Computer-aided diagnosis of lung nodule using gradient tree boosting and Bayesian optimization

Nishio, Motohiro; Nishizawa, Masahiko; Sugiyama, Osamu; Kojima, Ryosuke; Yakami, Masahiro; Kuroda, Tomohiro; Togashi, Kaori

doi:10.1371/journal.pone.0195875

Cited by 86 publications

(58 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Classification and feature importance. The classifier used in this work was Extreme Gradient Boosting (XGBoost), which is a scalable and accurate implementation of gradient boosted trees algorithms 60 that has been used for lung cancer related works 61,62 . A benefit of using gradient boosting is that after the boosted trees are constructed, it is possible to retrieve the importance scores for each feature, based on how useful or valuable each feature was in the construction of the boosted decision trees within the model.…”

Section: Dimensionality Reductionmentioning

confidence: 99%

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Pinheiro

Pereira

Dias

et al. 2020

Sci Rep

View full text Add to dashboard Cite

EGFR and KRAS are the most frequently mutated genes in lung cancer, being active research topics in targeted therapy. The biopsy is the traditional method to genetically characterise a tumour. However, it is a risky procedure, painful for the patient, and, occasionally, the tumour might be inaccessible. This work aims to study and debate the nature of the relationships between imaging phenotypes and lung cancer-related mutation status. Until now, the literature has failed to point to new research directions, mainly consisting of results-oriented works in a field where there is still not enough available data to train clinically viable models. We intend to open a discussion about critical points and to present new possibilities for future radiogenomics studies. We conducted high-dimensional data visualisation and developed classifiers, which allowed us to analyse the results for EGFR and KRAS biological markers according to different combinations of input features. We show that EGFR mutation status might be correlated to CT scans imaging phenotypes; however, the same does not seem to hold for KRAS mutation status. Also, the experiments suggest that the best way to approach this problem is by combining nodule-related features with features from other lung structures. Lung cancer is the cancer type leading the incidence and mortality rates 1,2. This is linked to the fact that it is often diagnosed in an advanced stage, with 15% or less chance of a 5-year survival 3 , which magnifies the importance of treatments for advanced-stage disease. In Non-small-cell lung cancer (NSCLC), which constitutes 85% of all cases of lung cancer, certain genomic biomarkers are now considered predictive biomarkers and critical for the prognostic 4. Epidermal Growth Factor Receptor (EGFR) and Kristen Rat Sarcoma Viral Oncogene Homolog KRAS are the most frequently mutated gene in lung cancer 5. EGFR mutated is present in 15 to 50% of NSCLC patients from never-smokers 5. The two most common EGFR mutations (deletions in exon 19 and the single amino acid substitution L858R in exon 21) correspond to approximately 85% of the EGFR mutations in NSCLC. The other low frequency mutations include: point mutations, deletions, insertions, and duplications correspond to approximated 15% of EGFR mutations in NSCLC 6. Unlike the previous marker, KRAS is associated with tobacco use, with only 5 to 10% of KRAS-mutant lung cancers arising in never or light smokers 5,7. Surgically treated NSCLC patients with EGFR mutations showed better disease-free survival (DFS) and overall survival (OS) and the opposite was verified for KRAS, with worse DFS and OS 8. For cytotoxic chemotherapy, the role of EGFR and KRAS as a predictive marker is still unclear 9 ; however, it appears that mutant KRAS may predict a lack of response to chemotherapy 10. Regarding target therapy, tumour driver mutations have reliable predictive value and, in fact, they guide treatment decision in clinics 11. EGFR gene is a paradigmatic example, since its activating mutations, namely those loca...

show abstract

Section: Dimensionality Reductionmentioning

confidence: 99%

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Pinheiro

Pereira

Dias

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…For comparison, the AUC and of two radiologists were 0.898 & 0.822, and accuracies were 0.838 and 0.717. The XGBoost algorithm obtained human-level receiver operator curve's and outranked the SVM (42). XGBoost has also been evaluated for use in Clinical Decision Support systems for MRI (43) and differentiating uterine leiomyomas from sarcomas (44).…”

Section: Tree Based Boosting -Xgboostmentioning

confidence: 99%

Machine Learning Principles for Radiology Investigators

Borstelmann

2020

Academic Radiology

View full text Add to dashboard Cite

“…The classifier used in this work was Extreme Gradient Boosting (XGBoost), which is a scalable and accurate implementation of gradient boosted trees algorithms 47 that has been used for lung cancer related works 48,49 . A benefit of using gradient boosting is that after the boosted trees are constructed, it is possible to retrieve the importance scores for each feature, based on how useful or valuable each feature was in the construction of the boosted decision trees within the model.…”

Section: Classification and Feature Importancementioning

confidence: 99%

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Pinheiro

Pereira

Dias

et al. 2019

Preprint

View full text Add to dashboard Cite

EGFR and KRAS are the most frequently mutated genes in lung cancer, being active research topics in targeted therapy. Biopsy is the traditional method to genetically characterise a tumour. However, it is a risky procedure, painful for the patient, and, occasionally, the tumour might be inaccessible. This work aims to study and debate the nature of the relationships between imaging phenotypes and lung cancer-related mutation status. Until now, the literature has failed to point to new research directions, mainly consisting of results-oriented works in a field where there is still not enough available data to train clinically viable models. We intend to open a discussion about critical points and to present new possibilities for future radiogenomics studies. We conducted high-dimensional data visualisation and developed classifiers, which allowed us to analyse the results for EGFR and KRAS biological markers according to different combinations of input features. We show that EGFR mutation status might be correlated to CT scans imaging phenotypes, however, the same does not seem to hold true for KRAS mutation status. Also, the experiments suggest that the best way to approach this problem is by combining nodule-related features with features from other lung structures. 2/103/10 7/10 is balanced individually for each fold using SMOTE-NC, avoiding data leakage. After parameter optimisation, probabilistic outputs of each model with optimal parameters were analysed using the AUC of Receiver Operating Characteristic (ROC). ROC is a probability curve and AUC represents degree or measure of separability, telling how much model is capable of distinguishing between classes. The ROC curve is plotted with True Positive Rate (TPR) against the False Positive Rate (FPR), usually with TPR on the y-axis and FPR on the x-axis. Experimental DesignWe designed four experiments in order to test and compare which type of input features allow to achieve better performance in gene mutation status prediction. We first trained a model that took nodule-related radiomic features as input. Then, for direct comparison purposes and to allow a modular evaluation, we split the semantic data into three parts: nodule, non-nodule and hybrid. The first one contains only nodular information, the second one contains only information external to the nodule and the third one is the combination of both. The split can be seen in detail in Table 2 of the supplementary material. 10/10

show abstract

Computer-aided diagnosis of lung nodule using gradient tree boosting and Bayesian optimization

Cited by 86 publications

References 34 publications

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Machine Learning Principles for Radiology Investigators

Identifying relationships between imaging phenotypes and lung cancer-related mutation status: EGFR and KRAS

Contact Info

Product

Resources

About