NEW YORK (360Dx) – Biotech firm Berg continued its push into cancer diagnostics with the presentation of several posters earlier this month at the annual meeting of the American Association for Cancer Research and the preparation of a forthcoming publication detailing an assay for distinguishing between prostate cancer and benign prostate hyperplasia.
The prostate assay is the closest to commercialization of the company's tests in oncology. It is currently focused on clinical validation of that test, said Michael Kiebish, Berg's chief precision medicine officer. He said that the company didn't have a definitive timeline for its commercialization, however.
Prostate cancer is a major area of interest for diagnostics developers with overdiagnosis a particular area of focus. While guidelines are equivocal on the benefits of screening with prostate-specific antigen, many men have their PSA levels tested regularly. But while a rising PSA can indicate cancer, in many cases it does not, and even when it does, the detected cancer is often slow growing and presents little risk. This leads to both unnecessary prostate biopsies and overtreatment of cancers found through these biopsies.
Berg's test is intended to address symptomatic patients to help doctors distinguish between cases of benign prostate hyperplasia (BPH) and prostate cancer, which could allow patients with the former to avoid biopsies. The emphasis on symptomatic patients marks a shift in focus from that of many existing tests, which are intended for use in patients who present with elevated PSA levels during regular screening.
"We recognize that prostate cancer screening has moved toward symptom-based screening," Kiebish said. "There is a population of men who have BPH, and they have a swollen, enlarged prostate, and they have symptoms. They will also have elevated PSA. So this test fits into that market where it gives options for men who are symptomatic to consult with their doctor to decide whether they need a biopsy."
The company and its collaborators published a study last year in the Journal of Molecular Biomarkers & Diagnosis that investigated the performance of three serum protein markers, filamin A, filamin B and keratin-19 (FLNA, FLNB and KRT19), for distinguishing between BPH and prostate cancer. Looking at a total of 503 symptomatic patients (34 with BPH, 190 who were healthy or had another benign disease, and 279 with prostate cancer) who underwent prostate biopsy, the researchers found that two of the three markers — FLNA and KRT19 — combined with patient age could distinguish between BPH and prostate cancer with an area under the curve of .70, compared to .58 for PSA alone.
At AACR this month, Berg presented a poster detailing the performance of the test in two cohorts — the first consisting of 183 symptomatic patients with BPH who had undergone one biopsy, and 112 prostate cancer patients; and the second consisting of 20 symptomatic BPH patients who had undergone repeat biopsies, and 221 prostate cancer patients undergoing radical prostatectomy.
In the first group, the researchers found that levels of serum FLNA and KRT19 along with patient age and prostate volume distinguished between BPH and prostate cancer with an AUC of .76, compared to .56 for PSA alone. The test had sensitivity of 80 percent, specificity of 57 percent, positive predictive value of 56 percent and negative predictive value of 82 percent.
In the second group, the test distinguished between BPH and prostate cancer with an AUC of .82, compared to .54 for PSA alone, and had sensitivity of 80 percent, specificity of 80 percent, PPV of 98 percent and NPV of 27 percent.
At an NPV of 82 percent, the test would identify a substantial number of men who could safely avoid biopsy. However, it would likely miss a number of cancers, as well.
Kiebish said that the company had additional patients it planned to add to both study groups (250 additional BPH and cancer patients to the first group, and 120 BPH patients to the second group) that would better power the studies and, hopefully, improve the test's performance.
"A high NPV is the focus, and with the increase of more BPH patients in the upcoming comparison we will be able to better power the final assessment," he said.
Kiebish said Berg is now planning to publish a separate study validating the assay's ability to distinguish between BPH and prostate cancer that will look at a larger cohort of patients enrolled at four clinical sites.
He said the company has not yet committed to a commercial strategy for the test and is open to a variety of potential routes, including offering it as a laboratory-developed test out of its CLIA facility or partnering with an outside firm to bring the test to market.
Diagnostics development is just one part of Berg's larger business, which includes drug research and development and health IT. Kiebish described the company's biomarker work as both supporting the other arms of its business and as an end in itself.
"We really invested in this [omics] technology to understand it and evolve it for our own clinical understanding, to connect causality of molecular markers to clinical phenotypes," he said. "Metabolism and our phenotypes are really governed by these omics. So if we can harness the power of this technology and our AI, we can define subpopulations of patients for response or clinical outcome."
For its oncology biomarker work, Berg has relied heavily on mass spec-based discovery looking for proteomic, lipidomic, metabolomic, and other markers in patient sample types, including serum, urine, tissue, and saliva.
The company does all its mass spec work in house, Kiebish said, and has more than a dozen, mostly high-resolution instruments in its discovery facility along with several triple quadrupole systems in its CLIA lab. The validation work in the Journal of Molecular Biomarkers & Diagnosis study was done using both immunoassays and multiple-reaction monitoring mass spec, and Kiebish said Berg was open to possibly taking tests to market on a mass spec platform.
Still in its early stages is another prostate cancer assay the company is pursuing for predicting patients likely to suffer recurrence or metastasis after radical prostatectomy. That assay stems from a large cohort analysis done in collaboration with the US Department of Defense, Kiebish said. In a poster presented at AACR, the researchers detailed a study of 385 prostate cancer patients who had undergone radical prostatectomy that found that a panel of two proteins, a metabolite and a phospholipid marker, and clinical measures could identify patients who would go on to have biochemical recurrence with an AUC of .89, a PPV of 36 percent and an NPV of 96 percent.
"We're looking forward to submitting a paper [on these markers] very soon, and then pursuing clinical validation of these markers," Kiebish said.
Berg and its collaborators also presented posters at AACR detailing ongoing work in pancreatic and breast cancer. The company is in the middle of a seven-year prospective study looking for proteomic, metabolomic, and lipidomic markers for early detection of the disease and prediction of response to therapy. It has enrolled 423 of a proposed 600 patients spanning a range of pancreatic cancer stages and benign prostate conditions and identified three markers — a serum protein, a serum lipid, and a urine lipid — that appear potentially useful for detecting pancreatic cancer.
The company also presented at AACR work from its breast cancer biomarker program, with a poster comparing proteomic measurements made in flash frozen tissue samples to those made in samples preserved in optimal cutting temperature compound.
In its breast cancer work, Berg is "analyzing several hundred tissues and plasmas from different subtypes," Kiebish said, adding that the company is looking to address questions around "molecular subtypes of the disease, response to therapy, and health disparities in underserved populations."