Data Description and Background A university medical center urology group was in
ID: 74154 • Letter: D
Question
Data Description and Background
A university medical center urology group was interested in the association between prostate specific antigen (PSA) and a number of prognostic clinical measurements in men with advanced prostate cancer. Data were collected on 97 men who were about to undergo radical prostectomies. Each line of data set provides information on 8 other variables for each person.
PSA is commonly used as a screening mechanism for detecting prostate cancer. However, to be an efficient screening tool it is important that we understand how PSA levels relate to factors
that may determine prognosis and outcome.
The PSA test measures the blood level of prostate-specific antigen, an enzyme produced by
the prostate. PSA levels under 4 ng/mL (nanograms per milliliter) are generally considered normal, while levels over 4 ng/mL are considered abnormal (although in men over 65 levels up to 6.5 ng/mL may be acceptable, depending upon each laboratory’s reference ranges). PSA levels between 4 and 10 ng/mL indicate a risk of prostate cancer higher than normal, but the risk does not seem to rise within this six-point range. When the PSA level is above 10 ng/mL, the association with cancer becomes stronger. However, PSA is not a perfect test. Some men with prostate cancer do not have an elevated PSA, and most men with an elevated PSA do not have prostate cancer. PSA levels can change for many reasons other than cancer. Two common causes of high PSA levels are enlargement of the prostate (benign prostatic hypertrophy (BPH)) and infection in the prostate (prostatitis).
Some of the variable names may look unfamiliar to you - please use resources on the web if you feel unsure as to what these variables measure. The section above is based on excerpts from Wikipedia.org, and you can also find variable definitions at http://www.prostate-cancer.org/resource/glossary.html.
For example, a large tumor may invade surrounding tissue and penetrate the wall of the prostate (variable svi and cp). Also, benign hyperplasia is associated with higher PSA levels, but is non-cancerous (variable bph).
The goal of the analysis is to develop a model for PSA to be used for inferential purposes. Your model should be parsimonious, that is a model that balances both explanatory power with simplicity. To this end, you may employ any of the methods learned in class. Write up a report (5-7 pages) describing how you obtained this model. Below is a list of things to address in the report.
- Are all the assumptions needed to fit the model satisfied? Do any transformations need to be applied to the response and/or explanatory variables in order to correct for any model deviations?
- Are there any outliers in the dataset? Are they adversely affecting the estimates obtained using the least squares method?
- Recall that the goal of model building is not to build the model that best fits your particular dataset, but rather a model that can generalize. Consequently, what is the method you will employ to select a model? How many variables will you use? How did you model them (aka via polynomial terms, interactions or transformations)?
Explanation / Answer
Prostate cancer starts from the prostate gland, which is the gland that is situated underneath the bladder. Prostate cancer usually grows very slowly in some men that even if it is not treated will not threaten their lives. In some men it grows very faster and distributes to other parts of the body. In this situation, it would be better to detect these cancers in their early stages and given with effective treatment. So, it has become the goal for using PSA testing to screen people with prostate cancer without any symptoms.
Prostate specific antigen is present in the blood released by prostate gland. The PSA levels are observed to be in high levels than normal levels in men with prostate cancer, with benign prostatic hyperplasia and inflammation of prostate.
Though PSA testing has helped tremendously in identifying the prostate cancer in the earlier stage and reduce the death rates in this regard, it is found to have increased unnecessary biopsies of prostate tissues and its treatment. So, there is a controversy regarding the usage of PSA testing for detecting prostate cancer in men with no symptoms.
An updates scientific report has stated that multiple variables were used for creating risk models for prostate cancer. These variables included PSA velocity, age, prostate volume, family history, DRE results, previous negative biopsies, free PSA levels and so on. Meta-analysis was done using these variables in creating several models that can detect the clinically significant prostate cancer. But, these models were decided to be sent for clinical efficiency before they are recommended to be used for concluding the detection and continue with biopsy.
European randomized study of screening for prostate cancer (ERSPC) has made use of lower PSA levels (2.5 to 3 ng/ml) along with the DRE and transrectal ultrasonography test results to accompany the biopsy results. The cutoff value of 4 ng/ml can be used for further recommending the patient for biopsy.
Biopsy is also referred only by examining the other factors like PSA density, PSA velocity, age and race associated PSA levels, free PSA levels and complexed PSA levels. However, data analysis by ERSPC suggests that PSA velocity has not found to be improving the cancer predicting capacity of PSA levels.
The predictive value of PSA levels greater than 4ng/ml was observed to be approximately 30 percent. The predictive value would be 25 percent when the PSA levels are between 4 and 10 ng/ml. If the PSA levels are higher than 10ng/ml then the predictive value will rise from 42 to 64 percent. But, it is well reported that 75 percent of detected cancers with PSA between 4 and 10 ng/ml were confined to the organ and was curable. If the PSA is greater than 10 ng/ml, then 50 percent of the cases are not organ confined.
It is the opinion of many experts that tumor volume greater than 0.5cm3 and with Gleason score greater than or equal to 7 showed higher risk for disease progression. It is also found that PSA levels can be less helpful to identify men having symptomatic benign prostatic hyperplasia.
PSA is found to be increasing rapidly in cases where the cancer is seen than in healthy men. Men with PSA velocity higher than 0.75ng/ml/year were at higher risk of being diagnosed with prostate cancer. PSA velocity could only slightly improve the predictive ability of PSA in detecting the cancer rather than detecting the high grade cancer.
The discriminating ability of PSA will be enhanced by the age specific ranges of PSA levels as PSA increases with age. But age specific detections led to increase in unnecessary biopsies and missing of clinically localized cancers.
DRE can only detect cancer in the posterior and lateral parts of the prostate gland. Transrectal ultrasonography is not the primary screening test for prostate cancer.
The model that is best recommended to accompany PSA testing in order to detect prostate cancer could have these variables to show the better results.
PSA velocity,
Age specific PSA
Free PSA
Repeated biopsies
Complexed PSA
DRE
PSA levels
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.