NEW YORK (GenomeWeb) – Investigators shared new data yesterday from their work advancing methods to screen for ovarian and endometrial cancers using a liquid biopsy test that detects tumor DNA in pap smear or uterine brush samples.
The results — published in Science Translational Medicine — demonstrate that even higher sensitivity is possible than what the group was able to achieve in its first experiments, either by using a different kind of sampling method or by combining pap smear analyses with liquid biopsy testing of a blood sample.
The methodology, which the group calls PapSeek, was initially developed by researchers at Johns Hopkins University, who published early results in 2013, using a DNA sequencing method invented at the institution called the Safe-Sequencing System, or Safe-SeqS.
According to the authors of the new study this week, more than 63,000 women are diagnosed with endometrial cancer in the US every year, and more than 11,000 of them die from the disease. "Ovarian cancer is less common but more lethal, affecting more than 22,000 women and killing about 14,000," the group wrote.
In the study this week, the team used a slightly updated protocol from their 2013 report, analyzing cell-free DNA in pap smear samples for mutations in a panel of 18 genes, as well as looking for aneuploidy or the presence of abnormal numbers of chromosomes in cells.
Collaborating with clinicians at McGill University in Canada, the group also extended the study to a new type of sampling: a method for collecting cells from deeper inside the uterus called a Tao brush.
Lucy Gilbert — director of Gynecologic Cancer Services at the McGill University Health Centre and a professor of Obstetrics and Gynecology and Oncology at McGill — spearheaded the addition of the Tao brush after learning about the Hopkins group's early results several years ago.
Because she came on after the Hopkins group's early development of their liquid biopsy approach, Gilbert wasn't necessarily invested in the proprietary genetic technology, but she said that through her clinical career she has become "obsessed, almost to the point of pathology," with ovarian cancer early detection.
She and her team had already been doing research on trying to offer earlier diagnosis, using a more aggressive implementation of the existing protein biomarker CA-125 and transvaginal ultrasound, but they were only able to push back the timing of diagnosis incrementally.
"I thought DNA was a match made in heaven," she said. "So, I contacted [Hopkins] and proposed [trying to go] even higher into the uterus." Biologically, Gilbert said, it makes sense. "The whole purpose of living for the fallopian tube is to get the egg into the uterus, and that's all it does — flush cells down, so if we can get closer to that, I thought we could pick up these [DNA markers] better."
In the study this week, she and her coauthors reported that they studied a total of 1,958 pap smear samples obtained from 1,658 women, including 658 endometrial or ovarian cancer patients and 1,002 healthy controls.
PapSeek detected 81 percent of the endometrial cancers (78 percent of those were early-stage cancers) and 33 percent of ovarian cancers, about a third of which were early-stage. The method was also nearly 99 percent specific overall, meaning it rarely yielded a positive result in the normal control participants.
Meanwhile, in 123 endometrial cancer patients studied using Tao brush samples, the approach identified 93 percent of the cancer cases. And in 51 ovarian cancer patients, 45 percent tested positive for cancer using the longer brush.
The Hopkins team also decided to look at whether combining PapSeek with blood-based DNA detection could boost sensitivity. The team has also developed blood- and urine-based ctDNA approaches — CancerSeek, a blood test that screens for eight cancer types, and UroSeek, a test designed to detect bladder and upper tract urothelial cancer from analysis of urine samples.
Promisingly, in a small group of 83 ovarian cancer patients in whom both plasma and pelvic samples were tested, the sensitivity of detecting ovarian cancer increased to 63 percent.
A central question for how successful the test might be if researchers are able to validate it prospectively is whether it can pick up the most dangerous cancers early enough that they can be cured.
The cohort only included only a handful of early-stage, high-grade ovarian cancer cases, consistent with the fact that these cancers are so often diagnosed at advanced stages. But, the authors were able to calculate that 36 percent of these cases were positive with combined Pap and plasma sample testing and 80 percent were positive in Tao brush samples.
Pap smears are already a mainstay of cervical cancer screening, having led to dramatic improvements in the detection and treatment of cervical cancers. The Tao brush is not commonly used in the United States but is approved by the US Food and Drug Administration for endometrial sampling.
According to Gilbert, the procedure is not much different than a normal pap smear and is something that clinicians would be able to do easily in the context of annual well visits.
"It's more painful than a standard pap test, but it's just three or four seconds … and we did a study that asked women, 'Would you be willing to have this done once a year?' And they said 'gladly,'" she said.
This is important, she argued, because the most dangerous cancers seem to occur frequently in women with no known risk factors. Ideally, a test based on the PapSeek method would be best applied in broad screening — for all women over 50 — rather than those with particular indications of a higher risk.
According to Gilbert, the predictive power the group saw in this study, especially with the Tao brush and with combined blood and pelvic sampling — is significantly better than current screening methods that use protein biomarkers. And, since it has been solidly established that catching these cancers (especially high-grade tumors) early, dramatically improves patient outcomes, the argument for clinical utility is also very straightforward.
"The cure rate for high-grade serous cancer has increased no more than 2 or 3 percent in 20 years," Gilbert said. "We have made these changes — we do immunotherapy, use biomarkers for precision treatments … but all that is firefighting," she added.
"I think that's shameful. So, before I die, we absolutely must nail this."
One of the main limitations of the results so far is that the group's current study was performed in patients who had already been diagnosed with cancer, many of whom had or would have tested positive using available protein assays. Because of that, its not possible to use the results to prove directly that the liquid biopsy method is superior to these existing protocols, even though detection rates were so high.
Another important question that will have to be answered is whether the method can distinguish cancers from benign lesions that might also feature mutations that are part of the PapSeek panel. If too many of these are mistaken for cancers, it could pose problems for both clinicians and patients.
Recent findings from DNA analysis of uterine lavage, or rinse samples, found that it was possible to detect women with confirmed endometrial cancer, but also that known cancer driver mutations were present in 51 of 95 women who had no histopathological evidence of cancer.
Gilbert said that she and her colleagues have already begun to validate the current findings prospectively, which will allow them to start answering these questions.