Geographic and Sex Distribution of LI-RADS Research Globally: A Cross-Sectional Meta-Research Study

Abstract

French

Background:

Hepatocellular carcinoma (HCC) has an uneven global distribution and a higher incidence in males. The Liver Imaging Reporting and Data System (LI-RADS) is a standardized imaging framework for HCC diagnosis. It is unclear whether the evidence used to inform LI-RADS reflects the global burden of HCC.

Purpose:

To determine whether the geographic and sex distribution of studies assessing LI-RADS and those included in the LI-RADS individual participant data (IPD) database reflect the global burden of HCC.

Materials and Methods:

We conducted a cross-sectional meta-research study comparing the country- and sex-specific HCC prevalence from GLOBOCAN 2022 with the distribution of studies eligible for and included in the LI-RADS IPD database. Studies were identified through a systematic search of 4 databases. Study Protocol: OSF page .

Results:

We identified 470 eligible studies including 98 014 patients; of these, 76 studies comprising 11 924 patients were included in the IPD database. Asian and African countries, excluding Republic of Korea, were underrepresented in both the eligible and the IPD dataset. North America was overrepresented. Female patients were under-represented in LI-RADS eligible studies (Z = −21.95, P < .0001) and in studies included in the IPD (Z = −9.02, P < .0001) compared to the global prevalence of HCC in females.

Conclusion:

LI-RADS research is disproportionately reported from some countries relative to HCC burden. Asia, Africa, and female patients remain underrepresented. This may affect the generalizability and diagnostic equity of the LI-RADS system and underscore the need for improved global inclusivity in LI-RADS research.

Keywords

LI-RADS hepatocellular carcinoma geographic representation disease burden sex representation meta-research

Introduction

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy, accounting for 906 000 new cases and 830 000 deaths globally in 2020.¹ Common risk factors for HCC include patients with chronic liver disease such as cirrhosis or chronic hepatitis B virus (HBV) and patients with a prior history of HCC. HCC occurs more frequently in male patients (3:1 M/F) and is unevenly distributed globally, with 76% of cases found in Asia, compared to 10.0% in Europe and 5% in the United States.² In the Eastern hemisphere, cirrhosis secondary to chronic HBV is the primary risk factor for HCC, while in the Western hemisphere, the leading risk factors include cirrhosis due to chronic hepatitis C virus (HCV), alcohol-associated liver disease (ALD), and metabolic dysfunction-associated steatotic liver disease (MASLD).³ Among patients with HCC in the United States, 50% to 60% are infected with HCV, 20% to 25% have ALD, and 20% to 30% exhibit features of the metabolic syndrome, including diabetes and obesity, which are closely linked to MASLD.³

The Liver Imaging Reporting and Data System ( LI-RADS ) was developed to standardize reporting and enhance diagnostic accuracy in liver imaging for patients at risk of HCC.⁴ Radiologists use LI-RADS to assign a relative probability of HCC, rated on a 1 to 5 scale, based on imaging findings on CT, MRI, and contrast-enhanced US (CEUS). The LI-RADS consortium consists of more than 400 contributors represented by 240 institutions from 38 countries. LI-RADS core content is available in 13 languages and is integrated into major clinical practice guidelines such as the American Association for the Study of Liver Diseases (AASLD), Organ Procurement and Transplantation Network (OPTN), and the European Association for the Study of the Liver (EASL), representing its criteria are widely used for imaging-based diagnosis of HCC.^5,6

The LI-RADS Living Systematic Review and Individual Participant Data (IPD) Meta-Analysis Project is an ongoing initiative which evaluates the performance of CT/MRI and CEUS LI-RADS diagnostic algorithms in patients at risk for HCC. The LI-RADS IPD currently consists of 76 studies from 30 institutions and 12 countries, encompassing more than 13 000 participant-level observations. The evidence base from the LI-RADS Living Systematic Review and IPD project has helped evaluate the diagnostic accuracy and inform revisions of the LI-RADS system. Accordingly, it is important that the evidence base included in the LI-RADS IPD database is representative of patients who are affected by HCC to ensure generalizability of the system for all patients.

A potential source of bias that may impact generalizability of the literature is an imbalance between the global burden of HCC and the geographic distribution of published LI-RADS studies. For instance, if most LI-RADS research originates from regions with a high proportion of HCC secondary to HBV infection, applying LI-RADS in populations with different predominant risk factors, such as MASLD, may impact different diagnostic performance. In addition, low- and middle-income countries which are often under-represented in randomized trials and biomedical research more broadly, and females who are historically underrepresented due to sex-based exclusions in study design may also impact the generalizability of published LI-RADS studies.^7,8 The purpose of this study is to determine whether the geographic and sex distribution of studies assessing LI-RADS and those included in the IPD database reflect the global burden of HCC.

Methods

Study Design

We performed a cross-sectional meta-research study examining the global burden of HCC in relation to the geographic and sex representation of studies eligible for or included in the LI-RADS Living Systematic Review and IPD Meta-Analysis Project. The study protocol is available on the OSF page .

The LI-RADS living systematic review project has approval from the institutional review board (IRB) at The Ottawa Hospital. The LI-RADS IPD project has assessed numerous accuracy measures pertaining to LI-RADS; this study represents a unique analysis with no overlap with prior work.^9-19 No personal health information was disclosed to the project team. This study integrated data from CANCER TODAY , a global database that reports cancer statistics up to 2022, using GLOBOCAN estimates on incidence, mortality, and prevalence for 36 cancer types, across 185 countries, stratified by age and sex.¹ The burden of HCC was determined by using the prevalence in each country from the GLOBOCAN database.²⁰ Prevalence was selected as the primary measure of disease burden because LI-RADS is applied in both newly diagnosed and existing cases of HCC. Prevalence therefore more closely reflects the population contributing to diagnostic accuracy studies and the LI-RADS IPD dataset. The sex distribution from eligible studies and available IPD was analyzed and compared to that in the GLOBOCAN database.

LI-RADS IPD Search and Inclusion Criteria

Search, screening, and inclusion methods are detailed on our OSF page .¹⁰ In brief, we searched MEDLINE, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), and Scopus (updated to February 2024) databases for studies evaluating the diagnostic accuracy of CT, MRI, or CEUS for HCC using LI-RADS (CT/MRI v2014, v2017, or v2018; CEUS v2016 or v2017), without language or publication type restrictions. The corresponding authors of each study identified for inclusion were contacted. Eligible studies included those reporting the percentage of HCC across LI-RADS categories 1 to 5 in high-risk patients.

All CT, MRI, and CEUS studies reporting the percentage of HCC for LI-RADS categories 1 to 5 in patients at high risk (hepatic cirrhosis, chronic HBV infection, current or prior HCC) were eligible for inclusion. CT, MRI, and CEUS techniques were evaluated for each study to determine concordance with the LI-RADS technical imaging guidelines.

Data Collection Process and Definitions for Data Extraction

Data from eligible studies included in the LI-RADS IPD were stored in DistillerSR, with information extracted on country of origin, sample size, and reported sex distribution. While the analysis focused on sex due to its availability in study demographics, the influence of gender could not be assessed because it was not reported in the included studies. The prevalence of primary liver cancer (ICD-10 C22), was obtained from the GLOBOCAN 2022 database for 185 countries and territories, as well as the male to female global proportion of HCC.²⁰ Data were extracted in duplicate by 2 assessors with disagreements resolved by consensus and when necessary, discussion with a third reviewer.

Outcome Measures

To quantify the available LI-RADS evidence, 2 measures of published LI-RADS evidence were used: the number of published LI-RADS studies eligible for inclusion into the LI-RADS IPD, and the number of studies for which data are available within the LI-RADS IPD Redcap database. Location of study was determined by identifying the country of the corresponding authors as well as the description of setting in the methods. For studies with patients from multiple countries, the study “counted” for each of these countries. Both measures of LI-RADS evidence base were calculated as a percentage of the number of studies from each country, divided by the total number of studies from all countries.

The estimated burden of disease for HCC was calculated by the number of individuals with HCC in a country, divided by the total number of individuals with HCC globally. To assess the representation of females in the eligible studies, including IPD, and the GLOBOCAN database, we calculated the proportion of female patients within each dataset. Specifically, we determined the number of females in each eligible study and in the GLOBOCAN database.

Statistical Analysis

To compare the burden of disease for HCC in each country to the percentage of LI-RADS studies eligible for inclusion to the LI-RADS IPD, and to the percentage of the studies from each country that is available within the IPD, we calculated the differences between percentages of prevalence in each country reported in GLOBOCAN, and eligibility percentage in our database. The possible range of outputs in differences ranged from −100% to +100%, representing maximum under-representation to maximum over-representation respectively. A z-test was conducted to assess for differences in sex proportions between LI-RADS-eligible studies and studies included in the IPD database to those reported in GLOBOCAN. All analyses were performed using R (version 4.3.1) and visualization using the ggplot2 package and mapmychart visualization software.^21-23

Results

The study selection process is summarized in the PRISMA flow diagram in Figure 1.

Figure 1.

PRISMA 2020 flow diagram of study selection for geographic, sex-based, and IPD subset analyses.

Table 1 presents the global distribution of primary liver cancers based on age-adjusted prevalence data from the GLOBOCAN 2022 database.

Table 1.

Global Distribution of Hepatocellular Carcinoma (HCC) by Region, with Total Number of Male and Female Cases.

Region	HCC cases (n = 459 693)	% of global HCC burden
Africa	40 830	8.9
Latin America and the Caribbean	21 360	4.6
North America	25 120	5.5
Europe	43 910	9.6
Oceania	2477	0.5
Asia	325 996	70.9
Total male cases	321 405	71.2
Total female cases	132 288	28.8
Global total	459 693	100

Table 2 presents the number and percentage of LI-RADS eligible studies and IPD studies contributed by each country. Of the 470 eligible studies, the largest proportion originated from China (n = 165; 35.1%), followed by the Republic of Korea (n = 115; 24.5%) and the United States (n = 69; 14.7%).

Table 2.

Name and Proportion of LI-RADS Eligible and Individual Participant Data (IPD) Studies by Country of Origin (Countries with ≥1 Eligible Study Only).

Country	Number of eligible studies in LI-RADS database	% of total eligible LI-RADS studies	Number of studies in IPD database	% of total IPD studies
Argentina	1	0.21	1	1.3
Belgium	1	0.21	0	0
Canada	14	2.98	5	6.6
China	165	35.11	13	17.1
Egypt	15	3.19	0	0
France	5	1.06	1	1.30
Germany	10	2.13	0	0
India	9	1.91	0	0
Iran	1	0.21	0	0
Italy	27	5.74	4	5.3
Japan	10	2.13	1	1.3
Poland	3	0.64	2	2.6
Republic of Korea	115	24.47	38	50
Romania	5	1.06	0	0
Singapore	1	0.21	0	0
Spain	2	0.43	0	0
Sweden	1	0.21	0	0
Switzerland	2	0.43	1	1.3
Taiwan	1	0.21	0	0
Thailand	2	0.43	0	0
Tunisia	1	0.21	0	0
Turkey	1	0.21	0	0
United Kingdom	2	0.43	1	1.3
United States of America	69	14.68	9	11.8

Figure 2 presents a forest plot comparing the difference between the percentage of global HCC burden per country and the percentage representation of each country in the LI-RADS eligible studies and in studies included in the IPD database. Most countries cluster tightly around the zero-difference mark, reflecting relatively proportional representation. However, a subset of countries demonstrates substantial divergence.

Figure 2.

Forest plot depicting discrepancies between Global HCC burden and country representation in LI-RADS research. This forest plot illustrates the difference between each country’s representation in LI-RADS research and contribution to the global hepatocellular carcinoma (HCC) burden. Blue dots represent the difference between the proportion of eligible studies originating from each country and that country’s share of the global HCC burden. Red dots represent the difference between the proportion of individual patient data (IPD) studies and the global burden. The x-axis indicates the percentage point difference (% Representation − % HCC Burden), with a dashed vertical line at 0 marking proportional alignment. Countries with a difference of 0.5% ± of 0 were excluded for readability purposes. Positive values reflect overrepresentation, while negative values indicate underrepresentation. Only countries with data available for at least one study type are shown. Select countries with extreme discrepancies are labeled.

Figure 3 illustrates the difference between each country’s share of the HCC burden and its representation in the LI-RADS database using a global heatmap.

Figure 3.

Geographic visualization of disparities between eligible study representation and global HCC burden by country. Global map displays the difference between each country’s share of global HCC burden and the proportion of eligible studies it contributed. Countries shaded in dark red are those that contributed far more studies than expected based on their disease burden (>1.51% overrepresentation). Countries shaded in orange represent a −0.99% to 1.51% difference. Beige shades indicate a <−1% difference. Gray areas indicate missing data. Image created with mapmychart.net.

Of the 76 studies included in the LI-RADS IPD database, the Republic of Korea accounted for 50.0% (n = 38), China for 17.1% (n = 13), and the United States for 11.8% (n = 9). Canada, Italy, and Poland contributed 5 (6.6%), 4 (5.3%), and 2 (2.6%) studies respectively.

Supplemental Table 1 presents a summary of primary liver cancer prevalence (both sexes, ages 15-85+) based on GLOBOCAN 2022 data, alongside the proportion of LI-RADS studies attributed to each country.

Several countries such as Bangladesh, Australia, and Angola were absent from the LI-RADS eligible study database. Albania, Armenia, Azerbaijan, Austria, Bahamas, Bahrain, Barbados, and Belize each had reported HCC cases ranging from 10 to 534, yet none were represented in the LI-RADS eligible studies or in studies included in the IPD database.

India accounted for 4.53% of the global HCC burden and contributed to 1.91% of eligible studies.

China, which carries the highest global burden of hepatocellular carcinoma (42.6%), was underrepresented in both the LI-RADS eligible study dataset and in studies included in the IPD database. The proportion of studies from China comprised only 35.1% of the eligible dataset and 17.1% of studies included in the IPD database, yielding a difference of −7.5% and −25.5%, respectively, when compared to its share of global HCC burden.

Republic of Korea contributed 115 eligible studies (24.47%) and 38 of IPD studies, accounting for 50.00% of all IPD studies, despite a global HCC burden of only 2.0%, yielding a difference of +24.5% (eligible studies) and +48.0% (studies included in the IPD database).

To visualize broader regional patterns, we plotted the difference between each region’s contribution to LI-RADS studies and its share of the global HCC burden. Positive values reflect overrepresentation, while negative values indicate underrepresentation. Differences in region-level representation can be seen in Figure 4.

Figure 4.

Difference between LI-RADS study representation and global hepatocellular carcinoma (HCC) burden by region. This plot displays the percentage point difference between each region’s representation in LI-RADS studies and its share of the global HCC burden. Blue points represent eligible studies; orange points represent studies contributing individual patient data (IPD). The vertical dashed line at 0% indicates proportional representation. Positive values reflect overrepresentation in LI-RADS research relative to disease burden, while negative values indicate underrepresentation.

Sex Representation in LI-RADS Eligible Studies Versus Global HCC Estimates

Among the 98 014 patients included in LI-RADS eligible studies, 26 035 were female, corresponding to 26.6% of the sample. Z tests of proportions are shown in Table 3.

Table 3.

Comparison of Female Representation in Eligible Studies and Studies Included in IPD Database Versus Global HCC Estimates from GLOBOCAN, with Z-Test of Proportions.

Dataset	Total patients	Female N (%)	z-Value	p-Value
GLOBOCAN (Global Estimate)	459 693	138 288 (30.08)	–	–
LI-RADS eligible studies	98 014	26 035 (26.56)	−21.95	<.0001
Studies included in IPD database	11 924	3130 (26.3)	−9.02	<.0001

Discussion

This study identifies underrepresentation of patients from regions with the highest global disease burden, most notably China, India, Japan, Egypt, and other countries across Asia and Africa in LI-RADS research. Our study also identified a significant underrepresentation of female patients in LI-RADS-eligible studies relative to global HCC prevalence.

In East and Southeast Asia, HCC is predominantly driven by HBV, often presenting in younger patients and associated with higher alpha-fetoprotein (AFP) levels.²⁴ Conversely, in Western nations such as the United States and Canada, which are overrepresented in the IPD database, HCC is more often associated with HCV, ALD, and MASLD.³ These etiologic differences may influence tumor biology and vascular profiles, and thereby impact imaging characteristics and LI-RADS classification.²⁵ HBV-related HCCs for example (common in East and Southeast Asia) present more frequently as well-differentiated, non-hypervascular tumors in early stages, which may not meet LR-5 criteria due to absent or subtle enhancement features.²⁶

The underrepresentation of China, a nation where nearly half of all global patients with HCC reside, is of particular importance. China’s health system has made substantial investments in liver cancer screening, particularly through ultrasound surveillance programs in HBV-infected populations, making it uniquely positioned to benefit from optimized imaging strategies.²⁷ Yet, the lack of Chinese IPD in LI-RADS research may limit the system’s performance in this setting, potentially reducing its value in clinical practice where it is most needed. LI-RADS is not yet widely adopted in China, in part due to variation in imaging standards and the use of national frameworks such as the China National Liver Cancer (CNLC) guidelines, updated in 2024.²⁸ Differences in contrast agent emphasis and inclusion of subcentimeter HCC may pose challenges for standardizing diagnostic criteria, underscoring the importance of including Chinese IPD in future efforts.

Another pattern in our results is the sharp drop in country representation from the eligible study level to the IPD level. For instance, Egypt, India, and Germany all had eligible studies but no representation in the IPD subset. Possible contributing factors, such as language, infrastructure, and data-sharing logistics, may prevent the inclusion of valuable participant-level data from these regions.²⁹ Overreliance on data from a narrow set of countries risks limiting the generalizability of LI-RADS toward specific populations.

Although HCC has a known male predominance (3:1 M/F) this does not account for the underrepresentation of female patients in diagnostic studies.² Although recent IPD evidence suggests that LI-RADS major feature performance is consistent across sex, these findings are derived from datasets with underrepresentation of female patients and several high-burden regions, and therefore may not fully capture potential effects of sex or geography on diagnostic accuracy in globally representative populations. Biological sex may influence the presentation and detection of HCC; for example, Liou et al reported that females were more likely to be diagnosed via surveillance (37.9% vs 29.6%, P = .003), resulting in detection at earlier stages with smaller tumors (median 30 vs 39.5 mm, P = .038) and lower rates of portal vein invasion and metastasis.³⁰ These differences suggest that sex-specific variations, whether biological, behavioral, or systemic, can influence tumor characteristics and imaging outcomes.³⁰ If diagnostic systems like LI-RADS are primarily developed and validated in male-predominant cohorts, they may not be applicable for female patients, increasing the risk of indeterminate or false-negative classifications. Despite this potential risk, a recent study by Adamo et al found no significant difference in the association of LI-RADS major features with HCC between males and females, which reduces concern related to the observed underrepresentation found in females in eligible studies and studies included in the IPD database.¹⁴ However, these conclusions remain dependent on the representativeness of the underlying dataset, underscoring the importance of ensuring global inclusivity in future LI-RADS research. In addition, longstanding issues regarding sex bias in radiologic research emphasize the need for sex-stratified validation of imaging criteria.³¹ In many regions, particularly North America and the Middle East, the incidence of MASLD related HCC is increasing, while the contribution of viral hepatitis is declining. As the etiologic drivers of HCC continue to transition away from viral causes toward MASLD, the traditional male predominance of HCC may attenuate over time, potentially narrowing the observed sex gap in disease incidence. These trends demonstrate the importance of ensuring that diagnostic systems such as LI-RADS are validated in contemporary and evolving patient populations, including those with MASLD-related liver disease.³²

Our study emphasizes an important message to the LI-RADS community and imaging researchers in general. Given that the LI-RADS IPD database is used to guide updates to the LI-RADS system, the observed imbalances call for methodological reforms and improved representation, both in study recruitment and in the curation of datasets used for meta-analysis. Furthermore, future research should adopt stratified sampling approaches when designing validation studies to ensure balanced representation across sex, geographic regions, and underlying disease etiology. Such representation is critical for developing diagnostic tools that are effective and generalizable across patient populations. In regions where HCC presents with distinct etiologic and phenotypic characteristics, such as HBV-driven disease in East Asia or aflatoxin-related HCC in sub-Saharan Africa, regionally adapted or etiology-specific modifications of LI-RADS may enhance diagnostic accuracy and clinical utility. Finally, journals and funding agencies can play an important role by recommending disaggregated reporting of sex and geography in diagnostic accuracy studies and incentivizing inclusion of underrepresented populations.

This study has several limitations. Our analysis relied on global HCC burden estimates from the GLOBOCAN 2022 database, which, while widely used, may be subject to inaccuracies. Notably, the GLOBOCAN classification groups HCC and intrahepatic cholangiocarcinoma (ICC) under the broader category of “liver cancer” (ICD-10 code C22), limiting the specificity of burden estimates for HCC alone. This may result in overestimation of true HCC prevalence in countries where ICC constitutes a higher proportion of “liver cancer” cases. However, HCC comprises approximately 90% of all primary liver cancers globally, minimizing the impact of this limitation.³³ Because GLOBOCAN reports liver cancer under a combined category that influences HCC and intrahepatic cholangiocarcinoma (iCCA), sex-based estimates may be slightly affected. iCCA has a sex distribution of 1.1:1 M/F, compared to 3:1 for HCC, which may increase the proportion of females in the global estimate.³⁴

Third, variability in cancer registration systems, diagnostic capacity, and healthcare infrastructure across countries introduces additional limitations. For example, the validity of GLOBOCAN estimates has been assessed for other cancer types, with studies showing that accuracy varies by cancer site, prognosis, and data availability.³⁵ It was found that cancers with low mortality or those detectable by screening (eg, thyroid, prostate, breast) had less accurate estimates compared to cancers with poor survival, highlighting the method’s limitations.³⁵

Conclusion

This study highlights major geographic and sex-based disparities in LI-RADS research, raising concern about its global applicability and diagnostic equity. High-burden countries are underrepresented while low-burden countries are overrepresented. Female patients are also significantly underrepresented compared to global HCC demographics. Addressing these gaps will likely require investment in data infrastructure and inclusive recruitment strategies to ensure LI-RADS reflects the diversity of the global HCC population.

Supplemental Material

sj-docx-1-caj-10.1177_08465371261429344 – Supplemental material for Geographic and Sex Distribution of LI-RADS Research Globally: A Cross-Sectional Meta-Research Study

Supplemental material, sj-docx-1-caj-10.1177_08465371261429344 for Geographic and Sex Distribution of LI-RADS Research Globally: A Cross-Sectional Meta-Research Study by Hoda Osman, Eric Lam, Christian B. van der Pol, Hala Talal Mahdi, Rawan Awad, Nabil Islam, Mostafa Alabousi, Jean-Paul Salameh, Mustafa R. Bashir, Andreu F. Costa, Haresh Naringrekar, Justin Presseau, An Tang and Matthew D. F. McInnes in Canadian Association of Radiologists Journal

Footnotes

ORCID iDs

Hoda Osman

Eric Lam

Christian B. van der Pol

Hala Talal Mahdi

Rawan Awad

Mostafa Alabousi

Andreu F. Costa

An Tang

Matthew D. F. McInnes

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Canadian Institutes of Health Research (CIHR) Operating Research Grant (#469052) to Matthew DF McInnes, Radiological Society of North America (RSNA) R&E Foundation Research Scholar Grant to Christian van der Pol, and Fonds de recherche du Québec en Santé (FRQ-S) Merit research scholarship (#376472) to An Tang.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Protocol Registration

Open Science Framework ().

Supplemental Material

Supplemental material for this article is available online.

References

Sung

Ferlay

Siegel

, et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209-249. doi:10.3322/caac.21660

De Muzio

Grassi

Dell’Aversana

, et al. A narrative review on LI-RADS algorithm in liver tumors: prospects and pitfalls. Diagnostics (Basel). 2022;12(7):1655. doi:10.3390/diagnostics12071655

Choo

Tan

Goh

Tai

Zhu

AX.

Comparison of hepatocellular carcinoma in Eastern versus Western populations. Cancer. 2016;122(22):3430-3446.

American College of Radiology. Liver imaging reporting & data system. 2020. Accessed July 3, 2024. https://www.acr.org/Clinical-Resources/Clinical-Tools-and-Reference/Reporting-and-Data-Systems/LI-RADS

American Association for the Study of Liver Diseases. AASLD. 2024. Accessed November 14, 2024. https://www.aasld.org/

Organ Procurement and Transplantation Network. OPTN. 2024. Accessed November 14, 2024. https://optn.transplant.hrsa.gov/

Røttingen

J-A

Regmi

Eide

, et al. Mapping of available health research and development data: what’s there, what’s missing, and what role is there for a global observatory? Lancet. 2013;382:1286-1307.

Emdin

Odutayo

Hsiao

, et al. Association between randomised trial evidence and global burden of disease: cross sectional study (Epidemiological Study of Randomized Trials—ESORT). BMJ. 2015;350:h117. doi:10.1136/bmj.h117

Naringrekar

Costa

Lam

, et al. Risk of bias in Liver Imaging Reporting and Data System studies using QUADAS-2. Can Assoc Radiol J. 2025;76(2):273-286.

10.

‌Abedrabbo

Lerner

Lam

, et al. Is concurrent LR-5 associated with a higher rate of hepatocellular carcinoma in LR-3 or LR-4 observations? An individual participant data meta-analysis. Abdom Radiol (NY). 2024;50(4):1533-1546.

11.

Goins

Jiang

van der Pol

, et al. Comparative performance of 2018 LI-RADS versus modified LIRADS (mLI-RADS): an individual participant data meta-analysis. J Magn Reson Imaging. 2024;60(3):1082-1091.

12.

Costa

McInnes

MDF

van der Pol

, et al. Individual participant data meta-analyses for diagnostic accuracy research: challenges and lessons learned from the LI-RADS IPD group. Radiol Imaging Cancer. 2024;6(3):e240015.

13.

‌Dawit

Lam

McInnes

MDF

, et al. LI-RADS CT and MRI ancillary feature association with hepatocellular carcinoma and malignancy: an individual participant data meta-analysis. Radiology. 2024;310(2):e231501.

14.

Adamo

Lam

Salameh

, et al. Do risk factors for HCC impact the association of CT/MRI LIRADS major features with HCC? An individual participant data meta-analysis. Can Assoc Radiol J. 2025;76(3):466-476.

15.

Goins

Jiang

van der Pol

, et al. Individual participant data meta-analysis of LR-5 in LI-RADS version 2018 versus revised LI-RADS for hepatocellular carcinoma diagnosis. Radiology. 2023;309(3):e231656.

16.

Goins

Adamo

Lam

, et al. Conversion strategy for LI-RADS category 5 observations across versions 2014, 2017, and 2018. Radiology. 2023;307(4):e222971.

17.

van der Pol

McInnes

MDF

Salameh

, et al. CT/MRI and CEUS LI-RADS major features association with hepatocellular carcinoma: individual patient data meta-analysis. Radiology. 2022;302(2):326-335.

18.

van der Pol

McInnes

MDF

Salameh

, et al. Impact of reference standard on CT, MRI, and contrast-enhanced US LI-RADS diagnosis of hepatocellular carcinoma: a meta-analysis. Radiology. 2022;303(3):544-545.

19.

van der Pol

Costa

Lam

Dawit

Bashir

McInnes

MDF.

Best practice for MRI diagnostic accuracy research with lessons and examples from the LI-RADS individual participant data group. J Magn Reson Imaging. 2023;60(1):21-28.

20.

Ferlay

Ervik

Lam

, et al. Global cancer observatory: cancer today (version 1.1). International Agency for Research on Cancer. 2024. Accessed September 3, 2024. https://gco.iarc.who.int/today

21.

R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. 2023. Accessed January 17, 2025. https://www.R-project.org/

22.

Wickham

ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York; 2016. Accessed February 17, 2025. https://ggplot2.tidyverse.org

23.

MapChart. Create your own custom map. 2024. Accessed January 28, 2025. https://www.mapchart.net/

24.

Chidambaranathan-Reghupaty

Fisher

Sarkar

Hepatocellular carcinoma (HCC): epidemiology, etiology and molecular classification. Adv Cancer Res. 2021;149:1-61.

25.

Chernyak

Fowler

RKG

, et al. LI-RADS: looking back, looking forward. Radiology. 2023;307(1):e222801.

26.

‌Yang

Chen

Huang

, et al. Correlation between CEUS LI-RADS categorization of HCC < 20 mm and clinic-pathological features. Insights Imaging. 2024;15(1):110.

27.

Chen

Zheng

Baade

, et al. Cancer statistics in China, 2015. CA Cancer J Clin. 2016;66(2):115-132.

28.

Wei

Liao

, et al. Interpretation of the updates of the Chinese guidelines for the diagnosis and treatment of primary liver cancer (CNLC-2024 Edition). Hepatoma Res. 2024;10:30.

29.

Hamilton

Hong

Fraser

Rowhani-Farid

Fidler

Page

MJ.

Prevalence and predictors of data and code sharing in the medical and Health Sciences: systematic review with meta-analysis of individual participant data. BMJ. 2023;382:e075767. doi:10.1136/bmj-2023-075767

30.

Liou

Tan

TJ-Y

Chen

Goh

Chang

Tan

CK.

Gender survival differences in hepatocellular carcinoma: is it all due to adherence to surveillance? A study of 1716 patients over three decades. JGH Open. 2023;7(5):377-386.

31.

Taliaferro

Agarwal

Coleman

, et al. Sex differences in radiation research. Int J Radiat Biol. 2024;100(3):466-485. doi:10.1080/09553002.2023.2283089

32.

Pan

Al Hassani

Ismail

, et al. Changing profile and burden of hepatocellular carcinoma in Arab Countries in 1990–2021. Ann Hepatol. 2025;30(2):102104.

33.

Yang

J-Q

Wang

X-G

Incidence trend and prognosis of intrahepatic cholangiocarcinoma: a study based on the seer database. Transl Cancer Res. 2023;12(11):3007-3015. doi:10.21037/tcr-23-1278

34.

Antwi

Mousa

Patel

Racial, ethnic, and age disparities in incidence and survival of intrahepatic cholangiocarcinoma in the United States; 1995-2014. Ann Hepatol. 2018;17(4):604-614.

35.

Antoni

Soerjomataram

Møller

Bray

Ferlay

An assessment of GLOBOCAN methods for deriving national estimates of cancer incidence. Bull World Health Organ. 2016;94(3):174-184. doi:10.2471/blt.15.164384