Interpretable Machine Learning Model for Locoregional Relapse Prediction in Oropharyngeal Cancers
Paul Giraud
- Fonction : Auteur
Philippe Giraud
- Fonction : Auteur
Eliot Nicolas
- Fonction : Auteur
Pierre Boisselier
- Fonction : Auteur
- PersonId : 760275
- ORCID : 0000-0003-0121-365X
Marc Alfonsi
- Fonction : Auteur
Michel Rives
- Fonction : Auteur
Etienne Bardet
- Fonction : Auteur
Valentin Calugaru
- Fonction : Auteur
- PersonId : 776041
- ORCID : 0000-0002-7156-9750
- IdRef : 163698457
Georges Noel
- Fonction : Auteur
Enrique Chajon
- Fonction : Auteur
Pascal Pommier
- Fonction : Auteur
Lionel Perrier
- Fonction : Auteur
- PersonId : 15600
- IdHAL : lionel-perrier
- ORCID : 0000-0003-4487-8723
- IdRef : 060772239
Xavier Liem
- Fonction : Auteur
- PersonId : 767372
- ORCID : 0000-0002-9448-3757
Anita Burgun
- Fonction : Auteur
- PersonId : 765172
- ORCID : 0000-0001-6855-4366
- IdRef : 094355207
Jean Emmanuel Bibault
- Fonction : Auteur
- PersonId : 814390
- ORCID : 0000-0002-1728-6776
- IdRef : 183965787
Résumé
Background: There is no evidence to support surgery or radiotherapy as the best treatment for resectable oropharyngeal cancers with a negative HPV status. Predictive algorithms may help to decide which strategy to choose, but they will only be accepted by caregivers and European authorities if they are interpretable. As a proof of concept, we developed a predictive and interpretable algorithm to predict locoregional relapse at 18 months for oropharyngeal cancers as a first step towards that goal. Methods: The model was based on clinical and Pyradiomics features extracted from the dosimetric CT scan. Intraclass correlation was used to filter out features dependant on delineation. Correlated redundant features were also removed. An XGBoost model was cross-validated and optimised on the HN1 cohort (79 patients), and performances were assessed on the ART ORL cohort (45 patients). The Shapley Values were used to provide an overall and local explanation of the model. Results: On the ART ORL cohort, the model trained on HN1 yielded a precision—or predictive positive value—of 0.92, a recall of 0.42, an area under the curve of the receiver operating characteristic of 0.68 and an accuracy of 0.64. The most contributory features were shape Voxel Volume, grey level size zone matrix Small Area Emphasis (glszmSAE), gldm Dependence Non Uniformity Normalized (gldmDNUN), Sex and Age. Conclusions: We developed an interpretable and generalizable model that could yield a good precision—positive predictive value—for relapse at 18 months on a different test cohort.
Domaines
Economies et financesFormat du dépôt | Notice |
---|---|
Type de dépôt | Article dans une revue |
Titre |
en
Interpretable Machine Learning Model for Locoregional Relapse Prediction in Oropharyngeal Cancers
|
Résumé |
en
Background: There is no evidence to support surgery or radiotherapy as the best treatment for resectable oropharyngeal cancers with a negative HPV status. Predictive algorithms may help to decide which strategy to choose, but they will only be accepted by caregivers and European authorities if they are interpretable. As a proof of concept, we developed a predictive and interpretable algorithm to predict locoregional relapse at 18 months for oropharyngeal cancers as a first step towards that goal. Methods: The model was based on clinical and Pyradiomics features extracted from the dosimetric CT scan. Intraclass correlation was used to filter out features dependant on delineation. Correlated redundant features were also removed. An XGBoost model was cross-validated and optimised on the HN1 cohort (79 patients), and performances were assessed on the ART ORL cohort (45 patients). The Shapley Values were used to provide an overall and local explanation of the model. Results: On the ART ORL cohort, the model trained on HN1 yielded a precision—or predictive positive value—of 0.92, a recall of 0.42, an area under the curve of the receiver operating characteristic of 0.68 and an accuracy of 0.64. The most contributory features were shape Voxel Volume, grey level size zone matrix Small Area Emphasis (glszmSAE), gldm Dependence Non Uniformity Normalized (gldmDNUN), Sex and Age. Conclusions: We developed an interpretable and generalizable model that could yield a good precision—positive predictive value—for relapse at 18 months on a different test cohort.
|
Auteur(s) |
Paul Giraud
, Philippe Giraud
, Eliot Nicolas
, Pierre Boisselier
, Marc Alfonsi
, Michel Rives
, Etienne Bardet
, Valentin Calugaru
, Georges Noel
, Enrique Chajon
, Pascal Pommier
, Magali Morelle
1
, Lionel Perrier
1
, Xavier Liem
, Anita Burgun
, Jean Emmanuel Bibault
1
GATE Lyon Saint-Étienne -
Groupe d'Analyse et de Théorie Economique Lyon - Saint-Etienne
( 1169842 )
- 93, chemin des Mouilles 69130 Écully 6, rue Basse des Rives 42023 Saint-Étienne cedex 02
- France
|
Langue du document |
Anglais
|
Nom de la revue |
|
Vulgarisation |
Non
|
Comité de lecture |
Oui
|
Audience |
Internationale
|
Date de publication |
2021-01
|
Volume |
13
|
Numéro |
1
|
Page/Identifiant |
57
|
Domaine(s) |
|
DOI | 10.3390/cancers13010057 |
PubMed Central | PMC7795920 |
Loading...