Unsupervised learning-based quantitative analysis of CT intratumoral subregions predicts risk stratification of bladder cancer patients
Rattachement africain : cn. Niveau de preuve : code pays fourni par la source.
Le résumé fourni par la source
BACKGROUND: Preoperative diagnosis of muscle invasion and American Joint Committee on Cancer (AJCC) stage plays a crucial role in guiding treatment strategies for bladder cancer (BCa). Utilizing quantitative analysis of tumor subregions via CT imaging holds promise in identifying high-risk populations. Developing and evaluating the performance of an unsupervised clustering algorithm-based intratumoral subregion radiomics model for distinguishing between bladder muscle invasion and AJCC stage. METHODS: This retrospective study included BCa patients who underwent CT imaging prior to transurethral resection of bladder tumor (TURBT) between January 2006 and December 2022, with all cases being histologically confirmed. Utilizing an unsupervised clustering learning method, the tumor region was classified into three intratumoral subregions, and radiomics features were extracted from both the whole tumor region and each intratumoral subregion. Following feature selection and nested cross-validation, seven predictive models were developed: clinicopathological model, whole-tumor model, intratumoral subregion 1 model, intratumoral subregion 2 model, intratumoral subregion 3 model, merged intratumoral subregion model, and fusion model. Predictive performance was evaluated using the area under the receiver operating characteristic curve (AUROC). RESULTS: A total of 1017 patients were included in this study, with 778 from center A (training cohort; median age, 67 years [IQR, 69-75 years]) and 239 from center B (external validation cohort; median age, 58.5 years [IQR, 66-74 years]). In the external validation cohort, the fusion model AUROC was highest for predicting muscle invasion and AJCC stage tasks, at 0.884 (95% CI, 0.842, 0.926) and 0.832 (95% CI, 0.768, 0.897), respectively. The merged intratumoral subregion model outperformed the whole-tumor model (AUROC, muscle invasion, 0.871 vs 0.804, p = 0.004; AJCC stage, 0.832 vs 0.804, p = 0.4). CONCLUSIONS: The subregional radiomics model based on preoperative CT shows potential value in predicting muscle invasion and AJCC stage. This noninvasive risk stratification tool may offer supportive insights into surgical approaches and treatment decisions for BCa patients.
Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.
Le contrôle bibliographique ouvert
DOI retrouvé dans Crossref DOI retrouvé ; titre concordant.
- Titre Crossref
- Unsupervised learning-based quantitative analysis of CT intratumoral subregions predicts risk stratification of bladder cancer patients
- Date Crossref
- 02/06/2025
- Éditeur
- Springer Science and Business Media LLC
- Type
- journal-article
Ce recoupement confirme des métadonnées liées au DOI. Il ne confirme ni la méthode ni les conclusions de l’étude, et il ne compte pas comme une seconde source scientifique indépendante.
Où se fait cette recherche
-
Shandong Provincial Hospital Department of Radiology pays non établi dans la noticeÉtablissement de santé
-
Shandong First Medical University pays non établi dans la noticeUniversité ou école supérieure
-
Qingdao University Department of Radiology pays non établi dans la noticeUniversité ou école supérieure
-
Affiliated Hospital of Qingdao University pays non établi dans la noticeÉtablissement de santé
-
Zhangqiu City People's Hospital pays non établi dans la noticeÉtablissement de santé
-
Capital Medical University Department of Radiology pays non établi dans la noticeUniversité ou école supérieure
-
Beijing Friendship Hospital pays non établi dans la noticeÉtablissement de santé
-
Peking University Department of General Surgery pays non établi dans la noticeUniversité ou école supérieure
-
Peking University Third Hospital pays non établi dans la noticeÉtablissement de santé
-
School of Radiology pays non établi dans la noticeUniversité ou école supérieure
-
The People's Hospital of Zhangqiu Area Department of Radiology pays non établi dans la noticeÉtablissement de santé
Department of Radiology — Shandong Provincial Hospital, Shandong First Medical University et Department of Radiology — Qingdao University, avec 8 autres affiliations.
Une affiliation ne permet pas de déduire la nationalité d’un auteur.