M Craigie1, J Squires2, K Miles3. 1. Department of Medical Imaging, Princess Alexandra Hospital, 199 Ipswich Road, Woolloongabba, 4102, QLD, Australia. Electronic address: michelle.craigie@uqconnect.edu.au. 2. Department of Medical Imaging, Princess Alexandra Hospital, 199 Ipswich Road, Woolloongabba, 4102, QLD, Australia. 3. Department of Medical Imaging, Princess Alexandra Hospital, 199 Ipswich Road, Woolloongabba, 4102, QLD, Australia; Institute of Nuclear Medicine, University College London, Euston Road, London, UK.
Abstract
AIM: To undertake a preliminary assessment of the potential for computed tomography (CT) measurement of tumour heterogeneity to stratify risk of nodal metastasis in patients with non-small cell lung cancer (NSCLC). MATERIALS AND METHODS: Tumour heterogeneity in CT images from combined positron-emission tomography (PET)/CT examinations in 150 consecutive patients with NSCLC was assessed using CT texture analysis (CTTA). The short axis diameter of the largest mediastinal node was also measured. Forty-two patients without distant metastases subsequently had tumour nodal status confirmed at surgery (n=26) or endobronchial ultrasound (EBUS; n=16). CTTA parameters and largest nodal diameter were related to nodal status using the rank correlation and the risk ratio for each nodal stage (>N0, >N1, >N2) was compared between patients categorised as high and low risk by CTTA or nodal size. The most significant predictor of nodal status was related to overall survival using Kaplan-Meier analysis. RESULTS: N-stage was more significantly correlated with CTTA than nodal diameter (Rs = -0.39, p=0.011, Rs = -0.45, p=0.0025, Rs = -0.40, p=0.0091 for normalised standard deviation (SD), normalised entropy and kurtosis respectively; Rs = -0.39, p=0.042 for nodal diameter). The presence of two or more high-risk CTTA values was the greatest risk factor for mediastinal metastasis (risk ratio: 11.0, 95% confidence interval: 1.56-77.8, p=0.0014) and was associated with significantly poorer overall survival (p=0.016). CONCLUSION: CTTA in NSCLC is related to nodal status in patients without distant metastases and has the potential to inform selection of investigative strategies for the assessment of mediastinal malignancy.
AIM: To undertake a preliminary assessment of the potential for computed tomography (CT) measurement of tumour heterogeneity to stratify risk of nodal metastasis in patients with non-small cell lung cancer (NSCLC). MATERIALS AND METHODS:Tumour heterogeneity in CT images from combined positron-emission tomography (PET)/CT examinations in 150 consecutive patients with NSCLC was assessed using CT texture analysis (CTTA). The short axis diameter of the largest mediastinal node was also measured. Forty-two patients without distant metastases subsequently had tumour nodal status confirmed at surgery (n=26) or endobronchial ultrasound (EBUS; n=16). CTTA parameters and largest nodal diameter were related to nodal status using the rank correlation and the risk ratio for each nodal stage (>N0, >N1, >N2) was compared between patients categorised as high and low risk by CTTA or nodal size. The most significant predictor of nodal status was related to overall survival using Kaplan-Meier analysis. RESULTS: N-stage was more significantly correlated with CTTA than nodal diameter (Rs = -0.39, p=0.011, Rs = -0.45, p=0.0025, Rs = -0.40, p=0.0091 for normalised standard deviation (SD), normalised entropy and kurtosis respectively; Rs = -0.39, p=0.042 for nodal diameter). The presence of two or more high-risk CTTA values was the greatest risk factor for mediastinal metastasis (risk ratio: 11.0, 95% confidence interval: 1.56-77.8, p=0.0014) and was associated with significantly poorer overall survival (p=0.016). CONCLUSION:CTTA in NSCLC is related to nodal status in patients without distant metastases and has the potential to inform selection of investigative strategies for the assessment of mediastinal malignancy.
Authors: Tai H Dou; Thibaud P Coroller; Joost J M van Griethuysen; Raymond H Mak; Hugo J W L Aerts Journal: PLoS One Date: 2018-11-02 Impact factor: 3.240