Introduction

Ovarian cancer is the eighth most common cancer in women worldwide and the leading cause of death among gynecological malignancies in developed countries. A major contributor to its high mortality is the difficulty of early detection: roughly 70% of ovarian cancers are diagnosed at an advanced stage (FIGO stage III or IV) because early-stage disease is typically asymptomatic or produces only vague, nonspecific symptoms. When a woman presents with an adnexal mass, the clinical challenge is to distinguish the relatively rare malignant tumour from the far more common benign conditions (functional cysts, endometriomas, mature cystic teratomas, cystadenomas, and fibromas) so that appropriate surgical planning and referral pathways can be initiated.

The Risk of Malignancy Index (RMI) was developed by Jacobs and colleagues in 1990 to address this triage dilemma. It is a simple, multiplicative scoring system that combines three readily available clinical parameters: ultrasound morphology score (U), menopausal status (M), and serum CA-125 level. The resulting product, RMI = U × M × CA-125, generates a continuous numeric value that stratifies women into risk categories for ovarian malignancy. At the widely adopted cut-off of 200, the RMI achieves a sensitivity of approximately 78-80% and a specificity of approximately 87-92% for detecting primary ovarian cancer.

The RMI has been endorsed by the National Institute for Health and Care Excellence (NICE) in clinical guideline CG122, by the Royal College of Obstetricians and Gynaecologists (RCOG), and by numerous national and international societies as the recommended first-line triage tool for adnexal masses. Its enduring utility lies in its simplicity, accessibility, and robust validation across diverse populations and clinical settings.

Historical Background and Development

Prior to the RMI, the evaluation of adnexal masses relied on a combination of clinical impression, imaging characteristics, and individual tumour markers, each assessed in isolation. No single parameter was sufficiently sensitive or specific on its own. Serum CA-125, discovered in the early 1980s, offered promise as an ovarian cancer biomarker but was limited by elevated values in numerous benign conditions (endometriosis, pelvic inflammatory disease, liver disease, heart failure) and by poor sensitivity for mucinous and early-stage tumours. Ultrasound could characterize mass morphology but suffered from inter-observer variability and overlapping features between benign and malignant lesions.

Jacobs and colleagues hypothesized that combining these imperfect parameters into a single composite score would yield superior discriminatory power. Their prospective study evaluated 143 women with adnexal masses undergoing surgical exploration at the King's College Hospital, London. By analyzing ultrasound features, menopausal status, and CA-125 levels in a multiplicative model, they demonstrated that the RMI outperformed any individual component in predicting malignancy. The multiplicative (rather than additive) design was deliberate: it amplifies the score dramatically when multiple risk indicators are simultaneously present, while keeping the score low when only one indicator is mildly abnormal.

Since its publication, the original RMI (now often referred to as RMI I) has been joined by modified versions (RMI II, RMI III, RMI IV) that adjust the weighting of individual components. However, multiple comparative studies have shown that no subsequent version consistently outperforms RMI I across all populations, and RMI I remains the most widely used and externally validated version. The NICE and RCOG recommendations are based on RMI I methodology.

The RMI Formula

The RMI is calculated as the product of three components:

RMI = U × M × CA-125

Where:

• U = Ultrasound morphology score (0, 1, or 3)
• M = Menopausal status multiplier (1 for premenopausal, 3 for postmenopausal)
• CA-125 = Serum CA-125 level in IU/mL (entered as the absolute value)

The product can range from 0 (when no ultrasound features are present, making U = 0) to several thousand or higher in postmenopausal women with suspicious ultrasound morphology and markedly elevated CA-125.

Component 1: Ultrasound Morphology Score (U)

The ultrasound score evaluates five morphological features that are associated with malignancy in adnexal masses:

Feature	Description
Multilocular cyst	A cystic mass with internal septations creating two or more compartments. Multiloculation suggests a complex cystic neoplasm rather than a simple functional cyst.
Solid areas	The presence of solid (echogenic, non-cystic) components within the mass. Solid areas raise concern because malignant ovarian tumours frequently contain solid tissue, whereas most benign cysts are purely cystic.
Metastases	Imaging evidence of metastatic disease, such as peritoneal implants, omental thickening ("omental cake"), or liver metastases. The presence of metastases essentially confirms advanced malignancy.
Ascites	Free fluid in the abdomen or pelvis beyond a physiological amount. While small amounts of free fluid are normal (particularly mid-cycle or in the cul-de-sac), moderate to large volumes of ascites in association with an adnexal mass are highly suspicious for malignancy.
Bilateral lesions	Masses present in both ovaries. Bilaterality increases the suspicion for malignancy (particularly metastatic disease to the ovaries, e.g., Krukenberg tumours) or primary epithelial ovarian cancer, which is bilateral in approximately 30-50% of advanced cases.

Ultrasound Score Assignment

The number of features present determines the ultrasound multiplier (U):

Features Present	U Score	Rationale
0 features	0	No suspicious morphological features. The mass appears simple and benign on ultrasound. Because U = 0, the entire RMI product becomes 0 regardless of menopausal status or CA-125 level. This reflects the very low malignancy risk of a sonographically simple cyst.
1 feature	1	A single suspicious feature is present. The score of 1 means the RMI equals M × CA-125. One isolated feature (e.g., a single septation or small amount of ascites) confers a modest but non-negligible risk.
2 or more features	3	Multiple suspicious features are present, substantially increasing the morphological suspicion for malignancy. The multiplier of 3 amplifies the RMI to reflect the higher pretest probability associated with complex, heterogeneous masses.

The binary counting approach (rather than a weighted scoring of individual features) was chosen for simplicity and reproducibility. It requires only that the examiner identifies the presence or absence of each feature, minimizing subjective judgment and inter-observer variability.

Component 2: Menopausal Status (M)

Menopausal status is the second multiplier in the RMI equation:

Status	M Score
Premenopausal	1
Postmenopausal	3

Menopause is defined in the RMI as amenorrhea for more than one year, or age 50 years or older in women who have had a hysterectomy.

The rationale for this multiplier is epidemiological. The incidence of ovarian cancer rises steeply with age, with more than 80% of epithelial ovarian cancers occurring in women over 50. In premenopausal women, adnexal masses are far more commonly benign (functional cysts, endometriomas, mature teratomas). In postmenopausal women, the ovaries should be quiescent, and any new adnexal mass carries a substantially higher pretest probability of malignancy. The multiplier of 3 for postmenopausal status reflects this threefold increase in malignancy risk.

Impact on the Formula

A postmenopausal woman with two or more ultrasound features (U = 3) and a CA-125 of 100 IU/mL has an RMI of 3 × 3 × 100 = 900, well above the high-risk threshold. The same ultrasound and CA-125 findings in a premenopausal woman yield an RMI of 3 × 1 × 100 = 300, which still exceeds the cut-off but is substantially lower. This differential weighting appropriately reflects the different pretest probabilities in the two populations.

Component 3: Serum CA-125

Cancer Antigen 125 (CA-125) is a glycoprotein encoded by the MUC16 gene, expressed on the surface of ovarian epithelial cells and released into the bloodstream when these cells proliferate or when the peritoneal mesothelium is irritated. CA-125 is the most widely used serum biomarker in ovarian cancer, and its absolute value in IU/mL is entered directly into the RMI formula.

Normal Range and Interpretation

The standard upper limit of normal for CA-125 is 35 IU/mL. Values above this threshold are found in approximately 80% of patients with advanced-stage epithelial ovarian cancer and approximately 50% of patients with early-stage (FIGO I) disease. However, CA-125 has several important limitations as a standalone marker:

• Elevated in benign conditions: Endometriosis, uterine fibroids, pelvic inflammatory disease, liver cirrhosis, heart failure, pleuritis, and peritonitis can all cause CA-125 elevations, sometimes to several hundred IU/mL.
• Elevated in other cancers: Pancreatic, breast, lung, gastric, and colon cancers can all produce CA-125 elevation.
• Normal in mucinous tumours: Mucinous ovarian cancers (10-15% of epithelial ovarian cancers) frequently have normal CA-125 levels. CEA or CA 19-9 may be more informative in these cases.
• Normal in early-stage disease: Approximately 50% of stage I ovarian cancers present with a normal CA-125.
• Physiological variation: CA-125 fluctuates with the menstrual cycle (higher during menstruation) and is elevated during pregnancy, particularly in the first trimester.

By incorporating CA-125 into the multiplicative RMI formula alongside ultrasound and menopausal status, the impact of false-positive CA-125 elevations is mitigated. A premenopausal woman with endometriosis may have a CA-125 of 80 IU/mL, but if her ultrasound shows a typical endometrioma with no other suspicious features (U = 1, M = 1), her RMI is only 1 × 1 × 80 = 80, below the high-risk threshold. The same CA-125 level in a postmenopausal woman with a complex mass (U = 3, M = 3) produces an RMI of 3 × 3 × 80 = 720, appropriately triggering referral.

Risk Stratification and Interpretation

The RMI produces a continuous numeric score that is stratified into risk categories to guide clinical management:

RMI Value	Risk Category	Recommended Action
< 25	Low Risk	Management in a general gynecology setting is appropriate. Conservative follow-up with interval ultrasound may be considered if the mass appears benign. Clinical correlation should still guide decisions.
25 - 200	Moderate Risk	Further evaluation warranted. Consider repeat imaging, additional tumour markers (HE4, ROMA), or specialist assessment. Surgery may be performed in a unit with access to gynecologic oncology support.
> 200	High Risk	High probability of malignancy. NICE and RCOG recommend referral to a specialist gynecologic oncology centre. Pre-operative staging workup (CT chest/abdomen/pelvis, further laboratory evaluation) is typically indicated.

The Cut-off of 200

The RMI cut-off of 200 is the most widely adopted and externally validated threshold. At this value, the original study reported a sensitivity of 85.4% and specificity of 96.9%. Subsequent meta-analyses across multiple populations have settled on pooled estimates of approximately 78-80% sensitivity and 87-92% specificity. This means that roughly 4 out of 5 malignant adnexal masses are correctly identified as high risk, while fewer than 1 in 10 benign masses are incorrectly flagged.

The Cut-off of 25

The lower cut-off of 25 is used to define the low-risk category. At this threshold, the negative predictive value is very high (>98%), meaning that women with an RMI below 25 have an extremely low probability of malignancy. This allows for confident conservative management without surgical intervention in many cases.

Diagnostic Performance

The RMI has been one of the most extensively validated clinical scoring systems in gynecologic oncology. Key performance data include:

Original Derivation (Jacobs et al., 1990)

• Study population: 143 women with adnexal masses undergoing surgery
• At cut-off 200: Sensitivity 85.4%, Specificity 96.9%
• Positive predictive value: 89.5%
• Correctly classified 81% of ovarian cancers as high risk

External Validation Meta-analyses

A systematic review encompassing over 20,000 patients across more than 100 studies has demonstrated the following pooled performance metrics for the RMI at a cut-off of 200:

• Pooled sensitivity: 78% (95% CI: 71-85%)
• Pooled specificity: 87% (95% CI: 83-91%)
• Positive likelihood ratio: ~6.0
• Negative likelihood ratio: ~0.25
• Area under the ROC curve: 0.87-0.92

These figures place the RMI among the best-performing simple clinical prediction tools in oncology. The positive likelihood ratio of 6.0 means that a score above 200 is approximately six times more likely to be observed in a woman with ovarian cancer than in a woman with a benign mass.

Epidemiology of Ovarian Cancer

Understanding the epidemiology of ovarian cancer provides essential context for interpreting the RMI and counseling patients about their risk:

• Global incidence: Approximately 314,000 new cases annually worldwide, making it the eighth most common cancer in women.
• Mortality: Approximately 207,000 deaths annually, the highest case-fatality ratio of any gynecologic malignancy.
• Age distribution: Median age at diagnosis is 63 years. Incidence increases with age, peaking in the seventh and eighth decades. Over 80% of cases occur in women aged 50 and older.
• Lifetime risk: Approximately 1.3% (1 in 78) in the general population, rising to 15-40% in women with BRCA1 mutations and 10-20% in BRCA2 carriers.
• Five-year survival: Overall approximately 47%. When diagnosed at stage I, five-year survival exceeds 90%. When diagnosed at stage III-IV (the majority), five-year survival falls to 20-35%.
• Histological subtypes: High-grade serous carcinoma is the most common (70%), followed by endometrioid (10%), clear cell (10%), mucinous (3-5%), and low-grade serous (5%).

Pathophysiology and Biology of Ovarian Cancer

The biology of ovarian cancer informs why the RMI components are diagnostically meaningful:

The Dualistic Model of Ovarian Cancer

Modern understanding divides epithelial ovarian cancers into two broad groups:

• Type I tumours: Low-grade serous, endometrioid, clear cell, and mucinous carcinomas. These tend to be confined to the ovary at diagnosis (early stage), grow slowly, and arise from well-defined precursor lesions (borderline tumours, endometriosis). They frequently present as large, complex, sometimes bilateral adnexal masses with moderate CA-125 elevation.
• Type II tumours: High-grade serous carcinomas, the most common and most lethal subtype. Recent evidence suggests that many high-grade serous carcinomas originate in the fallopian tube fimbriae (serous tubal intraepithelial carcinoma, STIC) rather than the ovarian surface epithelium. They are characterized by TP53 mutations (>96%), rapid growth, early peritoneal dissemination, and marked CA-125 elevation. These tumours frequently present at advanced stage with ascites, bilateral ovarian involvement, and omental or peritoneal metastases, all features captured by the RMI ultrasound score.

CA-125 in Tumour Biology

CA-125 (MUC16) is a transmembrane mucin glycoprotein expressed on the surface of Mullerian-derived epithelium (ovarian surface, fallopian tube, endometrium, endocervix) and the mesothelial lining of the peritoneum, pleura, and pericardium. In malignancy, proteolytic cleavage of the extracellular domain releases CA-125 into the bloodstream at levels proportional to tumour burden and peritoneal involvement. This is why CA-125 is most dramatically elevated in advanced-stage serous carcinomas with widespread peritoneal disease, and least elevated in encapsulated, early-stage mucinous tumours.

The Five Ultrasound Features in Clinical Detail

Multilocular Cyst

A multilocular cyst contains internal septations that divide it into two or more compartments. On ultrasound, thin regular septations (< 3 mm) are generally associated with benign cystadenomas or hydrosalpinx. Thick, irregular septations (> 3 mm) with solid papillary projections along the septa are more concerning for malignancy. In the RMI, any septation classifying the mass as multilocular is counted as a positive feature, regardless of septal thickness. This binary approach favors sensitivity over specificity.

Solid Areas

The presence of solid components within an adnexal mass is one of the strongest ultrasound predictors of malignancy. While some benign masses contain solid tissue (mature teratomas with Rokitansky nodules, fibromas, thecomas), the combination of solid and cystic areas in a mass of epithelial origin raises significant concern. Color Doppler assessment of solid components can add information: low-resistance, high-velocity arterial flow within solid areas (indicating neovascularization) further increases the suspicion for malignancy.

Metastases

Imaging evidence of metastatic spread, including peritoneal deposits, omental infiltration ("omental cake"), liver surface metastases, or malignant lymphadenopathy, is the most specific ultrasound feature for malignancy. When metastases are identified, the diagnosis of cancer is virtually certain, and the RMI primarily serves to formalize the referral pathway rather than add diagnostic information.

Ascites

Ascites in association with an adnexal mass is a red flag for malignancy, particularly advanced epithelial ovarian cancer. The pathophysiology of malignant ascites involves increased peritoneal vascular permeability mediated by vascular endothelial growth factor (VEGF) secreted by tumour cells, combined with lymphatic obstruction by peritoneal implants. However, clinicians must distinguish malignant ascites from physiological free fluid (small amounts in the pouch of Douglas during mid-cycle or early pregnancy) and ascites from non-malignant causes (liver disease, heart failure, nephrotic syndrome).

Bilateral Lesions

Bilaterality of adnexal masses increases the suspicion for malignancy. Approximately 30-50% of advanced epithelial ovarian cancers involve both ovaries. Bilateral masses also raise the possibility of metastatic disease to the ovaries (Krukenberg tumours from gastric or colorectal primary, bilateral endometrioid carcinomas in Lynch syndrome). However, bilateral benign masses are not uncommon: bilateral endometriomas, bilateral dermoid cysts, and bilateral functional cysts all occur with some frequency. The RMI counts bilaterality as one of the five features without attempting to determine its etiology.

RMI Versions: I, II, III, and IV

Following the original RMI (RMI I), several modifications have been proposed:

Version	Year	U Score (0 features / 1 feature / ≥2 features)	M Score (Pre / Post)	CA-125
RMI I	1990	0 / 1 / 3	1 / 3	Absolute value
RMI II	1996	1 / 1 / 4	1 / 4	Absolute value
RMI III	1999	1 / 1 / 3	1 / 3	Absolute value
RMI IV	2009	1 / 1 / 4	1 / 4	Absolute value × tumour size score

RMI II (Tingulstad et al., 1996) differs by assigning U = 1 (instead of 0) when no ultrasound features are present, and by using a higher U multiplier of 4 (instead of 3) for two or more features. The menopausal multiplier is also increased to 4. RMI II tends to produce higher absolute values but uses the same cut-off of 200.

RMI III (Tingulstad et al., 1999) is identical to RMI I in its U and M scoring but differs in that it assigns U = 1 when no features are present (instead of 0). This means the RMI can never be zero; even a woman with no ultrasound features, premenopausal status, and a low CA-125 will have a non-zero score.

RMI IV (Yamamoto et al., 2009) adds a fourth component, tumour size (S), to the formula, making it RMI = U × M × CA-125 × S. The tumour size score is 1 if the largest diameter is < 7 cm and 2 if it is ≥ 7 cm.

Comparative studies have generally found that no single version consistently outperforms the others across all populations and clinical contexts. RMI I remains the most widely used and best-validated version, and it is the version recommended by NICE and RCOG.

Differential Diagnosis of Adnexal Masses

The RMI must be interpreted in the context of the broad differential diagnosis of adnexal masses. Not all high-RMI scores represent primary ovarian cancer, and not all ovarian cancers produce high RMI values.

Benign Ovarian Masses

• Functional cysts (follicular, corpus luteum): The most common adnexal masses in premenopausal women. They are typically simple, unilocular, thin-walled, and resolve spontaneously within 1-3 menstrual cycles. RMI is typically very low or zero.
• Endometriomas: "Chocolate cysts" containing old blood with a characteristic ground-glass echogenicity on ultrasound. They can be bilateral and may mildly elevate CA-125. The RMI is usually below 200 but may enter the moderate range if CA-125 is significantly elevated.
• Mature cystic teratomas (dermoid cysts): Contain ectodermal, mesodermal, and endodermal elements. They have a highly characteristic ultrasound appearance (fat-fluid levels, Rokitansky nodule, "tip of the iceberg" sign) but may be scored as having solid areas, potentially raising the RMI.
• Serous cystadenomas: Benign epithelial tumours that can be large, multilocular, and sometimes bilateral. They may produce a moderate RMI score based on ultrasound morphology alone.
• Fibromas/Thecomas: Solid ovarian tumours that can be mistaken for malignancy on ultrasound due to their solid nature. They are associated with Meigs syndrome (ascites and pleural effusion), which can produce a very high RMI despite the benign histology.

Malignant Ovarian Masses

• Epithelial ovarian carcinoma: The most common ovarian malignancy (90% of cases). Subtypes include high-grade serous (most common), endometrioid, clear cell, mucinous, and low-grade serous. High-grade serous carcinomas typically produce the highest RMI values due to bilateral disease, solid areas, ascites, and markedly elevated CA-125.
• Borderline (low malignant potential) tumours: Account for 10-15% of epithelial ovarian neoplasms. They have an excellent prognosis and may produce intermediate RMI values. CA-125 is often only mildly elevated, and ultrasound may show complex cystic morphology without solid components or ascites.
• Germ cell tumours: Occur primarily in young women (teens and twenties). They include dysgerminomas, immature teratomas, yolk sac tumours, and choriocarcinomas. CA-125 is often normal; alternative markers (AFP, beta-hCG, LDH) are more useful. The RMI may underperform in this subgroup.
• Sex cord-stromal tumours: Include granulosa cell tumours, Sertoli-Leydig cell tumours, and fibromas. CA-125 is often normal; inhibin B and anti-Mullerian hormone (AMH) are more specific markers. The RMI may not capture these tumours reliably.
• Metastatic tumours to the ovary: Krukenberg tumours (typically from gastric or colorectal primary), breast cancer metastases, and lymphoma can present as ovarian masses. They are often bilateral and may elevate CA-125.

Role of Imaging in Conjunction with the RMI

Transvaginal Ultrasound (TVUS)

TVUS is the primary imaging modality for evaluating adnexal masses and provides the ultrasound data used in the RMI. In expert hands, TVUS can characterize the morphology of adnexal masses with high accuracy. The International Ovarian Tumor Analysis (IOTA) group has developed standardized terminology and classification rules ("simple rules," ADNEX model) that have been shown to outperform the RMI in some studies. However, these approaches require greater expertise in pattern recognition and are not as widely implemented as the RMI.

MRI Pelvis

Pelvic MRI is the most accurate single imaging modality for characterizing adnexal masses, with sensitivity and specificity for malignancy both exceeding 90% in experienced centers. MRI can distinguish fat (teratomas), blood products (endometriomas), and fibrous tissue (fibromas) with high confidence using T1, T2, fat-saturated, and diffusion-weighted sequences. For women with intermediate RMI scores (25-200) or sonographically indeterminate masses, MRI can provide additional characterization that influences management decisions.

CT Abdomen/Pelvis

CT is not the primary tool for characterizing adnexal masses (lower soft-tissue contrast than MRI) but is essential for pre-operative staging in women with suspected ovarian cancer (RMI > 200). CT of the chest, abdomen, and pelvis evaluates peritoneal disease burden, omental involvement, lymphadenopathy, and distant metastases to guide surgical planning and assess resectability.

PET-CT

PET-CT has a limited role in the initial evaluation of adnexal masses. It may be useful in assessing disease recurrence or in evaluating indeterminate findings on CT in patients with known malignancy. It is not recommended as a routine component of the initial workup.

Additional Biomarkers and Complementary Tools

HE4 (Human Epididymis Protein 4)

HE4 is a serum biomarker that complements CA-125 in ovarian cancer detection. Unlike CA-125, HE4 is not elevated in endometriosis and has fewer false positives in premenopausal women. It is particularly useful in distinguishing malignant from benign masses when CA-125 is equivocal. However, HE4 has limited sensitivity for mucinous and germ cell tumours.

ROMA (Risk of Ovarian Malignancy Algorithm)

The ROMA combines CA-125 and HE4 levels with menopausal status in a logistic regression model to generate a predicted probability of malignancy. It uses separate algorithms for premenopausal and postmenopausal women. Several studies have shown that ROMA and RMI have comparable overall diagnostic performance, though ROMA may offer improved specificity in premenopausal women where CA-125 false positives are common.

IOTA Simple Rules and ADNEX Model

The International Ovarian Tumor Analysis (IOTA) group has developed ultrasound-based classification systems that do not require CA-125 measurement. The "simple rules" system uses five B-features (benign) and five M-features (malignant) to classify masses as benign, malignant, or inconclusive. The ADNEX model is a multivariate logistic regression model that estimates the probability of five outcomes: benign, borderline, stage I ovarian cancer, stage II-IV ovarian cancer, and metastasis to the ovary. Both IOTA tools require specialized training in ovarian ultrasound and are not as universally applicable as the RMI.

Clinical Application: Step-by-Step

Applying the RMI in clinical practice follows a systematic approach:

1. Perform transvaginal ultrasound: Evaluate the adnexal mass for each of the five morphological features (multilocular cyst, solid areas, metastases, ascites, bilateral lesions). Count the number of features present.
2. Assign the ultrasound score (U): 0 features = U of 0; 1 feature = U of 1; 2 or more features = U of 3.
3. Determine menopausal status: Premenopausal = M of 1; Postmenopausal (amenorrhea > 1 year or age ≥ 50 after hysterectomy) = M of 3.
4. Obtain serum CA-125: Record the absolute value in IU/mL.
5. Calculate the RMI: Multiply U × M × CA-125.
6. Interpret the result: RMI < 25 = low risk (general gynecology management); RMI 25-200 = moderate risk (further evaluation); RMI > 200 = high risk (refer to gynecologic oncology).
7. Document and communicate: Record the individual component values and the total RMI in the patient's chart. Communicate the result and recommended pathway to the patient and the referring or receiving team.

RMI in Special Populations

Premenopausal Women

The RMI performs less well in premenopausal women than in postmenopausal women. This is because the premenopausal multiplier (M = 1) keeps the RMI relatively low even when ultrasound features and CA-125 are moderately elevated. Additionally, CA-125 is more commonly elevated in premenopausal benign conditions (endometriosis, pelvic inflammatory disease, menstruation). The ROMA algorithm or HE4 may provide incremental value in this group. Clinicians should maintain a low threshold for further investigation (MRI, specialist referral) in premenopausal women with complex adnexal masses even when the RMI is below 200.

Postmenopausal Women

The RMI performs best in postmenopausal women, its target population. In this group, the higher menopausal multiplier (M = 3) amplifies the score appropriately. Simple cysts in postmenopausal women (which would have U = 0 and RMI = 0) are generally benign, with a malignancy risk below 1% for unilocular cysts less than 5 cm. These can be managed with surveillance ultrasound. Complex masses in postmenopausal women with any elevated CA-125 quickly generate high RMI values that trigger appropriate referral.

Pregnant Women

Adnexal masses are found incidentally in 1-5% of pregnancies, most commonly during first-trimester ultrasound. The majority are functional (corpus luteum cysts) and resolve by the second trimester. CA-125 is physiologically elevated in the first trimester of pregnancy, reducing its specificity. The RMI has not been validated in pregnant women and should not be applied in this population. Management of persistent adnexal masses in pregnancy requires multidisciplinary input involving obstetrics, gynecologic oncology, and maternal-fetal medicine.

BRCA Mutation Carriers

Women with BRCA1 or BRCA2 mutations have a substantially elevated lifetime risk of ovarian cancer (15-40% for BRCA1, 10-20% for BRCA2). In these women, the pretest probability of malignancy for any adnexal mass is higher than in the general population. While the RMI can be calculated, clinicians should apply a lower threshold for referral and surgical intervention. Risk-reducing bilateral salpingo-oophorectomy is recommended after completion of childbearing for BRCA carriers, which may preclude the need for RMI assessment altogether.

Strengths of the RMI

• Simplicity: Only three readily available inputs are required: ultrasound morphology (routinely assessed during pelvic ultrasound), menopausal status (obtained from history), and serum CA-125 (a widely available blood test). No specialized software, complex formulas, or advanced imaging is needed.
• Extensive validation: The RMI has been validated in over 100 studies involving more than 20,000 patients worldwide, making it one of the most externally validated clinical prediction tools in oncology.
• Guideline endorsement: Recommended by NICE (CG122), RCOG, and multiple national and international gynecologic oncology societies as the first-line triage tool for adnexal masses.
• Effective triage: The RMI has been shown to improve the proportion of women with ovarian cancer who are operated on by gynecologic oncologists (rather than general gynecologists), which is associated with improved surgical outcomes and survival.
• Continuous output: Unlike binary decision tools, the RMI produces a numeric value that conveys the magnitude of risk, allowing clinicians to calibrate their response to the individual patient.
• Multiplicative design: The product formula appropriately amplifies the score when multiple risk factors coincide, while keeping it low when risk factors are isolated or minimal.

Limitations of the RMI

• Reduced sensitivity for early-stage and mucinous cancers: Approximately 20% of ovarian cancers are missed at the RMI cut-off of 200. Stage I cancers and mucinous tumours are the most likely to be missed because they may have few ultrasound features and normal or mildly elevated CA-125.
• Lower performance in premenopausal women: The premenopausal multiplier of 1 limits the score's ability to stratify risk in younger women, where benign causes of elevated CA-125 and complex masses are common.
• Does not identify non-epithelial tumours well: Germ cell tumours and sex cord-stromal tumours may not elevate CA-125 and may not produce the ultrasound features captured by the RMI.
• Ultrasound operator dependency: Although the RMI uses a simplified binary count of features, the initial identification of those features depends on the skill and experience of the ultrasound operator. Inter-observer variability in ultrasound assessment can affect scoring.
• Does not account for patient-specific risk factors: The RMI does not incorporate family history, BRCA mutation status, personal cancer history, or symptoms, all of which influence the pretest probability of malignancy.
• Binary feature assessment: The five ultrasound features are counted as present or absent without grading. A mass with one thin septation (low concern) and a mass with one large solid papillary projection (high concern) both receive U = 1.
• Zero score when U = 0: Because the formula is multiplicative, if no ultrasound features are present (U = 0), the RMI is always 0 regardless of menopausal status or CA-125 level. While this reflects the very low risk of sonographically simple cysts, rare malignancies can present as apparently simple cysts, particularly mucinous tumours and early serous carcinomas.
• Does not differentiate borderline from invasive tumours: The RMI cannot distinguish borderline ovarian tumours (which have an excellent prognosis and may be suitable for fertility-sparing surgery) from invasive carcinomas. Both may produce similar scores.

Impact on Clinical Outcomes

The implementation of the RMI into clinical pathways has had measurable effects on ovarian cancer care:

Centralization of Cancer Surgery

Studies from the United Kingdom, Scandinavia, and Australia have demonstrated that the introduction of RMI-based triage increases the proportion of women with ovarian cancer who are operated on in specialized gynecologic oncology centres, from approximately 40-50% to 70-85%. This centralization is clinically significant because surgery performed by gynecologic oncologists (compared to general gynecologists or general surgeons) is associated with higher rates of optimal cytoreduction, improved staging accuracy, and longer overall survival.

Reduction in Unnecessary Referrals

Without the RMI, a substantial proportion of women with benign adnexal masses are referred to tertiary centres unnecessarily, consuming specialist resources and causing patient anxiety. The RMI's specificity of approximately 87-92% means that fewer than 1 in 10 benign masses are incorrectly classified as high risk, allowing most benign conditions to be managed efficiently in general gynecology settings.

Facilitation of Multi-Disciplinary Discussion

The RMI provides a common, objective language for discussing adnexal mass management across disciplines. It allows primary care physicians, general gynecologists, radiologists, and gynecologic oncologists to communicate risk using a standardized metric rather than subjective impressions.

Clinical Pearls

• An RMI of 0 does not mean zero risk of malignancy. It means no ultrasound features were identified, making U = 0 and the product zero. If the clinical picture is concerning despite a simple-appearing cyst (unexplained weight loss, new-onset abdominal distension, strong family history), further investigation is still warranted.
• Always record the individual RMI components (U score, M score, CA-125 value) alongside the total RMI in the patient's chart. This allows the receiving specialist to understand the basis of the referral and reassess the components independently.
• In premenopausal women with endometriosis, CA-125 levels can reach 100-200 IU/mL. If ultrasound shows bilateral endometriomas (U = 1 for bilaterality, or potentially more features), the RMI can enter the moderate or even high range despite benign disease. Clinical context and MRI may be needed to avoid unnecessary oncology referral.
• Meigs syndrome (fibroma/thecoma with ascites and pleural effusion) is a classic benign mimic that can produce a very high RMI. The combination of a solid adnexal mass, ascites, and potentially elevated CA-125 yields a high score. MRI showing a solid mass with low T2 signal (characteristic of fibromas) can help differentiate this entity.
• The RMI cut-off of 200 was optimized for the overall population. In settings with a very high or very low prevalence of ovarian cancer, the positive and negative predictive values shift accordingly. Clinicians should consider local prevalence when interpreting results.
• CA-125 should ideally be drawn before any surgical intervention or tissue biopsy, as manipulation of the mass or peritoneum can artifactually elevate levels.
• In postmenopausal women with a simple unilocular cyst less than 5 cm and a normal CA-125, the risk of malignancy is less than 1%. These patients can be managed with surveillance ultrasound (e.g., repeat at 4-6 months) rather than surgical intervention.
• The RMI was designed as a triage tool, not a definitive diagnostic test. A high RMI should trigger referral and further workup, not automatic surgery. Conversely, a low RMI should not prevent further investigation if the clinical presentation is concerning.
• When applying the RMI across populations, be aware that CA-125 reference ranges may differ with ethnicity. Some studies have reported slightly higher baseline CA-125 values in women of Asian descent, which could affect the specificity of the cut-off.