Page 53 - ARNM-3-3

Advances in Radiotherapy
& Nuclear Medicine LuPSMA response patterns on PSMA PET

not attributable to the tumor. Remaining lesions were in SUV , individual lesion volume, and TTV on both
max
combined into a single total tumor burden contour, from 68 Ga-PSMA and piflufolastat F-18 PET. 8,11,12
which the total tumor volume (TTV), SUV , mean SUV Following TLW analysis, each patient’s overall response
max
(SUV mean ), and total lesional activity were calculated. was categorized as follows: “Responder,” when no lesions
The TLW was created using MIM Encore to track were categorized as “increasing;” “stable disease,” when
TM
individual lesions on pre- and post-treatment PSMA no “reducing,” “increasing” or new lesions were identified;
PET/CT scans, quantifying changes in lesion volume and “low-volume progressor,” when the proportion of
PSMA SUV . The TLW automatically matched lesions “increasing”/new disease was <50% of the TTV on post-
max
across time points, after which the user reviewed and treatment imaging; or “high-volume progressor,” when
adjusted matches to ensure accuracy. Each lesion was then the proportion of “increasing”/new disease was more than
classified into one of four categories, each represented by a 50% of the TTV on post-treatment imaging (Figure 1). The
colored contour (Figure 1): “Reducing” (green), defined as threshold of 50% to define “low-volume progressor” and
a ≥sf% decrease in volume or complete resolution; “stable” “high-volume progressor” was chosen a priori.
(yellow), defined as <30% change in volume or PSMA We evaluated the reproducibility of the TLW by
SUV ; “increasing” (red), defined as a ≥30% increase in performing duplicate analyses on each paired pre- and
max
volume or PSMA SUV ; or “new” (red), lesions identified post-treatment scan, with test–retest variability and
max
only on the post-treatment scan. repeatability coefficients calculated at the individual lesion
We selected a 30% threshold a priori. This is consistent level. Overall TLW response category repeatability was
with the European Organization for Research and also determined.
Treatment of Cancer FDG-PET criteria and PSMA PET Participants were additionally stratified according
consensus guidelines, and repeatability studies support to RECIP 1.0, based on changes in quantitative PSMA-
the use of a 30% cutoff for detecting significant changes derived TTV and the presence of new lesions. 4

2.5. Statistical analysis
A
The following clinical endpoints were assessed: PSA decline
from baseline; ≥ase reduction from baseline (PSA50) at
any time-point; PSA progression-free survival (PSA-PFS),
as defined by PCWG3 criteria; radiographic progression,
defined by Response Evaluation Criteria in Solid Tumors
(1.1) and PCWG3 criteria; and overall survival (OS).
9,13
Time-to-event endpoints (PSA-PFS and OS) were defined
B as the interval from the date of post-treatment imaging
to the event date or the date last known to be event free,
at which point the observation was censored. Multistate
modeling was used to account for potential bias related to
time on treatment. 14
Two-sided, exact binomial 95% confidence interval
(CI) were calculated for PSA response rates. Time-to-
C
event endpoints were measured, and median survival was
estimated using the Kaplan–Meier method. Overall TLW
response patterns were correlated with OS using univariable
Cox proportional hazards regression. p<5% were deemed
statistically significant, without adjustment for multiplicity.
Prognostic models based on TLW and RECIP 1.0 were
compared using the concordance index (C-index). The
reproducibility of the TLW was assessed using repeatability
Figure 1. The TLW analysis identifies reducing (green), stable (yellow), statistics. These were calculated from a hierarchical, linear,
and increasing or new (red) lesions on the end-of-treatment PET. Based mixed model that accounted for variance in measurements
on TLW, the overall response was categorized into (A) “responder,” at both the lesion-within-patient level (lower-level
(B) “low-volume progressor,” or (C) “high-volume progressor.” 15
Abbreviations: PET: Positron emission tomography; TLW: Traffic light clustering) and the patient level (higher-level clustering).
workflow. The model also included fixed effect terms for lesion type

Volume 3 Issue 3 (2025) 45 doi: 10.36922/ARNM025110011