Model database

Every published prediction model identified by the systematic review. Click a row for full extraction details, PROBAST domains, and a plain-language summary.

Outcome Predicted

Modelling Method

Year

Country

External validation

Calibration reported

Public model

PROBAST Risk of Bias

Showing 10 of 54 models

					Modelling Method				Publicly available
Cornelissen et al.	2026	The Netherlands	Surveillance	Tumour Growth	SVM	110	AUC 0.890	High risk of bias	No
Schouten et al.	2025	The Netherlands	Surveillance	Tumour Growth	Logistic Regression	110	AUC 0.850	Low risk of bias	Yes
Stastna et al.	2025	United Kingdom	Surveillance	Tumour Growth	Cox Regression	615	AUC 0.858	Low risk of bias	No
Chen et al.	2024	The Netherlands	Surveillance	Tumour Growth	ANN / Deep Learning	131	—	High risk of bias	Yes
Wang et al.	2023	USA	Surveillance	Tumour Growth	ANN / Deep Learning	103	AUC 0.770	High risk of bias	Yes
Itoyama et al.	2022	Japan	Surveillance	Tumour Growth	Random Forest	67	AUC 0.690	High risk of bias	No
Gadot et al.	2022	USA	Surveillance	Need for Intervention (SRS or Surgery)	Random Forest	124	AUC 0.830	High risk of bias	No
Hentschel et al.	2021	The Netherlands	Surveillance	Tumour Growth	Logistic Regression	1217	C-index 0.690	Low risk of bias	Yes
Varughese et al.	2012	Norway	Surveillance	Tumour Growth	Mixed-Effects	178	—	High risk of bias	No
Timmer et al.	2011	The Netherlands	Surveillance	Tumour Growth	Logistic Regression	240	AUC 0.723	Low risk of bias	Yes