PROSTATE SPECIFIC ANTIGEN BASED PROSTATE CANCER SCREENING: ACCUMULATING EVIDENCE OF EFFICACY BUT PERSISTENT UNCERTAINTY

Other and All-Cause Mortality among Men Diagnosed with Prostate Cancer in the PLCO Trial

Relationship Among Initial Serum Prostate Specific Antigen, Prostate Specific Antigen Progression and Prostate Cancer Detection at Repeat Screening Visits

Mass Screening for Prostate Cancer in Korea: A Population Based Study

US Trial shows no early mortality benefit from annual prostate cancer screening

The benefit of screening for prostate cancer using prostate-specific antigen (PSA) testing and digital rectal examination (DRE) is uncertain and is under evaluation in a randomized prospective trial, the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Although the final results are several years away, the initial round of screening is complete. We describe the population enrolled in the PLCO trial, their baseline PSA and DRE screening results, and diagnostic follow-up results during the first year of follow-up. A total of 38,350 men were randomly assigned to the screening arm of the PLCO trial from November 1993 through June 2001. Men were advised to seek diagnostic follow-up from their primary care provider if their DRE was suspicious for cancer and/or if their serum PSA level was higher than 4 ng/mL. PLCO trial staff obtained records related to diagnostic follow-up. Compliance with both screening tests was high (more than 89%). At screening, 7.5% of men had a positive DRE (i.e., suspicious for cancer) and 7.9% had a PSA level higher than 4 ng/mL. Of the men with positive screening tests, 74.2% underwent additional diagnostic testing, and 31.5% underwent a prostatic biopsy within 1 year. Overall, 1.4% of the men in the screening arm were diagnosed with prostate cancer, the majority of whom had clinically localized cancer. These compliance, biopsy, and cancer detection rates appear to be representative of contemporary practice patterns. The PLCO trial is evaluating PSA- and DRE-based screening for prostate cancer in a clinically valid manner. Whether such screening will result in a reduction of prostate cancer mortality cannot be answered until the randomized comparison is completed.

To describe the results of the first four rounds (T0-T3) of prostate cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial (designed to determine the value of screening in the four cancers), that for prostate cancer is evaluating whether annual screening with prostate-specific antigen (PSA) and a digital rectal examination (DRE) reduces prostate cancer-specific mortality. In all, 38 349 men aged 55-74 years were randomized to undergo annual screening with PSA (abnormal >4.0 ng/mL) and a DRE. The follow-up of abnormal screening results was at the discretion of subjects' physicians. PLCO staff obtained records related to diagnostic follow-up of positive screen results. Compliance with screening decreased slightly from 89% at baseline to 85% at T3. Both PSA positivity rates (range 7.7-8.8% at T0-T3) and DRE positivity rates (range 6.8-7.6% at T0-T3) were relatively constant over time. The positive predictive value (PPV) of a PSA level of >4.0 ng/mL decreased from 17.9% at T0 to 10.4-12.3% at T1-T3; the PPV for DRE (in the absence of a positive PSA test) was constant over time (2.9-3.6%). Cancer was diagnosed in 1902 men (4.9%). Screen-detected cancers at T0 (549) were more likely to be clinical stage III/IV (5.8%) and to have a Gleason score of 7-10 (34%) than screen-detected cancers at T1-T3 (1.5-4.2% stage III/IV and 24-27% Gleason score 7-10 among 1054 cases). The present findings on serial prostate screening are similar to those reported from other multi-round screening studies. Determining the effect of PSA screening on prostate cancer mortality awaits further follow-up.

The United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) recently reported a reduction in the average overall mortality among ovarian cancer patients screened with an annual sequential, multimodal strategy that tracked biomarker CA125 over time, where increasing serum CA125 levels prompted ultrasound. However, multiple cases were documented wherein serum CA125 levels were rising, but ultrasound screens were normal, thus delaying surgical intervention. A significant factor which could contribute to false negatives is that many aggressive ovarian cancers are believed to arise from epithelial cells on the fimbriae of the fallopian tubes, which are not readily imaged. Moreover, because only a fraction of metastatic tumors may reach a sonographically-detectable size before they metastasize, annual screening with ultrasound may fail to detect a large fraction of early-stage ovarian cancers. The ability to detect ovarian carcinomas before they metastasize is critical and future efforts towards improving screening should focus on identifying unique features specific to aggressive, early-stage tumors, as well as improving imaging sensitivity to allow for detection of tubal lesions. Implementation of a three-stage multimodal screening strategy in which a third modality is employed in cases where the first-line blood-based assay is positive and the second-line ultrasound exam is negative may also prove fruitful in detecting early-stage cases missed by ultrasound.

Ovarian cancer typically presents late and has a poor prognosis. Ovarian screening aims to reduce mortality from the disease by detecting and treating women at an earlier, presymptomatic stage. The paper by Menon et al.1 published in BJOG more than a decade ago, set the stage for a multimodal approach to ovarian cancer screening, whereby a combination of CA125 levels and ultrasound scan findings could identify women at risk of the disease. Subsequent studies have not only informed the current management of women with suspected ovarian cancer but will also determine whether ovarian cancer screening can save lives. This commentary documents the ovarian screening story so far from inception to current day as we await the results from UKCTOCS, the largest ovarian cancer screening trial conducted to date. Ovarian cancer is known as the ‘silent killer’ because it only becomes symptomatic in its late stages. Compared with other common cancers, women diagnosed with ovarian cancer have a relatively poor prognosis. The overall 5-year survival rate of 42.9% masks vast differences in prognosis for those women presenting with Stage I disease compared with those with Stage IV disease.2 Since the late 1980s there has been a steady improvement in 5-year survival in the relatively small group of women who present with Stage I disease from 80% in 1987–91 to 92% in 2005–08.3 Changes in survival for the much larger proportion of women who present with Stage III and IV disease have been more modest. The 5-year survival rate for those with stage IV disease, however, stands unchanged at around 5%.3 These dismal survival rates for most women with ovarian cancer have stimulated enormous research effort aimed at identifying and treating women with early-stage disease. The hypothesis is that ovarian cancer screening may pick up early, asymptomatic disease in apparently healthy women that is then more amenable to surgical cure than disease that presents with symptoms. A major challenge when screening an asymptomatic population is that the proportion of true positives is extremely low given that the age-standardised incidence of ovarian cancer is only 16 per 100 000.4 Furthermore, the consequence of a positive screen is referral for laparoscopy/laparotomy with its incumbent risks. Hence, it has been suggested to ensure an acceptable false-positive rate any test needs to achieve a specificity of over 99.6%.5 This equates to nine women being offered further investigation or treatment for every woman diagnosed with an ovarian malignancy. This degree of specificity has not been achieved by any individual marker alone. To overcome this obstacle a number of strategies incorporating multiple markers have been proposed since the 1980s. The CA125 antigen was discovered in the early 1980s as part of an effort to develop a monoclonal antibody that could be used for immunotherapy of ovarian cancer.6 In this regard CA125 appeared to be relatively specific to epithelial ovarian cancers, being expressed in up to 80% of these cancers. Furthermore, it was found in the serum of women with ovarian cancer, although it can also be raised in a number of other cancers as well as some benign conditions. Although targeting CA125 has not proved to be therapeutically successful,7, 8 it was to prove a much more useful diagnostic/predictive biomarker. Its promise as a marker of early disease was shown in studies led by Zurawski in the 1980s. Serum CA125 levels were rarely elevated in healthy postmenopausal women9 but in up to 20% of cases, raised CA125 could be seen at least 2 years before ovarian cancer diagnosis.10 Alone, a raised CA125 has a high specificity but the likelihood of a woman having an occult cancer with a raised CA125 (Positive Predictive Value; PPV) remains low. PPV and sensitivity are determined in large part by the cut-offs used. Early studies by Klug et al. had determined that if a cut-off of 35 U/ml was used, then 82% of women with cancer, 6% of women with benign disease and 1% of normal women would have a raised CA125. Increasing the cut-off to 65 U/ml saw the percentage of ovarian cancer patients with a positive CA125 fall to 73% but those with benign disease and normal disease drop to 2% and 1%, respectively. To ensure that no women with benign disease would be included would necessitate the reference value to be set at 220 U/ml, which would mean that nearly 40% of women with ovarian cancers would be missed. Clearly serum CA125 measurement as a sole test was unlikely to offer the assurance of a low false-positive and false-negative rate needed to roll-out population screening. The 1950s saw the development of medical ultrasound and with it the ability to characterise gynaecological masses. Fittingly it was one of the first uses demonstrated by the seminal work of Ian Donald, Professor of Midwifery and a pioneer of diagnostic ultrasound.11 However, it would not be until the 1970s that this diagnostic tool was perceived to be a viable method for early detection in its own right, rather than a simple adjunct to the clinical examination of the symptomatic woman. Advances in ultrasound technology meant that by the mid-1970s lesions of 1 cm in diameter, which would normally be undetectable clinically, could be defined by ultrasound.12 Furthermore, various criteria to distinguish benign from malignant lesions with accuracies ranging from 70 to 90% were proposed.13-15 In 1982, Campbell et al.16 suggested that because ultrasound was able to define ovarian size accurately, ovarian volume could be a useful discriminator for potential ovarian disease in postmenopausal women. This strategy was studied in a large-scale prospective study of 5479 women followed over 8 years.17 Campbell et al. were able to identify 100% of the women that developed primary ovarian cancer within this time frame. The specificity of this procedure was 97.1% with a PPV of 1.5%, i.e. odds of a positive screen result representing a primary cancer of 1 in 67. Although Campbell et al. were able to demonstrate the utility of transabdominal ultrasound to identify women with asymptomatic ovarian masses, the inability to satisfactorily distinguish those with benign from malignant lesions underlies the high false-positive rates. The first approach proposed to reduce false positives was that only women with increasing ovarian volumes at a second scan, who had previously demonstrated either ovarian volumes >96th centile or abnormal morphology, e.g. cystic areas, should be offered surgery.18 Under this strategy, the PPV rose to 2.0%. The second approach came with the advent of transvaginal and Doppler ultrasound. Transvaginal ultrasound provided a higher resolution and so had the ability to detect abnormal morphology in smaller ovaries, while Doppler ultrasound provided the ability to visualise areas of neovascularisation more common in malignant lesions. Small studies showed promising results for the detection of women with ovarian pathology and the distinction of benign from malignant disease using transvaginal ultrasonography.19 The University of Kentucky Ovarian Cancer Screening Project was the first large-scale prospective trial to investigate whether transvaginal ultrasound assessment of ovarian volume could be used in ovarian cancer screening in pre- and postmenopausal women. In total, 90 women (1.4%) with persistent abnormalities were offered surgery to remove the affected ovary; six of whom were discovered to have a primary ovarian cancer. Overall, this strategy resulted in a sensitivity of 85.7%, specificity of 98.7% and 14 false positives for every woman with screen-detected ovarian cancer. Forgoing the possibility that the removal of benign ovarian masses with the potential for malignant transformation may lead to a resultant late fall in mortality from ovarian cancer, there remained a need for strategies to reduce the proportion of false positives. Jacobs et al.5 were the first to describe a multimodal approach for ovarian cancer screening in 1988. In their initial study, 1010 healthy postmenopausal women were offered serum CA125 measurement and a clinical pelvic examination. If either the clinical or biochemical assessment was abnormal women were offered a transabdominal ultrasound scan to assess ovarian volume. When CA125 measurements were considered individually, specificity was 97%, but when ultrasonography was added as a secondary test, specificity increased to 99.8%. This two-stage strategy was validated by a larger prevalence study in which 22 000 women were screened. With this approach equally high specificity and PPVs were achieved for a sensitivity of 79% at 1 year of follow up when a cut-off of 30 U/ml was used.20 The same cohort was then randomised to annual incidence screening or no further screening and followed up over 8 years. Initially, women recruited to this study in the late 1980s were offered transabdominal ultrasonography on entry. Subsequently, as evidence of the superiority of the transvaginal route in terms of image resolution with high acceptability among patients became available, follow-up scans were offered transvaginally. In contrast to the Kentucky Ovarian Cancer Screening Project, only postmenopausal women who had an abnormal CA125 were offered an ultrasound. In this preselected group of women, ultrasonographical assessment led to a much higher PPV of 21.3% and 23.75%, respectively, when ovarian volume and morphology were considered separately.1 Furthermore, there appeared to be minimal loss in sensitivity or specificity with this approach (sensitivity 89.5% and 100%; specificity 93.75% and 93.95% when considering ovarian volume and morphology, respectively). This was a breakthrough as it demonstrated that the target of nine healthy women being investigated or treated for every woman with cancer identified by screening was not only attainable but could be improved upon. This landmark finding stimulated a new wave of research to refine the detection of early ovarian cancers by screening. While early detection is fundamental, it is only relevant if it can translate into lives saved. Further analysis showed that screening identified proportionately more early-stage ovarian cancer (31.3% versus 10%), although the difference between the two groups was not statistically significant (P = 0.171).21 Nevertheless, the histological grade was lower and length of survival following randomisation was longer for women with ovarian cancer in the screened group (P = 0.02 and P = 0.01, respectively). Despite large numbers, the study was underpowered to detect a difference in mortality among screened and unscreened women, setting the scene for a still larger trial to answer this crucial question. Until 1998 there had been 25 prospective studies of ovarian cancer screening but no comparison with an unscreened population.22 Both ultrasound-based and multimodal screening strategies appeared to be able to identify asymptomatic women with ovarian cancer at early stages, but the clinical significance of this in terms of survival and mortality remained unknown. Furthermore ultrasound-based techniques appeared to offer higher sensitivity although at the cost of lower specificity, than multimodal screening. Notably, most of the women who received diagnostic surgery as a result, were found to have a benign gynaecological condition. Again the clinical impact of surgery in these circumstances is unclear. To answer this question UKCTOCS (UK Collaborative Trial of Ovarian Cancer Screening), the largest randomised control trial to date of an ovarian cancer screening strategy, involving over 200 000 women followed over 7 years was established. This study primarily aimed to ascertain whether screening, be it by a multimodal or imaging-only approach, could have an impact on ovarian cancer mortality. The trial would also assess the physical and psychological impact of screening, volunteer compliance and cost-effectiveness to inform decisions about the use of a national screening programme. As a result of further analysis of data from the pilot study, the multimodal arm had a new weapon in its armamentarium: the Risk of Ovarian Cancer algorithm (ROCA).23 This algorithm stratifies patients according to their CA125 trajectory rather than using a single threshold of 30 U/ml and is recalculated with every successive CA125 result. It was designed to improve sensitivity for the detection of early-stage cancers where earlier intervention could offer the most survival benefit and where CA125 is >30 U/ml in only 57% of cases, compared with 84% of late-stage cancers.24 Using the ROCA, a woman with a low baseline CA125 that suddenly rose would be identified as at risk of ovarian cancer, even if the absolute level remained modest, while a woman with a static but high CA125 would be deemed low risk. Retrospective testing of the algorithm on a Swedish cohort demonstrated high levels of sensitivity and specificity, which were subsequently verified by a prospective cohort of over 16 000 women between 1995 and 2000.25-27 More recently, ROCA was employed by Lu et al.28 in a large single-arm prospective study, which like UKCTOCS triaged women deemed high risk by ROCA to transvaginal ultrasound. Encouragingly, this protocol resulted in <1% of women being referred for surgery, giving it a specificity of 99.9% (95% confidence interval 99.7–100%) and a PPV of 40% (95% confidence interval of 12.2–78%).28 This equates to just over two women requiring surgery for every woman diagnosed with cancer. The pilot study published in BJOG,1 although primarily aimed at providing the platform for general population screening for postmenopausal women in whom 90% of ovarian cancer occurs, also provided the basis for studies in a second group of women with a family history of ovarian cancer. Some women are diagnosed with gene faults that increase their risk of ovarian and also breast cancer, notably defects in the BRCA1 and BRCA2 genes, but others have unidentified gene faults and/or are labelled as ‘at risk’ by virtue of their strong family history. Given the poor prognosis of ovarian cancer presenting symptomatically, women at high risk of ovarian cancer are currently offered risk-reducing prophylactic salpingo-oophorectomy (RRSO) to prevent incident ovarian cancers and reduce their risk of premenopausal breast cancer. When performed in young women, such surgery inevitably results in infertility and premature menopause and so screening may facilitate women safely delaying surgery until after their natural menopause. The United Kingdom Familial Ovarian Cancer Screening Study (UKFOCSS) recruited 3563 women with a 10% or greater lifetime risk of ovarian cancer between 2002 and 2008 to determine whether screening could be a viable alternative to RRSO. Reports from the first phase of the UKFOCSS study suggest that a negative screen can reliably reassure women (99.9% probability) that they will not develop ovarian cancer in the subsequent year. Screening was able to detect incident disease with a sensitivity of >80% and may reduce the proportion of stage IIIc disease.29 Phase II of the study drew on the successes of the UKCTOCS by stratifying patients according to ROCA using a computerised system, which would prompt investigators to repeat blood tests in women with intermediate risk and organise scans/referrals in women who were high risk. Additionally, screening intervals were reduced to 4 months to reduce false negatives.29 These amendments to the study design resulted in no interval cancers in this phase with no significant increase in the proportion of women referred.30 PPV in this phase was 13%. All women diagnosed with ovarian cancer as a result of screening went on to have optimal cytoreduction at surgery and survival was increased by 43 months.30 However, until such time as the efficacy of screening in this context has been fully evaluated, RRSO will remain the recommended intervention in women with a genetic predisposition to ovarian cancer. This research1 has also benefitted women with suspected ovarian cancer. The search for screening strategies which could be translated from women with clinically apparent cysts to those with occult disease fortuitously led to the development of the ‘Risk of Malignancy’ Index (RMI).31 This simple algorithm combines three factors that are independently predictive for an increased risk of ovarian cancer; menopausal status (scored 1or 3), serum CA125 and ultrasound features of the cyst (scored 0, 1 or 3). The product of these three criteria results in a score that can be used to provide a quantitative and reproducible assessment of likelihood that a symptomatic patient does indeed have an ovarian malignancy. Cut-offs can be useful tools for clinical decision making, e.g. referral to tertiary gynaecological oncology centres for surgery. To put this into context an RMI of 200 equates to a likelihood ratio of 42.1 of malignant disease and a likelihood ratio of 0.22 of benign disease; a 280-fold difference.32 Although the effect of the use of the RMI on the prognosis of a woman diagnosed with ovarian cancer has not been formally assessed, many studies have demonstrated the survival benefit of being treated by a specialist gynaecological oncologist.33 Where resources are limited, adjusting cut-offs can ensure only those at highest risk will be referred to tertiary centres. RMI remains the most widely used and evidenced stratification tool34 and in the UK, National Institute for Health and Care Excellence (NICE) guidance suggests the use of this tool to assess all women presenting to primary and secondary care with suspected ovarian cancer. The beauty of the RMI is its simplicity, which has helped it to stand the test of time. Accurate ultrasound assessment of an adnexal mass is fundamental to the success of the RMI. Substituting the subjective assessment by a skilled sonographer35-37 with a score defined by the presence or absence of certain key features facilitates reproducibility. However, the diagnostic test accuracy of the RMI appears lower in external validation studies than the original publications.38, 39 The International Ovarian Tumour Analysis (IOTA) collaboration was formed to encourage the development of better indices by standardising ultrasound reporting across research groups and prospectively collecting a database of features observed in cancerous and benign cysts from which new models could be developed and validated. Initial reports suggest that IOTA models (including simple rules and LR2) demonstrate better sensitivity for the detection of ovarian cancer with a lower false-positive rate than the RMI.38, 40 The challenge is to demonstrate the translation of highly sophisticated ultrasound protocols into ‘real-life’ clinical practice in different centres around the world.41 Women's symptoms have also gained attention as the search for strategies to reduce mortality from ovarian cancer through early detection has gathered pace. To this end, Goff et al.42 developed a symptom index using questionnaire data from women with and without ovarian cancer who were participating in a screening study. Over two-thirds of women had symptoms between 3 and 36 months before diagnosis. Using the Goff index the sensitivity for ovarian disease is 66.7% for a specificity of 90%, which has been validated in a number of external studies.43-45 This has formed the basis of the recommendation in the UK by NICE that symptomatic women are identified as at risk of ovarian cancer and offered further assessment. Much of the criticism surrounding general population screening for ovarian cancer is that low PPV would lead to the overtreatment of large numbers of women. A recent study has demonstrated that the prevalence of ovarian cancer among symptomatic women is ten times higher than that seen in asymptomatic women.46 Furthermore, when offered assessment with CA125 and transvaginal ultrasound at presentation, symptomatic women had lower tumour burdens, which were then more amenable to complete surgical resection, though unfortunately not earlier stage disease than those presenting to gynaecological oncologists by traditional routes. It is suggested that detecting women with lower tumour burdens may account for the increased survival of women diagnosed with stage 3 disease in the first randomised control trials of ovarian cancer screening.47 Data from UKCTOCS showed that the risk of a woman with an adnexal mass developing ovarian cancer within 3 years is 1 in 22.48 Therefore screening asymptomatic women for ovarian cancer will no doubt increase the numbers of women with incidental findings. On the one hand this has offered an unprecedented opportunity to conduct a number of nested cohort studies to assess the natural history of the conditions readily detectable by ultrasonography. But on the other hand, incidental findings cause and will resources in terms of and/or surgery. The of cancers which would not have clinically is a in The incidental finding of cancer is Although has been widely as a risk for cancer in women with postmenopausal the significance of a in the asymptomatic and so its potential use as a screening tool for cancer, is from UKCTOCS suggest that the cut-off for the detection of women with asymptomatic cancer or needs to be in the of as to the used cut-off of in symptomatic But even with an of the risk of cancer in an asymptomatic woman remains low at just Further studies will be needed to the value of early detection of cancer, the acceptability of screening and In the of the roll-out of an ovarian cancer screening that pelvic the need for such studies will more saw a to the cause of ovarian cancer screening. The and Ovarian Cancer Screening which randomised women to ovarian cancer screening, that there was no overall mortality benefit for annual with CA125 and transvaginal ultrasound rate confidence interval Furthermore there was no difference in survival or stage at between the screened and unscreened Additionally, this intervention resulted in more women to from to potential ovarian cancer. in study design between and UKCTOCS provide the that screening may offer for women with ovarian cancer. of the protocol that the study used a single cut-off CA125 level rather than the ROCA to identify women at had clinical in to with positive and follow up for many years after the screening were which time they to Initial data from UKCTOCS have been with of the ovarian cancers identified being as providing evidence for a in screened have suggested that these strategies identify and I cancers which may have clinically apparent within a II cancers representing up to of ovarian cancers diagnosed in symptomatic women, over two-thirds of the ovarian cancers identified by ultrasonography in UKCTOCS were or I screening appears to be more specific for II cancers but screening have had a number of interval cancers all of which were II cancers. These findings are in with the current that ovarian cancers may in the rather than from ovarian the women diagnosed with II cancers, had no adnexal pathology and were diagnosed at an It has been suggested that differences in the that CA125 was used to determine screen in the two trials for the of sensitivity of the trial to detect early Although of the algorithm to the data to any data from the Risk Ovarian Screening Study where women from the general population were using the ROCA followed by transvaginal ultrasound scan if they as high showed specificity and high PPV for the detection of ovarian It is that the UKCTOCS trial will answer whether CA125 as a test will be to translate to in ovarian cancer mortality. If UKCTOCS does not demonstrate survival or mortality benefit from general population ovarian cancer screening, it is to that any randomised trials will be to this given the to such a study The of any benefit from ovarian cancer screening in may be by poor sensitivity of screening to detect true early disease. disease suggests that cancers 90% of the of opportunity where intervention may prognosis at and may have been present for between 8 and years before detectable by current CA125 current screening strategies may detect disease at a that is late to disease The prospective of serum in asymptomatic women may result in the of more relevant of early disease than have been suggested by the comparison of from women with clinically detectable cancers with studies aimed at identifying an from the serum of women who subsequently developed ovarian cancer in UKCTOCS are The years have seen a in the of ovarian cancer, its natural history and the of mortality. Despite in and there has been in the prognosis for women with disease. with the ability of ovarian cancer screening strategies to the of disease to earlier stages, which may translate to better clinical and reduced mortality. screening and ultrasound-based have been shown to be and specific for detecting ovarian cancer but whether these in will provide lead time to the of the disease remains to be success will on the of the multiple and to ensure that research will for better that can be translated into clinical It is that and to the and on the by the high compliance rates seen in the screening studies to work mortality from ovarian cancer. have no to and this was to this is a and is an at the Institute of Cancer University of

The prostate component of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial was undertaken to determine whether there is a reduction in prostate cancer mortality from screening using serum prostate-specific antigen (PSA) testing and digital rectal examination (DRE). Mortality after 7-10 years of follow-up has been reported previously. We report extended follow-up to 13 years after the trial. A total of 76 685 men, aged 55-74 years, were enrolled at 10 screening centers between November 1993 and July 2001 and randomly assigned to the intervention (organized screening of annual PSA testing for 6 years and annual DRE for 4 years; 38 340 men) and control (usual care, which sometimes included opportunistic screening; 38 345 men) arms. Screening was completed in October 2006. All incident prostate cancers and deaths from prostate cancer through 13 years of follow-up or through December 31, 2009, were ascertained. Relative risks (RRs) were estimated as the ratio of observed rates in the intervention and control arms, and 95% confidence intervals (CIs) were calculated assuming a Poisson distribution for the number of events. Poisson regression modeling was used to examine the interactions with respect to prostate cancer mortality between trial arm and age, comorbidity status, and pretrial PSA testing. All statistical tests were two-sided. Approximately 92% of the study participants were followed to 10 years and 57% to 13 years. At 13 years, 4250 participants had been diagnosed with prostate cancer in the intervention arm compared with 3815 in the control arm. Cumulative incidence rates for prostate cancer in the intervention and control arms were 108.4 and 97.1 per 10 000 person-years, respectively, resulting in a relative increase of 12% in the intervention arm (RR = 1.12, 95% CI = 1.07 to 1.17). After 13 years of follow-up, the cumulative mortality rates from prostate cancer in the intervention and control arms were 3.7 and 3.4 deaths per 10 000 person-years, respectively, resulting in a non-statistically significant difference between the two arms (RR = 1.09, 95% CI = 0.87 to 1.36). No statistically significant interactions with respect to prostate cancer mortality were observed between trial arm and age (P(interaction) = .81), pretrial PSA testing (P(interaction) = .52), and comorbidity (P(interaction) = .68). After 13 years of follow-up, there was no evidence of a mortality benefit for organized annual screening in the PLCO trial compared with opportunistic screening, which forms part of usual care, and there was no apparent interaction with age, baseline comorbidity, or pretrial PSA testing.

Gleason 6 Prostate Cancer: Translating Biology into Population Health

Purpose: Androgenic actions are thought to underlie the development of both male pattern baldness and prostate cancer. However, results from previous studies of the relationship between these two phenotypes have been inconsistent. Therefore, we investigated the association of male pattern baldness at age 45 years with overall and aggressive prostate cancer risk in a large, prospective cohort–the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Methods: We included 39,070 men, from the usual care and screening arms of the trial, who were cancer-free at start of follow-up. Hazard ratios (HRs) and 95% confidence intervals (95%CIs) were estimated using Cox proportional hazards regression models with age as the underlying time-scale. Results: During follow-up (median=2.78 years), 1,138 incident prostate cancer cases were diagnosed, 578 of which were aggressive (Gleason score ≥ 7, stage ≥ III, or fatal). Frontal + vertex balding at age 45 years, compared with no balding, was not associated with overall prostate cancer risk (HR=1.19, 95%CI: 0.98, 1.45), but was associated with increased risk of aggressive prostate cancer (HR=1.42, 95%CI: 1.10, 1.83). Adjustment for covariates did not significantly alter these estimates. Other classes of male pattern baldness were not associated with the overall or aggressive prostate cancer. Conclusion: Our analysis indicates that frontal + vertex balding at age 45 years increases the risk for future development of aggressive prostate cancer, and supports the possibility of overlapping pathogenesis. Citation Format: Cindy Ke Zhou, Ruth M. Pfeiffer, Sean D. Cleary, Heather J. Hoffman, Paul H. Levine, Lisa W. Chu, Ann W. Hsing, Michael B. Cook. Male pattern baldness increases the risk of aggressive prostate cancer: A prospective analysis of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3260. doi:10.1158/1538-7445.AM2014-3260

Pathologic findings following false-positive screening tests for ovarian cancer in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial

INTERVAL AFTER PROSTATE SPECIFIC ANTIGEN TESTING AND SUBSEQUENT RISK OF INCURABLE PROSTATE CANCER

Background: Aspirin may slow progression of cancer through the inhibition of platelet aggregation as experimental research has shown that platelets play a role in cancer cell dissemination and metastatic growth. While a 10% reduction in prostate cancer risk has been associated with regular aspirin use, only a few epidemiological studies have evaluated the association between aspirin use and prostate cancer mortality among prostate cancer cases. Methods: Participants included 3857 men, age 55-74, who were diagnosed with prostate cancer in the control arm of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial between 1993-2009. The frequency of aspirin use one year prior to baseline (1993-2001) was ascertained through a self-administered questionnaire. Death due to prostate cancer was ascertained by annual questionnaire and linkage to the National Death Index. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using COX proportional hazards models using age as the time metric and adjusting for clinical stage, Gleason score, primary treatment, smoking status, body mass index, race, history of heart attack, and history of stroke. Results: After a median of 5 years of follow-up, 136 cases died of prostate cancer. No significant association with prostate cancer mortality was observed between cases who reported pre-diagnostic daily or more than daily aspirin use compared to cases who reported no aspirin use (HR=0.77; 95% CI 0.48-1.25). Among those diagnosed with clinically advanced stage (III, IV) disease, an inverse association with prostate cancer mortality was observed for cases who reported daily or more than daily aspirin use (HR=0.37; 95% CI 0.15-0.92), while no association for daily or more than daily users was seen among those with clinically localized prostate cancer (HR=0.86; 95% CI 0.47-1.58). Conclusion: Pre-diagnostic use of aspirin was not significantly associated with prostate cancer mortality based on cases diagnosed in the PLCO control arm, although a reduction in risk of prostate cancer mortality was observed among those with advanced stage disease in a subgroup analysis. Citation Format: Sarah E. Daugherty, Ruth Pfeiffer, Armen Ghazarian, Grant Izmirlian, Phil Prorok, Katherine McGlynn. Frequency of aspirin use and prostate cancer mortality among prostate cancer cases in the control arm of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr LB-15. doi:10.1158/1538-7445.AM2013-LB-15

Biomarkers for diagnosis: looking for change.