Native myocardial T1 reference values in a tunisian cohort using MOLLI and SMART1Map Authors: Mariem Dali, Narjes Benameur, Mohamed Deriche, Hassen Hafsi, Seif Boukriba, Wassim Frikha, Younes Arous, Salam Labidi Published: 2026 ABSTRACT Native T1 values in cardiac magnetic resonance (CMR) are influenced by acquisition sequence, physiological factors, and site-specific variability, complicating clinical interpretation. The absence of population-specific reference values further limits diagnostic accuracy. This study aimed to establish native T1 reference ranges in a Tunisian population, evaluate sex- and age-related differences using MOLLI 5(3)3 and SMART1Map sequences, and assess inter-site variability of MOLLI measurements across two centers using identical scanners and protocols. Ninety-four healthy volunteers were enrolled: 64 subjects (24 females, 40 males; 5–77 years) were scanned at Site 1 to establish reference values, and 30 subjects (20 females, 10 males; 19–62 years), frequency-matched by age to a subset of Site 1, were scanned at Site 2 to evaluate site-related variability. MOLLI-derived T1 values showed a significant negative association with age (β = –0.84 ms/year, p = 0.012), whereas SMART1Map-derived values were not associated with age (β = 0.30 ms/year, p = 0.556). SMART1Map consistently yielded higher T1 values than MOLLI (global mean ± SD: 1184 ± 70 ms vs. 1049 ± 55 ms; p < 0.001). Sex-related differences were minimal, with only a small apical difference observed for MOLLI (p = 0.023). Multi-center MOLLI measurements showed similar T1 distributions across sites, with minor differences observed at the mid-ventricular level. These findings provide the first Tunisian-specific native myocardial T1 reference values at 1.5 T, highlighting sequence-dependent differences, age-related effects for MOLLI, minimal sex-related influence, and overall inter-site consistency. These findings support the use of locally derived, sequence-specific reference ranges to improve clinical interpretation.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-47324-w. Introduction Cardiac Magnetic Resonance (CMR) enables comprehensive assessment of myocardial structure, function, and tissue characterization through a variety of imaging sequences. However, consistent and objective quantification of tissue properties remains challenging due to reliance on relative signal intensities and variability in contrast enhancement patterns1–3. The 2018 revision of the Lake Louise Criteria (LLC) incorporated parametric mapping techniques, particularly native T1 mapping, as standardized diagnostic markers for myocardial inflammation, reinforcing their growing clinical importance and diagnostic robustness4. Native T1 mapping quantifies the longitudinal relaxation time (ms) on a pixel-by-pixel basis, providing a quantitative surrogate of myocardial tissue composition. This technique employs either inversion-recovery (IR) or saturation-recovery (SR) strategies, which sample longitudinal magnetization recovery at defined time points to generate a fitted T1 relaxation curve5. The choice of acquisition strategy directly impacts measurement accuracy, precision, and susceptibility to physiological influences. Among IR-based techniques, the Modified Look-Locker Inversion Recovery (MOLLI) sequence, introduced in 2004, has become one of the most widely adopted methods for myocardial T1 quantification6. MOLLI schemes are described using notations such as 5(3)3, which indicate the number of images acquired after each inversion pulse and the number of recovery heartbeats between inversion blocks. In the 5(3)3 scheme, five images are acquired following the first inversion pulse, followed by a three-heartbeat recovery period, and then three additional images after a second inversion pulse. This ECG-gated design determines the temporal sampling of magnetization recovery and influences fitting accuracy, heart-rate sensitivity, and breath-hold duration. Early MOLLI protocols, such as 5(3)3(3)3, provided dense sampling of the T1 recovery curve but required relatively long breath-holds. To improve acquisition efficiency, refinements such as the 5(3)3 scheme were introduced, reducing breath-hold duration while preserving adequate sampling7–9. Additional MOLLI variants, such as 2(2)4 and 4(1)3(1)2, were specifically optimized for post-contrast T1 mapping and extracellular volume (ECV) quantification10,11. In contrast, the Saturation Method using Adaptive Recovery Times for cardiac T1 mapping (SMART1Map) employs a SR approach, acquiring images at predefined saturation delay times across separate heartbeats. Unlike MOLLI, which relies on IR sampling followed by Look-Locker correction, SMART1Map directly models signal recovery using a SR exponential function. These fundamentally different sampling and modeling strategies produce systematic differences in measured T1 values and distinct sensitivities to heart rate (HR), inversion efficiency, and sequence timing. Consequently, sequence-specific reference ranges and cross-sequence validation are essential when interpreting or establishing normative values12,13. Despite its expanding clinical integration and proven diagnostic utility, native T1 mapping remains affected by multiple sources of variability, including acquisition protocol, scanner hardware, field strength, post-processing software, and physiological characteristics such as age and sex. As a result, substantial inter-study variability persists, complicating clinical interpretation and limiting the direct transferability of published reference values. Importantly, most available reference ranges are derived from non-local populations, limiting their direct applicability to specific ethnic or regional cohorts. To date, Tunisian-specific native myocardial T1 reference values have not been reported. The absence of population-adapted thresholds may hinder accurate diagnostic interpretation and risk misclassification in routine practice. Beyond sequence-related variability, population-specific factors also play a critical role. Establishing local reference ranges using standardized acquisition and analysis protocols is therefore essential to ensure accurate clinical interpretation within a given population. Such standardization requires harmonized imaging parameters, unified vendor platforms, consistent post-processing workflows, and the definition of reproducible summary metrics (e.g., mean ± Standard Deviation (SD)) to enhance comparability and clinical applicability14,15. Accordingly, this study aims to establish native myocardial T1 reference values for a Tunisian cohort at 1.5 T, with particular emphasis on sex-related differences and age-related trends. The performance of two widely used acquisition strategies, MOLLI 5(3)3 and SMART1Map, is evaluated to assess sequence-dependent variability. Furthermore, a multi-center comparison of MOLLI measurements is conducted using harmonized acquisition protocols, a unified vendor platform, and standardized post-processing software to examine inter-site reproducibility and generalizability. Literature review and current gaps in T1 mapping reference values The 2013 consensus statement from the T1 Mapping Development Group established methodological standards for myocardial T1 mapping, emphasizing site-specific reference values due to variability related to scanner vendor, field strength, pulse sequence, and population characteristics16. The document highlighted standardized acquisition protocols, careful Region of Interest (ROI) placement, correction for B1 inhomogeneity, and rigorous quality control to improve reproducibility. Importantly, it underscored that native T1 values are not interchangeable across platforms or sequences. Published native myocardial T1 reference values at 1.5 T are summarized in Table 1. Corresponding 3 T values are provided in Supplementary Table S1.Table 1Native myocardial T1 reference values at 1.5 T.AuthorYearNSequence (scheme)Vendor/ModelROI DefinitionEthnicity/nationalityReported T1 value (ms)Sex effectMessroghli et al. 17200615MOLLI 3(3)3(3)5Philips Healthcare Gyroscan Intera CVMid-ventricular SAX: endocardial and epicardial contours manually traced; blood pool excluded–980 ± 53NRPiecknick et al. 182013342SHMOLLISiemens Healthineers AvantoGlobal LV myocardium; multi-slice SAX average–962 ± 25M: 950 ± 20;F: 974 ± 23YesLiu et al. 1920131231MOLLI 3(3)3(3)5Siemens Healthineers Avanto and EspreeMid-ventricular SAX; central myocardial ROI; blood pool and epicardial fat excluded–977 ± 42M: 968 ± 38;F: 986 ± 45YesDabir et al. 202014102MOLLI 3(3)3(3)5Philips Healthcare Achieva / IngeniaMid-ventricular SAX; interventricular septum ROI (single slice); blood pool excluded–950 ± 21NRRauhalammi et al. 21201684MOLLI 3(3)3(3)5Siemens Healthineers Magnetom AvantoMid-ventricular SAX; interventricular septum ROI (single slice); blood pool excluded–944 ± 25M: 931.5 ± 22.2;F: 960.0 ± 20.3YesBulluck et al. 222017101MOLLI 5(3)3Siemens Healthineers Magnetom AeraMid-ventricular SAX; inferior interventricular septum ROI; blood pool excludedChinese1013 ± 27M: 1001 ± 23;F: 1025 ± 26YesLiu et al. 2320171291SHMOLLISiemens Healthineers Magnetom AvantoMid-ventricular SAX; septal or focal lesion ROI–941 ± 23NRRosmini et al. 24201894SHMOLLISiemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–957 ± 30M: 948 ± 26;F: 966 ± 31YesRosmini et al. 24201894MOLLI 5(3)3Siemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–1024 ± 39M: 1008 ± 33;F: 1043 ± 37YesRosmini et al. 24201894SASHASiemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–1144 ± 45M: 1120 ± 35;F: 1171 ± 41YesMatsumoto et al. 25201824SMART1MapGE Healthcare WaukeshaMid-ventricular SAX; interventricular septum ROIJapanese1227.3 ± 41.9NRMatsumoto et al. 25201824MOLLI 3(3)3(3)5GE Healthcare WaukeshaMid-ventricular SAX; interventricular septum ROIJapanese1014.8 ± 49.4NRGranitz et al. 26201960MOLLI 5(3)3Philips Healthcare IngeniaMid-ventricular SAX; left ventricular myocardium–995.8 ± 30.9M: 984.4 ± 26.4; F: 1004.6 ± 31.5YesBurkhardt et al. 27201934SMART1MapGE Healthcare MilwaukeeMid-ventricular SAX; interventricular septum and total LV myocardium; basal and apical slices additionally measured–1191–1221NRBurkhardt et al. 27201934MOLLI 3(3)3(3)5GE Healthcare MilwaukeeMid-ventricular SAX; interventricular septum and total LV myocardium; basal and apical slices additionally measured–990–1042NRAlseid et al. 282020102MOLLI 5(3)3Philips Healthcare IngeniaGlobal LV myocardium; mid-ventricular SAX–1018 ± 25NRMeloni et al. 292021100MOLLI 3(3)3(3)5GE Healthcare Signa ArtistGlobal LV myocardium; 16-segment AHA modelCaucasiansM: 993 ± 32;F: 1037 ± 27YesShaw et al. 30202250MOLLI 3(3)5Siemens Healthineers AeraGlobal LV myocardium; 16-segment AHA modelIndian946.86 ± 14.16NoRamirez et al. 31202328MOLLI 3(3)5Siemens Healthineers Magnetom AeraThree SAX slices (basal, mid-ventricular, apical); 16-segment AHA model; 10% endocardial and epicardial erosion; basal blood pool ROI includedLatin American1069.41 ± 38.73M: 1049;F: 1086YesAbdullah et al. 32202351MOLLI 5(3)3Siemens Healthineers AmiraBasal, mid-ventricular, and apical SAX slices; septal free-hand crescent-shaped ROI; artifacts excludedMalaysianM: 999.5 ± 29.3; F: 1017.6 ± 35.4YesÖzcan et al. 332024125MOLLI 5(3)3Siemens Healthineers Magnetom AeraGlobal LV myocardium; 16-segment AHA modelTurkish1016.2 ± 57.2M: 999.2 ± 39.0; F: 1035.3 ± 49.8YesCadour et al. 342025172MOLLI 5(3)3Siemens Healthineers (model not specified)Mid-ventricular SAX; AHA segments 7–12; septal segments 8 and 9 averaged; manual ROI with endocardial and epicardial borders excluded–M: 949.7 ± 18.3; F: 980.9 ± 26.2YesTammo et al. 352025312MOLLI (scheme not specified)Siemens Healthineers Magnetom Avanto FitSAX slices; interventricular septum onlyUkrainian1000.2NRRana et al. 362025491MOLLI 5(3)3Philips Healthcare AchievaGlobal LV myocardium; basal, mid-ventricular, and apical SAX slices; manual ROI excluding blood pool and pericardium–1004.7 ± 49.8NRLiukkonen et al. 37202647MOLLI 5(3)3Siemens Healthineers Magnetom SolaGlobal LV myocardium; 16-segment AHA model; endocardial and epicardial borders contoured; papillary muscles excludedFinnishM: 1020 ± 17;F: 1040 ± 29YesMOLLI: modified look-locker imaging; SAPPIRE: saturation pulse prepared heart rate independent inversion recovery; SASHA: saturation recovery single-shot acquisition; SMART1Map: saturation method using adaptive recovery times for cardiac T1 mapping; SHMOLLI: shortened modified look-locker imaging; ms = milliseconds; N = sample size; T = Tesla.ROI definitions differ across studies (global LV, septal, or segmental), contributing to variability in reported T1 values. Native myocardial T1 values are presented according to ROI, specific to 1.5 T acquisitions, and reported as mean ± SD unless otherwise indicated. Ranges (e.g., 1191–1221 ms) indicate the minimum and maximum T1 measured within the cohort. Table 1Native myocardial T1 reference values at 1.5 T.AuthorYearNSequence (scheme)Vendor/ModelROI DefinitionEthnicity/nationalityReported T1 value (ms)Sex effectMessroghli et al. 17200615MOLLI 3(3)3(3)5Philips Healthcare Gyroscan Intera CVMid-ventricular SAX: endocardial and epicardial contours manually traced; blood pool excluded–980 ± 53NRPiecknick et al. 182013342SHMOLLISiemens Healthineers AvantoGlobal LV myocardium; multi-slice SAX average–962 ± 25M: 950 ± 20;F: 974 ± 23YesLiu et al. 1920131231MOLLI 3(3)3(3)5Siemens Healthineers Avanto and EspreeMid-ventricular SAX; central myocardial ROI; blood pool and epicardial fat excluded–977 ± 42M: 968 ± 38;F: 986 ± 45YesDabir et al. 202014102MOLLI 3(3)3(3)5Philips Healthcare Achieva / IngeniaMid-ventricular SAX; interventricular septum ROI (single slice); blood pool excluded–950 ± 21NRRauhalammi et al. 21201684MOLLI 3(3)3(3)5Siemens Healthineers Magnetom AvantoMid-ventricular SAX; interventricular septum ROI (single slice); blood pool excluded–944 ± 25M: 931.5 ± 22.2;F: 960.0 ± 20.3YesBulluck et al. 222017101MOLLI 5(3)3Siemens Healthineers Magnetom AeraMid-ventricular SAX; inferior interventricular septum ROI; blood pool excludedChinese1013 ± 27M: 1001 ± 23;F: 1025 ± 26YesLiu et al. 2320171291SHMOLLISiemens Healthineers Magnetom AvantoMid-ventricular SAX; septal or focal lesion ROI–941 ± 23NRRosmini et al. 24201894SHMOLLISiemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–957 ± 30M: 948 ± 26;F: 966 ± 31YesRosmini et al. 24201894MOLLI 5(3)3Siemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–1024 ± 39M: 1008 ± 33;F: 1043 ± 37YesRosmini et al. 24201894SASHASiemens Healthineers Magnetom AvantoMid-ventricular SAX; AHA segment 8 (mid-anteroseptal)–1144 ± 45M: 1120 ± 35;F: 1171 ± 41YesMatsumoto et al. 25201824SMART1MapGE Healthcare WaukeshaMid-ventricular SAX; interventricular septum ROIJapanese1227.3 ± 41.9NRMatsumoto et al. 25201824MOLLI 3(3)3(3)5GE Healthcare WaukeshaMid-ventricular SAX; interventricular septum ROIJapanese1014.8 ± 49.4NRGranitz et al. 26201960MOLLI 5(3)3Philips Healthcare IngeniaMid-ventricular SAX; left ventricular myocardium–995.8 ± 30.9M: 984.4 ± 26.4; F: 1004.6 ± 31.5YesBurkhardt et al. 27201934SMART1MapGE Healthcare MilwaukeeMid-ventricular SAX; interventricular septum and total LV myocardium; basal and apical slices additionally measured–1191–1221NRBurkhardt et al. 27201934MOLLI 3(3)3(3)5GE Healthcare MilwaukeeMid-ventricular SAX; interventricular septum and total LV myocardium; basal and apical slices additionally measured–990–1042NRAlseid et al. 282020102MOLLI 5(3)3Philips Healthcare IngeniaGlobal LV myocardium; mid-ventricular SAX–1018 ± 25NRMeloni et al. 292021100MOLLI 3(3)3(3)5GE Healthcare Signa ArtistGlobal LV myocardium; 16-segment AHA modelCaucasiansM: 993 ± 32;F: 1037 ± 27YesShaw et al. 30202250MOLLI 3(3)5Siemens Healthineers AeraGlobal LV myocardium; 16-segment AHA modelIndian946.86 ± 14.16NoRamirez et al. 31202328MOLLI 3(3)5Siemens Healthineers Magnetom AeraThree SAX slices (basal, mid-ventricular, apical); 16-segment AHA model; 10% endocardial and epicardial erosion; basal blood pool ROI includedLatin American1069.41 ± 38.73M: 1049;F: 1086YesAbdullah et al. 32202351MOLLI 5(3)3Siemens Healthineers AmiraBasal, mid-ventricular, and apical SAX slices; septal free-hand crescent-shaped ROI; artifacts excludedMalaysianM: 999.5 ± 29.3; F: 1017.6 ± 35.4YesÖzcan et al. 332024125MOLLI 5(3)3Siemens Healthineers Magnetom AeraGlobal LV myocardium; 16-segment AHA modelTurkish1016.2 ± 57.2M: 999.2 ± 39.0; F: 1035.3 ± 49.8YesCadour et al. 342025172MOLLI 5(3)3Siemens Healthineers (model not specified)Mid-ventricular SAX; AHA segments 7–12; septal segments 8 and 9 averaged; manual ROI with endocardial and epicardial borders excluded–M: 949.7 ± 18.3; F: 980.9 ± 26.2YesTammo et al. 352025312MOLLI (scheme not specified)Siemens Healthineers Magnetom Avanto FitSAX slices; interventricular septum onlyUkrainian1000.2NRRana et al. 362025491MOLLI 5(3)3Philips Healthcare AchievaGlobal LV myocardium; basal, mid-ventricular, and apical SAX slices; manual ROI excluding blood pool and pericardium–1004.7 ± 49.8NRLiukkonen et al. 37202647MOLLI 5(3)3Siemens Healthineers Magnetom SolaGlobal LV myocardium; 16-segment AHA model; endocardial and epicardial borders contoured; papillary muscles excludedFinnishM: 1020 ± 17;F: 1040 ± 29YesMOLLI: modified look-locker imaging; SAPPIRE: saturation pulse prepared heart rate independent inversion recovery; SASHA: saturation recovery single-shot acquisition; SMART1Map: saturation method using adaptive recovery times for cardiac T1 mapping; SHMOLLI: shortened modified look-locker imaging; ms = milliseconds; N = sample size; T = Tesla.ROI definitions differ across studies (global LV, septal, or segmental), contributing to variability in reported T1 values. Native myocardial T1 values are presented according to ROI, specific to 1.5 T acquisitions, and reported as mean ± SD unless otherwise indicated. Ranges (e.g., 1191–1221 ms) indicate the minimum and maximum T1 measured within the cohort. Native myocardial T1 reference values at 1.5 T. 962 ± 25 M: 950 ± 20; F: 974 ± 23 977 ± 42 M: 968 ± 38; F: 986 ± 45 944 ± 25 M: 931.5 ± 22.2; F: 960.0 ± 20.3 1013 ± 27 M: 1001 ± 23; F: 1025 ± 26 957 ± 30 M: 948 ± 26; F: 966 ± 31 1024 ± 39 M: 1008 ± 33; F: 1043 ± 37 1144 ± 45 M: 1120 ± 35; F: 1171 ± 41 995.8 ± 30.9 M: 984.4 ± 26.4; F: 1004.6 ± 31.5 M: 993 ± 32; F: 1037 ± 27 1069.41 ± 38.73 M: 1049; F: 1086 1016.2 ± 57.2 M: 999.2 ± 39.0; F: 1035.3 ± 49.8 M: 1020 ± 17; F: 1040 ± 29 MOLLI: modified look-locker imaging; SAPPIRE: saturation pulse prepared heart rate independent inversion recovery; SASHA: saturation recovery single-shot acquisition; SMART1Map: saturation method using adaptive recovery times for cardiac T1 mapping; SHMOLLI: shortened modified look-locker imaging; ms = milliseconds; N = sample size; T = Tesla. ROI definitions differ across studies (global LV, septal, or segmental), contributing to variability in reported T1 values. Native myocardial T1 values are presented according to ROI, specific to 1.5 T acquisitions, and reported as mean ± SD unless otherwise indicated. Ranges (e.g., 1191–1221 ms) indicate the minimum and maximum T1 measured within the cohort. Across acquisition schemes, native myocardial T1 values are generally higher at 3 T compared with 1.5 T. At 1.5 T, MOLLI-derived myocardial T1 values are most commonly reported in the approximate range of 940–1050 ms depending on acquisition scheme, ROI definition, and population characteristics17,19,21,26,28,33. At 3 T, reported MOLLI-derived values are typically higher, frequently ranging between approximately 1150 and 1250 ms across studies using septal or global myocardial measurements38,40,43,45,52. This difference reflects the known dependence of longitudinal relaxation time on magnetic field strength. Importantly, the magnitude of this field-related shift may vary according to acquisition scheme, vendor-specific implementation, fitting approach, and underlying tissue properties. Even after controlling field strength, acquisition strategy remains a major determinant of measured T1 values. At 1.5 T, MOLLI schemes demonstrate moderate dispersion, SHMOLLI yields slightly lower values, and saturation-based techniques such as SASHA and SMART1Map consistently produce higher values. At 3 T, inter-sequence divergence becomes more pronounced, with differences exceeding 400 ms between IR and SR approaches. These findings underscore acquisition strategy as a dominant technical source of variability. Sex-related differences represent another reproducible determinant. Across multiple vendors and populations, females consistently demonstrate higher native myocardial T1 measurements than males. At 1.5 T, reported differences typically range from 20 to 40 ms, while at 3 T this pattern appears even more robust. The reproducibility of this finding supports consideration of sex-specific reference ranges. Substantial inter-population variability persists even within comparable acquisition frameworks. At 1.5 T, MOLLI-based values span from the mid-940 ms range in certain Asian and South Asian cohorts to values exceeding 1060 ms in Latin American populations. At 3 T, MOLLI 5(3)3 values range from approximately 1120 ms to 1270 ms in recent cohorts. Notably, newly published studies from 2024–2026 continue to demonstrate persistent inter-cohort variability despite increasing methodological standardization, reinforcing that myocardial native T1 behaves as a context-dependent biomarker rather than a universal constant. Collectively, available evidence demonstrates that native myocardial T1 reference ranges remain strongly dependent on field strength, sequence design, vendor platform, sex, and geographic population. Despite expanding international datasets, a rigorously acquired, vendor-specific native myocardial T1 reference dataset has not yet been established for North African populations. This absence of regional reference standards constitutes a clinically relevant and methodologically significant gap and provides the rationale for the present study. Material and methods Participant characteristics Group 1 comprised 64 healthy volunteers (24 females, 40 males) aged 5 to 77 years, selected to establish native T1 reference values across a broad age spectrum representative of the Tunisian population. This cohort included 27 pediatric subjects (< 18 years), aged 5 to 15 years, of whom 14 were ≤ 10 years old. All participants were screened using strict eligibility criteria to exclude conditions that could affect myocardial tissue characterization. Individuals with a history of cardiovascular disease, malignancy or prior cancer therapy (e.g., chemotherapy or radiotherapy), systemic disorders, or other relevant comorbidities were excluded. Comprehensive clinical evaluation and medical history review confirmed the absence of cardiovascular pathology. For pediatric participants, examinations were performed only when adequate breath-hold compliance was achievable, and parental consent was obtained for all minors. Additional exclusion criteria included chronic pulmonary or cardiac disease, smoking, implanted cardiac devices, pregnancy, arterial hypertension, and hearing impairment. Group 2 included 30 healthy adult controls (20 females, 10 males), aged 19–62 years (median 35 years), recruited to evaluate inter-site variability in native T1 measurements under comparable acquisition protocols. Participants met the same eligibility criteria as Group 1. Anthropometric data, including height and weight, were collected to calculate body surface area (BSA) and body mass index (BMI). Recruitment involved public calls and institutional advertisements, with clinical screening to confirm the absence of known or suspected disease, distinguishing “healthy volunteers” from patients without obvious pathology. Ethics statement The study protocol, including pediatric participants, was approved by the Institutional Review Board of the Principal Military Hospital of Tunis (approval number: 42/2023/CLPP/Hôpital Militaire de Tunis), in accordance with the Declaration of Helsinki. Written informed consent was obtained from all adult participants and from the parents or legal guardians of all minors prior to participation. To protect participant confidentiality, all data obtained were de-identified and securely archived. The study was conducted in accordance with ethical standards for research involving human participants, with full respect for participant autonomy, safety, and confidentiality. CMR protocol All examinations were performed using a 1.5 T system (SIGNA Artist, GE Healthcare) at both sites to ensure protocol uniformity through identical acquisition settings for cine imaging and MOLLI 5(3)3 sequences. Cine imaging was performed using ECG-gated breath-hold acquisitions to assess cardiac function, ventricular volumes, and myocardial mass. Images were acquired in three long-axis views (horizontal long-axis, vertical long-axis, and 3-chamber) and a contiguous stack of short-axis (SAX) slices covering the entire left ventricle. Imaging parameters included a matrix size of 156 × 192, echo time (TE) of 1.3 ms, repetition time (TR) of 3.1 ms, slice thickness of 6 mm, field of view (FOV) of 276 × 340 mm2, receiver bandwidth (BW) of 704 Hz, flip angle (FA) of 45°, asymmetric echo factor of 0.29, and 30 cardiac phases. Native T1 mapping was performed at basal, mid-ventricular, and apical SAX levels using both MOLLI 5(3)3 and SMART1Map sequences. All acquisitions were performed during breath-hold at end-expiration to minimize motion artifacts, with consistent slice positioning across sequences for each participant. Detailed acquisition parameters for T1 mapping sequences are summarized in Table 2.Table 2Typical CMR parameters performed on all participants.ParametersVendorGeneral ElectricModelSIGNA ARTISTField Strength1.5 TSequenceMOLLISMART1MapSchematic5(3)3–TR (ms)3.23.38FOV Varies according to the size and anatomy of volunteersBW(Hz)488.281325.508Slice thickness (mm)88TE (ms)1.441.51FA (⁰)3565Matrix size160⨯128160⨯128Time increment (ms)80– Table 2Typical CMR parameters performed on all participants.ParametersVendorGeneral ElectricModelSIGNA ARTISTField Strength1.5 TSequenceMOLLISMART1MapSchematic5(3)3–TR (ms)3.23.38FOV Varies according to the size and anatomy of volunteersBW(Hz)488.281325.508Slice thickness (mm)88TE (ms)1.441.51FA (⁰)3565Matrix size160⨯128160⨯128Time increment (ms)80– Typical CMR parameters performed on all participants. T1 mapping post-processing and analysis Post-processing was performed using \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{CVI}}^{42}$$\end{document} software. Anatomical images were registered, and the myocardium was segmented on SAX slices. Endocardial and epicardial contours were manually refined to minimize partial volume effects and account for respiratory misregistration. All T1 measurements were performed by a single experienced observer in a core laboratory to ensure methodological consistency. Voxel-wise T1 values (ms) were generated using sequence-specific exponential fitting models. ROI-level T1 values were calculated as the mean ± SD of voxel- wise estimates within each segmented myocardial region. No explicit propagation of voxel-wise fitting uncertainty was performed; therefore, ROI-level SD reflects both spatial heterogeneity and measurement-related variability. For MOLLI, T1 values were estimated using a three-parameter exponential recovery model:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{S}\left(\mathrm{TI}\right)=\text{ A }-\text{ B}\cdot \text{ exp }( - \frac{TI}{T1*} )$$\end{document}with a subsequent correction to obtain the true T1 value:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{T}1= {T1}^{*} \cdot (\frac{\mathrm{B}}{\mathrm{A}}-1 )$$\end{document} For SMART1Map, signal intensity was modeled using a two-parameter exponential function:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{S}\left(\mathrm{TI}\right)=\text{ A }\cdot (1- {e}^{- \frac{TI}{T1}} )$$\end{document} A border offset was applied during contour refinement to exclude myocardial edges and reduce partial volume contamination. T1 maps were visualized using standardized color scales; quantitative analyses were based solely on voxel-wise fitted values. Maps were interpreted alongside cine imaging for anatomical reference. Figure 1 illustrates the complete workflow for establishing native T1 reference values and assessing inter-site variability.Fig. 1Overall workflow for T1 value establishment and site comparison. Overall workflow for T1 value establishment and site comparison. Statistical analysis Robust statistical methods were employed to evaluate and compare native myocardial T1 estimates across sequences and participant groups. Paired t-tests were used to compare mean T1 values obtained from MOLLI and SMART1Map sequences. Sex-related differences in T1 values across myocardial regions were assessed using independent samples t-tests, with Levene’s test applied to confirm equality of variances. Age dependence was evaluated using Pearson correlation analysis and linear regression modeling between age and global native T1 values for each sequence. Inter-site variability was assessed using two-sample t-tests to compare T1 measurements acquired on identical MRI scanners at two separate hospitals. Statistical significance was defined as p < 0.05. All analyses were performed using IBM SPSS Statistics, Version 25 (IBM Corp., Armonk, NY, USA). Results A total of 81 individuals were initially screened for participation in Group 1. After application of the predefined exclusion criteria, 64 participants (24 females, 40 males) were included to establish native myocardial T1 reference values for the Tunisian population. The mean age of the cohort was 36.8 ± 21.1 years, with women representing 37.7% of participants. Seventeen individuals were excluded during the selection process: one due to claustrophobia, fourteen due to scheduling constraints, and two due to ECG gating failure during image acquisition; All exclusions were documented to ensure data integrity. Males were significantly younger than females (31.2 vs. 45.6 years, p < 0.001), whereas no significant differences were observed between sexes in weight, BMI, or indexed CMR parameters. For comparative analyses, only adult participants from Group 1 within a similar age range were included; pediatric participants were excluded. Although individual pairwise matching was not performed, frequency matching by age intervals ensured comparable age distribution between the groups. Statistical comparison of age between the adult subgroup of Group 1 (n = 30) and Group 2 (n = 30) confirmed no significant difference (40.2 ± 11.8 vs. 37.9 ± 11.5 years; p = 0.499, independent samples t-test). Sex distribution differed between groups (Group 1 adults: 20 males, 10 females; Group 2: 10 males, 20 females) and was not used as a matching variable, given that inter-site variability was the primary outcome of interest. Anthropometric and physiological parameters (height, weight, BMI, BSA, HR) were available only for Group 1; therefore, no inter-group comparisons were performed for these variables. A detailed comparison of Group 1 demographic and CMR characteristics by sex is presented in Table 3.Table 3Clinical and CMR-derived left and right ventricular parameters stratified by sex in group 1 (overall healthy cohort).Overall population (N = 64)Men (N = 40)Women (N = 24)P valueClinicalAge, years36.8 ± 21.131.2 ± 22.345.6 ± 15.5 < 0.001Weight, Kg69.6 ± 18.369.48 ± 21.169.74 ± 15.10.176BMI, Kg/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2}$$\end{document}25.1 ± 5.024.4 ± 4.426.0 ± 5.50.482Height, cm164.4 ± 14.3164.9 ± 19.2163.9 ± 5.90.76BSA,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2}$$\end{document}1.8 ± 0.31.79 ± 0.221.77 ± 0.160.65CMRLVEF (%)64.1 ± 10.160.7 ± 12.966.7 ± 6.30.025 LVEDV (ml)121.9 ± 31.7131.4 ± 33.5114.3 ± 28.80.215LVESV (ml)45.9 ± 18.352.8 ± 18.640.4 ± 16.50.311LV mass (g)66.6 ± 16.271.5 ± 24.564.6 ± 12.70.051LVEDF index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}70.3 ± 14.772.4 ± 15.968.6 ± 13.80.360LVESV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}26.5 ± 9.729.7 ± 11.923.9 ± 6.90.028RVEF (%)60.7 ± 11.659.9 ± 14.661.3 ± 8.80.432RVEDV (ml)123.3 ± 33.1135.5 ± 35.6113.0 ± 27.10.179RVESV (ml)51.8 ± 21.461.5 ± 24.943.7 ± 14.00.287RVEDV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}71.8 ± 18.575.8 ± 23.268.5 ± 13.10.115RVESV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}30.2 ± 18.033.9 ± 26.427.1 ± 7.80.147Values are expressed as mean ± SD. P values indicate differences between male and female subgroups. Abbreviations: BSA: body surface area; BMI: body mass index; LVEF: left ventricular ejection fraction; LVEDV: left ventricular end-diastolic volume; LVESV: left ventricular end-systolic volume; LVEDF index: LVEDV indexed to BSA; LVESV index: LVESV indexed to BSA; LV mass: left ventricular mass; RVEF: right ventricular ejection fraction; RVEDV: right ventricular end-diastolic volume; RVESV: right ventricular end-systolic volume; RVEDV index: RVEDV indexed to BSA; RVESV index: RVESV indexed to BSA. Table 3Clinical and CMR-derived left and right ventricular parameters stratified by sex in group 1 (overall healthy cohort).Overall population (N = 64)Men (N = 40)Women (N = 24)P valueClinicalAge, years36.8 ± 21.131.2 ± 22.345.6 ± 15.5 < 0.001Weight, Kg69.6 ± 18.369.48 ± 21.169.74 ± 15.10.176BMI, Kg/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2}$$\end{document}25.1 ± 5.024.4 ± 4.426.0 ± 5.50.482Height, cm164.4 ± 14.3164.9 ± 19.2163.9 ± 5.90.76BSA,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2}$$\end{document}1.8 ± 0.31.79 ± 0.221.77 ± 0.160.65CMRLVEF (%)64.1 ± 10.160.7 ± 12.966.7 ± 6.30.025 LVEDV (ml)121.9 ± 31.7131.4 ± 33.5114.3 ± 28.80.215LVESV (ml)45.9 ± 18.352.8 ± 18.640.4 ± 16.50.311LV mass (g)66.6 ± 16.271.5 ± 24.564.6 ± 12.70.051LVEDF index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}70.3 ± 14.772.4 ± 15.968.6 ± 13.80.360LVESV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}26.5 ± 9.729.7 ± 11.923.9 ± 6.90.028RVEF (%)60.7 ± 11.659.9 ± 14.661.3 ± 8.80.432RVEDV (ml)123.3 ± 33.1135.5 ± 35.6113.0 ± 27.10.179RVESV (ml)51.8 ± 21.461.5 ± 24.943.7 ± 14.00.287RVEDV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}71.8 ± 18.575.8 ± 23.268.5 ± 13.10.115RVESV index (ml/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{m}}^{2})$$\end{document}30.2 ± 18.033.9 ± 26.427.1 ± 7.80.147Values are expressed as mean ± SD. P values indicate differences between male and female subgroups. Abbreviations: BSA: body surface area; BMI: body mass index; LVEF: left ventricular ejection fraction; LVEDV: left ventricular end-diastolic volume; LVESV: left ventricular end-systolic volume; LVEDF index: LVEDV indexed to BSA; LVESV index: LVESV indexed to BSA; LV mass: left ventricular mass; RVEF: right ventricular ejection fraction; RVEDV: right ventricular end-diastolic volume; RVESV: right ventricular end-systolic volume; RVEDV index: RVEDV indexed to BSA; RVESV index: RVESV indexed to BSA. Clinical and CMR-derived left and right ventricular parameters stratified by sex in group 1 (overall healthy cohort). Values are expressed as mean ± SD. P values indicate differences between male and female subgroups. Abbreviations: BSA: body surface area; BMI: body mass index; LVEF: left ventricular ejection fraction; LVEDV: left ventricular end-diastolic volume; LVESV: left ventricular end-systolic volume; LVEDF index: LVEDV indexed to BSA; LVESV index: LVESV indexed to BSA; LV mass: left ventricular mass; RVEF: right ventricular ejection fraction; RVEDV: right ventricular end-diastolic volume; RVESV: right ventricular end-systolic volume; RVEDV index: RVEDV indexed to BSA; RVESV index: RVESV indexed to BSA. Evaluating T1 reference values: a comparison between SMART1Map and MOLLI 5(3)3 Native T1 values were compared between MOLLI 5(3)3 and SMART1Map at 1.5 T at basal, mid-ventricular, and apical levels (Table 4) to assess inter-sequence differences and regional distribution relevant to myocardial tissue characterization.Table 4Native T1 mapping for Basal, Mid, and Apical planes employing MOLLI and SMART1Map.SequenceGlobal T1Basal T1Mid-VentricularApicalMOLLI 5(3)31049 ± 551035 ± 661047 ± 541067 ± 53SMART1Map1184 ± 701204 ± 941176 ± 661180 ± 56Data are expressed as mean ± SD. Variability is described using SD without formal statistical comparison of dispersion. Table 4Native T1 mapping for Basal, Mid, and Apical planes employing MOLLI and SMART1Map.SequenceGlobal T1Basal T1Mid-VentricularApicalMOLLI 5(3)31049 ± 551035 ± 661047 ± 541067 ± 53SMART1Map1184 ± 701204 ± 941176 ± 661180 ± 56Data are expressed as mean ± SD. Variability is described using SD without formal statistical comparison of dispersion. Native T1 mapping for Basal, Mid, and Apical planes employing MOLLI and SMART1Map. Data are expressed as mean ± SD. Variability is described using SD without formal statistical comparison of dispersion. SMART1Map yielded consistently higher T1 values than MOLLI 5(3)3 across all myocardial levels (p < 0.001), consistent with a systematic inter-sequence offset attributable to differences in acquisition strategies (IR vs. SR). Accordingly, interpretation focused on the magnitude and spatial distribution of this offset rather than on comparative accuracy. In terms of variability, MOLLI demonstrated slightly lower within-slice SDs at basal and mid-ventricular levels, whereas SMART1Map showed somewhat greater dispersion, particularly at the basal slice. These differences are reported descriptively, as SD reflects the spread of measurements rather than sequence robustness. Both sequences exhibited modest inter-regional variation with comparable basal-to-apical gradients (32 ms for MOLLI and 24 ms for SMART1Map), suggesting similar spatial distribution patterns despite the systematic offset. Analysis was extended to the 16 standardized American Heart Association (AHA) cardiac segments (Fig. 2) to evaluate regional T1 patterns. Within each sequence, no substantial segment-to-segment differences were observed.Fig. 2T1 relaxation values (ms) distribution across every cardiac segment. The values in the central box range from the lowest to the highest quartile. The middle line represents the median. Outliers are represented as outliers. T1 relaxation values (ms) distribution across every cardiac segment. The values in the central box range from the lowest to the highest quartile. The middle line represents the median. Outliers are represented as outliers. For MOLLI, the lowest mean T1 values were consistently found in the inferior segments (1023 ± 64 ms at basal, 1033 ± 63 ms at mid-ventricular levels), while the highest values appeared in the anteroseptal segments (1046 ± 77 ms at basal, 1049 ± 58 ms at mid-ventricular levels). This pattern persisted at the apical level, with inferior segments showing the lowest values (1061 ± 70 ms) and septal segments the highest (1081 ± 50 ms). SMART1Map displayed slightly different regional trends: the lowest T1 values were measured in the anterolateral segment (1181 ± 10 ms at basal, 1167 ± 10 ms at mid-ventricular levels), whereas the highest values were observed in the inferoseptal and anteroseptal segments at basal and mid-ventricular levels, respectively. At the apical level, the anterior segment had the lowest values (1171 ± 71 ms), and the septal segment the highest (1188 ± 73 ms). Across all segments, SMART1Map values remained significantly higher than MOLLI values (p < 0.001), reflecting a systematic inter-sequence offset. Age-dependence of native myocardial T1 values Given the wide age range of the cohort (5–77 years), we assessed the relationship between age and global native myocardial T1. MOLLI-derived T1 values exhibited a significant negative correlation with age (r = –0.367, p = 0.012), with linear regression indicating a decrease of β = –0.84 ms/year, suggesting age contributes to variability in MOLLI measurements (Fig. 3). In contrast, SMART1Map-derived T1 values showed no significant association with age (r = 0.099, p = 0.556; β = 0.30 ms/year). These findings highlight sequence-specific age effects, with MOLLI demonstrating greater variability in older subjects and emphasizing the importance of considering age when interpreting MOLLI reference values.Fig. 3Age dependence of global native myocardial T1 estimates measured using MOLLI (blue) and SMART1Map (orange). Individual data points are shown along with linear regression lines and 95% confidence intervals (shaded areas). Age dependence of global native myocardial T1 estimates measured using MOLLI (blue) and SMART1Map (orange). Individual data points are shown along with linear regression lines and 95% confidence intervals (shaded areas). Gender-based variations in T1 reference values: male vs. female participants Analysis of native quantitative T1 values revealed no consistent sex-related differences across myocardial regions or sequences. Using the MOLLI sequence, T1 values were similar between males and females in the basal (1028 ± 62 ms vs. 1044 ± 71 ms) and mid-ventricular regions (1037 ± 47 ms vs. 1060 ± 61 ms), with no statistically significant differences. A statistically significant difference was identified only at the apical level (1050 ± 47 ms vs. 1090 ± 54 ms, p = 0.023). In contrast, SMART1Map showed no statistically significant sex-related differences across basal (1191 ± 71 ms vs. 1225 ± 12 ms), mid-ventricular (1164 ± 82 ms vs. 1184 ± 42 ms), or apical regions (1172 ± 46 ms vs. 1189 ± 42 ms) (Fig. 4).Fig. 4Sex-related differences in native T1 values (ms) for MOLLI and SMART1Map sequences at 1.5 T. Sex-related differences in native T1 values (ms) for MOLLI and SMART1Map sequences at 1.5 T. Segmental native T1 values were analyzed and illustrated in the bull’s-eye plot (Fig. 5) to provide a detailed view of gender-specific patterns. For MOLLI, men had a mean T1 of 1017 ± 67 ms in the basal anterior segment (Segment 1) versus 1057 ± 81 ms in women; this difference approached significance (Levene’s test p = 0.049 for equal variances, p = 0.053 for unequal variances). A statistically significant gender difference was observed in the mid-anterior segment (Segment 7, p = 0.004). In contrast, SMART1Map-derived T1 values showed no significant gender differences across the majority of segments. Levene’s test indicated no significant variance differences (p = 0.142), and independent t-tests were non-significant across segments 2–16 (p = 0.115).Fig. 5Comparison of (A) MOLLI and (B) SMART1Map sequences illustrating sex-related mean native T1 values (± SD) across left ventricular segments from basal to apical levels. Error bars represent 95% confidence intervals. Comparison of (A) MOLLI and (B) SMART1Map sequences illustrating sex-related mean native T1 values (± SD) across left ventricular segments from basal to apical levels. Error bars represent 95% confidence intervals. Generalizability of T1 reference values across hospitals Native T1 measurements are sensitive to acquisition parameters and post-processing methods. To explore the extent to which strict technical standardization may reduce inter-site variability, two independent hospitals each scanned 30 healthy subjects using the same 1.5 T system (SIGNA Artist, GE Healthcare), identical acquisition parameters (MOLLI 5(3)3), and the same post-processing software (CVI42). This design enabled a controlled descriptive assessment of T1 distributions under harmonized technical conditions. At the global level, T1 distributions appeared broadly similar between sites at the basal and apical levels, while modest rightward displacement was visually observed at the mid-ventricular level (Fig. 6). Overlayed histograms with Gaussian fits demonstrated substantial overlap between sites, particularly at basal and apical levels. Distribution widths were also similar, suggesting comparable intra-group variability.Fig. 6Overlaid histograms of native T1 values at basal, mid-ventricular, and apical levels for Hospital 1 and Hospital 2. Gaussian fits illustrate broadly similar distribution patterns across sites, with substantial overlap at basal and apical levels and a slight displacement at the mid-ventricular level. These observations are descriptive and reflect inter-group variability rather than direct inter-site reproducibility. Overlaid histograms of native T1 values at basal, mid-ventricular, and apical levels for Hospital 1 and Hospital 2. Gaussian fits illustrate broadly similar distribution patterns across sites, with substantial overlap at basal and apical levels and a slight displacement at the mid-ventricular level. These observations are descriptive and reflect inter-group variability rather than direct inter-site reproducibility. As these observations are derived from independent subject groups rather than repeated measurements in the same individuals, they should be interpreted as descriptive only. They do not establish inter-site reproducibility but rather indicate that, under standardized acquisition and processing conditions, T1 values fall within a similar range across centers, with some residual variability likely reflecting biological differences. Overall, these findings suggest that while technical standardization contributes to reducing variability, site-related and biological factors remain relevant. These results support the continued use of site-specific reference ranges, even when acquisition and processing protocols are closely aligned. Discussion Native T1 mapping has become a cornerstone of quantitative CMR, enabling non-invasive assessment of myocardial tissue composition. However, native T1 values are not absolute biological constants; rather, they are composite measurements influenced by sequence design, field strength, vendor implementation, post-processing methodology, and intrinsic biological variability. Even under tightly controlled acquisition conditions, measurable dispersion persists17–54. Consequently, the establishment of population- and sequence-specific reference ranges is essential to ensure accurate clinical interpretation and appropriate diagnostic thresholds. To date, normative native T1 data have not been available for the Tunisian population. This study therefore provides the first reference dataset for healthy Tunisian subjects across a wide age spectrum, acquired at 1.5 T using both MOLLI 5(3)3 and SMART1Map. The simultaneous evaluation of two distinct mapping techniques within the same cohort enables direct intra-population comparison while minimizing demographic confounding, thereby strengthening internal validity. The principal finding of this study is the systematic and statistically significant offset between MOLLI and SMART1Map measurements across all myocardial levels. MOLLI-derived values (approximately 1035–1067 ms depending on level) were consistently lower than SMART1Map values (approximately 1176–1204 ms), with p < 0.001 at all levels. This divergence reflects fundamental mechanistic differences between IR–based and SR–based approaches. MOLLI is known to underestimate true T1 due to factors such as magnetization transfer effects, inversion efficiency, HR sensitivity, and T2 dependence24,25,39,40,42. In contrast, SMART1Map, as a SR technique, is designed to provide a more direct quantification of longitudinal relaxation with reduced sensitivity to some of these confounders, albeit often yielding higher absolute values25,27,39. The magnitude and direction of the observed offset are consistent with prior reports using SMART1Map25,27, suggesting that the discrepancy represents a predictable sequence-related bias rather than biological variation. Importantly, without phantom calibration against a standardized reference, the present data cannot determine which sequence most accurately reflects true myocardial T1. The inter-sequence difference should therefore be interpreted as a systematic methodological offset rather than an indicator of superiority. Global MOLLI T1 values in our cohort (1049 ± 55 ms) fall within the spectrum reported in contemporary 1.5 T studies across European, Turkish, Latin American, Malaysian, Ukrainian, and Finnish populations28,31–37, while tending toward the upper range of published values. They are slightly higher than those described in some earlier Asian cohorts18,21,23. These inter-population differences are likely multifactorial, reflecting variation in age distribution, sex composition, ROI definition, slice selection strategy, and acquisition scheme. Subtle ethnic or population-specific myocardial characteristics cannot be excluded, but methodological heterogeneity remains a dominant explanatory factor in literature. The concordance of our SMART1Map values (1184 ± 70 ms) with limited reports from Japanese and European cohorts suggests that SR–based measurements may demonstrate relatively stable cross-population behavior when acquisition parameters are harmonized25,27. Collectively, these findings indicate that Tunisian native T1 values align with global reference ranges but underscore the necessity of local reference datasets for clinical deployment. Given the inclusion of pediatric participants, the observed negative association between age and MOLLI-derived T1 may reflect developmental myocardial maturation in addition to age-related remodeling. The linear regression across the entire age spectrum likely oversimplifies the underlying biology, as myocardial composition evolves nonlinearly from childhood through adulthood. Moreover, the slight sex imbalance, with younger males compared to females, may partially confound the association. Notably, MOLLI T1 values exhibited a negative correlation with age, whereas SMART1Map measurements remained age-independent. The decline in MOLLI T1 with advancing age may reflect subtle myocardial remodeling, including alterations in water content, interstitial composition, and magnetization transfer effects. Because IR sequences are sensitive to T2-related and magnetization transfer phenomena, small microstructural changes may disproportionately influence MOLLI-derived estimates. In contrast, the stability of SMART1Map suggests that SR–based techniques may more selectively capture intrinsic longitudinal relaxation. These sequence-specific age behaviors highlight that the biological interpretation of native T1 should always be contextualized within the acquisition methodology. Sex differences in native T1 have been widely reported, particularly in large MOLLI-based cohorts, where females typically exhibit higher values19,21,24,33,37,52. In our study, sex-related differences were limited to the apical level using MOLLI, with no significant differences at basal or mid-ventricular levels and no detectable sex effect with SMART1Map. The region-specific and modest magnitude of this finding suggests that sex effects in native T1 may be more nuanced than previously assumed and potentially influenced by regional myocardial characteristics, sample size, and methodological factors. The absence of differences with SMART1Map further supports the possibility that reported sex effects may, in part, reflect sequence-dependent sensitivities rather than purely biological dimorphism. Larger cohorts with balanced sex distribution will be necessary to clarify the robustness and clinical relevance of sex-specific reference ranges in this population. Despite rigorous harmonization, including identical 1.5 T scanners (GE SIGNA Artist), uniform MOLLI 5(3)3 acquisition parameters, and standardized post-processing with CVI42, modest inter-site differences were observed, particularly at the mid-ventricular level. These residual differences under standardized conditions suggest that variability cannot be fully attributed to hardware or acquisition parameters alone. Subtle operational factors, such as slice positioning, breath-hold stability, physiological state, or demographic differences between sites, may contribute. Overall, overlayed T1 distributions indicate substantial overlap between centers, highlighting high inter-site comparability despite localized variability. These findings align with previous evidence that technical harmonization substantially reduces but does not completely eliminate inter-site variability, with residual dispersion reflecting both biological heterogeneity and subtle procedural factors55. These findings support the integration of sequence-specific reference thresholds into clinical workflows, which may improve diagnostic precision and reduce misclassification in myocardial tissue characterization. Several limitations warrant consideration. The cohort size, although sufficient for preliminary reference estimation, limits statistical power for subgroup analyses, particularly age- and sex-stratified comparisons. The predominance of male participants may further constrain generalizability. In addition, the study was conducted across only two centers using a single vendor platform at 1.5 T, which may limit the broader applicability of the findings to other scanners and acquisition environments. Inter-site comparison was restricted to MOLLI acquisitions and did not include SMART1Map, precluding full cross-sequence reproducibility assessment. Measurements were performed by a single experienced reader, preventing formal evaluation of inter-observer variability. Finally, no histological validation was available, and the findings therefore remain based solely on imaging-derived biomarkers. Future studies should include larger, demographically balanced cohorts, multi-center designs with multiple vendors, and multi-reader reproducibility analyses. The inclusion of post-contrast T1 and extracellular volume measurements would enable more comprehensive tissue characterization and facilitate integration into clinical decision-making. In this context, the availability of sequence-specific native T1 reference values may provide a critical quantitative foundation for the development and calibration of advanced data-driven frameworks, enabling improved diagnostic performance and more reliable automated myocardial tissue characterization, as increasingly explored in recent deep learning–based studies56,57. Finally, phantom-based validation would be valuable to assess absolute accuracy and to further disentangle biological variability from sequence-dependent bias. Conclusion This study establishes sequence-specific native myocardial T1 reference values for a Tunisian population at 1.5 T using MOLLI 5(3)3 and SMART1Map. A consistent and significant offset was observed between the two sequences, with SMART1Map yielding higher T1 measurements across all myocardial regions, reflecting fundamental differences between SR and IR techniques. Age-related variation was detected only in MOLLI-derived measurements, indicating sequence-dependent sensitivity to physiological factors, while sex-related differences were minimal and limited to a regional finding. Despite strict acquisition standardization, modest inter-site variability persisted, suggesting that both biological and operational factors contribute to residual dispersion. These findings highlight the importance of sequence-specific interpretation of T1 measurements and underscore the need for population-adapted reference frameworks. The proposed values provide a clinically relevant basis for myocardial tissue characterization in this population but should be applied with caution across different populations, acquisition settings, vendors, and magnetic field strengths. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1