Simon Tanguay MD
Assistant Professor
Montreal General Hospital
McGill University

Early detection of prostate cancer is a priority to many physicians. Although the use of Prostate Specific Antigen (PSA) measurement has improved our ability to detect prostate cancer, this test remains limited by its lack of specificity. The use of variation in PSA measurement such as PSA velocity, PSA density or PSA age reference range has failed to significantly improve the accuracy of prostate cancer detection over the use of total PSA alone. The recognized variation in both free and alpha 1-antichemotrypsin bound PSA content observed in patients with different prostatic diseases was shown to improve our ability to detect prostate cancer.¹ Although numerous reports have confirmed the ability of the Free over Total (F/T) PSA ratio to improve prostate cancer detection, the benefit associated with the use of the complexed PSA (cPSA) measurements remains to be determined. Our study was designed to compare total PSA, F/T PSA ratio and cPSA in their ability to improve either sensitivity or specificity of prostate cancer detection.

Between January 1998 and October 1999 573 patients evaluated at the UROMED prostate cancer detection clinic accepted to have blood drawn at the time of their initial evaluation. This patient population was referred to our center exclusively by urologists because of suspicion of prostate cancer based either on an elevated PSA or an abnormal Digital Rectal Examination (DRE). The blood samples were drawn before any prostatic manipulation. The transrectal ultrasound was performed using standard technique and at least 6 peripheral zone biopsies were performed in each patient.

Of the 573 patients evaluated, 535 patients had a complete PSA evaluation performed. This evaluation included a Total and Free PSA using the Axsym assay (Abbott laboratories) and a Total and Complexed PSA using the Bayer Immuno 1 assay. Of the 535 patients evaluated, 204 (38.1%) had cancer detected and 331 had a benign histology. Patients with high grade Prostatic Intraepithelial Neoplasia (PIN) were included in the benign group for this analysis. As shown in Table 1, the mean and median values of total PSA, F/T PSA and cPSA were statistically different between patients with benign or malignant histology.

Table 1:

	CANCER n = 204	BENIGN n = 331	P value
Total PSA (ng/ml)
Mean	14.9	7.2	P< 0.000001
Median	7.7	6.0
Range	1.2 - 423.8	0.3 - 104.5
Free/Total PSA (%)
Mean	14.2	22.1	P< 0.000001
Median	12.5	20.8
Range	1.8 - 76.6	0.3 - 83.3
Complexed PSA (ng/ml)
Mean	13.7	6.4	P< 0.000001
Median	7.3	5.2
Range	1.1 - 363.8	0.2 - 84.9

The sensitivity and specificity obtained with the use of total PSA were evaluated. Using the standard cutoff value of 4.0 ng/ml, a sensitivity and specificity of 87% and 27% respectively was observed. As expected, if the PSA cutoff was lowered to 2.5 ng/ml, the sensitivity increased to 96% at the expense of a decrease in specificity to 13%. The use of digital rectal examination (DRE) alone resulted in a sensitivity of 48% with a specificity of 77%. On the other hand, using the PSA density (PSAd) cutoff of 0.15 resulted in a sensitivity and specificity of 68% and 75% respectively.

One of the challenges faced by clinicians is to improve on the sensitivity and specificity of prostate cancer detection. Both F/T PSA ratio and cPSA have been proposed as tools to improve on cancer detection. Table 2 illustrates the sensitivity and specificity obtained using different F/T PSA cutoff on the entire patient population.

Table 2:

F/T PSA	Sensitivity	Specificity
9%	30%	93%
12%	47%	86%
15%	63%	75%
21%	84%	50%
23%	89%	41%
25%	91%	36%
27%	94%	30%
30%	97%	19%
Total PSA > 4.0 ng/ml	87%	27%

Table 3 illustrates the results obtained using different cutoff of cPSA on the same patient population. Only a marginal improvement in specificity was observed using the cPSA.

Table 3

cPSA (ng/ml)	Sensitivity	Specificity
2.5	94%	17%
2.75	93%	19%
3.0	92%	22%
3.25	89%	24%
3.5	87%	28%
3.75	85%	29%
4.0	83%	33%
Total PSA > 4.0 ng/ml	87%	27%

In order to compare all PSA assays, Table 4 illustrates the specificity obtained for the total PSA, F/T PSA ratio and the cPSA at similar sensitivity levels.

Table 4

	Total PSA		F/T PSA		cPSA
Sensitivity	Cutoff	Specificity	Cutoff	Specificity	Cutoff	Specificity
85%	4.0	27%	21%	50%	3.75	29%
90%	3.5	22%	25%	36%	3.25	24%
95%	3.0	18%	27%	30%	2.5	17%

These results clearly illustrates the superiority of the F/T PSA ratio compared to both total and complexed PSA. In fact, little if any difference was observed between the total and complexed PSA. The benefit of the F/T PSA ratio over both complexed and total PSA is also clearly illustrated by the ROC curves in Figure 1.

Figure 1

DISCUSSION

PSA remains one of the most useful markers currently available for prostate cancer detection. However, despite its adequate sensitivity, the use of PSA is limited by significant lack of specificity. In consequence, the clinical assessment of patients with an elevated PSA value will result in the performance of unnecessary prostatic biopsies in a substantial number of men. The search for improvement in specificity led to the evaluation of multiple variation in PSA measurement such as PSA velocity, PSA density and PSA age reference ranges. Unfortunately, none of these clinical parameters were able to significantly improve our ability to detect prostate cancer.

Numerous studies have evaluated the variation in Free and complexed PSA levels in patients with benign and malignant diseases of the prostate.^2-4 A reduction in the proportion of Free PSA was demonstrated in the majority of patients with prostate cancer. For example, Carlson et al, demonstrated in a large cohort of patients with PSA values between 4 and 20 ng/ml that the F/T PSA ratio was able to stratify patients according to their risk of having prostate cancer.⁵ In fact, they reported a cancer detection rate of 56% in patients with a F/T PSA ratio of less than 7% compared to only 16% of cancer detected when the ratio is greater than 25%. From this evaluation, the authors concluded that significant improvement in both sensitivity and specificity was associated with the use of the F/T PSA ratio compared to the use of PSA alone.

Although several cutoffs for the F/T PSA ratio have been proposed, the cutoff selected should be individualized according to the expectations of each physician and patient. Therefore, if one wished to improve the sensitivity and detect more cancer while maintaining the same ratio of biopsy performed, our results would favor 27% as the ideal F/T PSA ratio cutoff. Using this cutoff, the specificity is maintained at a level similar to the use of PSA alone (30%) but the number of cancer detected would increase from 87% to 94%. On the other hand, if the goal is to avoid unnecessary biopsies while maintaining a similar prostate cancer detection rate, a cutoff of 21% should be used. Using this strategy, the cancer detection rate remains relatively stable with a sensitivity of 84%, but in return, 23% of biopsies could be avoided.

The characterization of the different PSA subsets present in circulation provided us with alternatives to the use of total PSA. The proportion of PSA bound to a1-antichomotrypsin constitutes the majority of the detectable PSA in circulation. In consequence, it is now well recognized that the proportion of cPSA is increased in patients with prostate cancer. The practical use of cPSA has the theoretical advantage of being able to provide the benefit observed with the F/T PSA ratio without the necessity to obtain more than one laboratory value. In fact, Brawer et al, initially reported an improved prostate cancer detection rate using the cPSA cutoff of 3.75 ng/ml when compared to the F/T PSA cutoff of 25%.⁶ However, in the recent article published in the Prostate Update, Brawer reported on the use of cPSA in a larger patient population and concluded that the enhanced specificity observed with cPSA was similar to the results observed with F/T PSA. In a recent publication, Maeda et al. reported that cPSA was more accurate in predicting the outcome of prostatic biopsy than both total or F/T PSA ratio.⁷ In addition, they demonstrated a significant improvement in using the adjusted cPSA for the transitional zone volume in patients with prostate smaller than 45 cc when compared to all other diagnostic modalities. However, one must keep in mind that this group previously reported no difference between the total and cPSA and that their patient population was relatively small with only 23 out of 137 patients evaluated having biopsy proven adenocarcinoma of the prostate.

Our results are consistent with the results reported by Stamey et al where no benefit could be shown for the use of cPSA.⁸ We believe that our patient population of 535 patients, of which 204 had cancer, is of sufficient size to limit bias related to sample size. The ROC clearly demonstrates the reduced performance of the cPSA compared to the F/T PSA ratio. However, it is possible, as suggested by other investigators that various combinations using cPSA to produce ratios may improve the sensitivity and specificity.

In conclusion, our results confirmed other reports where only the F/T PSA ratio is able to improve the sensitivity or specificity observed with the use of total PSA alone.


1. Polascik TJ, Oesterling JE and Partin AW.: Prostate specific antigen: A decade of discovery – what we have learned and where we are going. J Urol 162: 293-306, 1999.

2. Catalona WJ, Partin AW, Slawin KM, et al.: Use of the percentage of free prostate specific antigen to enhance differentiation of prostate cancer from benign prostatic disease: A prospective multicenter trial. JAMA 279: 1542-1547, 1998.

3. Veltri RW and Miller MC.: Free/Total PSA ratio improves differention of benign and malignant disease of the prostate: Critical analysis of two different test populations. Urology 53: 736-745, 1999.

4. Woodrum DL, Brawer MK, Partin AW, et al.: Interpretation of free prostate specific antigen clinical research studies for detection of prostate cancer. J Urol 159: 5-12, 1998.

5. Carlson GD, Calvanese CB, and Partin AW.: An algorithm combining age, total prostate-specific antigen (PSA), and percent free PSA to predict prostate cancer: results on 4298 cases. Urology 52: 455-461, 1998.

6. Brawer MK, Meyer GE, Letran JL, et al.: Measurement of complexed PSA improves specificity for early detection of prostate cancer. Urology 52: 372-378, 1998.

7. Maeda H, Arai Y, Aoki Y, et al.: Complexed Prostate-Specific antigen and its volume indexes in the detection of prostate cancer. Urology 54: 225-228, 1999.

8. Stamey TA, Yemoto CE: Examination of 3 molecular forms of serum prostate specific antigen for distinguishing negative from positive biopsy: Relationship to transition zone volume. J Urol 163: 119-126, 2000.