Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples

Luisa Ricaurte Archila; Lindsey Smith; Hanna Kaisa Sihvo; Ville Koponen; Sarah M. Jenkins; Donnchadh M. O'Sullivan; Maria Camila Cardenas Fernandez; Yaohong Wang; Priyadharshini Sivasubramaniam; Ameya Patil; Puanani E. Hopson; Imad Absah; Karthik Ravi; Taofic Mounajjed; Evan S. Dellon; Albert J. Bredenoord; Rish Pai; Christopher P. Hartley; Rondell P. Graham; Roger K. Moreira

doi:10.1016/j.modpat.2023.100285

Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples

Luisa Ricaurte Archila, Lindsey Smith, Hanna Kaisa Sihvo, Ville Koponen, Sarah M. Jenkins, Donnchadh M. O'Sullivan, Maria Camila Cardenas Fernandez, Yaohong Wang, Priyadharshini Sivasubramaniam, Ameya Patil, Puanani E. Hopson, Imad Absah, Karthik Ravi, Taofic Mounajjed, Evan S. Dellon, Albert J. Bredenoord, Rish Pai, Christopher P. Hartley, Rondell P. Graham, Roger K. Moreira

Research output: Contribution to journal › Article › peer-review

Abstract

We have developed an artificial intelligence (AI)–based digital pathology model for the evaluation of histologic features related to eosinophilic esophagitis (EoE). In this study, we evaluated the performance of our AI model in a cohort of pediatric and adult patients for histologic features included in the Eosinophilic Esophagitis Histologic Scoring System (EoEHSS). We collected a total of 203 esophageal biopsy samples from patients with mucosal eosinophilia of any degree (91 adult and 112 pediatric patients) and 10 normal controls from a prospectively maintained database. All cases were assessed by a specialized gastrointestinal (GI) pathologist for features in the EoEHSS at the time of original diagnosis and rescored by a central GI pathologist (R.K.M.). We subsequently analyzed whole-slide image digital slides using a supervised AI model operating in a cloud-based, deep learning AI platform (Aiforia Technologies) for peak eosinophil count (PEC) and several histopathologic features in the EoEHSS. The correlation and interobserver agreement between the AI model and pathologists (Pearson correlation coefficient [r_s] = 0.89 and intraclass correlation coefficient [ICC] = 0.87 vs original pathologist; r_s = 0.91 and ICC = 0.83 vs central pathologist) were similar to the correlation and interobserver agreement between pathologists for PEC (r_s = 0.88 and ICC = 0.91) and broadly similar to those for most other histologic features in the EoEHSS. The AI model also accurately identified PEC of >15 eosinophils/high-power field by the original pathologist (area under the curve [AUC] = 0.98) and central pathologist (AUC = 0.98) and had similar AUCs for the presence of EoE-related endoscopic features to pathologists’ assessment. Average eosinophils per epithelial unit area had similar performance compared to AI high-power field–based analysis. Our newly developed AI model can accurately identify, quantify, and score several of the main histopathologic features in the EoE spectrum, with agreement regarding EoEHSS scoring which was similar to that seen among GI pathologists.

Original language	English (US)
Article number	100285
Journal	Modern Pathology
Volume	36
Issue number	10
DOIs	https://doi.org/10.1016/j.modpat.2023.100285
State	Published - Oct 2023

Keywords

artificial intelligence
deep learning
digital pathology
EoE
eosinophilic esophagitis
eosinophils

ASJC Scopus subject areas

General Medicine

Access to Document

10.1016/j.modpat.2023.100285

Cite this

Ricaurte Archila, L., Smith, L., Sihvo, H. K., Koponen, V., Jenkins, S. M., O'Sullivan, D. M., Cardenas Fernandez, M. C., Wang, Y., Sivasubramaniam, P., Patil, A., Hopson, P. E., Absah, I., Ravi, K., Mounajjed, T., Dellon, E. S., Bredenoord, A. J., Pai, R., Hartley, C. P., Graham, R. P., & Moreira, R. K. (2023). Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples. Modern Pathology, 36(10), Article 100285. https://doi.org/10.1016/j.modpat.2023.100285

Ricaurte Archila, L, Smith, L, Sihvo, HK, Koponen, V, Jenkins, SM, O'Sullivan, DM, Cardenas Fernandez, MC, Wang, Y, Sivasubramaniam, P, Patil, A, Hopson, PE, Absah, I, Ravi, K, Mounajjed, T, Dellon, ES, Bredenoord, AJ, Pai, R, Hartley, CP, Graham, RP & Moreira, RK 2023, 'Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples', Modern Pathology, vol. 36, no. 10, 100285. https://doi.org/10.1016/j.modpat.2023.100285

@article{56471350982243e4adcba7dde120d792,

title = "Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples",

abstract = "We have developed an artificial intelligence (AI)–based digital pathology model for the evaluation of histologic features related to eosinophilic esophagitis (EoE). In this study, we evaluated the performance of our AI model in a cohort of pediatric and adult patients for histologic features included in the Eosinophilic Esophagitis Histologic Scoring System (EoEHSS). We collected a total of 203 esophageal biopsy samples from patients with mucosal eosinophilia of any degree (91 adult and 112 pediatric patients) and 10 normal controls from a prospectively maintained database. All cases were assessed by a specialized gastrointestinal (GI) pathologist for features in the EoEHSS at the time of original diagnosis and rescored by a central GI pathologist (R.K.M.). We subsequently analyzed whole-slide image digital slides using a supervised AI model operating in a cloud-based, deep learning AI platform (Aiforia Technologies) for peak eosinophil count (PEC) and several histopathologic features in the EoEHSS. The correlation and interobserver agreement between the AI model and pathologists (Pearson correlation coefficient [rs] = 0.89 and intraclass correlation coefficient [ICC] = 0.87 vs original pathologist; rs = 0.91 and ICC = 0.83 vs central pathologist) were similar to the correlation and interobserver agreement between pathologists for PEC (rs = 0.88 and ICC = 0.91) and broadly similar to those for most other histologic features in the EoEHSS. The AI model also accurately identified PEC of >15 eosinophils/high-power field by the original pathologist (area under the curve [AUC] = 0.98) and central pathologist (AUC = 0.98) and had similar AUCs for the presence of EoE-related endoscopic features to pathologists{\textquoteright} assessment. Average eosinophils per epithelial unit area had similar performance compared to AI high-power field–based analysis. Our newly developed AI model can accurately identify, quantify, and score several of the main histopathologic features in the EoE spectrum, with agreement regarding EoEHSS scoring which was similar to that seen among GI pathologists.",

keywords = "artificial intelligence, deep learning, digital pathology, EoE, eosinophilic esophagitis, eosinophils",

author = "{Ricaurte Archila}, Luisa and Lindsey Smith and Sihvo, {Hanna Kaisa} and Ville Koponen and Jenkins, {Sarah M.} and O'Sullivan, {Donnchadh M.} and {Cardenas Fernandez}, {Maria Camila} and Yaohong Wang and Priyadharshini Sivasubramaniam and Ameya Patil and Hopson, {Puanani E.} and Imad Absah and Karthik Ravi and Taofic Mounajjed and Dellon, {Evan S.} and Bredenoord, {Albert J.} and Rish Pai and Hartley, {Christopher P.} and Graham, {Rondell P.} and Moreira, {Roger K.}",

note = "Publisher Copyright: {\textcopyright} 2023",

year = "2023",

month = oct,

doi = "10.1016/j.modpat.2023.100285",

language = "English (US)",

volume = "36",

journal = "Modern Pathology",

issn = "0893-3952",

publisher = "Nature Publishing Group",

number = "10",

}

TY - JOUR

T1 - Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples

AU - Ricaurte Archila, Luisa

AU - Smith, Lindsey

AU - Sihvo, Hanna Kaisa

AU - Koponen, Ville

AU - Jenkins, Sarah M.

AU - O'Sullivan, Donnchadh M.

AU - Cardenas Fernandez, Maria Camila

AU - Wang, Yaohong

AU - Sivasubramaniam, Priyadharshini

AU - Patil, Ameya

AU - Hopson, Puanani E.

AU - Absah, Imad

AU - Ravi, Karthik

AU - Mounajjed, Taofic

AU - Dellon, Evan S.

AU - Bredenoord, Albert J.

AU - Pai, Rish

AU - Hartley, Christopher P.

AU - Graham, Rondell P.

AU - Moreira, Roger K.

PY - 2023/10

Y1 - 2023/10

N2 - We have developed an artificial intelligence (AI)–based digital pathology model for the evaluation of histologic features related to eosinophilic esophagitis (EoE). In this study, we evaluated the performance of our AI model in a cohort of pediatric and adult patients for histologic features included in the Eosinophilic Esophagitis Histologic Scoring System (EoEHSS). We collected a total of 203 esophageal biopsy samples from patients with mucosal eosinophilia of any degree (91 adult and 112 pediatric patients) and 10 normal controls from a prospectively maintained database. All cases were assessed by a specialized gastrointestinal (GI) pathologist for features in the EoEHSS at the time of original diagnosis and rescored by a central GI pathologist (R.K.M.). We subsequently analyzed whole-slide image digital slides using a supervised AI model operating in a cloud-based, deep learning AI platform (Aiforia Technologies) for peak eosinophil count (PEC) and several histopathologic features in the EoEHSS. The correlation and interobserver agreement between the AI model and pathologists (Pearson correlation coefficient [rs] = 0.89 and intraclass correlation coefficient [ICC] = 0.87 vs original pathologist; rs = 0.91 and ICC = 0.83 vs central pathologist) were similar to the correlation and interobserver agreement between pathologists for PEC (rs = 0.88 and ICC = 0.91) and broadly similar to those for most other histologic features in the EoEHSS. The AI model also accurately identified PEC of >15 eosinophils/high-power field by the original pathologist (area under the curve [AUC] = 0.98) and central pathologist (AUC = 0.98) and had similar AUCs for the presence of EoE-related endoscopic features to pathologists’ assessment. Average eosinophils per epithelial unit area had similar performance compared to AI high-power field–based analysis. Our newly developed AI model can accurately identify, quantify, and score several of the main histopathologic features in the EoE spectrum, with agreement regarding EoEHSS scoring which was similar to that seen among GI pathologists.

AB - We have developed an artificial intelligence (AI)–based digital pathology model for the evaluation of histologic features related to eosinophilic esophagitis (EoE). In this study, we evaluated the performance of our AI model in a cohort of pediatric and adult patients for histologic features included in the Eosinophilic Esophagitis Histologic Scoring System (EoEHSS). We collected a total of 203 esophageal biopsy samples from patients with mucosal eosinophilia of any degree (91 adult and 112 pediatric patients) and 10 normal controls from a prospectively maintained database. All cases were assessed by a specialized gastrointestinal (GI) pathologist for features in the EoEHSS at the time of original diagnosis and rescored by a central GI pathologist (R.K.M.). We subsequently analyzed whole-slide image digital slides using a supervised AI model operating in a cloud-based, deep learning AI platform (Aiforia Technologies) for peak eosinophil count (PEC) and several histopathologic features in the EoEHSS. The correlation and interobserver agreement between the AI model and pathologists (Pearson correlation coefficient [rs] = 0.89 and intraclass correlation coefficient [ICC] = 0.87 vs original pathologist; rs = 0.91 and ICC = 0.83 vs central pathologist) were similar to the correlation and interobserver agreement between pathologists for PEC (rs = 0.88 and ICC = 0.91) and broadly similar to those for most other histologic features in the EoEHSS. The AI model also accurately identified PEC of >15 eosinophils/high-power field by the original pathologist (area under the curve [AUC] = 0.98) and central pathologist (AUC = 0.98) and had similar AUCs for the presence of EoE-related endoscopic features to pathologists’ assessment. Average eosinophils per epithelial unit area had similar performance compared to AI high-power field–based analysis. Our newly developed AI model can accurately identify, quantify, and score several of the main histopathologic features in the EoE spectrum, with agreement regarding EoEHSS scoring which was similar to that seen among GI pathologists.

KW - artificial intelligence

KW - deep learning

KW - digital pathology

KW - EoE

KW - eosinophilic esophagitis

KW - eosinophils

UR - http://www.scopus.com/inward/record.url?scp=85170202872&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85170202872&partnerID=8YFLogxK

U2 - 10.1016/j.modpat.2023.100285

DO - 10.1016/j.modpat.2023.100285

M3 - Article

C2 - 37474003

AN - SCOPUS:85170202872

SN - 0893-3952

VL - 36

JO - Modern Pathology

JF - Modern Pathology

IS - 10

M1 - 100285

ER -

Performance of an Artificial Intelligence Model for Recognition and Quantitation of Histologic Features of Eosinophilic Esophagitis on Biopsy Samples

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this