Journal of Surgical Oncology 49: 172-175 (1992)
Prostate-Specific Antigen, Digital Rectal Examination, and Transrectal Ultrasound in Predicting the Probability of Cancer JOSEPH R. DRAGO, MD, AND JEFFREY P. YORK, MD From the Division of Urology and Surgery, The Ohio State University, Columbus
Over a 4% year period, 1,940 asymptomatic men were entered in a prostate cancer detection program consisting of digital rectal examination (DRE), prostate-specific antigen (PSA), and transrectal prostate ultrasound (TRUS). Four hundred and sixteen biopsies were performed resulting in the diagnosis of 79 cancers; 82% had clinically organ confined tumors. A recommendation for biopsy was made in 260 (62%) based on the TRUS alone, 55 (1 3%) by DRE alone, 92 (22%) when the DRE and TRUS were both abnormal, and in 9 (2.2%) cases when only PSA levels were elevated. The DRE, PSA, and TRUS were abnormal in 1,261 (65%), 989 (51%), and 1,552 (80%) of the patients with cancer, respectively. Prostate cancer detection increased as the serum PSA level increased above 4 ng/ml. The positive predictive value of both DRE and TRUS were significantly influenced by an elevated PSA, ( P = .042 and P < .00005, respectively). The results of this study support the idea that, although the prostate cancer detection rate is influenced by these three modalities and the detection rate of localized disease can be improved by early detection programs, its effect on mortality rates remains undefined at this time. 0 1992 Wiley-Liss, Inc.
KEYWORDS:organ-confined tumors, prostate cancer detection, digital rectal examination
INTRODUCTION The goal of prostate cancer detection is to identify men with potentially curable or clinically organ confined disease. The individual applications of digital rectal examination (DRE), transrectal ultrasound (TRUS), and prostate-specific antigen (PSA) in prostate cancer detection and staging have been well established. Additionally, TRUS appears to be twice as sensitive as DRE in the detection of prostate cancer. However, despite the fact that these three modalities have been clinically available for several years, their combined application in the detection of prostate cancer is just beginning to be appreciated. Herein we report The Ohio State University experience in detecting prostate cancer in 1,940 asymptomatic men using PSA, DRE, and TRUS.
entered into a program of DRE, TRUS, and PSA. DRE and TRUS was performed by a urologist. Either a Bruel & Kjaer 1846 (Marlborough, MA) or a Proscan ultrasound machine using similar 7.0 MHz transducers was used. The serum PSA determination was obtained prior to DRE or TRUS. PSA was analyzed by the monoclonal TANDEM-RPSA or IRMA-Count assays (Hybritech, Inc., San Diego, CA; Diagnostic Products, Los Angeles, CA). Repeat examinations were performed annually, and by the time of this analysis, 416 biopsies were done. Of these, 320 biopsies were conducted following the initial examinations, 80 followed the second year examinations, and 16 were performed on the basis of findings from the third annual examinations.
MATERIALS AND METHODS Between January 1986 and June 1990, 1,940 men without any known or suspected prostate pathology were 0 1992 Wiley-Liss, Inc.
Accepted for publication December 9, 1991 Address reprint requests to Dr. Joseph R . Drago, Professor and Chief, Division of Urology and Surgery, The Ohio State University, 456 W. Tenth Avenue, Columbus, OH 43210.
Use of PSA, DRE, and TRUS in Prostate Cancer TABLE 1. Negative BioDsv and Prostate-Swdfic Antigen (PSA) PSA level
PSA (%)
Total (%)
14 >4
287 (83) 59 (17) 346
355 (82.2) 77 (17.8) 432
TABLE 11. Relation of Prostate-Specific Antigen (PSA) and Diagnosis of Cancer in 79 Patients PSA level
14 >4
173
TABLE 111. Distribution of Prostate-Specific Antigen (PSA) by Pathologic Outcomes No. at PSA levela Pathological Stage
A B C D
Total
0.3-2
2-4
4-10
>I0
11 53 11 4
5 6 1
3
2 36 3 -
10
-
1
4 2
3 2
aNorrnal range Ohio State University laboratory = 0.3-4 ng/rnl.
PSA
22 (29%) 57 (71%) -
The relationship of PSA levels to other detection modalities and to the finding of cancer is shown in Table 11. Biopsy was performed based on TRUS alone in 260 In preparation for a biopsy, patients had a Fleets enema (62%), on DRE alone in 55 (13.1%), on both TRUS and and were administered either Noroxin, 400 mg, or Cipro, DRE in 92 (22%), and on PSA alone in 27 patients. 500 mg orally. Antibiotics were repeated 12 and 24 hours Among the 79 men with cancer, ultrasound had the after biopsy. Patients with a hypoechoic lesion on ultra- greatest yield in terms of cancers detected. Relationship sound underwent an ultrasound-guided biopsy using the of stage and PSA reveals a correlation with increasing Biopty gun (Bard, Atlanta, GA) or an 18 gauge “tru-cut” stage in general increasing PSA (Table 111). A hyponeedle of lesions measuring at least 5 mm in two projec- echoic area was identified by ultrasound alone or in comtions. A finger-guided biopsy using the Biopty gun was bination with abnormal DRE in 58 of the 79 who proved performed in patients who had only an abnormal DRE. to have cancer. A total of 47 of 79 cancers were suspected Biopsies were obtained using ultrasonic guidance alone, on the basis of the TRUS alone or in combination with an digitally directed alone, or both in 351/416 (85%), 48 abnormal DRE (Table IV). ( I l%), and 17 (4%), respectively. The positive predictive value, i.e., number of cancers The mean age of the study population was 64 years detected by a test relative to the number of biopsies rec(SD = 3.8) with a range of 55-70 years. The mean PSA ommended by that test, may be a better indicator of the value was 3.9 ng/ml (SD = 8.9) with a range of 0.6- efficiency of a detection tool. Overall, irrespective of the 100.0. The racial distribution of men biopsied shows a PSA value, a biopsy recommended by TRUS alone was preponderance of Caucasians (35 1; 85%) over blacks (65; least likely to prove to be cancer, with only 40 of 260 (1 3%) proving to be cancer. Similarly, the positive pre15%). Prostate volume was measured in these patients. The dictive value of the DRE was similarly low, with 7 of 55 PSA to volume ratio was also calculated and represents (13%) biopsies recommended by DRE alone yielding the PSA density (PSAD). cancer. When both TRUS and DRE were positive and without considering the PSA value, the positive predicRESULTS tive value was 32% (30 of 92). The predictive value of a recommendation to biopsy The distribution of the biopsy outcomes for the 416 men and the distribution of PSA levels according to out- based on TRUS alone increases threefold when biopsies come are shown in Table I. A total of 79 cancers (17%) of are restricted to men with PSA > 4.0 ng/ml (1 1% vs. 34%, those biopsied were detected, 92% of which were clini- P = .OOOOl). When only men with an elevated PSA cally organ confined. Other outcomes of the biopsy in- (> 4.0 ng/ml), abnormal TRUS, and abnormal DRE are cluded 142 (34%) patients with benign prostatic hyper- considered for biopsy, the predictive value of the diagplasia, 118 (27%) with normal tissue, 50 (13%) with nostic triad increases to 69%. The data in Table I1 show that the influence of PSA on inflammation, and 36 (9%) who had prostatic intraepithethe detection of cancer has an increasing specificity assolial neoplasia. PSA levels were increased in men with cancer. The ciated with increasing PSA level. At the highest levels of mean PSA level for cancer patients was 14 ng/ml com- PSA (> 10.0 ng/ml), the predictive values of a recompared to 3. I ng/ml for all other outcomes ( P < .001). The mendation to biopsy based on TRUS only rose to 60%. In overall incidence of cancer in men undergoing biopsy in those relatively few cases for which both TRUS and DRE this series based on whether their PSA value was 4 or are in agreement in recommending biopsy, the likelihood > 4 was 8% vs. 42%, respectively ( P < .00005) (Tables of cancer being detected when PSA is greater than 10.0 ng/ml is 84%. I, 11). 79
174
Drago and York
TABLE IV. Proportion of Biopsies Positive for Cancer According to Level of Prostate-Specific Antigen (PSA) Assay, Overall, and Transrectal Ultrasound (TRUS) and Digital Rectal Examination (DRE) Results" ~
All combinations TRUS and DRE negative DRE only positive TRUS only positive TRUS and DRE positive
>o.o
>2.0
791416 (18) 0 7/55 (13) 40/260 (21) 30192 (3 1.2)
78/262 (30) 0 6/55 (15) 251130 (17) 30160 (50)
* Number positive for cancer/total number in category; percent
PSA (ndml) >4.0 >6 0 59/137 (45) 119 (1 1) 5/20 (25) 21/65 (31.8) 2 1/35 (60)
47/91 (52) 3/9 (33) 3/11 (33) 6/40 (37) 2 1/25 (72)
>8.0
>10.0
40/66 (63) 2 / 5 (40) 216 (33) 1 1 125 (44) 17/20 (85)
34/51 (69) 314 (75) 216 (33) 8/14 (49) 13/16 (82)
positive for cancer in parentheses.
The high specificity for biopsies conducted when both TRUS and DRE were negative must be evaluated in light of the small numbers. All had PSA > 4.0 ng/ml and 7 of the 9 (72%) in this category had PSA > 10.0 ng/ml. In these patients with cancer the PSAD was greater than .6 in 63 of the 79 (89%).
DISCUSSION The clinical significance of PSA as a parameter in the diagnosis of cancer of the prostate has increased over the last several years. The high false-positive rates of the DRE and TRUS when used alone or in combination are well known [ 1-31. Incorporation of the PSA assay into the judgment process can significantly impact the falsenegative and false-positive rates. Cooner and associates reported the impact of the serum PSA value irrespective of the TRUS on the incidence of cancer in men with an abnormal DRE. Drago et al. have had similar findings as have others [4,5]. The use of the PSAD may be of benefit in determining which patient with an elevation of PSA requires a biopsy even when the TRUS is normal. In our patients when the PSA/prostate volume ratio is b .6 there was an 80% chance of finding cancer with the biopsy. When this ratio was a . 2 only 7% of the patients had cancer. Babaian and Camps [6] reported similar results, finding a 23.1 % and 19.6% cancer incidence in men with an abnormal DRE (regardless of TRUS) and an abnormal TRUS (regardless of DRE), respectively, if the PSA was 9 4 compared to 77.8% and 71.2%, respectively, if the PSA was >4. The overall cancer detection rate in the NPCDP study is 2.4%, which is comparable to other reports based on similar populations [7-91. The comparable cancer incidence in this series for men with an abnormal DRE or abnormal TRUS is 24% and 16%, respectively. This study at a single institution verifies that while TRUS detects approximately twice as many cancers as the DRE, it results in considerably more biopsies being performed when the decision to biopsy is based solely on the results of either test alone. This study reasserts the association of the degree of PSA elevation with the presence of prostate cancer.
While 66% of all cancers detected in this study had a PSA >4, the incidence of cancer increased as the PSA value increased whether the DRE or TRUS alone or combined were abnormal. Patients with either an abnormal DRE or TRUS and a PSA value >4 ng/ml had a greater likelihood of having cancer compared to those men with a PSA value s 4 (44.4% vs. 8.6%, P < .00005). In addition, there was a statistically significant difference ( P < 0.0004) in the cancer incidence in men with either one abnormal test (DRE or TRUS). The positive predictive value of both DRE and TRUS alone are significantly influenced by PSA if the value is 5 or > 4, P = .044 (6.4% vs. 31.6%) and P d .00005 (5.4% vs. 59.5%), respectively, and the diagnosis of cancer is made more frequently with PSA 3 4 and any abnormality of DRE and/or TRUS. The status of the PSA value (G4 or >4) also significantly influenced the positive predictive value even when both the DRE and the TRUS were abnormal, P < .00005. This finding of enhanced positive predictive value as the number of positive diagnostic tests increases has been reported by other investigators. Based on the data of this study and that of others, one can begin to question the need to biopsy lesions that are abnormal by neither DRE or TRUS and are associated with a PSA d 4. However, with such an approach the possibility of missing 4 6 % of potentially curable cancers is present. As more understanding of the inner gland on TRUS occurs, more of these inner gland lesions may be discovered. The respective cancer incidences in this series of 416 biopsies when the PSA is 9 4 are 6.6% for the DRE and 5.6% for the TRUS. In addition, an early detection program can substantially increase the detection rate of clinically organ-confined disease, which may have an impact on disease survival.
ACKNOWLEDGMENTS This work was supported in part by the University Urology Education & Research Foundation, Medical Research and Development Fund, and Surgical Research, Inc.
Use of PSA, DRE, and TRUS in Prostate Cancer
REFERENCES 1 . Mettlin C: National prostate cancer detection program. Paper pre-
2. 3. 4.
5.
sented before Third International Symposium on Transrectal Ultrasound, Chicago, September 23-24, 1988. Stamey TA, Yang M, Hay A, et al: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. N Engl J Med 3 1 7 :909-9 16, 1987. Lee F, Top-Pedersen ST, Siders DB: The role of transrectal ultrasound in the early detection of prostate cancer. CA 39:337-360, 1989. Drago JR, Badalament RA: Prostate cancer: Early detection and screening. In Resnick MI (ed): “Prostatic Ultrasonography.” Toronto: BC Decker, 1990. Scardino PT, Shinohara K: The value of transrectal ultrasonogra-
6. 7.
8. 9.
175
phy in staging of prostate cancer. In: “Transrectal Ultrasound in the Diagnosis and Management of Prostate Cancer.” Third International Symposium, 1988, pp. 30-34. Babaian RJ, Camps JL, Frangos DN, et al: Monoclonal prostatespecific antigen in untreated prostate cancer. Relationship to clinical stage and grade. Cancer 67:2200, 1991. Drago JR, Nesbitt JA, Badalament RA: Use of transrectal ultrasound in detection of prostatic carcinoma: A preliminary report. J Surg Oncol41:27&277, 1989. Cooner WH, Eggers GW, Lichtenstein P Prostate cancer: New hope for early diagnosis. Ala Med 56:13-16, 1987. Lee F, Littrup PJ, Torp-Pedersen ST, et al: Prostate cancer: Comparison of transrectal US and digital rectal examination for screening. Radiology 168:389-394, 1988.
Prostate-Specific Antigen, Digital Rectal Examination, and Transrectal Ultrasound in Predicting the Probability of Cancer JOSEPH R. DRAGO, MD, AND JEFFREY P. YORK, MD From the Division of Urology and Surgery, The Ohio State University, Columbus
Over a 4% year period, 1,940 asymptomatic men were entered in a prostate cancer detection program consisting of digital rectal examination (DRE), prostate-specific antigen (PSA), and transrectal prostate ultrasound (TRUS). Four hundred and sixteen biopsies were performed resulting in the diagnosis of 79 cancers; 82% had clinically organ confined tumors. A recommendation for biopsy was made in 260 (62%) based on the TRUS alone, 55 (1 3%) by DRE alone, 92 (22%) when the DRE and TRUS were both abnormal, and in 9 (2.2%) cases when only PSA levels were elevated. The DRE, PSA, and TRUS were abnormal in 1,261 (65%), 989 (51%), and 1,552 (80%) of the patients with cancer, respectively. Prostate cancer detection increased as the serum PSA level increased above 4 ng/ml. The positive predictive value of both DRE and TRUS were significantly influenced by an elevated PSA, ( P = .042 and P < .00005, respectively). The results of this study support the idea that, although the prostate cancer detection rate is influenced by these three modalities and the detection rate of localized disease can be improved by early detection programs, its effect on mortality rates remains undefined at this time. 0 1992 Wiley-Liss, Inc.
KEYWORDS:organ-confined tumors, prostate cancer detection, digital rectal examination
INTRODUCTION The goal of prostate cancer detection is to identify men with potentially curable or clinically organ confined disease. The individual applications of digital rectal examination (DRE), transrectal ultrasound (TRUS), and prostate-specific antigen (PSA) in prostate cancer detection and staging have been well established. Additionally, TRUS appears to be twice as sensitive as DRE in the detection of prostate cancer. However, despite the fact that these three modalities have been clinically available for several years, their combined application in the detection of prostate cancer is just beginning to be appreciated. Herein we report The Ohio State University experience in detecting prostate cancer in 1,940 asymptomatic men using PSA, DRE, and TRUS.
entered into a program of DRE, TRUS, and PSA. DRE and TRUS was performed by a urologist. Either a Bruel & Kjaer 1846 (Marlborough, MA) or a Proscan ultrasound machine using similar 7.0 MHz transducers was used. The serum PSA determination was obtained prior to DRE or TRUS. PSA was analyzed by the monoclonal TANDEM-RPSA or IRMA-Count assays (Hybritech, Inc., San Diego, CA; Diagnostic Products, Los Angeles, CA). Repeat examinations were performed annually, and by the time of this analysis, 416 biopsies were done. Of these, 320 biopsies were conducted following the initial examinations, 80 followed the second year examinations, and 16 were performed on the basis of findings from the third annual examinations.
MATERIALS AND METHODS Between January 1986 and June 1990, 1,940 men without any known or suspected prostate pathology were 0 1992 Wiley-Liss, Inc.
Accepted for publication December 9, 1991 Address reprint requests to Dr. Joseph R . Drago, Professor and Chief, Division of Urology and Surgery, The Ohio State University, 456 W. Tenth Avenue, Columbus, OH 43210.
Use of PSA, DRE, and TRUS in Prostate Cancer TABLE 1. Negative BioDsv and Prostate-Swdfic Antigen (PSA) PSA level
PSA (%)
Total (%)
14 >4
287 (83) 59 (17) 346
355 (82.2) 77 (17.8) 432
TABLE 11. Relation of Prostate-Specific Antigen (PSA) and Diagnosis of Cancer in 79 Patients PSA level
14 >4
173
TABLE 111. Distribution of Prostate-Specific Antigen (PSA) by Pathologic Outcomes No. at PSA levela Pathological Stage
A B C D
Total
0.3-2
2-4
4-10
>I0
11 53 11 4
5 6 1
3
2 36 3 -
10
-
1
4 2
3 2
aNorrnal range Ohio State University laboratory = 0.3-4 ng/rnl.
PSA
22 (29%) 57 (71%) -
The relationship of PSA levels to other detection modalities and to the finding of cancer is shown in Table 11. Biopsy was performed based on TRUS alone in 260 In preparation for a biopsy, patients had a Fleets enema (62%), on DRE alone in 55 (13.1%), on both TRUS and and were administered either Noroxin, 400 mg, or Cipro, DRE in 92 (22%), and on PSA alone in 27 patients. 500 mg orally. Antibiotics were repeated 12 and 24 hours Among the 79 men with cancer, ultrasound had the after biopsy. Patients with a hypoechoic lesion on ultra- greatest yield in terms of cancers detected. Relationship sound underwent an ultrasound-guided biopsy using the of stage and PSA reveals a correlation with increasing Biopty gun (Bard, Atlanta, GA) or an 18 gauge “tru-cut” stage in general increasing PSA (Table 111). A hyponeedle of lesions measuring at least 5 mm in two projec- echoic area was identified by ultrasound alone or in comtions. A finger-guided biopsy using the Biopty gun was bination with abnormal DRE in 58 of the 79 who proved performed in patients who had only an abnormal DRE. to have cancer. A total of 47 of 79 cancers were suspected Biopsies were obtained using ultrasonic guidance alone, on the basis of the TRUS alone or in combination with an digitally directed alone, or both in 351/416 (85%), 48 abnormal DRE (Table IV). ( I l%), and 17 (4%), respectively. The positive predictive value, i.e., number of cancers The mean age of the study population was 64 years detected by a test relative to the number of biopsies rec(SD = 3.8) with a range of 55-70 years. The mean PSA ommended by that test, may be a better indicator of the value was 3.9 ng/ml (SD = 8.9) with a range of 0.6- efficiency of a detection tool. Overall, irrespective of the 100.0. The racial distribution of men biopsied shows a PSA value, a biopsy recommended by TRUS alone was preponderance of Caucasians (35 1; 85%) over blacks (65; least likely to prove to be cancer, with only 40 of 260 (1 3%) proving to be cancer. Similarly, the positive pre15%). Prostate volume was measured in these patients. The dictive value of the DRE was similarly low, with 7 of 55 PSA to volume ratio was also calculated and represents (13%) biopsies recommended by DRE alone yielding the PSA density (PSAD). cancer. When both TRUS and DRE were positive and without considering the PSA value, the positive predicRESULTS tive value was 32% (30 of 92). The predictive value of a recommendation to biopsy The distribution of the biopsy outcomes for the 416 men and the distribution of PSA levels according to out- based on TRUS alone increases threefold when biopsies come are shown in Table I. A total of 79 cancers (17%) of are restricted to men with PSA > 4.0 ng/ml (1 1% vs. 34%, those biopsied were detected, 92% of which were clini- P = .OOOOl). When only men with an elevated PSA cally organ confined. Other outcomes of the biopsy in- (> 4.0 ng/ml), abnormal TRUS, and abnormal DRE are cluded 142 (34%) patients with benign prostatic hyper- considered for biopsy, the predictive value of the diagplasia, 118 (27%) with normal tissue, 50 (13%) with nostic triad increases to 69%. The data in Table I1 show that the influence of PSA on inflammation, and 36 (9%) who had prostatic intraepithethe detection of cancer has an increasing specificity assolial neoplasia. PSA levels were increased in men with cancer. The ciated with increasing PSA level. At the highest levels of mean PSA level for cancer patients was 14 ng/ml com- PSA (> 10.0 ng/ml), the predictive values of a recompared to 3. I ng/ml for all other outcomes ( P < .001). The mendation to biopsy based on TRUS only rose to 60%. In overall incidence of cancer in men undergoing biopsy in those relatively few cases for which both TRUS and DRE this series based on whether their PSA value was 4 or are in agreement in recommending biopsy, the likelihood > 4 was 8% vs. 42%, respectively ( P < .00005) (Tables of cancer being detected when PSA is greater than 10.0 ng/ml is 84%. I, 11). 79
174
Drago and York
TABLE IV. Proportion of Biopsies Positive for Cancer According to Level of Prostate-Specific Antigen (PSA) Assay, Overall, and Transrectal Ultrasound (TRUS) and Digital Rectal Examination (DRE) Results" ~
All combinations TRUS and DRE negative DRE only positive TRUS only positive TRUS and DRE positive
>o.o
>2.0
791416 (18) 0 7/55 (13) 40/260 (21) 30192 (3 1.2)
78/262 (30) 0 6/55 (15) 251130 (17) 30160 (50)
* Number positive for cancer/total number in category; percent
PSA (ndml) >4.0 >6 0 59/137 (45) 119 (1 1) 5/20 (25) 21/65 (31.8) 2 1/35 (60)
47/91 (52) 3/9 (33) 3/11 (33) 6/40 (37) 2 1/25 (72)
>8.0
>10.0
40/66 (63) 2 / 5 (40) 216 (33) 1 1 125 (44) 17/20 (85)
34/51 (69) 314 (75) 216 (33) 8/14 (49) 13/16 (82)
positive for cancer in parentheses.
The high specificity for biopsies conducted when both TRUS and DRE were negative must be evaluated in light of the small numbers. All had PSA > 4.0 ng/ml and 7 of the 9 (72%) in this category had PSA > 10.0 ng/ml. In these patients with cancer the PSAD was greater than .6 in 63 of the 79 (89%).
DISCUSSION The clinical significance of PSA as a parameter in the diagnosis of cancer of the prostate has increased over the last several years. The high false-positive rates of the DRE and TRUS when used alone or in combination are well known [ 1-31. Incorporation of the PSA assay into the judgment process can significantly impact the falsenegative and false-positive rates. Cooner and associates reported the impact of the serum PSA value irrespective of the TRUS on the incidence of cancer in men with an abnormal DRE. Drago et al. have had similar findings as have others [4,5]. The use of the PSAD may be of benefit in determining which patient with an elevation of PSA requires a biopsy even when the TRUS is normal. In our patients when the PSA/prostate volume ratio is b .6 there was an 80% chance of finding cancer with the biopsy. When this ratio was a . 2 only 7% of the patients had cancer. Babaian and Camps [6] reported similar results, finding a 23.1 % and 19.6% cancer incidence in men with an abnormal DRE (regardless of TRUS) and an abnormal TRUS (regardless of DRE), respectively, if the PSA was 9 4 compared to 77.8% and 71.2%, respectively, if the PSA was >4. The overall cancer detection rate in the NPCDP study is 2.4%, which is comparable to other reports based on similar populations [7-91. The comparable cancer incidence in this series for men with an abnormal DRE or abnormal TRUS is 24% and 16%, respectively. This study at a single institution verifies that while TRUS detects approximately twice as many cancers as the DRE, it results in considerably more biopsies being performed when the decision to biopsy is based solely on the results of either test alone. This study reasserts the association of the degree of PSA elevation with the presence of prostate cancer.
While 66% of all cancers detected in this study had a PSA >4, the incidence of cancer increased as the PSA value increased whether the DRE or TRUS alone or combined were abnormal. Patients with either an abnormal DRE or TRUS and a PSA value >4 ng/ml had a greater likelihood of having cancer compared to those men with a PSA value s 4 (44.4% vs. 8.6%, P < .00005). In addition, there was a statistically significant difference ( P < 0.0004) in the cancer incidence in men with either one abnormal test (DRE or TRUS). The positive predictive value of both DRE and TRUS alone are significantly influenced by PSA if the value is 5 or > 4, P = .044 (6.4% vs. 31.6%) and P d .00005 (5.4% vs. 59.5%), respectively, and the diagnosis of cancer is made more frequently with PSA 3 4 and any abnormality of DRE and/or TRUS. The status of the PSA value (G4 or >4) also significantly influenced the positive predictive value even when both the DRE and the TRUS were abnormal, P < .00005. This finding of enhanced positive predictive value as the number of positive diagnostic tests increases has been reported by other investigators. Based on the data of this study and that of others, one can begin to question the need to biopsy lesions that are abnormal by neither DRE or TRUS and are associated with a PSA d 4. However, with such an approach the possibility of missing 4 6 % of potentially curable cancers is present. As more understanding of the inner gland on TRUS occurs, more of these inner gland lesions may be discovered. The respective cancer incidences in this series of 416 biopsies when the PSA is 9 4 are 6.6% for the DRE and 5.6% for the TRUS. In addition, an early detection program can substantially increase the detection rate of clinically organ-confined disease, which may have an impact on disease survival.
ACKNOWLEDGMENTS This work was supported in part by the University Urology Education & Research Foundation, Medical Research and Development Fund, and Surgical Research, Inc.
Use of PSA, DRE, and TRUS in Prostate Cancer
REFERENCES 1 . Mettlin C: National prostate cancer detection program. Paper pre-
2. 3. 4.
5.
sented before Third International Symposium on Transrectal Ultrasound, Chicago, September 23-24, 1988. Stamey TA, Yang M, Hay A, et al: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. N Engl J Med 3 1 7 :909-9 16, 1987. Lee F, Top-Pedersen ST, Siders DB: The role of transrectal ultrasound in the early detection of prostate cancer. CA 39:337-360, 1989. Drago JR, Badalament RA: Prostate cancer: Early detection and screening. In Resnick MI (ed): “Prostatic Ultrasonography.” Toronto: BC Decker, 1990. Scardino PT, Shinohara K: The value of transrectal ultrasonogra-
6. 7.
8. 9.
175
phy in staging of prostate cancer. In: “Transrectal Ultrasound in the Diagnosis and Management of Prostate Cancer.” Third International Symposium, 1988, pp. 30-34. Babaian RJ, Camps JL, Frangos DN, et al: Monoclonal prostatespecific antigen in untreated prostate cancer. Relationship to clinical stage and grade. Cancer 67:2200, 1991. Drago JR, Nesbitt JA, Badalament RA: Use of transrectal ultrasound in detection of prostatic carcinoma: A preliminary report. J Surg Oncol41:27&277, 1989. Cooner WH, Eggers GW, Lichtenstein P Prostate cancer: New hope for early diagnosis. Ala Med 56:13-16, 1987. Lee F, Littrup PJ, Torp-Pedersen ST, et al: Prostate cancer: Comparison of transrectal US and digital rectal examination for screening. Radiology 168:389-394, 1988.
