EURURO Vol. 72 No. 6

E172 1030

Letter to the Editor
Re: Eric A. Klein, Arnon Chait, Jason M. Hafron, et al.
The Single-parameter, Structure-based IsoPSA Assay
Demonstrates Improved Diagnostic Accuracy for
Detection of Any Prostate Cancer and High-grade
Prostate Cancer Compared to a Concentration-based
Assay of Total Prostate-specific Antigen: A Preliminary
Report. Eur Urol 2017;72:942–9
The Most Appropriate IsoPSA Cutoff for Diagnosis of
Prostate Cancer
In their article on determination of IsoPSA cutoff value for
the diagnosis of prostate cancer, Klein et al reported a
cutoff point with sensitivity (Se) of 90% and specificity (Sp) of
48% to discriminate men with and without prostate cancer.
However, to determine the most appropriate cutoff value
for a test with continuous results such as IsoPSA, we need to
take into account the pretest probability (Ppr) of the disease
(prostate cancer) and an estimate of the cost of a false-
negative relative to a false-positive test result (C) . We
have shown that maximizing the weighted number needed
to misdiagnose (wNNM) corresponds to the most appro-
priate test cutoff value :
wNNM ¼
1
C Ppr 1 Se ð Þ þ 1 Ppr 1 Sp ð Þ
To calculate this index, we need to know estimates of
Ppr and C , as well as Se and Sp values corresponding to each
point on the receiver operating characteristic (ROC) curve.
Klein et al reported that the prevalence of prostate
cancer in their sample was 53%.
If we do not diagnose a patient with prostate cancer
(false-negative result), he returns in a worse condition
needing more extensive treatment. Conversely, designating
a healthy man as having prostate cancer (false-positive
result) would result in unnecessary tests, waste of
resources, time, and money, and psychological trauma to
him and his family members. Let us assume that the costs
associated with a false-negative result are 15 times those
for a false-positive result (C = 15) . To determine Se and
Sp corresponding to each point on the ROC curve (Fig. 2A in
, we scanned the figure using a digitizer (GTCO CalComp,
Scottsdale, AZ, USA).
The most appropriate cutoff value (maximum wNNM)
corresponds to a point at Se = 99% and Sp = 6% )
This value, however, changes depending on the
condition under which we want to use it. For example,
assuming prevalence of 5% (instead of 53%), the cutoff value
corresponds to a point at Se = 83% and Sp = 65% (retaining
C = 15); assuming prevalence of 53%, but C = 5 (instead of
15), the most appropriate cutoff value corresponds to a
point at Se = 93% and Sp = 44% . Having access to the
raw data, the authors can determine the IsoPSA levels
corresponding to each of the aforementioned points. Similar
analyses can be performed to determine the most
available at
journal homepage:
[(Fig._1)TD$FIG]
Fig. 1 – Receiver operating characteristic curve for discriminating men
with and without prostate cancer, regenerated from parts of Figure 2A
in the article by Klein et al . The open red circle is the point they
reported for discrimination. However, if the data presented in the
article (prevalence of 53%) are used, assuming C = 15, the most
appropriate cutoff value corresponds to another point (the green
diamond). Using the same assumptions but C = 5 (instead of 15), the
most appropriate cutoff value corresponds to another point (red solid
circle). Assuming a prostate cancer prevalence of 5% (instead of 53%)
and C = 15, the point corresponds to the orange solid square. Cutoff
values were calculated following Habibzadeh et al .
DOI of original article: .
0302-2838/# 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.