Annals of Oncology 17: 1234–1238, 2006

original article doi:10.1093/annonc/mdl120

Published online 9 June 2006

CA 125 half-life and CA 125 nadir during induction

chemotherapy are independent predictors of
epithelial ovarian cancer outcome: results of
a French multicentric study
J. M. Riedinger1*, J. Wafflart2, G. Ricolleau3, N. Eche4, H. Larbre5, J. P. Basuyau6, I. Dalifard7,
K. Hacene8 & M. F. Pichon9
1
Laboratoire de Biologie Médicale, Centre GF. Leclerc, Dijon, France; 2Laboratoire de Radioanalyse, Institut Bergonié, Bordeaux, France; 3Laboratoire d’ Oncobiologie,
Centre René Gauducheau, Nantes, France; 4Laboratoire de Biologie, Institut Claudius Regaud, Toulouse, France; 5Laboratoire de Biologie Clinique, Institut J Godinot,
Reims, France; 6Laboratoire de Biologie, Centre H Becquerel, Rouen, France; 7Laboratoire de Radioanalyse, Centre P. Papin, Angers, France; 8Service de Statistiques
Médicales, Centre René Huguenin, Saint-Cloud, France; 9Laboratoire d’Oncobiologie, Centre René Huguenin, Saint-Cloud, France

Received 13 July 2005; revised 22 March 2006; accepted 13 April 2006

Background: CA 125 assays enable treatment-response monitoring in ovarian cancer.

Patients and methods: A multicentric study of CA 125 kinetics under induction chemotherapy was performed
in 631 patients. CA 125 half-life was calculated by mono-compartmental logarithmic regression. Nadir CA 125
concentration and time to nadir were also studied. Survival analyses for disease-free survival (DFS) and overall
survival (OS) used univariate (Kaplan-Meier) and multivariate (Cox) models.
original
article

Results: For 553 stage IIC–IV patients, 459 (83.0%) relapsed and 444 (80.3%) died from cancer. Median (range)
follow up time was 32 months (2–214 months). Median (range) for CA 125 kinetics were: 263 kU/l (5–52000 kU/l)
before 1st course, 15.8 days (4.5–417.9 days) for CA 125 half-life, 16 kU/l (3–2610 kU/l) for nadir and 85 days
(0–361 days) for time to nadir. Pre-chemotherapy CA 125, its half-life, nadir concentration and time to nadir all had
a univariate prognostic value for DFS and OS (P < 0.0001). In Cox models, CA 125 half-life, residual tumour
(P < 0.0001 for both), nadir concentration (P = 0.0002) and stage (P = 0.0118) were the most powerful prognostic
factors for DFS. For OS, the significant variables were similar, with age ranking last (P = 0.0319).
Conclusion: Among well-established prognostic factors in ovarian cancers, CA 125 half-life and nadir concentration
bear a strong and independent prognostic value.
Key words: CA 125, ovarian cancer, survival

introduction CA 125 [2–7]. A first approach consisted in defining

In oncology, treatment responses need to be classified
according to standard definitions, such as the widely accepted
RECIST (Response Evaluation Criteria in Solid Tumours)
guideline [1]. However, in ovarian cancer patients without
measurable targets, responses cannot be assessed using the
RECIST criteria. The CA 125 tumour marker, elevated in 90%
of patients with advanced epithelial ovarian cancer, represents
an alternative to monitoring treatment responses.
Different prognostic and predictive criteria derived from
CA 125 variations under induction chemotherapy have been
investigated. Most of the previous studies of CA 125 under
chemotherapy included less than 100 patients and there is
still no agreement about the prognostic value of post-operative
*Correspondence to: Dr J. M. Riedinger, Laboratoire de Biologie Médicale,
Centre GF Leclerc, 1 rue du Pr Marion, 21034 Dijon cedex, France.
Tel: +33 03 80 73 75 10; Fax: +33 03 80 73 75 26; E-mail: jriedinger@dijon.fnclcc.fr
responses according to precise criteria for CA 125 decrease
[8–13]. Study of CA 125 half-life [2, 5, 7, 14–21], CA 125
return to normal range after 2 or 3 chemotherapy courses [5–7,
18–19, 22], nadir CA 125 concentrations [4, 23] and time to
reach the nadir [9] were also investigated. We performed
a multicentric study with a large cohort of ovarian cancer
patients during and after primary treatment to analyse the
long term prognostic value of the different criteria linked to
CA 125 kinetics.
patients and methods
patients
Six hundred and thirty-one patients with epithelial ovarian cancer from
eight French Cancer Centres were retrospectively studied. Their primary
treatment was performed between 1988 and 1996 and patient follow up
was censored at May 2005. Diagnostic and treatment procedures, identical

ª 2006 European Society for Medical Oncology

Annals of Oncology original article
in each Centre, followed standard methods included since 1997 in the Toulouse Centres for 199 patients, and ELSA CA 125 II
Standards Options and Recommendations of the French Fédération des (immunoradiometric assay, Cis Bio International, Gif sur Yvette,
Centres de Lutte Contre le Cancer [24]. Table 1 summarises patient France) was used for 327 patients monitored in Angers, Nantes, Reims,
characteristics. All patients underwent a primary laparotomy and received Saint-Cloud and Rouen Centres.
a first line of chemotherapy including platinum salts and All assays were run in duplicate with two levels of control sera in
cyclophosphamide. To ensure treatment homogeneity, patients treated each series. Samples showing more than 10% variation between
by taxanes were excluded from the study. duplicates were re-assayed. The threshold for CA 125 elevation indicated
Tumour staging followed the Fédération Internationale de Gynécologie by the manufacturers was 35 kU/l for all methods.
Obstétrique (FIGO) recommendations [25]. The number and location of
extra-abdominal metastatic sites were recorded. Histological type was
determined according to World Health Organisation classification and CA 125 kinetics
was grouped into serous, mucinous, endometroid, undifferentiated, clear Serum CA 125 was measured 1–4 days before the first chemotherapy
cell and of unknown histology [26]. Tumour grading was not included in the cycle. Mean interval between the date of surgery and the first
analyses because of insufficient reproducibility [24]. Residual tumour after chemotherapy course was 20 ± 7 days. The lowest CA 125 concentration
primary surgery was classified into three groups: absence, £ 1 cm or > 1 cm. reached during or after chemotherapy (nadir) and the time elapsed from
Induction chemotherapy started 2 to 4 weeks after surgery and patients start of chemotherapy to nadir were also recorded. CA 125 half-life
received 6 ± 1.9 (mean ± SD) chemotherapy courses. Treatment (T1/2) calculation, considered as an early indicator of tumour
evaluation was based on biological and clinical data recorded at each cycle chemo-sensitivity, was based on a mono-compartmental model and an
and on medical imaging. During and after chemotherapy, chest X-rays, exponential regression: T1/2 = Ln 2/s, where s is the slope of the
abdomino-pelvic scans, MRI or ultrasonography were performed regression line and Ln the natural logarithm. A minimum of three serial
according to identified targets. and aligned Ln CA 125 concentrations in the first six blood samples was
Serum CA 125 was measured prior to each cycle (on average each 3 weeks) required to enable regression line calculation. The regression was
using the same immunoassay during monitoring. considered significant only if its correlation coefficient r was ‡ 0.95,
allowing CA 125 half-life results to be taken into account. To ensure
unbiased calculations of CA 125 half-life, only cases with pre-chemotherapy
CA 125 assays CA 125 ‡ 100 kU/l were included in statistical analyses.
Assays were performed following manufacturer recommendations. All
Centres used immunoassays based on monoclonal Centocor/Fujirebio

antibodies but with different assay formats: IRMA-mat CA 125 II statistical methods
(immunoradiometric assay, DiaSorin, Antony, France) was used for Disease-free survival (DFS) was defined as the time elapsed from the date
105 patients from Bordeaux Centre, CPE CA 125 II (enzyme-immunoassay, of the first chemotherapy cycle to the date of clinically proven recurrence.
Cis Bio International, Gif sur Yvette, France) was used in Dijon and Overall survival (OS) was calculated from the date of the first chemotherapy
cycle to the date of cancer-related death. Patients without recurrences or
Table 1. Clinical and pathological characteristics of patients (n = 553) still alive were censored at study endpoint (May 30, 2005).
Since the post-operative CA 125 distribution was not Gaussian,
non-parametric statistics were used. The Chi-square test was used for
Age at diagnosis (years) Median 60 (range 24–84) comparison of observed frequencies. Receiver operating characteristics
FIGO stage curves (ROC) were used to define the thresholds of CA 125 kinetics
IIC 51 (9.2%) parameters in view of survival analyses.
III 21 (3.8%) Univariate survival analyses were performed using the Kaplan-Meier
IIIA 13 (2.4%) method and log-rank test [27]. Cox models were used for multivariate
IIIB 50 (9.0%) survival analyses [28]. For all statistical tests, the level of significance was
IIIC 322 (58.2%) P £ 0.05. Statistical softwares used were SAS v. 8.2 (Cary, NC, USA)
IV 96 (17.4%) and MedCalc v. 8.0 (Mariakerke, Belgium).
Histology
Serous 396 (71.6%)
Mucinous 33 (6.0%)
Endometrioid 49 (8.9%)
results
Undifferentiated 42 (7.6) patients
Clear cell 6 (1.0%)
Six hundred and thirty one patients were included in this
Unknown 27 (4.9%)
multicentric study. Thirty-three patients (5%) were removed
Residual
tumour mass
because of incomplete serial CA 125 measures, absence of
Absence 145 (26.2%) chemotherapy, treatment including taxanes, or deaths unrelated
£1 cm 139 (25.1%) to ovarian cancer. For prognostic studies, only stage IIC to IV
>1 cm 269 (48.7%) patients were included (n = 553). Their main characteristics
Follow-up time (months) Median 32 (range 2–214) are summarised in Table 1. Ascites was noted for 393 (62.3%)
patients. Eighty-four percent of the patients received
Recurrence 459 (83.0%)
chemotherapy including cyclophosphamide + platinum salts,
Endpoint status
16 % cyclophosphamide + platinum salts + anthracycline
Alive 109 (19.7%)
with a mean ± SD number of cycles of 6.1 ± 1.9 and 5.4 ±
Cancer specific deaths 444 (80.3%)
2.3 respectively.

Volume 17 | No. 8 | August 2006 doi:10.1093/annonc/mdl120 | 1235

original article
comparison of patients and CA 125 concentrations
in the different centres
Mean number of CA 125 determinations per patient was
12 ± 4. As three different immunoassays were used by the
participating centres, the homogeneity of patient
characteristics with regard to CA 125 results was tested.
Median CA 125 was 214 kU/l (range 5–52 000, n = 172) for
CPE CA 125 II, 230 kU/l (range 9–14 000, n = 278) for ELSA CA
125 II and 350 kU/l (range 25–21 200, n = 103) for IRMA-mat
CA 125 II (significant difference, P = 0.008). Pairwise
comparisons showed non significant differences between CPE
CA 125 II and ELSA CA 125 II methods but significant
differences with IRMA-mat CA 125 II. The IRMA method
was used by a single centre which included 94 of 103 (91.3%) of
stage IIIC or IV patients with 84.5% of residual tumour after
surgery, whereas the mean proportions were respectively
324 of 450 (72.0% stage IIIC or IV) and 321 of 450 (71.3%) with
residual tumour for the other methods. These proportions
are significantly different (P = 0.0001 for stage IIIC or IV and
P = 0.009 for residual tumour). It was thus considered that
the higher median CA 125 level obtained with IRMA-mat CA
125 II was due to the high proportion of advanced stages
and positive residual tumour, and consequently all CA
125 results were pooled for statistical analyses.
distribution of CA 125-related parameters during
ovarian cancer monitoring
For stage IIC–IV patients, median CA 125 measured before
first chemotherapy course was 263 kU/l, (range 5–52 000 kU/l),
median CA 125 half-life was 15.8 days (range 4.5–417.9 days),
median CA 125 nadir concentration was 16 kU/l (range 3–
2610 kU/l) and median time to nadir was 85 days (range 0–361
days). Median CA 125 before first chemotherapy course was
related to residual tumour mass: absence of residual tumour:
78 kU/l (range 5–2985 kU/l), residual tumour £ 1 cm: 165 kU/l
(range 18–1911 kU/l), residual tumour > 1 cm: 915 kU/l
(range 40–52000 kU/l) (P < 0.0001).
univariate survival analyses
Survival analyses were performed according to four criteria:
pre-chemotherapy CA 125, CA 125 half-life, CA 125 nadir and
Table 2. Results of Kaplan–Meier analysis (stage IIC-IV ovarian cancers)
Categories n Disease-free survival
Pre-chemotherapy CA 125 £230 kU/L 260
>230 kU/L 293
CA 125 half-life £14 days 211
>14 days 204
CA 125 nadir £20 kU/L 334
>20 kU/L 219
Time to nadir £72 days 225
>72 days 328
1236 | Riedinger et al.
Annals of Oncology
time to nadir. In order to delineate dichotomised subgroups for
Kaplan-Meier analysis, ROC curves for each parameter were
performed with relapse or death as classification variables.
With relapse as classification variable, area under ROC
curves (AUC) were 0. 680 (95% confidence interval (CI)
0.640–0.719) for pre-chemotherapy CA 125, 0.810 (95% CI
0.769–0.847) for CA 125 half-life, 0.731 (95% CI 0.692–0.767)
for CA 125 nadir and 0.677 (95% CI 0.636–0.716) for time
to nadir. With death as classification variable AUC and 95%
CI were 0.689 (0.648–0.727) for pre-chemotherapy CA 125,
0.770 (0.727–0.810) for CA 125 half-life, 0.721 (0.682–0.758)
for CA 125 nadir and 0.669 (0.628–0.708) for time to nadir.
Cut-offs obtained were the following: 230 kU/l
for pre-chemotherapy CA 125, 14 days for CA 125
half-time, 20 kU/l for CA 125 nadir and 72 days for time to
CA 125 nadir.
Among stages IIC–IV patients, 459 (83.0%) relapsed and
444 (80.3%) died from ovarian cancer. Median DFS was
19 months (range 1–214 months) and median OS was
32 months (range 2–214 months). The four parameters linked
to CA 125 kinetics had a strong prognostic value for DFS
and OS with P < 0.0001 (Table 2).
multivariate survival models
For the variable CA 125 half-time, only cases with validated
half-life calculation (i.e. pre-chemotherpy CA 125 ‡ 100 kU/l)
were included (n = 415). During follow up, 459 recurrences
and 444 deaths were recorded in the IIC–IV group, including
364 recurrences and 356 deaths in the subgroup with validated
half-life calculation. Clinico-pathological variables introduced
in the Cox models were: age, (continuous or dichotomised
with 50 years as cut-off), FIGO stage, residual tumour mass
and tumour histology in three groups (serous, mucinous or
undifferenciated). Parameters related to CA 125 kinetics were
introduced as continuous and as dichotomised variables. For
the whole population (n = 553), the models obtained were
identical for DFS and OS, with CA 125 nadir and residual
tumour ranking first (all with P < 0.0001) followed by age
as a continuous variable (P = 0.0236 for DFS and
P = 0.0019 for OS) and FIGO stage (P = 0.0236 for DFS
and P = 0.0072 for OS). Table 3 shows the results of Cox
models for DFS and OS in the subgroup with validated CA
125 half-life.
Overall survival
Median survival P value Median survival P value
estimate (months) estimate (months)
24.5 48.6
16.5 <0.0001 25.1 <0.0001
26.5 48.6
12.6 <0.0001 19.1 <0.0001
26.4 49.1
11.9 <0.0001 18.7 <0.0001
24.1 42.8
18.0 <0.0001 30.6 <0.0001
Volume 17 | No. 8 | August 2006
Annals of Oncology original article
Table 3. Results of Cox models for 415 patients with validated half-life

Variables b SE v2 P-value Categories HR 95% CI

Disease-free survival
CA 125 half-life 0.79645 0.13149 110.47 <0.0001 0: £14 days 1.000 –
1: >14 days 2.218 [1.714; 2.870]
Residual tumour 0.28394 0.08008 27.50 <0.0001 0: 0 1.000 –
1: £1 cm 1.328 [1.135; 1.554]
2: > 1 cm 1.765 [1.289; 2.415]
CA 125 nadir 0.46701 0.12910 13.52 0.0002 0: £20 kU/l 1.000 –
1: >20 kU/l 1.595 [1.239; 2.055]
FIGO stage 0.27846 0.11072 6.33 0.0118 0: IIC 1.000 –
1: III, IIIA, IIIB, IIIC 1.321 [1.063; 1.641]
2: IV [1.131; 2.694]
Overall survival
CA 125 half-life 0.71192 0.13006 101.60 <0.0001 0: £14 days 1.000 –
1: >14 days 2.038 [1.579; 2.630]
Residual tumour 0.26031 0.08119 23.86 <0.0001 0: 0 1.000 –
1: £1 cm 1.297 [1.106; 1.521]
2: > 1 cm 1.683 [1.224; 2.314]
CA 125 nadir 0.49894 0.12812 14.46 0.0001 0: £20 kU/l 1.000 –
1: >20 kU/l 1.647 [1.281; 2.117]
FIGO stage 0.27047 0.11310 5.88 0.0153 0: IIC 1.000 –
1: III, IIIA, IIIB, IIIC 1.311 [1.050; 1.636]
2: IV [1.103; 2.676]
Age 0.01121 0.00523 4.60 0.0319 years 1.011 [1.001; 1.022]

SE, standard error; HR, hazard ratio; CI, confidence interval.

Figure 1. Kaplan-Meier cumulative survival plot for OS according to Figure 2. Kaplan-Meier cumulative survival plot for OS according to CA
CA 125 half-life during induction chemotherapy in stage IIC–IV patients 125 nadir concentration during or after induction chemotherapy in stage
(n = 415). Cut-off 14 days, d £ 14 days, s > 14 days. IIC–IV patients (n = 553). Cut-off 20 kU/l, d £ 20 kU/l, s > 20 kU/l.

for this assumption. Rustin proposed another set of criteria

discussion
It is now widely accepted that the tumour marker CA 125 is
a predictive and prognostic factor in CA 125 positive ovarian
cancers. Different criteria derived from CA 125 kinetics during
primary treatment have been investigated [2, 5, 11, 17, 19].
When calculating CA 125 half-life most of the authors, except
Buller, used only two selected time points, i.e. before
chemotherapy and up to 3 months later. This implies that
the CA 125 decrease under chemotherapy always follows
a mono-compartmental model. Our results provide the basis
based on assigned proportions of CA 125 decrease with two or
three serial samples [10].
All objective criteria derived from CA 125 kinetics under
induction chemotherapy were compared in a large series of
patients with a long follow up time. Our results show that,
among other well-established prognostic factors in ovarian
cancers, two parameters bear a strong and independent
prognostic value: CA 125 half-life and CA 125 nadir
concentration. To be reliable, the CA 125 half-life should
be calculated from the start of chemotherapy on a minimum

Volume 17 | No. 8 | August 2006 doi:10.1093/annonc/mdl120 | 1237

original article
of three serial samples enabling the calculation of a regression
line and a correlation coefficient. This method is now widely
accessible through software provided with immunoassay
equipment or by using popular spreadsheets. Its only
drawback is when the pre-chemotherapy CA 125 is too
low (< 100 kU/l) to enable half-life calculations. However,
most of the ovarian cancers are detected in advanced
stages, 70% in stages III and IV, and show elevated CA
125 levels after primary surgery [6]. Furthermore, the CA
125 nadir, the second powerful criteria from our results,
is available even if CA 125 half-life cannot be ascertained.
It is noteworthy that the discriminant cut-off found for
CA 125 nadir (20 kU/l) for DFS and OS in stage IIC–IV
patients is significantly lower than the usual cut-off
proposed by manufacturers (35 kU/l). The latter was
established for untreated ovarian cancer patients
without taking into account patient menopausal status.
To conclude, CA 125 half life and nadir concentration are
powerful independent prognostic factors which warrants to be
compared to the classical prognostic factors in prospective
protocols of ovarian cancers treatment.
references
1. Therasse P, Susan G, Arbuck SG et al. New guidelines to evaluate the response
to treatment in solid tumors. J Natl Cancer lnst 2000; 92: 205–16.
2. Van Den Burg MEL, Lammes FB, Van Putten WLJ, Stoter G. Ovarian Cancer: the
prognostic value of the serum half-life of CA 125 during induction chemotherapy.
Gynecol Oncol 1988; 30: 307–312.
3. Makar AP, Kristensen GB, Kaern J et al. Prognostic value of pre- and
postoperative serum CA 125 levels in ovarian cancer: new aspects and
multivariate analysis. Obstet Gynecol 1992; 79: 1002–10.
4. Gard GB, Houghton CR. An assessment of the value of serum CA 125
measurements in the management of epithelial ovarian carcinoma. Gynecol
Oncol 1994; 53: 283–289.
5. Gadducci A, Zola P, Landoni F et al. Serum half-life of CA 125 during early
chemotherapy as an independent prognostic variable for patients with advanced
epithelial ovarian cancer: Results of a Multicentric Italian Study. Gynecol Oncol
1995; 58: 42–47.
6. Munstedt K, Krisch M, Sachsse S, Vahrson H. Serum CA 125 levels and survival
in advanced ovarian cancer. Arch Gynecol Obstet 1997; 259: 117–23.
7. Colakovic S, Lukic V, Mitrovic L et al. Prognostic value of CA125 kinetics and
half-life in advanced ovarian cancer. Int J Biol Markers 2000; 15: 147–52.
8. Makar AP, Kristensen GB, Bormer OP, Trope CG. Serum CA 125 level allows
early identification of nonresponders during induction chemotherapy. Gynecol
Oncol 1993; 49: 73–79.
9. Frasci G, Conforti S, Zullo F et al. A risk model for ovarian carcinoma
patients using CA 125. Time to normalization renders second-look laparotomy
redundant. Cancer 1996; 77: 1122–30.
1238 | Riedinger et al.
Annals of Oncology
10. Rustin GJS, Nelstrop AE, McClean P et al. Defining response of ovarian
carcinoma to initial chemotherapy according to serum CA 125. J Clin Oncol
1996; 14: 1545–1551.
11. Peters-Engl C Obermair A, Heinzl H et al. CA 125 regression after two completed
cycles of chemotherapy: lack of prediction for long-term survival in patients with
advanced ovarian cancer. Br J Cancer 1999; 81: 662–666.
12. Bridgewater JA, Nelstrop AE, Rustin GJS et al. Comparison of standard and
CA-125 response criteria in patients with epithelial ovarian cancer treated
with platinum or paclitaxel. J Clin Oncol 1999; 17: 501–508.
13. Rustin GJS. Use of CA-125 to assess response to new agents in ovarian cancer
trials. J Clin Oncol 2003; 21: 187s–193s.
14. Hawkins RE, Roberts K, Wiltshaw E et al. The prognostic significance of the half-
life of serum CA 125 in patients responding to chemotherapy for epithelial
ovarian carcinoma. Br J Obstet Gynaecol 1989; 96: 1395–9.
15. Hogberg T, Kagedal B. Serum half-life of the tumor marker CA 125 during
induction chemotherapy as a prognostic indicator for survival in ovarian
carcinoma. Acta Obstet Gynecol Scand 1990; 69: 423–9.
16. Markowska J, Kopczynski Z, Manys G, Szewierski Z. CA 125 in monitoring
chemotherapy of patients with ovarian cancer: early response to the treatment.
Neoplasma 1990; 37: 687–92.
17. Buller RE, Berman ML, Bloss JD et al. CA 125 regression: a model for epithelial
ovarian cancer response. Am J Obstetr Gynecol 1991; 165: 360–367.
18. Rosman M, Hayden CL, Thiel RP et al. Prognostic indicators for poor risk
epithelial ovarian carcinoma. Cancer 1994; 74: 1323–1328.
19. Buller RE, Vasilev S, DiSaia PJ. CA 125 kinetics: a cost-effective clinical tool
to evaluate clinical trial outcomes in the 1990s. Am J Obstetr Gynecol 1996;
174: 1241–125.
20. Riedinger JM, Barillot I, Coudert B et al. Prognostic value of CA 125 initial
half-life measured during first-line chemotherapy in 62 patients with epithelial
ovarian cancer stage III or IV. Bull Cancer 1996; 63: 654–663.
21. Gadducci A, Cosio S, Fanucchi A et al. The predictive and prognostic value of
serum CA 125 half-life during paclitaxel/platinum-based chemotherapy in
patients with advanced ovarian carcinoma. Gynecol Oncol 2004; 93: 131–6.
22. Ferrero JM, Largillier R, Ramaioli A et al. Prognostic value of early normalization
of CA 125 during chemotherapy in stages III and IV ovarian tumors. Bull Cancer
1997; 84: 722–728.
23. Crawford SM, Peace J. Does the nadir CA125 concentration predict a
long-term outcome after chemotherapy for carcinoma of the ovary? Ann Oncol
2005; 16(1): 47–50.
24. Standards Options et Recommandations. Ovarian Cancer. Kerbrat P, Lhomme C,
Fervers B et al. French National Fédération of cancer (FNCLCC). Br J Cancer
2001; 84 Suppl 2: 18–23.
25. International Federation of Gynaecology and Obstetrics (FIGO). Changes in
definitions of clinical staging for carcinoma of the cervix and ovary. Am J Obstet
Gynecol 1987; 156: 263–264.
26. World Health Organization Classification of Tumours. Pathology and Genetics of
Tumours of the Breast and Female Genital Organs. IARC Press: Lyon 2003.
27. Kaplan EL and Meier P. Nonparametric estimation from incomplete observations.
J Am Stat Assoc 1958; 53: 457–481.
28. Cox DR. Regression models and life tables (with discussion). J Royal Stat Soc
1972; B 34: 187–220.
Volume 17 | No. 8 | August 2006