A fractal-integrated CT score predicts high-risk pathology and early recurrence in pancreatic ductal adenocarcinoma
European Radiology · 03.11.2026
DOI: 10.1007/s00330-026-12992-9 ↗Kurz gefasst
Fachartikel in der Zeitschrift European Radiology, online erschienen am 03.11.2026. Verfasst von Jing Zhu und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, West China Hospital, Sichuan University, No. 37, Guoxue Alley, 610041, Chengdu, China. Die Arbeit nennt 30 Literaturangaben. Gefördert von: Sichuan Science and Technology Program, 1.3.5 Project for Disciplines of Excellence, West China Hospital, Sichuan University, Development project of Hainan provincial clinical medical center and post-doctoral station development project of Sanya.
Abstract (Originalsprache)
Objective
To develop and validate a noninvasive CT-based imaging score incorporating fractal analysis to predict high-risk pathological features in pancreatic ductal adenocarcinoma (PDAC) and to assess its prognostic value for early recurrence. Materials and methods This retrospective bicenter study included 514 patients with pathologically confirmed PDAC who underwent curative resection, divided into training ( n = 219), internal validation ( n = 94), external validation ( n = 100) and outcome datasets ( n = 101). The fractal dimension (FD) and lacunarity were extracted from portal venous-phase CT images via box-counting algorithms. Imaging predictors of high-risk pathological features, including high tumor grade, lymphovascular invasion (LVI) and perineural invasion (PNI), were identified by logistic regression and integrated into a comprehensive weighted imaging score. Cox regression analyses were performed to assess its prognostic value for early recurrence.
Results
Patients with high-risk pathological features exhibited greater FD and lacunarity. The final score included six imaging features: necrosis, peripancreatic fat infiltration, venous involvement, regional lymphadenopathy, FD and lacunarity. For any high-risk pathology, the fractal-integrated score achieved AUC of 0.94, 0.88 and 0.84 in the training, internal validation and external validation datasets, respectively. For high pathological burden, the AUCs were 0.92, 0.80, and 0.77. In the outcome dataset, the score was an independent predictor of early recurrence (HR = 2.71; p = 0.004) and improved prognostic discrimination when combined with pathologic predictors.
Conclusion
This CT-based score, integrating fractal biomarkers, enables accurate, noninvasive preoperative prediction of aggressive pathological features in PDAC and provides incremental value for early recurrence risk stratification. Key Points Question Can the fractal-based imaging features be used to predict high-risk pathological features and guide risk stratification for early recurrence in pancreatic ductal adenocarcinoma (PDAC)?
Findings
This CT-based score integrating fractal analysis predicted high-risk pathological features and was independent predictors of early recurrence and provided incremental prognostic value beyond pathologic factors . Clinical relevance This CT-based imaging score may help identify patients at high-risk of aggressive pathology before surgery and may further complement postoperative pathological risk stratification for early recurrence . Graphical Abstract
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Radiology, West China Hospital, Sichuan University, No. 37, Guoxue Alley, 610041, Chengdu, China.
- Department of Radiology, Affiliated Hospital of North Sichuan Medical College, 637000, Nanchong, China.
- CT Imaging Research Center, GE Healthcare China, 100176, Beijing, China.
- Division of Pancreatic Surgery, Department of General Surgery, West China Hospital, Sichuan University, Chengdu, China.
- Pancreatic Cancer Comprehensive Ward, ShangJin NanFu Hospital, Chengdu, China.
- Department of Radiology, West China Hospital, Sichuan University, No. 37, Guoxue Alley, 610041, Chengdu, China. songlab_radiology@163.com.
Förderung
- Sichuan Science and Technology Program
- 1.3.5 Project for Disciplines of Excellence, West China Hospital, Sichuan University
- Development project of Hainan provincial clinical medical center and post-doctoral station development project of Sanya
Zitieren
APA
Zhu, J., Yu, H., Ren, W., Zhang, Y., Song, B., & Huang, Z. (2026). A fractal-integrated CT score predicts high-risk pathology and early recurrence in pancreatic ductal adenocarcinoma. European Radiology. https://doi.org/10.1007/s00330-026-12992-9
Vancouver
Zhu J, Yu H, Ren W, Zhang Y, Song B, Huang Z. A fractal-integrated CT score predicts high-risk pathology and early recurrence in pancreatic ductal adenocarcinoma. European Radiology. 2026. doi:10.1007/s00330-026-12992-9
BibTeX
@article{Zhu2026,
author = {Zhu, Jing and Yu, Haopeng and Ren, Wei and Zhang, Yi and Song, Bin and Huang, Zixing},
title = {A fractal-integrated CT score predicts high-risk pathology and early recurrence in pancreatic ductal adenocarcinoma},
journal = {European Radiology},
year = {2026},
doi = {10.1007/s00330-026-12992-9}
}
EUVIMED