Grand Rounds

Could Existing Ultrasound Predictive Models be Used to Better Predict Ovarian Cancer?

BACKGROUND AND PURPOSE:

Screening for ovarian cancer in asymptomatic, low risk women is not recommended
However, research continues on predictive models, particularly how to assess ultrasound images in the presence of an adnexal mass
The IOTA group, founded in 1999, is a network of approximately 50 multidisciplinary centers, worldwide with aims that include
- Standardize terminology, definitions and measurements to describe sonographic features of adnexal masses
- Research that focus on the development of predictive models to estimate the risk of ovarian malignancy when an adnexal mass is present
Abramowicz et al. (AJOG, 2017) reviewed the performance characteristics of two new ultrasound-based scoring systems
- Simple Rules
- Assessment of Different NEoplasias in the adnexa (ADNEX)

Interpretation
- B= benign, M= Malignant
- If ≥1 M features present, absence of B features= tumor is malignant
- If ≥1 B features present, absence of M features= tumor is benign
- If both B and M features present = inconclusive
Benign features
- Unilocular cyst
- Solid component <7 mm in diameter
- Presence of acoustic shadows
- Smooth multilocular tumor with largest diameter <10cm
- No detectable color Doppler signal
Malignant features
- Irregular solid tumor
- Ascites
- ≥4 papillary structures
- Irregular multilocular mass >10 cm in diameter
- Strong color Doppler signal

A model capable of estimating the risk that an adnexal tumor is benign, borderline, early stage malignant (stage 1), late stage malignant (stage II-IV), or metastatic
ADNEX model Criteria:
- Age
- Serum CA125 (software still calculates risk even if no CA- 125 value is available)
- Type of center (oncology referral center or general hospital)
- Maximum diameter of lesion (mm)
- Proportion of solid tissue (largest solid component diameter divided by maximum lesion diameter)
- Presence of more than 10 cyst locules (yes/no)
- Number of intracavitary papillary projections (0,1,2,3,>3)
- Presence of acoustic shadows (yes/no)
- Presence of ascites (yes/no)

Simple Rules
- In meta-analysis, sensitivity and specificity were 93% and 81% respectively when inconclusives were considered malignant
- Negative predictive value is 94.9% if result followed up with expert opinion
- Model prospectively validated
- Simple Rules performance was similar even with different levels of ultrasound experience
ADNEX
- Results based on clinical and ultrasound data from almost 6000 women
- Model clinically validated
- Sensitivity and specificity were 96.5% and 71.3% respectively
- ROC curve AUC was 0.94
Simple rules and ADNEX model are validated, good predictive tests for the preoperative classification of adnexal tumors

Purpose of Simple Rules and ADNEX is to help improve triage and provide more educated decisions regarding management and referral to gyn oncologists in the presence of an adnexal mass
- Improved outcomes when malignant tumors are managed by gyn oncologists
ACOG/SGO committee opinion states that
- Sensitivity and specificity of 86% and 99% respectively for malignancy can be expected when the sonographic evaluation of an adnexal mass includes
  - Tumor size
  - Morphology
  - Wall characteristics
  - Presence of septae
- Referral to a gyn oncology is recommended when one suspects malignancy
  - Postmenopause: Elevated CA125 or presence of ascites, nodular or fixed pelvic mass or evidence of abdominal or distant metastasis
  - Premenopause: Very elevated CA125, ascites or evidence of abdominal or distant metastasis
- The authors conclude that Simple Rules or ADNEX would yield better results in discriminating benign from malignant adnexal masses
- Models are simple but not yet used in the US
  - Litigation concerns result in active management of adnexal masses
  - Simple Rules and ADNEX are not yet validated in the USA
- The study suggests the models should be validated in the US and introduced into medical practice

< Previous

Next >