Urologic Oncology: Seminars and Original Investigations
Seminar articleBladder (cancer treatment)Are nomograms needed in the management of bladder cancer?
Introduction
In the field of urologic oncology, the optimal treatment strategy for an individual patient is one that maximizes oncologic efficacy and survival while minimizing morbidity and any adverse impact on quality of life. However, management decisions are rarely ever straightforward and can be complicated by complex tumor biology, variable treatment approaches, as well as a paucity of definitive data (i.e., from randomized controlled trials) favoring one therapy over another. As a result, consensus on what constitutes optimal treatment can be difficult to reach, often generating controversy and confusion amongst clinicians and patients alike. This is certainly the case with bladder cancer, a heterogeneous disease with a course that varies greatly from patient to patient and requires multiple therapeutic approaches.
Effective patient counseling regarding the merits of alternative treatment strategies depend upon accurate estimation of relevant treatment outcomes (e.g., risk of treatment failure, rate of survival, or the incidence of complications). In this regard, decision aids, such as risk groupings, probability tables, and nomograms, have greatly enhanced our ability to predict endpoints and provide patients the data they require to make informed medical decisions. Nomograms are among the more accurate means of estimating risks and have been shown to surpass other classes of decision aids, including risk groupings, neural networks, or probability tables, in predictive accuracy [1], [2], [3], [4], [5], [6], [7]. In this review, we examine the current status and utility of nomograms in bladder cancer, comparing their predictive capacity against traditional prediction methods and also determining the need for novel nomograms.
Section snippets
Controversies in bladder cancer management
Treatment strategies in bladder cancer can be quite variable, and selection is determined by a combination of tumor-related factors (e.g., stage, grade, size, multifocality, associated CIS, vascular invasion, and histologic subtype), patient-specific characteristics (age, comorbidity, and utilities), as well as molecular markers (p53, pRB, p16, p21) and laboratory tests (cytology or FISH). All of these clinical parameters can impact disease course and determine the risks of tumor recurrence,
The evolution of generating predictions for clinical decision-making
With the lack of conclusive data favoring any treatment over another, physicians and patients alike rely on estimates of outcomes when considering bladder cancer risks and treatment choices. Historically, when a patient has wanted a prediction, physicians relied on their own individual knowledge and experience to provide patients with a prediction of an outcome, such as life expectancy or the chance of treatment success. Decision-making and patient counseling have since evolved with the
The utility of nomograms in bladder cancer
Nomograms demonstrate several advantages over other prediction tools: they incorporate multiple prognostic markers, generate individualized risk estimates, and are presented in an intuitive graphical format that avoids complex calculations. Not only can nomograms assist in basic treatment selection, but by offering highly accurate means of identifying bladder cancer patients at high risk for adverse outcomes (e.g., treatment failure or metastatic progression), they can also facilitate the
Nomogram vs. physician judgment
The current interest in developing nomograms for use in bladder cancer and other diseases is predicated on the assumption that they are superior to traditional methods of risk estimation, which include physician judgment and risk groupings. Before the advent of formal decision aids, management decisions were traditionally based upon physician judgment. In recommending a specific form of therapy to a patient, a clinician presumably believes that relevant outcomes for that treatment modality are
Nomogram versus risk classification
Because individual physician judgment is clearly not an ideal foundation for outcome prediction, many clinicians have turned to formal decision aids for patient counseling. Risk classifications, such as the AJCC TNM staging system or the WHO pathologic grading system, have enjoyed widespread popularity due to their simplicity and become integral to the management of bladder cancer. Traditional risk groupings determine prognosis by assigning patients into categories based on the presence or
Current status of bladder cancer nomograms
Traditional methods of risk estimation may not provide the best predictive accuracy or applicability to individual patients. Undoubtedly, individual clinicians' expert knowledge and advice are still crucial to patient counseling and the measurement of variables used in the prediction process, and risk groupings can certainly be valuable for making general observations about the bladder cancer population as a whole. However, the distinct advantages of nomograms have prompted greater interest in
Future directions for bladder cancer nomograms
Although nomograms currently represent the most accurate means of risk stratification, they do possess limitations which provide the opportunity for research and improvement. To date, no nomogram, or any other prediction method, has been able to predict clinical outcomes with 100% accuracy. The suboptimal predictive accuracy of current nomograms is partly due to the inherent unpredictability of bladder cancer, a heterogeneous disease with a highly variable course. However, there are modifiable
Conclusions
Bladder cancer is a heterogeneous disease with a course and prognosis that can vary greatly from patient to patient. Prediction of tumor behavior in the individual patient is therefore critical to patient counseling and selection of the treatment strategy that optimizes survival and quality of life. Assessment of the risk of disease recurrence, progression, or mortality in bladder cancer patients by traditional prediction methods, such as physician judgment or risk classification systems, can
References (24)
- et al.
A multivariate analysis of clinical and pathological factors that predict for prostate specific antigen failure after radical prostatectomy for prostate cancer
J Urol
(1995) - et al.
External validity of predictive models: a comparison of logistic regression, classification trees, and neural networks
J Clin Epidemiol
(2003) - et al.
Incorporating predictions of individual patient risk in clinical trials
Urol Oncol
(2004) - et al.
Recall and attitudes in patients with prostate cancer
Urology
(1999) - et al.
Ability of 2 pretreatment risk assessment methods to predict prostate cancer recurrence after radical prostatectomy: data from CaPSURE
J Urol
(2005) - et al.
Nomograms for bladder cancer
Eur Urol
(2008) - et al.
Nomograms including nuclear matrix protein 22 for prediction of disease recurrence and progression in patients with Ta, T1, or CIS transitional cell carcinoma of the bladder
J Urol
(2005) - et al.
A critical appraisal of logistic regression-based nomograms, artificial neural networks, classification and regression-tree models, look-up tables and risk-group stratification models for prostate cancer
BJU Int
(2007) Nomograms are superior to staging and risk grouping systems for identifying high-risk patients: preoperative application in prostate cancer
Curr Opin Urol
(2003)- et al.
Pretreatment nomogram for predicting the outcome of three-dimensional conformal radiotherapy in prostate cancer
J Clin Oncol
(2000)
Comparison of artificial neural networks with other statistical approaches: Results from medical data sets
Cancer
Nomograms provide improved accuracy for predicting survival after radical cystectomy
Clin Cancer Res
Cited by (16)
Risk prediction models of locoregional failure after radical cystectomy for urothelial carcinoma: External validation in a cohort of Korean patients
2014, International Journal of Radiation Oncology Biology PhysicsCitation Excerpt :A risk categories schema is constructed with the assumption that each risk group represents a homogenous population. The individual heterogeneity within a risk group leads to lower predictive accuracy when risk classifications are applied to a particular patient (24). Thus, heterogeneity in patient populations such as ethnicity or genetic backgrounds might lead to higher accuracy in models, although data on ethnic variability in bladder cancer are sparse and ambiguous.
Predicting global QoL after orthotopic neobladder or ileal conduit diversion: nomogram development
2023, Frontiers in OncologyMortality prediction models after radical cystectomy for bladder tumour: A systematic review and critical appraisal
2022, European Journal of Clinical InvestigationNomograms to predict individual prognosis of patients with primary signet ring cell carcinoma of the urinary bladder
2021, Translational Cancer Research