What if the screening test hurt more people than the cancer?

ovarian cancer ribbon

Every so often I get an e-mail forwarded to me recounting the story of a friend or acquaintance recently diagnosed with ovarian cancer. The cancer is almost always far advanced, and the prognosis is very grim.

The e-mail reveals that the cancer might have been diagnosed much earlier if only the woman had been given a simple blood test (CA125 test) or had an ultrasound. Readers are exhorted to press their doctors for both tests, so that if they develop ovarian cancer, it can be diagnosed early, when treatment is more likely to be successful. The e-mail makes it sound like the means of diagnosing ovarian cancer is here, but doctors are ignoring the possibilities.

The situation is far more complicated. Yes, a simple blood test or an ultrasound can lead to early detection of ovarian cancer. Unfortunately, though, it also leads to tremendous numbers of unnecessary surgeries and the complications that result. In fact, it is entirely possible that screening for ovarian cancer is more dangerous than not screening for ovarian cancer.

That is the central message of a new study published today in Lancet Oncology, Sensitivity and specificity of multimodal and ultrasound screening for ovarian cancer. You might not realize that if you read newspaper accounts of the study, which emphasize the number of cancers diagnosed. Many newspaper accounts don’t mention that for every woman diagnosed with ovarian cancer, many more had unnecessary major surgery and quite a few suffered serious complications as a result.

The study screened more than 100,000 postmenopausal women for ovarian cancer. Half had ultrasound and CA125 tests (multimodal screening); the other half had ultrasound alone. Ovarian cancer was detected in 87 women, 42 in the multimodal group and 45 in the ultrasound alone group. That sounds pretty good until you learn that in order to make those diagnoses, 942 women had surgery. In other words, 855 women had major abdominal surgery for no reason. Of those, 24 experienced major complications including perforation of an organ (requiring surgery for repair), hemorrhage, deep vein thrombosis, and pulmonary embolus.

There was a big difference in unnecessary surgery between the multimodal group and the ultrasound group. Of the 942 women who had surgery, 845 were from the ultrasound group. In other words, adding the CA125 blood test made the screening more accurate. Even so, for every woman in the multimodal group who had ovarian cancer, 2 additional women had surgery that they did not need. In the ultrasound group, for every case of ovarian cancer diagnosed, approximately 19 women underwent major abdominal surgery that was unnecessary.

Screening hurt far more women than were helped. For every woman who was diagnosed with ovarian cancer, 9 more had surgery that they didn’t need, and 2.8% of women who had unnecessary surgery sustained serious, life threatening surgical complications. That is a pretty dismal record for a screening test.

If we leave aside the ultrasound only group, the results in the multimodal group are far more encouraging. Only 97 underwent surgery, of whom 42 had ovarian cancer. As mentioned above, for every case of ovarian cancer diagnosed in the multimodal group, 1 woman had surgery that she didn’t need. Of those women who had unnecessary surgery, 4.2% sustained serious, life threatening complications.

What would happen if we instituted multimodal screening for all post menopausal women. For every 1 million women screened, 866 cases of ovarian cancer would be diagnosed, 1034 women would have unnecessary major abdominal surgery, of which 43 would sustain major, life threatening complications.

In addition, we do not know if the early diagnosis of ovarian cancer in these patients would improve outcome. Over half of the women diagnosed by screening already had advanced disease, so it is unlikely that screening improved their prognosis. Moreover, even early stage ovarian cancer is a dangerous disease, and many of these women are going to die anyway.

The ultimate value of a screening test is in lives saved, and that information is beyond the scope of this study. It is already clear, though, that for every life saved, 4 or more women will have unnecessary major abdominal surgery, some women will sustain life threatening complications, and inevitably, some women will die from complications of surgery that they did not need.

This study is large, comprehensive and well done, but it does not support mandatory screening for ovarian cancer. It demonstrates that large-scale screening is possible, and that early ovarian cancer can be diagnosed by screening. Unfortunately, it also shows that large-scale screening efforts results in substantial harm to more people than are helped. When the screening test is potentially more dangerous than the disease, it makes no sense to implement mandatory screening.