Evaluation of early chemotherapy response by combining static- and dynamic [18F]FDG-PET with diffusion-weighted MRI in subcutaneous patient-derived endometrial cancer mouse models

Abstract

Background: The combination of carboplatin and paclitaxel is the standard chemotherapy for treatment of high-risk and recurrent endometrial cancer. Evaluation of treatment response by diagnostic imaging is routinely carried out months after start of treatment, and is based on changes in tumor size or appearance of new metastases. The aim of this study was to evaluate early chemotherapeutic response in two subcutaneous endometrial cancer mouse models generated from patient-derived organoids using static- and dynamic [¹⁸F]fluorodeoxyglucose ([¹⁸F]FDG) positron emission tomography (PET) and diffusion-weighted (DW) magnetic resonance imaging (MRI). Mice were injected bilaterally with endometrioid endometrial cancer grade 3 (EEC G3), International Federation of Gynaecology and Obstetrics (FIGO) stage 3C1 (Model A) or stage 1B (Model B) organoids (n = 15 mice). The mice were randomized into treatment (combined carboplatin and paclitaxel, n_A=8 / n_B=6 tumors) or control (saline, n_A=8 / n_B=8 tumors) groups. During tumor progression, the mice underwent T2-weighted (T2w) MRI, DW-MRI and dynamic [¹⁸F]FDG-PET at baseline/Day 0 (start of treatment), Day 3 (early) and Day 10 (endpoint) using a sequential PET-MRI small-animal scanner.

Results: At endpoint, tumor volumes at T2w-MRI (vMRI) were lower in the treatment groups in both models (p ≤ 0.029). The tumor metabolic rate (MR_FDG) from dynamic PET, was significantly lower in the treatment group at the early timepoint (Day 3) and at the endpoint in Model A (p ≤ 0.042). In Model B, MR_FDG was similar for both groups at Day 3 and at endpoint (p≥0.217). The 10 tumor voxels with the highest standardised uptake value (SUV₁₀) from static [¹⁸F]-FDG-PET was significantly lower at endpoint in the treatment groups in both models (p ≤ 0.041), but not at the early timepoint (p≥0.083). Similarly, the tumor apparent diffusion coefficient (ADC_mean) was significantly higher indicating treatment response at endpoint for treatment groups in both models (p ≤ 0.036).

Conclusions: Multimodal imaging is feasible for evaluation of early signs of treatment response in preclinical subcutaneous endometrial cancer models. The novel MR_FDG dynamic PET imaging parameter seems most promising for detecting very early treatment response following chemotherapy.

Keywords: Chemotherapy; Diffusion-weighted MRI; Dynamic PET; Endometrial cancer; Metabolic rate of glucose; Organoid models; Preclinical imaging; Treatment response.

Fig. 1

Study design. Study overview and timeline for both EEC G3 O-PDX models (Model A and B). All mice underwent three MRI-PET imaging sessions, and three treatment cycles of either Carbo-PTX (treatment group) or saline (control group) starting when the biggest tumor in each mouse was approximately 100 mm³. The first treatment was given immediately following the baseline imaging (Day 0). Made with Biorender.com. Abbreviations: Carbo-PTX Carboplatin-paclitaxel, EEC G3 endometrioid endometrial cancer grade 3, MRI magnetic resonance imaging, O-PDX Organoid-based patient-derived xenograft model, PET: positron emission tomography

Fig. 2

Representative tumor images. MRI and static [¹⁸F]FDG-PET images from a representative tumor from each model (control group) and timepoint. White arrows indicate tumor location. The vMRI for, EEC G3 Model A (A) was 78.1 mm³ (Baseline), 106.4 mm³ (Early) and 193.2 mm³ (Endpoint), For EEC G3 Model B (D) the vMRI was 96.8 mm³ (Baseline), 102.2 mm³ (Early) and 127.5 mm³ (Endpoint) mm³. Scale ADC-map: 0-0.0025 mm²/s. Scale PET: 0-3.5 SUV. Abbreviations: ADC apparent diffusion coefficient, D day, EEC G3 endometrioid endometrial cancer grade 3, [¹⁸F]FDG fluorodeoxyglucose, MRI magnetic resonance imaging, PET positron emission tomography, SUV standardised uptake value, T2w T2-weighted, vMRI MRI-derived tumor volume

Fig. 3

MRI-derived imaging markers. Quantification of vMRI and ADC_mean from the MR images from EEC G3 Model A (A-B, E-F) and EEC G3 Model B (C-D, G-H) at the different timepoints. Median and interquartile range are displayed for the bar graphs. Statistical significance (p-value < 0.05) is indicated by *. The fold change (%) was calculated from the baseline values. Abbreviations: ADC apparent diffusion coefficient, Carbo-PTX Carboplatin-paclitaxel, D day, DWI diffusion-weighted imaging, EEC G3 endometrioid endometrial cancer grade 3, ns not significant, vMRI tumor volume derived from T2w-MRI

Fig. 4

PET-derived imaging markers. Quantification of static and dynamic parameters derived from the PET images for EEC G3 Model A (A-B, E-F) and EEC G3 Model B (C-D, G-H) for the different timepoints. Median and interquartile range are displayed for the bar graphs. Statistical significance (p-value < 0.05) is indicated by *. The fold change (%) was calculated from the baseline values. Abbreviations: Carbo-PTX Carboplatin-paclitaxel, D day, EEC G3 endometrioid endometrial cancer grade 3, [¹⁸F]FDG fluorodeoxyglucose, MR_FDG, metabolic rate of FDG, ns not significant, SUV₁₀ standardised uptake value of 10 hottest tumor voxels