Diagnostic accuracy of ADC values as a supplementary tool in differentiating hepatocellular adenoma from focal nodular hyperplasia: a systematic review and meta-analysis

Diagnostic accuracy of ADC values as a supplementary tool in differentiating hepatocellular adenoma from focal nodular hyperplasia: a systematic review and meta-analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Systematic Review Diagnostic accuracy of ADC values as a supplementary tool in differentiating hepatocellular adenoma from focal nodular hyperplasia: a systematic review and meta-analysis Alisa Mohebbi, Mehrad Zare, Afshin Mohammadi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8799129/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Accurate differentiation between hepatocellular adenoma (HCA) and focal nodular hyperplasia (FNH) is essential due to divergent management and risk profiles. Diffusion-weighted MRI (DWI) with apparent diffusion coefficient (ADC) quantification has been proposed as a non-contrast method to distinguish these lesions. Methods A systematic search of PubMed, Embase, Web of Science, and Cochrane Library was performed through August 2025 following PRISMA-DTA guidelines. Studies comparing ADC values in HCA and FNH, reporting measurements for at least five lesions per group, were included. Quality assessment used the QUADAS-2 tool. Random-effects meta-analysis pooled mean ADC values, calculated standardized mean differences (SMD), and assessed heterogeneity (I²). Sensitivity analyses, subgroup analyses by study design, and publication bias evaluation using trim-and-fill were conducted. Results Thirteen studies comprising 524 lesions (195 HCA, 329 FNH) met inclusion criteria. Pooled mean ADC was 1.44×10⁻³ mm²/s (95% CI: 1.34–1.55) for FNH and 1.35×10⁻³ mm²/s (95% CI: 1.23–1.47) for HCA, yielding an absolute difference of 0.10×10⁻³ mm²/s (p = 0.073) and SMD of 0.49 (95% CI: − 0.02–0.99). Relative ADC difference was 7.6% (p = 0.109). Heterogeneity was high (I² = 84.5%). Prospective studies showed reduced heterogeneity (I² = 0%) but non-significant mean difference (0.21×10⁻³ mm²/s). Trim-and-fill attenuated effect size further, confirming non-significance. Conclusion ADC values exhibit only modest, non-significant differences between HCA and FNH, with substantial heterogeneity limiting clinical applicability. Standardized, prospective multicenter studies and multiparametric MRI protocols remain necessary to improve non-contrast differentiation of these lesions. Nuclear Medicine & Medical Imaging Focal nodular hyperplasia (FNH) Apparent diffusion coefficient (ADC) Diffusion-weighted imaging (DWI) Hepatocellular adenoma (HCA) Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Distinguishing between hepatocellular adenoma (HCA) and focal nodular hyperplasia (FNH) represents one of the most consequential diagnostic challenges in hepatic imaging due to their profoundly different clinical implications, risk profiles, and management paradigms [ 1 ]. While both lesions may initially appear similar on routine imaging, their biological behavior diverges markedly, with critical consequences for patient outcomes [ 2 ]. This diagnostic distinction becomes particularly crucial when considering atypical HCA subtypes, which present unique challenges and carry varying degrees of risk for hemorrhage and malignant transformation [ 3 , 4 ]. The clinical significance of accurate differentiation cannot be overstated. HCA, particularly certain genetic subtypes, poses substantial risks including spontaneous bleeding and malignant transformation to hepatocellular carcinoma [ 5 , 6 ]. According to systematic reviews, the overall frequency of hemorrhage in HCA patients reaches 27.2%, with rupture and intraperitoneal bleeding occurring in 17.5% of cases [ 7 ]. Risk factors for bleeding are well-established and include lesion size of 5 cm or greater, location in the left hepatic lobe, and exophytic growth patterns [ 8 – 10 ]. Even more concerning is the potential for malignant transformation, which occurs in 4.2% of HCAs, with males exhibiting a higher risk compared to females [ 11 ]. Beta-catenin activated HCAs represent a particularly high-risk subtype, with up to 47% of males with these lesions developing hepatocellular carcinoma [ 12 ]. In contrast, FNH represents a benign, polyclonal lesion with no risk of malignant transformation or spontaneous hemorrhage [ 13 – 15 ]. This fundamental difference necessitates entirely divergent management strategies. While FNH typically does not require follow-up, HCA management depends on size, subtype, and patient demographics, often necessitating surgical resection or intensive monitoring [ 16 , 17 ]. The mortality associated with emergency resection for ruptured HCA ranges from 5–10%, compared to less than 1% for elective surgery [ 18 ]. Current imaging modalities face significant limitations in achieving this critical differentiation. Conventional ultrasonography, computed tomography, and standard magnetic resonance imaging demonstrate substantial overlap in enhancement patterns between HCA and FNH, particularly in smaller lesions or atypical presentations. The central scar, traditionally considered pathognomonic for FNH, is present in only 50% of cases and is frequently absent in lesions smaller than 3 cm [ 19 ]. Furthermore, HCA may exhibit features that mimic FNH, including similar enhancement patterns and the occasional presence of central scarring [ 20 ]. In addition, the reliance on contrast-enhanced imaging presents additional challenges; gadolinium-based contrast agents, while improving diagnostic accuracy, are contraindicated in patients with renal impairment, during pregnancy, or in those with contrast allergies. Even with hepatocyte-specific contrast agents like gadoxetate disodium, which represent the current non-invasive gold standard, diagnostic overlap persists, particularly with certain HCA subtypes that may demonstrate iso- or hyperintense uptake during hepatobiliary phase imaging [ 3 , 21 ]. Moreover, the substantial cost implications and the need for specialized expertise limit widespread accessibility to these advanced contrast agents. Advanced magnetic resonance techniques, particularly diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) quantification, have generated considerable interest as potential non-invasive biomarkers for distinguishing HCA from FNH. Several individual investigations have reported encouraging results, demonstrating statistically significant differences in ADC values between these lesions and proposing threshold cutoffs with promising diagnostic accuracy. In contrast, some studies have failed to confirm these findings, instead highlighting substantial overlap in ADC measurements that precludes reliable discrimination in routine practice. The present systematic review and meta-analysis therefore seeks to resolve these discrepancies by rigorously synthesizing all available evidence on ADC values in HCA versus FNH. This work aims to clarify whether ADC measurements can offer reproducible, contrast-free differentiation of these two benign hepatocellular lesions or whether their diagnostic promise is undermined by overlapping diffusion profiles. Ultimately, the findings of this study will inform whether quantitative DWI adoption as a supplementary imaging tool in clinical algorithms is clinically beneficial for focal liver lesion characterization. Materials and Methods Methods adhered to the comprehensive recommendations of the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy. The conduct and reporting of this review were guided by the PRISMA-DTA and PRISMA-Search (PRISMA-S) statements for systematic reviews and meta-analyses of diagnostic accuracy. To enhance reproducibility, the study protocol was prospectively registered on the Open Science Framework (OSF) at ( https://osf.io/vwq4f/ ) (See Supplementary A) before study initiation. Search strategy: Two reviewers independently designed the electronic search strategies using EMTREE and MeSH terms. They then jointly refined and harmonized their search syntax. Any discrepancies or disagreements were resolved through detailed discussion and, when necessary, consultation with a third reviewer to achieve consensus and optimize the search strategy. The final search was conducted on August 30, 2025, across multiple databases, including PubMed, Web of Science, Embase, and the Cochrane Library, without language restrictions to maximize inclusivity and minimize language bias. The full search strategy is provided in Supplementary B to promote transparency and reproducibility. In addition, the reference lists of all eligible studies identified in the database search were screened to further enhance the completeness of the literature search. All retrieved citations were subsequently imported into EndNote for management and de-duplication. Eligibility criteria: Studies were eligible for inclusion only if they reported ADC on a minimum of five discrete lesions, thereby ensuring adequate lesion sampling to support meaningful statistical comparison. To maintain methodological homogeneity, studies employing alternative imaging modalities such as contrast-enhanced sequences, elastography, or radiotracer techniques were deliberately excluded. We further refined our selection to include exclusively those investigations that reported HCA and FNH groups separately, thereby avoiding analytical confounding introduced by aggregated benign lesion cohorts. Conversely, studies limited to a single lesion type (only HCA or FNH) were also excluded, as our principal aim was to perform direct, within‐study comparisons of ADC values between these two entities. In cases where multiple publications drew upon the same patient cohort, only the report presenting the most comprehensive dataset was retained, thereby eliminating the risk of patient or data duplication and preserving the integrity of effect estimates. Finally, to maximize literature capture and mitigate the risk of publication bias, the bibliographies of all included articles were screened for additional pertinent studies, ensuring that our synthesis encompassed the full scope of available evidence. The study selection process consisted of a detailed screening of titles and abstracts, followed by an independent full-text assessment of all potentially eligible articles by two reviewers. Any disagreements about study inclusion were addressed through in-depth discussion until a consensus was achieved. This collaborative procedure strengthened the methodological rigor and reliability of the study selection. Risk of bias assessment: We conducted a rigorous study-level quality appraisal for all articles included in this systematic review and meta-analysis to enhance the robustness and trustworthiness of the findings. Two reviewers independently evaluated risk of bias using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool, a widely accepted and validated framework for appraising diagnostic accuracy research. QUADAS-2 examines four key domains that underpin the internal validity of diagnostic studies: (A) Patient selection: This domain assesses whether the enrolled participants are representative of the intended target population, thereby ensuring that the patients and lesions studied appropriately reflect the clinical context in which the diagnostic test is to be applied. (B) Index test: This domain focuses on the performance and interpretation of the index test—here, DWI and ADC measurement—evaluating whether they were conducted and interpreted in a standardized, unbiased, and reproducible manner. (C) Reference standard: This domain evaluates the adequacy and consistency of the histopathologic reference standard used, including its ability to correctly classify the target condition. In several studies, radiologist blinding to clinical and histopathological outcomes was incompletely reported, representing a possible source of bias. The assessment also considers whether the reference standard was applied uniformly across all participants. (D) Flow and timing: This domain examines the interval and sequence between the index test and the reference standard, ensuring that the time gap is appropriate and that any changes in the patient’s clinical status do not compromise the validity of the diagnostic comparison. Data extraction: A comprehensive Excel-based data extraction form was developed to capture all information required for the systematic review. This form was carefully structured to ensure clarity and consistency across all extracted variables. The extraction template included several categories, such as general study characteristics (author names, year of publication, country of origin, study design, b-values, the number of participating radiologists, and their level of experience). Population-related data encompassed the number of HCA and FNH lesions and their corresponding ADC values. Two reviewers independently extracted data from each eligible study to maintain the rigor and reliability of the process. The extracted information was entered into an Excel spreadsheet to preserve a standardized format for subsequent comparison and analysis. Following independent extraction, the datasets from both reviewers were cross-checked for accuracy. Any discrepancies identified during this reconciliation were initially addressed through discussion between the two reviewers. If consensus could not be reached, a third reviewer was consulted to adjudicate unresolved disagreements and establish the final data. Data synthesis: Statistical analyses were conducted using STATA (version 17.0) and MedCalc (version 23.2). A random-effects model was applied to combine results across studies, accounting for between-study variability. Sensitivity analyses were performed to examine the influence of individual studies on the overall findings, assessing the robustness of the results when particular studies or parameters were omitted. Publication bias was evaluated using formal statistical tests to detect whether studies with positive findings were more likely to be published. When publication bias was suspected, the trim-and-fill method was applied to provide adjusted pooled estimates. Subgroup analyses and meta-regression were undertaken to explore potential sources of heterogeneity and to identify factors that might modify the association between ADC values and diagnostic adequacy. Substantial statistical heterogeneity was defined as an I² statistic greater than 50%, indicating notable variation among study results. The certainty of the evidence was appraised using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach, which systematically rates the quality of evidence for each outcome based on considerations including risk of bias, inconsistency, indirectness, imprecision, and publication bias. A p-value of less than 0.05 was considered indicative of statistical significance. Results Search Results and Study Characteristics: Thirteen studies comprising a total of 524 lesions (195 HCA and 329 FNH) published between 2010 and 2024 met inclusion criteria for the meta-analysis [ 22 – 34 ]. The demographic data consistently indicated a predominantly female patient population (65–85% across various studies) with an average age range of 35–45 years aligning with the expected demographic presentation of these lesions. Geographically, two studies were conducted in Asia, nine in Europe, and two in the U.S. (Table 1 ). The study selection process is detailed in the PRISMA flowchart (Fig. 1). Table 1 Characteristics of studies included in the meta-analysis. Name Study type Country Field strength (T) b-value (s/mm²) Slice thickness Field of view Readers FNH lesions Adenoma lesions Sandrasegaran et al. (2009) Retrospective USA 1.5 T 400 6.2 mm 360–400 mm × 360–400 mm 1 6 3 Miller et al. (2010) Retrospective USA 1.5 T 500 6 mm 360–400 mm × 360–400 mm 4 (16,12,7,3 YOE) 43 9 Agnello et al. (2012) Retrospective France 1.5 T 600 6 mm 320–400 mm × 320–400 mm 2 (7 and 27 YOE) 54 36 Filipe et al. (2012) Retrospective Portugal 1.5 T 700 8 mm 380 × 380 mm × 380–380 mm NR 10 4 Doblas et al. (2013) Prospective France 1.5 T 500 4 mm 320 × 320 mm × 320–320 mm 2 (7 and 3 YOE) 20 9 Morelli et al. (2013) Retrospective Germany 1.5 T 800 3 mm 370 × 265 mm 379 × 265 mm 2 (> 10 YOE) 8 10 Parsai et al. (2014) Retrospective England 1.5 T 1000 5 mm 400–450 mm × 400–450 mm 2 (8 and 10 YOE) 8 15 Hennedige et al. (2015) Retrospective Singapore 1.5 T 500 5mm NR 2 + 2 (> 10 YOE + 8 and 2) 17 5 Budjan et al. (2017) Retrospective Germany 3 T 1000 4 mm 340 × 240 mm × 340 − 280 mm 2 (8 and 4 YOE) 3 4 Jerjir et al. (2017) Prospective Belgium 3 T 800 NR NR 1 (4 YOE) 13 8 Colagrande et al. (2018) Retrospective Italy 1.5 T 1000 5 mm 300–420 mm × 300–420 mm 2 (10 and 5 YOE) 22 5 Zarghampour et al. (2018) Retrospective China 1.5 T 1000 8 mm 320–400 mm × 320–400 mm 1 (> 2 YOE) 65 81 Rybczynska et al. (2024) Retrospective Poland 1.5 T 1500 NR NR 2 60 6 YOE = Years of Experience NR = Not Reported Risk of Bias Assessment: The risk of bias in the included studies was rigorously assessed using the QUADAS-2 technique, concentrating on four domains: Patient Selection, Index Test, Reference Standard, and Flow and Timing. Two independent evaluators performed assessments and reconciled conflicts by consensus or, if unresolved, through a third-party arbitrator. The majority of research effectively delineated patient selection and implemented explicit inclusion/exclusion criteria, hence assuring population representativeness. The index test (DWI/ADC measurement) exhibited inconsistent reporting concerning radiologist blinding and ROI placement, hence posing a potential risk for measurement bias. The reference standard was predominantly histology; nevertheless, the poor documentation of radiologists' blinding to pathological outcomes may have introduced bias. The flow and time between imaging and the reference standard were suitable in the majority of research; nevertheless, 11 out of 13 studies employed retrospective designs, hence elevating the risk of selection and information bias. Significant heterogeneity remained evident, even when categorized by risk of bias (I² >80% for both low-risk and high-risk categories), and neither grouping achieved statistical significance for the mean ADC difference. This indicates that methodological quality alone does not completely account for the diversity in diagnostic performance found among trials (Fig. 2). Diagnostic Performance: The pooled ADC measurement for FNH was 1.44 × 10⁻³ mm²/s, with a 95% confidence interval (CI) of 1.34 to 1.55 × 10⁻³ mm²/s. In contrast, HCA lesions yielded a slightly lower mean ADC of 1.35 × 10⁻³ mm²/s (95% CI: 1.23 to 1.47 × 10⁻³ mm²/s). When directly comparing these values, the absolute difference was 0.10 × 10⁻³ mm²/s. This difference approached but did not reach conventional statistical significance (p = 0.073), and the 95% CI ranged from − 0.01 to 0.20 × 10⁻³ mm²/s. The observed between-study heterogeneity for this analysis was high (I² = 84.5%) (Fig. 3A). Standardizing the difference by the pooled standard deviation produced a standardized mean difference of 0.49 (95% CI: -0.02 to 0.99). In relative terms, FNH lesions demonstrated a 7.63% higher ADC value compared to HCA lesions, although this proportional difference failed to achieve statistical significance (p = 0.109; 95% CI: − 1.70% to + 16.95%) (Fig. 3B). Clinically, a 7–8% elevation in ADC might be perceptible on high-resolution diffusion‐weighted sequences, but such a modest difference is dwarfed by overlapping diffusion characteristics of individual lesions and is unlikely to serve as a reliable diagnostic cutoff in isolation. The leave-one-out sensitivity analysis demonstrated that no single study exerted an undue influence on the overall pooled estimate, indicating that the meta-analytic results were not driven by any particular investigation or outlier study. This finding enhances confidence in the stability of the effect size estimate and suggests that the conclusions are not dependent on the inclusion or exclusion of any specific study within the dataset. When studies were stratified by methodological design, notable differences in heterogeneity patterns emerged. Retrospective studies maintained considerable heterogeneity (I² = 86.9%) with a mean difference of 0.08 × 10⁻³ mm²/s (95% CI: -0.03 to 0.20), indicating persistent variability in ADC measurements across retrospective investigations. In stark contrast, prospective studies achieved complete homogeneity (I² = 0.0%) with a larger mean difference of 0.21 × 10⁻³ mm²/s (95% CI: -0.02 to 0.43). This resolution of heterogeneity in prospective studies suggests that standardized, protocol-driven data collection may significantly reduce between-study variability, although the confidence interval still crosses zero, indicating non-significant differences. When studies were categorized according to their overall risk of bias assessment using QUADAS-2 criteria, both low-risk studies (I² = 82.4%, MD = 0.09, 95% CI: -0.06 to 0.24) and high-risk studies (I² = 87.7%, MD = 0.00, 95% CI: -0.10 to 0.30) maintained substantial heterogeneity. However, neither reached statistical significance. This pattern suggests that methodological quality, as assessed by traditional bias assessment tools, may not be the primary driver of the observed heterogeneity in ADC measurements. Formal evaluation of publication bias using Egger's regression test yielded a non-significant result (p = 0.496), suggesting no statistically detectable small-study effects or asymmetry in the funnel plot distribution. However, the trim-and-fill adjustment resulted in a reduction of the pooled mean difference from 0.10 × 10⁻³ mm²/s to 0.08 × 10⁻³ mm²/s, representing a 20% attenuation of the effect estimate. More importantly, this adjustment shifted the statistical significance from borderline (p = 0.073) to clearly non-significant (p = 0.152), moving the result further from the conventional alpha threshold. This change suggests that even minimal publication bias, if present, could eliminate the marginal statistical trend observed in the primary analysis. Certainty of Evidence: The certainty of evidence for ADC-based differentiation between HCA and FNH was assessed using the GRADE framework, which evaluates risk of bias, inconsistency, indirectness, imprecision, and publication bias. The risk of bias is moderate due to the prevalence of retrospective designs and incomplete reporting of blinding procedures. Inconsistency is high, attributed to significant heterogeneity between studies (I² reaching 84.5% overall), alongside considerable methodological and technical variability. Indirectness is low; studies directly compared ADC in HCA and FNH using histopathological confirmation. Imprecision is moderate, indicated by overlapping confidence intervals for mean and standardized mean differences, with relative ADC differences failing to achieve statistical significance. Publication bias was evaluated through Egger's test and trim-and-fill adjustment, revealing no significant asymmetry in the funnel plot. The adjusted effect estimates decreased in magnitude but continued to lack statistical significance. The evidence for a clinically significant difference in ADC values between HCA and FNH is characterized by low to moderate certainty, largely attributable to ongoing heterogeneity and the absence of statistically significant results, despite the application of rigorous methodology and thorough literature synthesis (Table 2 ). Table 2 GRADE assessment of the certainty of evidence. Outcome Number of studies Number of lesions included Results Risk of bias Applicability and indirectness Inconsistency Imprecision Publication bias Strength of effect size certainty ADC difference: HCA vs. FNH (overall) 13 524 Figure 3 Moderate Low risk Seriously inconsistent Imprecise Low risk Weak strength Low ⊕⊕ ADC difference: Retrospective studies 11 470 Figure 3 Moderate Low risk Seriously inconsistent Imprecise Low risk Weak strength Low ⊕⊕ ADC difference: Prospective studies 2 54 Figure 3 Low risk Low risk Consistent Imprecise Weak strength Moderate ⊕⊕⊕ ADC difference: Low risk of bias studies 6 180 Figure 3 Low risk Low risk Seriously inconsistent Imprecise Low risk Weak strength Low ⊕⊕ ADC difference: High risk of bias studies 7 344 Figure 3 High risk Low risk Seriously inconsistent Imprecise Low risk Weak strength Very Low ⊕ Discussion The present systematic review and meta-analysis, encompassing thirteen studies and 524 focal liver lesions (195 HCA and 329 FNH), reveals that DWI with ADC quantification provides only modest discrimination between these two entities. Although FNH lesions exhibited slightly higher pooled ADC values (1.44 × 10⁻³ mm²/s) compared with HCA (1.35 × 10⁻³ mm²/s), the absolute difference of 0.10 × 10⁻³ mm²/s failed to achieve statistical significance (p = 0.073), and the proportional difference of 7.6% likewise lacked significance (p = 0.109), the magnitudes unlikely to support ADC role in clinical discrimination given overlapping confidence intervals. While ADC quantification offers a rapid, contrast-free metric of tissue cellularity and microstructure, its discriminatory power between HCA and FNH is inherently limited by overlapping diffusion profiles. From a clinical perspective, the marginal ADC disparity observed here aligns with emergent literature indicating substantial overlap of ADC values across solid hepatocellular lesions. A meta-analysis of sixteen studies demonstrated only moderate diagnostic accuracy of ADC for differentiating malignant from benign hepatic lesions, noting considerable overlap in ADC distributions among solid benign entities such as HCA and FNH. In particular, mean ADC values of HCA and FNH closely approximated those of malignant tumors (e.g., hepatocellular carcinoma). This overlap likely reflects shared microstructural features including similar cellularity and extracellular matrix composition across solid benign and malignant lesions, which diminish diffusion differences [ 35 ]. Early optimism regarding DWI for liver lesion characterization stemmed from studies reporting significant ADC differences between malignant and benign lesions; however, many of these included cystic or hemangiomatous lesions, which artificially inflate effect sizes due to their markedly different diffusion properties. When analysis is confined to solid benign lesions, such as HCA vs. FNH, the discriminatory power of ADC diminishes substantially. The pooled AUC of ADC for distinguishing solid malignant from benign lesions was only 0.82 (95% CI: 0.78–0.85), with sensitivity and specificity around 78% and 74%, respectively. In included studies, head-to-head comparisons of HCA versus FNH, reported ADC thresholds varied substantially but none achieved both high sensitivity and specificity across independent cohorts, reflecting methodological inconsistencies in b-value selection, region-of-interest (ROI) placement, and post-processing algorithms [ 35 ]. In addition, assessment of statistical heterogeneity in this analysis underscores the impact of study design on pooled estimates. Retrospective studies demonstrated substantial heterogeneity (I² = 86.9%) and a smaller mean ADC difference of 0.08 × 10⁻³ mm²/s, whereas prospective studies achieved complete homogeneity (I² = 0%) with a larger but still non-significant mean difference of 0.21 × 10⁻³ mm²/s. This contrast suggests that standardized, protocol‐driven prospective data collection may reduce variability in ADC measurements, highlighting the methodological importance of harmonized imaging acquisition and analysis methods. The limited diagnostic yield of ADC quantification alone mandates that DWI be considered a supplementary tool rather than a standalone modality for HCA versus FNH differentiation. Clinicians should prioritize hepatobiliary phase imaging with gadoxetic acid, reserving ADC measurements for research settings or as part of a broader multiparametric protocol. In resource-limited contexts where contrast agents are contraindicated or unavailable, ADC may contribute incremental information, but physicians must interpret values cautiously and in conjunction with lesion morphology, patient demographics, and clinical history. As future research direction, advances in quantitative image analysis namely texture analysis and radiomics offer promise to capture subtle signal heterogeneity beyond mean ADC or simple intensity thresholds. On gadoxetic acid, texture analysis parameters such as skewness and entropy on hepatobiliary phase images have demonstrated area under curve values of 0.87 to 0.98 for distinguishing HCA from FNH, with sensitivity up to 90.6% and specificity of 100% in select cohorts [ 36 ]. Radiomic features extracted from DWI and multiple sequences may further enhance classification, though external validation and standardization of feature extraction pipelines are required before clinical translation. In addition, convolutional neural networks (CNNs) and deep learning frameworks have emerged as powerful tools for liver lesion classification on multi-sequence MRI. DenseNet architectures trained on hepatocyte-specific contrast-enhanced images achieved AUCs of 0.91 to 1.00 for FNH differentiation and overall classification accuracies exceeding 94% for benign versus malignant lesions [ 37 ]. More comprehensive AI systems integrating detection, segmentation, and classification modules (e.g., LiAIDS) have demonstrated F1-scores of 0.94 for benign lesions, outperforming junior radiologists and matching senior experts in multicenter evaluations [ 38 ]. Preliminary AI models trained with specific scoring systems to isolate FNH achieved sensitivity of 76.9% and specificity of 96.6%, indicating potential for automated diagnosis in focused tasks, though larger datasets and inclusion of diverse lesion subtypes are necessary to bridge the performance gap with experienced radiologists [ 39 ]. Despite rigorous methodology and comprehensive synthesis, several limitations of this meta-analysis warrant consideration. First, the predominance of retrospective, single-center studies [ 40 ] limited methodological uniformity and increased susceptibility to selection bias. Retrospective designs frequently lacked predefined sampling frameworks or formal sample-size calculations, potentially underpowering analyses and inflating type II error when ADC differences were marginal. Second, ROI placement methodology varied markedly across studies: some investigators employed small circular ROIs positioned in presumed most homogeneous lesion regions, while others delineated whole-lesion volumes. Such differences in sampling strategy, combined with inconsistent exclusion of areas of hemorrhage, necrosis, or fatty infiltration, may have biased ADC estimates and inflated interstudy variability. Third, the heterogeneity of lesion subtypes, particularly the molecularly defined HCA variants (HNF1α-inactivated, inflammatory, β-catenin–activated) was underexplored; studies rarely stratified lesions by subtype, precluding meaningful subgroup analysis of ADC performance in high-risk β-catenin lesions that may exhibit diffusion characteristics closer to HCC. Conclusion In this meta-analysis, DWI-derived ADC values demonstrated only a modest, non-significant difference between FNH and HCA. Prospective, multicenter studies with standardized DWI parameters are needed to clarify the true diagnostic value of diffusion metrics. Meanwhile, multiparametric MRI incorporating hepatocyte-specific contrast agents remains the non-invasive gold standard for differentiating HCA from FNH. Declarations All authors contributed significantly to this article. Funding: None Conflict of interest: None Animal study: N/A Acknowledgment: The authors acknowledge all contributors to this study. References Mounajjed T (2021) Hepatocellular Adenoma and Focal Nodular Hyperplasia. Clin Liver Dis (Hoboken) 17(4):244–248 Myers L, Ahn J (2020) Focal Nodular Hyperplasia and Hepatic Adenoma: Evaluation and Management. Clin Liver Dis 24(3):389–403 Kim TH, Woo S, Ebrahimzadeh S, McInnes MDF, Gerst SR, Do RK (2023) Hepatic Adenoma Subtypes on Hepatobiliary Phase of Gadoxetic Acid-Enhanced MRI: Systematic Review and Meta-Analysis. AJR Am J Roentgenol 220(1):28–38 Poetter-Lang S, Ba-Ssalamah A, Bastati N, Ba-Ssalamah SA, Hodge JC, Brancatelli G, Paradis V, Vilgrain V (2024) Hepatocellular adenoma update: diagnosis, molecular classification, and clinical course. Br J Radiol 97(1163):1740–1754 Aziz H, Underwood PW, Gosse MD, Afyouni S, Kamel I, Pawlik TM (2024) Hepatic adenoma: evolution of a more individualized treatment approach. J Gastrointest Surg 28(6):975–982 Klompenhouwer AJ, de Man RA, Dioguardi Burgio M, Vilgrain V, Zucman-Rossi J, Ijzermans JNM (2020) New insights in the management of Hepatocellular Adenoma. Liver Int 40(7):1529–1537 van Aalten SM, de Man RA, JN IJ, Terkivatan T (2012) Systematic review of haemorrhage and rupture of hepatocellular adenomas. Br J Surg 99(7):911–916 Bieze M, Phoa SS, Verheij J, van Lienden KP, van Gulik TM (2014) Risk factors for bleeding in hepatocellular adenoma. Br J Surg 101(7):847–855 Cacciatori FA, Rodrigues PD, Fontes PRO (2023) Bleeding hepatocellular adenoma: historical series and outcomes. Rev Col Bras Cir 50:e20233549 Thomeer MG, Broker M, Verheij J, Doukas M, Terkivatan T, Bijdevaate D, De Man RA, Moelker A, JN IJ (2016) Hepatocellular adenoma: when and how to treat? Update of current evidence. Th Adv Gastroenterol 9(6):898–912 Stoot JH, Coelen RJ, De Jong MC, Dejong CH (2010) Malignant transformation of hepatocellular adenomas into hepatocellular carcinomas: a systematic review including more than 1600 adenoma cases. HPB (Oxford) 12(8):509–522 González IA, Torbenson M, Sharifai N, Byrnes K, Chatterjee D, Kakar S, Yeh MM, Wu TT, Zhang X, Jain D (2023) Clinicopathologic characterization of hepatocellular adenomas in men: a multicenter experience. Hum Pathol 138:24–33 Perrakis A, Vassos N, Grützmann R, Croner RS (2017) What is Changing in Indications and Treatment of Focal Nodular Hyperplasia of the Liver. Is There Any Place for Surgery? Ann Hepatol 16(3):333–341 Kim JV, Wu GY (2024) Focal Nodular Hyperplasia: A Comprehensive Review with a Particular Focus on Pathogenesis and Complications. J Clin Transl Hepatol 12(2):182–190 Virgilio E, Cavallini M (2018) Managing Focal Nodular Hyperplasia of the Liver: Surgery or Minimally-invasive Approaches? A Review of the Preferable Treatment Options. Anticancer Res 38(1):33–36 Balabaud C, Al-Rabih WR, Chen PJ, Evason K, Ferrell L, Hernandez-Prera JC, Huang SF, Longerich T, Park YN, Quaglia A, Schirmacher P, Sempoux C, Thung SN, Torbenson M, Wee A, Yeh MM, Yeh SH, Le Bail B, Zucman-Rossi J (2013) P. Bioulac-Sage, Focal Nodular Hyperplasia and Hepatocellular Adenoma around the World Viewed through the Scope of the Immunopathological Classification, Int J Hepatol (2013) 268625 Frenette C, Mendiratta-Lala M, Salgia R, Wong RJ, Sauer BG, Pillai A (2024) ACG Clinical Guideline: Focal Liver Lesions. Am J Gastroenterol 119(7):1235–1271 Vijay A, Elaffandi A, Khalaf H (2015) Hepatocellular adenoma: An update. World J Hepatol 7(25):2603–2609 Venturi A, Piscaglia F, Vidili G, Flori S, Righini R, Golfieri R, Bolondi L (2007) Diagnosis and management of hepatic focal nodular hyperplasia. J Ultrasound 10(3):116–127 Agarwal S, Fuentes-Orrego JM, Arnason T, Misdraji J, Jhaveri KS, Harisinghani M, Hahn PF (2014) Inflammatory hepatocellular adenomas can mimic focal nodular hyperplasia on gadoxetic acid-enhanced MRI. AJR Am J Roentgenol 203(4):W408–W414 Guo Y, Li W, Cai W, Zhang Y, Fang Y, Hong G (2017) Diagnostic Value of Gadoxetic Acid-Enhanced MR Imaging to Distinguish HCA and Its Subtype from FNH: A Systematic Review. Int J Med Sci 14(7):668–674 Agnello F, Ronot M, Valla DC, Sinkus R, Van Beers BE, Vilgrain V (2012) High-b-value diffusion-weighted MR imaging of benign hepatocellular lesions: quantitative and qualitative analysis. Radiology 262(2):511–519 Budjan J, Sauter EA, Zoellner FG, Lemke A, Wambsganss J, Schoenberg SO, Attenberger UI (2018) Diffusion kurtosis imaging of the liver at 3 Tesla: in vivo comparison to standard diffusion-weighted imaging. Acta Radiol 59(1):18–25 Colagrande S, Calistri L, Grazzini G, Nardi C, Busoni S, Morana G, Grazioli L (2018) MRI features of primary hepatic lymphoma. Abdom Radiol (NY) 43(9):2277–2287 Doblas S, Wagner M, Leitao H, Daire J-L, Sinkus R, Vilgrain V, Beers B (2013) Determination of Malignancy and Characterization of Hepatic Tumor Type With Diffusion-Weighted Magnetic Resonance Imaging Comparison of Apparent Diffusion Coefficient and Intravoxel Incoherent Motion-Derived Measurements. Invest Radiol 48:722–728 Filipe JP, Curvo-Semedo L, Casalta-Lopes J, Marques MC, Caseiro-Alves F (2013) Diffusion-weighted imaging of the liver: usefulness of ADC values in the differential diagnosis of focal lesions and effect of ROI methods on ADC measurements. Magma 26(3):303–312 Hennedige TP, Hallinan JT, Leung FP, Teo LL, Iyer S, Wang G, Chang S, Madhavan KK, Wee A, Venkatesh SK (2016) Comparison of magnetic resonance elastography and diffusion-weighted imaging for differentiating benign and malignant liver lesions. Eur Radiol 26(2):398–406 Jerjir N, Bruyneel L, Haspeslagh M, Quenet S, Coenegrachts K (2017) Intravoxel incoherent motion and dynamic contrast-enhanced MRI for differentiation between hepatocellular adenoma and focal nodular hyperplasia. Br J Radiol 90(1076):20170007 Miller FH, Hammond N, Siddiqi AJ, Shroff S, Khatri G, Wang Y, Merrick LB, Nikolaidis P (2010) Utility of diffusion-weighted MRI in distinguishing benign and malignant hepatic lesions. J Magn Reson Imaging 32(1):138–147 Morelli JN, Michaely HJ, Meyer MM, Rustemeyer T, Schoenberg SO, Attenberger UI (2013) Comparison of Dynamic and Liver-Specific Gadoxetic Acid Contrast-Enhanced MRI versus Apparent Diffusion Coefficients. PLoS ONE 8(6):e61898 Parsai A, Zerizer I, Roche O, Gkoutzios P, Miquel ME (2015) Assessment of diffusion-weighted imaging for characterizing focal liver lesions. Clin Imaging 39(2):278–284 Rybczynska DN, Markiet KE, Pienkowska JM, Szurowska E, Frydrychowski A (2024) To assess the quantitative features of focal liver lesions in gadoxetic acid enhanced MRI and to determine whether these features can accurately differentiate benign form malignant lesions. Eur J Radiol 171:111288 Sandrasegaran K, Akisik FM, Lin C, Tahir B, Rajan J, Aisen AM (2009) The value of diffusion-weighted imaging in characterizing focal liver masses. Acad Radiol 16(10):1208–1214 Zarghampour M, Fouladi DF, Pandey A, Ghasabeh MA, Pandey P, Varzaneh FN, Khoshpouri P, Shao N, Pan L, Grimm R, Kamel IR (2018) Utility of volumetric contrast-enhanced and diffusion-weighted MRI in differentiating between common primary hypervascular liver tumors. J Magn Reson Imaging 48(4):1080–1090 Nalaini F, Shahbazi F, Mousavinezhad SM, Ansari A, Salehi M (2021) Diagnostic accuracy of apparent diffusion coefficient (ADC) value in differentiating malignant from benign solid liver lesions: a systematic review and meta-analysis. Br J Radiol 94(1123):20210059 Cannella R, Rangaswamy B, Minervini MI, Borhani AA, Tsung A, Furlan A (2019) Value of Texture Analysis on Gadoxetic Acid-Enhanced MRI for Differentiating Hepatocellular Adenoma From Focal Nodular Hyperplasia. AJR Am J Roentgenol 212(3):538–546 Stollmayer R, Budai BK, Tóth A, Kalina I, Hartmann E, Szoldán P, Bérczi V, Maurovich-Horvat P, Kaposi PN (2021) Diagnosis of focal liver lesions with deep learning-based multi-channel analysis of hepatocyte-specific contrast-enhanced magnetic resonance imaging. World J Gastroenterol 27(35):5978–5988 Ying H, Liu X, Zhang M, Ren Y, Zhen S, Wang X, Liu B, Hu P, Duan L, Cai M, Jiang M, Cheng X, Gong X, Jiang H, Jiang J, Zheng J, Zhu K, Zhou W, Lu B, Zhou H, Shen Y, Du J, Ying M, Hong Q, Mo J, Li J, Ye G, Zhang S, Hu H, Sun J, Liu H, Li Y, Xu X, Bai H, Wang S, Cheng X, Xu X, Jiao L, Yu R, Lau WY, Yu Y, Cai X (2024) A multicenter clinical AI system study for detection and diagnosis of focal liver lesions. Nat Commun 15(1):1131 Kantarcı M, Kızılgöz V, Terzi R, Kılıç AE, Kabalcı H, Durmaz Ö, Tokgöz N, Harman M, Sağır Kahraman A, Avanaz A, Aydın S, Elpek G, Yazol M, Aydınlı B (2025) Evaluating artificial intelligence for a focal nodular hyperplasia diagnosis using magnetic resonance imaging: preliminary findings. Diagn Interv Radiol 31(5):405–415 Vossen JA, Buijs M, Liapi E, Eng J, Bluemke DA, Kamel IR (2008) Receiver operating characteristic analysis of diffusion-weighted magnetic resonance imaging in differentiating hepatic hemangioma from other hypervascular liver lesions. J Comput Assist Tomogr 32(5):750–756 Additional Declarations The authors declare no competing interests. Supplementary Files SupplementaryA.docx SupplementaryB.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8799129","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":586498901,"identity":"6de87bf9-c9bc-45c6-b712-860b7e4c0d55","order_by":0,"name":"Alisa Mohebbi","email":"","orcid":"","institution":"School of Medicine, Tehran University of Medical Sciences, Tehran, Iran.","correspondingAuthor":false,"prefix":"","firstName":"Alisa","middleName":"","lastName":"Mohebbi","suffix":""},{"id":586498902,"identity":"0a997f2f-30d9-4216-9cff-4f948108c6a6","order_by":1,"name":"Mehrad Zare","email":"","orcid":"","institution":"School of Medicine, Tehran University of Medical Sciences, Tehran, Iran.","correspondingAuthor":false,"prefix":"","firstName":"Mehrad","middleName":"","lastName":"Zare","suffix":""},{"id":586498903,"identity":"e849b8f6-aa01-4d6e-bea4-726e470ccb9c","order_by":2,"name":"Afshin Mohammadi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIiWNgGAWjYFACHjDJ2AbED+CCCURqYTYgTUsDAwObBFHO0m3vPfjpZo6dbB8D87Nq3py6xAb2ww8YHu7BrcXszLlk6dxtycZtDGxmt3m3HU5s4EkzYEh4hkfLjRwDoBbmRKBfQFoOJDYw5AD9cgCvFuPfudvqgVrYvxXzbgM6jP8NQS1mQFsOA7XwmDHzAq1rkCBky5kzZta5244btzHzFEvO3XbYuE3imcEBvFqO9xjfzt1WLTu/vX3jh7fb6mT7+ZMfPvyBRwsCMENpNiAmSsMoGAWjYBSMAtwAAGBoT5NR9fb8AAAAAElFTkSuQmCC","orcid":"","institution":"Radiology Department, Faculty of Medicine, Urmia University of Medical Sciences, Urmia, Iran","correspondingAuthor":true,"prefix":"","firstName":"Afshin","middleName":"","lastName":"Mohammadi","suffix":""}],"badges":[],"createdAt":"2026-02-05 16:05:05","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8799129/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8799129/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102216414,"identity":"6fc1069d-8c3c-4294-ac17-75a704f058d1","added_by":"auto","created_at":"2026-02-09 12:57:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":757599,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 1.\u003c/strong\u003e PRISMA 2020 flow diagram for study selection.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/ca20f275214db12b3c8691e1.png"},{"id":102216315,"identity":"92b97dec-b839-4288-b320-719fe7bb48e5","added_by":"auto","created_at":"2026-02-09 12:56:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":67541,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 2.\u003c/strong\u003e Quality assessment of included studies using the QUADAS-2 tool.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/3f2dcb6d2345d9a259ab25a0.png"},{"id":102216440,"identity":"038115c2-97c7-4580-a497-abe03fb08d10","added_by":"auto","created_at":"2026-02-09 12:57:10","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":70842,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 3. \u003c/strong\u003eForest plot comparing pooled mean differences in apparent diffusion coefficient (ADC) values between focal nodular hyperplasia (FNH) and hepatocellular adenoma (HCA) across all included studies.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/0e2945445a5da2d4f9acd582.png"},{"id":102216401,"identity":"7d41391b-e477-4dd2-ac62-9875929f1329","added_by":"auto","created_at":"2026-02-09 12:56:56","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":70017,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 4. Forest plot illustrating percentage difference in mean ADC values between FNH and HCA.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/497906cf339e591dcf046406.png"},{"id":102216640,"identity":"a13f538b-425e-47c1-91b8-edec986b1a03","added_by":"auto","created_at":"2026-02-09 12:58:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1484433,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/652f3626-be05-49ee-9637-13a04d8f99ef.pdf"},{"id":102216544,"identity":"f8a9b801-8f3d-4c6a-af38-eec49f795e4a","added_by":"auto","created_at":"2026-02-09 12:57:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":18099,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryA.docx","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/173866e006f4cd57db844bce.docx"},{"id":102216295,"identity":"a11d383d-94d0-468e-a844-aa714c157245","added_by":"auto","created_at":"2026-02-09 12:56:51","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":8182,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryB.docx","url":"https://assets-eu.researchsquare.com/files/rs-8799129/v1/4e9527be5d9f4361b5283e26.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eDiagnostic accuracy of ADC values as a supplementary tool in differentiating hepatocellular adenoma from focal nodular hyperplasia: a systematic review and meta-analysis\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDistinguishing between hepatocellular adenoma (HCA) and focal nodular hyperplasia (FNH) represents one of the most consequential diagnostic challenges in hepatic imaging due to their profoundly different clinical implications, risk profiles, and management paradigms [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. While both lesions may initially appear similar on routine imaging, their biological behavior diverges markedly, with critical consequences for patient outcomes [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. This diagnostic distinction becomes particularly crucial when considering atypical HCA subtypes, which present unique challenges and carry varying degrees of risk for hemorrhage and malignant transformation [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe clinical significance of accurate differentiation cannot be overstated. HCA, particularly certain genetic subtypes, poses substantial risks including spontaneous bleeding and malignant transformation to hepatocellular carcinoma [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. According to systematic reviews, the overall frequency of hemorrhage in HCA patients reaches 27.2%, with rupture and intraperitoneal bleeding occurring in 17.5% of cases [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Risk factors for bleeding are well-established and include lesion size of 5 cm or greater, location in the left hepatic lobe, and exophytic growth patterns [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Even more concerning is the potential for malignant transformation, which occurs in 4.2% of HCAs, with males exhibiting a higher risk compared to females [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Beta-catenin activated HCAs represent a particularly high-risk subtype, with up to 47% of males with these lesions developing hepatocellular carcinoma [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. In contrast, FNH represents a benign, polyclonal lesion with no risk of malignant transformation or spontaneous hemorrhage [\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. This fundamental difference necessitates entirely divergent management strategies. While FNH typically does not require follow-up, HCA management depends on size, subtype, and patient demographics, often necessitating surgical resection or intensive monitoring [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The mortality associated with emergency resection for ruptured HCA ranges from 5\u0026ndash;10%, compared to less than 1% for elective surgery [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCurrent imaging modalities face significant limitations in achieving this critical differentiation. Conventional ultrasonography, computed tomography, and standard magnetic resonance imaging demonstrate substantial overlap in enhancement patterns between HCA and FNH, particularly in smaller lesions or atypical presentations. The central scar, traditionally considered pathognomonic for FNH, is present in only 50% of cases and is frequently absent in lesions smaller than 3 cm [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Furthermore, HCA may exhibit features that mimic FNH, including similar enhancement patterns and the occasional presence of central scarring [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. In addition, the reliance on contrast-enhanced imaging presents additional challenges; gadolinium-based contrast agents, while improving diagnostic accuracy, are contraindicated in patients with renal impairment, during pregnancy, or in those with contrast allergies. Even with hepatocyte-specific contrast agents like gadoxetate disodium, which represent the current non-invasive gold standard, diagnostic overlap persists, particularly with certain HCA subtypes that may demonstrate iso- or hyperintense uptake during hepatobiliary phase imaging [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Moreover, the substantial cost implications and the need for specialized expertise limit widespread accessibility to these advanced contrast agents.\u003c/p\u003e \u003cp\u003eAdvanced magnetic resonance techniques, particularly diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) quantification, have generated considerable interest as potential non-invasive biomarkers for distinguishing HCA from FNH. Several individual investigations have reported encouraging results, demonstrating statistically significant differences in ADC values between these lesions and proposing threshold cutoffs with promising diagnostic accuracy. In contrast, some studies have failed to confirm these findings, instead highlighting substantial overlap in ADC measurements that precludes reliable discrimination in routine practice. The present systematic review and meta-analysis therefore seeks to resolve these discrepancies by rigorously synthesizing all available evidence on ADC values in HCA versus FNH. This work aims to clarify whether ADC measurements can offer reproducible, contrast-free differentiation of these two benign hepatocellular lesions or whether their diagnostic promise is undermined by overlapping diffusion profiles. Ultimately, the findings of this study will inform whether quantitative DWI adoption as a supplementary imaging tool in clinical algorithms is clinically beneficial for focal liver lesion characterization.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003eMethods adhered to the comprehensive recommendations of the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy. The conduct and reporting of this review were guided by the PRISMA-DTA and PRISMA-Search (PRISMA-S) statements for systematic reviews and meta-analyses of diagnostic accuracy. To enhance reproducibility, the study protocol was prospectively registered on the Open Science Framework (OSF) at (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://osf.io/vwq4f/\u003c/span\u003e\u003cspan address=\"https://osf.io/vwq4f/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) (See Supplementary A) before study initiation.\u003c/p\u003e \u003cp\u003eSearch strategy:\u003c/p\u003e \u003cp\u003eTwo reviewers independently designed the electronic search strategies using EMTREE and MeSH terms. They then jointly refined and harmonized their search syntax. Any discrepancies or disagreements were resolved through detailed discussion and, when necessary, consultation with a third reviewer to achieve consensus and optimize the search strategy. The final search was conducted on August 30, 2025, across multiple databases, including PubMed, Web of Science, Embase, and the Cochrane Library, without language restrictions to maximize inclusivity and minimize language bias. The full search strategy is provided in Supplementary B to promote transparency and reproducibility. In addition, the reference lists of all eligible studies identified in the database search were screened to further enhance the completeness of the literature search. All retrieved citations were subsequently imported into EndNote for management and de-duplication.\u003c/p\u003e \u003cp\u003eEligibility criteria:\u003c/p\u003e \u003cp\u003eStudies were eligible for inclusion only if they reported ADC on a minimum of five discrete lesions, thereby ensuring adequate lesion sampling to support meaningful statistical comparison. To maintain methodological homogeneity, studies employing alternative imaging modalities such as contrast-enhanced sequences, elastography, or radiotracer techniques were deliberately excluded. We further refined our selection to include exclusively those investigations that reported HCA and FNH groups separately, thereby avoiding analytical confounding introduced by aggregated benign lesion cohorts. Conversely, studies limited to a single lesion type (only HCA or FNH) were also excluded, as our principal aim was to perform direct, within‐study comparisons of ADC values between these two entities. In cases where multiple publications drew upon the same patient cohort, only the report presenting the most comprehensive dataset was retained, thereby eliminating the risk of patient or data duplication and preserving the integrity of effect estimates. Finally, to maximize literature capture and mitigate the risk of publication bias, the bibliographies of all included articles were screened for additional pertinent studies, ensuring that our synthesis encompassed the full scope of available evidence.\u003c/p\u003e \u003cp\u003eThe study selection process consisted of a detailed screening of titles and abstracts, followed by an independent full-text assessment of all potentially eligible articles by two reviewers. Any disagreements about study inclusion were addressed through in-depth discussion until a consensus was achieved. This collaborative procedure strengthened the methodological rigor and reliability of the study selection.\u003c/p\u003e \u003cp\u003eRisk of bias assessment:\u003c/p\u003e \u003cp\u003eWe conducted a rigorous study-level quality appraisal for all articles included in this systematic review and meta-analysis to enhance the robustness and trustworthiness of the findings. Two reviewers independently evaluated risk of bias using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool, a widely accepted and validated framework for appraising diagnostic accuracy research. QUADAS-2 examines four key domains that underpin the internal validity of diagnostic studies: (A) Patient selection: This domain assesses whether the enrolled participants are representative of the intended target population, thereby ensuring that the patients and lesions studied appropriately reflect the clinical context in which the diagnostic test is to be applied. (B) Index test: This domain focuses on the performance and interpretation of the index test\u0026mdash;here, DWI and ADC measurement\u0026mdash;evaluating whether they were conducted and interpreted in a standardized, unbiased, and reproducible manner. (C) Reference standard: This domain evaluates the adequacy and consistency of the histopathologic reference standard used, including its ability to correctly classify the target condition. In several studies, radiologist blinding to clinical and histopathological outcomes was incompletely reported, representing a possible source of bias. The assessment also considers whether the reference standard was applied uniformly across all participants. (D) Flow and timing: This domain examines the interval and sequence between the index test and the reference standard, ensuring that the time gap is appropriate and that any changes in the patient\u0026rsquo;s clinical status do not compromise the validity of the diagnostic comparison.\u003c/p\u003e \u003cp\u003eData extraction:\u003c/p\u003e \u003cp\u003eA comprehensive Excel-based data extraction form was developed to capture all information required for the systematic review. This form was carefully structured to ensure clarity and consistency across all extracted variables. The extraction template included several categories, such as general study characteristics (author names, year of publication, country of origin, study design, b-values, the number of participating radiologists, and their level of experience). Population-related data encompassed the number of HCA and FNH lesions and their corresponding ADC values.\u003c/p\u003e \u003cp\u003eTwo reviewers independently extracted data from each eligible study to maintain the rigor and reliability of the process. The extracted information was entered into an Excel spreadsheet to preserve a standardized format for subsequent comparison and analysis. Following independent extraction, the datasets from both reviewers were cross-checked for accuracy. Any discrepancies identified during this reconciliation were initially addressed through discussion between the two reviewers. If consensus could not be reached, a third reviewer was consulted to adjudicate unresolved disagreements and establish the final data.\u003c/p\u003e \u003cp\u003eData synthesis:\u003c/p\u003e \u003cp\u003eStatistical analyses were conducted using STATA (version 17.0) and MedCalc (version 23.2). A random-effects model was applied to combine results across studies, accounting for between-study variability. Sensitivity analyses were performed to examine the influence of individual studies on the overall findings, assessing the robustness of the results when particular studies or parameters were omitted. Publication bias was evaluated using formal statistical tests to detect whether studies with positive findings were more likely to be published. When publication bias was suspected, the trim-and-fill method was applied to provide adjusted pooled estimates. Subgroup analyses and meta-regression were undertaken to explore potential sources of heterogeneity and to identify factors that might modify the association between ADC values and diagnostic adequacy. Substantial statistical heterogeneity was defined as an I\u0026sup2; statistic greater than 50%, indicating notable variation among study results. The certainty of the evidence was appraised using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach, which systematically rates the quality of evidence for each outcome based on considerations including risk of bias, inconsistency, indirectness, imprecision, and publication bias. A p-value of less than 0.05 was considered indicative of statistical significance.\u003c/p\u003e "},{"header":"Results","content":"\u003cp\u003eSearch Results and Study Characteristics:\u003c/p\u003e \u003cp\u003eThirteen studies comprising a total of 524 lesions (195 HCA and 329 FNH) published between 2010 and 2024 met inclusion criteria for the meta-analysis [\u003cspan additionalcitationids=\"CR23 CR24 CR25 CR26 CR27 CR28 CR29 CR30 CR31 CR32 CR33\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. The demographic data consistently indicated a predominantly female patient population (65\u0026ndash;85% across various studies) with an average age range of 35\u0026ndash;45 years aligning with the expected demographic presentation of these lesions. Geographically, two studies were conducted in Asia, nine in Europe, and two in the U.S. (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The study selection process is detailed in the PRISMA flowchart (Fig.\u0026nbsp;1).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of studies included in the meta-analysis.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eName\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStudy type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eField strength (T)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eb-value (s/mm\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSlice thickness\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eField of view\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eReaders\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eFNH lesions\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eAdenoma lesions\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSandrasegaran et al. (2009)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.2 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e360\u0026ndash;400 mm \u0026times;\u003c/p\u003e \u003cp\u003e360\u0026ndash;400 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMiller et al. (2010)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e360\u0026ndash;400 mm \u0026times;\u003c/p\u003e \u003cp\u003e360\u0026ndash;400 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4 (16,12,7,3 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAgnello et al. (2012)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFrance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e600\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e320\u0026ndash;400 mm \u0026times; 320\u0026ndash;400 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (7 and 27 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFilipe et al. (2012)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePortugal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e700\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e380 \u0026times; 380 mm \u0026times;\u003c/p\u003e \u003cp\u003e380\u0026ndash;380 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDoblas et al. (2013)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eProspective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFrance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e320 \u0026times; 320 mm \u0026times;\u003c/p\u003e \u003cp\u003e320\u0026ndash;320 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (7 and 3 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMorelli et al. (2013)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGermany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e370 \u0026times; 265 mm\u003c/p\u003e \u003cp\u003e379 \u0026times; 265\u0026nbsp; mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (\u0026gt;\u0026thinsp;10 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eParsai et al. (2014)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEngland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e400\u0026ndash;450 mm \u0026times; 400\u0026ndash;450 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (8 and 10 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHennedige et al. (2015)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSingapore\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2\u0026thinsp;+\u0026thinsp;2 (\u0026gt;\u0026thinsp;10 YOE\u0026thinsp;+\u0026thinsp;8 and 2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBudjan et al. (2017)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGermany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e340 \u0026times; 240 mm \u0026times;\u003c/p\u003e \u003cp\u003e340\u0026thinsp;\u0026minus;\u0026thinsp;280 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (8 and 4 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJerjir et al. (2017)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eProspective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBelgium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1 (4 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eColagrande et al. (2018)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eItaly\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e300\u0026ndash;420 mm \u0026times;\u003c/p\u003e \u003cp\u003e300\u0026ndash;420 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2 (10 and 5 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZarghampour et al. (2018)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e320\u0026ndash;400 mm \u0026times;\u003c/p\u003e \u003cp\u003e320\u0026ndash;400 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1 (\u0026gt;\u0026thinsp;2 YOE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e81\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRybczynska et al. (2024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrospective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePoland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"10\"\u003eYOE\u0026thinsp;=\u0026thinsp;Years of Experience\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"10\"\u003eNR\u0026thinsp;=\u0026thinsp;Not Reported\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eRisk of Bias Assessment:\u003c/p\u003e \u003cp\u003eThe risk of bias in the included studies was rigorously assessed using the QUADAS-2 technique, concentrating on four domains: Patient Selection, Index Test, Reference Standard, and Flow and Timing. Two independent evaluators performed assessments and reconciled conflicts by consensus or, if unresolved, through a third-party arbitrator. The majority of research effectively delineated patient selection and implemented explicit inclusion/exclusion criteria, hence assuring population representativeness. The index test (DWI/ADC measurement) exhibited inconsistent reporting concerning radiologist blinding and ROI placement, hence posing a potential risk for measurement bias. The reference standard was predominantly histology; nevertheless, the poor documentation of radiologists' blinding to pathological outcomes may have introduced bias. The flow and time between imaging and the reference standard were suitable in the majority of research; nevertheless, 11 out of 13 studies employed retrospective designs, hence elevating the risk of selection and information bias.\u003c/p\u003e \u003cp\u003eSignificant heterogeneity remained evident, even when categorized by risk of bias (I\u0026sup2; \u0026gt;80% for both low-risk and high-risk categories), and neither grouping achieved statistical significance for the mean ADC difference. This indicates that methodological quality alone does not completely account for the diversity in diagnostic performance found among trials (Fig.\u0026nbsp;2).\u003c/p\u003e \u003cp\u003eDiagnostic Performance:\u003c/p\u003e \u003cp\u003eThe pooled ADC measurement for FNH was 1.44 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s, with a 95% confidence interval (CI) of 1.34 to 1.55 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s. In contrast, HCA lesions yielded a slightly lower mean ADC of 1.35 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s (95% CI: 1.23 to 1.47 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s). When directly comparing these values, the absolute difference was 0.10 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s. This difference approached but did not reach conventional statistical significance (p\u0026thinsp;=\u0026thinsp;0.073), and the 95% CI ranged from \u0026minus;\u0026thinsp;0.01 to 0.20 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s. The observed between-study heterogeneity for this analysis was high (I\u0026sup2; = 84.5%) (Fig.\u0026nbsp;3A).\u003c/p\u003e \u003cp\u003eStandardizing the difference by the pooled standard deviation produced a standardized mean difference of 0.49 (95% CI: -0.02 to 0.99). In relative terms, FNH lesions demonstrated a 7.63% higher ADC value compared to HCA lesions, although this proportional difference failed to achieve statistical significance (p\u0026thinsp;=\u0026thinsp;0.109; 95% CI: \u0026minus;\u0026thinsp;1.70% to +\u0026thinsp;16.95%) (Fig.\u0026nbsp;3B). Clinically, a 7\u0026ndash;8% elevation in ADC might be perceptible on high-resolution diffusion‐weighted sequences, but such a modest difference is dwarfed by overlapping diffusion characteristics of individual lesions and is unlikely to serve as a reliable diagnostic cutoff in isolation.\u003c/p\u003e \u003cp\u003eThe leave-one-out sensitivity analysis demonstrated that no single study exerted an undue influence on the overall pooled estimate, indicating that the meta-analytic results were not driven by any particular investigation or outlier study. This finding enhances confidence in the stability of the effect size estimate and suggests that the conclusions are not dependent on the inclusion or exclusion of any specific study within the dataset.\u003c/p\u003e \u003cp\u003eWhen studies were stratified by methodological design, notable differences in heterogeneity patterns emerged. Retrospective studies maintained considerable heterogeneity (I\u0026sup2; = 86.9%) with a mean difference of 0.08 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s (95% CI: -0.03 to 0.20), indicating persistent variability in ADC measurements across retrospective investigations. In stark contrast, prospective studies achieved complete homogeneity (I\u0026sup2; = 0.0%) with a larger mean difference of 0.21 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s (95% CI: -0.02 to 0.43). This resolution of heterogeneity in prospective studies suggests that standardized, protocol-driven data collection may significantly reduce between-study variability, although the confidence interval still crosses zero, indicating non-significant differences.\u003c/p\u003e \u003cp\u003eWhen studies were categorized according to their overall risk of bias assessment using QUADAS-2 criteria, both low-risk studies (I\u0026sup2; = 82.4%, MD\u0026thinsp;=\u0026thinsp;0.09, 95% CI: -0.06 to 0.24) and high-risk studies (I\u0026sup2; = 87.7%, MD\u0026thinsp;=\u0026thinsp;0.00, 95% CI: -0.10 to 0.30) maintained substantial heterogeneity. However, neither reached statistical significance. This pattern suggests that methodological quality, as assessed by traditional bias assessment tools, may not be the primary driver of the observed heterogeneity in ADC measurements.\u003c/p\u003e \u003cp\u003eFormal evaluation of publication bias using Egger's regression test yielded a non-significant result (p\u0026thinsp;=\u0026thinsp;0.496), suggesting no statistically detectable small-study effects or asymmetry in the funnel plot distribution. However, the trim-and-fill adjustment resulted in a reduction of the pooled mean difference from 0.10 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s to 0.08 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s, representing a 20% attenuation of the effect estimate. More importantly, this adjustment shifted the statistical significance from borderline (p\u0026thinsp;=\u0026thinsp;0.073) to clearly non-significant (p\u0026thinsp;=\u0026thinsp;0.152), moving the result further from the conventional alpha threshold. This change suggests that even minimal publication bias, if present, could eliminate the marginal statistical trend observed in the primary analysis.\u003c/p\u003e \u003cp\u003eCertainty of Evidence:\u003c/p\u003e \u003cp\u003eThe certainty of evidence for ADC-based differentiation between HCA and FNH was assessed using the GRADE framework, which evaluates risk of bias, inconsistency, indirectness, imprecision, and publication bias. The risk of bias is moderate due to the prevalence of retrospective designs and incomplete reporting of blinding procedures. Inconsistency is high, attributed to significant heterogeneity between studies (I\u0026sup2; reaching 84.5% overall), alongside considerable methodological and technical variability. Indirectness is low; studies directly compared ADC in HCA and FNH using histopathological confirmation. Imprecision is moderate, indicated by overlapping confidence intervals for mean and standardized mean differences, with relative ADC differences failing to achieve statistical significance. Publication bias was evaluated through Egger's test and trim-and-fill adjustment, revealing no significant asymmetry in the funnel plot. The adjusted effect estimates decreased in magnitude but continued to lack statistical significance.\u003c/p\u003e \u003cp\u003eThe evidence for a clinically significant difference in ADC values between HCA and FNH is characterized by low to moderate certainty, largely attributable to ongoing heterogeneity and the absence of statistically significant results, despite the application of rigorous methodology and thorough literature synthesis (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eGRADE assessment of the certainty of evidence.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"11\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of studies\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNumber of lesions included\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eResults\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRisk of bias\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eApplicability and indirectness\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInconsistency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecision\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePublication bias\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eStrength of effect size\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003ecertainty\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC difference: HCA vs. FNH (overall)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e524\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFigure 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSeriously inconsistent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecise\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eWeak strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eLow \u0026oplus;\u0026oplus;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC difference: Retrospective studies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e470\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFigure 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSeriously inconsistent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecise\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eWeak strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eLow \u0026oplus;\u0026oplus;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC difference: Prospective studies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFigure 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eConsistent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecise\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eWeak strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eModerate \u0026oplus;\u0026oplus;\u0026oplus;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC difference: Low risk of bias studies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFigure 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSeriously inconsistent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecise\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eWeak strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eLow \u0026oplus;\u0026oplus;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC difference: High risk of bias studies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e344\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFigure 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHigh risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSeriously inconsistent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprecise\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLow risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eWeak strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eVery Low \u0026oplus;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe present systematic review and meta-analysis, encompassing thirteen studies and 524 focal liver lesions (195 HCA and 329 FNH), reveals that DWI with ADC quantification provides only modest discrimination between these two entities. Although FNH lesions exhibited slightly higher pooled ADC values (1.44 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s) compared with HCA (1.35 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s), the absolute difference of 0.10 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s failed to achieve statistical significance (p\u0026thinsp;=\u0026thinsp;0.073), and the proportional difference of 7.6% likewise lacked significance (p\u0026thinsp;=\u0026thinsp;0.109), the magnitudes unlikely to support ADC role in clinical discrimination given overlapping confidence intervals.\u003c/p\u003e \u003cp\u003eWhile ADC quantification offers a rapid, contrast-free metric of tissue cellularity and microstructure, its discriminatory power between HCA and FNH is inherently limited by overlapping diffusion profiles. From a clinical perspective, the marginal ADC disparity observed here aligns with emergent literature indicating substantial overlap of ADC values across solid hepatocellular lesions. A meta-analysis of sixteen studies demonstrated only moderate diagnostic accuracy of ADC for differentiating malignant from benign hepatic lesions, noting considerable overlap in ADC distributions among solid benign entities such as HCA and FNH. In particular, mean ADC values of HCA and FNH closely approximated those of malignant tumors (e.g., hepatocellular carcinoma). This overlap likely reflects shared microstructural features including similar cellularity and extracellular matrix composition across solid benign and malignant lesions, which diminish diffusion differences [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eEarly optimism regarding DWI for liver lesion characterization stemmed from studies reporting significant ADC differences between malignant and benign lesions; however, many of these included cystic or hemangiomatous lesions, which artificially inflate effect sizes due to their markedly different diffusion properties. When analysis is confined to solid benign lesions, such as HCA vs. FNH, the discriminatory power of ADC diminishes substantially. The pooled AUC of ADC for distinguishing solid malignant from benign lesions was only 0.82 (95% CI: 0.78\u0026ndash;0.85), with sensitivity and specificity around 78% and 74%, respectively. In included studies, head-to-head comparisons of HCA versus FNH, reported ADC thresholds varied substantially but none achieved both high sensitivity and specificity across independent cohorts, reflecting methodological inconsistencies in b-value selection, region-of-interest (ROI) placement, and post-processing algorithms [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn addition, assessment of statistical heterogeneity in this analysis underscores the impact of study design on pooled estimates. Retrospective studies demonstrated substantial heterogeneity (I\u0026sup2; = 86.9%) and a smaller mean ADC difference of 0.08 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s, whereas prospective studies achieved complete homogeneity (I\u0026sup2; = 0%) with a larger but still non-significant mean difference of 0.21 \u0026times; 10⁻\u0026sup3; mm\u0026sup2;/s. This contrast suggests that standardized, protocol‐driven prospective data collection may reduce variability in ADC measurements, highlighting the methodological importance of harmonized imaging acquisition and analysis methods.\u003c/p\u003e \u003cp\u003eThe limited diagnostic yield of ADC quantification alone mandates that DWI be considered a supplementary tool rather than a standalone modality for HCA versus FNH differentiation. Clinicians should prioritize hepatobiliary phase imaging with gadoxetic acid, reserving ADC measurements for research settings or as part of a broader multiparametric protocol. In resource-limited contexts where contrast agents are contraindicated or unavailable, ADC may contribute incremental information, but physicians must interpret values cautiously and in conjunction with lesion morphology, patient demographics, and clinical history.\u003c/p\u003e \u003cp\u003eAs future research direction, advances in quantitative image analysis namely texture analysis and radiomics offer promise to capture subtle signal heterogeneity beyond mean ADC or simple intensity thresholds. On gadoxetic acid, texture analysis parameters such as skewness and entropy on hepatobiliary phase images have demonstrated area under curve values of 0.87 to 0.98 for distinguishing HCA from FNH, with sensitivity up to 90.6% and specificity of 100% in select cohorts [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Radiomic features extracted from DWI and multiple sequences may further enhance classification, though external validation and standardization of feature extraction pipelines are required before clinical translation. In addition, convolutional neural networks (CNNs) and deep learning frameworks have emerged as powerful tools for liver lesion classification on multi-sequence MRI. DenseNet architectures trained on hepatocyte-specific contrast-enhanced images achieved AUCs of 0.91 to 1.00 for FNH differentiation and overall classification accuracies exceeding 94% for benign versus malignant lesions [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. More comprehensive AI systems integrating detection, segmentation, and classification modules (e.g., LiAIDS) have demonstrated F1-scores of 0.94 for benign lesions, outperforming junior radiologists and matching senior experts in multicenter evaluations [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Preliminary AI models trained with specific scoring systems to isolate FNH achieved sensitivity of 76.9% and specificity of 96.6%, indicating potential for automated diagnosis in focused tasks, though larger datasets and inclusion of diverse lesion subtypes are necessary to bridge the performance gap with experienced radiologists [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eDespite rigorous methodology and comprehensive synthesis, several limitations of this meta-analysis warrant consideration. First, the predominance of retrospective, single-center studies [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e] limited methodological uniformity and increased susceptibility to selection bias. Retrospective designs frequently lacked predefined sampling frameworks or formal sample-size calculations, potentially underpowering analyses and inflating type II error when ADC differences were marginal. Second, ROI placement methodology varied markedly across studies: some investigators employed small circular ROIs positioned in presumed most homogeneous lesion regions, while others delineated whole-lesion volumes. Such differences in sampling strategy, combined with inconsistent exclusion of areas of hemorrhage, necrosis, or fatty infiltration, may have biased ADC estimates and inflated interstudy variability. Third, the heterogeneity of lesion subtypes, particularly the molecularly defined HCA variants (HNF1α-inactivated, inflammatory, β-catenin\u0026ndash;activated) was underexplored; studies rarely stratified lesions by subtype, precluding meaningful subgroup analysis of ADC performance in high-risk β-catenin lesions that may exhibit diffusion characteristics closer to HCC.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn this meta-analysis, DWI-derived ADC values demonstrated only a modest, non-significant difference between FNH and HCA. Prospective, multicenter studies with standardized DWI parameters are needed to clarify the true diagnostic value of diffusion metrics. Meanwhile, multiparametric MRI incorporating hepatocyte-specific contrast agents remains the non-invasive gold standard for differentiating HCA from FNH.\u003c/p\u003e"},{"header":"Declarations","content":" \u003cp\u003eAll authors contributed significantly to this article.\u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eNone\u003c/p\u003e \u003cp\u003eConflict of interest: None\u003c/p\u003e \u003cp\u003eAnimal study: N/A\u003c/p\u003e\u003ch2\u003eAcknowledgment:\u003c/h2\u003e \u003cp\u003eThe authors acknowledge all contributors to this study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMounajjed T (2021) Hepatocellular Adenoma and Focal Nodular Hyperplasia. Clin Liver Dis (Hoboken) 17(4):244\u0026ndash;248\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMyers L, Ahn J (2020) Focal Nodular Hyperplasia and Hepatic Adenoma: Evaluation and Management. Clin Liver Dis 24(3):389\u0026ndash;403\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim TH, Woo S, Ebrahimzadeh S, McInnes MDF, Gerst SR, Do RK (2023) Hepatic Adenoma Subtypes on Hepatobiliary Phase of Gadoxetic Acid-Enhanced MRI: Systematic Review and Meta-Analysis. AJR Am J Roentgenol 220(1):28\u0026ndash;38\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoetter-Lang S, Ba-Ssalamah A, Bastati N, Ba-Ssalamah SA, Hodge JC, Brancatelli G, Paradis V, Vilgrain V (2024) Hepatocellular adenoma update: diagnosis, molecular classification, and clinical course. Br J Radiol 97(1163):1740\u0026ndash;1754\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAziz H, Underwood PW, Gosse MD, Afyouni S, Kamel I, Pawlik TM (2024) Hepatic adenoma: evolution of a more individualized treatment approach. J Gastrointest Surg 28(6):975\u0026ndash;982\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlompenhouwer AJ, de Man RA, Dioguardi Burgio M, Vilgrain V, Zucman-Rossi J, Ijzermans JNM (2020) New insights in the management of Hepatocellular Adenoma. Liver Int 40(7):1529\u0026ndash;1537\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Aalten SM, de Man RA, JN IJ, Terkivatan T (2012) Systematic review of haemorrhage and rupture of hepatocellular adenomas. Br J Surg 99(7):911\u0026ndash;916\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBieze M, Phoa SS, Verheij J, van Lienden KP, van Gulik TM (2014) Risk factors for bleeding in hepatocellular adenoma. Br J Surg 101(7):847\u0026ndash;855\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCacciatori FA, Rodrigues PD, Fontes PRO (2023) Bleeding hepatocellular adenoma: historical series and outcomes. Rev Col Bras Cir 50:e20233549\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThomeer MG, Broker M, Verheij J, Doukas M, Terkivatan T, Bijdevaate D, De Man RA, Moelker A, JN IJ (2016) Hepatocellular adenoma: when and how to treat? Update of current evidence. Th Adv Gastroenterol 9(6):898\u0026ndash;912\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStoot JH, Coelen RJ, De Jong MC, Dejong CH (2010) Malignant transformation of hepatocellular adenomas into hepatocellular carcinomas: a systematic review including more than 1600 adenoma cases. HPB (Oxford) 12(8):509\u0026ndash;522\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonz\u0026aacute;lez IA, Torbenson M, Sharifai N, Byrnes K, Chatterjee D, Kakar S, Yeh MM, Wu TT, Zhang X, Jain D (2023) Clinicopathologic characterization of hepatocellular adenomas in men: a multicenter experience. Hum Pathol 138:24\u0026ndash;33\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerrakis A, Vassos N, Gr\u0026uuml;tzmann R, Croner RS (2017) What is Changing in Indications and Treatment of Focal Nodular Hyperplasia of the Liver. Is There Any Place for Surgery? Ann Hepatol 16(3):333\u0026ndash;341\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim JV, Wu GY (2024) Focal Nodular Hyperplasia: A Comprehensive Review with a Particular Focus on Pathogenesis and Complications. J Clin Transl Hepatol 12(2):182\u0026ndash;190\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVirgilio E, Cavallini M (2018) Managing Focal Nodular Hyperplasia of the Liver: Surgery or Minimally-invasive Approaches? A Review of the Preferable Treatment Options. Anticancer Res 38(1):33\u0026ndash;36\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBalabaud C, Al-Rabih WR, Chen PJ, Evason K, Ferrell L, Hernandez-Prera JC, Huang SF, Longerich T, Park YN, Quaglia A, Schirmacher P, Sempoux C, Thung SN, Torbenson M, Wee A, Yeh MM, Yeh SH, Le Bail B, Zucman-Rossi J (2013) P. Bioulac-Sage, Focal Nodular Hyperplasia and Hepatocellular Adenoma around the World Viewed through the Scope of the Immunopathological Classification, Int J Hepatol (2013) 268625\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrenette C, Mendiratta-Lala M, Salgia R, Wong RJ, Sauer BG, Pillai A (2024) ACG Clinical Guideline: Focal Liver Lesions. Am J Gastroenterol 119(7):1235\u0026ndash;1271\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVijay A, Elaffandi A, Khalaf H (2015) Hepatocellular adenoma: An update. World J Hepatol 7(25):2603\u0026ndash;2609\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVenturi A, Piscaglia F, Vidili G, Flori S, Righini R, Golfieri R, Bolondi L (2007) Diagnosis and management of hepatic focal nodular hyperplasia. J Ultrasound 10(3):116\u0026ndash;127\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAgarwal S, Fuentes-Orrego JM, Arnason T, Misdraji J, Jhaveri KS, Harisinghani M, Hahn PF (2014) Inflammatory hepatocellular adenomas can mimic focal nodular hyperplasia on gadoxetic acid-enhanced MRI. AJR Am J Roentgenol 203(4):W408\u0026ndash;W414\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo Y, Li W, Cai W, Zhang Y, Fang Y, Hong G (2017) Diagnostic Value of Gadoxetic Acid-Enhanced MR Imaging to Distinguish HCA and Its Subtype from FNH: A Systematic Review. Int J Med Sci 14(7):668\u0026ndash;674\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAgnello F, Ronot M, Valla DC, Sinkus R, Van Beers BE, Vilgrain V (2012) High-b-value diffusion-weighted MR imaging of benign hepatocellular lesions: quantitative and qualitative analysis. Radiology 262(2):511\u0026ndash;519\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBudjan J, Sauter EA, Zoellner FG, Lemke A, Wambsganss J, Schoenberg SO, Attenberger UI (2018) Diffusion kurtosis imaging of the liver at 3 Tesla: in vivo comparison to standard diffusion-weighted imaging. Acta Radiol 59(1):18\u0026ndash;25\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColagrande S, Calistri L, Grazzini G, Nardi C, Busoni S, Morana G, Grazioli L (2018) MRI features of primary hepatic lymphoma. Abdom Radiol (NY) 43(9):2277\u0026ndash;2287\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoblas S, Wagner M, Leitao H, Daire J-L, Sinkus R, Vilgrain V, Beers B (2013) Determination of Malignancy and Characterization of Hepatic Tumor Type With Diffusion-Weighted Magnetic Resonance Imaging Comparison of Apparent Diffusion Coefficient and Intravoxel Incoherent Motion-Derived Measurements. Invest Radiol 48:722\u0026ndash;728\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFilipe JP, Curvo-Semedo L, Casalta-Lopes J, Marques MC, Caseiro-Alves F (2013) Diffusion-weighted imaging of the liver: usefulness of ADC values in the differential diagnosis of focal lesions and effect of ROI methods on ADC measurements. Magma 26(3):303\u0026ndash;312\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHennedige TP, Hallinan JT, Leung FP, Teo LL, Iyer S, Wang G, Chang S, Madhavan KK, Wee A, Venkatesh SK (2016) Comparison of magnetic resonance elastography and diffusion-weighted imaging for differentiating benign and malignant liver lesions. Eur Radiol 26(2):398\u0026ndash;406\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJerjir N, Bruyneel L, Haspeslagh M, Quenet S, Coenegrachts K (2017) Intravoxel incoherent motion and dynamic contrast-enhanced MRI for differentiation between hepatocellular adenoma and focal nodular hyperplasia. Br J Radiol 90(1076):20170007\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiller FH, Hammond N, Siddiqi AJ, Shroff S, Khatri G, Wang Y, Merrick LB, Nikolaidis P (2010) Utility of diffusion-weighted MRI in distinguishing benign and malignant hepatic lesions. J Magn Reson Imaging 32(1):138\u0026ndash;147\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorelli JN, Michaely HJ, Meyer MM, Rustemeyer T, Schoenberg SO, Attenberger UI (2013) Comparison of Dynamic and Liver-Specific Gadoxetic Acid Contrast-Enhanced MRI versus Apparent Diffusion Coefficients. PLoS ONE 8(6):e61898\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParsai A, Zerizer I, Roche O, Gkoutzios P, Miquel ME (2015) Assessment of diffusion-weighted imaging for characterizing focal liver lesions. Clin Imaging 39(2):278\u0026ndash;284\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRybczynska DN, Markiet KE, Pienkowska JM, Szurowska E, Frydrychowski A (2024) To assess the quantitative features of focal liver lesions in gadoxetic acid enhanced MRI and to determine whether these features can accurately differentiate benign form malignant lesions. Eur J Radiol 171:111288\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSandrasegaran K, Akisik FM, Lin C, Tahir B, Rajan J, Aisen AM (2009) The value of diffusion-weighted imaging in characterizing focal liver masses. Acad Radiol 16(10):1208\u0026ndash;1214\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZarghampour M, Fouladi DF, Pandey A, Ghasabeh MA, Pandey P, Varzaneh FN, Khoshpouri P, Shao N, Pan L, Grimm R, Kamel IR (2018) Utility of volumetric contrast-enhanced and diffusion-weighted MRI in differentiating between common primary hypervascular liver tumors. J Magn Reson Imaging 48(4):1080\u0026ndash;1090\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNalaini F, Shahbazi F, Mousavinezhad SM, Ansari A, Salehi M (2021) Diagnostic accuracy of apparent diffusion coefficient (ADC) value in differentiating malignant from benign solid liver lesions: a systematic review and meta-analysis. Br J Radiol 94(1123):20210059\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCannella R, Rangaswamy B, Minervini MI, Borhani AA, Tsung A, Furlan A (2019) Value of Texture Analysis on Gadoxetic Acid-Enhanced MRI for Differentiating Hepatocellular Adenoma From Focal Nodular Hyperplasia. AJR Am J Roentgenol 212(3):538\u0026ndash;546\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStollmayer R, Budai BK, T\u0026oacute;th A, Kalina I, Hartmann E, Szold\u0026aacute;n P, B\u0026eacute;rczi V, Maurovich-Horvat P, Kaposi PN (2021) Diagnosis of focal liver lesions with deep learning-based multi-channel analysis of hepatocyte-specific contrast-enhanced magnetic resonance imaging. World J Gastroenterol 27(35):5978\u0026ndash;5988\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYing H, Liu X, Zhang M, Ren Y, Zhen S, Wang X, Liu B, Hu P, Duan L, Cai M, Jiang M, Cheng X, Gong X, Jiang H, Jiang J, Zheng J, Zhu K, Zhou W, Lu B, Zhou H, Shen Y, Du J, Ying M, Hong Q, Mo J, Li J, Ye G, Zhang S, Hu H, Sun J, Liu H, Li Y, Xu X, Bai H, Wang S, Cheng X, Xu X, Jiao L, Yu R, Lau WY, Yu Y, Cai X (2024) A multicenter clinical AI system study for detection and diagnosis of focal liver lesions. Nat Commun 15(1):1131\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKantarcı M, Kızılg\u0026ouml;z V, Terzi R, Kılı\u0026ccedil; AE, Kabalcı H, Durmaz \u0026Ouml;, Tokg\u0026ouml;z N, Harman M, Sağır Kahraman A, Avanaz A, Aydın S, Elpek G, Yazol M, Aydınlı B (2025) Evaluating artificial intelligence for a focal nodular hyperplasia diagnosis using magnetic resonance imaging: preliminary findings. Diagn Interv Radiol 31(5):405\u0026ndash;415\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVossen JA, Buijs M, Liapi E, Eng J, Bluemke DA, Kamel IR (2008) Receiver operating characteristic analysis of diffusion-weighted magnetic resonance imaging in differentiating hepatic hemangioma from other hypervascular liver lesions. J Comput Assist Tomogr 32(5):750\u0026ndash;756\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Tehran University of Medical Sciences","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Focal nodular hyperplasia (FNH), Apparent diffusion coefficient (ADC), Diffusion-weighted imaging (DWI), Hepatocellular adenoma (HCA)","lastPublishedDoi":"10.21203/rs.3.rs-8799129/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8799129/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAccurate differentiation between hepatocellular adenoma (HCA) and focal nodular hyperplasia (FNH) is essential due to divergent management and risk profiles. Diffusion-weighted MRI (DWI) with apparent diffusion coefficient (ADC) quantification has been proposed as a non-contrast method to distinguish these lesions.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA systematic search of PubMed, Embase, Web of Science, and Cochrane Library was performed through August 2025 following PRISMA-DTA guidelines. Studies comparing ADC values in HCA and FNH, reporting measurements for at least five lesions per group, were included. Quality assessment used the QUADAS-2 tool. Random-effects meta-analysis pooled mean ADC values, calculated standardized mean differences (SMD), and assessed heterogeneity (I\u0026sup2;). Sensitivity analyses, subgroup analyses by study design, and publication bias evaluation using trim-and-fill were conducted.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThirteen studies comprising 524 lesions (195 HCA, 329 FNH) met inclusion criteria. Pooled mean ADC was 1.44\u0026times;10⁻\u0026sup3; mm\u0026sup2;/s (95% CI: 1.34\u0026ndash;1.55) for FNH and 1.35\u0026times;10⁻\u0026sup3; mm\u0026sup2;/s (95% CI: 1.23\u0026ndash;1.47) for HCA, yielding an absolute difference of 0.10\u0026times;10⁻\u0026sup3; mm\u0026sup2;/s (p\u0026thinsp;=\u0026thinsp;0.073) and SMD of 0.49 (95% CI: \u0026minus;\u0026thinsp;0.02\u0026ndash;0.99). Relative ADC difference was 7.6% (p\u0026thinsp;=\u0026thinsp;0.109). Heterogeneity was high (I\u0026sup2; = 84.5%). Prospective studies showed reduced heterogeneity (I\u0026sup2; = 0%) but non-significant mean difference (0.21\u0026times;10⁻\u0026sup3; mm\u0026sup2;/s). Trim-and-fill attenuated effect size further, confirming non-significance.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eADC values exhibit only modest, non-significant differences between HCA and FNH, with substantial heterogeneity limiting clinical applicability. Standardized, prospective multicenter studies and multiparametric MRI protocols remain necessary to improve non-contrast differentiation of these lesions.\u003c/p\u003e","manuscriptTitle":"Diagnostic accuracy of ADC values as a supplementary tool in differentiating hepatocellular adenoma from focal nodular hyperplasia: a systematic review and meta-analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-09 12:54:11","doi":"10.21203/rs.3.rs-8799129/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1be2411e-2347-4cab-acb7-73889ac065c6","owner":[],"postedDate":"February 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":62399373,"name":"Nuclear Medicine \u0026 Medical Imaging"}],"tags":[],"updatedAt":"2026-02-09T12:54:11+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-09 12:54:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8799129","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8799129","identity":"rs-8799129","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}