"A new test can help doctors identify ovarian cancer more accurately and cut down on instances of unnecessary surgery," BBC News reports.
The BBC accurately reflects the findings of researchers who developed new tests for ovarian cancer. These tests use clinical and ultrasound findings to assess whether tumours are benign or malignant and, if they were malignant, the likely stage of the cancer (how far the cancer has spread).
More accurate testing could lead to women with ovarian cancer getting the most appropriate treatment. In younger women, it may also help ensure, where possible, their fertility is preserved in some cases.
The series of tests used by the researchers are based on a "prediction model" developed using information from more than 3,000 women who had a "mass" seen on their ovaries at ultrasound. These masses were then surgically removed and examined in the laboratory.
This model was able to distinguish well between benign and malignant tumours, as well as assessing the stage of any spread.
Although the researchers suggest the predictive model (called ADNEX) could be improved and used as a second-stage test to distinguish between different types of ovarian tumours, it is not a screening test. Screening for ovarian cancer does not presently take place in the UK.
Where did the story come from?
The study was carried out by researchers from the University of Leuven in Belgium in collaboration with researchers from across Europe and the International Ovarian Tumour Analysis group.
It was funded by the Flemish government, the UK National Institute for Health Research, the Swedish Medical Research Council, funds administered by Malmö University Hospital and Skåne University Hospital, the Malmö General Hospital Foundation for Fighting Against Cancer, and two Swedish government grants.
The BBC reported the story well, emphasising that, in its present state, the research could help ensure women get the appropriate treatment, rather than providing the basis for a general screening programme.
What kind of research was this?
This was a cohort study of women with at least one ovarian mass (tumour) that needed to be surgically removed. The researchers wanted to find a way to predict risk, using ultrasound features and other patient characteristics to help distinguish between different types and stages of ovarian tumour (including benign and malignant) prior to surgery.
Knowing the likely stage and grade of individual cases of ovarian cancers would allow teams to optimise treatments, which should lead to improved outcomes. Also, in younger women, it may offer the opportunity to preserve fertility.
If a cancer is at a very early stage, it can be possible to treat it by removing the ovaries but leaving the womb intact. The woman would then still have the option to conceive through IVF with donor eggs or eggs removed prior to surgery.
What did the research involve?
The researchers studied ultrasound and clinical data from 3,506 women with an ovarian mass who had an ultrasound before the mass was removed by surgery. The mass was examined in the laboratory and classified into one of five tumour types.
The researchers used all of this data to create a "prediction model", which they called Assessment of Different Neoplasias in the Adnexa (ADNEX), to help distinguish between:
- benign tumours (not cancerous)
- borderline tumours (tumours that normally grow slowly and have low malignant potential)
- stage I invasive tumours (cancer is only in the ovaries)
- stage II to IV invasive tumours (the cancer has spread to other organs)
- secondary metastatic ovarian tumours (where the cancer didn't start in the ovaries, but has spread to them from somewhere else in the body)
The researchers tested their prediction model to see whether it could distinguish between these different types of tumours on a further 2,403 women. They used their findings from these women to update their prediction model.
What were the basic results?
The ADNEX prediction model contained three clinical and six ultrasound predictors:
- blood levels of cancer antigen-125 (a tumour marker that can be raised in ovarian cancer)
- the type of centre the woman was being treated at (oncology centre or other hospital)
- maximum diameter of the mass proportion of solid tissue
- more than 10 cyst locules (making the mass look like a cluster of grapes)
- number of papillary projections (where the wall of the mass projects into the mass itself)
- acoustic shadows (loss of sound echo behind a sound-absorbing structure)
- ascites (presence of abnormal free fluid in the abdomen)
The researchers looked at how well the prediction tool was able to distinguish between the different types of tumour.
The tool was found to be able to distinguish between benign and malignant tumours. The area under the curve (AUC) for distinguishing between all benign and malignant tumours was 0.94 (AUC can range from 0 to 1, with 1 being a perfect test, with no false positives or false negatives). An AUC of 0.94 shows good distinguishing power between benign and malignant tumours.
When looking at the ability to distinguish between benign and different stages of cancerous tumour, the AUC ranged from 0.85 for benign versus borderline, to 0.99 for benign versus stage II to IV ovarian cancer.
The tool had a variable level of accuracy distinguishing between different types and states of cancer. For example, the AUCs ranged from 0.71 for stage I versus secondary metastatic, and 0.75 for borderline versus stage I, to 0.95 for borderline versus stage II to IV.
The performance of a test depends on the cut-off you select. When the cut-off was set so women tested positive for a malignant tumour if the ADNEX said they had a 10% risk or more of the tumour being malignant, the tool had a sensitivity of around 96.5%, which is the proportion of women with a malignant tumour accurately detected as malignant. It also had a specificity of 71.3%, which is the proportion of those with a benign tumour accurately detected as benign.
This means that at this cut-off the test has a very low "false negative" rate, but quite a high "false positive" rate. This is important because it means that while the test could be very useful in detecting potentially harmful cancers, around 30% of women with benign tumours were also given a positive test result.
How did the researchers interpret the results?
The researchers concluded that, "The ADNEX model discriminates well between benign and malignant tumours and offers fair to excellent discrimination between four types of ovarian malignancy.
"The use of ADNEX has the potential to improve triage and management decisions, and so reduce morbidity and mortality associated with adnexal [ovarian] pathology."
This study describes a new way of distinguishing between benign and malignant ovarian tumours, and for distinguishing between different stages of malignant tumours.
The reseachers found their prediction model was able to discriminate well between benign and malignant tumours overall. But it showed more variable accuracy for distinguishing between different stages of ovarian cancer – for example, between borderline, stage I and stage II to IV ovarian cancers and secondary metastatic tumours.
As the researchers point out, a potential limitation of their study is they were only able to study tumours from women who were about to undergo surgery to remove the tumour.
They were unable to study women with ovarian masses who were deemed not to require, or were not suitable for, surgery and who underwent "expectant management" (watching and waiting). They say information on women who are managed conservatively started being collected in 2013.
The researchers hope the ADNEX tool will be able to aid decisions about treatment of ovarian cancer and improve outcomes.
It should be noted ADNEX is not a screening test and screening for ovarian cancer does not presently take place in the UK.