GENERAL COMMENTARY article

Front. Oncol., 28 April 2025

Sec. Cancer Imaging and Image-directed Interventions

Volume 15 - 2025 | https://doi.org/10.3389/fonc.2025.1527723

Commentary: An insight to PDAC tumor heterogeneity across pancreatic subregions using computed tomography images

  • Department of Thoracic Surgery, Shenyang Chest Hospital & Tenth People’s Hospital, Shenyang, Liaoning, China

A Commentary on
An insight to PDAC tumor heterogeneity across pancreatic subregions using computed tomography images

By Javed S, Qureshi TA, Wang L, Azab L, Gaddam S, Pandol SJ and Li D (2024) Front. Oncol. 14:1378691. doi: 10.3389/fonc.2024.1378691

Introduction

I am writing to express my appreciation for the recent publication by Javed et al., titled”An insight to PDAC tumor heterogeneity across pancreatic subregions using computed tomography images” (1). This study offers a comprehensive exploration into the spatial heterogeneity of pancreatic ductal adenocarcinoma (PDAC) tumors using radiomics and machine learning techniques. The authors demonstrated that CT imaging can non-invasively differentiate tumors in the head from those in the body and tail regions, achieving promising classification accuracy. While the findings have significant implications for enhancing PDAC diagnosis and treatment strategies, I would like to offer some comments that could further enrich the discussion.

Discussion

The authors identified distinct radiomic features that differentiate PDAC tumors based on their pancreatic subregions. However, as acknowledged in the study, the integration of imaging features with genomic and molecular profiles was not explored. Existing research has shown that combining radiomic and molecular data can lead to more precise stratification of tumors and better prediction of treatment responses (2, 3). Currently, the molecular characteristics of non-small cell lung cancer (NSCLC) form the basis for clinical diagnosis and treatment of lesions (e.g., EGFR, KRAS), and the phenotypic associations between molecular subtypes and imaging semantic features have been widely studied. A recent study on NSCLC generated 10 meaningful gene sets through RNA sequencing analysis, including the epidermal growth factor (EGF) pathway (4). An imaging-genomics map was created, with 32 statistically significant correlations between semantic imaging features and gene sets. For example, nodule attenuation and margins are associated with late cell-cycle genes, and the gene set representing the EGF pathway is significantly correlated with the presence of ground-glass opacity and irregular nodules or nodules with ill-defined margins. The authors can refer to this study to further improve their work. A deeper investigation into the correlation between radiomic features and known molecular subtypes could enhance the predictive power of the proposed model and potentially guide personalized therapeutic strategies.

The authors primarily focused on the morphological and textural characteristics of PDAC tumors without considering the tumor microenvironment (TME). The TME plays a crucial role in PDAC progression and can significantly influence imaging features (5). For instance, desmoplastic stroma, a hallmark of PDAC, might contribute to variations in signal intensity across different pancreatic regions. Non-invasive imaging techniques for characterizing the TME have not yet entered routine clinical practice. Cancer staging and treatment response assessment still mainly rely on morphological imaging, focusing on measuring the size of the primary tumor and its metastases. Positron emission tomography (PET) is a clinically established imaging technique for assessing molecular characteristics within the TME. For example, prostate-specific membrane antigen (PSMA) PET has been successfully applied in clinical practice. However, PET can only detect one molecular characteristic at a time. In contrast, magnetic resonance imaging (MRI) has the potential to more comprehensively characterize the TME, as it can be used for non-invasive evaluation of the structure, function, and molecular characteristics of tumors (6). Multiparametric MRI (mpMRI) combines multiple MRI sequences, each capturing unique features of the TME, thereby providing a more detailed insight into the tumor lesion through a single examination. Simultaneous analysis of multiple tissue characteristics more effectively addresses the complexity of the TME, including its spatial and temporal heterogeneity, and reduces the risk of missing information relevant to treatment. The authors may consider incorporating multimodal imaging techniques into their research. In the future, features related to the TME may offer a more comprehensive understanding of tumor heterogeneity.

While the Naïve Bayes classifier achieved a commendable accuracy of 86%, it is worth considering that tumor boundaries, especially between the head and body regions, may not always be well-defined. Misclassification near these boundaries suggests that anatomical overlap may influence the radiomic features. Employing advanced deep learning models or ensemble approaches might improve classification performance, particularly for tumors located at the border between subregions (7, 8). The current study analyzed CT scans from a single time point. Given PDAC’s aggressive nature, it would be valuable to investigate how radiomic features evolve over time and correlate these changes with treatment response and disease progression (911). Longitudinal radiomics could offer new biomarkers for early detection of treatment resistance or disease recurrence.

To translate the findings of Javed et al. into clinical practice, several practical challenges need to be addressed. First, the integration of radiomics and molecular data requires powerful computational tools as well as expertise in imaging and genomics. Collaborative efforts among radiologists, oncologists, and bioinformaticians can facilitate the development of integrated models. Second, multimodal imaging approaches, such as PET/MRI, may require additional resources and expertise. Investing in advanced imaging technologies and training healthcare professionals can help overcome these challenges. Finally, longitudinal studies require consistent follow-up and data collection. Establishing standardized protocols for data collection and analysis can ensure the reliability and reproducibility of radiomic features over time.

Conclusion

In summary, Javed et al.’s study provides a valuable contribution to the understanding of PDAC heterogeneity using non-invasive imaging techniques. However, incorporating molecular data, TME features, and longitudinal analysis could further strengthen the study’s clinical impact. I hope the authors will consider these suggestions in their future research to expand on their promising findings.

Author contributions

JC: Writing – original draft. YF: Writing – review & editing.

Funding

The author(s) declare that no financial support was received for the research and/or publication of this article.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Generative AI was used in the creation of this manuscript.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. Javed S, Qureshi TA, Wang L, Azab L, Gaddam S, Pandol SJ, et al. An insight to PDAC tumor heterogeneity across pancreatic subregions using computed tomography images. Front Oncol. (2024) 14:1378691. doi: 10.3389/fonc.2024.1378691

PubMed Abstract | Crossref Full Text | Google Scholar

2. Moffitt RA, Marayati R, Flate EL, Volmar KE, Loeza SG, Hoadley KA, et al. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet. (2015) 47:1168–78. doi: 10.1038/ng.3398

PubMed Abstract | Crossref Full Text | Google Scholar

3. Majzner RG, Theruvath JL, Nellan A, Heitzeneder S, Cui Y, Mount CW, et al. CAR T cells targeting B7-H3, a pan-cancer antigen, demonstrate potent preclinical activity against pediatric solid tumors and brain tumors. Clin Cancer Res. (2019) 25:2560–74. doi: 10.1158/1078-0432.CCR-18-0432

PubMed Abstract | Crossref Full Text | Google Scholar

4. Zhou M, Leung A, Echegaray S, Gentles A, Shrager JB, Jensen KC, et al. Non-small cell lung cancer radiogenomics map identifies relationships between molecular and imaging phenotypes with prognostic implications. Radiology. (2018) 286:307–15. doi: 10.1148/radiol.2017161845

PubMed Abstract | Crossref Full Text | Google Scholar

5. Durante M, Orecchia R, Loeffler JS. Charged-particle therapy in cancer: clinical uses and future perspectives. Nat Rev Clin Oncol. (2017) 14:483–95. doi: 10.1038/nrclinonc.2017.30

PubMed Abstract | Crossref Full Text | Google Scholar

6. Hoffmann E, Masthoff M, Kunz WG, Seidensticker M, Bobe S, Gerwing M, et al. Multiparametric MRI for characterization of the tumour microenvironment. Nat Rev Clin Oncol. (2024) 21:428–48. doi: 10.1038/s41571-024-00891-1

PubMed Abstract | Crossref Full Text | Google Scholar

7. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. (2017) 542:115–8. doi: 10.1038/nature21056

PubMed Abstract | Crossref Full Text | Google Scholar

8. Liu C, Pang M. Extracting lungs from CT images via deep convolutional neural network based segmentation and two-pass contour refinement. J Digit Imaging. (2020) 33:1465–78. doi: 10.1007/s10278-020-00388-0

PubMed Abstract | Crossref Full Text | Google Scholar

9. Yun G, Kim YH, Lee YJ, Kim B, Hwang JH, Choi DJ. Tumor heterogeneity of pancreas head cancer assessed by CT texture analysis: association with survival outcomes after curative resection. Sci Rep. (2018) 8:7226. doi: 10.1038/s41598-018-25627-x

PubMed Abstract | Crossref Full Text | Google Scholar

10. Villaruz LC, Cobo M, Syrigos K, Mavroudis D, Zhang W, Kim JS, et al. A phase II study of nab-paclitaxel and carboplatin chemotherapy plus necitumumab in the first-line treatment of patients with stage IV squamous non-small cell lung cancer. Lung Cancer. (2019) 136:52–6. doi: 10.1016/j.lungcan.2019.08.009

PubMed Abstract | Crossref Full Text | Google Scholar

11. Luminari S, Manni M, Galimberti S, Versari A, Tucci A, Boccomini C, et al. Response-adapted postinduction strategy in patients with advanced-stage follicular lymphoma: the FOLL12 study. J Clin Oncol. (2022) 40:729–39. doi: 10.1200/JCO.21.01234

PubMed Abstract | Crossref Full Text | Google Scholar

Keywords: pancreatic ductal adenocarcinoma (PDAC), CT, pancreatic subregions, radiomics, pancreas cancer

Citation: Cao J and Feng Y (2025) Commentary: An insight to PDAC tumor heterogeneity across pancreatic subregions using computed tomography images. Front. Oncol. 15:1527723. doi: 10.3389/fonc.2025.1527723

Received: 19 November 2024; Accepted: 10 April 2025;
Published: 28 April 2025.

Edited by:

Sharon R Pine, University of Colorado Anschutz Medical Campus, United States

Reviewed by:

Touseef Ahmad Qureshi, Cedars Sinai Medical Center, United States

Copyright © 2025 Cao and Feng. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Yong Feng, eHlmZW5neW9uZzE2M0AxNjMuY29t

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.