Anda di halaman 1dari 6




Uji Diagnostik Risk of Ovarian Malignancy Algorithm (ROMA) dalam Menentukan

Keganasan Ovarium
Di RSUP dr. Mohammad Hoesin Palembang

Forbes D1, Agustiansyah P1, Sastradinata I1, Theodorus2

1 Departement of Obstetric and Gynecology
2 Medical Research Unit Faculty of Medicine Universitas Sriwijaya
Dr. Mohammad Hoesin Palembang Hospital
Abstract Abstrak
Background: Ovarian cancer was one of the leading Latar Belakang: Kanker ovarium menduduki urutan ke
causes of cancer for women and about 70% was diagnose enam terbanyak dari tumor ganas pada wanita dan
in advanced stage. Unfortunately, there is no methods for hampir 70% terdiagnosis pada stadium lanjut,
early screening available to diagnose ovarian malignancy sayangnya sampai sekarang tidak ada metode penapisan
preoperatively. diagnosis dini yang akurat untuk mendiagnosis
Objective: To compare diagnostic value of Risk of keganasan ovarium
ovarian malignancy algorithm (ROMA) to Tujuan: Membandingkan nilai diagnostik ROMA
histopathology (gold standard) in diagnose ovarian terhadap histopatologi (gold standard) dalam
malignancy. menentukan keganasan tumor ovarium.
Methods: Diagnostic test was performed at dr. Metode: Penelitian uji diagnostik dilakukan di RSUP dr.
Mohammad Hoesin Hospital Palembang during June Mohammad Hoesin Palembang selama periode Juni
2016 November 2016, 61 women with ovarian tumor 2016 November 2016, sebanyak 61 wanita dengan
was included in this study, 4 subject was excluded tumor ovarium dimasukkan sebagai subjek penelitian, 4
because of discrepancy in diagnosis intraoperatively. pasien dieksklusi karena perbedaan diagnosis saat
Data then analized with SPSS version 21.0 and Med-calc intraoperatif. Data kemudian dianalisis dengan
statistic. menggunakan software SPSS versi 21.0 dan Med-calc
Results: From 57 subject which fulfill the inclusion and statistic.
exclusion criteria. Patient was divided in two group Hasil: Dari 57 pasien yang memenuhi kriteria inklusi
which is pre and postmenonpause. Analysis with ROC dan eksklusi, Pasien dibagi menjadi dua kelompok yaitu
curve was performed, the ROMA optimal cutoff in this kelompok premenopause dan menopause. Dilakukan
study was 23,7% in premenopause and 48,15% in analisis dengan kurva ROC didapatkan cut off optimal
menopause. With the optimal cutoff, the sensitivity was ROMA pada penelitian ini yaitu 23,7% untuk kelompok
79,41% and specivicity was 75%, positive predictive premenopause dan 48,15% untuk kelompok menopause.
value wa 73,07% and negative predictive value 83,77% Dilakukan uji diagnostik, didapatkan sensitivitas 79,41%
with accuracy 76,92% in diagnosing ovarian malignancy. dan spesifisitas 75%, nilai duga positif adalah 73,07%
Compare to RMI-3, the sensitivity was 65,5% and dan nilai duga negative 83,77% dengan nilai akurasi
specivicity was 85,7% with accuracy 75,44%. 76,92% dalam mendiagnosa keganasan ovarium.
Conclusion: ROMA and RMI-3 test have no difference Dibandingkan dengan RMI-3, didapatkan nilai
in diagnostic value for diagnose ovarian malignancy. sensitifitas 65,5% dan spesifisitas 85,7% dengan nilai
Both of the test cannot be used as diagnostic tools in akurasi 75,44%
diagnose and predicting malignancy in ovarian tumor Simpulan: Pemeriksaan ROMA dan RMI-3 memiliki
Keywords: Ovarian cancer, HE4, CA125, Risk of nilai diagnostik yang hampir sama dalam mendiagnosa
ovarian malignancy algorithm/ ROMA, Risk of ovarian keganasan pada tumor ovarium dengan nilai akurasi
malignancy index/RMI yang hampir sama. Keduanya belum dapat dijadikan alat
Correspondence: Darlin Forbes, Obstetric and diagnostik dalam menegakkan atau memprediksi
Gynecology Department dr. Moh. Hoesin Hospital, keganasan ovarium.
Palembang. Email: Kata kunci: kanker ovarium, HE-4, CA-125, Risk of
ovarian malignancy algorithm, Risk of ovarian
malignancy index.
Korespondensi: Darlin Forbes, Departemen Obstetrik
dan Ginekologi RSUP dr. Moh. Hoesin, Palembang.

BACKGROUND et al. reported higher sensitivity and specificity
Ovarian cancer is the sixth most common in ROMA 4,5
malignancy in women after uterine cervical Karlsen et al. reported a high sensitivity
carcinoma, breast, colorectal, skin and (94.8%) and specificity (75%) results of ROMA
lymphoma cancer. Up to 70% of ovarian cancer in diagnosing ovarian cancer. Molina et al. also
is diagnosed at later stage and have spread into reported a better sensitivity (90.1%) and
upper abdominal cavity (Stage III) or wider specificity (87.1%) results of ROMA.6
(Stage IV) with 5 years survival rate at 15-20%, Risk of Malignancy Index (RMI) is one of
whereas survival rate at Stage I and Stage II are the mostly used method in identifying
predicted at lower rate of 90% and 70%.1,2 malignancy and considered as a simple method
Patients are mostly diagnosed at later stage as which uses menopause status, ultrasound and
early diagnosing tool is still not available. One CA125 level. Jacobs et al. obtained sensitivity of
of the mostly used tumor marker is cancer 85.4% and specificity of 96.9% using cut off
antigen (CA) 125. Up to now, CA 125 is the best 200. However, Andriata et al. obtained a
tumor marker available in diagnosing and different results in using the same RMI method:
monitoring ovarian cancer patients. However, sensitivity of 8.4% and specificity of 76.9%.7
CA 125 increases only in 80% of patients in late Anton et al. from Brazil did the same
stages and 50% of patients in early stages. About comparison of ROMA and RMI and no
20% of patients of early stages ovarian cancer significant difference was found between
have normal CA 125 values.2 diagnostic values of ROMA and RMI.8
Several biomarkers have been tested lately as Normal value of biomarkers such as CA125
alternatives or additional markers to differentiate and HE4 varies in different population. Pauler et
benign from malignant tumor. Human al. reported a difference in normal value of
Epididymis 4 (HE4) is a promising biomarker to CA125 in Caucasian and Asian. Several studies
be used.HE4, a glycoprotein, is over expressed have been done to find normal values of these
in ovarian cancer particularly 3 markers in different populations. This difference
Moore et al. designed Risk of Ovarian in normal values can alter the outcome of
Malignancy Algorithm (ROMA), using blood ROMA. Therefore, Karen et al. proposed
test algorithm, as a simpler biomarker compared different cut off values for different population
to RMI (risk malignancy index) that requires to accommodate this variation of normal values
ultrasonography.They reported significant in different population.10
increase in sensitivity and specificity when HE4 Based on outcomes and differences of
and CA125 are used in combination. In a further previous studies, we intended to do diagnostic
study done to compare ROMA and RMI, Moore test on ROMA and RMI-3

METHODS value, negative predictive value, likelihood ratio
Diagnostic tests and cross sectional design were were calculated by med-calc statistic.
used on 61 women with ovarian cancer and were RESULT
planned for operative procedure. This research a. Demographic
was conducted at Obstetrics and Gynecologic Demography of samples were depicted in table
Department of Dr. Mohammad Hoesin 1. Mean age of samples was 40,51 years old
Palembang Hospital from June 2016 to with majority of the patients were from pre-
November 2016. menopause group (56.1%) and higher tendency
Inclusion criteria used are all women must be found in multipara (43.9%). Malignant
diagnosed with ovarian cancer based on histopathology was found in 34 patients
anamnesis, physical examination and (42.1%), whereas borderline histopathology was
ultrasonography, they are planned for operative found in 5 patients (8.7%). Samples were further
procedures, and they are agree to participate in categorized to benign and malignant
this research by signing informed consent. histopathology group. Benign histopathology
Exclusion criteria used are all women diagnosed group was dominated by serous type (7 patients
with non-gynecologic malignancies, women on / 25%) followed by mucinous type.
pregnancy, women diagnosed with kidney Demography characteristics of samples were
failure or intraoperative mass of non-ovarian depicted in table 1. In malignant
origin. histopathology group, it was dominated by
After all inclusion criteria were met, data mucinous type (12 patients / 41.4%) followed by
were collected which consisted of identity, serous (30.9%).
gestation age, parity, education level, Tabel 1. Characteristic

occupation, smoking , contraception, physical Frequency

N %
examination, ultrasonography, post-operative 40,5116,32
Age , meanSD
CA125 and HE4. CA 125 and HE4 tests were Pre-menopause 32 56,1
Menopause 25 43,9
carried out by using architect reagent and Not Married 11 19,3
Primipara 21 36,8
histopathology results were masked. ROMA Multipara 25 43,9
score was calculated by software downloaded House wives 31 54,4
Enterpreneur 5 8,8
from Teacher 7 12,3
Civil Servant 2 3,5
Data were processed by SPSS 21.0 software. Student 12 21,1
Descriptive data were analyzed by SPSS. Cut off SD 13 22,8
SLTP 10 17,5
point value of ROMA was determined by SLTA 22 38,6
D3/S1 12 21,1
Receiving Operating Characteristic (ROC) Total 57 100

curve. Sensitivity, specificity, positive predictive

Table 2 showed that median of HE4 was In the pre-menopause group, optimal cut off
highest in the malignant histopathology group point of ROMA was obtained at 23.7% with
(460.1298.13). Highest average of HE4 was sensitivity of 72.72% and specificity of 76.19%,
found in the epithelium histopathologic type positive predictive value of 61.54% and negative
(324.4180.50). This conforms to the theory that predictive value of 84.21% as shown in the
there is overexpression in the epithelial cell. figure 2 below.
Highest HE4 value was observed at menopause
group (41296.96) which indicated that
menopause status attributed to the increase in
Table 2. Mean value of Human Epididymis Protein 4 (HE4)
Characteristic N Mean SD
Malignant 29 460,12 98,13
Benign 28 97,06 10,46

Jenis Sel Figure 2. ROC curve of ROMA in premenopausal group

Epitelial 50 324,41 80,50
Non epithelial 7 79,18 6,79 Diagnostic test was done with alternative cut
Status Menopause off. Samples with borderline histopathology
Menopause 25 412,89 96,96
Premenopause 32 179,33 43,27 was included as malignant group. Results
obtained were then compared to those of
In this study, alternative cut off point of standard cut off. Comparison made can be seen
ROMA was determined by using ROC curve. in Table 3 below
Patients were divided into two groups of pre- Table 3. Diagnostic value of ROMA with borderline included
Cut Off study Cut Off
menopause and menopause. For the menopause Standard
Sensitivity, % 72,41 82,7
group, optimal cut off point of ROMA was Specificity, % 75 64,2
PPV, % 75 70,5
obtained at 58.15% with sensitivity of 47.15%, NPV, % 72,41 78,2
Accuracy, % 73,68 73,68
specificity of 71.4%, positive predictive value of
86.6% and negative predictive value of 50% as If samples with borderline histopathology
were not included as malignant group, better
shown in figure 1 below.
diagnostic results were obtained. These can be
seen at Table 4 below.
Table 4. Diagnostic value of ROMA with borderline excluded
Cut Off this Cut Off
study Standard
Sensitivity, % 72,41 82,7
Specificity, % 75 64,2
PPV, % 75 70,5
NPV, % 72,41 78,2
Accuracy, % 73,68 73,68
Figure 1. ROC curve of ROMA on menopause group

Both ROMA results with and without borderline resulting in different diagnostic values. Hence,
histopathology were compared to RMI-3 with this study is to find an alternative cut off value
cut off > 200. RMI-3 have better sensitivity and to be compared with standard cut off value and
specificity compared to ROMA results without RMI which is often use as an diagnostic tool.7,10-
borderline histopathology included as shown in
Table 5 and 6 . In this research, by using ROC analysis,
Table 5. Diagnostic value of RMI-3 vs ROMA alternative vs alternative cut off values were obtain at 23,7%
ROMA standard with borderline included
for pre-menopause and 48.15% for menopause
Benign vs Malignant ROMA ROMA RMI-3
+ Borderline alternative standard (72.41% sensitivity, 75% specificity, 73.68%
Sensitivity, % 65,5 82,7 65,5
Specificity, % 85,7 64,2 85,7 accuracy) if patients with borderline
PPV, % 82,6 70,5 82,6
NPV, % 70,5 78,2 70,5
Accuracy, % 75,44 73,68 75,44
histopathology were included as malignant
group. Better results were obtained (79.41%
Table 6. Diagnostic value of RMI-3 vs ROMA alternative vs sensitivity, 75% specificity, 76.68% accuracy) if
ROMA standard with borderline excluded
patients with borderline histopathology were not
Benign vs Malignant ROMA ROMA RMI-3
alternative standard included as malignant group. Using standard cut
Sensitivity, % 79,41 91,67 70,83
Specificity, % 75 64,2 85,7
PPV, % 73,07 68,75 90.47
off, ROMA has better sensitivity (82.7%
NPV, % 83,77 90,00 77,41
Accuracy, % 76,92 76,92 78,84
including borderline; 91.67% if borderline
DISCUSSION RMI-3 diagnostic value was also improved if

Almost of 70% ovarian cancer was diagnosed at borderline histopathology was not included as

later stage with 5 years survival rate at about 15- malignant group: 70.83% for sensitivity; 85.7%

20 %, whereas survival rate at stage I and stage for specificity; and 78.84% for accuracy.

II were predicted at 90% and 70%. Sensitivity and specificity of this study have
ROMA, a test using combination of CA125, lower sensitivity and specificity compared to
HE4, and menopause status is an effective Moores. These differences can be attributed to
diagnostic tool to diagnose ovarian cancer. The different demographic data. Menopause patients
effectiveness of ROMA to be used as pre- were more dominant proportion in Moores
operative diagnosis tool in patients with pelvic (53.3%) and Van Gorps (53.2%)10-13
mass have been proven by several studies, In this research, median age was 40.51 years
though there are still doubts of it superiority old, which is close to Winartos median age of
because of variations in cut off values to 41 years old, and there was not much difference
diagnose malignancy. in proportion of menopause status. There were
Gorp et al. and Anton et al. have shown differences in the dominant type of tumor.
different cut off values for different population Mucinous type of ovarian carcinoma was more

dominant ( 41%) in this research compared to patients with a pelvic mass. Gynecol Oncol.
Moores, Molinas and Karlsens in which 5. Moore RG, Brown AK, MillerMC, Skates S,
serous type of ovarian cancer was more Allard WJ, Verch T, et al. The use of multiple
novel tumor biomarkers for the detection of
dominant (>75%).4,5,9 ovarian carcinoma in patients with a pelvic
mass. Gynecol Oncol. 2008;108:402-8.
RMI method was first designed by Jacobs et 6. Moore RG, McMeekin DS, Brown AK,
DiSilvestro P, Miller MC, Allard WJ, et al. A
al. by using ultrasonography combined with
novel multiple marker bioassay utilizing HE4
CA125 value. In this study, no difference was and CA125 for the prediction of ovarian cancer
in patients with a pelvic mass. Gynecol Oncol.
found between RMI and ROMA. However, 2009;112:40-6.
accuracy of RMI3 was the highest which is 7. Andriahta Z, Saleh AZ, Sastradinata I. Akurasi
Uji Diagnostik Risk of Malignancy Index dan
78.85% in the samples with borderline Indeks Novel dalam Memprediksi Keganasan
Ovarium. Thesis Departemen Obstetrik dan
histopathology not included as malignant.13 Ginekologi RSMH.
The shortcomings in this study were smaller 8. Anton C, Carvalho FM, Oliviera EI, Maciel
GAR, Baracat EC, Carvalho JP. A coparison of
samples compared to previous studies and no CA 125, HE4, risk of ovarian malignancy
algorithm (ROMA) and the risk of malignancy
FSH measurement done to differentiate index (RMI) for the classification of ovarian
menopause and pre-menopause samples masses. Clinic.2012:67:437-41
9. Molina R, Escudero JM, Auge JM, Filella X,
CONCLUSION Foj L, Torne A, et al. HE4 a novel tumour
marker for ovarian cancer: comparison with CA
Diagnostic value of ROMA with alternative cut 125 and ROMA algorithm in patients with
off had lower sensitivity compared to standard gynaecological diseases. Tumor Biol. 2011;
cut off but the accuracy was the same. Both of 10. Winarto H, Laihad JB, Nuranna L.
Modification of cut off values for HE4, CA 125,
the test cannot be used to predict or diagnose the risk of ovarian malignancy index and the
ovarian malignancy. risk of malignancy algorithm for ovarian cancer
detection in Jakarta Indonesia. Asian Pac J
REFERENCES Cancer Prev. 2014: 15:1949-53
1. Aziz MF. Gynecological cancer in Indonesia. J 11. Hamed EO, Ahmed H, Seedek OB,
Gynecol Oncol. 2009;20:1:8-10 Mohhammed AM, Alla AA, Ghaffar HM.
2. Tavasolli FA. Tumours of the ovary and Significance of HE4 estimation in comparison
peritoneum. in: Devilee P . WHO Classification with CA125 in diagnosis of ovarian cancer and
of umours pathology & genetics tumours of the assessment of treatment response. Diagnostic
breast and female genital organs. Lyon : IARC Pathol. 2013; 11:1-8.
Press.; 2003: 114-192. 12. Pitta DD, Sarian LO, Barreta A, Campos EA, et
3. Hellstrom I, Raycraft J, Hayden-Ledbetter M, all. Symptoms, CA125 and HE4 for the
Ledbetter JA, Schummer M, McIntosh M, et al. preoperative prediction of ovarian malignancy
The HE4 (WFDC2) protein is a biomarker for in Brazilian women with ovarian masses. BMC
ovarian carcinoma. Cancer Res. 2003;63:3695- Cancer. 2013;13:423:1-11.
700. 13. Chudezka A, Ploska AC, Menkiszak J,
4. Karlsen MA, Sandhu N, Hgdall C, Rzechula AS, Stjona A, Byra E, et al.
Christensen, et al. Evaluation of HE4, CA125, Preoperative diagnostic performance of ROMA
risk of ovarian malignancy algorithm (ROMA) (risk of ovarian malignancy algorithm) in
and risk of malignancy index (RMI) as relation to etiopathogenesis of epithelial ovarian
diagnostic tools of epithelial ovarian cancer in tumor. J Mol Biomark Diagn. 2013: S4-003