Abstract B53: Predicting efficacy of chemopreventive agents in animal tumor assays by statistical modeling

NCI's Division of Cancer Prevention has over 25 years entered some 800 agents into a program designed to test cancer preventive efficacy. Early in the testing pathway are 2 sequential critical steps: (1) in vitro/in vivo morphologic assays and, (2) for agents successful in step 1, cancer-preventive testing (measured in terms of tumor incidence and multiplicity reduction) in animal tumor assays. The ultimate intention is to follow successful testing in animals with entry into clinical trial testing in humans. To facilitate the optimal selection of appropriate morphologic tests for specific disease-site animal tumor models, we undertook the current project. We evaluated the success of our strategy by determining how accurately the earlier stage (morphologic) assays predict efficacy in the later-stage (animal tumor) assays. Focusing on 210 agents that were tested in both morphologic and animal tumor assays, we carried out statistical modeling of how well the 6 most commonly used morphologic assays predicted drug efficacy in animal tumor models. Using multimodel inference, three statistical models were generated to evaluate the ability of these 6 morphologic assays to predict tumor outcomes in 3 different sets of animal tumor assays: (1) all tumor types, (2) breast cancer only, and (3) colon cancer only. Using this statistical modeling approach, each morphologic assay was assigned a value reflecting how strongly it predicted outcomes in each of the 3 different sets of animal tumor assays. The goal is to use predictive models such as these to guide our future decision-making with respect to selection of preventive agents as well as morphologic and animal tumor assays. In this manner, we hope to improve the efficiency of our overall approach to chemoprevention agent development.

Citation Format: Barbara K. Dunn, Vernon E. Steele, Richard M. Fagerstrom, Carol F. Topp, David Ransohoff, Leslie G. Ford, Barnett S. Kramer. Predicting efficacy of chemopreventive agents in animal tumor assays by statistical modeling. [abstract]. In: Proceedings of the Thirteenth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2014 Sep 27-Oct 1; New Orleans, LA. Philadelphia (PA): AACR; Can Prev Res 2015;8(10 Suppl): Abstract nr B53.

Predictive Value Tools as an Aid in Chemopreventive Agent Development

BACKGROUND

Over 25 years, the National Cancer Institute's Division of Cancer Prevention has entered some 800 agents into a chemopreventive agent testing program. Two critical steps involve: 1) in vitro/in vivo morphologic assays and 2) animal tumor assays (incidence/multiplicity reduction). We sought to determine how accurately the earlier-stage (morphologic) assays predict efficacy in the later-stage (animal tumor) assays.

METHODS

Focusing on 210 agents tested in both morphologic and animal tumor assays, we carried out statistical modeling of how well the six most commonly used morphologic assays predicted drug efficacy in animal tumor assays. Using multimodel inference, three statistical models were generated to evaluate the ability of these six morphologic assays to predict tumor outcomes in three different sets of animal tumor assays: 1) all tumor types, 2) mammary cancer only, and 3) colon cancer only. Using this statistical modeling approach, each morphologic assay was assigned a value reflecting how strongly it predicted outcomes in each of the three different sets of animal tumor assays.

RESULTS

We demonstrated differences in the predictive value of specific morphologic assays for positive animal tumor assay results. Some of the morphologic assays were strongly predictive of meaningful positive efficacy outcomes in animal tumor assays representing specific cancer types, particularly the aberrant crypt focus (ACF) assay for colon cancer. Moreover, less strongly predictive assays can be combined and sequenced, resulting in enhanced composite predictive ability.

CONCLUSIONS

Predictive models such as these could be used to guide selection of preventive agents as well as morphologic and animal tumor assays, thereby improving the efficiency of our approach to chemopreventive agent development.

Abstract 2813: Chemoprevention of mammary cancer: Modeling predictive values of short-term morphologic assays for efficacy in animal tumor assays

Background: Understanding the ability of chemopreventive agent efficacy in morphologic (in vitro/in vivo) assays to predict efficacy in animal (mouse and rat) in vivo tumor assays has not been well studied. The Chemopreventive Agent Development Research Group (CADRG) in the U.S. NCI's Division of Cancer Prevention has, over 25 years, tested approximately 800 agents for potential chemopreventive activity. Of these agents, a subset of 146 that were tested in both morphologic and mammary gland tumor assays constitutes the focus of our current project. Our goal is to gain deeper insight into the relevant predictive values.

Materials and Methods: Two critical steps are involved in the early stages of the CADRG testing pathway: (1) in vitro/in vivo morphologic assays and, for agents successful in these, (2) testing for tumor prevention (measured in terms of tumor incidence and multiplicity reduction) in animal tumor assays. Ultimately, our goal is to test agents that successfully decrease tumor incidence and multiplicity in animal tumor assays in humans. In the project presented here, we evaluated the predictive values of earlier-stage (morphologic) assays for efficacy in later-stage (animal tumor, specifically mammary tumor) assays. We used statistical modeling to determine how well the six most commonly used morphologic assays predicted efficacy of the 146 tested agents in mouse and rat mammary gland tumor assays. For this purpose multimodel inference was applied to ordinal logistic regression.

Results: We evaluated the ability of the six morphologic assays to predict tumor outcomes in the mouse and rat mammary gland cancer assays. In our statistical modeling approach, each morphologic assay was assigned a value describing how strongly it predicted outcomes in the mammary gland tumor assays. Specific morphologic assays (mouse mammary organ culture/MMOC and human foreskin epithelial cell/HFE morphologic assays) in combination yielded predictive values that are reasonably suggestive of the efficacy demonstrated by chemopreventive agents in the mouse and rat mammary gland tumor assays.

Conclusions: Predictive models such as these should prove useful in guiding future decision-making regarding agent selection and morphologic and animal tumor assay use. Our current strategy is to gain insight into the Predictive Value approach by: (1) identifying the mammary gland tumor assays that best reflect anti-tumor efficacy in animals; (2) teasing apart those classes of agents that exhibit the highest predictive values overall; and (3) examining those classes of agents that exhibit the highest efficacy in specific mammary gland tumor assays.

Citation Format: Barbara K. Dunn, Vernon E. Steele, Richard M. Fagerstrom, Carol F. Topp, Barnett S. Kramer. Chemoprevention of mammary cancer: Modeling predictive values of short-term morphologic assays for efficacy in animal tumor assays. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2813. doi:10.1158/1538-7445.AM2015-2813

Abstract P5-11-06: Understanding predictive values of short-term morphologic assays of cancer chemoprevention for efficacy in animal mammary gland tumor assays

Background: The predictive value of chemopreventive agent efficacy in morphologic (in vitro/in vivo) assays for efficacy in animal (mouse and rat) in vivo tumor assays is not well characterized. Over a 25-year period, the Chemopreventive Agent Development Research Group in the U.S. NCI’s Division of Cancer Prevention has tested approximately 800 agents for potential chemopreventive activity. The current project focuses on a subset of 146 that were tested in both morphologic and mammary gland tumor assays in order to gain a deeper understanding of the relevant predictive value.

Materials and Methods: The early stages of the testing pathway involve two critical steps: (1) in vitro/in vivo morphologic assays and, for agents successful in these, (2) testing for tumor prevention (measured in terms of tumor incidence and multiplicity reduction) in animal tumor assays. The ultimate goal is to test agents that successfully decrease tumor incidence and multiplicity in animal tumor assays in humans. In the current project we evaluated the predictive values of the earlier-stage (morphologic) assays for efficacy in the later-stage (animal tumor, specifically mammary tumor) assays. Statistical modeling to determine how well the six most commonly used morphologic assays predicted efficacy of the 146 tested agents in mouse and rat mammary gland tumor assays was carried out by multimodel inference applied to ordinal logistic regression.

Results: The ability of these six morphologic assays to predict tumor outcomes was evaluated in the mouse and rat mammary gland cancer assays. Based on this statistical modeling, each morphologic assay was assigned a value describing how strongly it predicted outcomes in the mammary gland tumor assays. Selected morphologic assays (the mouse mammary organ culture (MMOC) and human foreskin epithelial cell (HFE) morphologic assays) in combination give a predictive value that meaningfully forecasts results for chemopreventive efficacy in the mouse and rat mammary gland tumor assays.

Conclusions: These predictive models can be used to guide our future decision-making with respect to agent selection as well as morphologic and animal tumor assay use. Our future work is focused on deepening our understanding of our Predictive Value approach by: (1) identifying the mammary gland tumor assays that best reflect anti-tumor efficacy in animals; (2) teasing apart those classes of agents that exhibit the highest predictive values overall; and (3) examining which agent classes show the highest efficacy in specific mammary gland tumor assays.

Citation Format: Barbara K Dunn, Vernon E Steele, Carol F Topp, Richard M Fagerstrom, Barnett S Kramer. Understanding predictive values of short-term morphologic assays of cancer chemoprevention for efficacy in animal mammary gland tumor assays [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P5-11-06.

In North America, some ovarian cancers express the oncogenes of preventable human papillomavirus HPV-18

Some researchers in other regions have recommended human papillomavirus (HPV) vaccination to reduce risk of ovarian cancer, but not in North America, where evidence has previously suggested no role for HPV in ovarian cancer. Here we use a large sample of ovarian cancer transcriptomes (RNA-Seq) from The Cancer Genome Atlas (TCGA) database to address whether HPV is involved with ovarian cancer in North America. We estimate that a known high-risk type of HPV (type 18) is present and active in 1.5% of cases of ovarian epithelial cancers in the US and Canada. Our detection methods were verified by negative and positive controls, and our sequence matches indicated high validity, leading to strong confidence in our conclusions. Our results indicate that previous reports of zero prevalence of HPV in North American cases of ovarian cancer should not be considered conclusive. This is important because currently used vaccines protect against the HPV-18 that is active in ovarian tumors and, therefore, may reduce risk in North America of cancers of the ovaries as well as of the cervix and several other organ sites.

Results of the two incidence screenings in the National Lung Screening Trial

BACKGROUND

The National Lung Screening Trial was conducted to determine whether three annual screenings (rounds T0, T1, and T2) with low-dose helical computed tomography (CT), as compared with chest radiography, could reduce mortality from lung cancer. We present detailed findings from the first two incidence screenings (rounds T1 and T2).

METHODS

We evaluated the rate of adherence of the participants to the screening protocol, the results of screening and downstream diagnostic tests, features of the lung-cancer cases, and first-line treatments, and we estimated the performance characteristics of both screening methods.

RESULTS

At the T1 and T2 rounds, positive screening results were observed in 27.9% and 16.8% of participants in the low-dose CT group and in 6.2% and 5.0% of participants in the radiography group, respectively. In the low-dose CT group, the sensitivity was 94.4%, the specificity was 72.6%, the positive predictive value was 2.4%, and the negative predictive value was 99.9% at T1; at T2, the positive predictive value increased to 5.2%. In the radiography group, the sensitivity was 59.6%, the specificity was 94.1%, the positive predictive value was 4.4%, and the negative predictive value was 99.8% at T1; both the sensitivity and the positive predictive value increased at T2. Among lung cancers of known stage, 87 (47.5%) were stage IA and 57 (31.1%) were stage III or IV in the low-dose CT group at T1; in the radiography group, 31 (23.5%) were stage IA and 78 (59.1%) were stage III or IV at T1. These differences in stage distribution between groups persisted at T2.

CONCLUSIONS

Low-dose CT was more sensitive in detecting early-stage lung cancers, but its measured positive predictive value was lower than that of radiography. As compared with radiography, the two annual incidence screenings with low-dose CT resulted in a decrease in the number of advanced-stage cancers diagnosed and an increase in the number of early-stage lung cancers diagnosed. (Funded by the National Cancer Institute; NLST ClinicalTrials.gov number, NCT00047385.).

Results of initial low-dose computed tomographic screening for lung cancer

BACKGROUND

Lung cancer is the largest contributor to mortality from cancer. The National Lung Screening Trial (NLST) showed that screening with low-dose helical computed tomography (CT) rather than with chest radiography reduced mortality from lung cancer. We describe the screening, diagnosis, and limited treatment results from the initial round of screening in the NLST to inform and improve lung-cancer-screening programs.

METHODS

At 33 U.S. centers, from August 2002 through April 2004, we enrolled asymptomatic participants, 55 to 74 years of age, with a history of at least 30 pack-years of smoking. The participants were randomly assigned to undergo annual screening, with the use of either low-dose CT or chest radiography, for 3 years. Nodules or other suspicious findings were classified as positive results. This article reports findings from the initial screening examination.

RESULTS

A total of 53,439 eligible participants were randomly assigned to a study group (26,715 to low-dose CT and 26,724 to chest radiography); 26,309 participants (98.5%) and 26,035 (97.4%), respectively, underwent screening. A total of 7191 participants (27.3%) in the low-dose CT group and 2387 (9.2%) in the radiography group had a positive screening result; in the respective groups, 6369 participants (90.4%) and 2176 (92.7%) had at least one follow-up diagnostic procedure, including imaging in 5717 (81.1%) and 2010 (85.6%) and surgery in 297 (4.2%) and 121 (5.2%). Lung cancer was diagnosed in 292 participants (1.1%) in the low-dose CT group versus 190 (0.7%) in the radiography group (stage 1 in 158 vs. 70 participants and stage IIB to IV in 120 vs. 112). Sensitivity and specificity were 93.8% and 73.4% for low-dose CT and 73.5% and 91.3% for chest radiography, respectively.

CONCLUSIONS

The NLST initial screening results are consistent with the existing literature on screening by means of low-dose CT and chest radiography, suggesting that a reduction in mortality from lung cancer is achievable at U.S. screening centers that have staff experienced in chest CT. (Funded by the National Cancer Institute; NLST ClinicalTrials.gov number, NCT00047385.).

Abstract 2595: Predictive value tools as an aid in chemopreventive agent development

Over the past 25 years the Chemoprevention Agent Development Research Group of the Division of Cancer Prevention in the NCI has entered approximately 800 agents into a program designed to test for cancer preventive efficacy. The first 2 key steps in the testing pathway are in vitro/in vitro morphologic assays and, for agents successful in these, anti-tumor testing (tumor incidence and multiplicity reduction) in animal tumor models. Ultimately, agents that successfully decrease tumor incidence and multiplicity in animal models could progress to testing in humans. The goal of the current project is to evaluate the success of our program by determining how accurately the earlier stage assays (morphologic assays) predict efficacy in the later-stage assays (animal tumor assays). In our approach the 208 agents that were tested in both morphologic and animal tumor assays in this program were included in our analysis. Statistical modeling of the input from the 6 most commonly used morphologic assays into predicting the efficacy of agents in animal tumor models was carried out by multimodel inference. The statistical models that were generated predicted the ability of these 6 morphologic assays to predict tumor outcomes in animal models representing (1) all tumor types, (2) breast cancer only, and (3) colon cancer only. The statistical models assigned each morphologic assay a value describing how strongly it predicted outcomes in each of the 3 animal tumor assay settings. The development of these predictive models will guide our future decision-making with respect to agent selection as well as morphologic and animal tumor assay use. Our goal is to improve the efficiency of the overall process of chemoprevention agent development.

Citation Format: Barbara K. Dunn, Vernon E. Steele, Richard M. Fagerstrom, Carol F. Topp, David Ransohoff, Christopher Cunningham, Ronald A. Lubet, Leslie G. Ford, Barnett S. Kramer. Predictive value tools as an aid in chemopreventive agent development. [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 2595. doi:10.1158/1538-7445.AM2013-2595

Reduced lung-cancer mortality with low-dose computed tomographic screening

BACKGROUND

The aggressive and heterogeneous nature of lung cancer has thwarted efforts to reduce mortality from this cancer through the use of screening. The advent of low-dose helical computed tomography (CT) altered the landscape of lung-cancer screening, with studies indicating that low-dose CT detects many tumors at early stages. The National Lung Screening Trial (NLST) was conducted to determine whether screening with low-dose CT could reduce mortality from lung cancer.

METHODS

From August 2002 through April 2004, we enrolled 53,454 persons at high risk for lung cancer at 33 U.S. medical centers. Participants were randomly assigned to undergo three annual screenings with either low-dose CT (26,722 participants) or single-view posteroanterior chest radiography (26,732). Data were collected on cases of lung cancer and deaths from lung cancer that occurred through December 31, 2009.

RESULTS

The rate of adherence to screening was more than 90%. The rate of positive screening tests was 24.2% with low-dose CT and 6.9% with radiography over all three rounds. A total of 96.4% of the positive screening results in the low-dose CT group and 94.5% in the radiography group were false positive results. The incidence of lung cancer was 645 cases per 100,000 person-years (1060 cancers) in the low-dose CT group, as compared with 572 cases per 100,000 person-years (941 cancers) in the radiography group (rate ratio, 1.13; 95% confidence interval [CI], 1.03 to 1.23). There were 247 deaths from lung cancer per 100,000 person-years in the low-dose CT group and 309 deaths per 100,000 person-years in the radiography group, representing a relative reduction in mortality from lung cancer with low-dose CT screening of 20.0% (95% CI, 6.8 to 26.7; P=0.004). The rate of death from any cause was reduced in the low-dose CT group, as compared with the radiography group, by 6.7% (95% CI, 1.2 to 13.6; P=0.02).

CONCLUSIONS

Screening with the use of low-dose CT reduces mortality from lung cancer. (Funded by the National Cancer Institute; National Lung Screening Trial ClinicalTrials.gov number, NCT00047385.).

The National Lung Screening Trial: overview and study design

The National Lung Screening Trial (NLST) is a randomized multicenter study comparing low-dose helical computed tomography (CT) with chest radiography in the screening of older current and former heavy smokers for early detection of lung cancer, which is the leading cause of cancer-related death in the United States. Five-year survival rates approach 70% with surgical resection of stage IA disease; however, more than 75% of individuals have incurable locally advanced or metastatic disease, the latter having a 5-year survival of less than 5%. It is plausible that treatment should be more effective and the likelihood of death decreased if asymptomatic lung cancer is detected through screening early enough in its preclinical phase. For these reasons, there is intense interest and intuitive appeal in lung cancer screening with low-dose CT. The use of survival as the determinant of screening effectiveness is, however, confounded by the well-described biases of lead time, length, and overdiagnosis. Despite previous attempts, no test has been shown to reduce lung cancer mortality, an endpoint that circumvents screening biases and provides a definitive measure of benefit when assessed in a randomized controlled trial that enables comparison of mortality rates between screened individuals and a control group that does not undergo the screening intervention of interest. The NLST is such a trial. The rationale for and design of the NLST are presented.

Lung cancer: interobserver agreement on interpretation of pulmonary findings at low-dose CT screening

PURPOSE

To evaluate agreement among radiologists on the interpretation of pulmonary findings at low-dose computed tomographic (CT) screening examinations for lung cancer.

MATERIALS AND METHODS

Institutional review board approval and informed consent were obtained. HIPAA guidelines were followed. Sixteen radiologists from the 10 National Lung Screening Trial screening centers of the National Cancer Institute's Lung Screening Study network reviewed image subsets from 135 baseline low-dose screening CT examinations in 135 trial participants (89 men, 46 women; mean age, 62.7 years +/- 5.4 [standard deviation]). Interpretations were classified into one of four of the following categories: noncalcified nodule 4 mm or larger in greatest transverse dimension (positive screening result); noncalcified nodule smaller than 4 mm in greatest transverse dimension (negative screening result); calcified, benign nodule (negative screening result); or no nodule (negative screening result). A recommendation for follow-up evaluation was obtained for each case. Interobserver agreement was evaluated by using the multirater kappa statistic and by using response frequencies and descriptive statistics.

RESULTS

Multirater kappa values ranged from 0.58 (for agreement among all four classifications; 95% confidence interval: 0.55, 0.61) to 0.64 (for agreement on classification as a positive or negative screening result; 95% confidence interval: 0.62, 0.65). The average percentage of reader pairs in agreement on the screening result per case (percentage agreement) was 82%. There was wide variation in the total number of abnormalities detected and classified as pulmonary nodules, with differences of up to more than twofold among radiologists. For cases classified as positive, multirater kappa for follow-up recommendations was 0.35.

CONCLUSION

Interobserver agreement was moderate to substantial; potential for considerable improvement exists. Clinical trial registration no. NCT00047385.

Ovarian Cancer Screening in Women With a Family History of Breast or Ovarian Cancer

OBJECTIVE

To evaluate positive predictive values of CA 125 or transvaginal ultrasonography screening for ovarian cancer according to family history of breast or ovarian cancer.

METHODS

In the screening arm of a randomized controlled trial of screening compared with usual care, 28,460 women with family history data received baseline and annual CA 125 and transvaginal ultrasonography examinations. We analyzed CA 125 and transvaginal ultrasonography results from the first four rounds of screening. We classified women as average (n=22,687), moderate (n=2,572), or high (n=2,163) risk based on family history, or high risk due to a personal history of breast cancer (n=1,038). Cancers were identified by active follow-up of women with abnormal screening results and annual questionnaires. We calculated positive predictive values for screening combinations.

RESULTS

Similar proportions (4.8-5.0%) of women in each group had abnormal screening results. Higher-risk women were more likely than lower-risk women to undergo biopsy after a positive screen. Screening identified 43 invasive ovarian cancers. The positive predictive values for abnormal screening results were 0.7% in average-risk, 1.3% in moderate-risk, and 1.6% in high-risk groups; one ovarian cancer occurred among the breast cancer survivors. The positive predictive values for postbaseline abnormal screening results were also higher in the higher-risk groups. The positive predictive values did not significantly differ across risk groups.

CONCLUSION

Probabilities of abnormal annual CA 125 and transvaginal ultrasonography screens were similar across groups based on family history of breast or ovarian cancer. However, ovarian cancer was more likely to be diagnosed after an abnormal screening result among women at higher family history-based risk than among women at lower risk.

LEVEL OF EVIDENCE

I.

Prostate specific antigen changes as related to the initial prostate specific antigen: data from the prostate, lung, colorectal and ovarian cancer screening trial

PURPOSE

Annual screening with PSA, although of unproven benefit, is currently used for prostate cancer early detection. A large fraction of screened men have low (less than 2 ng/ml) initial PSA. The yield over time of positive PSA screens (ie more than 4 ng/ml) in these men has not been well characterized in large cohorts in the United States.

MATERIALS AND METHODS

Men in the screening arm of the PLCO received baseline PSA and annual tests for 5 years. 30,495 of these men had baseline PSA 4 ng/ml or less. We estimated the cumulative probability of converting to PSA greater than 4 at years 1 through 5 as a function of baseline PSA.

RESULTS

Among men with baseline PSA less than 1 ng/ml, 1.5% converted by year 5 (95% CI 1.2-1.7). Among men with baseline PSA of 1.0 to 1.99 ng/ml, 1.2% (95% CI 0.9-1.3) and 7.4% (95% CI 6.8-8.1) converted by year 1 and 5, respectively. A total of 33.5% and 79% of men with initial PSA of 2.0 to 2.99 and 3.0 to 4.0, respectively, converted by year 5. Of men with baseline PSA less than 1 ng/ml converting to PSA more than 4 ng/ml, 8% were diagnosed with cancer within 2 years of conversion. About 10% of men with baseline PSA less than 1 ng/ml and negative baseline DRE had a positive DRE within 3 years.

CONCLUSIONS

For men choosing PSA screening, screening every 5 years for baseline PSA less than 1 ng/ml and every 2 years for PSA 1 to 2 ng/ml could result in a 50% reduction in PSA tests and in less than 1.5% of men missing earlier positive screens, but with an unknown effect on prostate cancer mortality.

Ovarian cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial: findings from the initial screen of a randomized trial

OBJECTIVE

Ovarian cancer screening with transvaginal ultrasound (TVU) and CA-125 was evaluated in the Prostate, Lung, Colorectal and Ovarian (PLCO) Trial.

STUDY DESIGN

This was a randomized controlled trial of screening versus usual care. Baseline screening results are reported.

RESULTS

Of 39,115 women randomized to receive screening, 28,816 received at least 1 test. Abnormal TVU was found in 1338 (4.7%), and abnormal CA-125 in 402 (1.4%). Twenty-nine neoplasms were identified (26 ovarian, 2 fallopian, and 1 primary peritoneal neoplasm). Nine were tumors of low malignant potential and 20 were invasive. The positive predictive value for invasive cancer was 3.7% for an abnormal CA-125, 1.0% for an abnormal TVU, and 23.5% if both tests were abnormal.

CONCLUSION

The effect of screening on ovarian cancer mortality in the PLCO cohort has yet to be evaluated and will require longer follow-up. Screening identified both early- and late-stage neoplasms, and the predictive value of both tests was relatively low.

Estimating the cumulative risk of a false-positive test in a repeated screening program

The goal of screening tests for a chronic disease such as cancer is early detection and treatment with a consequent reduction in mortality from the disease. Screening tests, however, might produce false positive and false-negative results. With an increasing number of screening tests, it is clear that the risk of a false-positive screen, a finding with potentially significant emotional, financial, and health costs, also increases. Elmore et al. (1998, New England Journal of Medicine 338, 1089-1096), Christiansen et al. (2000, Journal of the National Cancer Institute 92, 1657-1666), and Gelfand and Wang (2000, Statistics in Medicine 19, 1865-1879) investigated this problem under the somewhat unrealistic assumption that the choice of making the decision to drop out at the kth screen does not depend upon the results of the earlier k - 1 screens. In this article we obtain sufficient and necessary conditions for their assumption to hold and use one of them to provide a method for testing the validity of the assumption. A new model which does not depend on their assumption is introduced. The maximum likelihood estimator of the cumulative risk of receiving a false-positive screen under the new model is derived and its asymptotic normality is proved. The extension of the new model by incorporating covariate information is also considered. We apply our testing method and the new model to data from the breast cancer screening trial of the Health Insurance Plan of Greater New York.

Final results of the Lung Screening Study, a randomized feasibility study of spiral CT versus chest X-ray screening for lung cancer

The Lung Screening Study (LSS) was a pilot study designed to assess the feasibility of conducting a large scale randomized controlled trial (RCT) of low radiation dose spiral computed tomography (LDCT) versus chest X-ray (CXR) for lung cancer screening. Baseline results of LSS have been previously reported. Here, we report on the findings at the year one screen and on the final results of the LSS study. A total of 1660 subjects were randomized to the LDCT arm and 1658 to the CXR arm. Compliance with screening declined from 96% at baseline to 86% at year one in the LDCT arm and declined from 93% at baseline to 80% at year one in the CXR arm. Positivity rates for the year one screen were 25.8% for LDCT and 8.7% for CXR. Cancer yield was significantly less at year one for LDCT, 0.57%, than at baseline, 1.9%; cancer yield for CXR increased from 0.45% at baseline to 0.68% at year one. Forty lung cancers in the LDCT arm and 20 in the CXR arm were diagnosed over the study period. Stage I cancers comprised 48% of cases in the LDCT arm and 40% in the CXR arm. A total of 16 stage III-IV cancers were observed in the LDCT arm versus nine in the CXR arm. The LSS has established the feasibility of a RCT comparing annual spiral CT to chest X-ray for lung cancer screening.

Quality control of cancer screening examination procedures in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial

Investigators for the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial describe quality control procedures for the digital rectal examination, ovarian palpation examination, transvaginal ultrasound, chest X-ray, and flexible sigmoidoscopy. These cancer screening tests are subjective and difficult to standardize. PLCO quality control procedures aim to measure and, where possible, reduce variation, across examiner and screening center, with respect to cancer screening test performance. Initial protocols stressed examiner qualifications, experience, and training; equipment specifications; examination procedures; and definitions for positive tests. The PLCO quality assurance subcommittee developed a final quality assurance plan, which included central approval and registration of PLCO examiners, direct observation of screening test performance during periodic site visits by the National Cancer Institute and coordinating center auditors, periodic analysis of screening test data, and procedures for independently duplicating or reviewing selected examinations. For each modality, the periodic data analyses examine the test-positive and the test-inadequate proportions and aim to identify divergent centers or examiners. Procedures for duplicating examinations specify feasible sample sizes for precise estimates of agreement between examiners, at each center, for each screening test modality, and over a 1-year period. These quality control procedures will help characterize the consistency and reliability of the PLCO cancer screening tests.

Design and evolution of the data management systems in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial

This paper describes the design and evolution of the data management systems developed in support of the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial. These systems span platforms from stand-alone computers to distributed systems on local area networks to mainframes. Allowing all of these systems to share appropriate information electronically introduces integration, synchronization, testing, and support challenges. For each platform, applications were developed to handle data entry, editing, trial management, reporting, telecommunications, and data sharing. Approaches to issues such as level of data access, integration with other, existing applications, and handling the expansion of the protocol are discussed.

Lung cancer mortality in the Mayo Lung Project: impact of extended follow-up

BACKGROUND

The Mayo Lung Project (MLP) was a randomized, controlled clinical trial of lung cancer screening that was conducted in 9211 male smokers between 1971 and 1983. The intervention arm was offered chest x-ray and sputum cytology every 4 months for 6 years; the usual-care arm was advised at trial entry to receive the same tests annually. No lung cancer mortality benefit was evident at the end of the study. We have extended follow-up through 1996.

METHODS

A National Death Index-PLUS search was used to assign vital status and date and cause of death for 6523 participants with unknown information. The median survival for lung cancer patients diagnosed before July 1, 1983, was calculated by use of Kaplan-Meier estimates. Survival curves were compared with the log-rank test.

RESULTS

The median follow-up time was 20.5 years. Lung cancer mortality was 4.4 (95% confidence interval [CI] = 3.9-4.9) deaths per 1000 person-years in the intervention arm and 3.9 (95% CI = 3.5-4.4) in the usual-care arm (two-sided P: for difference =.09). For participants diagnosed with lung cancer before July 1, 1983, survival was better in the intervention arm (two-sided P: =.0039). The median survival for patients with resected early-stage disease was 16.0 years in the intervention arm versus 5.0 years in the usual-care arm.

CONCLUSIONS

Extended follow-up of MLP participants did not reveal a lung cancer mortality reduction for the intervention arm. Similar mortality but better survival for individuals in the intervention arm indicates that some lesions with limited clinical relevance may have been identified in the intervention arm.

Estimation of post-lead-time survival under dependence between lead-time and post-lead-time survival

Early detection of cancer by screening advances the date of diagnosis, but may or may not alter time to death. Screening programme need to assess the true benefit of screening, that is, the length of time by which survival has been extended, beyond merely the time by which the diagnosis is advanced (lead-time). One method is to estimate the distribution of the time survived post-lead-time using total survival time data for screen-detected cancer cases, under the assumption of independence of the lead-time and the past-lead-time survival. However, it seems biologically reasonable that the lead-time and the post-lead-time survival are positively correlated. This paper investigates the consequences of departures from independence of lead-time and post-lead-time survival on estimation of post-lead-time survival. We introduce a new model that involves dependence between the lead-time and the post-lead-time survival. We show that the new model can be converted to the model discussed by Xu and Prorok. We consider the non-parametric maximum likelihood estimator of the post-lead-time survival under the new model. We apply the method to data from the HIP (Health Insurance Plan of Greater New York) breast cancer screening trial. We make comparisons with the survival of cancer cases not detected by screening, such as interval cases, cases among individuals who refused screening, and randomized control cases.

Prenatal sonographic assessment of the fetal thorax: normal values

Fetal thoracic circumference and thoracic length measurements were obtained by ultrasonographic examination on 576 women between 16 and 40 weeks gestation. Nomograms for thoracic circumference and thoracic length with respect to gestational age were developed. Growth of both thoracic parameters was observed to be linear throughout gestation. In normal pregnancy the ratios of thoracic circumference to abdominal circumference and thoracic length to humerus length remained virtually constant at 0.89 and 0.93, respectively.

Marked decline in trabecular bone mineral content in healthy men with age: lack of association with sex steroid levels

To define the association of age-related changes in bone mineral content to gonadal function in normal men, we measured radial (largely cortical) and vertebral (largely trabecular) bone mineral content (BMC), testosterone (total and free), estrone and estradiol-17B levels in 62 healthy subjects, ages 30 to 92. Radial BMC fell 2 to 3.4% per decade but vertebral trabecular BMC declined more rapidly at 12% per decade. Of the sex steroids measured the only statistically significant change occurred in free testosterone levels which decreased with age (r = -.57, P less than .0001). Free testosterone levels correlated significantly with trabecular vertebral BMC (r = .458, P less than .0002) but not with bone mineral measures at the predominantly cortical radial sites. However, by multiple regression analysis free testosterone did not add to the effect of age on vertebral BMC. There were no associations of total testosterone, estrone, or estradiol levels to bone mineral content at any of the three sites measured in these healthy men. Age-related declines in male gonadal function do not appear to be of primary importance in male age-related bone loss.

Twin pregnancy: routine use of ultrasound examinations in the prenatal diagnosis of intrauterine growth retardation and discordant growth

A prospective study was undertaken in 42 normal twin pregnancies to determine the value of routine ultrasound evaluation in the prenatal diagnosis of intrauterine growth retardation (IUGR) and intra-pair growth discordancy in twin gestations. Use of multiple sonographic parameters significantly improved diagnostic accuracy when compared with the use of single parameters, such as biparietal diameter (BPD) alone in the prepartum diagnosis of these two entities.

Deficiencies of clinical trials of alcohol withdrawal

Eighty-one therapeutic trials of alcohol withdrawal were found that have been published in English since 1954; controls were randomized in 29 (RCTs). Two thousand three hundred thirteen patients were randomized. Variable pretreatment description prevented estimates of delirium tremens and convulsion prevalence, but only four deaths were reported. Endpoints were thus entirely subjective in these moderately ill patients. Protocol quality of the RCTs was graded by a previously developed system for evaluating adequacy of descriptions, blinding, and essential measurements. Mean score obtained was .49 +/- .03 (1 SE). (A perfect paper would score 1.00.) Data presentations and statistical analyses scored .18 +/- .03. There was little evidence of improvement of scores over time. Papers lacked confidence intervals, proper handling of dropouts, and adequate details of side effects. In five RCTs, six comparisons showed that benzodiazepines are clearly superior to placebo (p less than .001), but conclusions about comparisons with other drugs were not possible. In none of eight "negative" comparisons was the probability of a type II error (beta) considered. Discovery of more effective symptomatic agents or methods of reducing the death rate will require more rigid protocols and analyses as well as larger studies to allow the use of more critical endpoints such as occurrence of delirium tremens, convulsions, or death.

A survey of clinical trials of antibiotic prophylaxis in colon surgery: evidence against further use of no-treatment controls

To evaluate the use of antibiotics given prophylactically of colon surgery, we examined 26 trials published from 1965 to 1980 in which patients given various antibiotic regiments were compared with controls given no antibiotic treatment. In 22 (85 per cent of these trials) antibiotics reduced postoperative wound infection (p less than 0.05 in 14). Combining the results of the trials published from 1965 to 1975 reveals a 95 per cent confidence interval from the true difference in infection rates of 14 +/- 6 per cent (36 per cent for control group vs. 22 per cent for treatment group) and the true difference in death rates of 6.7 +/- 4.4 per cent (11.2 per cent for control group vs 4.5 per cent for treatment group). Yet trials employing control groups given no treatment continue to be reported. Since the use of such controls is justified only when no effective alternative therapy exists, we believe that any further trials of antibiotic prophylaxis in colon surgery should employ a previously proved standard. However, steadily increasing efficacy of treatment means that comparisons of new therapies with standard therapies will become prohibitively expensive because of the large number of patients required.