Radiomics of soft tissue sarcoma metastases to assess prognostic factors related to intrapatient intertumor heterogeneity

Soft tissue sarcomas are a heterogenous group of mesenchymal neoplasms with variable malignant potential. Radiomics has been increasingly applied to imaging studies of soft tissue sarcomas for a variety of purposes, however, almost all studies have focused on analysis of the primary tumor and mostly on magnetic resonance imaging (MRI). Many studies have focused on using radiomics to predict the histological grade of soft tissue sarcomas (1-10). Others have focused on neoadjuvant chemotherapy or radiotherapy treatment response assessment (11-15). Some have concentrated on predicting the risk for metastatic disease (16-20). Some have evaluated risk of local recurrence (21-24). A few have assessed associations with tumor gene expression (8,25,26). More soft tissue sarcoma radiomics studies are centered on predicting patient prognosis and outcomes (27-30). To our knowledge, only one other paper has used radiomics to assess soft tissue sarcoma lung metastases on computed tomography (CT) to predict chemotherapy efficacy (31).

Geady et al. assess intrapatient radiomic variability in multi-metastatic soft tissue sarcoma patients to predict treatment response (32). This was a retrospective study of 397 multi-metastatic soft tissue sarcoma patients enrolled as part of the SARC021 trial (33) obtained through the Sarcoma Alliance for Research through Collaboration (SARC) with a Data Use Agreement (DUA), comprised of 165 leiomyosarcoma (41%), 56 liposarcoma (14%), 47 undifferentiated pleomorphic sarcoma (12%), and 129 other soft tissue sarcoma subtypes (33%). This study assessed 1,072 metastases on CT and introduces a new radiomics metric, Measured Intrapatient Radiomic Variability (MIRV), to quantify intertumor heterogeneity across all metastatic lesions within a patient. Prior studies have suggested correlations with geographically based radiomics assessments of texture and intratumoral heterogeneity, which may in turn serve as indicators for biological tumor aggressiveness (34). MIRV attempts to capture pairwise differences across metastatic lesions as a means to quantitatively express tumor heterogeneity on a per-patient level.

Radiomics features were extracted from metastatic lesions contoured on pre-treatment CT scans. Statistical analysis was then performed to evaluate the association of MIRV with overall survival (OS) (all patients); with volumetric treatment response data for a subset of 85 patients with 312 metastases using tumor-specific response classification (TSRC) for pulmonary metastases (all tumors in a patient met a predefined 33% volume reduction threshold) and response evaluation criteria in solid tumors (RECIST) 1.1 for the full cohort; and with liquid biopsy circulating tumor deoxyribonucleic acid (ctDNA) detection (35) in a subset of 53 leiomyosarcoma patients with 164 metastases.

MIRV (indicating greater intertumor heterogeneity) had a strong negative correlation with complete tumor response; a moderate correlation with volumetric response, independent of baseline tumor volume; and a moderate correlation with post-treatment ctDNA positivity. Although there was no significant association between MIRV and OS in the full cohort, higher MIRV was significantly associated with worse survival in leiomyosarcoma patients. As ctDNA is thought to reflect overall tumor burden, an association with baseline volumetrics was anticipated; however, it would appear that MIRV assessments capture features of intratumoral heterogeneity not reflected in rudimentary size-based assessments alone given that the MIRV Distance was not significantly influenced by baseline tumor volume. These findings suggest that MIRV may capture additional information about tumor biology that may impact treatment response and patient prognosis in ways that other traditional prognostic markers do not.

There are several weaknesses of this paper. There is no specific breakdown of the types of metastases included in this analysis. Although most are presumably pulmonary metastases, the non-pulmonary metastases would be of interest as a subset analysis. Also, including non-pulmonary metastases might also increase heterogeneity in the MIRV metric that might not otherwise be present in an analysis of pulmonary metastases only. The study also does not specify whether the CT scans were non-contrast or contrast-enhanced studies or both. It would be useful to know if there are effects on the MRIV metric based on the CT being performed with or without contrast. By focusing on the MIRV metric, this study does not assess other radiomics parameters that could have associations with volumetric response, ctDNA, or OS. This study also does not provide any assessment of the primary tumors and tumor grade is omitted from the multivariable survival model. The study also only looks at OS, not recurrence-free or progression-free survival. The study also mentions construction of predictive models for receiver operating characteristic (ROC) curve analysis with and without MIRV and RECIST, however, the details of model construction are not elaborated upon. Finally, although 397 patients are included in this study, 268 (67.5%) are cases of metastatic leiomyosarcoma, liposarcoma, or undifferentiated pleomorphic sarcoma. This means that only a small subset of the other 50+ subtypes of soft tissue sarcoma (36) are poorly represented or not included at all.

However, the authors should be congratulated for being only the second study to assess the radiomics of soft tissue sarcoma metastases and the first to examine intrapatient intertumor heterogeneity with the introduction of the novel MIRV metric. Further study could strengthen the demonstrated associations with neoadjuvant chemotherapy treatment response and for ctDNA and OS in the leiomyosarcoma subgroup and reveal potential future applications of radiomics and the MIRV metric for metastatic disease for specific soft tissue sarcoma subtypes or other primary tumors. The concept of quantitative radiomic assessment of tumor heterogeneity may also have utility for evaluation of syndromes with multiple tumors, such as neurofibromatosis, or hereditary cancer syndromes, such as Li-Fraumeni or multiple endocrine neoplasia.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Translational Cancer Research. The article has undergo external peer review.

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2223/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2223/coif). G.R.M. reports paid consultant for Canon Medical Systems, USA (ended Jan 2025) and Clinical Imaging (journal) Subject Editor for Emergency and Musculoskeletal Imaging. B.K.K.F. received prior consulting fees from Mendaera, Inc.; RSNA and institutional support (UCSF) stipends for conference and meeting attendance. B.K.K.F. is the Chair of the RSNA Resident and Fellow Committee; member of the American Board of Radiology Initial Certification Advisory Committee for Diagnostic Radiology, of the RSNA Education Council, and of the RSNA Annual Meeting Program Planning Committee; Associate Editor with Radiology: Imaging Cancer; and Associate Editor for Artificial Intelligence in Radiology with Frontiers in Radiology. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Peeken JC, Spraker MB, Knebel C, et al. Tumor grading of soft tissue sarcomas using MRI-based radiomics. EBioMedicine 2019;48:332-40. [Crossref] [PubMed]
2. Zhang Y, Zhu Y, Shi X, et al. Soft Tissue Sarcomas: Preoperative Predictive Histopathological Grading Based on Radiomics of MRI. Acad Radiol 2019;26:1262-8. [Crossref] [PubMed]
3. Wang H, Chen H, Duan S, et al. Radiomics and Machine Learning With Multiparametric Preoperative MRI May Accurately Predict the Histopathological Grades of Soft Tissue Sarcomas. J Magn Reson Imaging 2020;51:791-7. [Crossref] [PubMed]
4. Xu W, Hao D, Hou F, et al. Soft Tissue Sarcoma: Preoperative MRI-Based Radiomics and Machine Learning May Be Accurate Predictors of Histopathologic Grade. AJR Am J Roentgenol 2020;215:963-9. [Crossref] [PubMed]
5. Yan R, Hao D, Li J, et al. Magnetic Resonance Imaging-Based Radiomics Nomogram for Prediction of the Histopathological Grade of Soft Tissue Sarcomas: A Two-Center Study. J Magn Reson Imaging 2021;53:1683-96. [Crossref] [PubMed]
6. Dou Y, Li X, Tao J, et al. Prediction of high-grade soft-tissue sarcoma using a combined intratumoural and peritumoural MRI-based radiomics nomogram. Clin Radiol 2023;78:e1032-40. [Crossref] [PubMed]
7. Zhang L, Yang Y, Wang T, et al. Intratumoral and peritumoral MRI-based radiomics prediction of histopathological grade in soft tissue sarcomas: a two-center study. Cancer Imaging 2023;23:103. [Crossref] [PubMed]
8. Zhu YF, Li YS, Zhang Y, et al. Radiomics model based on intravoxel incoherent motion and diffusion kurtosis imaging for predicting histopathological grade and Ki-67 expression level of soft tissue sarcomas. Acta Radiol 2023;64:2541-51. [Crossref] [PubMed]
9. Schmitz F, Voigtländer H, Jang H, et al. Predicting the Malignancy Grade of Soft Tissue Sarcomas on MRI Using Conventional Image Reading and Radiomics. Diagnostics (Basel) 2024;14:2220. [Crossref] [PubMed]
10. Wang B, Guo H, Zhang M, et al. Prediction of soft tissue sarcoma grading using intratumoral habitats and a peritumoral radiomics nomogram: a multi-center preliminary study. Front Oncol 2024;14:1433196. [Crossref] [PubMed]
11. Crombé A, Périer C, Kind M, et al. T(2) -based MRI Delta-radiomics improve response prediction in soft-tissue sarcomas treated by neoadjuvant chemotherapy. J Magn Reson Imaging 2019;50:497-510. [Crossref] [PubMed]
12. Chen S, Li N, Tang Y, et al. Radiomics Analysis of Fat-Saturated T2-Weighted MRI Sequences for the Prediction of Prognosis in Soft Tissue Sarcoma of the Extremities and Trunk Treated With Neoadjuvant Radiotherapy. Front Oncol 2021;11:710649. [Crossref] [PubMed]
13. Peeken JC, Asadpour R, Specht K, et al. MRI-based delta-radiomics predicts pathologic complete response in high-grade soft-tissue sarcoma patients treated with neoadjuvant therapy. Radiother Oncol 2021;164:73-82. [Crossref] [PubMed]
14. Fields BKK, Demirjian NL, Cen SY, et al. Predicting Soft Tissue Sarcoma Response to Neoadjuvant Chemotherapy Using an MRI-Based Delta-Radiomics Approach. Mol Imaging Biol 2023;25:776-87. [Crossref] [PubMed]
15. Sherminie LPG, Jayatilake ML, Hewavithana PB, et al. Morphometry-based radiomics for predicting prognosis in soft tissue sarcomas of extremities following radiotherapy. Radiography (Lond) 2024;30:1501-7. [Crossref] [PubMed]
16. Vallières M, Freeman CR, Skamene SR, et al. A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Phys Med Biol 2015;60:5471-96. [Crossref] [PubMed]
17. Liang HY, Yang SF, Zou HM, et al. Deep Learning Radiomics Nomogram to Predict Lung Metastasis in Soft-Tissue Sarcoma: A Multi-Center Study. Front Oncol 2022;12:897676. [Crossref] [PubMed]
18. Casale R, De Angelis R, Coquelet N, et al. The Impact of Edema on MRI Radiomics for the Prediction of Lung Metastasis in Soft Tissue Sarcoma. Diagnostics (Basel) 2023;13:3134. [Crossref] [PubMed]
19. Hu Y, Wang X, Yue Z, et al. Radiomics of multi-parametric MRI for the prediction of lung metastasis in soft-tissue sarcoma: a feasibility study. Cancer Imaging 2024;24:119. [Crossref] [PubMed]
20. Yang M, Jin J. Radiomics model for predicting distant metastasis in soft tissue sarcoma of the extremities and trunk treated with surgery. Clin Transl Oncol 2025;27:2307-15. [Crossref] [PubMed]
21. Tagliafico AS, Bignotti B, Rossi F, et al. Local recurrence of soft tissue sarcoma: a radiomic analysis. Radiol Oncol 2019;53:300-6. [Crossref] [PubMed]
22. Liu S, Sun W, Yang S, et al. Deep learning radiomic nomogram to predict recurrence in soft tissue sarcoma: a multi-institutional study. Eur Radiol 2022;32:793-805. [Crossref] [PubMed]
23. Casale R, Varriano G, Santone A, et al. Predicting risk of metastases and recurrence in soft-tissue sarcomas via Radiomics and Formal Methods. JAMIA Open 2023;6:ooad025. [Crossref] [PubMed]
24. Wang Y, Wang T, Zheng F, et al. Fusion of Deep Transfer Learning and Radiomics in MRI-Based Prediction of Post-Surgical Recurrence in Soft Tissue Sarcoma. J Imaging Inform Med 2025; Epub ahead of print. [Crossref] [PubMed]
25. Crombé A, Bertolo F, Fadli D, et al. Distinct patterns of the natural evolution of soft tissue sarcomas on pre-treatment MRIs captured with delta-radiomics correlate with gene expression profiles. Eur Radiol 2023;33:1205-18. [Crossref] [PubMed]
26. Yang Y, Zhang L, Wang T, et al. MRI Fat-Saturated T2-Weighted Radiomics Model for Identifying the Ki-67 Index of Soft Tissue Sarcomas. J Magn Reson Imaging 2023;58:534-45. [Crossref] [PubMed]
27. Peeken JC, Neumann J, Asadpour R, et al. Prognostic Assessment in High-Grade Soft-Tissue Sarcoma Patients: A Comparison of Semantic Image Analysis and Radiomics. Cancers (Basel) 2021;13:1929. [Crossref] [PubMed]
28. Fadli D, Kind M, Michot A, et al. Natural Changes in Radiological and Radiomics Features on MRIs of Soft-Tissue Sarcomas Naïve of Treatment: Correlations With Histology and Patients' Outcomes. J Magn Reson Imaging 2022;56:77-96. [Crossref] [PubMed]
29. Crombé A, Lucchesi C, Bertolo F, et al. Integration of pre-treatment computational radiomics, deep radiomics, and transcriptomics enhances soft-tissue sarcoma patient prognosis. NPJ Precis Oncol 2024;8:129. [Crossref] [PubMed]
30. Yu Y, Guo H, Zhang M, et al. Multi-institutional validation of a radiomics signature for identification of postoperative progression of soft tissue sarcoma. Cancer Imaging 2024;24:59. [Crossref] [PubMed]
31. Miao L, Ma ST, Jiang X, et al. Prediction of the therapeutic efficacy of epirubicin combined with ifosfamide in patients with lung metastases from soft tissue sarcoma based on contrast-enhanced CT radiomics features. BMC Med Imaging 2022;22:131. [Crossref] [PubMed]
32. Geady C, Bannon JJ, Reza S, et al. Measured intrapatient radiomic variability as a predictor of treatment response in multi-metastatic soft tissue sarcoma patients. Sci Rep 2025;15:27838. [Crossref] [PubMed]
33. Tap WD, Papai Z, Van Tine BA, et al. Doxorubicin plus evofosfamide versus doxorubicin alone in locally advanced, unresectable or metastatic soft-tissue sarcoma (TH CR-406/SARC021): an international, multicentre, open-label, randomised phase 3 trial. Lancet Oncol 2017;18:1089-103. [Crossref] [PubMed]
34. Varghese BA, Fields BKK, Hwang DH, et al. Spatial assessments in texture analysis: what the radiologist needs to know. Front Radiol 2023;3:1240544. [Crossref] [PubMed]
35. Madanat-Harjuoja LM, Klega K, Lu Y, et al. Circulating Tumor DNA Is Associated with Response and Survival in Patients with Advanced Leiomyosarcoma. Clin Cancer Res 2022;28:2579-86. [Crossref] [PubMed]
36. Soft Tissue and Bone Tumours. 5th ed. WHO Classification of Tumours. Lyon, France: IARC Publications; 2020.