Digital pathology, what is the future?

Survey of liver pathologists to assess attitudes towards digital pathology and artificial intelligence

AIMS

A survey of members of the UK Liver Pathology Group (UKLPG) was conducted, comprising consultant histopathologists from across the UK who report liver specimens and participate in the UK National Liver Pathology External Quality Assurance scheme. The aim of this study was to understand attitudes and priorities of liver pathologists towards digital pathology and artificial intelligence (AI).

METHODS

The survey was distributed to all full consultant members of the UKLPG via email. This comprised 50 questions, with 48 multiple choice questions and 2 free-text questions at the end, covering a range of topics and concepts pertaining to the use of digital pathology and AI in liver disease.

RESULTS

Forty-two consultant histopathologists completed the survey, representing 36% of fully registered members of the UKLPG (42/116). Questions examining digital pathology showed respondents agreed with the utility of digital pathology for primary diagnosis 83% (34/41), second opinions 90% (37/41), research 85% (35/41) and training and education 95% (39/41). Fatty liver diseases were an area of demand for AI tools with 80% in agreement (33/41), followed by neoplastic liver diseases with 59% in agreement (24/41). Participants were concerned about AI development without pathologist involvement 73% (30/41), however, 63% (26/41) disagreed when asked whether AI would replace pathologists.

CONCLUSIONS

This study outlines current interest, priorities for research and concerns around digital pathology and AI for liver pathologists. The majority of UK liver pathologists are in favour of the application of digital pathology and AI in clinical practice, research and education.

Digital pathology with artificial intelligence analyses provides greater insights into treatment-induced fibrosis regression in NASH

BACKGROUND & AIMS

Liver fibrosis is a key prognostic determinant for clinical outcomes in non-alcoholic steatohepatitis (NASH). Current scoring systems have limitations, especially in assessing fibrosis regression. Second harmonic generation/two-photon excitation fluorescence (SHG/TPEF) microscopy with artificial intelligence analyses provides standardized evaluation of NASH features, especially liver fibrosis and collagen fiber quantitation on a continuous scale. This approach was applied to gain in-depth understanding of fibrosis dynamics after treatment with tropifexor (TXR), a non-bile acid farnesoid X receptor agonist in patients participating in the FLIGHT-FXR study (NCT02855164).

METHOD

Unstained sections from 198 liver biopsies (paired: baseline and end-of-treatment) from 99 patients with NASH (fibrosis stage F2 or F3) who received placebo (n = 34), TXR 140 μg (n = 37), or TXR 200 μg (n = 28) for 48 weeks were examined. Liver fibrosis (qFibrosis®), hepatic fat (qSteatosis®), and ballooned hepatocytes (qBallooning®) were quantitated using SHG/TPEF microscopy. Changes in septa morphology, collagen fiber parameters, and zonal distribution within liver lobules were also quantitatively assessed.

RESULTS

Digital analyses revealed treatment-associated reductions in overall liver fibrosis (qFibrosis®), unlike conventional microscopy, as well as marked regression in perisinusoidal fibrosis in patients who had either F2 or F3 fibrosis at baseline. Concomitant zonal quantitation of fibrosis and steatosis revealed that patients with greater qSteatosis reduction also have the greatest reduction in perisinusoidal fibrosis. Regressive changes in septa morphology and reduction in septa parameters were observed almost exclusively in F3 patients, who were adjudged as 'unchanged' with conventional scoring.

CONCLUSION

Fibrosis regression following hepatic fat reduction occurs initially in the perisinusoidal regions, around areas of steatosis reduction. Digital pathology provides new insights into treatment-induced fibrosis regression in NASH, which are not captured by current staging systems.

LAY SUMMARY

The degree of liver fibrosis (tissue scarring) in non-alcoholic steatohepatitis (NASH) is the main predictor of negative clinical outcomes. Accurate assessment of the quantity and architecture of liver fibrosis is fundamental for patient enrolment in NASH clinical trials and for determining treatment efficacy. Using digital microscopy with artificial intelligence analyses, the present study demonstrates that this novel approach has greater sensitivity in demonstrating treatment-induced reversal of fibrosis in the liver than current systems. Furthermore, additional details are obtained regarding the pathogenesis of NASH disease and the effects of therapy.

Digital Pathology Enables Automated and Quantitative Assessment of Inflammatory Activity in Patients with Chronic Liver Disease

Traditional histological evaluation for grading liver disease severity is based on subjective and semi-quantitative scores. We examined the relationship between digital pathology analysis and corresponding scoring systems for the assessment of hepatic necroinflammatory activity. A prospective, multicenter study including 156 patients with chronic liver disease (74% nonalcoholic fatty liver disease-NAFLD, 26% chronic hepatitis-CH etiologies) was performed. Inflammation was graded according to the Nonalcoholic Steatohepatitis (NASH) Clinical Research Network system and METAVIR score. Whole-slide digital image analysis based on quantitative (I-score: inflammation ratio) and morphometric (C-score: proportionate area of staining intensities clusters) measurements were independently performed. Our data show that I-scores and C-scores increase with inflammation grades (p < 0.001). High correlation was seen for CH (ρ = 0.85-0.88), but only moderate for NAFLD (ρ = 0.5-0.53). I-score (p = 0.008) and C-score (p = 0.002) were higher for CH than NAFLD. Our MATLAB algorithm performed better than QuPath software for the diagnosis of low-moderate inflammation (p < 0.05). C-score AUC for classifying NASH was 0.75 (95%CI, 0.65-0.84) and for moderate/severe CH was 0.99 (95%CI, 0.97-1.00). Digital pathology measurements increased with fibrosis stages (p < 0.001). In conclusion, quantitative and morphometric metrics of inflammatory burden obtained by digital pathology correlate well with pathologists' scores, showing a higher accuracy for the evaluation of CH than NAFLD.

Digital pathology: accurate technique for quantitative assessment of histological features in metabolic-associated fatty liver disease

BACKGROUND

Histological evaluation of metabolic-associated fatty liver disease (MAFLD) biopsies is subjective, descriptive and with interobserver variability.

AIMS

To examine the relationship between different histological features (fibrosis, steatosis, inflammation and iron) measured with automated whole-slide quantitative digital pathology and corresponding semiquantitative scoring systems, and the distribution of digital pathology measurements across Fatty Liver Inhibition of Progression (FLIP) algorithm and Steatosis, Activity and Fibrosis (SAF) scoring system METHODS: We prospectively included 136 consecutive patients who underwent liver biopsy for MAFLD at three Spanish centres (January 2017-January 2020). Biopsies were scored by two blinded pathologists according to the Non-alcoholic Steatohepatitis (NASH) Clinical Research Network system for fibrosis staging, the FLIP/SAF classification for steatosis and inflammation grading and Deugnier score for iron grading. Proportionate areas of collagen, fat, inflammatory cells and iron deposits were measured with computer-assisted digital image analysis. A test-retest experiment was performed for precision repeatability evaluation.

RESULTS

Digital pathology showed strong correlation with fibrosis (r = 0.79; P < 0.001), steatosis (r = 0.85; P < 0.001) and iron (r = 0.70; P < 0.001). Performance was lower when assessing the degree of inflammation (r = 0.35; P < 0.001). NASH cases had a higher proportion of collagen and fat compared to non-NASH cases (P < 0.005), whereas inflammation and iron quantification did not show significant differences between categories. Repeatability evaluation showed that all the coefficients of variation were ≤1.1% and all intraclass correlation coefficient values were ≥0.99, except those of collagen.

CONCLUSION

Digital pathology allows an automated, precise, objective and quantitative assessment of MAFLD histological features. Digital analysis measurements show good concordance with pathologists´ scores.