EURURO Vol. 72 No. 3

sought to evaluate whether the change in clinical risk,

assessed from initial biopsy to final surveillance biopsy,

predicted the presence of adverse pathologic findings at

delayed radical prostatectomy (RP).

We identified study participants under a prior University

of California, San Francisco (UCSF) institutional review

board–approved protocol. Most men were selected based

on strict UCSF AS criteria: PSA 10 ng/mL, clinical stage T2

or lower, 33% positive biopsy cores, and 50% positivity

within a single core. Patients with relatively favorable risk

profiles outside of strict AS criteria (eg, higher volume

Gleason 3 + 3, low-volume Gleason 3 + 4) desiring surveil-

lance were also included. All pathology slides obtained from

outside institutions were reviewed by experienced aca-

demic pathologists. Men on AS between 1993 and 2013 who

consented to prospective data collection and who did not

receive definitive treatment for a minimum of 6 mo were

included and participated in a surveillance program, as

previously reported

[2]

. The Cancer of the Prostate Risk

Assessment (CAPRA) score, an extensively validated risk

assessment instrument, was calculated, as described

previously, for all patients at initial diagnosis and following

most recent biopsy (Supplementary Table 1)

[3,4]

. Patients

with high-risk PCa (CAPRA

5 and/or Gleason score 4 + 3)

at diagnosis and those without subsequent clinical follow-

up at our institution were excluded.

We evaluated the change in CAPRA score and its

constituent components among all patients from diagnosis

to last follow-up biopsy, using descriptive statistics and

contingency tables with

values based on chi-square tests.

Definitive treatment included RP, radiation therapy, andro-

gen deprivation therapy alone, or ablative therapy. Among

patients who received RP, adverse pathology was defined as

the presence of primary Gleason pattern 4 + 3, pathologi-

cal T stage T3a or higher, and/or lymph node positivity—

pathologic end points demonstrated to predict significant

future clinical events

[5] .

The difference in CAPRA score

from diagnosis to last biopsy was then used as a primary

explanatory variable. A patient with, for example, a CAPRA

score of 1 on diagnostic biopsy and 4 on a third surveillance

biopsy would experience a net change of 3 points. Other

covariates that could affect the response variable were

included in the models, including PSA density, CAPRA score

at the time of last biopsy, and time from diagnosis to

prostatectomy. Among patients ultimately receiving

delayed RP, we used receiver operating characteristic

analysis to compare the effect of individual clinical

parameter reclassification with CAPRA score change on

the prediction of adverse pathology. Mann-Whitney

statistics were used to compare the area under the curve

(AUC).

Overall, 735 patients met the inclusion criteria and were

followed for a median of 52 mo. Mean age at diagnosis was

62 yr, and the median PSA was 5.2 ng/ml. At diagnosis,

577 patients (79%) met strict UCSF AS criteria, and 85% were

low risk (CAPRA 0–2), whereas 15% were classified as

intermediate risk (CAPRA 3–5). The complete baseline

clinical and demographic characteristics of the cohort

are shown in Supplementary Table 2. When assessed on

a continuous scale, CAPRA score was unchanged in

192 patients and decreased in 74. Shift in CAPRA score

occurred due to multidirectional changes in biopsy Gleason

score in 413 patients (56%), in PSA in 297 (40%), and in

percentage of positive cores in 278 (38%). Moreover, 97

(13%) experienced reclassification by Gleason score alone,

156 (21%)

[1_TD$DIFF]

by PSA alone, and 29 (4%) were reclassified based

[5_TD$DIFF]

changes

[6_TD$DIFF]

in tumor volume. In total, 282 (38%) had a

change in one parameter alone, 166 (23%) had changes in

two parameters, and 42 (5.7%) had changes in three

parameters

( Fig. 1

). In a multivariable logistic regression

model, unit increases in CAPRA score were significantly

associated with the occurrence of adverse pathology (odds

ratio [OR]: 1.60; 95% confidence interval [CI], 1.25–2.04;

0.001). In addition, clinical risk (CAPRA) following last

biopsy was also independently associated with adverse

pathology (OR: 1.52; 95% CI, 1.21–1.92;

0.001)

( Table 1

Magnitude of CAPRA score change (AUC: 0.72) outper-

formed individual PSA progression (AUC: 0.64;

= 0.03),

and change in percentage of cores positive for cancer (AUC:

0.64;

= 0.04) for the prediction of adverse pathology at

delayed RP.

Numerous surrogate end points during surveillance have

been proposed to identify individuals at risk for harboring

sufficiently aggressive disease warranting treatment.

Changes in biopsy Gleason grade or tumor volume are

widely regarded by surveillance protocols as an indication to

pursue treatment; these events occur in a reliable proportion

of men with each successive biopsy

[6,7] .

The occurrence of

Gleason grade reclassification has been attributed to several

factors including the detection of higher grade or volume

tumor as a consequence of initial biopsy inaccuracy and the

contribution of genuine cancer progression over time;

however, the directionality of sampling limitations does

not uniformly favor underdetection

This has been reflected

in RP series in which misclassification at initial biopsy has

resulted in both pathologic upgrade and downgrade

[8] .

Con-

sequently, reclassification metrics based solely on a single

clinical parameter may incompletely account for an individ-

ual’s risk of clinically significant disease progression and

expose many to early treatment.

We evaluated the change in clinical risk among men with

low- and intermediate-risk PCa managed with AS at a single

Table 1 – Multivariable logistic regression models of adverse

surgical pathology among men treated with radical prostatectomy

following initial active surveillance (

= 169)

Variable

OR 95% CI

value

CAPRA score at last biopsy (per 1 U)

1.52 1.21–1.92

0.001

Log PSA density

0.83 0.44–1.58

0.57

Age (per 5 yr)

1.15 0.89–1.49

0.29

Time to RP (yr)

1.17 0.99–1.39

0.07

Magnitude of CAPRA change (per 1 U)

1.60 1.25–2.04

0.001

Log PSA density

0.96 0.52–1.80

0.91

Age (per 5 yr)

1.24 0.96–1.61

0.10

Time to RP (yr)

1.18 0.99–1.39

0.06

CAPRA = Cancer of the Prostate Risk Assessment; CI = confidence interval;

OR = odds ratio;

PSA = prostate-specific antigen;

RP = radical

prostatectomy.

E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 3 2 9 – 3 3 2

330