Elsevier

Physica Medica

Volume 32, Issue 12, December 2016, Pages 1672-1680
Physica Medica

Original paper
Characterization of cervical lymph-nodes using a multi-parametric and multi-modal approach for an early prediction of tumor response to chemo-radiotherapy

https://doi.org/10.1016/j.ejmp.2016.09.003Get rights and content

Highlights

  • Characterization of cervical lymph nodes for the prediction of treatment response.

  • Image-based features estimated from T2w-MRI and DW-MRI before and at mid-RT.

  • ADC and textural features before treatment are the best single predictor.

  • Combination of T2w-MRI and DW-MRI can better predict the final response.

Abstract

Purpose

In the treatment of Head-and-Neck Squamous Cell Carcinoma (HNSCC), the early prediction of residual malignant lymph nodes (LNs) is currently required. Here, we investigated the potential of a multi-modal characterization (combination of CT, T2w-MRI and DW-MRI) at baseline and at mid-treatment, based on texture analysis (TA), for the early prediction of LNs response to chemo-radiotherapy (CRT).

Methods

30 patients with pathologically confirmed HNSCC treated with CRT were considered. All patients underwent a planning CT and two serial MR examinations (including T2w and DW images), one before and one at mid-CRT. For each patient the largest malignant LN was selected and within each LN, morphological and textural features were estimated from T2w-MRI and CT, besides a quantification of the apparent diffusion coefficient (ADC) from DW-MRI. After a median follow-up time of 26.6 months, 19 LNs showed regional control, while 11 LNs showed regional failure at a median time of 4.6 months. Linear discriminant analysis was used to test the accuracy of the image-based features in predicting the final response.

Results

Pre-treatment features showed higher predictive power than mid-CRT features, the ADC having the highest accuracy (80%); CT-based indices were found not predictive. When ADC was combined with TA, the classification performance increased (accuracy = 82.8%). If only T2w-MRI features were considered, the best combination of pre-CRT indices and their variation reached an equivalent accuracy (81.8%).

Conclusion

Our results may suggest that TA on T2w-MRI and ADC can be combined together to obtain a more accurate prediction of response to CRT.

Introduction

Treatment of locally advanced Head-and-Neck Squamous Cell Carcinoma (HNSCC) is a challenging task, since it affects the overall quality of life of patients, and, despite chemo-radiotherapy (CRT) has been accepted as a standard management, the 5 year survival rate of these patients remains below 50% [1]. In this context, a timely identification of patients at risk of residual malignant lymph nodes (LNs) can help in choosing the best treatment [2]. However, the early prediction of treatment response presents many challenges and it is currently a hot topic in this research area, since reproducible and robust findings have not been established yet [3].

Recently, many studies have reported image-based criteria for treatment assessment, considering Computed Tomography (CT), morphological or functional Magnetic Resonance Imaging (MRI) [3], [4], [5], [6], [7], [8], [9]. The morphological criteria principally regard volume regression [5], [7], [9], the presence of focal abnormalities [3] and nodal density [8]. However, very heterogeneous results in the prediction performance using these criteria were reported in literature [10].

Great attention was given to Diffusion-weighted MRI (DW-MRI) [11] in the evaluation of diffusion properties of the primary tumor and malignant nodes, related to the treatment response [12]. In fact, it was reported that primary tumors and metastatic nodes characterized by lower pre-treatment values of apparent diffusion coefficient (ADC) can better respond to the CRT [13], [14], [15]. Moreover, ADC changes evaluated both during and early after CRT were found predictive of treatment response at the primary and nodal site [6], [10], [16].

Very few works [17], [18] tried to perform a multi-parametric analysis, but they were limited to a combination of ADC and morphological features or ADC and PET findings. Moreover, morphological evaluations were performed using qualitative findings or very simple quantitative approaches, able to characterize nodes density or shape [3], [9]. To our knowledge, there are no attempts to predict treatment response using advanced image analysis techniques, such as texture analysis, on anatomical images which can provide quantitative information about structural properties of the tissue, related to the spatial pattern. Only a very recent work proposed textural features applied on parametric maps derived on Dynamic Contrast Enhanced (DCE)-MRI for the prediction of tumor response [19].

Aim of this work was thus to investigate the potential of a multi-modal characterization descriptive of morphological, structural and functional properties at baseline and at mid-CRT, based on advanced image processing methods, for the early prediction of LNs response to CRT in patients with HNSCC. In particular, shape-related indices were calculated on T2w-MRI, textural features were extracted from T2w-MRI and planning CT, and ADC values from DW-MRI.

Section snippets

Patient population

From October 2011 to March 2014, a total of 30 patients with histologically proven HNSCC were included retrospectively in the present study. This retrospective analysis was performed on an existing data-set of a prospective study conducted at Regina Elena National Cancer Institute and previously approved by the ethics committee of the hospital. The primary end point of the study, currently closed, was to investigate the ability of DWI to predict the treatment response of HNSCC patients after

Patient population

Nineteen LNs (63.3%) showed RC after a median follow-up time of 26.6 months (range, 7.3, 50.2). Eleven LNs (36.7%) showed RF at a median time of 4.6 months (range, 1.5, 19.5), four of which had a persistence of the lesion at the first MR scan performed 8 weeks after CRT. Three of the 11 patients with RF developed also a primary tumor recurrence. In two patients, PET-CT acquired 12 weeks after the end of CRT documented a recurrence in a LN different from the largest one. The patients with RC and RF

Discussion

In this paper, a multi-modal and multi-parametric characterization of cervical lymph nodes was proposed to predict the specific response to chemo-radiotherapy in patients treated for HNSCC, using quantitative image-based features calculated before and at mid treatment.

The patient population analysis demonstrated that the two groups were homogeneous regarding baseline characteristics. On the contrary, the image-based characterization showed significant differences, especially in terms of pre-RT

Conclusion

The present work suggested that a multi-modal approach, aimed at characterizing cervical nodes with different textural and quantitative features, can early predict treatment response in patients with HNSCC with good accuracy. When textural features estimated on T2w-MRI were combined with basal ADC, the prediction power increased. Performance equivalent to ADC, which can be considered in this context as the gold-standard biomarker, can be obtained by the combination of textural features, both

Funding sources

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References (34)

  • D.P. Noij et al.

    Whole-body-MR imaging including DWIBS in the work-up of patients with head and neck squamous cell carcinoma: A feasibility study

    Eur J Radiol

    (2014)
  • J.A. Bonner et al.

    Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck

    N Engl J Med

    (2006)
  • A.D. King et al.

    Cervical nodal metastases from head and neck squamous cell carcinoma: MRI criteria for treatment assessment

    Head Neck

    (2016)
  • R. Davuluri et al.

    Image guided volumetric response during chemoradiotherapy for head and neck squamous cell carcinoma as a predictor of disease outcomes

    Am J Otolaryngol

    (2016)
  • Y. Ding et al.

    Intravoxel incoherent motion imaging kinetics during chemoradiotherapy for human papillomavirus-associated squamous cell carcinoma of the oropharynx: preliminary results from a prospective pilot study

    NMR Biomed

    (2015)
  • G. Sanguineti et al.

    Volumetric change of human papillomavirus-related neck lymph nodes before, during, and shortly after intensity-modulated radiation therapy

    Head Neck

    (2012)
  • Belli.C. Zerbetto et al.

    Early volume variation of positive lymph nodes assessed by in-room mega voltage CT images predicts risk of loco-regional relapses in head and neck cancer patients treated with intensity-modulated radiotherapy

    Acta Oncol

    (2015)
  • Cited by (23)

    • Quantitative magnetic resonance imaging on hybrid magnetic resonance linear accelerators: Perspective on technical and clinical validation

      2020, Physics and Imaging in Radiation Oncology
      Citation Excerpt :

      There is growing evidence that diffusion weighted (DW) MRI may provide prognostic information to predict RT treatment outcome in head and neck cancer (HNC) [7–11]. Also for other tumor sites, such as brain tumors [12,13], cervical cancer [14] or rectal cancer [15], the prognostic value of DW MRI for RT outcome prognosis has been shown. The majority of these studies correlated functional biomarkers derived from functional MRI before the start of RT with observed outcome following therapy.

    • PETER PHAN: An MRI phantom for the optimisation of radiomic studies of the female pelvis

      2020, Physica Medica
      Citation Excerpt :

      Moreover, the lack of a standard intensity scale for MR images necessitates the use of additional pre-processing, such as intensity normalisation and bias field correction [7]. Early clinical demonstrations of radiomics in MRI have been seen in prostate [8] and breast cancer [9], as well as glioblastoma multiforme [10] and head-and-neck squamous cell carcinoma [11]. Promising results have also been seen in female pelvic tumours, where preoperative radiomic features were seen to be predictive of lymph-vascular space invasion [12] and of lymph node metastasis [13] in cervical cancer patients.

    • Functional imaging early during (chemo)radiotherapy for response prediction in head and neck squamous cell carcinoma; a systematic review

      2019, Oral Oncology
      Citation Excerpt :

      An ADCmean increase after 1 week of treatment was reported higher in LRC than in LRF patients [47]. At 2–3 weeks two studies did not find significant differences in ΔADCmean between LRC and LRF patients [52,53], whereas three other studies showed an overall trend towards a higher ADCmean increase at 3 weeks intratreatment in LRC (22%[45], >25%[54], 100% increase [56]) compared with LRF (7%[45], not specified [54], 38% increase [56]). An ADCmean increase of 25% at 4 weeks was reported in LRC [55].

    View all citing articles on Scopus
    1

    E.S. and S.M. equally contributed to the work.

    View full text