Government & Nonprofit

A national cancer institute sponsored screening trial for prostatic, lung, colorectal, and ovarian cancers

A national cancer institute sponsored screening trial for prostatic, lung, colorectal, and ovarian cancers
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  589 A National Cancer Institute Sponsored Screening Trial for Prostatic, Lung, Colorectal, and Ovarian Cancers z arnett S. Kramer M.D. M.P.H. Iohn Gohagan Ph.D. Philip zy . Prorok Ph.D. and Charles Smart M.D. zyxwvu Background. Ovarian cancer is the fifth most com- mon cause of cancer-related death in American women. The median age at diagnosis is about 62 years; incidence rises rapidly after age 60. Pelvic examination has been the primary method for detection of ovarian carcinoma. It is insensitive for the detection of early disease, how- ever: most women present with disease beyond the pelvis (Stages I11 and IV) and are not curable with existing tech- niques. Two new technologies may be useful as screening tools for earlier detection of ovarian cancer. CA 125 is an antigenic determinant expressed on an ovarian cancer cell line. Transvaginal ultrasound (TVUS) mages the ova- ries from within the vagina and can be performed by a technician in about 10 minutes. In small preoperative studies of women with ovarian masses, serum CA 125 levels have been elevated (typically above 35 U/ml) in Over two-thirds of cases and in up to 50 of Stage I cases. The test is not absolutely specific: elevations have been reported with pregnancy, endometriosis, menstruation, benign ovarian tumors, and with cancers of the breast, colon, pancreas, lung, stomach, and liver. Nevertheless, the specificity of CA 125 in postmenopausal women has been reported at about 95 or more. TVUS provides higher resolving power for ovarian abnormalities than transabdominal ultrasound or physical examination; however, experience with it is limited. CA 125 and TVUS may be complementary. Conclusions. For these reasons, the National Cancer Institute is planning a randomized trial of all Presented at the Perspectives on Ovarian Cancer in Older-Aged women: Current Knowledge and Recommendations for Research working Conference, Bethesda, Maryland, November zyxwvu 0-21, 199 . From the National Cancer Institute, Division of Cancer Preven- tion and Control, Early Detection and Community Oncology Pro- gram and Biometry Branch, National Institutes of Health, Bethesda, Maryland. The authors thank Dr. Ken Chu for his comments during the Planning stages of the study and for his review of this manuscript. Address for reprints: Bamett S. Kramer, M.D., M.P.H., National Cancer Institute, EPN Building, Room 300, 9000 Rockville Pike, Be- thesda, MD 20892. Accepted for publication July 20, 1992. three tests versus routine medical care in women of ages 60-74 years. This is part of a larger trial to determine the efficacy of screening for lung, colorectal, and ovarian cancers in women, and for lung, colorectal, and prostatic cancers in men. Seventy-four thousand women will be randomized in the study. Cancer 1993; 71:589-93. Key words: ovarian cancer, screening, randomized trial, CA 125, transvaginal ultrasound. Several criteria should be met in recommending routine screening on a mass scale in asymptomatic individuals for a given disease.',' First, the disease should have seri- ous consequences for the popula tion-that is, there should be a significant burden of disease at the public health level. Some have added the criterion that the disease should have a reasonably prevalent preclinical phase in the population; little or nothing is known on this matter with regard to many neoplasms, including ovarian cancer.' Second, there should be a recognizable preclinical early stage of the disease. Finally, treatment at this stage should reduce mortality. In the case of ovarian cancer, the first criterion is clearly met relative to most cancers (Fig. 1). Ovarian cancer ranks as the seventh most common cancer in incidence and the fifth most common cause of cancer- related mortality in American women. The 5-year sur- vival rate has been relatively unchanged at about 35% for nearly three decade^.^ The burden of disease is par- ticularly heavy in the elderl~.~ he yearly incidence per 100,000 women is about 14 at age 50-54, rising to about 25 at age 60-64, and about 35 for ages 70-74. Likewise, the mortality rates per 100,000 for these three age groups are about 7, 15, and 24, respectively. The second criterion is likely to be true, since currently exist- ing tests can detect ovarian cancer before the onset of symptoms in at least a proportion of cases. The third criterion is the most problematic. Although duration of survival after diagnosis is longer for early stages of dis-  590 zyxwvusrqpo ANCER Supplement january zyxwvus 5, 2993, Volume 71, No. 2 Age (years) Figure 1 Age-specific ovarian cancer incidence and mortality in the United States from 1984-1988 (SEER data). zyxwvutsr ease than for advanced stages, measurement of survival duration in screening programs is subject to lead time bias (early diagnosis may prolong the duration of aware- ness of the disease without changing the date of death) and to length bias (asymptomatic cancers diagnosed at screening are often biologically less aggressive than cancers diagnosed because of symptom^).^ There are to date no data to show that treatment of the early stages of ovarian cancer detected at screening would lower the cause-specific mortality rate.6 Although there are other forms of evidence, the most definitive proof of the effect on cause-specific mor- tality is a prospective randomized trial of the screening tests compared to standard medical practice. To that end, the National Cancer Institute (NCI) plans a large multicenter randomized investigation of screening for ovarian cancer. The trial is part of a study of screening for cancers at four organ sites: Prostate, Lung, Colon- rectum, and Ovary (termed the PLCO Trial). These four sites were chosen for similar reasons: 1) each cames a large burden of disease for the American public in terms of mortality, 2) there are available tests capable of de- tecting disease before the onset of symptoms, which usually herald advanced disease, 3) curative potential of available treatments is restricted to early stages of disease, and 4) efficacy of screening in terms of mortal- ity reduction is unknown. In the case of ovarian cancer, the cure rate for Stage I is 80-90 .’** In addition, some of the current technologies for early detection are very expensive, and they are coming into wide use without a definitive test of efficacy. These include endoscopy for colorectal cancer and transvaginal sonography for ovarian cancer. Study Design A key feature of the PLCO Trial is the incorporation of screening efforts for multiple organ sites. This plan is based on the philosophy that if screening is to be effec- tive at the population level, there must be an efficiency of scale and convenience, Accordingly, we envision early detection centers that will screen each subject in the study group for three organ sites: prostate, lung, colon-rectum in men; and lung, colon-rectum, and ovary in women. It is intended that screening will be performed by technologists or physicians during a sin- gle visit to the study centers and will be completed within 2 hours. The screening tests for each organ site are listed in Table 1 A total of 148,000 subjects 60-74 years of age, half of them women, will be randomly assigned to a screened group versus a control group receiving standard medical care. With the exception of sigmoidoscopy, which will be done at baseline and then at year 3, all screening tests will be performed annually for four examinations in the screened group. All sub- jects will be followed for a minimum of 10 years after entry. We estimate the total costs of the trial to be about $60 million expressed in 1990 dollars, or about $88 mil- lion dollars over 16 years if the inflation rate were to remain stable at zyx 9b per annum. Based on the risk of death from ovarian cancer in women aged 60-74 at study entry, the power to detect a 35 decrease in cause-specific mortality in a sample of this size would be 88 . Calculations were made assum- ing a 10-year follow-up using a one-sided alpha level (Type 1 error) of 0.05. If the benefit of screening on mortality were only 25 , a total of 168,000 women would have to be randomized to achieve a power of 90%. If it became apparent during the study that it were necessary to do so, the female population base of this trial could be supplemented, or a meta-analysis of data from this trial and other relevant studies could be done to increase power. Screening Tests for Ovarian Cancer in the PLCO Trial zyxwv Physical Examination The most commonly used method for the detection of ovarian cancer, the bimanual pelvic examination, is in- adequate because of its inability to detect early disease. Despite its widespread use, about two-thirds of patients have advanced disease (Stages 111 or IV) at initial diag- nosis.’ Surprisingly little is known about the accuracy of the pelvic examination in detecting ovarian cancer. The sensitivity and specificity in detection of adnexal masses has not been well established. A recent study  Ovarian Cancer Screening Trial/Kruner et zyxwvusrq l. 591 Table zyxwvuts . Screening Tests in the PLCO Trial Screening tests zyxwvutsr   prostate Digital rectal examination Lung Chest x-ray Colon-rectum Prostatic specific antigen 60-an flexible sigmoidoscopy initial, then Pelvic examination Transvaginal ultrasound at 3 years) Ovary CA 125 zyxwvutsrqp used a synthetic model of the female pelvis that could be fitted with adnexal masses of various sizes up to 6 zyxw   4 cm.*O Using medical residents as the examining physicians, the sensitivity and specificity were 67 and O , respectively. Sensitivity increased with size of the mass and with medical school experience doing pel- vic examinations. It is possible that the accuracy of pel- vic examination in actual practice would be worse than in this study, since the model simulated a thin woman not necessarily representative of the general popda- tion, and patient discomfort that might interfere with the clinical situation was not an issue. In any case, the above information suggests that pelvic examination is much too insensitive to serve as the sole screening test to detect curable stages of ovarian carcinoma. CA 125 Ovarian cancer cell lines have been used as immuno- gens to identify tumor-associated antigens. One such antigenic glycoprotein, CA 125, has received consider- able attention as a potential screening too1.6,1'*12 erum levels of CA 125 are elevated above 35 U/ml in about BOY0 of patients with ovarian cancer. zyxwv s with most tu- mor markers released into the peripheral circulation, however, levels are correlated with tumor burden. In one study, stored sera from a cohort of healthy volun- teers were retrieved after 105 women of the cohort had been diagnosed with ovarian cancer. Only 12 of 24 am- ples from women with surgical Stage I disease (that is, confined to ovarian tissue) had CA 125 evels above 35 u/ml, with a lead time of 18 months.13 A review of 14 studies in the literature reveals that a preoperative evel of CA 125 was elevated above 35 in 45 of 92 Stage cases'' (Table 2). Hence, the sensitivity of the assay to detect disease confined to the ovaries appears to be about 50 at maximum. More recently, a prospective study in Sweden has been reported in which 5550 women aged 40 years and above were followed with annual CA 125 determina- tions. Sweden has among the highest incidence rates of ovarian cancer in the world. Those women with ele- vated levels were followed with pelvic examination and transabdominal ultrasound (175 women). Of these, six ultimately were diagnosed with ovarian cancer: two in Stage 1, two in Stage 11, and two in Stage 111. This repre- sents a positive predictive value of 3.4 for an elevated CA 125. In addition to sensitivity, the specificity of an assay and the underlying prevalence of disease in the popula- tion being screened play a major role in the positive predictive value of a screening test. Elevations of CA 125 have been reported in pregnancy, endometriosis, menstruation, benign ovarian tumors, and in cancers of the breast, colorectum, pancreas, lung, stomach, and liver. The specificity of the test in asymptomatic women is about 95 . Because of the spectrum of diseases that can cause false elevations of CA 125, he specificity of the test probably is better in postmenopausal than in premenopausal women. Nevertheless, even presuming that the specificity is 99 and that 80 of women with ovarian cancer have an elevation of CA 125, he calcu- lated positive predictive value would be only 2.3 , given the low prevalence of the disease in asymptom- atic American women. Therefore, the assay fails as a stand-alone screening test in such women. It would fare much better in a population known to have adnexal masses; but, strictly speaking, it would then be serving as a diagnostic test. Transvaginal Ultrasound Although transabdominal ultrasonography has been studied as a potential screening tool for ovarian cancer, transvaginal ultrasonography (TVUS) offers improved image quality.6 This is because the ultrasound probe can be positioned closer to the area of interest, allowing higher-frequency sound waves. In addition, transabdo- minal ultrasonography requires significant patient prep- aration, including filling the bladder before the exarni- nation. Transvaginal ultrasonography is performed after the woman has emptied her bladder. Criteria for concern on a TVUS are based on mor- Table 2. Elevated Preoperative CA 125 Levels Greater Than 35 U/ml) in Patients With Ovarian Cancer by FIGO Stage FIGO stage I I Ill lv Total Positive/total 45/92 47/53 174/191 70/74 572/678 Percent 49 89 91 95 84 Adapted from Jacobs Bast R The CA-125 tumour-associated anhgen. a review of the literature. Hum Reprod 1989, 41-12.  592 CANCER Supplement junuury 15, 2993, Volume 71, No. 2 zyxv phologic pattern as well as Uniform hypoecho- genic or entirely cystic patterns are of little concern, as opposed to complex or solid patterns. The size criteria for abnormality differ between premenopausal and postmenopausal women. The upper limit of normal size is about 18 zyxwvut n3 n an examination obtained a week after the subject's menses. In postmenopausal women, the upper limit of normal is about 8 cm3. Because of cyclic changes in the size of the ovaries during the men- strual cycle, a single abnormal TVUS by size criteria is less specific in premenopausal women. In a prospective screening study of TVUS in 1000 asymptomatic women older than 40 years of age, 39 of 422 (9.2 ) premenopausal women had an abnormal initial TVUS by size or morphologic criteria.' Only 16 (3.89'0) had persisting abnormalities on a repeat exami- nation 1 week after the next menstrual period. All 16 were offered exploratory laparotomy, and 13 con- sented. Although all 13 had abnormal findings at sur- gery and measured ovarian dimensions were correlated closely with those found at TVUS, none of the 13 had malignancy. Of 578 postmenopausal women screened, 15 had abnormal sonograms (2.6 ). All 15 underwent exploratory laparotomy, and one was found to have adenocarcinoma involving the ovary. This was a meta- static lesion from a primary colon cancer that had been in remission for two years. All findings in the other 14 patients were benign, including an endometrioma and a cervical myoma. Of interest, all 15 subjects with en- larged ovaries on TVUS also underwent bimanual ex- amination under anesthesia before surgery; seven had palpable abnormalities. Hence, TVUS is likely to be much more sensitive than pelvic examination for screening purposes, but there is no accurate informa- tion on their relative specificities. zyxwvu ge Range of the Subjects to be Screened As noted earlier, the subjects to be screened will be of ages 60-74 years at study entry. Since the median age of ovarian cancer at diagnosis is 62 years, more than half of the at-risk population would be encompassed by such a strategy. More importantly in a mass screening program, the screening test performs much better in a population with a high prevalence of the targeted dis- ease than in a lower-risk population. Figure 1 shows how rapidly the risk for development of ovarian cancer rises with age. The figure shows incidence by age; the prevalence of ovarian cancer in the population is un- known, but can be roughly calculated as the product of the incidence and the duration of disease. For example, if the incidence of early-stage curable ovarian cancer is 10 per 100,000 per year and the duration of detectable disease at a curable stage is 5 years, the prevalence of disease that could be diagnosed at a curable stage would be about 50 per 100,000. If one assumes the sensitivity of a screening procedure to pick up as much as 50 of ovarian cancers at a curable stage (an opti- mistic estimate), with a specificity of 99 (also optimis- tic), in a population with a prevalence rate of 50 per 100,000, one can use Bayes theorem to calculate the positive predictive value (PPV) of a screening proce- dure. In this case, the PPV would be 2.4 . Using the same screening procedure in a population with half the prevalence of early-stage disease (25 per 100,000) would cut the PPV of the test by nearly zy 0 to about 1.2 , assuming the same sensitivity and specificity. Ifa laparotomy were to be performed on each person with a positive screen, this would roughly translate into twice the number of laparotomies to find and resect each early-stage cancer. Also, the cost per year of life saved would go up dramatically, since twice the num- ber of screens would be necessary to detect an equiva- lent number of early cancers. Nevertheless, once the sensitivity and specificity of the screening tests are determined from a well designed trial with adequate long-term follow-up, the PPV and approximate cost effectiveness could be calculated for age ranges of women not included in the trial (for exam- ple, ages 50-59 years). A caveat for this calculation is that the sensitivity and specificity are the same in each age range. The situation becomes more complicated in women under the age of 50 years. The specificity of tests such as CA 125, TVUS, nd physical examination are likely to be lower in premenopausal than in postmenopausal women. This is because of the cyclic changes in ovarian size with phase of the menstrual cycle and an additional spectrum of conditions found in premenopausal women which may cause false elevations of the CA 125 level (see above). Another issue is whether screening women for ovarian cancer up to age 74 is likely to benefit them, given life expectancy and competing causes of death. In point of fact, average life expectancy for such women is substantial. As shown in Table 3, average life expec- tancy for a woman who has reached age 75 is 12 addi- Table 3. Age Versus Expected Survival in Elderly US. Residents Age (yr) Expected average survival (yr) 65 18 75 12 85 7  Ovarian Cancer Screening Trial/Kramer et al. 593 tional years. A woman who reaches the age of 85 has an average additional life expectancy of about 7 years. These may actually be underestimates for a screening program, since seriously disabled'women who have co- morbid disease limiting their life expectancy would be unlikely targets in mass screening efforts. Nevertheless, the majority of women at these ages are in relatively good health. zyxwvutsr Summary and Conclusions It should be clear from the discussion above that neither bimanual pelvic examination, CA 125, nor TVUS is likely to suffice as a stand-alone screening test for ovar- ian cancer in American women. The intent of this study is to use all three tests to see if an appropriate algorithm can be determined for mass screening efforts. To some extent, the tests may be complementary. CA 125 is likely to be tumor burden-driven, but may be able to detect at least some of the lesions before they have me- tastasized. It seem to be positive in no more than 50 of the early clinical stages of ovarian cancer. Ultrasound and physical examination may be useful for picking up CA 125-negative tumors and the TVUS may be particu- larly helpful in defining the pattern of ovarian abnor- mality. An algorithm that incorporates two or three tests may prove the most functional. For example, a modest elevation of the CA 125 is unlikely to be suffi- cient impetus for a laparotomy in the absence of an abnormal physical examination or TVUS. Nevertheless, a modestly elevated CA 125 that doubles over several months may warrant an invasive procedure such as lap- aroscopy even if the physical examination and TVUS are unrevealing. The trial will also allow for the first time an accurate determination of the sensitivity and specificity of the three screening modalities in asymp- tomatic postmenopausal women. These in turn will permit a calculation of the positive predictive value of the tests in postmenopausal women in the United States. The study also will help determine whether the new technologies that are creeping into the manage- ment of women without a definitive test of efficacy should play a role in mass screening efforts designed to decrease mortality from ovarian cancer. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Hulka 8. Cancer screening: degrees of proof and practical appli- cation. zyxwv ancer 1988; 62:1776-89. Wilson J Jungner G. Principles and practice of screening for disease. Public Health Papers 34. Geneva: World Health Organi- zation, 1968. Siverberg E, Boring C, Squires T. Cancer statistics 1990. Cancer Miller BA, zyxw ies LA, Hankey BF, et al. Cancer statistics review 1973-1989. U.S. Department of Health and Human Services. NIH publication no. 92-2789, 1992:Ovary (XX. 1-7). Feinleib M, Zelen M. Some pitfalls in the evaluation of screening programs. Arch Environ Health 1969; 19:412-5. Bourne T, Reynolds D, Campbell S. Ovarian cancer screening. Eur zyxwvu   Cancer 1991; 27:655-9. Dembo A, Davy M, Stenwig A, Berle E, Bush R, Kjorstad K. Prognostic factors in patients with stage 1 epithelial ovarian cancer. Obstet Gynecol 1990; 75:263-73. Young R, Walton L, Ellenberg S, Homesley G, Wolbanks z   Decker D, et al. Adjuvant therapy in stage I and stage I1 epithe- lial ovarian cancer. N Engl Med 1990; 3221021-7. van Nagell JJ Higgins R, Donaldson E, Gallion H, Powell D, Pavlik E, et al. Transvaginal sonography as a screening method for ovarian cancer: a report of the first 1000 cases screened. Cancer 1990; 65573-7. Ferris AM, Schapira M, Young M. Accuracy of pelvic examina- tion. Ann Ititern zyxwv ed 1991; 114:522. Jacobs I Bast R. The CA-125 tumour-associated antigen: a re- view of the literature. Hum Reprod 1989; 4:l-12. Bates S. Clinical applications of serum tumor markers. Ann ln fern Med 1991; 115:623-38. Zurawski V, Orjaseter H nerson A, Jellum E. Elevated serum CA-125 levels prior to diagnosis of ovarian neoplasia: relevance to early detection of ovarian cancer. Int Cuncer 1988; 42:677- 80. Einhorn N, Sjovall K, Schoenfeld D, Eklund G, Knapp RB Jr., Zurawski V. Early detection of ovarian cancer using the CA-125 radioimmunoassay RIA). Proc Am Soc Clin Oncol 1990; 9:157. Granberg S Wikland M, Jansson I. Macroscopic characterization of ovarian tumors and the relation to the histological diagnosis: criteria to be used for ultrasound evaluation. Gynecol Oncol Brody J Cassell C. Normal human aging theory, demography and epidemiology. In: Cassell CK, Butler RN, editors. Geriatric medicine: principles and practice. New York Springer-Verlag, 1990; 4O:lO-23. 1989; 35:139-44. 1990 16-27.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks