Markers for endometriosis
A waste of time and money ?
The search for non-invasive markers for the diagnosis of endometriosis remains a problem. It does cost a lot of research money and energy which -to the best of my knowledge- is a waste of time and money. Indeed, I do not expect a breakthrough in a reasonable time span notwithstanding the efforts invested.
That so many researchers invest so much effort and money is intriguing. Somebody has to make a mistake. Therefore a brief discussion or the arguments pro and con seems appropriate.
The arguments pro
Theoretically it would be nice to have a marker and simple blood test to diagnose endometriosis . It could solve the problem of the delay in diagnosis. Maybe early diagnosis (and therapy) could prevent progression to a more severe form. A marker for endometriosis would permit to give medical therapy without the need to do a laparoscopy for diagnosis and to avoid surgery.
Such a test or marker for endometriosis will also be a money making machine. First it will be massively used in all women with pelvic pain, even a little dysmenorrhea, and/or infertility. Secondly, it will boost medical therapy, which will be given to prevent progression. This claim however is based on very weak and unsubstantiated evidence.
The arguments con and the mistakes
Without defining which type of endometriosis, a marker is not useful
‘A marker for endometriosis’ sounds good. However the use of the word endometriosis (intentionally) hides that subtle, typical , cystic and deep endometriosis are different pathologies. It assumes that endometriosis is one progressive disease. This progression however has not been demonstrated for typical deep and cystic endometriosis . On the contrary most of these lesions are no longer progressive when the diagnosis is made. Also transition from one form to the other has never been demonstrated
The major problem is subtle endometriosis. For us this is a natural condition occurring intermittently in all women (see paragraph in recent review and our recent editoreal in Fertil steril ). Some however ‘erroneously) continue to consider subtle lesions as pathology that will inevitably progress.
CA125, in the early follicular phase has a sensitivity and specificity of over 90% for cystic ovarian and for deep endometriosis, as we published more than 20 years ago. In clinical practice however, this test is no longer used since the diagnostic accuracy of ultrasound is much better for large cystic ovarian endometriosis, while is is unknown what the diagnostic accuracy is for smaller deep endometriosis. Anyway the added value for a big deep endometriosis nodule which is felt with one finger, or for a large cyst, which is felt and seen on ultrasound is limited.
Several excellent reviews on potential markers were written. Not one is really useful. It would be surprising that a marker will diagnose small superficial typical endometriosis. And even if this would exist, therapy would be surgical destruction. Indeed whether medical therapy can prevent progression has not been demonstrated. What is worse, is that it is not clear that in some women taking medical treatment deep endometriosis progresses to much more severe forms.
The statistical mistakes
A significant difference is a research observation on the mean of 2 populations. A test is a prediction of one person in the toal population. Men are taller than women (highly significant) does not mean that height can be used to predict whether it is a man or a woman.
Second a statistical rule of thumb, says that if differences are large and thus not obvious in small groups of 10-20, it will never t become a useful test. The alternative is the collection of large amounts of data, in order to evaluate whether a combination of tests could yield something useful.
Third, a clear distinction should be made between a diagnostic and a screening test. A diagnostic test has a 100% sensitivity ( not 1 diagnosis missed ) and 100% specificity (not 1 erroneous diagnosis). This is clinically useful to plan therapy. The value of a screening test should be considered carefully. Example a test with 90% sensitivity (10% of diagnoses are missed) and 90% specificity (10% of diagnoses are wrong). If used for deep endometriosis, the result is that 10% of women that need surgery are not operated, and that 10% of women are operated without a reason. Before clinically useful the added value after history taking, clinical exam and ultrasound should be demonstrated.
For these reasons we do consider a diagnostic test for endometriosis not realistic unless well defined for which type of endometriosis.