Risk prediction models for symptomatic patients with bladder and kidney cancer: a systematic review

Abstract

Background: Timely diagnosis of bladder and kidney cancer is key to improving clinical outcomes. Given the challenges of early diagnosis, models incorporating clinical symptoms and signs may be helpful to primary care clinicians when triaging at-risk patients.

Aim: To identify and compare published models that use clinical signs and symptoms to predict the risk of undiagnosed prevalent bladder or kidney cancer.

Design and setting: Systematic review.

Method: A search identified primary research reporting or validating models predicting the risk of bladder or kidney cancer in MEDLINE and EMBASE. After screening identified studies for inclusion, data were extracted onto a standardised form. The risk models were classified using TRIPOD guidelines and evaluated using the PROBAST assessment tool.

Results: The search identified 20 661 articles. Twenty studies (29 models) were identified through screening. All the models included haematuria (visible, non-visible, or unspecified), and seven included additional signs and symptoms (such as abdominal pain). The models combined clinical features with other factors (including demographic factors and urinary biomarkers) to predict the risk of undiagnosed prevalent cancer. Several models (n = 13) with good discrimination (area under the receiver operating curve >0.8) were identified; however, only eight had been externally validated. All of the studies had either high or unclear risk of bias.

Conclusion: Models were identified that could be used in primary care to guide referrals, with potential to identify lower-risk patients with visible haematuria and to stratify individuals who present with non-visible haematuria. However, before application in general practice, external validations in appropriate populations are required.

Keywords: bladder cancer; early diagnosis; kidney cancer; risk prediction; systematic review.

Figure 1.

PRISMA flow diagram. UTCa = urinary tract cancer.

Figure 2.

RoB assessment using PROBAST framework. ^{a a}For each study, RoB is shown for model development and validation separately. RoB is assessed over four domains (D1: population, D2: risk factors, D3: outcome, D4: analysis), the overall results for each study are shown on the right. 1, 2, 3, and 4 refer to TRIPOD classification for each identified study. a, b, and c refer to models developed by the same author group. RoB = risk of bias.

Figure 3.

Model discrimination, AUROC. ^{a a} Models are split into groups describing the development population and within each group are ordered by the number of risk factors used. Study type (development, internal and external validation), type of haematuria used in model, and study setting are indicated on the plot. A, b, and c refer to models developed by the same author group. Each model is labelled according to its development study; however, the discrimination measured in several external validations ^{, ,} of these models are also included in this summary plot (see supplementary Table S6 for details). AUROC = area under the receiver operating curve. NVH = non-visible haematuria. RF = risk factor. VH = visible haematuria.