Qi Zhang1, Xiaoduo Yu1, Jieying Zhang1, Han Ouyang1, Xinming Zhao1, and Lizhi Xie2
1Department of Radiology, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical, Beijing, China, 2GE healthcare, China, Beijing, China
Synopsis
Cervical cancer is
the leading cause of death in gynecological malignancy around the world. The
most important prognostic factors include stage at diagnosis,
histological subtype, tumor differentiation and et al, which are
critical to make the optimal treatment strategies. However,
there are still huge challenges in accurate assessment tumor characteristics
even by clinical examination and biopsy. DWI has showed promising results in assessment tumor
characteristics in cervical cancer. Diffusion kurtosis imaging (DKI), an
extension of DWI, is more sensitive to tissue heterogeneity and water
exchange. This study demonstrated that DKI-derived parameters are helpful
in assessment histological features of cervical cancer.
Purpose
To
prospectively investigate the additional value of diffusion kurtosis imaging
(DKI) in characterization of cervical carcinoma.Method
Sixty-three
consecutive patients with histological-confirmed cervical cancer underwent pelvic
MRI with DKI with 3 b values (0, 1000 and 2000s/mm2) prior to receiving any
tumor related treatment at 3.0T MR scanner(GE Discovery 750, USA). ADC values
derived from 0 and 1000s/mm2 and parameters derived from DKI including
MD, a corrected apparent diffusion coefficient, and MK, the deviation of the
signal decay from a monoexponential model were measured using the Function tool
DKI program independently. A two-way analysis of variance (ANOVA, SPSS version 21.0,
USA) was used to assess differences between histological subtypes (squamous
cell carcinoma: n=44; adenocarcinoma: n=19) and grades of tumors (low grade:
n=32; high grade: n=31) for each of these parameters. The Spearman correlation
analysis was used to evaluate the correlation between these parameters and
tumor FIGO stage, then the significant parameters were compared between groups
of early stage (FIGO IB-IIA: n=24) and advanced stage (FIGO IIB and above:
n=39). ROC was performed to obtain the area under the receiver operating
characteristic curve (AUC) to assess the diagnostic capacities of these
parameters in characterization of cervical carcinoma.Result
The MD and ADC values
were significantly lower in squamous cell carcinoma than adenocarcinoma (P=0.005
and 0.009, respectively), but no significant differences were found between
tumor grade (P=0.460). Only MD value was inversely related to tumor FIGO
stage (r=-0.317, P=0.011) and there was significantly differences for MD
value between tumor with early stage and advanced stage (P=0.003).
Moreover, no differences were observed in MK for different pathological
features tested. The ADC values showed better diagnostic performance in differentiation
of tumor histological subtype with AUC of 0.749 at the optimal cut-off of 1.19×10-3mm2/s. An optimal cut-off MD values of 1.27×10-3mm2/s for distinguishing early stage and
advanced stage in cervical carcinoma has a sensitivity of 79.49% , the
specificity of 66.67% and the accuracy of 74.62%.Discussion and Conclusion
The significant differences
of MD and ADC values were observed in this study between squamous cell
carcinoma and adenocarcinoma indicating that these parameters were sensitive to
microstructure properties of tumors, which was consistent to some previous
studies. In addition, only MD was significantly lower in tumor with advanced
stage than those with early stage in cervical carcinoma likely related to the
more restricted water diffusion in advanced stage tumor with more cellular
packed tumor environment, which added value over conventional DWI for staging
cervical carcinoma. Overall, MD and ADC provide further and more complementary
information in differentiating histological subtype and FIGO stage for cervical
carcinoma aiding in confidence in diagnosis and staging tumor in clinic.Acknowledgements
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References
[1]
Wang M, Perucho J, Chan Q, et al. Diffusion Kurtosis Imaging in the Assessment
of Cervical Carcinoma[J]. Acad Radiol,2019.
[2]
Zhu L, Pan Z, Ma Q, et al. Diffusion Kurtosis Imaging Study of Rectal
Adenocarcinoma Associated with Histopathologic Prognostic Factors: Preliminary
Findings[J]. Radiology,2017,284(1):66-76.
[3]
Wang P, Thapa D, Wu G, et al. A study on diffusion and kurtosis features of
cervical cancer based on non-Gaussian diffusion weighted model[J]. Magn Reson
Imaging,2018,47:60-66.
[4]
Bai Y, Lin Y, Tian J, et al. Grading of Gliomas by Using Monoexponential,
Biexponential, and Stretched Exponential Diffusion-weighted MR Imaging and
Diffusion Kurtosis MR Imaging[J]. Radiology,2016,278(2):496-504.