Ovarian Cancer Screening by MikeJenny


									ANTICANCER RESEARCH 26: 4793-4802 (2006)


                                       Ovarian Cancer Screening

                            1Departmentof Obstetrics and Gynaecology, Hammersmith Hospital,
                             Queen Charlotte’s and Chelsea Hospital, London, W12 0HS;
                        2Department of Obstetrics and Gynaecology, The Simpson Centre for

                     Reproductive Health, Royal Infirmary of Edinburgh, Edinburgh, EH16 4SA;
     3Department of Obstetrics and Gynaecology, Central Lancashire Teaching Hospitals, Preston, PR2 9HT, U.K.;
                            4Scientific Institute of Public Health, 1050, Brussels, Belgium;
                                  5European Cancer Network, IARC, Lyon, France;
              6Department of Obstetrics and Gynaecology, Coombe Women’s Hospital, Dublin, Ireland;
                 7Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford, U.K.;
            8Department of Obstetrics and Gynaecology, University Hospital of Ioannina, Ioannina, 45100;
        9Department of Obstetrics and Gynaecology, University Hospital of Thessaloniki, Thessaloniki, Greece

Abstract. Purpose: This review presents the current                23,000 new cases (2) and 14,000 deaths per annum (3).
knowledge on ovarian cancer screening (OCS) together with          Approximately 6,100 new cases were registered in the
a cumulative evaluation of the efficacy across diverse             England and Wales in 1997, making it the fourth most
screening approaches based on up-to-date results. Materials        common site of cancer in women (after breast, colorectal
and Methods: An electronic literature search was conducted         and lung) in the UK (4). The estimated incidence and
targeting reports on the effectiveness of the various screening    mortality from OC worldwide is presented in Figure 1 (5).
strategies. Results: Twenty-two prospective studies, 18               The overall five-year survival rate is poor, at about 30%
cohorts and 4 randomised control trials (RCTs) were                for women >65 years (6), with minimal improvement in this
retrieved. Where possible, sensitivity, specificity and            figure over the past 20-30 years, despite the introduction of
predictive values, as well as the number needed to treat for       new chemotherapy agents. Survival is much better, around
each cancer (NNT) and each stage I disease [NNT (I)]               94% at five years, for women whose disease is localised to
detected are reported. Conclusion: The multimodal approach         the ovaries (FIGO stage I), but less than a third of cases
that incorporates CA125 as a primary and ultrasound as a           according to US statistics are currently diagnosed at this
secondary test appears to be superior to other strategies.         stage (6). Hence, the high mortality rates are attributed to
However, the conclusive end-point that would provide the           the fact that over 70% of women are diagnosed with stage II
acceptable level of evidence for the introduction of screening     or worse, where the 5-year survival is significantly less (6)
should be a reduction in mortality. Up-to-date, there is no        (Figure 2).
such evidence to justify this.                                        The purpose of this review is to present a
                                                                   comprehensive and concise overview of the evidence base
Ovarian cancer (OC) is the seventh most common cancer in           of ovarian cancer screening (OCS) in terms of
women worldwide. It is the fifth leading cause of cancer           methodology, application and possible restrictions, as well
death in U.S. women (1), accounting for approximately              as to present the existing literature and up-to-date
                                                                   evidence regarding OCS and the effectiveness of each
                                                                   available screening strategy in the general low-risk
                                                                   population. All the studies identified in the literature have
Correspondence to: Prof. Evangelos Paraskevaidis, Department of
                                                                   been scrutinised with regard to study populations,
Obstetrics and Gynaecology, University Hospital of Ioannina,
45110, Ioannina, Greece. Tel: +30-26510-99224 / +30-26510-
                                                                   methodologies and diagnostic performance criteria. These
99467, Fax: +30-26510-99224, e-mail: vangelispar@hotmail.com       results are presented and summarised so that the reader
                                                                   can draw conclusions on the effectiveness of different
Key Words: Ovarian cancer, screening, ultrasound, CA125, review.   screening techniques.

0250-7005/2006 $2.00+.40                                                                                                   4793
                                           ANTICANCER RESEARCH 26: 4793-4802 (2006)

Figure 1. Estimated incidence of and mortality from ovarian cancer worldwide (Globocan 2000).

Risk factors. Ovarian cancer is extremely rare in women
younger than 25 years (1 in 100,000), with a steep increase
after 50 years (32 per 100,000 in 55-years-old women) and a
peak and plateau after 70 years (75 per 100,000) (4).
   Approximately 90% of OCs are sporadic with no
apparent hereditary risk factors. At least 10% of the total
are estimated to have a hereditary background due to
mutations in the BRCA genes (7). The majority will have
only one direct relative with OC with only 1% of all women
with OC having a more extensive family history. Women
with two or more family members have a 10-15% life-time
risk of developing OC, which is approximately ten times the
risk of women with no family history.
   Increased parity, oral contraceptive use, tubal ligation,
hysterectomy with ovarian conservation and breast-feeding
are associated with lower risk of OC. Risk is increased
among women who have used fertility drugs (8). However, it
is not yet clear whether this must be attributed to the drugs
or to specific forms of infertility that predispose to OC                 Figure 2. Correlation of stage and 5-year survival incidence in ovarian
themselves. The age at menarche, menopause and first live                 cancer (5).
birth, as well as the duration of oestrogen replacement
treatment are unrelated to the risk of OC (8). There are                 Rationale of screening for ovarian cancer. The dramatic
only few and conflicting data regarding the potential                    impact of cervical screening programmes in reducing
association between polycystic ovarian syndrome and OC,                  mortality from cervical cancer cannot be reproduced in OC.
but a significant increase in the risk does not seem likely (9).         According to the WHO definition, an effective screening test

                                            Kyrgiou et al: Ovarian Cancer Screening (Review)

Table I. High-risk population.

Family history related to high risk of developing ovarian cancer (OC) is characterised by the presence of either of the following criteria:
a. ≥2 first degree* relatives with OC
b. 1 first degree relative with OC and 1 first degree relative with breast cancer diagnosed under 50 years of age
c. 1 first degree relative with OC and 2 first degree relatives with breast cancer diagnosed under 60 years of age
d. 1 first degree relative with OC and 3 first degree relatives with colorectal cancer (at least one diagnosed under the age of 50)
e. 1 affected individual with one of the known ovarian cancer-predisposing genes
f. In cases of paternal transmision, where disease presents to second degree** relatives with an unaffected interverning male relative who has an
   affected sister, criteria a,b,and c are modified.

*A first-degree relative is a mother, sister or daughter.
**A second-degree relative is a grandmother, grand-daughter, aunt or niece.

must address an important health issue, for which there                      Screening in this high-risk population consists of annual
should be a pre-invasive or early invasive phase, with the                or more frequent screening with serum CA125 and
progression from early to late disease being clear.                       transvaginal ultrasound, from the age of 30 years until the
Additionally, it should combine both high sensitivity and                 completion of desired family size. After the completion of
specificity, should have low morbidity, the intervention should           childbearing and by the age of 40 years, women should be
have a significant impact on prognosis and, finally, the                  offered the choice of either prophylactic salpingo-
screening should be cost-effective (10).                                  oophorectomy (17-18) or continued screening. The increased
   Despite circumstantial evidence, there are no                          number of pre-menopausal women in this population is
premalignant lesions for epithelial OC and the extent to                  associated with increased FP rates. However, ovarian
which benign lesions may become malignant disease is not                  screening does not eliminate the risk for primary peritoneal
known (11). Histologically benign ovarian tumours have                    serous papillary carcinoma (18) that is considered a
proven to have some molecular and immunological                           phenotypic variant of familial OC and available techniques
characteristics of the malignant counterparts (12). The                   have not been proven to reliably detect it (19).
natural history of how epithelial OC develops is not known,
but indirect evidence suggests that the progression time may              Screening tests and their diagnostic accuracy. There is no
be as short as 2 years. (13).                                             uniform screening protocol for OC. A variety of tests have
   In order to be acceptable, an OCS strategy must achieve                been used, based on the combination of clinical
a minimum positive predictive value (PPV) of 10%, i.e., no                examination, imaging techniques and biochemical markers.
more than nine false positives (FP) for each true positive                   The bimanual pelvic examination.The bimanual pelvic
(TP). It follows that extremely high specificity (Sp) is a                examination has low sensitivity for detection of early stage
prerequisite (14). Additionally, more than one screening                  cancer, and its widespread application has not resulted in a
test can be used in combination, in parallel or sequentially              significant shift to the detection of earlier stage ovarian
to balance sensitivity (S) and specificity.                               cancer (20).
                                                                             Transvaginal (TVS) or transabdominal (TAS) sonography
Target populations. The majority of OCs is sporadic and                   assessment. The sensitivity of TVS approaches 100%.
occurs in the general population. In order to minimize the                However, it still is an insufficiently specific and rather
FP rate and unnecessary surgical intervention, screening has              expensive first-line screening method (21).
currently focused on certain target populations:                             Serum Cancer Antigen 125 (CA125). CA125, the most
a) Women over the age of 50 years. The life-time risk of                  extensively studied marker, is a high molecular weight
developing OC in these women reaches 1.3%                                 glycoprotein expressed by most epithelial ovarian cancers. Its
(approximately 1 in 70 women) and 3-4% if there is a first-               sensitivity is low, as the percentage of women with stage I
degree relative with the disease. According to current data,              disease with an elevated CA125 level (>35 units/ml) is only
screening surveillance is not recommended for this                        23 to 50% (22, 23). Benign conditions, such as peritoneal
population (15).                                                          irritation from endometriosis or inflammatory disease and
b) Women with a strong family history of breast and/or ovarian            uterine fibroids, pregnancy and menstruation, or even liver
cancer. The median age for diagnosis is 52 years in these                 disease and heart failure are associated with high titres (23-
women and the presence of BRCA1 mutation carries a 16%                    27). Yet, a CA125 level ≥30 U/ml in an asymptomatic post-
lifetime risk for developing ovarian cancer (95 % CI: 6-28%)              menopausal woman is associated with a 36-fold increased
(16) (Table I).                                                           risk of OC in the subsequent years (23). A more

                                             ANTICANCER RESEARCH 26: 4793-4802 (2006)

sophisticated approach to interpreting CA125 results use                    statistical values, such as S, SP, PPV, NPV and also in the
serial CA125 levels rather than a fixed CA125 cut-off level                 NNT and NNT (I) for each study was feasible and could
(35 U/ml) (23, 26). Serial CA125 levels associated with                     lead to a better understanding of the limitations and
malignancy tend to rise with time, while, they tend to remain               appropriate use of various screening modalities.
static or decrease with time in case of benign disease (28,
29). Mathematical modelling of individual CA125 trends                      Prospective studies using ultrasound strategy as primary
increases the accuracy of CA125 (26).                                       screening test
                                                                            a) Ultrasound (US) as the sole screening modality in both
Materials and Methods                                                       primary and secondary testing. This approach was used in 8
                                                                            studies (17, 31, 34-39) involving a total of 51,888 women
A detailed search of electronic literature databases was conducted,         with either a median age of 54 years or an age over 25 years
targeting reports dealing with the diagnostic accuracy of different         with a family history for OC. The number of women who
strategies in OCS. Recent developments in all the aspects of OCS            had to undergo surgery in order to identify 1 case of OC
and current beliefs were recorded from the present literature.
                                                                            varied from 10.6 to 65.2 and for stage I disease from 16.4 to
   Prospective studies on OCS of the general, low-risk
asymptomatic population which evaluated the various screening               65.2. The test had low PPV (0-9.5%) and high NPV (99.97-
strategies and provided data on their effectiveness [S, Sp, PPV,            99.99%). The specificity was high (almost 98.9%) and
negative predictive value (NPV), TP, FP, true negative (TN), false          sensitivity ranged from 81% to 88.9% (Table III).
negative (FN), stage at detection, number of patients needed to             b) The colour Doppler imaging (CDI) and US one-level
treat for every cancer detected (NNT) and every cancer detected             approach. Two studies tested this strategy in 6,377 post-
in stage I (NNT (I))] were identified.                                      menopausal women (40, 41). The test showed very good
   The identified studies were classified with regard to the strategy
                                                                            results as far as NNT (I) and NNT (3 to 9.5) were
applied in five broad categories. The first author, publication year,
journal, country and region, type of strategy used (primary and             concerned. NPV was high, reaching 99.85%, but PPV was
secondary test), threshold applied for positive screens/operations          low (10.5-33.3%). It appears that the Sp reached 99.9% and
and subsequent surgical investigation and the main characteristics          S was only 33.3% (Table III).
of the participants (age, menopause status, etc.) were recorded.            c) The use of US as a primary test plus another test. Four
Furthermore, the total sample size screened, the number of                  studies (32, 42-44) on a total of 58,472 post-menopausal
positive primary and secondary tests and the total number of                women applied this approach. One study (44) tested the use
positive screens, the TP, FP, TN, FN cases, as well as the number
                                                                            of the Fine Needle Aspiration (FNA) as a secondary test,
(and stage) of invasive carcinomas detected and the number of
borderline and granulosa tumours were also indicated.                       another the use of CDI (32) and the others evaluated the
   The sensitivity, specificity, positive and negative predictive value,    use of CA125 combined or not with TVS (42, 43). There
as well as number needed to treat for each cancer and for each              was discrepancy among the S in various studies (50-100%).
stage I cancer detected were recorded when provided or calculated           However, the Sp was approximately 99.5%, PPV 9.1-10.5%
where applicable. The prospective screening strategies’                     and NPV close to 100%. NNT ranged from 9.5 to 14.7 and
performance was calculated using standard parameters.                       NNT (1) from 10 to 19 (Table III).
   Due to the heterogeneity in inclusion criteria, recruitment,
characteristics of the patients, design, thresholds and policy applied
for positive screens, as well as due to the lack of data across             Prospective trials using CA125 as the primary test
different studies, the data derived from the included studies were          a) CA125. CA125 was the sole screening modality in one study
not integrated into a formal meta-analysis. Thus, combined                  (25) that included 5,550 women older than 40 years. This
sensitivity, specificity, positive and negative predictive value, as well   study showed very low PPV (3.4%), low S (50%), good Sp
as positive and negative likelihood ratios could not be obtained.           (96.9%) and high NNT (29.2) and NNT (I) (87.5) (Table III).
                                                                            b) The multimodal screening approach (MMS). This strategy
Results                                                                     includes the measurement of serum CA125 as the primary
                                                                            screening test and US (transvaginal or transabdominal) as
Twenty-two prospective studies evaluating the efficacy of                   secondary and was evaluated in 6 studies (23, 30, 45-48) in a
OCS using five different strategies were identified, including              total of 38,518 women. Sensitivity ranged from 37.5-100%,
four randomised controlled trials (RCTs) (30-33) and 18                     Sp from 98.5-99.9%, PPV from 6.3-26.8% and NPV from
prospective cohort studies (17, 23, 25, 34-48) (Table II). The              99.91-100%. The NNT was 3.7 to 16.5 and the NNT (I) 9.7
following studies, by van Nagel et al. (34), van Nagel et al.               to 16.5 (Table III).
(36) and DePriest et al. (35), as well as by Jacobs et al. (30),
Jacobs et al. (23) and Jacobs et al. (47) reported overlapping              Current trials. Two randomised controlled trials are currently
results of duplicate studies referring to same population.                  underway, designed to provide evidence and assess the
   The total effectiveness of each test could not be                        mortality benefit of OCS in the general population. The
calculated. However an estimation of the ranges in                          United Kingdom Collaborative Trial of Ovarian Cancer

                                                Kyrgiou et al: Ovarian Cancer Screening (Review)

Table II. Characteristics of the studies investigated.

Study                       Country     Population                          Participants Primary screening    Secondary screening

Buys et al. (33)            USA         55-74 yr with no Hx of cancer          39,115   CA125 (≥35),
                                                                                        TVS or both
Jacobs et al. (30) †        UK          ≥45 yr and post-menopausal             10,958   CA125 >30              TAS/TVS if
                                                                                                               CA125 >30 (3 screens)
Tabor et al. (31)           Denmark 46-65 yr                                      435   TVS                    repeat TVS (3-8 weeks)
Parkes et al. (32)          UK      50-64 yr of age                             2,953   TVS                    CDI if repeat TVS
                                                                                                               after 6 weeks (+)

Cohort studies
van Nagel et al. (34)*      USA         all ≥50 yr and ≥25 yr                  14,469   TVS                    repeat TVS
                                        with family Hx                                                         (4-6 weeks, mean 4 weeks)
Hayashi et al. (17)         Japan       post-menopausal ≥50 yr                 23,451   TVS                    repeat TVS (4-8 weeks)
DePriest et al. (35)*       USA         all >50 yr and >30 with family Hx       6,470   TVS                    repeat TVS
                                                                                                               (4-6 weeks, mean 2.29)
van Nagel et al. (36)*      USA         all ≥50 yr or >25 yr with family Hx   8,500     TVS                    repeat TVS (4-6 weeks)
Campbell et al. (37)        UK          >45 and <44 with family Hx            5,479     TAS annually-3 screens repeat TAS (3-6 weeks)
Millo et al. (38)           Italy       >45 yr or post-menopausal (mean 50 yr) 500      US (not specified)     repeat US
Goswamy et al. (39)         UK          39-78 yr, post-menopausal             1,084     TAS                    repeat TAS
Vuento et al. (41)          Finland     56-61 post-menopausal (mean 59 yr)    1,364     CDI [TVS               repeat sonogram
                                                                                        (96%)-TAS(4%)          (1-3 months later)
Kurjak et al. (40)          Croatia     ≥40 yr (56% post-menopausal)            5,013   TVS CDI                repeat TVS CDI
Einhorn et al. (25)         Sweden      ≥40 yr                                  5,550   CA125 ≥30 OR 35
Jacobs et al. (23)†         UK          ≥45 yr and post-menopausal             22,000   CA125 (≥30)            TAS if CA125(+)
Adonakis et al. (45)        Greece      ≥45 yr (mean 58y)                       2,000   CA125 (≥35) and        TVS
                                                                                        pelvic examination
Grover et al. (46)          Australia ≥40 yr or younger with                    2,550   CA125 (>35) and        TAS or/and TVS
                                      family history (mean 51 yr)                       pelvic examination
Jacobs et al. (47)†         UK        >45 yr and post-menopausal               22,000   CA125 (≥30)            TAS if CA125(+)
Jacobs et al. (48)          UK        >45 yr and post-menopausal                1,010   CA125 (>30) and        TAS if CA125 or pelvic
                                      (mean 54 yr)                                      pelvic examination     examination(+)
Sato et al. (42)            Japan     asymptomatic >30 yr                      51,550   TVS                    TVS+CA125
Holbert et al. (43)         USA       post-menopausal 30-89 yr                    428   TVS                    CA125 if TVS (+) and repeat
                                                                                                               TVS 1 and 3 months if CA125 (–)
Schincincaglia et al. (44) Italy        50-69 and post-menopausal               3,541   TAS                    FNA/biopsy

*,† Dublicate studies, TVS: transvaginal scan, CDI: colour Doppler imaging, TAS: transabdominal scan, FNA: fine needle aspiration, US:
ultrasound scan, Hx: history.

Screening (UKC-TOCS) was initiated in 2001. This study                       This two-arm randomized trial incorporates US and CA125
will involve 200,000 post-menopausal women and is designed                   using a combined strategy (CA125 and TVS) for 4 years and
to document the impact of screening on mortality using a                     CA125 alone for an additional 2 years in the study arm. Out
three-arm randomised design, to determine the physical                       of the 78,237 women enrolled in the trial, 39,115 were
morbidity of OCS, the resource implications of screening,                    randomised to the intervention/screening arm of which
the feasibility of population screening as reflected by the                  approximately 28,803 were finally eligible. Out of the 29
uptake of invitations and compliance rates with annual                       malignant neoplasms identified, 9 were tumours of low
screening and to establish a serum bank for future                           malignant potential. CA125 alone was abnormal in 24% and
assessment of novel tumour markers. Results are still                        TVS in 45% of the neoplasms detected, whereas both tests
awaited (14).                                                                were abnormal in 9 cases (31%). The PPV ranged from
   The NIH Prostate, Lung, Colorectal and Ovarian                            1.6% for the TVS screening method to 4% for the CA125
Screening Trial (PLCO) was conducted in the USA and                          approach and better results were achieved (PPV=26.5%) if
completed the enrolment of approximately 74,000 women at                     both tests were abnormal. The CA125 approach
10 centres in the USA between 1992 and 2001. Initial                         demonstrated better efficacy in detecting 1 case of OC
baseline screening data have been recently reported (33).                    (NNT=3.9 compared to NNT=24.3 for TVS screening) and

                                              ANTICANCER RESEARCH 26: 4793-4802 (2006)

Table III. Effectiveness of various screening modalities.

Study                              (+)screens/ TP           FN   FP      TN     Sensitivity Specificity   PPV       NPV       NNT     NNT(I)

Primary US – secondary US ± other test
    van Nagel et al. (34)*         180          17 (11)      4   163   14285      81.0%       98.9%       9.4%     99.97%      10.6     16.4
    Hayashi et al. (17)             95           6 (5)      NP   NP      NP        NA          NA          NA        NA        15.8      19
    DePriest et al. (35)*           90           6 (5)       1    84    6379      85.7%       98.7%       6.7%     99.98%      15.0     18.0
    van Nagel et al. (36)*         121           8 (6)       1   113    8378      88.9%       98.7%       6.6%     99.99%      15.1     20.2
    Tabor et al. (31)                9           0 (0)      NP     9     NP        NA          NA         0.0%       NA        NA       NA
    Campbell et al. (37)           326           5 (5)      NP   321     NP        NA          NA         1.5%       NA        65.2     65.2
    Millo et al. (38)                4           0 (0)      NP     4     NP        NA          NA         0.0%       NA        NA       NA
    Goswamy et al. (39)            NP            1(1)       NP   NP      NP        NA          NA          NA        NA        NA       NA

Primary color Doppler imaging – secondary US ± colour Doppler imaging
    Vuento et al. (41)               3     1 (1)    2        2      1359          33.3%       99.9%       33.3%    99.85%      3.0      3.0
    Kurjak et al. (40)               38    4 (4)   NP       34       NP            NA          NP         10,5%      NA        9.5      9.5
    Primary CA125 (only)
    Einhorn et al. (25)             175    6 (2)    6      169      5369          50.0%       96.9%       3.4%     99.89%      29.2     87.5

Primary CA125 – secondary US
    Jacobs et al. (30)†                 29       6 (3)      10   23    10919      37.5%       99.8%       20.7%     99.91%      4.8      9.7
    Jacobs et al. (23)†                 NP      49 (16)     NP   NP     NP         NA          NA          NA         NA       NA       NA
    Adonakis et al. (45)                33       2 (2)       0   30    1967      100.0%       98.5%       6.3%     100.00%     16.5     16.5
    Grover et al. (46)                  16         0         0    2     NP         NA          NA          NA         NA       NA       NA
    Jacobs et al. (47)†                 41       11 (3)      8   30    21951      57.9%       99.9%       26.8%     99.96%      3.7     13.7
    Jacobs et al. (48)                  13       1 (1)       0   12     997      100.0%       98.8%       7.7%     100.00%     13.0     13.0

Primary US – secondary other test
    Sato et al. (42)                    324     22 (17)      0   300     NP      100.0%        NA          6.8%      NA        14.7      19
    Parkes et al. (32)                  10       1 (1)       1    9     2942      50.0%       99.7%       10.0%    99.97%      10.0     10.0
    Holbert et al. (43)                  11      1 (1)      NP    10     NP        NA          NA          9.1%      NA        11.0     11.0
    Schincincaglia et al. (44)           19      2 (0)       0    17    3522     100.0%       99.5%       10.5%    100.00%      9.5     NA

Buys et al. (33)                                            NP           NP        NA          NA                    NA
   CA125                                 62     16(1)             46                                       4.0%                 3.9     62
   TVS                                  535     22 (9)           513                                       1.6%                24.3     59
   both                                 27      9 (1)            18                                       26.5%                  3      27

*,† Dublicate studies, US: ultrasound, TVS: transvaginal scan, TP: true positive, FN: false negative, TN: true negative, FN: false negaive,
PPV:positive predictive value, NPV: negative predictive value, NNT: number needed to treat, NNT(I): number needed to treat to detect 1 case of
stage I OC, NA: non applicable, NP: not provided. Numbers in parentheses correspond to stage I disease.

this number was even lower if both tests were positive                   identify a test with adequate sensitivity for OC that can
(NNT=3). Accordingly, 62 women had to undergo surgery                    detect the disease long before it causes symptoms. Moreover,
in order to identify 1 case of stage I disease if CA125 was              as surgery is thought to be significantly beneficial mainly for
abnormal, 59 if TVS was not reassuring, and 27 if both tests             stage I disease, it is essential to focus our screening policy
were abnormal. If only subjects in whom both tests were                  on identifying patients who are at this early stage.
abnormal were evaluated, 20 out of the 29 OCs (and 12 out                   Based on the overall data of the prospective screening
of the 20 invasive) would have been missed. Follow-up will               studies in post-menopausal women described above, the MMS
continue for 16 years from randomisation in order to assess              approach appears to be superior to other strategies, based on
the health status and cause of death (33).                               the number of women that would undergo surgery in order to
                                                                         detect 1 case of OC (3.7- 16.5) and stage I OC (9.7-16.5).
Discussion                                                                  The CDI screening approach seemed to achieve the best
                                                                         effectiveness, with a range of 3.0 to 9.5 women needed to be
The key to altering the outcome for OC is intervention early             treated to find both 1 case of OC and 1 case of OC at stage I
in the natural history of the disease. The challenge is to               disease. However, it could be argued that there are only two

                                        Kyrgiou et al: Ovarian Cancer Screening (Review)

small studies (6,377 participants) for the use of CDI as a         investigation. Other possible novel biomarkers, such as
screening test for OC, whereas the efficacy of the MMS was         osteopontin (53, 54), prostasin (55), Human Kallikrein 6
studied in 38,518 participants. Future involvement of power        (56-58) vascular endothelial growth factor (VEGF) (59, 60)
Doppler and three-dimensional US as secondary tests seems          and serum soluble epidermal growth factor receptor
promising, but further studies are necessary to assess its         (sEGFR/sErbB1) (61, 62) are also under investigation.
possible role on differentiating complex masses.
   The least efficient method for OCS appears to be the use        The technology for ovarian imaging is developing rapidly and
of CA125 as the sole screening modality. Because of its low        holds promise for future improvement in the morphological
sensitivity and specificity, the number of women that have to      and vascular characterisation and differentiation of complex
undergo surgery in order to detect 1 case of OC and stage I        ovarian masses.
OC is unacceptably high (29.2 and 87.5, respectively) (25).
   It seems that the screening modalities currently used for the   Conclusion
detection of OC and especially for the detection of the stage
I disease, are not as efficient as expected. Preliminary data      Overall, this review suggests that the MMS approach shows
form the PLCO study (33) suggest detection of 1 neoplasm           superior specificity and PPV, while US alone as primary
and 1 invasive epithelial cancer for 994 and 1,517 subjects        screening test increases sensitivity. MMS appears superior
screened, respectively, but the effect of repeated annual          regarding the number of women needed to treat for each
screens is still awaited and remains currently unknown.            case of cancer detected and the stage of the disease at
Further follow-up will hopefully provide valuable data.            detection. Evidence on the future use of CDI seems
                                                                   promising in the two studies using this strategy, but more
Management of a positive screening result. Almost 10-20            studies are needed in order to establish the superiority of
women with positive screening for OC undergo an                    CDI as a screening test.
operation for each case of cancer detected, which is usually          Preliminary baseline results of the PLCO trial do not
not in stage I. As a result, numerous women will be                suggest a need to revise the ovarian screening guideline of
overtreated and might suffer from iatrogenic morbidity. The        the US Preventive Services Task Force (14), which declares
surgical method with the lowest morbidity is preferred.            that ‘routine screening for ovarian cancer by ultrasound, the
Laparoscopic surgery decreases post-operative morbidity,           measurement of tumour markers, or pelvic examination is
but compromises outcomes with the theoretical danger of            not recommended’. Screening asymptomatic post-
dissemination of the disease. On the other hand, even              menopausal women for OC do not meet the WHO’s
multiple laparoscopic biopsies might not be able to exclude        effective screening criteria, as the benefits of screening in
OC, unless the sonographically suspicious ovary is removed         terms of lives saved and alleviation of pain and suffering do
laparoscopically.                                                  not appear to be outweighed by the social costs of
   The risk of malignancy index (RMI) (49), based on               unnecessary investigations and treatments (13).
menopausal status, ultrasonographic findings and serum                Screening issues were recently addressed by the European
CA125 level is a simple, easily applicable method in the           Commission in its recommendation published in December
primary evaluation of patients with adnexal masses. The            2003 that states: ‘No screening test is scientifically justified
RMI cut-off level of 200 shows sensitivity that reaches            to be offered to people with no symptoms in an organized
nearly 75%, with specificity little beyond 90% (50). It has        population-based program before it has been shown in RCTs
been suggested that when the RMI is less than 50, the              to decrease disease specific mortality’. The conclusive end-
procedure can be performed laparoscopically, when                  point that would provide the acceptable level of evidence for
between 50-200 with laparotomy and when above 200 the              public health policy formulation and the introduction of
patient should be referred to a cancer centre (51).                screening should be a reduction in mortality. It remains to
                                                                   be shown whether the detection of early stage cancer leads
Research in progress – future developments. While future           to increased survival rates (63).
development in biological markers or imaging techniques
might offer higher diagnostic accuracy, evidence, particularly     References
mortality benefit, is awaited from prospective randomised
population-based studies. Only preliminary baseline data           1 Harries M and Gore M: Part I: chemotherapy for epithelial
                                                                     ovarian cancer-treatment at first diagnosis. Lancet Oncol 3:
from the PLCO study have been published to date (33).
                                                                     529-536, 2002.
   Further evaluation of serum proteins appears promising          2 Greenlee RT, Hill-Harmon MB, Murray T and Thun M:
for the detection of markers that could aid in the early             Cancer statistics. CA Cancer J Clin 51: 15-36, 2001.
detection of this malignancy (52). Among those, plasma             3 American Cancer Society. Cancer Facts and Figures: 2000. New
lysophosphatidic acid is currently undergoing extensive              York: American Cancer society, 2000.

                                          ANTICANCER RESEARCH 26: 4793-4802 (2006)

4 Quinn M, Babb P, Brock A, Kirby L and Jones J: Cancer trends          21 Crump C, Martin W, McIntosh, Urban N, Anderson G and
   in England and Wales 1950-1999. National Statistics SMPS No.            Karlan B: Ovarian cancer tumor marker behaviour in
   66. Website: http://www.statistics.gov.uk/downloads/theme_health/       asymptomatic healthy women. Implications for screening. Cancer
   cancertrends_5099.pdf.                                                  Epidemiology, Biomarkers & Prevention 9: 1107-1111, 2000.
5 Ferlay J, Bray F, Pisani P and Parkin DM: GLOBOCAN 2000:              22 Mann WJ, Patsner B, Cohen H and Loesch M: Pre-operative
   Cancer Incidence, Mortality and Prevalence Worldwide,                   serum CA125 levels in patients with surgical stage I invasive
   Version 1.0. IARC CancerBase No. 5. Lyon, IARC Press, 2001.             ovarian adenocarcinoma. J Natl Cancer Inst 80(3): 208-209, 1988.
6 National Cancer Institute. SEER Cancer Statistics Review              23 Jacobs IJ, Skates S, Davies AP, Woolas RP, Jeyerajah A,
   1975-2000. Website: http://www.seer.cancer.gov/csr/1975_2000/           Weidemann P, Sibley K and Oram DH: Risk of diagnosis of
   results_merged/sect_20_ovary.pdf.                                       ovarian cancer after raised serum CA 125 concentration: a
7 Malander S, Ridderheim M, Masback A, Loman N,                            prospective cohort study. Br Med J 313: 1355-1358, 1996.
   Kristoffersson U, Olsson H, Nilbert M and Borg A: One in 10          24 Komai K and Nishida T: Tumor marker in ovarian cancer. Gan
   ovarian cancer patients carry germ line BRCA1 or BRCA2                  To Kagaku Ryoho 29(3): 481-486, 2002.
   mutations: results of a retrospective study in Southern Sweden.      25 Einhorn N, Sjovall K, Knapp RC, Hall P, Scully RE, Bast RC Jr
   Eur J Cancer 40(3): 422-428, 2004.                                      and Zurawski VR Jr: Prospective evaluation of serum CA125
8 Whittemore AS, Harris R and Itnyre J: Characteristics relating           levels for early detection of ovarian cancer. Obstetrics and
   to ovarian cancer risk: collaborative analysis of 12 US case-           Gynecology 80(1): 14-18, 1992.
   control studies. II. Invasive epithelial ovarian cancers in white    26 Skates SJ, Menon U, Mac Donald N, Rosenthal AN, Oram DH,
   women. Collaborative Ovarian Cancer Group. Am J Epidemiol               Knapp RC and Jacobs IJ: Calculation of the Risk of Ovarian
   136(10): 1184-1203, 1992.                                               Cancer from Serial CA125 values for preclinical detection in
9 Balen A: Polycystic ovary syndrome and cancer. Hum Reprod                postmenopausal women. J Clin Oncol 21: 206-210, 2003.
   update 7(6): 522-525, 2001.                                          27 D’Aloia A, Faggiano P, Aurigemma G, Bontempi L, Rugggeri
10 Wilson JMG and Jugner G: Principles and Practice of                     G, Metra M, Nodari S and Dei Cas L: Serum levels of
   Screening for Disease. Geneva: World Health Organization, 14-           carboydrate antigen 125 in patients with chronic heart failure:
   39, 1968.                                                               relation to clinical severity, hemodynamic and Doppler
11 Hoskins WJ: Histologic changes in the ovaries of cancer prone           echocardiographic abnormalities, and short-term prognosis. J
   women: an indication of premalignant change? J Natl Cancer              Am Coll Cardiol 41: 1805-1811, 2003.
   Inst 88(24): 1790-1791, 1996.                                        28 Skates SJ and Singer DE: Quantifying the potential benefit of
12 Puls LE, Powell DE, DePriest PD, Gallion HH, Hunter JE,                 CA 125 screening for ovarian cancer. J Clin Epidemio 44: 365-
   Kryscio RJ and van Nagell JR Jr: Transition from benign to              380, 1991.
   malignant epithelium in mucinous and serous ovarian                  29 Skates SJ, Xu FJ, Yu YH, Sjovall K, Einhorn N, Chang Y, Bast
   cystadenocarcinoma. Gynecol Oncol 47(1): 53-57, 1992.                   RC Jr and Knapp RC: Towards an optimal algorithm for
13 Fung MF, Bryson P, Johnston M and Chambers A: Cancer                    ovarian cancer screening with longitudinal tumor markers.
   Care Ontario Practice Guidelines Initiative Gynecology Cancer           Cancer 76(10 suppl): 2004-2010, 1995.
   Disease Site Group. Screening postmenopausal women for               30 Jacobs IJ, Skates SJ, MacDonald N, Menon U, Rosenthal AN,
   ovarian cancer: a systematic review. J Obstet Gynaecol Can              Davies AP, Woolas R, Jeyarajah AR, Sibley K, Lowe DG and
   26(8): 717-728, 2004.                                                   Oram DH: Screening for ovarian cancer: a pilot randomized
14 Jacobs IJ and Menon U: Progress and challenges in screening             controlled trial. Lancet 353: 1207-1210, 1999.
   for early detection of ovarian cancer. Mol Cell Proteomics 3(4):     31 Tabor A, Jensen FR, Bock JE and Hogdall CK: Feasibility
   355-366, 2004.                                                          study of a randomized trial of ovarian cancer screening. Journal
15 U.S. Preventive Services Task Force. Guide to Clinical Preventive       of Medical Screening 1: 215-219, 1994.
   Services, 2nd ed. Baltimore: Williams & Wilkins, 1996.               32 Parkes CA, Smith D, Wald NJ and Bourne TH: Feasibility
16 Hartge P, Struewing JP, Wacholder S, Brody LC and Tucker                study of a randomized trial of ovarian cancer screening among
   MA: The prevalence of common BRCA1 and BRCA2 mutations                  the general population. Journal of Medical Screening 1: 209-
   among Ashkenazi Jews. Am J Hum Genet 64(4): 963-970, 1999.              214, 1994.
17 Hayashi H, Yaginuma Y, Kitamura S, Saitou Y, Miyamoto T,             33 Buys SS, Partridge E, Greene MH, Prorok PC, Reding D, Riley
   Komori H, Wada K and Ishikawa M: Bilateral oophorectomy                 TL, Hartge P, Fagerstrom RM, Ragard LR, Chia D, Izmirlian
   in asymptomatic women over 50 years old selected by ovarian             G, Fouad M, Johnson CC and Gohagan JK: PLCO Project
   cancer screening. Gynecol Obstetr Invest 47: 58-64, 1999.               Team: Ovarian cancer Screening in the Prostate, Lung,
18 Swisher E: Prophylactic surgery and other strategies for                Colorectal and Ovarian (PLCO) cancer screening trial: findings
   reducing the risk for familial ovarian cancer. Curr Treat               from the initial screen of a randomized trial. Am J Obst
   Options Oncol 4: 105-110, 2003.                                         Gynecol 193: 1630-1639, 2005.
19 Karlan BY, Baldwin RL, Lopez-Luevanos E, Raffel LJ, Barbuto          34 van Nagell JR Jr, DePriest PD, Reedy MB, Gallion HH, Ueland
   D, Narod S and Platt LD: Peritoneal serous papillary carcinoma,         FR, Pavlik EJ and Kryscio RJ: The efficacy of transvaginal
   a phenotypic variant of familial ovarian cancer: implications for       sonographic screening in asymptomatic women at risk for ovarian
   ovarian cancer screening. Am J Obst Gynecol 180: 917-928, 1999.         cancer. Gynecologic Oncology 77: 350-356, 2000.
20 Smith HL and Oi RH: Detection of malignant ovarian                   35 De Priest PD, Gallion HH, Pavlik EJ, Kryscio RJ and van
   neoplasms: a review of the literature. I. Detection of the patient      Nagell JR Jr: Transvaginal sonography as a screening method
   at risk; clinical, radiological and cytological detection. Obstet       for the detection of early ovarian cancer. Gynecologic Oncology
   Gynecol Surv 39: 313-328, 1984.                                         65(3): 408-414, 1997.

                                          Kyrgiou et al: Ovarian Cancer Screening (Review)

36 van Nagell JR Jr, Gallion HH, Pavlik EJ and De Priest PD:          52 Bandera CA, Ye B and Mok SC: New technologies for the
   Ovarian cancer screening. Cancer 76(suppl. 10): 2086-2091, 1995.      identification of markers for early detection of ovarian cancer.
37 Campbell S, Bhan V, Royston P, Whitehead MI and Collins               Curr Opin Obstet Gynecol 15(1): 51-55, 2003.
   WP: Transabdominal ultrasound screening for early ovarian          53 Kim JH, Skates SJ, Uede T, Wong KK, Shorge JO, Feltmate
   cancer. BMJ 299(6712): 1363-1367, 1989.                               CM, Bekowitz RS, Carmer DW and Mok SC: Osteopontin as a
38 Millo R, Facca MC and Alberico S: Sonographic evaluation of           potential diagnostic biomarker for ovarian cancer. JAMA
   ovarian volume in postmenopausal women: a screening test for          287(13): 1671-1679, 2002.
   ovarian cancer? Clin Exper Obst Gynecol 16(2-3): 72-78, 1989.      54 Schorge JO, Drake RD, Lee H, Skates SJ, Rajanbalu R, Miller
39 Goswamy RK, Campbell S and Whitehead MI: Screening for                DS, Kim JH, Cramer DW, Berkowitz RS and Mok SC:
   ovarian cancer. Clin Obst Gynaecol 10: 621-643, 1983.                 Osteopontin as an adjunct to CA125 in detecting reccurent
40 Kurjak A, Shalan H, Kupesic S, Kosuta D, Sosic A, Benic S,            ovarian cancer. Clin Cancer Res 10(10): 3474-3478, 2004.
   Ilijas M, Jukic S and Predanic M: An attempt to screen             55 Mok SC, Chao J, Skates S, Wong K, Yiu gK, Muto MG,
   asymptomatic women for ovarian and endometrial cancer with            Berkowitz RS and Cramer DW: Prostasin, a potential serum
   transvaginal color and pulsed Doppler sonography. J of                marker for ovarian cancer: identification through microarray
   Ultrasound Medicine 13(4): 295-301, 1994.                             technology. J Natl Cancer Inst 93(19): 1437-1439, 2001.
41 Vuento MH, Pirhonen JP, Makinen JI, Laippala PJ, Gronroos          56 Ghosh MC, Grass L, Soosaipillai A, Sotiropoulou G and
   M and Salmi TA: Evaluation of ovarian findings in                     Diamandis EP: Human Kallikrein 6 degrades extracellular mtrix
   asymptomatic post-menopausal women with color Doppler                 proteins and may enhance the metastatic potential of tumor
   ultrasound. Cancer 76(7): 1214-1218, 1995.                            cells.Tumor Biol 25(4): 193-192, 2004.
42 Sato S, Yokoyama Y, Sakamoto T, Futagami M and Saito Y:            57 Ni X, Zhang W, Huang KC, Wang Y, Ng SK, Mok SC,
   Usefulness of mass screening for ovarian carcinoma using              Berkowitz RS and Ng SW: Characterisation of human kallikrein
   transvaginal sonography. Cancer 89(3): 582-588, 2000.                 6/ protease M expression in ovarian Cancer. Br J Cancer 91(4):
43 Holbert TR: Screening transvaginal ultrasonography of                 725-731, 2004.
   poatmenopausal women in a private office setting. Am J Obst        58 Diamandis EP, Scorilas A, Fracchioli S, Van Gramberen M, De
   Gynecol 170(6): 1699-1703, 1994.                                      Bruijn H, Henrik A, Soosaipillai A, Grass L, Yousef GM,
44 Schincincaglia P, Brondelli L, Cicognani A, Buzzi G, Orsini LF,       Stenman UH, Massobrio M, Van Der Zee AG,Vergote I and
   Bovicelli L, Jasonni VM and Bucchi L: A feasibility study of          Katsaros D: Human kallikrein 6 (hK6): a new potential serum
   ovarian cancer screening: does fine-needle aspiration improve         biomarker for diagnosis and prognosis of ovarian carcinoma J
   ultrasound specificity? Tumori 80(3): 181-187, 1994.                  Clin Oncol 21(6): 1035-1043, 2003.
45 Adonakis GL, Paraskevaidis E, Tsiga S, Seferiadis K and Lolis      59 Nishida N, Yano H, Komai K, Nishida T and Kojiro M: Vascular
   DE: A combined approach for the early detection of ovarian            endothelial growth factor C and vascular endothelial growth
   cancer in asymptomatic women. Eur J Obst Gynecol Reprod               factor Receptor 2 are related closely to the prognosis of patients
   Biol 65(2): 221-225, 1996.                                            with ovarian carcinoma. Cancer 101(6): 1364-1374, 2004.
46 Grover S, Quinn MA, Weiderman P, Koh H, Robinson HP,               60 Brustmann H: Vascular endothelial growth factor expression in
   Rome R and Cauchi M: Screening for ovarian cancer using               serous ovariancarcinoma: relationship with topoisomerase II
   serum CA 125 and vaginal examination: report on 2550 females.         alpha and prognosis. Gynecol Oncol 95(1): 16-22, 2004.
   Int J Gynecol Cancer 5: 291-295, 1995.                             61 Baron AT, Cora EM, Lafky JM, Boardman CH,Buenafe MC,
47 Jacobs IJ, Davies AP, Bridges J, Stabile I, Fay T, Lower A,           Rademaker A, Liu D, Fiscman DA, Podratz KC and Maihle
   Grudzinskas JG and Oram D: Prevalence screening for ovarian           NJ: Soluble epidermal growth factor receptor (sEGFR/sErbB1)
   cancer in postmenopausal women by CA 125 measurement and              as a potential risk, screening, and diagnostic serum biomarker
   ultrasonography. BMJ 306: 1030-1034, 1993.                            of epithelial ovarian cancer. Cancer Epidemiol Biomarkers Prev
48 Jacobs I, Stabile I, Bridges J, Kemsley P, Reynolds C,                12(2): 103-113, 2003.
   Grudzinskas J and Oram D: Multimodal approach to screening         62 Baron AT, Boardman CH, Lafky JM, Rademaker A, Liu D,
   for ovarian cancer. Lancet 1(8580): 268-271, 1988.                    Fischman DA, Podratz KC and Maihle NJ: Soluble epidermal
49 Jacobs I, Oram D, Fairbanks J, Turner J, Frost C and                  growth factor receptor (SEG-FR) and cancer antigen 125 (CA-
   Grudzinskas JG: A risk-of-malignancy index incorporating CA           125) as screening and diagnostic tests for epithelial ovarian
   125, ultrasound and menopausal status for the accurate                cancer Cancer Epidemiol Biomarkers Prev 14(2): 306-318, 2005.
   preoperative diagnosis of ovarian cancer. Br J Obstet Gynecol      63 Alexander-Sefre F, Menon U and Jacobs IJ: Ovarian cancer
   97: 922-929, 1990.                                                    screening. Hosp Med 63(4): 210-213, 2002.
50 Andersen ES, Knudsen A, Rix P and Johansen B: Risk of
   malignancy index in the preoperative evaluation of patients with
   adnexal masses. Gynecol Oncol 90: 109-112, 2003.
51 Ma S, Shen K and Lang J: A risk of malignancy index in                                                    Received March 21, 2006
   preoperative diagnosis of ovarian cancer. Chin Med J 116: 396-                                          Revised September 14, 2006
   399, 2003.                                                                                             Accepted September 19, 2006


To top