Sonographers' self-reported visualization of normal postmenopausal ovaries on transvaginal ultrasound is not reliable: results of expert review of archived images from UKCTOCS.

OBJECTIVE: In the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS), self-reported visualization rate (VR) of the ovaries by the sonographer on annual transvaginal sonographic (TVS) examinations was a key quality control (QC) metric. The objective of this study was to assess self-reported VR using expert review of a random sample of archived images of TVS examinations from UKCTOCS, and then to develop software for measuring VR automatically. METHODS: A single expert reviewed images archived from 1000 TVS examinations selected randomly from 68 931 TVS scans performed in UKCTOCS between 2008 and 2011 with ovaries reported as 'seen and normal'. Software was developed to identify the exact images used by the sonographer to measure the ovaries. This was achieved by measuring caliper dimensions in the image and matching them to those recorded by the sonographer. A logistic regression classifier to determine visualization was trained and validated using ovarian dimensions and visualization data reported by the expert. RESULTS: The expert reviewer confirmed visualization of both ovaries (VR-Both) in 50.2% (502/1000) of the examinations. The software identified the measurement image in 534 exams, which were split 2:1:1 providing training, validation and test data. Classifier mean accuracy on validation data was 70.9% (95% CI, 70.0-71.8%). Analysis of test data (133 exams) provided a sensitivity of 90.5% (95% CI, 80.9-95.8%) and specificity of 47.5% (95% CI, 34.5-60.8%) in detecting expert confirmed visualization of both ovaries. CONCLUSIONS: Our results suggest that, in a significant proportion of TVS annual screens, the sonographers may have mistaken other structures for normal ovaries. It is uncertain whether or not this affected the sensitivity and stage at detection of ovarian cancer in the ultrasound arm of UKCTOCS, but we conclude that QC metrics based on self-reported visualization of normal ovaries are unreliable. The classifier shows some potential for addressing this problem, though further research is needed. © 2017 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of the International Society of Ultrasound in Obstetrics and Gynecology.

Prostasin, a potential serum marker for ovarian cancer: identification through microarray technology.

BACKGROUND: Screening biomarkers for ovarian cancer are needed because of its late stage at diagnosis and poor survival. We used microarray technology to identify overexpressed genes for secretory proteins as potential serum biomarkers and selected prostasin, a serine protease normally secreted by the prostate gland, for further study. METHODS: RNA was isolated and pooled from three ovarian cancer cell lines and from three normal human ovarian surface epithelial (HOSE) cell lines. Complementary DNA generated from these pools was hybridized to a microarray slide, and genes overexpressed in the cancer cells were identified. Real-time quantitative polymerase chain reaction was used to examine prostasin gene expression in ovarian cancer and HOSE cell lines. Anti-prostasin antibodies were used to examine prostasin expression and to measure serum prostasin by an enzyme-linked immunosorbent assay in 64 case patients with ovarian cancer and in 137 control subjects. Previously determined levels of CA 125, an ovarian cancer marker, were available from about 70% of all subjects. All statistical tests were two-sided. RESULTS: Prostasin was detected by immunostaining more strongly in cancerous ovarian epithelial cells and stroma than in normal ovarian tissue. The mean level of serum prostasin was 13.7 microg/mL (95% confidence interval [CI] = 10.5 to 16.9 microg/mL) in 64 case patients with ovarian cancer and 7.5 microg/mL (95% CI = 6.6 to 8.3 microg/mL) in 137 control subjects (P<.001, after adjustment for the subject's age, year of collection, and specimen quality). In 14 of 16 case patients with both preoperative and postoperative serum samples, postoperative prostasin levels were statistically significantly lower than preoperative levels (P =.004). In 37 case patients with nonmucinous ovarian cancer and in 100 control subjects for whom levels of CA 125 and prostasin were available, the combination of markers gave a sensitivity of 92% (95% CI = 78.1% to 98.3%) and a specificity of 94% (95% CI = 87.4% to 97.7%) for detecting ovarian cancer. CONCLUSIONS: Prostasin is overexpressed in epithelial ovarian cancer and should be investigated further as a screening or tumor marker, alone and in combination with CA 125.

Comparative morphometric study of tumor vasculature in human squamous cell carcinomas and their xenotransplants in athymic nude mice.

When human tumor xenotransplants into nude mice are used as experimental models, it is important to know whether their microvascular anatomy is rather host or tumor specific. Therefore a morphometric comparison of the vascular network in human squamous cell carcinomas and their xenotransplants was carried out. Biopsies were taken from surgical specimens of three squamous cell carcinomas of the oral cavity. Part of the material was processed for histology and the rest was cut into 1-mm3 cubes and transplanted s.c. into the lateral thorax of athymic nude mice [NCr/Sed(nu/nu)]. The microvascular architecture of original tumors and of three first, one second, and one late generation xenografts was compared. Capillaries were identified in original human tumors by anti-factor VIII staining and in xenografts with antilaminin staining. The median distances between interphase tumor cells and blood vessels were determined and were found to be much longer in original human tumors than in xenografts, ranging from 81 microns to 99 microns and 53 microns to 65 microns, respectively. However, the characteristic qualitative histology of tumors appeared to be preserved in xenotransplants. Analysis of the topographic distribution of mitotic figures revealed that in both original tumors and xenotransplants proliferation of tumor cells was concentrated around blood vessels. Again, vascular distances in original tumors were significantly longer than in xenotransplants. In addition, xenotransplants into nude mice from a long passaged cell line from a human pharyngeal squamous cell carcinoma, FaDu, was investigated. FaDu showed a rare-fication of the capillary network with increasing tumor volume, but a constant median distance of mitotic figures from blood vessels. In conclusion, the pattern of spatial distribution of proliferating tumor cells as well as differentiation characteristics appear to be retained in xenograft tumors, but the density of the vascular system is host specific. This has to be taken into account when physiological parameters of blood supply are studied in xenotransplanted tumors.

Analysis of c-erbB-2 expression in breast carcinomas with clinical follow-up.

Various monoclonal antibodies reactive with protooncogene products or tumor-associated antigens have been utilized to investigate breast carcinoma biology or antigen expression with potential prognostic relevance. Murine monoclonal antibody TA1, generated by immunization of BALB/c mice with whole c-erbB-2 (neu) transformed NIH/3T3 cells, recognizes the extracellular domain of the c-erbB-2 protein and binds a Mr 185,000 protein by immunoprecipitation. Using avidin-biotin-peroxidase techniques and monoclonal antibody TA1, 313 archival primary adenocarcinomas of the breast were evaluated for c-erbB-2 overexpression; 290 of these were used for multiparametric statistical analysis. Historical, clinical (age, laterality), histological (nuclear grade, tumor size, lymph node status, lymphatic or blood invasion), and hormone receptor data as well as clinical outcome (minimal follow-up, 6 years; median follow-up, 8.5 years) were compared to TA1 staining. For these 290 patients Cox regression multivariate analysis showed the strongest correlation between lymph node status or estrogen receptor status and overall survival (P = 0.0001 and 0.049, respectively). TA1 staining did not significantly correlate with survival (P = 0.395). However, univariate analysis of certain patient subpopulations showed a significant correlation if the examined tumors were subdivided into negative or focally reactive and those with greater than or equal to 40% cellular reactivity. For T3, T4 patients, strong TA1 immunoreactivity correlated with a shortened disease-free survival (log rank P = 0.0018; Wilcoxon p = 0.0078) and overall survival (log rank P = 0.0002; Wilcoxon P = 0.0013). For these patients the overall survival at 6 years was markedly different between the strongly reactive tumors (0%) and the negative to weakly reactive tumors (55%). In lymph node-positive patients a trend between high TA1 reactivity and a worse overall survival was also noted (log rank P = 0.128; Wilcoxon P = 0.054), with a 6-year survival of 42% in the strongly reactive tumors (n = 16) and 65% in the negative to weakly reactive carcinomas (n = 105). No correlation between TA1 immunoreactivity and other historical, clinical, and histological features were noted. c-erbB-2 overexpression as measured by immunohistochemical techniques, therefore, may have clinical significance in certain patient subpopulations.

Stage IA to IIB mediastinal Hodgkin's disease: three-dimensional volumetric assessment of response to treatment.

From 1979 to 1986, the response to treatment of 53 patients with stage IA to IIB mediastinal Hodgkin's disease was evaluated by three-dimensional volumetric analysis using thoracic computed tomographic (CT) scans. The mean initial volume of mediastinal disease in 34 patients treated with mantle and para-aortic irradiation was 166 mL, whereas for 19 patients treated with two to six cycles of multiagent chemotherapy and mantle and para-aortic irradiation the mean initial volume was 446 mL. Preliminary data suggested that patients with mediastinal volumes of less than 200 mL had a lower mediastinal relapse rate (13%) than patients with volumes greater than 200 mL (32%). For 12 patients receiving six cycles of nitrogen mustard, vincristine, procarbazine, and prednisone (MOPP), those with a greater than 85% reduction in volume 1 to 2 months after chemotherapy had a lower incidence of mediastinal relapse (zero of six, 0%) compared with patients having 85% or less reduction in volume (four of six, 67%). The primary value of this technique is that it provides a sensitive assessment of response to treatment and may aid in monitoring the effectiveness of a given treatment.

Prospective study of the outcomes of ambulatory patients with excessive warfarin anticoagulation.

BACKGROUND: Warfarin sodium therapy is highly effective in preventing thromboembolism. Its major toxic effect is hemorrhage, the risk of which increases with the international normalized ratio (INR). Data on the rate of major hemorrhage and the rate of INR decay after an episode of excessive anticoagulation therapy would help guide management of elevated INRs in the outpatient setting. METHODS: We prospectively followed up outpatients in an anticoagulant therapy unit from April 24, 1995, through March 1, 1996. Study patients had to be taking warfarin for longer than 1 month and have an INR target range of 2.0 to 3.0. Consecutive outpatients with an INR greater than 6.0 were identified and compared with a randomly selected concurrent set of patients whose INR was in the target range. Major hemorrhage was defined as fatal, intracranial, or requiring hospitalization and transfusion of at least 2 U of blood. RESULTS: One hundred fourteen patients with INRs greater than 6.0 were identified and compared with 268 patients with INRs in the target range. None of the patients had clinically apparent bleeding at the time of the INR measurement, and none received phytonadione (vitamin K1). Patients did not differ significantly in age, sex, indication, or duration of warfarin therapy. Ten patients with an INR greater than 6.0 (8.8%; 95% confidence interval, 4.3%-15.5%) sought medical attention for abnormal bleeding, and 5 of these experienced a major hemorrhage during 14-day follow-up (4.4%; 95% confidence interval, 1.4%-9.9%) compared with none of the patients with an in-range INR (P<.001). Thirty-three percent of patients with INRs greater than 6.0 had INRs less than 4.0 within 24 hours, 55% within 48 hours, 73% within 72 hours, and nearly 90% within 96 hours of temporary discontinuation of warfarin therapy. CONCLUSIONS: Outpatients with INRs greater than 6.0 face a significant short-term risk of major hemorrhage. Randomized trials are needed to determine the net benefit of preventive treatment with phytonadione.

Factors influencing serum CA125II levels in healthy postmenopausal women.

Our objective was to identify factors that correlate with CA125 concentrations in healthy postmenopausal women and to introduce recommendations for reporting and interpreting individual CA125 assay results. We analyzed repeated serum CA125 levels, as measured by the CA125II assay, in 18,748 postmenopausal women who participated in the ST: Bartholomew's/Royal London Hospital Ovarian Cancer screening trial from 1986 to 1994 and were not diagnosed with ovarian cancer during the 12-year follow-up period. We found that race is a substantial predictor of normal levels of CA125, with average CA125II concentration from African (median, 9.0; 95% range, 4.0-26.0 units/ml) and Asian women (median, 13.0; range, 5.9-33.3 units/ml) lower than that in Caucasian women (median, 14.2; range, 6.0-41.0 units/ml; P < 0.001). Women with a hysterectomy have lower CA125II values (median, 13.6; range 5.5-39.0 units/ml; P < 0.001), and women with a prior cancer diagnosis other than ovarian cancer have higher levels of CA125 II (median, 16.0; range, 6.0-49.0 units/ml; P < 0.003). Regular smoking and caffeine consumption decrease CA125 levels (P < 0.001). A woman's age, age at menarche, age at menopause, and history of a previous ovarian cyst (P < 0.05) are also predictive of baseline CA125 levels. Parity, history of hormone replacement therapy or unilateral oopherectomy, and previous use of oral contraceptives or talcum powder are not significant predictors of CA125 concentrations (P > 0.05). We concluded that clinically significant differences in individual patient characteristics need to be reflected in the screening algorithms that use CA125II so that designed performance characteristics (sensitivity and specificity) are maintained in practice.

A dose response analysis of injury to cranial nerves and/or nuclei following proton beam radiation therapy.

The low tolerance of the central nervous system (CNS) limits the radiation dose which can be delivered in the treatment of many patients with brain and head and neck tumors. Although there are many reports concerning the tolerance of the CNS, few have examined individual substructures of the brain and fewer still have had detailed dose information. This study has both. A three dimensional planning system was used to develop the combined proton beam/photon beam treatments for 27 patients with skull-base tumors. The cranial nerves and their related nuclei were delineated on the planning CT scans and the radiation dose to each was determined from three dimensional dose distributions. In the 594 CNS structures (22 structures/patient in 27 patients), there have been 17 structures (in 5 patients) with clinically manifest radiation injury, after a mean follow-up time of 74 months (range 40-110 months). From statistical analyses, dose is found to be a significant predictor of injury. Using logistic regression analysis, we find that, for each cranial nerve, at 60 Cobalt Gray Equivalent (CGE) the complication rate is 1% (0.5-3% with 95% confidence) and that the 5% complication rate occurs at 70 CGE (64-81 CGE with 95% confidence). The slope of the dose response curve (at 50%) is 3.2 (2.2-5.4 with 95% confidence). No significant relationship between dose and latency period for nerve injury was found.

Clinical implications of heterogeneity of tumor response to radiation therapy.

Heterogeneity of response of tumor tissue to radiation clearly exists. Major parameters include histopathologic type, size (number of tumor rescue units (TRUs)), hemoglobin concentration, cell proliferation kinetics and immune rejection reaction by host. Further, normal and presumably tumor tissue response is altered in certain genetic diseases, e.g. ataxia telangiectasia. Any assessment of response of tumor tissue to a new treatment method or the testing of a new clinical response predictor is optimally based upon a narrow strata, viz., uniform with respect to known parameters of response, e.g. size, histological type. Even among tumors of such a clinically defined narrow strata, there will be residual heterogeneity with respect to inherent cellular radiation sensitivity, distributions of pO2, (SH), cell proliferation etc. The value of a response predictor of an individual tumor will be determined by the heterogeneity of values for these and or other characteristics and by the coefficient of variation (CV) of the measured values of the individual parameters. Heterogeneity of one or more parameters of response is reflected in the slope of the dose response curve for local control, viz. the greater the heterogeneity the less steep the slope. To examine for this effect, the slope of dose response curves for control of model tumors of 10(8) tumor rescue units (TRU) and the SF2 = 0.5 (survival fraction after a single dose of 2 Gy) has been used to assess the impact of inter- and intra-tumoral variation of SF2 on slope, defined as gamma 50 values. The gamma 50 is the increase in local control expressed in percent points for a one percentage increment in dose, at the mid-point on the dose-response curve. The gamma 50 was 6.5 for CV = 0.0. For inter-tumoral CVs of 10%, 20% and 40%, the gamma 50 rapidly decreased to 2.4, 1.3 and 0.7. Intra-tumoral variation was less important, viz., for CVs of 10%, 20%, and 40% the gamma 50 values were reduced to 5.3, 3.8 and 2.2. Combining inter- and intra-tumoral variation reduced the gamma 50 only slightly below that for inter-tumoral variation alone. For example, were the CV 10% for inter- and intra-tumoral variation, the gamma 50 would be 2.1 as compared to 2.4 for inter-tumoral variation alone. The number of TRUs also affects slope, viz. gamma 50 increased from 1 to 9.7 as the TRU number increased from 10(1) to 10(12).(ABSTRACT TRUNCATED AT 400 WORDS)

Three-dimensional volumetric assessment of response to treatment: stage I and II diffuse large cell lymphoma of the mediastinum.

From 1981 to 1986, 12 patients with Stage I and II diffuse large cell lymphoma of the mediastinum were treated with 4 or more cycles of multiagent chemotherapy and for nine patients this was followed by mediastinal irradiation. The response to treatment was assessed by three-dimensional volumetric analysis utilizing thoracic CT scans. The initial mean tumor volume of the five patients relapsing was 540 ml in contrast to an initial mean tumor volume of 360 ml for the seven patients remaining in remission. Of the eight patients in whom mediastinal lymphoma volumes could be assessed 1-2 months after chemotherapy prior to mediastinal irradiation, the three patients who have relapsed had volumes of 292, 92, and 50 ml (mean volume 145 ml) in contrast to five patients who have remained in remission with residual volume abnormalities of 4-87 ml (mean volume 32 ml). Four patients in prolonged remission with CT scans taken one year after treatment have been noted to have mediastinal tumor volumes of 0-28 ml with a mean value of 10 ml. This volumetric technique to assess the extent of mediastinal large cell lymphoma from thoracic CT scans appears to be a useful method to quantitate the amount of disease at presentation as well as objectively monitor response to treatment.

Quantifying the potential benefit of CA 125 screening for ovarian cancer.

We develop a stochastic model of screening for ovarian cancer with serum levels of the CA 125 radioimmunoassay, a tumor-specific marker. The natural history of the disease is constructed using a four stage model characterized by a multivariate distribution for duration in each stage. Preserving its important features, we simplify the model to a function of two parameters; the average duration in stage I and the coefficient of variation of duration in each stage. A yearly screening program is superimposed using exponential CA 125 growth curves which result in stage-specific sensitivities corresponding to values reported in the literature. By implementing a computer simulation of the stochastic model, we estimate the benefit due to screening. This benefit is expressed as expected years of life saved per case of ovarian cancer. The model incorporates the stochastic nature of the disease process, allows easy analysis of repeated screens, and automatically accounts for correlation between subsequent tests. It provides the basis for planning optimal screening strategies with CA 125 testing.

Effects of pentobarbital anesthesia on the energy metabolism of murine tumors studied by in vivo 31P nuclear magnetic resonance spectroscopy.

The effects of pentobarbital anesthesia on the energy metabolism of FSaII and MCaIV foot tumors in mice were studied by 31P MRS. Using an 8.5 T spectrometer, in vivo spectra were obtained in 15 animals before and after pentobarbital anesthesia (0.05 mg/g ip). The average phosphocreatine/inorganic phosphate ratios (PCr/Pi) with and without pentobarbital were similar for both tumor histologies. Effects on individual tumors, however, were greater than 20% in 9/15 animals and greater than 50% in 6/15 animals. Pentobarbital anesthesia increased the variability of tumor intracellular pH, and the phosphomonoester/nucleotide triphosphate (PME/NTP) and nucleotide triphosphate/inorganic phosphate ratios (NTP/Pi). When examining the average in a cohort, pentobarbital anesthesia had no significant effect on the PCr/Pi, PME/NTP, NTP/Pi ratios or the pH. However, approximately equal to 50% of individual tumors do have significant changes in these parameters. The anesthesia-induced variability of tumor energy metabolism may explain the decrease in TCD50 observed in previous studies using multifraction radiation.

Fractionated proton radiation therapy of chordoma and low-grade chondrosarcoma of the base of the skull.

Sixty-eight patients with chordoma or low-grade chondrosarcoma at the base of the skull received fractionated high-dose postoperative radiation delivered with a 160-MeV proton beam. Protons have favorable physical characteristics which allow the delivery of high doses of radiation to these critically located tumors. The methods employed for these treatments are described. These patients have been followed for at least 17 months and for a median of 34 months. The median tumor dose was 69 CGE (cobalt Gy equivalent): CGE is the dose in proton Gy multiplied by 1.1, which is the relative biological effectiveness for protons compared to cobalt-60. The daily dose was 1.8 to 2.1 CGE. For this group the 5-year actuarial local control rate is 82% and disease-free survival rate is 76%. The incidence of treatment-related morbidity has been acceptable.

Clinical interest in determinations of cellular radiation sensitivity.

Determinations of cell sensitivity in terms of survival fraction after doses employed in clinical radiation therapy, say 1-3 Gy, are of increasing interest to clinicians as they provide direct experimental data which can be employed without reference to models of cell inactivation. SF2 values are expected ultimately to prove valuable as response predictors. Even so, SF2 values would surely be combined with other predictors also under development to give the best feasible estimate of response of tumor and normal tissue. There are, however, several concerns with the SF2 data currently available. These include: SF2 depends upon the cell system employed (established cell lines vs primary cultures) and the method of assaying survival fraction (colony formation vs population growth); dose-response curves for inactivation of tumors characterized by the reported distribution of SF2 values are, in many instances, not close to those judged to obtain in clinical practice; the broad distribution of SF2 values indicates a rather flatter dose-response curve for tumor control or normal tissue than seems true from clinical experience. There appears to be a potential for clinical gain by determination of sensitivity of normal tissues in order to identify patients who are of increased sensitivity (for example heterozygotes for AT, 5-oxoprolinuria, etc.). Although the absolute SF2 values obtained by current technologies of culturing human cells often appear to be poorly related to values expected from observed radiation response in patients, intensive research on cell viability assays will almost certainly yield more realistic results.

Tumor markers in screening for ovarian cancer.

Tumor markers are used for multiple purposes in clinical care, including screening asymptomatic subjects, differential diagnosis of symptomatic patients, treatment planning, prognosis during and immediately following treatment, and monitoring for recurrence. Generally, tumor markers have found most clinical utility in monitoring for recurrence of disease (1). Bast and coinvestigators discovered CA125 in 1979 using monoclonal antibody techniques (2), and subsequently demonstrated its utility in monitoring treatment and recurrence of ovarian cancer (3). CA125 is the most widely used ovarian tumor marker, and is currently approved in the United States for monitoring of disease to determine if second-look surgery is required. Tumor markers have not gained wide acceptance for early detection of disease with the one exception of PSA for prostate cancer in the U.S. The lack of acceptance is mainly because of the difficult hurdles a screening strategy must overcome, and few tumor markers have shown sufficient promise in overcoming these hurdles to put them to the test in a randomized controlled trial. Because of the low incidence of most cancers, sample sizes for prospective randomized screening trials are huge, so that sufficient numbers of disease specific events occur by the end of the trial. The significant costs entailed in clinical trials of this size imply that only very promising approaches to screening warrant prospective investigation. For ovarian cancer, three large trials are underway, two trials are planning to randomize 120,000 women followed for 7-8 yr (4,5), and an NCI trial will randomize 74,000 women followed for 16 yr (6).

Importance of tumor size and repopulation for radiocurability of skin cancer.

Data on 946 skin cancers treated by radiation were used to estimate the importance of repopulation. Six different treatment regimes were used from a single dose to 74 Gy given in 47 fractions. High local control of the small skin cancers (L 1 cm) was independent of dose fractionation. For large tumors, only 74 Gy in 47 fractions was the optimal treatment. Time factor analysis showed a steep increase in the NTD50 values between day 28 and 65 of treatment. This implies that tumor clonogen repopulation starts around 4 weeks of treatment. The present results showed a three-component dose response curve instead of the two-component curves which were found for head and neck and bladder cancer.

Survival analysis in familial ovarian cancer, a case control study.

OBJECTIVE: familial ovarian cancer patients have been found to differ from sporadic cases, clinically as well as in the molecular make-up of the tumour. Here, a case control study is performed to analyse potential differences in survival. STUDY DESIGN: 31 families with a strong history of ovarian and/or breast cancer presenting to a family cancer clinic 44 ovarian cancer patients were included. Each patient was matched for age and stage with controls from a cancer registry. Survival rates and the effect of several prognostic factors were analysed. RESULTS: median survival in the study group differed significantly from controls. A survival benefit for familial cases was maintained up to 5 years after diagnosis. Long-term survival was equally poor in both groups. CONCLUSION: the difference in survival between familial ovarian cancer cases and matched controls may reflect differences in biological behaviour. This may have important implications for the management and prevention of familial ovarian cancer.

Risk of diagnosis of ovarian cancer after raised serum CA 125 concentration: a prospective cohort study.

OBJECTIVE: To determine the risk of invasive epithelial ovarian cancer and fallopian tube cancer associated with a raised concentration of the tumour marker CA 125 in asymptomatic postmenopausal women. DESIGN: Serum CA 125 concentration was measured annually in study participants for one to four years. Participants with a concentration > or = 30 U/ml were recalled for abdominal ultrasonography. Follow up was by annual postal questionnaire. SETTING: General practice, occupational health departments, ovarian cancer screening unit in a teaching hospital. SUBJECTS: 22,000 volunteers, all postmenopausal women > or = 45 years of age; recruited between 1 June 1986 and 1 May 1990. INTERVENTION: Surgical investigation if the ultrasound examination was abnormal. MAIN OUTCOME MEASURES: Cumulative and relative risk of developing an index cancer (invasive epithelial cancer of the ovary or fallopian tube) after a specified CA 125 result. RESULTS: 49 index cancers developed in the study population during a mean follow up of 6.76 years. The overall cumulative risk of developing an index cancer was 0.0022 for the entire study population and was lower for women with a serum CA 125 concentration < 30 U/ml (cumulative risk 0.0012) but was appreciably increased for women with a concentration > or = 30 U/ml (0.030) and > 100 U/ml (0.149). Compared with the entire study population the relative risk of developing an index cancer within one year and five years was increased 35.9-fold (95% confidence interval 18.3 to 70.4) and 14.3-fold (8.5 to 24.3) respectively after a serum CA 125 concentration > or = 30 U/ml and 204.8-fold (79.0 to 530.7) and 74.5-fold (31.1 to 178.3) respectively after a concentration > or = 100 U/ml. CONCLUSION: CA 125 is a powerful index of risk of ovarian and fallopian tube cancer in asymptomatic postmenopausal women.