Cargando…

A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis

BACKGROUND & AIMS: Magnetic resonance cholangiopancreatography (MRCP) for evaluation of biliary disease currently relies on subjective assessment with limited prognostic value because of the lack of quantitative metrics. Artificial intelligence-enabled quantitative MRCP (MRCP+) is a novel techni...

Descripción completa

Detalles Bibliográficos
Autores principales: Vuppalanchi, Raj, Are, Vijay, Telford, Alison, Young, Liam, Mouchti, Sofia, Ferreira, Carlos, Kettler, Carla, Gromski, Mark, Akisik, Fatih, Chalasani, Naga
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Elsevier 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10472223/
https://www.ncbi.nlm.nih.gov/pubmed/37663118
http://dx.doi.org/10.1016/j.jhepr.2023.100834
_version_ 1785100030241144832
author Vuppalanchi, Raj
Are, Vijay
Telford, Alison
Young, Liam
Mouchti, Sofia
Ferreira, Carlos
Kettler, Carla
Gromski, Mark
Akisik, Fatih
Chalasani, Naga
author_facet Vuppalanchi, Raj
Are, Vijay
Telford, Alison
Young, Liam
Mouchti, Sofia
Ferreira, Carlos
Kettler, Carla
Gromski, Mark
Akisik, Fatih
Chalasani, Naga
author_sort Vuppalanchi, Raj
collection PubMed
description BACKGROUND & AIMS: Magnetic resonance cholangiopancreatography (MRCP) for evaluation of biliary disease currently relies on subjective assessment with limited prognostic value because of the lack of quantitative metrics. Artificial intelligence-enabled quantitative MRCP (MRCP+) is a novel technique that segments biliary anatomy and provides quantitative biliary tree metrics. This study investigated the utility of MRCP+ as a prognostic tool for the prediction of clinical outcomes in primary sclerosing cholangitis (PSC). METHODS: MRCP images of patients with PSC were post-processed using MRCP+ software. The duration between the MRCP and clinical event (liver transplantation or death) was calculated. Survival analysis and stepwise Cox regression were performed to investigate the optimal combination of MRCP+ metrics for the prediction of clinical outcomes. The resulting risk score was validated in a separate validation cohort and compared with an existing prognostic score (Mayo risk score). RESULTS: In this retrospective study, 102 patients were included in a training cohort and a separate 50 patients formed a validation cohort. Between the two cohorts, 34 patients developed clinical outcomes over a median duration of 3 years (23 liver transplantations and 11 deaths). The proportion of bile ducts with diameter 3–5 mm, total bilirubin, and aspartate aminotransferase were independently associated with transplant-free survival. Combined as a risk score, the overall discriminative performance of the MRCP+ risk score (M+BA) was excellent; area under the receiver operator curve 0.86 (95% CI: 0.77, 0.95) at predicting clinical outcomes in the validation cohort with a hazard ratio 5.8 (95% CI: 1.5, 22.1). This was superior to the Mayo risk score. CONCLUSIONS: A composite score combining MRCP+ with total bilirubin and aspartate aminotransferase (M+BA) identified PSC patients at high risk of liver transplantation or death. Prospective studies are warranted to evaluate the clinical utility of this novel prognostic tool. IMPACT AND IMPLICATIONS: Primary sclerosis cholangitis (PSC) is a disease of the biliary tree where inflammation and fibrosis cause areas of narrowing (strictures) and expansion (dilatations) within the biliary ducts leading to liver failure and/or cancer (cholangiocarcinoma). In this study, we demonstrate that quantitative assessment of the biliary tree can better identify patients with PSC who are at high risk of either death or liver transplantation than a current blood-based risk score (Mayo risk score).
format Online
Article
Text
id pubmed-10472223
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher Elsevier
record_format MEDLINE/PubMed
spelling pubmed-104722232023-09-02 A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis Vuppalanchi, Raj Are, Vijay Telford, Alison Young, Liam Mouchti, Sofia Ferreira, Carlos Kettler, Carla Gromski, Mark Akisik, Fatih Chalasani, Naga JHEP Rep Research Article BACKGROUND & AIMS: Magnetic resonance cholangiopancreatography (MRCP) for evaluation of biliary disease currently relies on subjective assessment with limited prognostic value because of the lack of quantitative metrics. Artificial intelligence-enabled quantitative MRCP (MRCP+) is a novel technique that segments biliary anatomy and provides quantitative biliary tree metrics. This study investigated the utility of MRCP+ as a prognostic tool for the prediction of clinical outcomes in primary sclerosing cholangitis (PSC). METHODS: MRCP images of patients with PSC were post-processed using MRCP+ software. The duration between the MRCP and clinical event (liver transplantation or death) was calculated. Survival analysis and stepwise Cox regression were performed to investigate the optimal combination of MRCP+ metrics for the prediction of clinical outcomes. The resulting risk score was validated in a separate validation cohort and compared with an existing prognostic score (Mayo risk score). RESULTS: In this retrospective study, 102 patients were included in a training cohort and a separate 50 patients formed a validation cohort. Between the two cohorts, 34 patients developed clinical outcomes over a median duration of 3 years (23 liver transplantations and 11 deaths). The proportion of bile ducts with diameter 3–5 mm, total bilirubin, and aspartate aminotransferase were independently associated with transplant-free survival. Combined as a risk score, the overall discriminative performance of the MRCP+ risk score (M+BA) was excellent; area under the receiver operator curve 0.86 (95% CI: 0.77, 0.95) at predicting clinical outcomes in the validation cohort with a hazard ratio 5.8 (95% CI: 1.5, 22.1). This was superior to the Mayo risk score. CONCLUSIONS: A composite score combining MRCP+ with total bilirubin and aspartate aminotransferase (M+BA) identified PSC patients at high risk of liver transplantation or death. Prospective studies are warranted to evaluate the clinical utility of this novel prognostic tool. IMPACT AND IMPLICATIONS: Primary sclerosis cholangitis (PSC) is a disease of the biliary tree where inflammation and fibrosis cause areas of narrowing (strictures) and expansion (dilatations) within the biliary ducts leading to liver failure and/or cancer (cholangiocarcinoma). In this study, we demonstrate that quantitative assessment of the biliary tree can better identify patients with PSC who are at high risk of either death or liver transplantation than a current blood-based risk score (Mayo risk score). Elsevier 2023-06-29 /pmc/articles/PMC10472223/ /pubmed/37663118 http://dx.doi.org/10.1016/j.jhepr.2023.100834 Text en © 2023 The Authors https://creativecommons.org/licenses/by/4.0/This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
spellingShingle Research Article
Vuppalanchi, Raj
Are, Vijay
Telford, Alison
Young, Liam
Mouchti, Sofia
Ferreira, Carlos
Kettler, Carla
Gromski, Mark
Akisik, Fatih
Chalasani, Naga
A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title_full A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title_fullStr A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title_full_unstemmed A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title_short A composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
title_sort composite score using quantitative magnetic resonance cholangiopancreatography predicts clinical outcomes in primary sclerosing cholangitis
topic Research Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10472223/
https://www.ncbi.nlm.nih.gov/pubmed/37663118
http://dx.doi.org/10.1016/j.jhepr.2023.100834
work_keys_str_mv AT vuppalanchiraj acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT arevijay acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT telfordalison acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT youngliam acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT mouchtisofia acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT ferreiracarlos acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT kettlercarla acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT gromskimark acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT akisikfatih acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT chalasaninaga acompositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT vuppalanchiraj compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT arevijay compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT telfordalison compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT youngliam compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT mouchtisofia compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT ferreiracarlos compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT kettlercarla compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT gromskimark compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT akisikfatih compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis
AT chalasaninaga compositescoreusingquantitativemagneticresonancecholangiopancreatographypredictsclinicaloutcomesinprimarysclerosingcholangitis