Computerized segmentation and measurement of malignant pleural mesothelioma

Our plain-English summary of Computerized segmentation and measurement of malignant pleural mesothelioma, by W.F. Sensakovic, S.G. Armato, C. Straus, et al (Medical Physics, 2011).

The study

• Authors: W.F. Sensakovic, S.G. Armato, C. Straus, R.Y. Roberts, P. Caligiuri, A. Starkey, H.L. Kindler
• Journal: Medical Physics
• Published: 2011;38(1):238-44
• Cited by: 42 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Purpose: The current linear method to track tumor progression and evaluate treatment efficacy is insufficient for malignant pleural mesothelioma (MPM). A volumetric method for tumor measurement could improve the evaluation of novel treatments, but a fully manual implementation of volume measurement is too tedious and time-consuming. This manuscript presents a computerized method for the three-dimensional segmentation and volumetric analysis of MPM.

Methods: The computerized MPM segmentation method segments the lung parenchyma and hemithoracic cavities to define the pleural space. Nonlinear diffusion and a k-means classifier are then implemented to identify MPM in the pleural space. A database of 31 computed tomography scans from 31 patients with pathologically confirmed MPM was retrospectively collected. Three observers independently outlined five randomly selected sections in each scan. The Jaccard similarity coefficient (J) between each of the observers and between the observer-defined and computer-defined segmentations was calculated. The computer-defined and the observer-defined segmentation areas (averaged over all observers) were both calculated for each axial section and compared using Bland-Altman plots.

Results: The median J value among observers averaged over all sections was 0.517. The median J between the computer-defined and manual segmentations was 0.484. The difference between these values was not statistically significant. The area delineated by the computerized method demonstrated variability and bias comparable to the tumor area calculated from manual delineations.

Conclusions: A computerized method for segmentation and measurement of MPM was developed. This method requires minimal initialization by the user and demonstrated good agreement with manually drawn outlines and area measurements. This method will allow volumetric tracking of tumor progression and may improve the evaluation of novel MPM treatments.

In plain English

This study described a computer method for outlining and measuring malignant pleural mesothelioma on chest CT scans and tested it against outlines drawn by hand. The researchers retrospectively collected 31 scans from 31 patients with pathologically confirmed malignant pleural mesothelioma. After exclusions, 66 scan sections from 19 patients were used for testing. Three observers (two attending radiologists and one resident) outlined tumor on randomly chosen sections. Overlap between outlines was scored from 0 (no overlap) to 1 (identical). The median overlap between observers was 0.517, and between the computer and the observers it was 0.484. This difference was not statistically significant (p=0.731), but the study's statistical power was low (0.224). On average the computer method overestimated tumor area by 593 mm2 compared with the observers' average. That bias fell between the biases seen when pairs of observers were compared with each other (366 and 910 mm2). The authors concluded that the method showed good agreement with manual outlines and reduced observer time from hours to minutes. They added that its outlines still need manual refinement, especially at the lung bases. They named the small testing database as the major limitation.

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 retrospective testing of a computerized method for outlining and measuring tumor on CT scans, compared against outlines drawn by human observers
• Who was studied: 31 CT scans from 31 patients with pathologically confirmed malignant pleural mesothelioma (26 males, 5 females, aged 50 to 83). After excluding scans with prominent calcifications (6) or surgical intervention (2), 23 scans remained. The final test set was 66 sections from 19 patients.
• Main results:
  • The median overlap (Jaccard coefficient) among the three observers was 0.517. The median overlap between the computer outlines and the manual outlines was 0.484.
  • The computer-versus-observer overlap was not statistically significantly different from the observer-versus-observer overlap (p=0.731). Statistical power, calculated retrospectively, was 0.224.
  • The computer method overestimated tumor area by 593 mm2 on average compared with the mean of the observers' outlines. The bias between observers A and C was 366 mm2, and between observers B and C it was 910 mm2.
  • Of 11 sections that the observers agreed contained no disease, the computer method correctly identified 3 (27.3%) as containing no disease.
  • The largest computer segmentation errors tended to occur in the lung bases and the spaces between the ribs.
• The authors' conclusion: The authors concluded that the method needs minimal input from the user and agreed well with manual outlines and area measurements. They reported that it reduced observer time from hours to minutes, but said its outlines are not yet sufficient for direct use in staging or prognosis and will require manual refinement.
• Limitations the authors state:
  • The testing database was small, which the authors called the major limitation.
  • Statistical power was low because of the small number of test sections. The authors estimated that 644 sections would be needed for adequately powered testing.
  • Only sections where the observers agreed on the existence and position of tumor were used, which further reduced the testing database.
  • Several sections were taken from each patient. The authors noted that drawing a single section per patient would ensure the samples were not correlated.
  • The method was validated against two-dimensional outlines rather than three-dimensional observer-drawn volumes.

Topics

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

Medical terms in this study, explained in our Lexicon: Pleural Mesothelioma, Parenchyma.

Related studies

• The influence of initial outlines on manual segmentation — W.F. Sensakovic, A. Starkey, et al. (2010)
• Three-dimensional stereoscopic volume rendering of malignant pleural mesothelioma — N.M. Mollberg, N.M. Parsad, et al. (2012)
• Variability of tumor area measurements for response assessment in malignant pleural mesothelioma — Z.E. Labby, C. Straus, et al. (2013)
• Characterization of mesothelioma and tissues present in contrast-enhanced thoracic CT scans — N. Corson, W.F. Sensakovic, et al. (2011)
• Lung volume measurements as a surrogate marker for patient response in malignant pleural mesothelioma — Z.E. Labby, S.G. Armato, et al. (2013)
• Convolutional Neural Networks for Segmentation of Malignant Pleural Mesothelioma: Analysis of Probability Map Thresholds (CALGB 30901, Alliance) — M. Shenouda, E. Gudmundsson, et al. (2023)

Browse more research in Medical Physics, or return to medical research.