Assessment of algorithmic performance in PET-CT segmentation of solitary lung tumors
Duc Trung Le, Dániel Szabó, Sándor Aranyi, Márton Mikó, Zita Képes, Iván Mihovk, Anna Rebeka Kovács, Iván Gábor Nagy, Bence Farkas, István Hajdú, György Trencsényi, Lajos Trón, Miklós Emri, Gábor Opposits
Abstract
PET-CT (Positron emission tomography–computed tomography), a nuclear medicine imaging technique is preferably used for lung tumor detection. However, manual image analysis may be time-consuming. To enhance efficiency, many automated segmentation algorithms have been developed.
Introduction
Lung cancer remains the leading cause of cancer-related death worldwide, accounting for the highest mortality rates among both men and women [1]. Early and accurate diagnosis plays a crucial role in improving patient outcomes [2].
Materials and methods
This retrospective methodological study was conducted using 221 previously acquired PET-CT images of patients with lung tumors at University of Debrecen Hospital.
Results
The nnU-Net model 535 and RECOMIA segmentation predictions were compared with the manual masks using radiomics features agreement. Table 2 reports Brown’s method combined p-values based on uncorrected p-values, FDR and Bonferroni-adjusted p-values and the sums of absolute value of correlations.
Discussion
In this research, we proposed a radiomics-based comparison method to evaluate segmentation algorithms and applied it to two deep learning-based algorithms nnU-Net and RECOMIA to compare their performance on PET-CT solitary lung tumor segmentation.
Conclusion
In this research, we proposed a radiomics-based method to evaluate segmentation algorithms and applied it to compare the performance of two deep learning-based algorithms nnU-Net and RECOMIA for PET-CT solitary lung tumor segmentation.
Citation: Le DT, Szabó D, Aranyi S, Mikó M, Képes Z, Mihovk I, et al. (2026) Assessment of algorithmic performance in PET-CT segmentation of solitary lung tumors. PLoS One 21(11): e0360849. https://doi.org/10.1371/journal.pone.0360849
Editor: Carmelo Caldarella, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, ITALY
Received: January 20, 2026; Accepted: October 1, 2026; Published: November 5, 2026
Copyright: © 2026 Le et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data analyzed in this study, including sensitive patient information, will be made available upon reasonable request from the University of Debrecen Data Protection Officer in accordance with applicable ethical and data protection regulations. Address: University of Debrecen Data Protection Officer, Kassai út campus, H-4028 Debrecen, Kassai u. 26. Chancellery IV, ground floor 34, website: https://unideb.hu/en/data-protection-officer-dpo, email: adatvedelmi.tisztviselo@unideb.hu, phone: +36 52,512,700/74465. The source codes are available at https://github.com/opgabor/nnUNet-vs-RECOMIA.
Funding: This work was supported by the Ministry of Culture and Innovation of Hungary through the National Research, Development and Innovation Fund (Nemzeti Kutatási, Fejlesztési és Innovációs Alap) under the TKP2021-NKTA funding scheme (TKP2021-NKTA-34 to ME). Scanomed Ltd. provided support in the form of salaries for authors MM and BF. The funders had no additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are described in the ‘author contributions’ section.
Competing interests: The authors declare that they have no competing interests.
