FDG PET versus CT radiomics to predict outcome in malignant pleural mesothelioma patients

Our plain-English summary of FDG PET versus CT radiomics to predict outcome in malignant pleural mesothelioma patients, by M. Pavic, M. Bogowicz, J. Kraft, et al (EJNMMI Research, 2020).

The study

• Authors: M. Pavic, M. Bogowicz, J. Kraft, D. Vuong, M. Mayinger, S.G.C. Kroeze, M. Friess, T. Frauenfelder, N. Andratschke, M. Huellner, W. Weder, M. Guckenberger, S. Tanadini-Lang, I. Opitz
• Journal: EJNMMI Research
• Published: 2020;10(1):81
• Cited by: 38 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Background: Careful selection of malignant pleural mesothelioma (MPM) patients for curative treatment is of highest importance, as the multimodal treatment regimen is challenging for patients and harbors a high risk of substantial toxicity. Radiomics-a quantitative method for image analysis-has shown its prognostic ability in different tumor entities and could therefore play an important role in optimizing patient selection for radical cancer treatment. So far, radiomics as a prognostic tool in MPM was not investigated.

Materials And Methods: This study is based on 72 MPM patients treated with surgery in a curative intent at our institution between 2009 and 2017. Pre-treatment Fluorine-18 fluorodeoxyglucose (FDG) PET and CT scans were used for radiomics outcome modeling. After extraction of 1404 CT and 1410 FDG PET features from each image, a preselection by principal component analysis was performed to include only robust, non-redundant features for the cox regression to predict the progression-free survival (PFS) and the overall survival (OS). Results were validated on a separate cohort. Additionally, SUVmax and SUVmean, and volume were tested for their prognostic ability for PFS and OS.

Results: For the PFS a concordance index (c-index) of 0.67 (95% CI 0.52-0.82) and 0.66 (95% CI 0.57-0.78) for the training cohort (n = 36) and internal validation cohort (n = 36), respectively, were obtained for the PET radiomics model. The PFS advantage of the low-risk group translated also into an OS advantage. On CT images, no radiomics model could be trained. SUV max and SUV mean were also not prognostic in terms of PFS and OS.

Conclusion: We were able to build a successful FDG PET radiomics model for the prediction of PFS in MPM. Radiomics could serve as a tool to aid clinical decision support systems for treatment of MPM in future.

In plain English

This retrospective study from University Hospital Zurich tested whether computer analysis of pre-treatment scan images, called radiomics, could predict progression-free survival and overall survival in patients with malignant pleural mesothelioma. It included 72 patients treated between 2009 and 2017 with surgery intended to cure. They were split by treatment date into two groups of 36, one to build models and one to test them. The study did not examine asbestos exposure. A model built from FDG PET scan features predicted progression-free survival in the test group with a concordance index of 0.66 (95% confidence interval 0.57 to 0.78). It also separated test patients into risk groups with progression-free survival of 12.5 versus 9 months. Models built from CT scans did not hold up in the test group, and neither did a PET model built for overall survival. The standard PET measures SUVmax and SUVmean were not prognostic, while tumor volume was prognostic for overall survival. The authors concluded they had built a successful FDG PET model for progression-free survival. They noted the small number of patients and recommended validation with more data from multiple centers.

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: Retrospective single-institution cohort study developing and validating prognostic radiomics models (training cohort and separate validation cohort split by date of treatment)
• Who was studied: 72 patients with malignant pleural mesothelioma (MPM), selected from 123 referred to University Hospital Zurich, Switzerland, between 2009 and 2017. All had biopsy-confirmed diagnosis and curative-intent treatment including surgery (extrapleural pneumonectomy or pleurectomy/decortication). 60 of 72 received induction chemotherapy. There were 36 patients in the training cohort and 36 in the validation cohort.
• Exposure: No asbestos exposure was assessed. The predictors studied were image features from pre-treatment FDG PET and non-contrast CT scans: 1404 CT and 1410 FDG PET radiomic features, plus SUVmax, SUVmean and tumor volume.
• Main results:
  • Of the FDG PET features, 780 out of 1410 (55.3%) were stable against differences in how observers outlined the tumor, tested on 11 MPM cases.
  • The FDG PET radiomics model for progression-free survival had a mean c-index of 0.67 (95% CI 0.52–0.82) in 5-fold cross validation in the training cohort, and a c-index of 0.66 (95% CI 0.57–0.78) in the validation cohort.
  • The PET model split patients into groups with different progression-free survival. In the validation cohort this was 12.5 vs. 9 months (p less than 0.001). In the training cohort it was 11 vs. 7 months, with the p-value printed incompletely.
  • The PET progression-free survival model also discriminated for overall survival, with a c-index of 0.66 (95% CI 0.52–0.80).
  • The FDG PET model built for overall survival had a mean c-index of 0.72 (95% CI 0.62–0.84) in training cross validation, but showed poor prognostic power in the validation cohort.
  • CT radiomics models for progression-free and overall survival could not be successfully validated. Validation c-indices were 0.54 (0.44–0.67) for progression-free survival and 0.59 (0.47–0.74) for overall survival.
  • SUVmax had c-index 0.55 (0.44–0.60) for progression-free survival and 0.54 (0.42–0.68) for overall survival. SUVmean had c-index 0.52 (0.40–0.66) for progression-free survival and 0.48 (0.39–0.61) for overall survival in the text. All of these were in the training cohort.
  • Tumor volume was not prognostic for progression-free survival (c-index 0.60 training, 0.57 validation). It was prognostic for overall survival, with c-index 0.62 (0.52–0.72) in training and 0.63 (0.50–0.75) in validation.
• The authors' conclusion: The authors concluded that they built a successful FDG PET radiomics model for predicting progression-free survival in MPM, that no CT-based model with sufficient discriminative power could be built, and that the model should be validated with more data in a multi-centre setting.
• Limitations the authors state:
  • Scans were acquired on different machines with different settings, which can affect the robustness of radiomic features. Subgroups with uniform protocols were too small to study or correct for this.
  • The number of patients was low.
  • Many features were analyzed for only 36 training patients, creating a risk of overfitting.
  • Only a simple regression model was used. A more complex model might perform better but would require a much larger cohort.

Topics

Medical terms in this study, explained in our Lexicon: Pleural Mesothelioma, CT Scan, Progression-Free Survival, Overall Survival, Concordance Index.

Related studies

• Influence of inter-observer delineation variability on radiomics stability in different tumor sites — M. Pavic, M. Bogowicz, et al. (2018)
• Epigenetic profiles distinguish pleural mesothelioma from normal pleura and predict lung asbestos burden and clinical outcome — B.C. Christensen, E.A. Houseman, et al. (2009)
• Prognostic nutritional index predicts outcomes of malignant pleural mesothelioma — Z.H. Yao, G.Y. Tian, et al. (2013)
• MiR-score: a novel 6-microRNA signature that predicts survival outcomes in patients with malignant pleural mesothelioma — M.B. Kirschner, Y.Y. Cheng, et al. (2015)
• Downregulation of miR-99a/let-7c/miR-125b miRNA cluster predicts clinical outcome in patients with unresected malignant pleural mesothelioma — A. Truini, S. Coco, et al. (2017)
• Loss of BAP1 expression is associated with genetic mutation and can predict outcomes in gallbladder cancer — T. Hirosawa, M. Ishida, et al. (2018)

Browse return to medical research.