Scholarlink

Background <p>The aim was to improve the realism of patient models for Monte-Carlo based evaluation of image-based quantification of tumour volume and activity in [<sup>68</sup>Ga]Ga-DOTA-TOC PET and [<sup>177</sup>Lu]Lu-DOTA-TATE SPECT images.</p> Methods <p>Three phantoms from the XCAT population were used as basis. Tumour structures were obtained from segmentation of patient <sup>68</sup>Ga-PET images and positioned at relevant positions in the phantoms. Respiratory motion was modelled, and the centre-of-mass of tumours displaced according to the organ motion in the respiration. Tumour activity concentrations were sampled from a log-normal distribution derived from patient data. Intra-organ non-uniform activity concentration was modelled for spleen, liver, and intestines. The full phantom models included all these aspects. For assessment of the impact of the modelled effects, simulations were also carried out for phantoms without respiratory motion (full-without-motion) and with spherical lesions without background variability (standard models). Evaluation was focussed on the impact of the models on the errors in tumour activity concentrations estimated from SPECT and PET images with and without partial-volume correction (PVC).</p> Results <p>The full models were successfully realized and were found to yield a visual appearance that more closely resembled that of patients, compared to the full-without-motion and standard models. For tumours in the volume range 0 mL to 10 mL, the mean relative errors obtained for <sup>177</sup>Lu-SPECT without PVC were − 51%, -38%, and − 34% for the full models, the full-without-motion models, and the standard models, respectively. Corresponding results for <sup>68</sup>Ga-PET without PVC were − 34%, -25%, and − 23%. Thus, for both PET and SPECT, the negative bias in estimated tumour activity concentrations demonstrated a clear sensitivity to the taken modelling approach.</p> Conclusions <p>Respiratory motion is an important factor for modelling realistic patient images, in the context of quantitative SPECT and PET. Non-spherical tumours and background variability has a minor, but measurable impact.</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

Improved patient models for simulation of clinically realistic 68Ga-SSTR PET and 177Lu-PRRT SPECT studies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月25日

查看全文
  • Johan Gustafsson,
  • Philip Kalaitzidis,
  • Selma Curkic Kapidzic,
  • Michael Ljungberg,
  • Katarina Sjögreen Gleisner

摘要

Background

The aim was to improve the realism of patient models for Monte-Carlo based evaluation of image-based quantification of tumour volume and activity in [68Ga]Ga-DOTA-TOC PET and [177Lu]Lu-DOTA-TATE SPECT images.

Methods

Three phantoms from the XCAT population were used as basis. Tumour structures were obtained from segmentation of patient 68Ga-PET images and positioned at relevant positions in the phantoms. Respiratory motion was modelled, and the centre-of-mass of tumours displaced according to the organ motion in the respiration. Tumour activity concentrations were sampled from a log-normal distribution derived from patient data. Intra-organ non-uniform activity concentration was modelled for spleen, liver, and intestines. The full phantom models included all these aspects. For assessment of the impact of the modelled effects, simulations were also carried out for phantoms without respiratory motion (full-without-motion) and with spherical lesions without background variability (standard models). Evaluation was focussed on the impact of the models on the errors in tumour activity concentrations estimated from SPECT and PET images with and without partial-volume correction (PVC).

Results

The full models were successfully realized and were found to yield a visual appearance that more closely resembled that of patients, compared to the full-without-motion and standard models. For tumours in the volume range 0 mL to 10 mL, the mean relative errors obtained for 177Lu-SPECT without PVC were − 51%, -38%, and − 34% for the full models, the full-without-motion models, and the standard models, respectively. Corresponding results for 68Ga-PET without PVC were − 34%, -25%, and − 23%. Thus, for both PET and SPECT, the negative bias in estimated tumour activity concentrations demonstrated a clear sensitivity to the taken modelling approach.

Conclusions

Respiratory motion is an important factor for modelling realistic patient images, in the context of quantitative SPECT and PET. Non-spherical tumours and background variability has a minor, but measurable impact.