rulseh AM. MRI contrast agents and retention in the brain: review of contemporary knowledge and recommendations to the future. Insights Imaging 2024;15:179. [Crossref] [PubMed]
11. 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]
12. Carrabba N, Amico MA, Guaricci AI, Carella MC, Maestrini V, Monosilio S, Pedrotti P, Ricci F, Monti L, Figliozzi S, Torlasco C, Barison A, Baggiano A, Scatteia A, Pontone G, Dellegrottaglie SWorking Group on Cardiac Magnetic Resonance of the Italian Society of Cardiology. CMR Mapping: The 4th-Era Revolution in Cardiac Imaging. J Clin Med 2024;13:337. [Crossref] [PubMed]
13. 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]
14. Burrage MK, Shanmuganathan M, Masi A, Hann E, Zhang Q, Popescu IA, Soundarajan R, Leal Pelado J, Chow K, Neubauer S, Piechnik SK, Ferreira VM. Cardiovascular magnetic resonance stress and rest T1-mapping using regadenoson for detection of ischemic heart disease compared to healthy controls. Int J Cardiol 2021;333:239-45. [Crossref] [PubMed]
15. Yimcharoen S, Zhang S, Kaolawanich Y, Tanapibunpon P, Krittayaphong R. Clinical assessment of adenosine stress and rest cardiac magnetic resonance T1 mapping for detecting ischemic and infarcted myocardium. Sci Rep 2020;10:14727. [Crossref] [PubMed]
16. Moscatelli S, Pozza A, Leo I, Ielapi J, Scatteia A, Piana S, Cavaliere A, Reffo E, Di Salvo G. Importance of Cardiovascular Magnetic Resonance Applied to Congenital Heart Diseases in Pediatric Age: A Narrative Review. Children (Basel) 2024;11:878. [Crossref] [PubMed]
17. Klaus JB, Goerke U, Klarhöfer M, Keerthivasan MB, Jung B, Berzigotti A, Ebner L, Roos J, Christe A, Obmann VC, Huber AT. MRI Dixon Fat-Corrected Look-Locker T1 Mapping for Quantification of Liver Fibrosis and Inflammation-A Comparison With the Non-Fat-Corrected Shortened Modified Look-Locker Inversion Recovery Technique. Invest Radiol 2024;59:754-60. [Crossref] [PubMed]
18. Messroghli DR, Moon JC, Ferreira VM, Grosse-Wortmann L, He T, Kellman P, Mascherbauer J, Nezafat R, Salerno M, Schelbert EB, Taylor AJ, Thompson R, Ugander M, van Heeswijk RB, Friedrich MG. Clinical recommendations for cardiovascular magnetic resonance mapping of T1, T2, T2* and extracellular volume: A consensus statement by the Society for Cardiovascular Magnetic Resonance (SCMR) endorsed by the European Association for Cardiovascular Imaging (EACVI). J Cardiovasc Magn Reson 2017;19:75. [Crossref] [PubMed]
19. Han D, Lin A, Kuronuma K, Gransar H, Dey D, Friedman JD, Berman DS, Tamarappoo BK. Cardiac Computed Tomography for Quantification of Myocardial Extracellular Volume Fraction: A Systematic Review and Meta-Analysis. JACC Cardiovasc Imaging 2023;16:1306-17. [Crossref] [PubMed]
20. Schindler TH, Fearon WF, Pelletier-Galarneau M, Ambrosio G, Sechtem U, Ruddy TD, et al. Myocardial Perfusion PET for the Detection and Reporting of Coronary Microvascular Dysfunction: A JACC: Cardiovascular Imaging Expert Panel Statement. JACC Cardiovasc Imaging 2023;16:536-48. [Crossref] [PubMed]
21. Mosteller RD. Simplified calculation of body-surface area. N Engl J Med 1987;317:1098. [Crossref] [PubMed]
22. Mewton N, Opdahl A, Choi EY, Almeida AL, Kawel N, Wu CO, Burke GL, Liu S, Liu K, Bluemke DA, Lima JA. Left ventricular global function index by magnetic resonance imaging--a novel marker for assessment of cardiac performance for the prediction of cardiovascular events: the multi-ethnic study of atherosclerosis. Hypertension 2013;61:770-8. [Crossref] [PubMed]
23. Svanholm H, Starklint H, Gundersen HJ, Fabricius J, Barlebo H, Olsen S. Reproducibility of histomorphologic diagnoses with special reference to the kappa statistic. APMIS 1989;97:689-98. [Crossref] [PubMed]
24. Sara JD, Widmer RJ, Matsuzawa Y, Lennon RJ, Lerman LO, Lerman A. Prevalence of Coronary Microvascular Dysfunction Among Patients With Chest Pain and Nonobstructive Coronary Artery Disease. JACC Cardiovasc Interv 2015;8:1445-53. [Crossref] [PubMed]
25. Camici PG, d'Amati G, Rimoldi O. Coronary microvascular dysfunction: mechanisms and functional assessment. Nat Rev Cardiol 2015;12:48-62. [Crossref] [PubMed]
26. Yang Z, Liu Y, Li Z, Feng S, Lin S, Ge Z, Fan Y, Wang Y, Wang X, Mao J. Coronary microvascular dysfunction and cardiovascular disease: Pathogenesis, associations and treatment strategies. Biomed Pharmacother 2023;164:115011. [Crossref] [PubMed]
27. Schwartzkopff B, Motz W, Frenzel H, Vogt M, Knauer S, Strauer BE. Structural and functional alterations of the intramyocardial coronary arterioles in patients with arterial hypertension. Circulation 1993;88:993-1003. [Crossref] [PubMed]
28. Brush JE Jr, Cannon RO 3rd, Schenke WH, Bonow RO, Leon MB, Maron BJ, Epstein SE. Angina due to coronary microvascular disease in hypertensive patients without left ventricular hypertrophy. N Engl J Med 1988;319:1302-7. [Crossref] [PubMed]
29. Rubinshtein R, Yang EH, Rihal CS, Prasad A, Lennon RJ, Best PJ, Lerman LO, Lerman A. Coronary microcirculatory vasodilator function in relation to risk factors among patients without obstructive coronary disease and low to intermediate Framingham score. Eur Heart J 2010;31:936-42. [Crossref] [PubMed]
30. Lanza GA, Crea F. Primary coronary microvascular dysfunction: clinical presentation, pathophysiology, and management. Circulation 2010;121:2317-25. [Crossref] [PubMed]
31. Forleo C, Carella MC, Basile P, Mandunzio D, Greco G, Napoli G, Carulli E, Dicorato MM, Dentamaro I, Santobuono VE, Memeo R, Latorre MD, Baggiano A, Mushtaq S, Ciccone MM, Pontone G, Guaricci AI. The Role of Magnetic Resonance Imaging in Cardiomyopathies in the Light of New Guidelines: A Focus on Tissue Mapping. J Clin Med 2024;13:2621. [Crossref] [PubMed]
32. Galea N, Marchitelli L, Pambianchi G, Catapano F, Cundari G, Birtolo LI, Maestrini V, Mancone M, Fedele F, Catalano C, Francone M. T2-mapping increase is the prevalent imaging biomarker of myocardial involvement in active COVID-19: a Cardiovascular Magnetic Resonance study. J Cardiovasc Magn Reson 2021;23:68. [Crossref] [PubMed]
33. 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. steering committee of the iPower study. 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]
34. 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]
35. 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]
36. Bairey Merz CN, Handberg EM, Shufelt CL, Mehta PK, Minissian MB, Wei J, Thomson LE, Berman DS, Shaw LJ, Petersen JW, Brown GH, Anderson RD, Shuster JJ, Cook-Wiens G, Rogatko A, Pepine CJ. A randomized, placebo-controlled trial of late Na current inhibition (ranolazine) in coronary microvascular dysfunction (CMD): impact on angina and myocardial perfusion reserve. Eur Heart J 2016;37:1504-13. [Crossref] [PubMed]