Kareem Wahid1, Abdallah Mohammed1, Lisanne Van Dijk1, Sara Ahmed1, Renjie He1, Baher Elgohari1, Yao Ding2, Jihong Wang2, Stephen Lai3, and Clifton Fuller1
1Radiation Oncology, MD Anderson Cancer Center, Houston, TX, United States, 2Radiation Physics, MD Anderson Cancer Center, Houston, TX, United States, 3Head and Neck Surgery, MD Anderson Cancer Center, Houston, TX, United States
Synopsis
Current methodology in
analyzing pathological lymph node radiotherapy response in head and neck cancer
relies on qualitative visual inspection of geometrical features. A quantitative
approach that takes into account dynamic changes in underlying lymph node
pathology could be an important supplemental clinical tool in radiotherapy
treatment planning. Herein, we investigate dynamic contrast enhanced MRI quantitative
parametric maps as possible biomarker candidates for lymph node radiotherapy response
by analyzing multi-timepoint images of head and neck cancer patients.
Introduction
Head and neck cancer (HNC) is a devastating disease that affects a large
portion of individuals every year. Conventional treatment involves radiotherapy
which targets the gross tumor volume (GTV). Malignant cervical lymph nodes are considered
as part of the GTV in HNC and viewed as negative prognostic indicators of disease
recurrence, distant metastases, and survival rates 1. The development of noninvasive imaging biomarkers
for use in treatment planning to detect early nodal response to therapy could
have the potential to improve radiotherapy efficacy in HNC such as through dose
modification. Dynamic contrast enhanced (DCE)-MRI is a functional imaging
technique that has been linked to tissue
perfusion and microvascular status
through quantitative parametric maps 2 but has only been modestly investigated in lymph
nodes with a focus on static comparisons between benign and malignant lesions 3,4 . Therefore, we attempt a proof of
concept study to determine if DCE-MRI parametric maps can be candidates for temporal
biomarkers of mid-therapy lymph node response.
Methods
Patients were enrolled under an IRB-approved protocol as part of an ongoing prospective clinical trial. All patients underwent two MRI scans: a baseline pre-therapy scan before beginning radiotherapy and a mid-therapy scan 3-4 weeks after radiotherapy initiation. Complete response (CR) and non-CR were determined through response evaluation criteria in solid tumors (RECIST1.1) criteria 5. Manual segmentations of pathologic lymph nodes receiving irradiation as part of the GTV were performed by a radiologist using T2-weighted MRI anatomical sequences for each time point, co-registered to DCE-MRI sequences, and propagated to respective quantitative DCE parametric maps
(Figure 1). Mean values for 11 quantitative (derived from pharmacokinetic models) and semi-quantitative parametric maps including the area under the curve in tissue (AUC), normalized AUC (NAUC), Tofts Model (TM) 6, Extended Tofts Model (ETM) were calculated for each segmented lesion at each timepoint. These maps are referred to as AUCc, AUCs, NAUCc, NAUCs, TMKep, TMKtrans, TMVe, ETMKep, ETMKtrans, ETMVe, and ETMVp. The Wilcoxon signed-rank test was used to assess temporal variation of quantitative parametric maps for all lesions. The Mann-Whitney U test was used to compare lesions that achieved CR at mid-therapy with those that did not. Changes in parametric maps were also associated with volume through a Pearson correlation to determine the relationship between these parameters and a more commonly available clinical metric. Results
A total of 27
patients with 38 lymph node lesions were included in the final analysis. 14 lesions
had a CR while 24 lesions had a non-CR. Mid-therapy AUCc, AUCs, and NAUCs were
significantly higher than pre-therapy values for all lesions (p<0.005). TMKep,
and ETMKep were significantly lower than pre-therapy values for all lesions
(p<0.05). Pre-therapy NAUCc, TMKtrans, TMVe, ETMKtrans, ETMVe, and ETMVp
were not significantly different than mid-therapy values for all lesions (p>0.05).
Changes in AUCc, AUCs, TMKep, and ETMKep were significantly higher for CR
lesions when compared to non-CR lesions (p<0.005). Changes in NAUCc, NAUCs, TMVe,
and ETMVe were significantly lower for CR lesions when compared to non-CR
lesions (p<0.05). Changes in TMKtrans, ETMKtrans, and ETMVp were not significantly
different for CR lesions when compared to non-CR lesions (p>0.05). No strong
correlations for any of the changes in parametric maps were found for changes
in volume, (r = -0.12, -0.28, -0.08, -0.13, 0.35, 0.10, -0.07, 0.25, -0.02,
-0.04, -0.01 for AUCc, AUCs, NAUCc, NAUCs, TMKep, TMKtrans, TMVe, ETMKep,
ETMKtrans, ETMVe, and ETMVp respectively). Discussion
Herein we have shown that certain quantitative parametric maps derived from DCE-MRI demonstrate significantly measurable changes during radiotherapy and can differentiate lymph nodes with mid-treatment CR or non-CR. These parametric maps have been linked to vascularity and blood flow distribution in previous literature, demonstrating their underlying pathological significance 2. Our results highlight the potential of these parameters to help target therapeutic intervention based on their temporal changes as indicators of response. Moreover, we demonstrate that these parametric maps are not strongly correlated to reductions in volume, a commonly available geometric quantification of tumor response, and therefore may yield novel information into the underlying pathological response of lymph node metastasis when compared to traditional qualitative metrics. Conclusions
We suggest DCE-MRI
contains underlying pathological information related to changes in lymph nodes,
primarily related to vascularity, and may yield additional or supplementary information
for clinician to use in the future for radiation therapy planning. Quantitative
imaging such as that achieved by DCE-MRI may be a useful tool to help target
personalized medicine therapeutics in oncology by indicating therapeutic response
potential of malignant tissues. Building off these preliminary results, we will
expand our analysis to a larger cohort of patients with external validation, probe
deeper into the underlying pathological significance of these parametric map
changes, and utilize parametric maps in predictive modeling of patient clinical
outcomes. Acknowledgements
Supported by a training
fellowship from The University of Texas Health Science Center at Houston Center for Clinical and Translational Sciences
TL1 Program (Grant No. TL1 TR003169).References
1. Leemans, C. R., Tiwari, R., Nauta, J.
J., Waal, I. V. D. & Snow, G. B. Recurrence at the primary site in head and
neck cancer and the significance of neck lymph node metastases as a prognostic
factor. Cancer 73, 187-190 (1994).
2. Chikui,
T. et al. The principal of dynamic
contrast enhanced MRI, the method of pharmacokinetic analysis, and its
application in the head and neck region. International
journal of dentistry 2012
(2012).
3. Yan,
S. et al. Characterization of
cervical lymph nodes using DCE-MRI: differentiation between metastases from SCC
of head and neck and benign lymph nodes. Clinical
hemorheology and microcirculation 64,
213-222 (2016).
4. Cintra,
M. B., Ricz, H., Mafee, M. F. & Santos, A. C. d. Magnetic resonance
imaging: dynamic contrast enhancement and diffusion-weighted imaging to
identify malignant cervical lymph nodes. Radiologia
brasileira 51, 71-75 (2018).
5. Eisenhauer,
E. A. et al. New response evaluation
criteria in solid tumours: revised RECIST guideline (version 1.1). European journal of cancer 45, 228-247 (2009).
6. Sourbron,
S. P. & Buckley, D. L. On the scope and interpretation of the Tofts models
for DCE‐MRI. Magnetic resonance in
medicine 66, 735-745 (2011).