Elsevier

Clinical Radiology

Volume 67, Issue 2, February 2012, Pages 157-164
Clinical Radiology

Original Paper
Tumour heterogeneity in oesophageal cancer assessed by CT texture analysis: Preliminary evidence of an association with tumour metabolism, stage, and survival

https://doi.org/10.1016/j.crad.2011.08.012Get rights and content

Aim

To undertake a pilot study assessing whether tumour heterogeneity evaluated using computed tomography texture analysis (CTTA) has the potential to provide a marker of tumour aggression and prognosis in oesophageal cancer.

Materials and methods

In 21 patients, unenhanced CT images of the primary oesophageal lesion obtained using positron-emission tomography (PET)-CT examinations underwent CTTA. CTTA was carried out using a software algorithm that selectively filters and extracts textures at different anatomical scales between filter values 1.0 (fine detail) and 2.5 (coarse features) with quantification as entropy and uniformity (measures image heterogeneity). Texture parameters were correlated with average tumour 2-[18F]-fluoro-2-deoxy-d-glucose (FDG) uptake [standardized uptake values (SUVmean and SUVmax)] and clinical staging as determined by endoscopic ultrasound (nodal involvement) and PET-CT (distant metastases). The relationship between tumour stage, FDG uptake, and texture with survival was assessed using Kaplan–Meier analysis.

Results

Tumour heterogeneity correlated with SUVmax and SUVmean. The closest correlations were found for SUVmean measured as uniformity and entropy with coarse filtration (r = –0.754, p < 0.0001; and r = 0.748, p = 0.0001 respectively). Heterogeneity was also significantly greater in patients with clinical stage III or IV for filter values between 1.0 and 2.0 (maximum difference at filter value 1.5: entropy: p = 0.027; uniformity p = 0.032). The median (range) survival was 21 (4–34) months. Tumour heterogeneity assessed by CTTA (coarse uniformity) was an independent predictor of survival [odds ratio (OR)=4.45 (95% CI: 1.08, 18.37); p = 0.039].

Conclusion

CTTA assessment of tumour heterogeneity has the potential to identify oesophageal cancers with adverse biological features and provide a prognostic indicator of survival.

Introduction

The incidence of oesophageal cancer is notably high in Asia, Iceland, as well as the United Kingdom and United States.1 In 2010, the American Cancer Society estimated, approximately 16,640 new oesophageal cancer cases annually in the United States.2 The incidence is also rising in other western countries.3 Most people diagnosed with oesophageal cancer present with advanced disease. Despite the use of neoadjuvant chemotherapy and radiotherapy, the prognosis for patients with oesophageal cancer remains poor. The 5 year survival for tumours remaining within the oesophageal wall is about 40%; for tumours involving the adventitia of the oesophagus it is only 4%; and for patients without nodal involvement it is about 40%, diminishing to approximately 3% for those with nodal metastases.4 Novel means of stratifying risk for patients with oesophageal cancer could potentially lead to optimization of management strategies with improved outcomes.

Heterogeneity of the tumour microenvironment is a well-recognized feature of malignancy that is associated with adverse tumour biology. In particular, heterogeneity of the tumour blood supply results in the formation of hypoxic voids, which in turn are associated with oxidative stress, promotion of survival factors, and genomic instability.5 Hypoxia-inducible factors (HIFs) are important biochemical mediators in tissue responses to hypoxia. HIFs are transcription factors that up-regulate a range of processes associated with increased tumour aggression and treatment resistance.6 A heterogeneous blood supply will also impact on treatment response due to poor delivery of chemotherapeutic agents to areas of low vascularity.7 Hence, a non-invasive imaging method for assessment of tumour heterogeneity could potentially provide a biomarker for prognosis and treatment response.

Computed tomography (CT) remains the initial imaging method for clinical staging of oesophageal cancer, evaluating local spread into adjacent structures (T4 disease) and the presence of distant metastases (M1). If these features are absent on CT, subsequent endoscopic ultrasonography (EUS) is used to distinguish T1 tumours from higher-stage lesions and detect nodal involvement (N1 disease), whilst 2-[18F]-fluoro-2-deoxy-d-glucose (FDG) positron-emission tomography computed tomography (PET-CT) is most helpful in the detection of previously occult distant metastases.8 A meta-analysis has also shown that the intensity of FDG uptake also can provide a marker of likely survival and treatment response.9 One study has also reported that high FDG uptake is associated with poorly differentiated tumours.10 However, despite potential relevance to tumour biology, patient survival, and treatment response, none of these imaging methods provides objective assessment of tumour heterogeneity.

By assessing the distribution of grey-levels, coarseness, and regularity within a lesion, computed image analysis algorithms have the potential to provide additional morphological information relating to tumour heterogeneity. These texture analyses also have the advantage of quantifying tumour heterogeneity, something that cannot be achieved reliably by simple visual analysis. Recently, texture analysis performed on CT images of lung tumours has identified parameters reflecting tumour heterogeneity that were associated with advanced disease, increased FDG uptake on PET, and poor survival.11, 12, 13 The association between CT texture and FDG uptake likely reflects increased glucose metabolism in response to regional hypoxia as HIFs are known to up-regulate glucose metabolic pathways.6

As a preliminary step to ascertain whether tumour heterogeneity assessed with CT texture analysis (CTTA) has the potential to provide a similar marker of tumour aggression and predict patient survival in oesophageal cancer, the aim of the present pilot study was to correlate CTTA with staging, glucose metabolism, and patient survival.

Section snippets

Methods and materials

This single institution pilot study comprised a retrospective evaluation of image data from patients undergoing PET-CT and EUS for staging oesophageal cancer. The need for formal ethics evaluation for this retrospective evaluation was waived by the Institutional Review Board.

Results

Based on PET, CT, and EUS imaging, the number of patients with stage ≤II and >II were eight and 13, respectively (stage II, n = 8; stage III, n = 8; and stage IV, n = 5). Of the 21 patients with oesophageal cancer 14 had ADC and seven had SCC at histology. The median (range) tumour SUVmean and SUVmax for all the patients was 4.8 (1.9–9.8) and 11.4 (2.5–32.5). Table 1 summarizes patient demographics (clinical stage, histology, and SUV). Twelve of 21 patients died within 24 months of their PET-CT. The

Discussion

Structural heterogeneity is a recognized feature of malignancy and methods for measuring tumour heterogeneity on histological images have been described previously including application to prediction of progression to invasive cancer in patients with Barrett’s oesophagus.16 However, measurement of tumour heterogeneity on diagnostic images has to date, received relatively little attention particularly in CT, with few studies highlighting its importance in magnetic resonance imaging (MRI).17, 18

References (25)

  • L.E. Quint et al.

    Staging esophageal cancer

    Cancer Imaging

    (2008)
  • L.L. Pana et al.

    Prognostic significance of SUV on PET/CT in patients with esophageal cancer; a systematic review and meta-analysis

    Eur J Gastroenterol Hepatol

    (2009)
  • Cited by (276)

    • A primer on artificial intelligence in pancreatic imaging

      2023, Diagnostic and Interventional Imaging
    • Heterogeneity of Lung Density in Patients With Chronic Thromboembolic Pulmonary Hypertension (CTEPH)

      2022, Academic Radiology
      Citation Excerpt :

      Heterogeneity in tumor tissue is a well-recognized feature of malignancy (28,29). Histogram parameters can quantify the heterogeneity in the tissue and show the implications for diagnosis, treatment effect, identification of drug targets, and survival (15). However, the findings of histogram analysis have not been elucidated in patients with respiratory diseases, except in patients with lung cancer (30,31).

    View all citing articles on Scopus
    View full text