Radiomic signatures from T2W and DWI MRI are predictive of tumour hypoxia in colorectal liver metastases

(2023) 14:133 Bodalal et al. Insights into Imaging https://doi.org/10.1186/s13244-023-01474-x ORIGINAL ARTICLE Open Access Radiomic signatures from T2W and DWI MRI are predictive of tumour hypoxia in colorectal liver metastases Zuhir Bodalal1,2, Nino Bogveradze1,2,3, Leon C. ter Beek4, Jose G. van den Berg5, Joyce Sanders5, Ingrid Hofland6, Stefano Trebeschi1,2, Kevin B. W. Groot Lipman1,2, Koen Storck1, Eun Kyoung Hong1,2, Natalya Lebedyeva1, Monique Maas1, Regina G. H. Beets‑Tan1,2, Fernando M. Gomez1,7*   and Ieva Kurilova1 Abstract Background Tumour hypoxia is a negative predictive and prognostic biomarker in colorectal cancer typically assessed by invasive sampling methods, which suffer from many shortcomings. This retrospective proof-of-principle study explores the potential of MRI-derived imaging markers in predicting tumour hypoxia non-invasively in patients with colorectal liver metastases (CLM). Methods A single-centre cohort of 146 CLMs from 112 patients were segmented on preoperative T2-weighted (T2W) images and diffusion-weighted imaging (DWI). HIF-1 alpha immunohistochemical staining index (high/low) was used as a reference standard. Radiomic features were extracted, and machine learning approaches were imple‑ mented to predict the degree of histopathological tumour hypoxia. Results Radiomic signatures from DWI b200 (AUC = 0.79, 95% CI 0.61–0.93, p = 0.002) and ADC (AUC = 0.72, 95% CI 0.50–0.90, p = 0.019) were significantly predictive of tumour hypoxia. Morphological T2W TE75 (AUC = 0.64, 95% CI 0.42–0.82, p = 0.092) and functional DWI b0 (AUC = 0.66, 95% CI 0.46–0.84, p = 0.069) and b800 (AUC = 0.64, 95% CI 0.44–0.82, p = 0.071) images also provided predictive information. T2W TE300 (AUC = 0.57, 95% CI 0.33–0.78, p = 0.312) and b = 10 (AUC = 0.53, 95% CI 0.33–0.74, p = 0.415) images were not predictive of tumour hypoxia. Conclusions T2W and DWI sequences encode information predictive of tumour hypoxia. Prospective multicentre studies could help develop and validate robust non-invasive hypoxia-detection algorithms. Critical relevance statement Hypoxia is a negative prognostic biomarker in colorectal cancer. Hypoxia is usually assessed by invasive sampling methods. This proof-of-principle retrospective study explores the role of AI-based MRIderived imaging biomarkers in non-invasively predicting tumour hypoxia in patients with colorectal liver metastases (CLM). Key points • Tumour hypoxia is a valuable prognostic/predictive biomarker in colorectal cancer. • This proof-of-principle study demonstrates non-invasive associations between MR-radiomics and tumour hypoxia. • Radiomics from DWI b200, ADC, and T2W TE75 predicted tumour hypoxia. *Correspondence: Fernando M. Gomez Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Bodalal et al. Insights into Imaging (2023) 14:133 Page 2 of 13 Keywords Colorectal cancer, Colorectal liver metastasis, Hypoxia, MRI, Radiomics Graphical Abstract Radiomic signatures from T2W and DWI MRI are predictive of tumour hypoxia in colorectal liver metastases Hypoxia is a negative prognostic biomarker in colorectal cancer. Hypoxia is usually assessed by invasive sampling methods. This proof-of-principle retrospective study explores the role of AI-based MRI-derived imaging biomarkers in non-invasively predicting tumour hypoxia in patients with colorectal liver metastases (CLM). Insights Imaging (2023) Bodalal Z, Bogveradze N, ter Beek LC et al. DOI: 10.1186/s13244-023-01474-x Introduction Tumour hypoxia is a valuable prognostic and predictive biomarker in colorectal cancer (CRC) [1–4]. Genomewide and microRNA analyses have shown that CRC tumours with a hypoxic microenvironment showed significantly worsened clinical outcomes, especially disease-free survival [5, 6]. Resistance to chemotherapy, radiotherapy, and immunotherapy has been associated with tumour hypoxia in other solid tumours [7–13]. Furthermore, when evaluating the role of hypoxia on local therapies such as percutaneous ablation, transarterial chemotherapy, and radioembolisation, several studies have found that it also negatively impacts resistance to the treatment and/or induces a more aggressive clonal cell selection [14–17]. Hypoxia-driven elevation of proangiogenic factors is a hallmark of colorectal cancer and its liver metastasis [18]. Specifically, in the treatment of colorectal liver metastases, the hypoxic status has also been shown to influence the resistance to antiangiogenic drugs [19] or their susceptibility regarding radiation therapy, thereby impacting the dosimetric planning for Y-90 radioembolisation [20]. Moreover, in the era of immune therapies, hypoxia has also been revealed to play a major role in the current understanding of the immune microenvironment of colorectal liver metastases [21]. Polarographic electrodes inserted directly into the tumour are considered the gold standard for measuring tumour hypoxia. However, in the routine clinical workflow, histopathological analysis is more commonly performed, with tissue hypoxia markers, such as hypoxiainducible factor-1 (HIF-1) alpha, being the most relevant [22]. Both approaches suffer from similar shortcomings: invasiveness, limitation to accessible tumours, and inability to take tumour heterogeneity into account [23]. Moreover, these methods cannot provide longitudinal information on changes in the oxygenation of the microenvironment. Developing a non-invasive, robust imaging-based technique to assess tumour hypoxia would improve patient selection, treatment monitoring, and treatment modification. As medical image analysis research has gained recognition in the clinical world, increasingly relevant applications of radiomics coupled with machine learning have Bodalal et al. Insights into Imaging (2023) 14:133 emerged. Radiomic features and signatures have been associated with long-term prognosis and response to local and systemic therapy [24–28]. Prominent among the use cases for radiomics has been the domain of radiogenomics, where morphological phenotypes are linked to the underlying tumour genotype [29, (...truncated)