Top

Open Access 25-04-2024 | Positron Emission Tomography | Oncology

Development and evaluation of two open-source nnU-Net models for automatic segmentation of lung tumors on PET and CT images with and without respiratory motion compensation

Authors: Montserrat Carles, Dejan Kuhn, Tobias Fechter, Dimos Baltas, Michael Mix, Ursula Nestle, Anca L. Grosu, Luis Martí-Bonmatí, Gianluca Radicioni, Eleni Gkika

Published in: European Radiology

Abstract

Objectives

In lung cancer, one of the main limitations for the optimal integration of the biological and anatomical information derived from Positron Emission Tomography (PET) and Computed Tomography (CT) is the time and expertise required for the evaluation of the different respiratory phases. In this study, we present two open-source models able to automatically segment lung tumors on PET and CT, with and without motion compensation.

Materials and methods

This study involved time-bin gated (4D) and non-gated (3D) PET/CT images from two prospective lung cancer cohorts (Trials 108237 and 108472) and one retrospective. For model construction, the ground truth (GT) was defined by consensus of two experts, and the nnU-Net with 5-fold cross-validation was applied to 560 4D-images for PET and 100 3D-images for CT. The test sets included 270 4D- images and 19 3D-images for PET and 80 4D-images and 27 3D-images for CT, recruited at 10 different centres.

Results

In the performance evaluation with the multicentre test sets, the Dice Similarity Coefficients (DSC) obtained for our PET model were DSC(4D-PET) = 0.74 ± 0.06, improving 19% relative to the DSC between experts and DSC(3D-PET) = 0.82 ± 0.11. The performance for CT was DSC(4D-CT) = 0.61 ± 0.28 and DSC(3D-CT) = 0.63 ± 0.34, improving 4% and 15% relative to DSC between experts.

Conclusions

Performance evaluation demonstrated that the automatic segmentation models have the potential to achieve accuracy comparable to manual segmentation and thus hold promise for clinical application. The resulting models can be freely downloaded and employed to support the integration of 3D- or 4D- PET/CT and to facilitate the evaluation of its impact on lung cancer clinical practice.

Clinical relevance statement

We provide two open-source nnU-Net models for the automatic segmentation of lung tumors on PET/CT to facilitate the optimal integration of biological and anatomical information in clinical practice. The models have superior performance compared to the variability observed in manual segmentations by the different experts for images with and without motion compensation, allowing to take advantage in the clinical practice of the more accurate and robust 4D-quantification.

Key Points

Lung tumor segmentation on PET/CT imaging is limited by respiratory motion and manual delineation is time consuming and suffer from inter- and intra-variability.
Our segmentation models had superior performance compared to the manual segmentations by different experts.
Automating PET image segmentation allows for easier clinical implementation of biological information.

Smith TA (1998) FDG uptake, tumour characteristics and response to therapy: a review. Nuclear Med Commun 19:97–105. https://doi.org/10.1097/00006231-199802000-00002CrossRef

Coleman RE (2002) Value of FDG-PET scanning in management of lung cancer. Lancet 359:1361–1362. https://doi.org/10.1016/s0140-6736(02)08388-5CrossRefPubMed

UyBico SJ, Wu CC, Suh RD, Le NH, Brown K, Krishnam MS (2010) Lung cancer staging essentials: the new tnm staging system and potential imaging pitfalls. Radiographics 30:1163–1181CrossRefPubMed

Pommier P, Touboul E, Chabaud S et al (2010) Impact of (18)F-FDG PET on treatment strategy and 3D radiotherapy planning in non-small cell lung cancer: a prospective multicenter study. AJR Am J Roentgenol 195:350–355CrossRefPubMed

Postmus PE, Kerr KM, Oudkerk M et al (2017) Early and locally advanced non-small-cell lung cancer (nsclc): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 28:iv1–iv21CrossRefPubMed

Manafi-Farid R, Askari E, Shiri I et al (2022) [(18)F]FDG-PET/CT radiomics and artificial intelligence in lung cancer: technical aspects and potential clinical applications. Semin Nucl Med 52:759–780. https://doi.org/10.1053/j.semnuclmed.2022.04.004CrossRefPubMed

Carles M, Fechter T, Radicioni G et al (2021) FDG-PET radiomics for response monitoring in non-small-cell lung cancer treated with radiation therapy. Cancers (Basel) 13:814. https://doi.org/10.3390/cancers13040814

Kong F, Hu C, Machta M et al (2021) OA02.04 randomized phase II trial (RTOG1106) on midtreatment PET/CT guided adaptive radiotherapy in locally advanced non-small cell lung cancer. J Thoracic Oncol 16:S104–S105CrossRef

Vinod SK, Hau E (2020) Radiotherapy treatment for lung cancer: current status and future directions. Respirology 25:61–71CrossRefPubMed

10.

Nestle U, Schimek-Jasch T, Kremp S et al (2020) Imaging-based target volume reduction in chemoradiotherapy for locally advanced non-small-cell lung cancer (PET-plan): a multicentre, open-label, randomised, controlled trial. Lancet Oncol 21:581–592CrossRefPubMed

11.

Cooke SA, de Ruysscher D, Reymen B et al (2023) [¹⁸F]FDG-PET guided vs whole tumour radiotherapy dose escalation in patients with locally advanced non-small cell lung cancer (PET-Boost): results from a randomised clinical trial. Radiother Oncol 181:109492. https://doi.org/10.1016/j.radonc.2023.109492CrossRefPubMed

12.

Gkika E, Grosu AL, Nestle U (2023) The use of 18 F-FDG PET/CT for radiotherapy treatment planning in non-small cell lung cancer: a mini-review. Precision Cancer Med 6:1–7.CrossRef

13.

Bettinardi V, Picchio M, Di Muzio N, Gianolli L, Gilardi MC, Messa C (2010) Detection and compensation of organ/lesion motion using 4D-PET/CT respiratory gated acquisition techniques. Radiother Oncol 96:311–316. https://doi.org/10.1016/j.radonc.2010.07.014CrossRefPubMed

14.

Büther F, Vehren T, Schäfers KP, Schäfers M (2016) Impact of data-driven respiratory gating in clinical PET. Radiology 281:229–238CrossRefPubMed

15.

Huang TC, Chou KT, Wang YC, Zhang G (2014) Motion freeze for respiration motion correction in PET/CT: a preliminary investigation with lung cancer patient data. BioMed Res Int https://doi.org/10.1155/2014/167491

16.

Chang G, Chang T, Pan T, Clark JW, Mawlawi OR (2010) Joint correction of respiratory motion artifact and partial volume effect in lung/thoracic PET/CT imaging. Med Phys 37:6221–6232CrossRefPubMedPubMedCentral

17.

Dawood M, Büther F, Lang N, Schober O, Schäfers KP (2007) Respiratory gating in positron emission tomography: a quantitative comparison of different gating schemes. Med Phys 34:3067–3076CrossRefPubMed

18.

Nehmeh SA, Erdi YE, Rosenzweig KE et al (2003) Reduction of respiratory motion artifacts in PET imaging of lung cancer by respiratory correlated dynamic PET: methodology and comparison with respiratory gated PET. J Nucl Med 44:1644–1648PubMed

19.

Park SJ, Ionascu D, Killoran JJ et al (2008) Evaluation of the combined effects of target size, respiratory motion and background activity on 3D and 4D PET/CT images. Phys Med Biol 53:3661–3679CrossRefPubMed

20.

Chirindel A, Adebahr S, Schuster D et al (2015) Impact of 4D- 18FDG- PET/CT imaging on target volume delineation in SBRT patients with central versus peripheral lung tumors. Multi-reader comparative study. Radiother Oncol 115:335–341CrossRefPubMed

21.

Heath E, Unkelbach J, Oelfke U (2009) Incorporating uncertainties in respiratory motion into 4D treatment plan optimization. Med Phys 36:3059–3071. https://doi.org/10.1118/1.3148582CrossRefPubMed

22.

Bittermann G, Scheifele C, Prokic V et al (2013) Description of a method: computer generated virtual model for accurate localisation of tumour margins, standardised resection, and planning of radiation treatment in head & neck cancer surgery. J Craniomaxillofac Surg 41:279–281. https://doi.org/10.1016/j.jcms.2012.10.011CrossRefPubMed

23.

Carles M, Torres-Espallardo I, Alberich-Bayarri A et al (2016) Evaluation of PET texture features with heterogeneous phantoms: complementarity and effect of motion and segmentation method. Phys Med Biol 62:652–9251CrossRefPubMed

24.

Carles M, Bach T, Torres-Espallardo I, Baltas D, Nestle U, Martí-Bonmatí L (2018) Significance of the impact of motion compensation on the variability of pet image features. Phys Med Biol 63:065013CrossRefPubMed

25.

Zaidi H, El Naqa I (2010) PET-guided delineation of radiation therapy treatment volumes: a survey of image segmentation techniques. Eur J Nucl Med Mol Imaging 37:2165–2187. https://doi.org/10.1007/s00259-010-1423-3

26.

Morra J, Tu Z, Toga A, Thomson P (2010) Machine learning for brain image segmentation. Biomedical image analysis and machine learning technologies: applications and techniques, edited by Gonzalez FA and Romero E, IGI Global, 102–126. https://doi.org/10.4018/978-1-60566-956-4.ch005

27.

Kao D, Cheze Le Rest C, Jaouen V, Hatt M (2019) Artificial intelligence, machine (deep) learning and radio(geno)mics: definitions and nuclear medicine imaging applications. Eur J Nucl Med Mol Imaging 46:2630–2637. https://doi.org/10.1007/s00259-019-04373-w

28.

Litjens G, Kooi T, Bejnordi BE et al (2017) A survey on deep learning in medical image analysis. Med Image Anal 42:60–88CrossRefPubMed

29.

Halder A, Dey D, Sadhu AK (2020) Lung nodule detection from feature engineering to deep learning in Thoracic CT images: a comprehensive review. J Digit Imaging 33:655–677. https://doi.org/10.1007/s10278-020-00320-6CrossRefPubMedPubMedCentral

30.

Liu X, Li KW, Yang R, Geng LS (2021) Review of deep learning based automatic segmentation for lung cancer radiotherapy. Front Oncol 11:717039. https://doi.org/10.3389/fonc.2021.717039CrossRefPubMedPubMedCentral

31.

Wang S, Mahon R, Weiss E et al (2022) Automated lung cancer segmentation using a PET and CT dual-modality deep learning neural network. Int J Radiat Oncol Biol Phys 115:529–539. https://doi.org/10.1016/j.ijrobp.2022.07.2312CrossRefPubMed

32.

Xiang D, Zhang B, Lu Y, Deng S (2023) Modality-specific segmentation network for lung tumor segmentation in PET-CT images. IEEE J Biomed Health Inform 27:1237–1248. https://doi.org/10.1109/JBHI.2022.3186275

33.

Nestle U, Adebahr S, Kaier K et al (2020) Quality of life after pulmonary stereotactic fractionated radiotherapy (SBRT): results of the phase II STRIPE trial. Radiother Oncol 148:82–88CrossRefPubMed

34.

Martinez-Movilla A, Mix M, Torres-Espallardo I et al (2022) Comparison of protocols with respiratory-gated (4D) motion compensation in PET/CT: open-source package for quantification of phantom image quality. EJNMMI Phys 9:80. https://doi.org/10.1186/s40658-022-00509-4CrossRefPubMedPubMedCentral

35.

Nestle U, De Ruysscher D, Ricardi U et al (2018) ESTRO ACROP guidelines for target volume definition in the treatment of locally advanced non-small cell lung cancer. Radiother Oncol 127:1–5. https://doi.org/10.1016/j.radonc.2018.02.023CrossRefPubMed

36.

Carles M, Fechter T, Nemmer U et al (2015) Feasibility of a semi-automated contrast-oriented algorithm for tumor segmentation in retrospectively gated PET images: phantom and clinical validation. Phys Med Biol https://doi.org/10.1088/0031-9155/60/24/9227

37.

Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18:203–211. https://doi.org/10.1038/s41592-020-01008-zCrossRefPubMed

38.

Zou KH, Warfield SK, Bharatha A et al (2004) Statistical validation of image segmentation quality based on a spatial overlap index. Acad Radiol 11:178–189CrossRefPubMedPubMedCentral

39.

Huttenlocher DP, Klanderman GA, Rucklidge WJ (1993) Comparing images using the Hausdorff distance. IEEE Trans Pattern Anal Mach Intell 15:850–863. https://doi.org/10.1109/34.232073CrossRef

40.

Zhang F, Wang Q, Li H (2020) Automatic segmentation of the gross target volume in non-small cell lung cancer using a modified version of ResNet. Technol Cancer Res Treat 19:1533033820947484. https://doi.org/10.1177/1533033820947484CrossRefPubMedCentral

41.

Jiang J, Hu YC, Liu CJ et al (2019) Multiple resolution residually connected feature streams for automatic lung tumor segmentation from CT images. IEEE Trans Med Imaging 38:134–144. https://doi.org/10.1109/TMI.2018.2857800CrossRefPubMed

42.

Bi N, Wang J, Zhang T et al (2019) Deep learning improved clinical target volume contouring quality and efficiency for postoperative radiation therapy in non-small cell lung cancer. Front Oncol 9:1192. https://doi.org/10.3389/fonc.2019.01192CrossRefPubMedPubMedCentral

43.

Wang C, Tyagi N, Rimner A et al (2019) Segmenting lung tumors on longitudinal imaging studies via a patient- specific adaptive convolutional neural network. Radiother Oncol 131:101–107. https://doi.org/10.1016/j.radonc.2018.10.037CrossRefPubMed

44.

Zhao X, Li L, Lu W, Tan S (2018) Tumor co-segmentation in PET/CT using multi-modality fully convolutional neural network. Phys Med Biol 64:015011. https://doi.org/10.1088/1361-6560/aaf44bCrossRefPubMedPubMedCentral

45.

Zhong Z, Kim Y, Zhou L et al (2018) 3D fully convolutional networks for co-segmentation of tumors on PET-CT images. Proc IEEE Int Symp Biomed Imaging 2018:228–231. https://doi.org/10.1109/isbi.2018.8363561PubMedPubMedCentral

46.

Kao YS, Yang J (2022) Deep learning-based auto-segmentation of lung tumor PET/CT scans: a systematic review. Clin Transl Imaging 10:217–223. https://doi.org/10.1007/s40336-022-00482-zCrossRef

47.

TIME. [Online] http://cmp.felk.cvut.cz/projects/LungTIME/, 2009. Accessed 14 Nov 2017

48.

ELCAP. [Online] http://www.via.cornell.edu/databases/lungdb.html/, 2013. Accessed 25 Nov 2017

49.

LISS. [Online] http://isc.cs.bit.edu.cn/MLMR/LISS.html/, 2015. Accessed 26 Nov 2017

Title: Development and evaluation of two open-source nnU-Net models for automatic segmentation of lung tumors on PET and CT images with and without respiratory motion compensation
Authors: Montserrat Carles
Dejan Kuhn
Tobias Fechter
Dimos Baltas
Michael Mix
Ursula Nestle
Anca L. Grosu
Luis Martí-Bonmatí
Gianluca Radicioni
Eleni Gkika
Publication date: 25-04-2024
Publisher: Springer Berlin Heidelberg
Keywords: Positron Emission Tomography
Computed Tomography
Computed Tomography
Lung Cancer
Lung Cancer
Lung Cancer
Artificial Intelligence
Published in: European Radiology
Print ISSN: 0938-7994
Electronic ISSN: 1432-1084
DOI: https://doi.org/10.1007/s00330-024-10751-2

At a glance: The ONWARDS insulin icodec trials

Springer Medicine

Development and evaluation of two open-source nnU-Net models for automatic segmentation of lung tumors on PET and CT images with and without respiratory motion compensation

Abstract

Objectives

Materials and methods

Results

Conclusions

Clinical relevance statement

Key Points

At a glance: The ONWARDS insulin icodec trials

Springer Medicine

Abstract

Objectives

Materials and methods

Results

Conclusions

Clinical relevance statement

Key Points

Please log in to get access to this content