ESTRO meets Asia 2024 - Abstract Book

S117

Interdisciplinary – Education in radiation oncology

ESTRO meets Asia 2024

contrast-enhanced imaging of the entire dataset sample, we developed single radiomics models based on LASSO+SVM with the average AUC of 0.78 、 0.77 、 0.79, respectively, to predict the development of PsP. The multimodal radiomics combined with machine learning strategy incorporating T1WI+T2WI+T1ce showed excellent classification performance, with an AUC of 0.86 after 5-fold cross-validation, and the accuracy, sensitivity, specificity, and F1-score of the model were 84.75% 、 90.71% 、 71.58% 、 87.67% 、 89. 13%, respectively.

Conclusion:

Traditional MRI-based radiomics models have demonstrated the ability to differentiate between PsP and TP in patients with HGG following standard treatment. To enhance the performance of the classification model, radiomics features from multi-sequence images were combined with SVM to construct a multimodal radiomics model. Among the various combinations, the radiomics model derived from the fusion of three sequences (T1WI+T2WI+T1CE) exhibited the highest discriminatory power, with an AUC of 0.86. This surpassed the performance of models constructed with two sequences or single-modality radiomics models. The multimodal radiomics model provides clinicians with a valuable tool for accurately identifying PsP at an early stage. This assists in the formulation of effective individualized treatment strategies for HGG patients after standard treatment.

Keywords: high-grade glioma,pseudoprogression,radiomics

References:

[1]Liao H, Bi L, Wei J, et al. Evaluation of apoptosis induced by exposure to antineoplastic drugs in peripheral blood lymphocytes of nurses [J]. Mol Med Rep, 2017, 16(6): 8103-8109.

[2]Topkan E, Topuk S, Oymak E, et al. Pseudoprogression in patients with glioblastoma multiforme after concurrent radiotherapy and temozolomide [J]. Am J Clin Oncol, 2012, 35(3): 284-9.

[3]Chaskis C, Neyns B, Michotte A, et al. Pseudoprogression after radiotherapy with concurrent temozolomide for high-grade glioma: clinical observations and working recommendations [J]. Surg Neurol, 2009, 72(4): 423-8.

[4]Rowe L S, Butman J A, Mackey M, et al. Differentiating pseudoprogression from true progression: analysis of radiographic, biologic, and clinical clues in GBM [J]. J Neurooncol, 2018, 139(1): 145-152.

[5]Holland E C. Glioblastoma multiforme: the terminator [J]. Proceedings of the National Academy of Sciences, 2000, 97(12): 6242-6244.

[6]Rusthoven K E, Olsen C, Franklin W, et al. Favorable prognosis in patients with high-grade glioma with radiation necrosis: the University of Colorado reoperation series [J]. International Journal of Radiation Oncology* Biology* Physics, 2011, 81(1): 211-217.

[7]Okada H, Weller M, Huang R, et al. Immunotherapy response assessment in neuro-oncology: a report of the RANO working group [J]. The Lancet Oncology, 2015, 16(15): e534-e542.

[8]Johnson D R, Glenn C A, Javan R, et al. Congress of Neurological Surgeons systematic review and evidence-based guidelines update on the role of imaging in the management of progressive glioblastoma in adults [J]. Journal of Neuro-Oncology, 2021: 1-27.

[9]Park Y W, Choi D, Park J E, et al. Differentiation of recurrent glioblastoma from radiation necrosis using diffusion radiomics with machine learning model development and external validation [J]. Sci Rep, 2021, 11(1): 2913.