Cancer Blood Test Passes Proof-of-Principle Trial

A multicancer diagnostic blood test showed early potential as a screen for asymptomatic cancers in women with no history of cancer, according to results of a first-of-its kind prospective study.

Of the 10,006 women who participated in the study, 134 had positive results with the blood screening and 26 were eventually found to have cancers. The tumors represented 10 different types of cancer, including seven types of cancer that currently have no standard diagnostic test. The screening did not confer a substantial risk of unnecessary follow-up tests or interventions, as only three patients without cancer had surgery after positive tests.

Participation in the trial did not discourage women from routine cancer screening, which detected an additional 24 cancers, Nickolas Papadopoulos, PhD, of Johns Hopkins Medicine, reported during the virtual meeting of the American Association for Cancer Research.

“It is feasible for a minimally invasive blood test to safely detect several types of cancers in patients not previously known to have cancer, enabling treatment with intent to cure in at least a subset of patients,” said Papadopoulos. “This advance informs the design of randomized trials to establish clinical utility, cost effectiveness, and benefit-to-risk ratio of future tests.”

The study was published simultaneously in Science.

The blood-based screen by itself had a positive predictive value of 19%, which increased to 41% with the addition of whole-body PET/CT imaging, said invited discussant David G. Huntsman, MD, founder and chief medical officer of Contextual Genomics, a cancer genomics and bioinformatics company. Most of the cancers detected by the test were localized or had only regional involvement. When combined with regular screening methods, the test doubled the number of cancers detected.

However, the results raised a number of questions left unanswered in the presentation: What kinds of usual screening were used? How were false-positive cases ruled out? How will the test’s sensitivity and negative predictive value be determined without use of whole-body MRI or some other method? How long until a meaningful survival analysis can be performed? What is the optimal testing schedule? Will such screening contribute to overdiagnosis and overtreatment?

Study details

The study evaluated the performance of the DETECT-A blood test, which was developed by Papadopoulos and colleagues at Johns Hopkins. The test can identify 1,933 bases covering regions of 16 genes that are commonly mutated in cancer, as well as nine cancer-associated proteins. A more advanced version of the test, known as CancerSEEK, was developed after the study began.

Most studies of blood tests, or liquid biopsies, for cancer have been modest in scale and involved patients known to have cancer. The DETECT-A trial limited enrollment to women without known cancer and without any symptoms of cancer. Cancers detected in that setting would likely be less advanced and smaller as compared with cancers assessed by case-control studies of liquid biopsies, said Papadopoulos.

The study participants were 65 to 75 and had multiple comorbid conditions. All were enrolled through the Geisinger Health System of Pennsylvania and New Jersey. A positive DETECT-A test was followed by a confirmatory test. If that was positive, the patient underwent a diagnostic PET/CT scan. Each patient was followed for at least 12 months after testing. Patients were encouraged to continue routine cancer screenings.

During the course of the study, 96 patients had cancer diagnoses: 26 identified by DETECT-A, 24 by routine screening, and 46 by other methods. The DETECT-A identified one or more cases of lymphoma and lung, breast, thyroid, kidney, colorectal, appendiceal, uterine, and ovarian cancers, as well as carcinoma of unknown primary.

Papadopoulos said 57% of the cancers detected by DETECT-A were identified by mutations and the rest by the presence of cancer-associated protein. Two-thirds of the DETECT-A cancers were localized or regional; statistical modeling predicted that only 37% of the cancers would be at such early stages of development at detection, Huntsman said.

Of the 26 cancers identified by DETECT-A, confirmatory PET/CT was performed in 15 cases, and nine patients had tumors removed surgically. Papadopoulos said false-positive results with the blood test led to PET/CT imaging in 1% of cases, and 0.22% of all patients screened had surgical procedures as a result of false-positive results (one patient with colonic polyps and high-grade dysplasia that could not be removed endoscopically; one case of appendiceal carcinoma in situ; and 10-cm ovarian mucinous cystadenoma).

Papadopoulos said the results showed that a blood test can be used prospectively to:

• Detect cancers not previously detected by other means
• Intervene with curative-intent therapy
• Detect cancers without discouraging adherence to standard-of-care screening methods
• Avoid frequent futile invasive follow-up tests

The results require confirmation in different patient populations (including men and different age groups), he acknowledged. Studies evaluating the CancerSEEK test, under commercial development by Thrive Earlier Detection (of which Papadopoulos and two other co-investigators are founders), have already begun.

Last Updated May 01, 2020