The prediction of any future event is fascinating and yet fraught with difficulty. Clinicians have always struggled with being able to accurately predict outcomes for patients and to balance the risks of any intervention. Traditionally, doctors relied on clinical judgment to predict outcomes but this has been shown to be error prone and highly variable between physicians(1). One technique for improving upon the accuracy of clinical judgment in cancer management is the use of prognostic factors such as tumour volume, PSA, clinical stage and pathological stage. These factors can then be used to stratify patients into risk groups, such as D’Amico et al.(2), enabling the surgeon to better predict the natural history of disease and advise on potential interventions most appropriate to each risk group. This method of stratifying patients into risk groups often loses significant data as it converts continuous variables into categorical variables. Additionally, scientists and clinicians are finding increasing numbers of prognostic risk factors in cancers. As the number of prognostic factors increases the prediction of disease outcomes will become ever more complex and risk groups are likely to be inaccurate for individual patients.
Nomograms
An alternative to risk groups is the use of nomograms. A nomogram is a graphic representation of a mathematical formula or algorithm that incorporates several predictors modelled as continuous variables to predict a particular endpoint(3). The use of prediction tools, such as nomograms, enables physicians and patients to gain a more accurate prediction tailored towards their individual risk profile. The main concern over the use of nomograms relates to data coming from an external population group which may not be applicable to a physician’s local population. However, a recent large multicentre external validation of bladder cancer nomograms from MSKCC showed that it outperformed the American Joint Committee on Cancer/TNM staging system(4).
Memorial Sloan-Kettering Cancer Centre
A good example of the use of online predictive tools based on nomograms can be found at Memorial Sloan-Kettering Cancer Centre (MSKCC) website. The researchers at MSKCC have constructed a free to access website with a collection of cancer prediction tools for a wide variety of specialties. Significantly for urologists, they have developed predictive tools based on nomograms for bladder, prostate and renal cancers (links below).
The website has an excellent layout and after the initial prognostic data is entered the results can be printed, form part of the patient record, and inform and empower patient decision making at outpatient visits. The researchers and web developers have done an outstanding job at making the nomograms easy to use and all of the predictive tools are supported by published articles in leading peer reviewed journals.
Our challenge is to integrate and externally validate predictive tools such as these into our clinical practice. Hopefully, this will also enable us to create better datasets and predictive tools for our local populations and ultimately improve patient care.
MSKCC prediction tools
http://www.mskcc.org/cancer-care/prediction-tools
Bladder Cancer
http://www.mskcc.org/cancer-care/adult/bladder/prediction-tools
The bladder cancer nomogram that has been developed can be used to calculate the probability that patient will not have a recurrence of bladder cancer at 5 years following cystectomy for transitional cell carcinoma, squamous cell carcinoma or pure adenocarcinoma of the bladder.
Prostate Cancer
http://www.mskcc.org/cancer-care/adult/prostate/prediction-tools
For prostate cancer management there are a number of nomograms available including those for:
- patients who are pre-treatment and considering brachytherapy or radical prostatectomy,
- patients post radical prostatectomy (recurrence after surgery)
- patients post radical prostatectomy considering radiation therapy post surgery
- patients who are hormone refractory
As well as prostate cancer nomograms, MSKCC have made available online calculators for assessing PSA doubling time and tumour volume.
Renal Cell Carcinoma
http://www.mskcc.org/cancer-care/adult/kidney/prediction-tools
The renal cell cancer nomogram is based on histology (chromophobe, conventional and papillary), symptoms and tumour stage and size. These factors can be used to predict the chance of being cancer free at 5 years post surgery.
Authors
Mr Ivo Dukic, ST5 in Urology, United Kingdom
Dr Tiffany S. Berrington, Foundation Year 1 Doctor, United Kingdom
References
1. D’Amico AV, Whittington R, Malkowicz SB, Schultz D, Blank K, Broderick GA, et al. Biochemical Outcome After Radical Prostatectomy, External Beam Radiation Therapy, or Interstitial Radiation Therapy for Clinically Localized Prostate Cancer. JAMA. 1998 Sep 16;280(11):969–74.
2. Shariat SF, Karakiewicz PI, Godoy G, Lerner SP. Use of nomograms for predictions of outcome in patients with advanced bladder cancer. Ther Adv Urol. 2009 Apr;1(1):13–26.
3. Stephenson AJ, Kattan MW. Nomograms for prostate cancer. BJU International. 2006 Jun 8;98(1):39–46.
4. Nuhn P, May M, Sun M, Fritsche H-M, Brookman-May S, Buchner A, et al. External validation of postoperative nomograms for prediction of all-cause mortality, cancer-specific mortality, and recurrence in patients with urothelial carcinoma of the bladder. Eur. Urol. 2012 Jan;61(1):58–64.