Scholarlink

Objective <p>This retrospective single-center study evaluated and compared the prognostic value of distribution parameters of standardized uptake value (SUV) in whole-body tumor volume (WTV) on PSMA PET/CT prior to <sup>177</sup>Lu-PSMA radioligand therapy (PSMA RLT) for early treatment response and overall survival (OS).</p> Methods <p>For 59 patients with metastatic castration-resistant prostate cancer baseline clinical and PSMA PET/CT characteristics (Gleason score, time-to-therapy, age, PSA, prior therapies, metastasis sites, WTV) and SUV distribution parameters in WTV (SUVmean, SUVmedian, SUVmax, skewness, kurtosis and interquartile range (IQR)) prior to PSMA RLT were retrospectively collected. The prognostic value for early biochemical and/or imaging progression (PD) (according to PCWG3 and RECIP) as well as OS was evaluated using univariate logistic or Cox regression. SUV distribution parameters significant in univariate analyses were further evaluated using multivariate logistic or Cox regression adjusted for significant baseline characteristics. Model performance was assessed by cross-validation (5-folds, 10 repeats), quantified by area under the curve (AUC) and Harrell’s C, and compared with Wilcoxon signed-rank tests. Addition of other parameters into a regression model was assessed with likelihood ratio tests.</p> Results <p>31 patients (53%) showed early PD after 2 cycles. Median follow up was 20 [2–38] months, with 18 patients (31%) being alive at last follow-up. Univariately, PD correlated with age, prior chemotherapy (CTx), SUVmean, SUVmedian, SUVmax and IQR (<i>p</i> ≤ 0.05) while OS correlated with SUVmean, SUVmedian, SUVmax, IQR (<i>p</i> ≤ 0.01) as well as age on a trending level (<i>p</i> = 0.06). IQR was more predictive of PD (AUC 0.95 [0.8-1.0], <i>p</i> &lt; 0.001, corrected for age and CTx) and OS (C 0.8 [0.62–0.97], <i>p</i> &lt; 0.001, corrected for age) compared to SUVmean (AUC 0.86 [0.58-1.0] ; C 0.66 [0.41–0.87]), SUVmedian (AUC 0.86 [0.66-1.0]; C 0.61 [0.36–0.92]) and SUVmax (AUC 0.84 [0.59-1.0]; C 0.63 [0.44–0.88]). Adding SUVmean, SUVmedian or SUVmax to the IQR regression model did not significantly increase its predictive value for PD or OS (<i>p</i> ≥ 0.27).</p> Conclusions <p>On PSMA PET/CT before PSMA RLT the dispersion parameter IQR of the SUV distribution in WTV demonstrates strong potential for predicting early progression and overall survival and surpasses the prognostic value of established measures of central tendency (SUVmean and SUVmedian) and SUVmax.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prognostic value of whole-body tumor SUV dispersion on baseline PSMA PET prior to PSMA radioligand therapy

  • 期刊论文

  • 发表日期：

    2025年12月5日

查看全文
  • Felix Kind,
  • Ursula Nemer,
  • Katia Brüggemann,
  • Cordula A. Jilg,
  • Philipp T. Meyer,
  • Michael Mix,
  • Martin T. Freitag

摘要

Objective

This retrospective single-center study evaluated and compared the prognostic value of distribution parameters of standardized uptake value (SUV) in whole-body tumor volume (WTV) on PSMA PET/CT prior to 177Lu-PSMA radioligand therapy (PSMA RLT) for early treatment response and overall survival (OS).

Methods

For 59 patients with metastatic castration-resistant prostate cancer baseline clinical and PSMA PET/CT characteristics (Gleason score, time-to-therapy, age, PSA, prior therapies, metastasis sites, WTV) and SUV distribution parameters in WTV (SUVmean, SUVmedian, SUVmax, skewness, kurtosis and interquartile range (IQR)) prior to PSMA RLT were retrospectively collected. The prognostic value for early biochemical and/or imaging progression (PD) (according to PCWG3 and RECIP) as well as OS was evaluated using univariate logistic or Cox regression. SUV distribution parameters significant in univariate analyses were further evaluated using multivariate logistic or Cox regression adjusted for significant baseline characteristics. Model performance was assessed by cross-validation (5-folds, 10 repeats), quantified by area under the curve (AUC) and Harrell’s C, and compared with Wilcoxon signed-rank tests. Addition of other parameters into a regression model was assessed with likelihood ratio tests.

Results

31 patients (53%) showed early PD after 2 cycles. Median follow up was 20 [2–38] months, with 18 patients (31%) being alive at last follow-up. Univariately, PD correlated with age, prior chemotherapy (CTx), SUVmean, SUVmedian, SUVmax and IQR (p ≤ 0.05) while OS correlated with SUVmean, SUVmedian, SUVmax, IQR (p ≤ 0.01) as well as age on a trending level (p = 0.06). IQR was more predictive of PD (AUC 0.95 [0.8-1.0], p < 0.001, corrected for age and CTx) and OS (C 0.8 [0.62–0.97], p < 0.001, corrected for age) compared to SUVmean (AUC 0.86 [0.58-1.0] ; C 0.66 [0.41–0.87]), SUVmedian (AUC 0.86 [0.66-1.0]; C 0.61 [0.36–0.92]) and SUVmax (AUC 0.84 [0.59-1.0]; C 0.63 [0.44–0.88]). Adding SUVmean, SUVmedian or SUVmax to the IQR regression model did not significantly increase its predictive value for PD or OS (p ≥ 0.27).

Conclusions

On PSMA PET/CT before PSMA RLT the dispersion parameter IQR of the SUV distribution in WTV demonstrates strong potential for predicting early progression and overall survival and surpasses the prognostic value of established measures of central tendency (SUVmean and SUVmedian) and SUVmax.