Freely available convolutional neural network-based quantification of PET/CT lesions is associated with survival in patients with lung cancer

Pablo Borrelli; José Luis Loaiza Góngora; Reza Kaboteh; Johannes Ulén; Olof Enqvist; Elin Trägårdh; Lars Edenbrandt

doi:10.1186/s40658-022-00437-3

Freely available convolutional neural network-based quantification of PET/CT lesions is associated with survival in patients with lung cancer

EJNMMI Phys. 2022 Feb 3;9(1):6. doi: 10.1186/s40658-022-00437-3.

Authors

Pablo Borrelli^#¹, José Luis Loaiza Góngora^#¹, Reza Kaboteh¹, Johannes Ulén², Olof Enqvist^{2

3}, Elin Trägårdh^{4

5}, Lars Edenbrandt^{6

7}

Affiliations

¹ Department of Clinical Physiology, Region Västra Götaland, Sahlgrenska University Hospital, Gothenburg, Sweden.
² Eigenvision AB, Malmö, Sweden.
³ Department of Electrical Engineering, Chalmers University of Technology, Gothenburg, Sweden.
⁴ Department of Translational Medicine and Wallenberg Centre for Molecular Medicine, Lund University, Malmö, Sweden.
⁵ Department of Clinical Physiology and Nuclear Medicine, Skåne University Hospital, Malmö, Sweden.
⁶ Department of Clinical Physiology, Region Västra Götaland, Sahlgrenska University Hospital, Gothenburg, Sweden. lars.edenbrandt@gu.se.
⁷ Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden. lars.edenbrandt@gu.se.

^# Contributed equally.

Abstract

Background: Metabolic positron emission tomography/computed tomography (PET/CT) parameters describing tumour activity contain valuable prognostic information, but to perform the measurements manually leads to both intra- and inter-reader variability and is too time-consuming in clinical practice. The use of modern artificial intelligence-based methods offers new possibilities for automated and objective image analysis of PET/CT data.

Purpose: We aimed to train a convolutional neural network (CNN) to segment and quantify tumour burden in [¹⁸F]-fluorodeoxyglucose (FDG) PET/CT images and to evaluate the association between CNN-based measurements and overall survival (OS) in patients with lung cancer. A secondary aim was to make the method available to other researchers.

Methods: A total of 320 consecutive patients referred for FDG PET/CT due to suspected lung cancer were retrospectively selected for this study. Two nuclear medicine specialists manually segmented abnormal FDG uptake in all of the PET/CT studies. One-third of the patients were assigned to a test group. Survival data were collected for this group. The CNN was trained to segment lung tumours and thoracic lymph nodes. Total lesion glycolysis (TLG) was calculated from the CNN-based and manual segmentations. Associations between TLG and OS were investigated using a univariate Cox proportional hazards regression model.

Results: The test group comprised 106 patients (median age, 76 years (IQR 61-79); n = 59 female). Both CNN-based TLG (hazard ratio 1.64, 95% confidence interval 1.21-2.21; p = 0.001) and manual TLG (hazard ratio 1.54, 95% confidence interval 1.14-2.07; p = 0.004) estimations were significantly associated with OS.

Conclusion: Fully automated CNN-based TLG measurements of PET/CT data showed were significantly associated with OS in patients with lung cancer. This type of measurement may be of value for the management of future patients with lung cancer. The CNN is publicly available for research purposes.

Keywords: Computer-assisted analysis; Prognosis; Total lesion glycolysis; Tumour burden.

Grants and funding

ALFGBG-720751/ALF