EURURO Vol. 72 No. 6

943 1030

1. Introduction
Prostate-specific antigen (PSA) is arguably the most
successful blood-based cancer biomarker to date. Despite
criticism PSA has transformed the landscape of early
detection, screening, and management of prostate cancer
(PCa) in the last few decades. PSA is distinct from virtually
all other cancer biomarkers because of its almost exclusive
specificity to the prostate, allowing direct assessment of
physiological conditions in the gland with a simple blood
test. Unfortunately, PSA is tissue- but not cancer-specific,
and overdiagnosis and overtreatment of PSA-detected,
biologically insignificant cancers are widely recognized as
key limitations in its clinical utility .
The vast majority of currently available protein-based
cancer biomarkers are defined as normal or abnormal
according to their concentration in body fluids. This
definition owes more to the wide availability of low-cost
and convenient technology such as enzyme-linked immu-
nosorbent assays than to biological reasons. Indeed, an
increase in biomarker concentration in blood or other body
fluids could be due to a plethora of unrelated physiological
mechanisms such as increased cell membrane permeability
or inflammation resulting in cell death. Furthermore, many
cancer-related proteins undergo alterations to their struc-
ture, including conformational changes due to point
mutations, truncations, and post-translational modifica-
tions such as glycosylation resulting from the altered
metabolism of cancer cells. These structural changes may
result in modified interactions with other proteins in the
blood, offering an opportunity for improved methods of
detection.
Recognition of structural changes to PSA, such as free PSA
(which signifies differences in interaction of PSA with a1-
antichymotrypsin and pro-PSA, a specific isoform of
PSA have better diagnostic accuracy than measure-
ment of the PSA parent protein alone. However, as the
molecular evolution of cancer may result in changes in
structural isoforms of a biomarker over time in the same
patient, and in differences in which isoforms are present
among individual patients, the diagnostic accuracy of even
these structurally altered PSA proteins has limitations. The
lack of perfect sensitivity of the currently available next-
generation PSA assays such as PHI and 4Kscore may be
attributable to the fact that they measure only a few known
isoforms of PSA that are informative only if they are present
in a given patient at a given time. Thus, since it is known
that multiple isoforms of PSA that are not measured by
current assays exist a method that detects multiple
PSA isoforms without a priori knowledge of which are
present in a given sample is likely to have better diagnostic
accuracy than existing assays.
Here we describe our initial clinical experience with
IsoPSA, previously known as PSA/SIA , as a new blood-
based assay for detection of PCa. IsoPSA is a structure-based
(rather than concentration-based) assay that agnostically
interrogates the entire spectrum of structural changes of
complex PSA (ACT-PSA). We report on the performance of
IsoPSA in a multi-institutional prospective study of US men
referred for prostate biopsy on the basis of currently
accepted clinical criteria. The endpoints of the study were
the ability of the IsoPSA assay to identify the risk of any PCa
(defined as Gleason 6) versus no cancer and of high-grade
PCa (defined as Gleason 7) versus low-grade PCa or benign
disease in comparison to a standard concentration-based
assay for total PSA.
2. Patients and methods
2.1. Patient population and specimen collection
This institutional review board–approved, multicenter prospective
study enrolled men scheduled for prostate biopsy because of a rising
PSA level or suspicious digital rectal examination (DRE). Five academic
and community urology centers across the USA (Cleveland Clinic; Louis
Stokes VA Medical Center; Kaiser Permanente Northwest; Michigan
Institute of Urology; and Chesapeake Institute of Urology) collected
heparin-plasma for IsoPSA between August 2015 and December
2016. Samples were collected within 30 d before to biopsy, processed
according to Early Detection Research Network (EDRN) guidelines ,
and frozen at 80 8C until analysis. The primary study endpoints of this
preliminary study were the presence or absence of cancer and cancer
grade as detected by 12-core transrectal ultrasound (TRUS) or magnetic
resonance imaging (MRI)-TRUS fusion biopsy. Exclusion criteria includ-
ed serum PSA < 2 ng/ml; recent (<72 h) prostate manipulation, including
DRE; recent ( < 2 wk) urinary tract infection and/or prostatitis; recent
( <30 d) prostate surgery, urinary catheterization, prostate infarction, or
endoscopic evaluation; and other urinary tract malignancy. Because
IsoPSA measures PSA structure rather than concentration, men on 5a-
reductase inhibitors (5ARIs), which are known to affect PSA concentra-
tion, were not excluded. Histopathologic evaluation of the biopsy
specimens was performed by each site according to local standards.
Overall, 434 samples were collected, with 173 exclusions: 84 because of
prolonged storage ( >90 d), 22 because of canceled biopsies, 21 because
of serum PSA < 2 ng/ml, 19 because of a breach in sample collection
protocol, 21 because of shipping delays, and six because of other reasons,
leaving a final analytical cohort of 261 samples. Signed informed consent
was obtained from all enrollees. Demographic data and clinical
information for the analytical cohort are shown in
2.2. Laboratory methods
Frozen plasma samples were shipped to Cleveland Diagnostics (Cleve-
land, OH, USA) and all testing was performed and reported naı¨ve to
pathology outcome. On receipt, the samples were thawed and
immediately added to IsoPSA reagent tubes. The reagent tubes were
vortexed, centrifuged, and subjected to the IsoPSA assay (the IsoPSA
assay is for research use only in the USA as of February 2017), which
comprises two steps: partitioning of plasma samples in an aqueous two-
phase system (IsoPSA RUO reagent, Cleveland Diagnostics), followed by
measurement of free and total PSA concentrations in each of the two
aqueous phases (referred to as top or bottom). An aliquot was removed
from each phase and the total and free PSA concentrations were
measured using US Food and Drug Administration–approved clinical
assays (Cobas e411, Roche Diagnostics, Indianapolis, IN, USA). The
relative robustness of the IsoPSA assay and its reliance on only standard
clinical PSA assays is a distinct advantage for its eventual use in
distributed environments.
The IsoPSA assay readout, or test parameter K, is calculated as:
K ¼
complex PSA ½ bottom
complex PSA ½ top ¼
total PSA ½ bottom free PSA ½ bottom
total PSA ½ top free PSA ½ top
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 9 4 2 – 9 4 9 943