Malignant pleural mesothelioma segmentation for photodynamic therapy planning

Our plain-English summary of Malignant pleural mesothelioma segmentation for photodynamic therapy planning, by W. Brahim, M. Mestiri, N. Betrouni, et al (Computerized Medical Imaging and Graphics, 2018).

The study

• Authors: W. Brahim, M. Mestiri, N. Betrouni, K. Hamrouni
• Journal: Computerized Medical Imaging and Graphics
• Published: 2018;65:79-92
• Cited by: 4 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Medical imaging modalities such as computed tomography (CT) combined with computer-aided diagnostic processing have already become important part of clinical routine specially for pleural diseases. The segmentation of the thoracic cavity represents an extremely important task in medical imaging for different reasons. Multiple features can be extracted by analyzing the thoracic cavity space and these features are signs of pleural diseases including the malignant pleural mesothelioma (MPM) which is the main focus of our research. This paper presents a method that detects the MPM in the thoracic cavity and plans the photodynamic therapy in the preoperative phase. This is achieved by using a texture analysis of the MPM region combined with a thoracic cavity segmentation method. The algorithm to segment the thoracic cavity consists of multiple stages. First, the rib cage structure is segmented using various image processing techniques. We used the segmented rib cage to detect feature points which represent the thoracic cavity boundaries. Next, the proposed method segments the structures of the inner thoracic cage and fits 2D closed curves to the detected pleural cavity features in each slice. The missing bone structures are interpolated using a prior knowledge from manual segmentation performed by an expert. Next, the tumor region is segmented inside the thoracic cavity using a texture analysis approach. Finally, the contact surface between the tumor region and the thoracic cavity curves is reconstructed in order to plan the photodynamic therapy. Using the adjusted output of the thoracic cavity segmentation method and the MPM segmentation method, we evaluated the contact surface generated from these two steps by comparing it to the ground truth. For this evaluation, we used 10 CT scans with pathologically confirmed MPM at stages 1 and 2. We obtained a high similarity rate between the manually planned surface and our proposed method. The average value of Jaccard index was about 0.79 and we obtained a value of 0.88 using the Dice index.

Copyright © 2017 Elsevier Ltd. All rights reserved.

In plain English

This was a technical study of computer image analysis. It did not study asbestos exposure. The authors developed and tested a method for analysing chest CT scans of people with malignant pleural mesothelioma, to help plan photodynamic therapy (a light-based treatment used with surgery) before the operation. The method outlined the chest cavity, identified tumor by its texture on the scan, and mapped where tumor met the chest cavity boundary. The authors used 30 non-contrasted CT scans from Calmette Hospital, Lille, France (23 male, 7 female, aged 47 to 70). Compared with manual tracing, the chest cavity outline scored an average Jaccard index of 0.935 and Dice index of 0.966 (1 means perfect overlap). In 10 scans with pathologically confirmed stage 1 and 2 mesothelioma, the planned treatment surface scored about 0.79 (Jaccard) and 0.88 (Dice). The authors reported that results depended strongly on manually placed starting points and that some tumor was missed. They called the method promising but said it needs further improvement and evaluation.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Development and evaluation of a computerized CT image segmentation method for planning photodynamic therapy, compared against manual segmentation
• Who was studied: 30 non-contrasted CT scans from the Calmette Hospital, Lille, France (23 male, 7 female, ages 47 to 70). The treatment-surface planning step was evaluated on 10 CT scans with pathologically confirmed malignant pleural mesothelioma at stage 1 and 2.
• Main results:
  • Chest (thoracic) cavity outlining compared with manual segmentation gave an average Jaccard index of 0.935 and an average Dice index of 0.966.
  • Similarity percentage for the chest cavity outline was 91.88% using the female shape model and 92.02% using the male shape model.
  • The chest cavity outline was less accurate in the upper thorax, where similarity varied from 71.91% to 91.94%.
  • For the tumor segmentation step, the best patch combination gave a Jaccard index of about 0.73, over-segmentation around 17% and under-segmentation about 18% (as stated in the text).
  • In the 10 stage 1 and 2 mesothelioma scans, the automatically planned treatment surface compared with the manually planned surface gave an average Jaccard index of about 0.79 and a Dice index of 0.88.
  • For the planned treatment surface in these 10 scans, the average under-segmentation rate was around 10.88% and the average over-segmentation rate was about 19.55%.
  • Generating the treatment surface took an average processing time of around 312 seconds.
• The authors' conclusion: The authors concluded that their method seems promising and could be added to the mesothelioma treatment protocol after further improvements and evaluations.
• Limitations the authors state:
  • The result of the chest cavity segmentation depends strongly on the manually placed starting points. Misplacing them caused large deformation of the outline.
  • Under-segmentation (missing parts of the tumor) persisted and needs further improvement of the algorithm.
  • Over-segmentation occurred in the mediastinal region, near the sternum and spinal regions.
  • Segmentation in the upper thorax was less accurate than in other parts of the thorax.
  • Users had to correct contact points manually to fix errors in the planned surface.

Topics

The National Library of Medicine indexes this study under subjects our site covers: Lung cancer, Mesothelioma.

Medical terms in this study, explained in our Lexicon: Pleural Mesothelioma, Pleural Cavity, CT Scan.

Related studies

• Three-dimensional finite-element based deformable image registration for evaluation of pleural cavity irradiation during photodynamic therapy — R. Penjweini, M.M. Kim, et al. (2017)
• Intrapleural Photodynamic Therapy for Mesothelioma: What Place and Which Future? — C. Munck, S.R. Mordon, et al. (2015)
• Illumination profile characterization of a light device for the dosimetry of intra-pleural photodynamic therapy for mesothelioma — C. Munck, S. Mordon, et al. (2016)
• Real-time light dosimetry for intra-cavity photodynamic therapy: Application for pleural mesothelioma treatment — N. Betrouni, C. Munck, et al. (2017)
• Phase III randomized trial of surgery with or without intraoperative photodynamic therapy and postoperative immunochemotherapy for malignant pleural mesothelioma — H.I. Pass, B.K. Temeck, et al. (1997)
• Radical pleurectomy and intraoperative photodynamic therapy for malignant pleural mesothelioma — J.S. Friedberg, M.J. Culligan, et al. (2012)

Browse return to medical research.