Sensitivity and Specificity in Prostate Cancer Screening Methods and Strategies
The debate regarding the effectiveness of current prostate cancer screening strategies and the necessity of population based testing is becoming a major issue within the medical community. While Prostate Specific Antigen (PSA) testing is the most widespread form of prostate cancer screening, its specificity of 63.1% and low sensitivity of 34.9% calls for a statistically improved and more effective method to more accurately and consistently detect the ninth most common form of cancer. New technology has enabled medical researchers to develop more reliable, less invasive screening methods. When analyzing alternatives to the PSA test, five classifications were used to assess the effectiveness of the method: sensitivity and specificity, availability, ease of use, possible side effects, and financial burden on the patient, with an emphasis on sensitivity and specificity. Several prostate cancer screening strategies, such as the Digital Rectal Exam (DRE), were compared and contrasted. Such methods were shown to be too invasive, without providing conclusive diagnostic evidence. Other methods demonstrated unacceptable sensitivity and specificity values, were not effective in diagnosing pre-Stage or Stage I cancer, required extremely expensive instruments, or relied too heavily on operator experience and doctor expertise. Serum proteomics came forth as the most effective method with an expected sensitivity and specificity of 99.6% at the time of its release. Developments of Serum proteomics will be evaluated as a potential screening process for the early detection of prostate cancer. Although the new screening strategy described is not available to the general public presently, it holds the promise of more efficient prostate cancer screening in the near future.
The effectiveness of prostate cancer screening presently does not measure up to expected medical standards due to the high number of false positive and false negative results. A false test result has the potential to significantly impact a patient, for a substantial amount of time, prior to conclusive evidence for a diagnosis. According to the results of a recent study published in the American Journal of Medicine, a considerable group of patients with suspicious Prostate Specific Antigen (PSA) test results, proven to be non-cancerous by a biopsy, still suffered psychologically for several of the following weeks, despite the fact that they were indeed perfectly healthy. This anxiety can be a heavy burden on how one lives life, not to mention the financial costs of follow-up testing in the immediate future as well as in years to come. False positive results from cancer screening tend to impact patients psychologically and financially causing more questions, anxiety, and expenses than necessary (Marcus 2004). If a screening method that was significantly more accurate than the PSA test existed, prostate cancer screening would be a clearly beneficial and worthwhile practice. So, where exactly do current methods stand at the present time? Do future methods hold the potential to resolve these issues? How can prostate cancer screening be improved?
The purpose of this review is to evaluate current methods for prostate cancer screening with respect to their effectiveness and to research methods that have the potential to improve upon the sensitivity and specificity of future prostate cancer testing. Based on this research of developing methods, conclusions will be drawn as to which future screening strategy holds the most promise.
Sensitivity and specificity measure the number of false positives and false negatives, and are useful in evaluating the effectiveness of screening methods. Sensitivity is the number of true positive results divided by the sum of the true positive results and false negative results (refer to equation 1). Specificity is the number of true negative results divided by the sum of the true negative results and false positive results (refer to equation 2, Prostate Cancer Research Institute 2001). Our primary goal is to find a screening strategy that will improve sensitivity without sacrificing specificity.
At the present time, the PSA test and Digital Rectal Examination (DRE) are the most widely used forms of prostate cancer screening. The PSA test is simply a blood test, widely available to the general population. It is inexpensive, costing the patient roughly $30 to $60, and there are no risks or side effects (Crawford et. al. 1999). Though this would appear to be the ideal screening strategy, the sensitivity is 34.9% and the specificity is 63.1%. These values use age-specific reference ranges, which are more exact than the general PSA test (Marcus 2004). The DRE is also inexpensive, costing approximately $28 (Crawford et. al. 1999). Though it is readily available by appointment in a doctor's office, there is discomfort for the patient and a risk of slight bleeding (American Cancer Society 2006). With the DRE test, the experience of the doctor is of utmost importance, yet, new, inexperienced urologists often perform the test. The sensitivity is 27.1% and the specificity is 49.0% (Marcus 2004). Although these two tests are often combined with each other, the sensitivity is still a low 38.0%, and the specificity is 87.9% (Marcus 2004).
Mass spectrometry marks the next wave in prostate cancer detection, proven effective in the identification and analysis of varying biomarkers. In 1987, Matrix Assisted Laser Desorption/Ionization (MALDI) was introduced as a superior method of mass spectrometry analysis by increasing ease and sensitivity of spectrometer usage. The process consists of a molecule sample intermixed with a matrix compound and exposed to laser energy to produce sample ionization (Lennon 1997). Serum proteomics is identified as a promising screening strategy to detect prostate cancer in its earliest stages using MALDI to produce a spectrum of protein patterns, later analyzed with current databases for diagnosis.
An evaluation of the effectiveness of today's most commonly used methods and potential future strategies for prostate cancer screening is essential in resolving the problem at hand. A clear definition of effectiveness is required to properly examine the many aspects of all potential methods found. The primary components of effectiveness used in this study consist of the following: sensitivity and specificity, availability, side effects, ease of use, and cost. Based on the issues of the problem, sensitivity and specificity were awarded the most attention and considered of substantial importance in the appraisal of tests both currently used and those in development.
The parameters of the problem, as limited to sensitivity and specificity, are defined as the measure of false negatives and false positives, respectively (Prostate Cancer Research Institute 2001). Availability is assessed in regards to when and where the procedure will take place, for example, if testing is restricted to laboratories only, or if it will not reach clinical use for an unspecified period of time. Although side effects are rarely an issue, they are taken into account due to the impact on an individual's examination experience. Ease of use involves the degree of invasiveness of the test upon the patient, as well as the level of operator experience required to perform such a test. Finally, cost is identified as the expense for the patient, not the healthcare facility.
These elements are important in the clear presentation of the criteria for effectiveness. In regards to the plethora of screening possibilities, these conditions were used to narrow down the initial search results and focus on one final solution: the method of serum proteomics. Throughout the process of researching these methods, information obtained directly from specialists was considered invaluable in conjunction with peer-reviewed articles from highly respected journals. The interpretation of effectiveness is key in this study and crucial in defining the confines of this review.
Following an extensive assessment of future prostate cancer screening methods, it has been concluded that serum proteomics will provide the most promising results in terms of increasing sensitivity without sacrificing specificity. The serum proteomic process is illustrated in Figure 1. A specific cell possesses thousands of different proteins, which are distinguished through the method of serum proteomics (U.S. National Institutes of Health). Although currently available in laboratories only, serum proteomics focuses on analyzing patterns of proteins within a given blood or serum sample using mass spectrometry in order to uncover pre-Stage and Stage I cancer (Conrads et. al. 2003). Pathological states can often be exposed through this method, most importantly by identifying changes in the serum proteome. Due to the fact that proteins can be altered by any tissue saturated by blood, abnormalities or pathological states in both surrounding organs and tissues can be examined for such irregularities through this single test (Petricoin et. al. 2002). By observing the entire proteome, serum proteomics increases its sensitivity and specificity from testing that focuses on a single biomarker, such as future urine testing that concentrates specifically on the PCA3 gene, a gene over expressed in prostate cancer tissue by a factor of approximately 34 (Strum 2005).
Serum proteomics testing begins with a blood sample, provided by the patient, applied to a protein-binding chip (refer to Figure 1). With only a fraction of proteins adherent to the chromatographic surface of the chip, these proteins are exposed to an acid solution to allow for ionization and dried, unbound proteins are washed away. The remaining complex consists of multiple protein samples arranged in rows on the protein chip. Then placed into a vacuum chamber, each sample is desorbed into numerous ionized proteins through the use of laser energy. These particles are expelled down the chamber towards an oppositely charged detector (refer to Figure 2). Time-of-flight values are computed by the time lapse between the initial launch of the ions and the time at which they reach the electrode. As the calculated time-of-flight of the individual particles directly correlates to their mass to charge ratios (m/z), specialized software is employed to evaluate these mass to charge ratios. The data is run through a mass spectrometer, producing a mass spectrum arranged by ionic size, due to the fact that small ions travel at greater speeds. Each protein possesses a specific pattern of peptides, confirming the importance of the spectrum (U.S. National Institutes of Health). Using artificial intelligence, the protein pattern spectrum is compared with input data and spectrums from both affected and unaffected populations and the presence of cancer is determined at this point. At the present time, proteomic pattern analyses have demonstrated sensitivities and specificities with respect to prostate cancer consistently exceeding that of 90% (Conrads et. al. 2003). However, evidence from ovarian cancer testing suggests that higher resolution mass spectrometry could increase the sensitivity and specificity of such results. Greater resolution accounts for an increase in reproducibility and enhanced clarity by increasing the number of mass spectrum peaks. Lower resolutions of mass spectrometry that are currently used cannot distinguish between ions with similar mass/charge ratios as well, producing a single peak from multiple ion coalescence (Petricoin et. al. 2002). The diagnostic benefits of entirely replacing low-resolution SELDI by that of higher resolution, however, are yet to be determined.
While serum proteomics with mass spectrometry incorporates obvious advantages for diagnostic purposes, several limitations must be noted. As human proteins undergo post-translational modifications, there is speculation that such variations could alter mass spectrometry results. Other possible restrictions include the disintegration of proteins into smaller, analogous components and the necessity of operator experience for related software and expensive equipment.
Discussion and Conclusions
Serum proteomics is an invaluable testing method for prostate cancer that satisfies every category of effectiveness and surpasses all current methods. Sensitivity and specificity values, although repeatedly extending beyond 90%, will reach a required minimal level of 99.6% prior to clinical use, expected by 2008 (Conrads et. al. 2003). As can be inferred from data previously referenced, in comparison to current methods, these statistics top the sensitivity and specificity for PSA and DRE combined by 61.6% and 11.7% respectively. The simplicity of the process adds to the effectiveness in that testing is quick and efficient for all populations, with the ability to screen for an assortment of different cancers and diseases through a single test. Although the exam is fairly expensive as compared with PSA and DRE, estimated at $100-$200, its impressively high sensitivity and specificity will eliminate unnecessary additional treatments or emotional suffering for both cancer and non-cancer patients (Conrads et. al. 2003). In addition, there are no known side effects at this time. Dr. David K. Ornstein, author of "Serum Proteomic Patterns for Detection of Prostate Cancer" from the Journal of the National Cancer Institute, observes that this new technology may revolutionize the way men are diagnosed with prostate cancer (Lang 2002).
The purpose of this inquiry is to demonstrate the inferiority of the PSA test and DRE when compared to potential screening techniques. This investigation focuses on the sensitivity and specificity of various screening methods and provides guidelines for further evaluating those methods. Based on the extensive research that has been performed in this area, serum proteomics is recommended as a promising and distinct future strategy.
Currently the PSA test and DRE are the most commonly used forms of prostate cancer screening in existence. However, research shows that sensitivity is significantly lower than 50% for the PSA test, DRE, and for both tests combined, thus resulting in a great deal of false negatives (Marcus 2004). The PSA test returns many false positives due to contributing factors that have the potential to elevate PSA levels such as inflammation, infection, and benign prostate hyperplasia (BPH). Other general factors that increase PSA levels include age, race, and family history. Since every individual has a normal unique PSA level, the setting of the standard normal level at 4 ng/mL often leads to inconclusive results, lowering the sensitivity and specificity levels for the process as a whole (Beth).
False test results can also occur using the DRE in situations where patients exhibit inflammation of the prostate or BPH (Beth). The experience of the urologist performing the test is another critical factor in maintaining a high level of sensitivity and specificity. Despite the fact that PSA and DRE are often used in conjunction with one another, they continue to yield incorrect test results. This situation invokes a great deal of concern in the healthcare industry.
Causing the deaths of 30,350 American men in the year 2005 alone, prostate cancer is clearly a serious medical issue that is worthy of further investigation (American Cancer Society 2006). Serum proteomics has been identified as a potential screening strategy to detect prostate cancer in the earliest stage using patterns of proteins as a diagnostic fingerprint. A mass spectrometry system identifies the protein patterns of the prostate and through artificial intelligence analyzes and detects prostate cancer with the utmost efficiency (Conrads et. al. 2003). It currently has a sensitivity of 95% and a specificity of 96% (Marcus 2004). This method has been proven to be considerably superior to the PSA test and DRE for prostate cancer screening purposes. Although the PSA test and DRE combined show an increase in specificity, the combined sensitivity value remain drastically lower than serum proteomics. This new screening technique gives credible results after the initial test, rather than forcing the patient into follow-up testing, as do the methods of today.
Serum proteomics will have an outstanding sensitivity and specificity at the time of expected release in 2008. A blood test with the ability to detect pre-Stage and Stage I cancer holds as the latest promising technology. The advances in diagnostic imaging using mass spectrometry combined with proteomic pattern profiling will increase the overall sensitivity and specificity of prostate cancer screening and reduce uncalled-for biopsies (Rosenblatt et. al. 2004).
We would like to thank Dr. Joe LeDoux for his guidance and encouragement as our mentor. We would also like to thank Dr. William Catalona of Northwestern University, Dr. Patrick Walsh of Johns Hopkins Hospital, and Dr. Wendy Newstetter along with the Biomedical Engineering department at Georgia Tech for their advice and support.References
American Cancer Society. "Can Prostate Cancer Be Found Early?" American Cancer Society,
Inc. 26 June 2006. 18 July 2006. .
Beth Israel Deaconess Medical Center. "PSA Test and Digital Rectal Exam (DRE)" 25 Jan. 2005
Conrads, Thomas. et. al. (2003) Cancer Diagnosis Using Proteomic Patterns. NCI Journal Expert
Rev Mol Diagn 3, 411-420.
Crawford, ED. et. al. (1999) Efficiency of Prostate-Specific Antigen and Digital Rectal
Examination in Screening, Using 4.0 ng/ml and Age-Specific Reference Range as a Cutoff for Abnormal Values. The Prostate Journal 38.4, 296-302.
Creel, Howard. (1993) Prospects for the Analysis of High Molar Mass Polymers Using MALDI
Mass Spectrometry. Trends in Polymer Science. 1.11, 336-342.
Lang, Leslie. "Protein Patterns May Aid Prostate Cancer Detection" UNC School of Medicine.
15 Oct. 2002. 3 Feb. 2006. .
Lennon, John. "Matrix Assisted Laser Desorption Ionization Time-of-flight Mass Spectrometry."
University of Washington. 13 June 1997. 6 Jan. 2006. .
Marcus, Amy Dockser. "The Hidden Toll of Cancer Testing: Research Shows That Even Benign
Results Have Emotional and Financial Consequences." The Wall Street Journal Online, D1 (2004). 22 Jan. 2006. .
Petricoin, E. et. al. (2002) Serum Proteomic Patterns for Detection of Prostate Cancer. Journal of
the National Cancer Institute 94.20, 1576-78.
Prostate Cancer Research Institute. "Glossary of Prostate Cancer Related Terms" Prostate Cancer Research Institute. Aug. 2001. 28 Jan. 2006 < http://www.prostatecancer.org/resource/glossary.html>.
Rosenblatt, Kevin. et. al. "Serum Proteomics in Cancer Diagnosis and Management." Annual
Review of Medicine. 55, 97-112 (2004). 7 Jan. 2007. .
Strum, Stephen. et. al. "What Every Doctor Who Treats Male Patients Should Know" Prostate
Cancer Research Institute. May 2005. 19 Dec. 2006. .
U.S. National Institutes of Health. "Clinical Proteomic Technologies for Cancer Tutorial."
National Cancer Institute. 30 Jan. 2006 .Works Consulted
American Cancer Society: "How is Prostate Cancer Diagnosed?" American Cancer Society, Inc.
10 Dec. 2005. 24 Jan. 2006. .
American Cancer Society. "New Prostate Cancer Marker Found." American Cancer Society, Inc.
3 April 2002. 23 Jan. 2006. .
Ciatto, S. et. al. "Comparing Modalities of Screening for Prostate Cancer: Digital Rectal
Examination + Transrectal Ultrasonography vs. Prostate-Specific Antigen." Tumori 81.4,
225-9 (1995). 18 Jan. 2006. .
Cohen, Steven. et. al. "PSA Screening." Medicine and Health Rhode Island. Mar. 2005.
26 Jan. 2006. .
Fradet, Y. et. al. (2004) uPM3, A New Molecular Urine Test for the Detection of Prostate Cancer. Journal of Urology 64.2, 311-315.
Lamm M.D., Tita. Bostwick Laboratories. Telephone Interview. 30 Jan. 2006.
"MALDI Mass Spectrometry." University of Southern Mississippi. 1995. 6 Jan. 2006.
Sauer, Mark. "Doctors Debate Whether Current Medical Methods Are an Effective Way of
Detecting Prostate Cancer." Union Tribune. 17 Jan. 2006. 23 Jan. 2006. .
Sreekumar, Arun. et. al. "Humoral Immune Response to Methylacyl-CoA Racemase and
Protate Cancer." Journal of the National Cancer Institute 96.11, 834-843 (2004).
26 Jan. 2006. .
"The Prostate." Wiley-Liss, Inc. 45.1, 19-35 (2001). 21 Aug. 2000 .
United States. Office of Technology Assessment. "Cost Effectiveness of Prostate Cancer
Screening in Elderly Men." Washington: GPO, 1995. 23 Jan. 2006. .
William Catalona, M.D. Northwestern University and Patrick Walsh, M.D. John Hopkins
Hospital. Telephone Interview. 30 Jan. 2006.