New strategies in detecting early-stage ovarian cancer

Ovarian cancer usually presents with vague symptoms, often after the disease has metastasized. According to authors publishing in Journal of Clinical Oncology, overall, 70%–75% of cases are detected at stage III or IV, when the cure rate is <30%.^[] Currently, only 25%–30% of patients are diagnosed at stage I or II.

Nonetheless, neither serum biomarkers nor transvaginal ultrasonography (TVU) alone is specific or sensitive enough to detect the disease, according to the National Cancer Institute (NCI).^[] Manual pelvic examination also lacks sensitivity for the detection of early-stage disease, with no evidence of decreased mortality. 

For these reasons, it’s worthwhile for physicians to evaluate the shortcomings of current screening methods and to take note of some research into multimodal tests, per the NCI.

CA-125 serum testing

The most commonly reported CA-125 reference threshold that indicates a positive screening test is 35 U/mL. Notably, this threshold was used in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Screening Trial (CA-125 levels were measured along with TVU).

Elevated CA-125 levels are not specific to ovarian cancer, as they are also present in non-gynecological cancers, liver disease, endometriosis, congestive heart failure, and pleural/peritoneal fluid accumulation, and during the first trimester of pregnancy. According to the results of two nested case-control studies using serum banks, the sensitivity for CA-125 levels of at least 35 U/mL was 20%–57% for cases occurring within the first 3 years of follow-up, and the specificity was 95%.

Multimodal screening

UKCTOCS

In The United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS), researchers compared screening outcomes in three groups of postmenopausal women aged 50 to 74 years who received 7 to 11 rounds of annual screening: TVU alone (n=50,623), multimodal screening with CA-125 testing plus TVU (n=50,625), and no screening (control; n=101,299). 

Longitudinal CA-125 measurements were assessed using the Risk of Ovarian Cancer Algorithm (ROCA), which defines a positive test as one where there is a statistically significant increase in a woman’s serial CA-125 measurements. The aim of this approach is to use subtle in-person changes to detect cancer earlier. In the trial, ROCA was the primary screen, and TVU was the secondary screen.

PLCO (nested analysis)

The PLCO trial previously mentioned also reflected multimodal screening, utilizing six annual tests for serum CA-125 (threshold: 35 U/mL) and four annual screens with TVU (n=39,105 for screening; n=39,111 for standard care). 

Moreover, of the 3,285 women with false-positive results, 1,080 received surgical follow-up, with 15% experiencing at least one serious complication.

NROSS

The article in the Journal of Clinical Oncology provided an update on the Normal Risk Ovarian Screening Study (NROSS). This study evaluated use of a two-stage screening approach in a population of 7,856 postmenopausal women at conventional hereditary risk for ovarian cancer. ROCA was used to evaluate CA-125 each year, and patients were followed up with TVU if directed by ROCA. 

The authors observed that the two-stage screening strategy with CA-125 followed by TVU had adequate specificity to achieve a positive predictive value of 50%. Only two operations were required to detect each case of ovarian cancer, a finding that was consistent with the three to four operations required in the UKCTOCS, using a similar strategy. 

The authors commented on the implications of their results for the detection of cancers originating in the fallopian tubes. A fraction of high-grade serous ovarian cancers can arise from the fallopian tubes, where there are no anatomic barriers to prevent transperitoneal spread. It may thus be particularly difficult to detect small, early-stage tubal cancers.

Although the fraction of cancers arising from the fallopian tube in women at conventional genetic risk is not known, the investigators expressed the potential of their approach: “[O]ur data suggest that the majority of ovarian cancers can be detected in early stage regardless of their site of origin.” 

They concluded that although the NROSS trial was not sufficiently powered to determine mortality, the high specificity, positive predictive value, and marked stage shift bolster further research for this approach.

Current guidance

Asymptomatic high-risk women who have a false-negative screening test are at risk of delaying preventive treatments.

Similarly, the USPSTF concluded with moderate certainty that the harms of screening using TVU or CA-125, or both, outweighed any potential benefits. There is no reduction in mortality from ovarian cancer. Harms, judged moderate to substantial, include false-positive results, which can result in unnecessary diagnostic surgery.^[]

Multi-marker blood assays

Proteins are the functional end products of genes and play a role in establishing the cancer phenotype. Consequently, proteomics may prove useful in understanding the molecular mechanisms linked to cancer. Single and multi-marker blood assays have been forwarded as potential diagnostic tools for detecting ovarian malignancies. 

Authors of a review article on screening and diagnostics tools for ovarian cancer acknowledged that CA-125 is the most important ovarian cancer biomarker discovered to date.^[] “However, there remains an ongoing search for alternative protein biomarkers for use in coordination or in comparison with CA-125,” they said. “The majority of these tests have exhibited increased accuracy for the detection of ovarian cancer over CA-125 alone while maintaining high specificity.“

They cite the work by Yale researchers, who identified a six-marker panel, consisting of CA-125, leptin, prolactin, osteopontinin, insulin-like growth factor 2, and macrophage inhibitory factor. 

Other protein techniques

Traditional methods used for quantification of serum proteins such as CA-125 or others include radiolabeled immunoassays or enzyme-based immunoassays (ELISAs). In addition, fluorescently labeled antibody-conjugated beads plus flow cytometry methods have been used in recent studies to identify ovarian cancer-specific protein levels in serum.

Researchers used a proximity extension assay combined with next-generation sequencing and multivariate modeling to identify an 11-plasma protein panel (ALPP, CXCL8, DPY30, IL6, IL12, KRT19, PAEP, TSPAN1, SIGLEC5, VTCN1, and WFDC2). Testing showed that this panel was capable of distinguishing early- from late-stage high-grade serous carcinoma with an AUC of 0.81. It also discriminated ovarian cancer patients (stage I–IV) from healthy individuals with a sensitivity of 75% and a specificity of 100%.

In commenting on these investigations, the review authors wrote: “So-called ‘perception-based’ approaches offer an additional new tool for classifying disease states, whereby an array of non-specific responses can be combined and processed with machine learning to produce a singular response that is highly specific and accurate.”

One example they cite is the development of a nano-sensing array consisting of modified semiconducting single-walled carbon nanotubes. This type of array detected high-grade serous carcinoma with higher sensitivity and specificity vs traditional serum and TVU-based approaches.

Emerging imaging techniques

The review article also addresses imaging diagnostics. Two major areas of research include the enhancement of current TVU imaging methods to improve accuracy, and the development of new imaging modalities that can accurately detect malignant ovarian masses. 

TVU can also be paired with photoacoustic imaging (PAI) to enhance diagnostic accuracy. PAI collects real-time functional and molecular information from tissues without using radiation or exogenous contrast. PAI allows for the high-resolution detection of angiogenesis and might detect neovascularization in early-stage ovarian cancer. Current PAI methods, however, can only penetrate tissues up to 5 cm, and resolution declines with increased depth. Consequently, TVU co-registration is still required.

Imaging modalities other than TVU that may be able to pick up pathologic changes that are characteristic of early ovarian tumors include MRI augmented with AI and PET combined with CT. 

Hysteroscopy and laparoscopy approaches may also be useful to screen for early ovarian cancer, with diagnosis made from a sampling of endoluminal cells. Cytuity is an FDA-cleared hysteroscopic catheter with a balloon-like mechanism on the distal end of the catheter that collects tubal epithelial cells at the level of the ostium of the fallopian tube.