Radiomics and visual analysis for predicting success of transplantation of heterotopic glioblastoma in mice with MRI

Abstract

Background: Quantifying tumor growth and treatment response noninvasively poses a challenge to all experimental tumor models. The aim of our study was, to assess the value of quantitative and visual examination and radiomic feature analysis of high-resolution MR images of heterotopic glioblastoma xenografts in mice to determine tumor cell proliferation (TCP).

Methods: Human glioblastoma cells were injected subcutaneously into both flanks of immunodeficient mice and followed up on a 3 T MR scanner. Volumes and signal intensities were calculated. Visual assessment of the internal tumor structure was based on a scoring system. Radiomic feature analysis was performed using MaZda software. The results were correlated with histopathology and immunochemistry.

Results: 21 tumors in 14 animals were analyzed. The volumes of xenografts with high TCP (H-TCP) increased, whereas those with low TCP (L-TCP) or no TCP (N-TCP) continued to decrease over time (p < 0.05). A low intensity rim (rim sign) on unenhanced T1-weighted images provided the highest diagnostic accuracy at visual analysis for assessing H-TCP (p < 0.05). Applying radiomic feature analysis, wavelet transform parameters were best for distinguishing between H-TCP and L-TCP / N-TCP (p < 0.05).

Conclusion: Visual and radiomic feature analysis of the internal structure of heterotopically implanted glioblastomas provide reproducible and quantifiable results to predict the success of transplantation.

Keywords: Experimental study; Glioblastoma; Magnetic resonance imaging; Radiomic feature analysis; Tumor cell proliferation.

Fig. 1

Successful tumor growth after implanting a xenograft in the left flank of an immunodeficient female mouse. A Coronal reconstruction from the 3D volume precontrast T1w MR-data set with a 0.3-mm slice thickness showing the abdomen, pelvis, and thighs of the mouse. In prone position the animal was fixed with stretched legs. The tumor in the right groin had a maximum diameter of 9.7 × 8.2 mm (white arrow). B—D A triplet of the magnified tumor is given. The T2w image B shows an inhomogeneous internal structure. Areas with linear signal voids (black arrow) within the tumor are seen as a correlate for histologically confirmed neovascular proliferation. T1w precontrast images C showing a rim (white arrowheads) isointense to the neighboring muscle (white asterix). Inhomogeneous enhancement of the tumor after administering of contrast agent D

Fig. 2

Transplant failure. A Coronal reconstruction of the 3D T1w volume precontrast MR-data set with a 0.3-mm slice thickness showing the abdomen, pelvis, and thighs of the mouse and the localization of the xenograft (white arrow) near a muscle bundle (white asterix). Eighty-eight days after cell transplantation, the xenograft measured 5.4 × 1.4 mm, had low signal intensity and was clearly distinguishable from the subcutaneous fatty tissue and the muscle. B—D A triplet of the zoomed xenograft is given. In contrast to the successful tumor growth in Fig. 2, the inner structure of the xenograft is homogeneous with high signal in the T2w image B and low signal in the T1w precontrast image C. A relatively homogeneous enhancement of the entire lesion could be seen after administering contrast agent D. No tumor cells could be detected in the histological work-up

Fig. 3

Scatter plot with fitted regression function for quantitative parameters: Volume A; Area of the contrast enhancing part B; T2w Signal Intensity C, and T1w Signal Intensity D. Group A = tumors with high tumor cell proliferation (H-TCP) and group B = tumors with low or no tumor cell proliferation (L-TCP/N-TCP). The solid, bold lines display the fitted regression line for each variable with darkened areas representing the 95% confidence interval. Note that the volume increased in H-TCP, whereas in 100% of N-TCP and 80% of L-TCP a continuous decrease in volume was observed over time (p < 0.05)

Fig. 4

Summary of the significantly different features of qualitative and quantitative MR measurements. The 21 tumors were divided into two groups: high tumor cell proliferation (Ki67 ≥ 10%—50%), and low or no tumor cell proliferation (Ki67 < 10%). The features `Signal intensity of the rim on precontrast T1w`, `Thickness of enhancing margin`, and `Volume` turned out to be significantly different between the two groups at the end of the observation period (p < 0.05) (table last column). Characteristic MR lesion morphology per feature is shown schematically in the graphics for both groups. In addition, the predictor value for each of the three MR features is provided at 8 and 12 weeks after transplantation (3rd and 4th column of the table). Note: AUC = area under the curve; acc = accuracy; sens = sensitivity; spec = specificity