AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH

AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (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];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH View ORCID Profile Dylan Windell , View ORCID Profile Alastair Magness , View ORCID Profile Cayden Beyer , View ORCID Profile Helena Thomaides Brears , View ORCID Profile Sarah Larkin , View ORCID Profile Kezia Hobson , View ORCID Profile Paul Aljabar , View ORCID Profile Kenneth Fleming , View ORCID Profile Eve Fryer , View ORCID Profile Timothy Kendall , Reema Kainth , Phil Wakefield , View ORCID Profile Caitlin Langford , View ORCID Profile Pierre Bedossa , View ORCID Profile Robert Goldin doi: https://doi.org/10.1101/2025.04.23.25326290 Dylan Windell 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Dylan Windell For correspondence: dylan.windell{at}perspectum.com Alastair Magness 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Alastair Magness Cayden Beyer 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Cayden Beyer Helena Thomaides Brears 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Helena Thomaides Brears Sarah Larkin 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sarah Larkin Kezia Hobson 1 Perspectum Ltd , Oxford, UK 2 Section for Pathology, Imperial College , London, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kezia Hobson Paul Aljabar 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Paul Aljabar Kenneth Fleming 1 Perspectum Ltd , Oxford, UK 3 Emeritus Fellow, Green Templeton College, University of Oxford , UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kenneth Fleming Eve Fryer 1 Perspectum Ltd , Oxford, UK 4 Department of Cellular Pathology, John Radcliffe Hospital , OUH NHS FT, Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Eve Fryer Timothy Kendall 1 Perspectum Ltd , Oxford, UK 5 Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Timothy Kendall Reema Kainth 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Phil Wakefield 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Caitlin Langford 1 Perspectum Ltd , Oxford, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Caitlin Langford Pierre Bedossa 6 LiverPat , Paris, France Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Pierre Bedossa Robert Goldin 1 Perspectum Ltd , Oxford, UK 2 Section for Pathology, Imperial College , London, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Robert Goldin Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Background & Aims Annotation of liver biopsies, for disease staging is increasingly aided by digital pathology, however existing systems do not quantify inflammation and steatosis within an anatomical framework. We developed an AI system to quantify portal tracts (PT) and disease features and their regional distribution in Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD)/Metabolic Dysfunction-Associated Steatohepatitis (MASH) and Autoimmune Hepatitis (AIH). Methods Digitised images of haematoxylin and eosin-stained specimens were pooled from 4 clinical cohorts (n=390: 89 MASLD, 238 MASH, 63 AIH). Portal tracts, regional steatosis, interface hepatitis, portal and lobular inflammation were quantified using a proprietary AI system and scored by expert pathologists. Results The percentage of steatosis was higher in MASH (7.5%) compared to MASLD (3.2%, p<0.0001). Lobular regions had larger steatotic vesicles (260 vs 190mm 2 , p<0.0001). AI-derived steatosis quantification correlated with manual grading (r s =0.72; p<0.0001). The inflammatory cell number (ICN) was 2-fold higher in AIH compared to MASLD/MASH in interface [390 vs 140; p<0.0001], portal [4600 vs 1500], and lobular [1500 vs 650] regions. Severity of portal inflammation from manual grading correlated well with ICN count at PT (r s =0.71; p<0.0001) but not lobular regions (r s =<0.1). For equivalent grades of portal inflammation, the ICN was up to 3-fold higher in AIH than in MASLD/MASH (r s =0.71; p<0.0001). Conclusion Despite equivalent pathologist portal inflammation grades, the digital burden of inflammation was significantly higher in AIH than MASLD/MASH. This digital system quantifies PT, inflammation, and steatosis, providing powerful decision support for pathologists in AIH diagnosis and MASH staging. LAY SUMMARY The detection of portal tracts is an important part of liver biopsy sample quality control and histological scoring. Using artificial intelligence, this study demonstrates a system that automatically detects and quantifies portal tracts and surrounding patterns of inflammation and steatosis. AI found that inflammation was in similar regions but was higher in autoimmune hepatitis than in metabolic dysfunction-associated steatohepatitis, despite similar grading from manual scoring. This AI system provides granular information that can aid biopsy grading and provide insights into liver disease progression and diagnosis. Introduction Metabolic dysfunction-associated steatotic liver disease (MASLD) affects one third of the global population 1 and is characterised by increased fat in and around hepatocytes 2 . Progression to steatohepatitis (MASH), with concurrent liver inflammation, hepatocyte ballooning, and increased risk of fibrosis, is estimated to occur in 5-14% of adults and projected to increase 1 , 3 , 4 . Autoimmune hepatitis (AIH) is rarer, prevalent in 16 cases per 100,000 people globally 5 , but progresses rapidly once diagnosed. Liver biopsy remains the reference standard for diagnosing AIH and staging MASLD/MASH, particularly in complex cases 2 , 6 – 8 . Therapeutics for AIH are relatively well established but the metabolic comorbidities and multiplicity of disease features in MASH complicate drug efficacy and have accelerated development of anti-steatotic 9 , appetite suppressing 10 , anti-inflammatory 11 , anti-apoptotic 12 , and anti-fibrotic 13 candidates. Resolution of histological disease features are primary endpoints in MASH late-phase clinical trials. While clearly improved by blinded and centralised liver histology reading, both inter– and intra-rater variability still affect biopsy-based evaluation 14 – 17 . This highlights a continued need for standardisation of biopsy sampling and consensus in grading 18 . Identification of portal tracts (PT) in liver biopsy samples allows the assessment of biopsy adequacy and staging of disease based on the relative anatomical distribution of liver disease features 19 – 21 . Identification of PT is challenging because they vary greatly in their size, structure, and shape 20 , and their appearance is affected by biopsy needle size, cutting angle 22 , 23 , and disease stage 24 – 26 . In recent audits 20 , 27 , at least 50% of biopsies fail adequacy criteria, so improving PT detection is also desirable. Artificial Intelligence (AI)-based systems have been developed that complement manual annotation and pathology grading of liver biopsy samples 28 . Existing systems for detecting and segmenting PT have been effective and accurate, with examples being considered by the FDA for application in MASH clinical trials 29 . Automated systems can show high degrees of correlation between manual annotations and automatic measures 30 – 36 , and are increasingly able to distinguish and segment anatomical features such as interface hepatitis and bile ducts 37 . However, current systems have been developed for fibrosis, and this prevents identification of inflammation, ballooning, and steatosis, which require a different staining agent (Haematoxylin & Eosin, H&E, rather than Masson’s Trichrome or Picrosirius Red). Furthermore, despite the accuracy of these systems, they lack an anatomical framework to assist with interpretation and grading of pathology. We set out to address these shortcomings by creating an AI system capable of detecting PT and quantifying pathological features relative to PT and other landmarks from biopsy samples stained with H&E. Using this new method, we aimed to comprehensively evaluate the burden of steatosis and of inflammation in a large dataset of MASLD/MASH and AIH liver biopsies. Additional objectives were to examine regional distributions of these features, at PT and in liver lobules, and associations to disease severity from manual grading. Materials and Methods Study Design and Populations Data was obtained from four independent clinical cohorts detailed previously 38 – 40 . Inclusion of digitized datasets was based on adequacy of biopsy samples 41 , 42 , in terms of stain and sample quality, and on confirmation of AIH 43 and MASLD/MASH based on clinical criteria and pathology scoring independent of this work 41 , 42 , 44 ( Supplementary Figure 1 ). Participants with MASLD/MASH or AIH were enrolled at routine patient visits in secondary or tertiary care, at sites in the United Kingdom, the United States, Poland, and Japan. Patients with any other known chronic liver disease were excluded from this analysis. Where biopsies from multiple timepoints were available, only the baseline timepoint was selected to avoid duplication. All patients and volunteers provided written informed consent. All parent studies were conducted in accordance with the ethical principles of the Declaration of Helsinki 2013 and the Good Clinical Practice guidelines and all studies approved by ethical committees (Institutional review at trial sites for NCT03551522 , Ethics committee at Yokohama City Hospital for UMIN000026145, West Midlands – Black Country Research Ethics Committee 14/WM/0010 for ISRCTN39463479 , Komisja Bioetyczna przy Instytucie Pomnik-Centrum Zdrowia Dziecka 11/KBE/2016 for NCT03198104 ). Pathology Protocol Histological scoring Histological WSI were split between four pathologists with over 100 years of combined experience, blinded to all clinical data (EF, KF, TK & RG, Supplementary Table 5 ). Biopsy sample adequacy was assessed using the definition outlined by the Royal College of Pathologists 45 . All MASLD/MASH and AIH WSI were scored using the Ishak scoring system 46 for portal, interface and lobular inflammation and the NAS scoring system 47 for lobular inflammation, fibrosis and steatosis. AI system for portal tract detection and quantification of steatosis and inflammation We developed a proprietary system of interrelated AI systems from H&E images that could identify landmark features, detect PT, perform regional demarcation, and spatial analysis to quantify steatotic vesicles and inflammatory cells ( Figure 1 ). The following measurements were derived within portal, interface, periportal, and lobular regions: Download figure Open in new tab Figure 1 – AI pipeline to segment inflammation and steatosis in relation to portal tract location. Equidistant regions (100µm) from PT allow for detailed quantification of steatosis and inflammation in portal, periportal and lobular regions of liver samples. Steatosis Fraction (SF) = Area of steatosis / Area of region Vesicle Hepatocyte Fraction (VHF) = Total number of steatotic vesicles / Total number of hepatocytes Inflammatory Cell Number (ICN) = Total number of Inflammatory cells Inflammatory Burden (IB) = Number of inflammatory cells / Area of region Statistical analyses Descriptive statistics were used to summarize cohort characteristics. Median and IQR (Interquartile Range) were used to describe all non-normally distributed variables. Frequency and percentage were used to describe categorical variables. Differences between AI measurements across disease states were assessed using Wilcoxon rank-sum test. Correlations between AI measurements and pathologist scoring were assessed using Spearman’s rank correlation. Grouped scores between pathologists were either compared using Kruskal Wallace tests or, if there were less than 3 scores within each category, a Wilcoxon rank-sum test was performed. Correlations between AI measurements in different regions were assessed using Pearson’s correlation coefficient. All statistical analyses were performed in R Studio version 4.2.2. Values of p<0.05 were considered statistically significant. Results Study populations A total of 390 individuals were recruited from secondary care and paediatric clinics between February 2014 and March 2019 ( Supplementary Figure 1 ). Their mean age was 51 years, 48% were male, with mean BMI of 29.5kg/m 2 ( Supplementary Table 1 ). Overall, 89/390 (23%, all adults) had been diagnosed with MASLD, 238/390 (61%, all adults) with MASH and 63/390 (16%, of which 3% adults) had AIH. In those with MASLD/MASH, known comorbidities were obesity (51%), diabetes (35%), hypertension (21%) and dyslipidaemia (33%). These comorbidities were rare in those with AIH. Treatment for AIH (56%) was considered as ongoing with either Steroids or Azathioprinum. In those with MASH, 286 had “at-risk MASH” (NAS activity score ≥4 and ≥2 fibrosis stage). Similarly, for AIH disease, 7 had moderate-severe disease (≥2 lobular and portal inflammation score). Every individual underwent a liver biopsy. Biopsy samples were scored manually by 1-4 pathologists and quantified with the AI-based quantification system of PT detection and liver tissue characterisation described previously ( Figure 1 ). Portal tract quantification using AI The AI system identified a total of 7359 PT across the 390 digitized whole slide images (WSI), with a mean number of 19 PT per WSI and 7 PT per 10mm of liver sampled ( Supplementary Table 2 ). 378/390 (97%) WSI met the minimum 6-11 PT/sample and 325/390 (83%) WSI met the minimum 10-12 PT/sample quality criterion recommended for pathological assessment 21 . The PT count correlated with length of sample (Spearman’s rho 0.71, p<0.0001) ( Supplementary Figure 2 ). PT detection by AI was comparable in performance with manual detection by each pathologist ( Supplementary Table 3 ). Agreement in PT detection by different pathologists ranged in F1 score from 0.58 ±0.16 to 0.72 ±0.11. Whereas for AI, the F1 score was 0.66 ±0.18 against the consensus detection achieved by all pathologists. Quantification of steatosis in MASLD and MASH The AI system delineated 1,658,463 steatotic vesicles from the subset of 327 WSI with MASLD or MASH, with a mean 3732 vesicles per sample for MASLD cases and 5572 vesicles per sample for MASH ( Supplementary Table 2 ). Across the whole of the tissue sampled, the % area with steatotic vesicles (steatosis fraction, SF) was lower in MASLD cases compared to the SF in people with MASH (3.2% (IQR 5.6) vs 7.5% (IQR 7.1), p<0.0001) ( Figure 2 , Supplementary Figure 3 ). Similarly, the ratio of vesicles to hepatocytes (the vesicle hepatocyte fraction) was lower in MASLD (10.2%; IQR 12.1) than in MASH (20.1% IQR 17.2). The SF correlated with the vesicle hepatocyte fraction (Spearman’s rho=0.9; p<0.0001). The size of steatotic vesicles also differed. Median area of steatotic vesicles was 190μm 2 (IQR 110) in those with MASLD and 260μm 2 (IQR 100) in MASH (p<0.0001) ( Figure 3 ). Download figure Open in new tab Figure 2 – Steatosis in MASLD (n=89) and MASH (n=238). (A) Example case with an interactive steatosis overlay of steatosis fraction represented on a colorimetric scale (blue-red) at specific regions from PT (grey). (B) Steatosis fraction (%) in lobular regions in MASLD and MASH. (C) Vesicle to hepatocyte fraction (%) in lobular regions in MASLD and MASH. Download figure Open in new tab Figure 3 – Regional analysis of steatosis in MASLD (n=89) and MASH (n=238). (A) Steatotic vesicles were quantified within areas defined by regions positioned equidistant from the PT, at intervals of 100μm, in a case with MASLD. (B) Steatosis fraction (%) in MASLD and MASH. (C) Size (µm 2 ) of steatotic vesicles in MASLD and MASH. Median is represented as a solid line and the upper/lower interquartile range is represented as a dashed line. Regional distribution of steatosis In the immediate vicinity of the PT, up to 100μm from the PT, the SF was lower compared to further away in lobular regions, with higher levels of steatosis observed at 500μm from the portal boundary ( Figure 3 ). This was observed for both MASLD and MASH. Significant differences were measured in all regions measured up to 1mm from each PT. The biggest difference in steatotic vesicle aggregation was at approximately 900μm from the PT, with SF of 2.7% (IQR 5.9) for MASLD and 7.9% (IQR 4.2) for MASH (p<0.0001). The difference in vesicle size between MASLD and MASH was preserved at all distances from the PT. Association with steatosis grading from manual pathology reads Expert pathologist grading for steatosis did not differ significantly between MASLD and MASH ( Supplementary Figure 3 ). Both AI-derived metrics for steatosis quantification showed significant correlation with steatosis grading in lobular regions in both MASLD (for SF r s =0.61 and for VHF r s =0.64; p<0.0001] and MASH (for SF r s =0.7 and for VHF r s =0.72; p<0.0001] across all 3 pathologists together or individually. Quantification of inflammation in MASLD/MASH and AIH The AI system identified 1.5 million inflammatory cells at PT and 0.4 million inflammatory cells in lobular regions, and 0.11 million cells in interface regions across all 390 WSI. Inflammatory cells were detected in the liver tissue of both MASLD or MASH individuals (ICN: 4300 mean number of cells per WSI) and in those with AIH (ICN: 12000 mean number of cells per WSI) ( Supplementary Table 2 ). The inflammatory burden (IB) was higher in AIH compared to MASLD/MASH at PT (2700 cells/mm 2 (IQR 2800), vs 1400 cells/mm 2 (IQR 1800), p<0.0001) and in lobular regions (60 cells/mm 2 ; IQR 90 vs 40 cells/mm 2 ; IQR 50, p<0.001) ( Figure 4 ). Download figure Open in new tab Figure 4 – Comparison of inflammation in MASLD/MASH (n=327) versus AIH (n=63). (A) Example of a case with MASLD; both segmented inflammatory cells (ICN) and inflammatory burden (IB) were quantified and overlayed with the anatomical location of the PT and represented with a colorimetric scale. (B) The IB (cells/mm 2 ) and (C) ICN in portal and lobular regions in MASLD/MASH and AIH. Regional distribution of inflammation There was significantly more inflammation in all regions extending from PT in AIH cases compared to MASLD/MASH, but the regional distribution was similar in both diseases. Maximal ICN and IB were in the PT for both AIH and MASLD/MASH ( Supplementary Table 2, Figure 5 ). At the immediate vicinity of the PT, up to 100μm from the PT, IB was suggestive of interface hepatitis in AIH (IB: 150 cells/mm 2 (IQR 130) ( Figure 5 ). Lower levels of IB were measured in MASLD/MASH cases at the interface (IB: 80 cells/mm 2 (IQR 90), p<0.0001). The ICN in the same region was also higher in AIH (390 cells (IQR 580) compared to 140 cells (IQR 220) in MASLD/MASH, p<0.0001). Download figure Open in new tab Figure 5 – Regional analysis performed across WSI for inflammatory burden and inflammatory cell numbers in MASLD/MASH (n=327) and AIH (n=63). (A) Inflammatory metrics are extracted from individual regions positioned equidistant from the PT, at intervals of 100μm, in an AIH liver biopsy. (B) Inflammatory burden and (C) inflammatory cell number in MASLD/MASH and AIH, across equidistant regions. Median is represented as a solid line and the upper/lower interquartile range is represented as a dashed line. In both AIH and MASLD/MASH, the IB was lower at distances of 200–600μm from the PT (2.2-fold lower relative to the interface region, for AIH, and 1.8-fold for MASLD/MASH), but then increased again in lobular regions further away in AIH (700–900μm; 1.3-fold higher relative to the IB at 200μm from PT). In contrast, the ICN decreased gradually across lobular regions in both AIH and MASLD/MASH. Association with scoring of inflammation from manual pathology reads Grading of inflammation in the MASLD/MASH samples showed good agreement between pathologists for portal tracts and interface region but was less aligned for lobular regions ( Supplementary Figure 4 ). Severity of portal inflammation from manual scoring correlated well with ICN count at PT (r s =0.46-0.73; p<0.01), but only weakly with IB (r s =0.16-0.28; all non-significant). Pathologist scorings for both Ishak and NAS for lobular inflammation correlated weakly with both IB and ICN (r s =0.15-0.35; p<0.05 – p<0.34). Manual grading from single pathologists showed similar trends. For all regions, severity of inflammation from manual pathology correlated better with ICN than with IB ( Figure 6 ). In addition, increasing grade of inflammation was more consistent with increasing ICN and IB at the PT and interface regions. Crucially, however, for equivalent grades of portal inflammation the ICN was up to 3.1-fold higher in AIH than in MASLD/MASH and this pattern was seen at portal, interface and lobular regions ( Figure 6 , Supplementary Table 4 ). Of note, ICN at PT correlated well with severity of interface hepatitis in both AIH (r s =0.68, p<0.0001) and MASLD/MASH (r s =0.56, p<0.0001) ( Supplementary Figure 5 ). Download figure Open in new tab Figure 6 – Inflammation quantified by AI in relation to manual grading by a single pathologist for MASLD/MASH (n=327) and for AIH (n=63). (A) Shows grades relative to IB (B) Shows Ishak grades relative to ICN. Grades shown are based on Ishak scoring at portal and interface regions and by either Ishak or NAS grading at the lobules. Discussion Our aim was to develop an AI solution to identify and quantify PT, steatosis, and inflammation to evaluate 390 MASLD/MASH and AIH cases. The resulting AI system accurately identified and quantified PT and delineated equidistant 100μm regions surrounding them. AI detected more steatosis and larger steatotic vesicles in MASH than in MASLD. Higher levels of inflammation were found across all AIH cases compared to MASLD/MASH cases. Regional distribution of inflammation across regions was similar between MASLD/MASH and AIH and was higher at the PT and interface. However, overall, greater degrees of inflammation were detected in the AIH cases, even for equivalent grades of inflammation from manual pathology. This comprehensive set of detailed metrics can assist pathology readers in differentiating liver disease types, monitoring disease progression and treatment response. The consistent and unbiased assessment offered by such AI/machine learning systems makes them valuable as decision support tools for pathologists. Manual identification of PT can be time-consuming and challenging due to their variable appearance. This AI system has the potential to speed up this process by directing pathologists’ attention to likely PT. This is particularly relevant for highly fibrotic cases where vascular structures merge together and portal tracts expand resulting in a lack of clear boundary between individual portal tracts 48 . This pipeline could be applied to perform automatic quality control on liver biopsy samples by enumerating total number of PT. Biopsy length is often used as an alternative to PT count for sample adequacy 49 and this AI system could additionally assist by automatically calculating the biopsy length and the full portal tract per linear cm 22 . In the NAS scoring system, steatosis scores are performed by assessing the percentage of hepatocytes that contain lipid deposits. Our AI system reports the macrovesicular steatosis fraction (SF) by area and microvesicular steatosis is excluded. Despite this, when comparing the SF to the ratio of macrovesicular steatosis to segmented hepatocytes, the correlation was very strong. This suggests that the SF metric is equivalent to what is graded manually. Our data showed that SF is increased in MASH as compared to MASLD across periportal and lobular regions, however, due to the low number of advanced MASH cases (Fibrosis score of 4) this was not explored 50 . Our system revealed a difference in the median steatotic vesicle size between MASLD and MASH across periportal and lobular areas. Micro– and macrovesicular steatosis exist in both MASLD and MASH, yet there is no clear consensus on whether the pattern of steatotic vesicle size varies between the disease states 51 . These larger droplets in MASH might indicate increased steatogenesis or an increased rate of coalescence, which might be expected to correlate with progressive disease activity in MASH 52 . However, the clinical significance of steatosis is unclear and is not related directly reassociated to mortality or outcomes 53 , 54 . Overall, there was broad agreement between AI metrics and pathologist steatosis grading for both MASLD and MASH cases. However, a striking feature of our results is that our SF measurements are under 33%, which does not appear consistent with pathologist steatosis scores greater than 1. Despite this, low SF measurements would match that of low MRI-PDFF that is often recorded 55 . There is evidence of pathologists overestimating steatosis, especially in severe cases 56 which may contribute to this observed difference. The clinical impact of steatosis not directly associated to mortality or outcomes 53 , 54 . There is evidence of pathologists overestimating steatosis, especially in severe cases 56 which may contribute to this observed difference, however a direct comparison of our systems measurements with steatosis grades should be interpreted with caution due to the differences in measurement methods. The clinical impact of steatosis not directly associated to mortality or outcomes 53 , 54 , however in drug efficacy trials for MASH reduction in liver steatosis can lead to disease activity resolution 53 , 54 . This AI system could support manual grading of steatosis in such trials. Overall, our findings show that, as expected, the total level of inflammation is elevated in AIH compared to MASLD/MASH and that AI metrics inflammation correlate with manual grading of inflammation, as long as there is substantial inflammatory cell burden. AI metrics for inflammation (for both ICN and IB) provided more quantitative information of regional distribution than is possible from manual pathology. We found that cases with the same pathologist score may have very different amounts of inflammation depending on the disease aetiology. For example, a portal inflammatory score (Ishak) of 3 or 4 for an AIH case may have considerably higher ICN or IB than a MASLD/MASH case with the same inflammation grade. A pattern of IB was observed in both AIH and MASLD/MASH where inflammation was greater in interface regions and higher 700-900μm from the PT. Similarly, the ICN tended to reduce with increasing distance from the PT with a slight peak of inflammation from 500-600μm from PT. This regional distribution may reflect the biology of the localised inflammatory response. In portal and interface regions, ICN and IB generally increase with higher pathologist scores. In lobular regions there is no correlation, but this may reflect both the notoriously low inter-rater variability for lobular inflammation and the low levels of inflammation found in lobular regions of MASLD/MASH cases 57 . Although portal inflammation is not typically included classical MASLD/MASH scoring systems such as NAS, it is frequently linked with disease progression 58 – 61 and used used to support the disease resolution decisions in MASH trial endpoints. The scoring systems for chronic hepatitis include both portal and lobular scorings in addition to the portal-parenchymal interface. This periportal inflammation correlates well with fibrosis progression in biopsies in chronic hepatitis C 62 and understanding the potential role of interface hepatitis in MASH progression requires further investigation 59 . A study by Balitzer et al. revealed that without the addition of new stains as well as a change to inflammatory evaluation, the current biopsy methodology for scoring AIH cases led to under-scoring, especially for inflammatory activity 63 ; 17% of cases would have been mis-classified as non-AIH. Using our system, the accurate quantification of inflammation at the interface region and graphical identification of PT with higher levels of inflammation will allow pathologists to more easily measure and identify interface hepatitis, avoid misclassification and potentially identify incidental findings. Further training and test datasets will improve our current system in addition to comparing our system measurements with additional pathologists and scorings. The septa and connective tissue are outlined in highly fibrotic tissue with our AI system, however, this may skew PT enumeration. AI metrics shown to correlate with pathologist grading herein, should nevertheless be interpreted with some caution due to the differences in measurement methods. Furthermore, integration of these tools with further pathological feature detection, such as hepatocyte ballooning, will provide pathologists with vital information. Even though the AIH cases were fewer than the MASLD/MASH cases in this study, we consistently found elevated levels of inflammation in AIH compared to MASLD/MASH. Our AIH cohort was predominantly pediatric (79%) and may exhibit fewer typical disease features than that found in adults 64 . With further dataset training across different populations and multiple clinical trials, our system will be more accurate at spatially analyzing detected features to determine which biological elements of a portal tract and surrounding elements are present in a given region of tissue. Conclusion Portal tract characterization is fundamental for histological scoring and establishing sample adequacy of liver biopsies. This AI system generates detailed visual overlays to assist pathological review, as well as producing automated quantifications of PT, inflammation and steatosis across portal, interface and lobular regions. AI systems show promise and could improve the reproducibility and accuracy of liver biopsy evaluation and consensus decisions in clinical trials. While the use of non-invasive approaches to quantify inflammation are becoming more prevalent, the low levels of IB and ICN that are present in MASLD and MASH will require the high degree of sensitivity as afforded by AI to measure small, anticipated effect sizes that are likely when assessing the impact of anti-inflammatory treatments. Data availability statement The portal tract identification system described for the first time in this study is the intellectual property of Perspectum Ltd and the subject of the Patent Cooperation Treaty PCT/IB2025/0530307. Summary data is included in the manuscript or uploaded as online supplemental information. Anonymised individual patient data can be shared upon request or as required by law and/or regulation with qualified external researchers. Approval of such requests is at the discretion of the study sponsors and is dependent on the nature of the request, the merit of the research proposed, the availability of the data, and the intended use of the data. Conflicts of interest DW/AM/CB/HTB/SL/KH/PA/RK/PW and CL are employees at Perspectum Ltd. KF/EF/TK and RG are consultants for Perspectum Ltd. All other co-authors have no conflicts of interest to declare relevant to this work. Manuscript includes use of data generated via collaboration with; Yokohama City University Hospital, CymaBay Therapeutics Inc (acquired by Gilead Sciences), University of Birmingham and Children’s Memorial Health Institute in Warsaw (IPCZD). Two trials of which data was utilised for this study were sponsored by Perspectum Ltd (Study 3&4). Please refer to the accompanying ICMJE disclosure forms for further details. Financial Support Statement This paper presents data from independent research funded by Innovate UK (Project number: 101679) and the Eureka Eurostars 2 Grant (E!10124). Author Contributions Conceptualization – DW/AM (Equal). Writing (original draft) – DW (Lead), AM (Supporting), SL (Supporting). Data curation – AM (Lead), CB/KHDW/KF/EF/TK/RG (Supporting). Investigation – DW (Lead), AM/CB (Supporting). Formal analysis – AM/CB (Equal). Preparation – DW (Lead), AM/CB/SL (Supporting), Visualisation – AM (Lead), CB/DW (Supporting). Review & Editing – DW/HTB (Equal), AM/HTB/SL/PA/KF/EF/TK/RK/PW/CL/PB/RG (Supporting). Acknowledgements We would like to thank the patients recruited into the studies used in this article for their participation. Study 1 was registered with registered with clinicaltrials.gov ( NCT03551522 ). Study 2 was registered with UMIN clinical trials registry (UMIN000026145). Study 3 was registered with the ISRCTN registry (ISCRTN39463479) and the National Institute for Health Research portfolio (15912). The collection of study 3 pathology data was funded by Innovate UK (Project number: 101679). Study 4 was registered with clinicaltrials.gov ( NCT03198104 ) and funded by the Eureka Eurostars 2 Grant (E!10124). List of abbreviations AI Artificial Intelligence AIH Autoimmune Hepatitis DL Deep Learning H&E Haematoxylin and Eosin IB Inflammatory Burden ICN Inflammatory Cell Number SF Steatosis Fraction MASLD Metabolic Dysfunction-associated Steatotic Liver Disease MASH Metabolic Dysfunction-associated Steatohepatitis ML Machine Learning MRI Magnetic Resonance Imaging NAS NAFLD Activity Score PDFF Proton Density Fat-Fraction PT Portal Tract(s) WSI Whole Slide Image(s) References 1. ↵ Harrison SA , Gawrieh S , Roberts K , et al. Prospective evaluation of the prevalence of non-alcoholic fatty liver disease and steatohepatitis in a large middle-aged US cohort . Journal of hepatology . 2021 ; 75 ( 2 ): 284 – 291 . OpenUrl PubMed 2. ↵ Rinella ME , Lazarus JV , Ratziu V , et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature . Hepatology . 2023 ; 78 ( 6 ): 1966 – 1986 . OpenUrl CrossRef PubMed 3. ↵ Younossi ZM , Golabi P , Paik JM , Henry A , Van Dongen C , Henry L . The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review . Hepatology . 2023 ; 77 ( 4 ): 1335 – 1347 . OpenUrl CrossRef PubMed 4. ↵ Le P , Tatar M , Dasarathy S , et al. Estimated Burden of Metabolic Dysfunction–Associated Steatotic Liver Disease in US Adults, 2020 to 2050 . JAMA Network Open . 2025 ; 8 ( 1 ): e2454707 – e2454707 . OpenUrl 5. ↵ Hahn JW , Yang HR , Moon JS , et al. Global incidence and prevalence of autoimmune hepatitis, 1970–2022: a systematic review and meta-analysis . EClinicalMedicine . 2023 ; 65 6. ↵ Liver EAftSoT . EASL clinical practice guidelines: autoimmune hepatitis . Journal of hepatology . 2015 ; 63 ( 4 ): 971 – 1004 . OpenUrl CrossRef PubMed 7. Dalekos GN , Gatselis NK , Zachou K , Koukoulis GK . NAFLD and autoimmune hepatitis: do not judge a book by its cover . European journal of internal medicine . 2020 ; 75 : 1 – 9 . OpenUrl PubMed 8. ↵ Rinella ME , Neuschwander-Tetri BA , Siddiqui MS , et al. AASLD Practice Guidance on the clinical assessment and management of nonalcoholic fatty liver disease . Hepatology . 2023 ; 77 ( 5 ): 1797 – 1835 . OpenUrl CrossRef PubMed 9. ↵ Harrison S , Bedossa P , Guy C , et al. A Randomized Controlled Phase 3 Trial of Resmetirom in NASH with Liver Fibrosis . New England Journal of Medicine . 2024 ; 10. ↵ Loomba R , Hartman ML , Lawitz EJ , et al. Tirzepatide for metabolic dysfunction–associated steatohepatitis with liver fibrosis . New England Journal of Medicine . 2024 ; 391 ( 4 ): 299 – 310 . OpenUrl CrossRef PubMed 11. ↵ Francque SM , Bedossa P , Ratziu V , et al. A randomized, controlled trial of the pan-PPAR agonist lanifibranor in NASH . New England Journal of Medicine . 2021 ; 385 ( 17 ): 1547 – 1558 . OpenUrl CrossRef PubMed 12. ↵ Loomba R , Lawitz E , Mantry PS , et al. The ASK1 inhibitor selonsertib in patients with nonalcoholic steatohepatitis: a randomized, phase 2 trial . Hepatology . 2018 ; 67 ( 2 ): 549 – 559 . OpenUrl CrossRef PubMed 13. ↵ Harrison SA , Ruane PJ , Freilich B , et al. A randomized, double-blind, placebo-controlled phase IIa trial of efruxifermin for patients with compensated NASH cirrhosis . Jhep Reports . 2023 ; 5 ( 1 ): 100563 . OpenUrl PubMed 14. ↵ Sanyal AJ , Chalasani N , Kowdley KV , et al. Pioglitazone, vitamin E, or placebo for nonalcoholic steatohepatitis . New England Journal of Medicine . 2010 ; 362 ( 18 ): 1675 – 1685 . OpenUrl CrossRef PubMed Web of Science 15. Sanyal AJ , Brunt EM , Kleiner DE , et al. Endpoints and clinical trial design for nonalcoholic steatohepatitis . Wiley Online Library ; 2011 . 16. Harrison SA , Bashir MR , Guy CD , et al. Resmetirom (MGL-3196) for the treatment of non-alcoholic steatohepatitis: a multicentre, randomised, double-blind, placebo-controlled, phase 2 trial . The Lancet . 2019 ; 394 ( 10213 ): 2012 – 2024 . OpenUrl 17. ↵ Tai D , Bashir M , Taub R , Ren Y , Chng E , Harrison S . Impact of resmetirom-mediated reductions in liver volume and steatosis compared with placebo on the quantification of fibrosis using second harmonic generation in a serial liver biopsy study . J Hepatol . 2022 ; 77 ( Suppl 1 ): S32 – 3 . OpenUrl 18. ↵ Brunt EM , Clouston AD , Goodman Z , et al. Complexity of ballooned hepatocyte feature recognition: defining a training atlas for artificial intelligence-based imaging in NAFLD . Journal of hepatology . 2022 ; 76 ( 5 ): 1030 – 1041 . OpenUrl PubMed 19. ↵ Boyd A , Cain O , Chauhan A , Webb GJ . Medical liver biopsy: background, indications, procedure and histopathology . Frontline gastroenterology . 2020 ; 11 ( 1 ): 40 – 47 . OpenUrl Abstract / FREE Full Text 20. ↵ Fryer E , Wang LM , Verrill C , Fleming K . How often do our liver core biopsies reach current definitions of adequacy? Journal of clinical pathology . 2013 ; 66 ( 12 ): 1087 – 1089 . OpenUrl Abstract / FREE Full Text 21. ↵ Langford CR , Goldinger MH , Treanor D , et al. Improved pathology reporting in NAFLD/NASH for clinical trials . Journal of clinical pathology . 2022 ; 75 ( 2 ): 73 – 75 . OpenUrl FREE Full Text 22. ↵ Crawford AR , Lin X-Z , Crawford JM . The normal adult human liver biopsy: a quantitative reference standard . Hepatology . 1998 ; 28 ( 2 ): 323 – 331 . OpenUrl CrossRef PubMed Web of Science 23. ↵ Sporea I , Gherhardt D , Popescu A , et al. Does the size of the needle influence the number of portal tracts obtained through percutaneous liver biopsy? Annals of hepatology . 2015 ; 11 ( 5 ): 691 – 695 . OpenUrl 24. ↵ Brunt EM . Liver biopsy reliability in clinical trials: thoughts from a liver pathologist . Journal of hepatology . 2020 ; 73 ( 6 ): 1310 – 1312 . OpenUrl PubMed 25. Lambrecht RW , Sterling RK , Naishadham D , et al. Iron levels in hepatocytes and portal tract cells predict progression and outcomes of patients with advanced chronic hepatitis C . Gastroenterology . 2011 ; 140 ( 5 ): 1490 – 1500 . e3. OpenUrl CrossRef PubMed Web of Science 26. ↵ Brunt EM , Janney CG , Di Bisceglie AM , Neuschwander-Tetri BA , Bacon BR . Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions . Official journal of the American College of Gastroenterology| ACG . 1999 ; 94 ( 9 ): 2467 – 2474 . OpenUrl 27. ↵ Davison BA , Harrison SA , Cotter G , et al. Suboptimal reliability of liver biopsy evaluation has implications for randomized clinical trials . Journal of hepatology . 2020 ; 73 ( 6 ): 1322 – 1332 . OpenUrl PubMed 28. ↵ Perveen S , Shahbaz M , Keshavjee K , Guergachi A . A systematic machine learning based approach for the diagnosis of non-alcoholic fatty liver disease risk and progression . Scientific reports . 2018 ; 8 ( 1 ): 2112 . OpenUrl PubMed 29. ↵ Yu H , Sharifai N , Jiang K , et al. Artificial intelligence based liver portal tract region identification and quantification with transplant biopsy whole-slide images . Computers in Biology and Medicine . 2022 ; 150 : 106089 . OpenUrl PubMed 30. ↵ Ratziu V , Hompesch M , Petitjean M , et al. Digital pathology and artificial intelligence in non-alcoholic steatohepatitis: current status and future directions . Journal of hepatology . 2023 ; 31. Sanyal AJ , Jha P , Kleiner DE . Digital pathology for nonalcoholic steatohepatitis assessment . Nature Reviews Gastroenterology & Hepatology . 2024 ; 21 ( 1 ): 57 – 69 . OpenUrl PubMed 32. Forlano R , Mullish BH , Giannakeas N , et al. High-throughput, machine learning–based quantification of steatosis, inflammation, ballooning, and fibrosis in biopsies from patients with nonalcoholic fatty liver disease . Clinical Gastroenterology and Hepatology . 2020 ; 18 ( 9 ): 2081 – 2090 . e9. OpenUrl 33. Liu F , Goh GBB , Tiniakos D , et al. qFIBS: an automated technique for quantitative evaluation of fibrosis, inflammation, ballooning, and steatosis in patients with nonalcoholic steatohepatitis . Hepatology . 2020 ; 71 ( 6 ): 1953 – 1966 . OpenUrl CrossRef PubMed 34. Munsterman ID , van Erp M , Weijers G , et al. A novel automatic digital algorithm that accurately quantifies steatosis in NAFLD on histopathological whole-slide images . Cytometry Part B: Clinical Cytometry . 2019 ; 96 ( 6 ): 521 – 528 . OpenUrl PubMed 35. Vanderbeck S , Bockhorst J , Kleiner D , Komorowski R , Chalasani N , Gawrieh S . Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies . Human pathology . 2015 ; 46 ( 5 ): 767 – 775 . OpenUrl PubMed 36. ↵ Pulaski H , Harrison SA , Mehta SS , et al. Clinical validation of an AI-based pathology tool for scoring of metabolic dysfunction-associated steatohepatitis . Nature Medicine . 2024 : 1 – 8 . 37. ↵ Tsai H-W , Chiou C-Y , Yang W-J , et al. Lymphocyte-Infiltrated Periportal Region Detection With Structurally-Refined Deep Portal Segmentation and Heterogeneous Infiltration Features . IEEE Open Journal of Engineering in Medicine and Biology . 2024 ; 38. ↵ Andersson A , Kelly M , Imajo K , et al. Clinical utility of MRI biomarkers for identifying NASH patients’ high risk of progression: A multi-center pooled data and meta-analysis . Clinical Gastroenterology and Hepatology . 2021 ; 39. Socha P , Shumbayawonda E , Roy A , et al. Quantitative digital pathology enables automated and quantitative assessment of inflammatory activity in patients with autoimmune hepatitis . Journal of Pathology Informatics . 2024 ; 15 : 100372 . OpenUrl PubMed 40. ↵ Dennis A , Kelly MD , Fernandes C , et al. Correlations between MRI biomarkers PDFF and cT1 with histopathological features of non-alcoholic steatohepatitis . Frontiers in endocrinology . 2021 ; 11 : 575843 . OpenUrl PubMed 41. ↵ Imajo K , Tetlow L , Dennis A , et al. Quantitative multiparametric magnetic resonance imaging can aid non-alcoholic steatohepatitis diagnosis in a Japanese cohort . World journal of gastroenterology . 2021 ; 27 ( 7 ): 609 . OpenUrl PubMed 42. ↵ McDonald N , Eddowes PJ , Hodson J , et al. Multiparametric magnetic resonance imaging for quantitation of liver disease: a two-centre cross-sectional observational study . Scientific reports . 2018 ; 8 ( 1 ): 9189 . OpenUrl PubMed 43. ↵ Janowski K , Shumbayawonda E , Cheng L , et al. Quantitative multiparametric MRI as a non-invasive stratification tool in children and adolescents with autoimmune liver disease . Scientific Reports . 2021 ; 11 ( 1 ): 15261 . OpenUrl PubMed 44. ↵ Alkhouri N , Beyer C , Shumbayawonda E , et al. Decreases in cT1 and liver fat content reflect treatment-induced histological improvements in MASH . Journal of Hepatology . 2025 ; 82 ( 3 ): 438 – 445 . OpenUrl PubMed 45. ↵ Heydtmann M , Davies S , Hübscher S , Wyatt J . Guidelines on the use of liver biopsy in clinical practice from the British Society of Gastroenterology, the Royal College of Radiologists and the Royal College of Pathology . Gut . 2020 ; 1382 : 69 . OpenUrl 46. ↵ Knodell RG , Ishak KG , Black WC , et al. Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis . Hepatology . 1981 ; 1 ( 5 ): 431 – 435 . OpenUrl CrossRef PubMed Web of Science 47. ↵ Brunt EM , Kleiner DE , Wilson LA , Belt P , Neuschwander-Tetri BA , Network NCR . Nonalcoholic fatty liver disease (NAFLD) activity score and the histopathologic diagnosis in NAFLD: distinct clinicopathologic meanings . Hepatology . 2011 ; 53 ( 3 ): 810 – 820 . OpenUrl CrossRef PubMed Web of Science 48. ↵ Vollmar B , Wolf B , Siegmund S , Katsen AD , Menger MD . Lymph vessel expansion and function in the development of hepatic fibrosis and cirrhosis . The American journal of pathology . 1997 ; 151 ( 1 ): 169 . OpenUrl PubMed Web of Science 49. ↵ Cholongitas E , Senzolo M , Standish R , et al. A systematic review of the quality of liver biopsy specimens . American journal of clinical pathology . 2006 ; 125 ( 5 ): 710 – 721 . OpenUrl CrossRef PubMed 50. ↵ Brunt EM . Nonalcoholic steatohepatitis . Copyright© 2004 by Thieme Medical Publishers, Inc ., 333 Seventh Avenue, New …; 2004 : 3 – 20 . 51. ↵ Dempsey JL , Ioannou GN , Carr RM . Mechanisms of lipid droplet accumulation in steatotic liver diseases . Thieme Medical Publishers, Inc .; 2023 : 52. ↵ Gluchowski NL , Becuwe M , Walther TC , Farese Jr RV . Lipid droplets and liver disease: from basic biology to clinical implications . Nature reviews Gastroenterology & hepatology . 2017 ; 14 ( 6 ): 343 – 355 . OpenUrl PubMed 53. ↵ Angulo P , Kleiner DE , Dam-Larsen S , et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease . Gastroenterology . 2015 ; 149 ( 2 ): 389 – 397 . e10. OpenUrl CrossRef PubMed 54. ↵ Younossi ZM , Mangla KK , Berentzen TL , et al. Liver histology is associated with long-term clinical outcomes in patients with metabolic dysfunction–associated steatohepatitis . Hepatology Communications . 2024 ; 8 ( 6 ): e0423 . OpenUrl 55. ↵ Caussy C , Reeder SB , Sirlin CB , Loomba R . Noninvasive, quantitative assessment of liver fat by MRI-PDFF as an endpoint in NASH trials . Hepatology . 2018 ; 68 ( 2 ): 763 – 772 . OpenUrl CrossRef PubMed 56. ↵ Hall AR , Dhillon AP , Green AC , et al. Hepatic steatosis estimated microscopically versus digital image analysis . Liver International . 2013 ; 33 ( 6 ): 926 – 935 . OpenUrl PubMed 57. ↵ Fukusato T , Fukushima J , Shiga J , et al. Interobserver variation in the histopathological assessment of nonalcoholic steatohepatitis . Hepatology research . 2005 ; 33 ( 2 ): 122 – 127 . OpenUrl PubMed 58. ↵ Brunt EM , Kleiner DE , Wilson LA , et al. Portal chronic inflammation in nonalcoholic fatty liver disease (NAFLD): a histologic marker of advanced NAFLD— clinicopathologic correlations from the nonalcoholic steatohepatitis clinical research network . Hepatology . 2009 ; 49 ( 3 ): 809 – 820 . OpenUrl CrossRef PubMed Web of Science 59. ↵ Dhingra S , Mahadik JD , Tarabishy Y , May SB , Vierling JM . Prevalence and clinical significance of portal inflammation, portal plasma cells, interface hepatitis and biliary injury in liver biopsies from patients with non-alcoholic steatohepatitis . Pathology . 2022 ; 54 ( 6 ): 686 – 693 . OpenUrl PubMed 60. Carotti S , Vespasiani-Gentilucci U , Perrone G , Picardi A , Morini S . Portal inflammation during NAFLD is frequent and associated with the early phases of putative hepatic progenitor cell activation . Journal of Clinical Pathology . 2015 ; 68 ( 11 ): 883 – 890 . OpenUrl Abstract / FREE Full Text 61. ↵ Rakha EA , Adamson L , Bell E , et al. Portal inflammation is associated with advanced histological changes in alcoholic and non-alcoholic fatty liver disease . Journal of Clinical Pathology . 2010 ; 63 ( 9 ): 790 – 795 . OpenUrl Abstract / FREE Full Text 62. ↵ Ghany MG , Kleiner DE , Alter H , et al. Progression of fibrosis in chronic hepatitis C . Gastroenterology . 2003 ; 124 ( 1 ): 97 – 104 . OpenUrl CrossRef PubMed Web of Science 63. ↵ Balitzer D , Shafizadeh N , Peters MG , Ferrell LD , Alshak N , Kakar S . Autoimmune hepatitis: review of histologic features included in the simplified criteria proposed by the international autoimmune hepatitis group and proposal for new histologic criteria . Modern pathology . 2017 ; 30 ( 5 ): 773 – 783 . OpenUrl CrossRef PubMed 64. ↵ Mau B , Hakar M , Lin HC , Davis JL . A review of histopathologic features of pediatric autoimmune liver disease . Clinical Liver Disease . 2022 ; 20 ( 4 ): 116 – 119 . OpenUrl PubMed View the discussion thread. Back to top Previous Next Posted April 25, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH Dylan Windell , Alastair Magness , Cayden Beyer , Helena Thomaides Brears , Sarah Larkin , Kezia Hobson , Paul Aljabar , Kenneth Fleming , Eve Fryer , Timothy Kendall , Reema Kainth , Phil Wakefield , Caitlin Langford , Pierre Bedossa , Robert Goldin medRxiv 2025.04.23.25326290; doi: https://doi.org/10.1101/2025.04.23.25326290 Share This Article: Copy Citation Tools AI portal tract detection and characterisation for a regional analysis of steatosis and inflammation in MASLD, MASH, and AIH Dylan Windell , Alastair Magness , Cayden Beyer , Helena Thomaides Brears , Sarah Larkin , Kezia Hobson , Paul Aljabar , Kenneth Fleming , Eve Fryer , Timothy Kendall , Reema Kainth , Phil Wakefield , Caitlin Langford , Pierre Bedossa , Robert Goldin medRxiv 2025.04.23.25326290; doi: https://doi.org/10.1101/2025.04.23.25326290 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Pathology Subject Areas All Articles Addiction Medicine (568) Allergy and Immunology (863) Anesthesia (300) Cardiovascular Medicine (4435) Dentistry and Oral Medicine (444) Dermatology (382) Emergency Medicine (608) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1509) Epidemiology (15227) Forensic Medicine (30) Gastroenterology (1124) Genetic and Genomic Medicine (6597) Geriatric Medicine (668) Health Economics (997) Health Informatics (4534) Health Policy (1368) Health Systems and Quality Improvement (1613) Hematology (540) HIV/AIDS (1264) Infectious Diseases (except HIV/AIDS) (15916) Intensive Care and Critical Care Medicine (1103) Medical Education (623) Medical Ethics (146) Nephrology (667) Neurology (6599) Nursing (346) Nutrition (998) Obstetrics and Gynecology (1144) Occupational and Environmental Health (957) Oncology (3332) Ophthalmology (974) Orthopedics (369) Otolaryngology (420) Pain Medicine (436) Palliative Medicine (130) Pathology (663) Pediatrics (1693) Pharmacology and Therapeutics (691) Primary Care Research (711) Psychiatry and Clinical Psychology (5447) Public and Global Health (9230) Radiology and Imaging (2198) Rehabilitation Medicine and Physical Therapy (1370) Respiratory Medicine (1196) Rheumatology (593) Sexual and Reproductive Health (712) Sports Medicine (530) Surgery (712) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a002ab5fcedb4807',t:'MTc3OTUyNDcyMA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();