Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study
Abstract The Arteriovenous Access Stage (AVAS) classification simplifies information about suitability of vessels for vascular access (VA). It’s been previously validated in a clinical study. Here, AVAS performance was tested against multiple ultrasound mapping measurements using machine learning. A...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-86456-3 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585727535218688 |
|---|---|
| author | Katerina Lawrie Petr Waldauf Peter Balaz Radoslav Bortel Ricardo Lacerda Emma Aitken Krzysztof Letachowicz Mario D’Oria Vittorio Di Maso Pavel Stasko Antonio Gomes Joana Fontainhas Matej Pekar Alena Srdelic VAVASC Study Group Stephen O’Neill |
| author_facet | Katerina Lawrie Petr Waldauf Peter Balaz Radoslav Bortel Ricardo Lacerda Emma Aitken Krzysztof Letachowicz Mario D’Oria Vittorio Di Maso Pavel Stasko Antonio Gomes Joana Fontainhas Matej Pekar Alena Srdelic VAVASC Study Group Stephen O’Neill |
| author_sort | Katerina Lawrie |
| collection | DOAJ |
| description | Abstract The Arteriovenous Access Stage (AVAS) classification simplifies information about suitability of vessels for vascular access (VA). It’s been previously validated in a clinical study. Here, AVAS performance was tested against multiple ultrasound mapping measurements using machine learning. A prospective multicentre international study (NCT04796558) with patient recruitment from March 2021-July 2024. Demographics, risk factors, vessels parameters, types of predicted and created VA (pVA, cVA) were collected. We modelled pVA and cVA using the Random Forest algorithm. Model performance was estimated and compared using Bayesian generalized linear models. ROC AUC with 95% credible intervals was the performance metric. 1151 patients were included. ROC AUC for pVA prediction by AVAS was 0.79 (0.77;0.82) and by mapping was 0.85 (0.83;0.88). ROC AUC for cVA prediction by AVAS was 0.71 (0.69;0.74) and by mapping was 0.8 (0.78;0.83). Using AVAS with other parameters increased the ROC AUC to 0.87 for pVA (0.84;0.89) and 0.82 (0.79;0.84) for cVA. Using mapping with other parameters increased the ROC AUC to 0.88 for pVA (0.86;0.91) and 0.85 (0.83;0.88) for cVA. Multiple mapping measurements showed higher performance at VA prediction than AVAS. However, AVAS is simpler and quicker, so may be preferable for routine clinical practice. |
| format | Article |
| id | doaj-art-9d4ea12be8274877a8ea6d9c129191b2 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-9d4ea12be8274877a8ea6d9c129191b22025-01-26T12:33:17ZengNature PortfolioScientific Reports2045-23222025-01-0115111410.1038/s41598-025-86456-3Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre studyKaterina Lawrie0Petr Waldauf1Peter Balaz2Radoslav Bortel3Ricardo Lacerda4Emma Aitken5Krzysztof Letachowicz6Mario D’Oria7Vittorio Di Maso8Pavel Stasko9Antonio Gomes10Joana Fontainhas11Matej Pekar12Alena Srdelic13VAVASC Study Group14Stephen O’Neill15Department of Transplantation Surgery, Institute for Clinical and Experimental MedicineThird Faculty of Medicine, Charles UniversityThird Faculty of Medicine, Charles UniversityFaculty of Electrical Engineering, Czech Technical University in PragueRL Vascular Surgery and Interventional Radiology, Private PracticeDepartment of Renal Surgery, Queen Elizabeth University HospitalDepartment of Nephrology and Transplantation Medicine, Wroclaw Medical UniversityDivision of Vascular and Endovascular Surgery, Cardio-Thoracic-Vascular Department, University Hospital of TriesteNephrology and Dialysis Unit, Department of Medicine, ASUGI - University Hospital of TriesteAdNa s.r.o., Vascular Surgery ClinicDepartment of General Surgery, Hospital Professor Doutor Fernando FonsecaDepartment of General Surgery, Hospital Professor Doutor Fernando FonsecaCentre for Vascular and Mini-invasive Surgery, Hospital AGELDivision of Nephrology and Haemodialysis, Internal Medicine Department, University Hospital of SplitNephrology and Dialysis Unit, Department of Medicine, ASUGI - University Hospital of TriesteCentre for Medical Education, Queen’s University BelfastAbstract The Arteriovenous Access Stage (AVAS) classification simplifies information about suitability of vessels for vascular access (VA). It’s been previously validated in a clinical study. Here, AVAS performance was tested against multiple ultrasound mapping measurements using machine learning. A prospective multicentre international study (NCT04796558) with patient recruitment from March 2021-July 2024. Demographics, risk factors, vessels parameters, types of predicted and created VA (pVA, cVA) were collected. We modelled pVA and cVA using the Random Forest algorithm. Model performance was estimated and compared using Bayesian generalized linear models. ROC AUC with 95% credible intervals was the performance metric. 1151 patients were included. ROC AUC for pVA prediction by AVAS was 0.79 (0.77;0.82) and by mapping was 0.85 (0.83;0.88). ROC AUC for cVA prediction by AVAS was 0.71 (0.69;0.74) and by mapping was 0.8 (0.78;0.83). Using AVAS with other parameters increased the ROC AUC to 0.87 for pVA (0.84;0.89) and 0.82 (0.79;0.84) for cVA. Using mapping with other parameters increased the ROC AUC to 0.88 for pVA (0.86;0.91) and 0.85 (0.83;0.88) for cVA. Multiple mapping measurements showed higher performance at VA prediction than AVAS. However, AVAS is simpler and quicker, so may be preferable for routine clinical practice.https://doi.org/10.1038/s41598-025-86456-3Arteriovenous accessClassification systemDialysisMappingRandom forestRenal replacement therapy |
| spellingShingle | Katerina Lawrie Petr Waldauf Peter Balaz Radoslav Bortel Ricardo Lacerda Emma Aitken Krzysztof Letachowicz Mario D’Oria Vittorio Di Maso Pavel Stasko Antonio Gomes Joana Fontainhas Matej Pekar Alena Srdelic VAVASC Study Group Stephen O’Neill Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study Scientific Reports Arteriovenous access Classification system Dialysis Mapping Random forest Renal replacement therapy |
| title | Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study |
| title_full | Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study |
| title_fullStr | Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study |
| title_full_unstemmed | Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study |
| title_short | Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study |
| title_sort | machine learning validation of the avas classification compared to ultrasound mapping in a multicentre study |
| topic | Arteriovenous access Classification system Dialysis Mapping Random forest Renal replacement therapy |
| url | https://doi.org/10.1038/s41598-025-86456-3 |
| work_keys_str_mv | AT katerinalawrie machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT petrwaldauf machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT peterbalaz machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT radoslavbortel machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT ricardolacerda machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT emmaaitken machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT krzysztofletachowicz machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT mariodoria machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT vittoriodimaso machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT pavelstasko machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT antoniogomes machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT joanafontainhas machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT matejpekar machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT alenasrdelic machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT vavascstudygroup machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy AT stephenoneill machinelearningvalidationoftheavasclassificationcomparedtoultrasoundmappinginamulticentrestudy |