Cervical cancer is the most common gynecological malignancies in the world, responsible for an estimation of 265,672 deaths in 2012, 88% of which occurred in the developing countries [
]. Currently, concurrent chemo-radiotherapy (CCRT) is considered the standard treatment option for patients with locally advanced cervical cancer. Studies have shown that CCRT improves overall and progression-free survival, as well as decreases local and distant recurrence, yet at the expense of varying degree of toxicity and morbidity [
]. In addition, due to tumor heterogeneity, it is unlikely that all cancers will respond equally to a specific treatment regimen [
]. Thus, successful treatment that leads to a better outcome in cervical cancer necessitates accurate clinical evaluation. In recent year, increasing efforts were devoted to the early detection and prognostic assessments of the treatment-associated change in tumor burden [
Several functional imaging techniques, such as diffusion-weighted (DW) magnetic resonance imaging (MRI), dynamic contrast-enhanced (DCE) MRI, 18 F–fluorodeoxyglucose-positron emission tomography (FDG-PET), have been introduced to evaluate the effect of chemo-radiotherapyin cervical cancer [
]. These functional imaging techniques have demonstrated superior capability to conventional imaging in terms of identifying biological or molecular changes which occur prior to visible or measureable change in tumor size [
]. However, most of these technologies are not used in routine surveillance due to increased radiation burden, potential contrast agent’s adverse reaction, long scan time and high cost, especially for patients who require long-term follow-ups. These limitations underline the need for a reliable, convenient and low-cost functional imaging biomarker for tumor response to treatment.
Sonoelastography is a well-established imaging modality that could provide the information of tissue elasticity (stiffness) that is complementary to the morphology and vascularity information provided by conventional sonographic examinations [
]. An important property of tissue is its intrinsic elasticity, which changes under the influence of pathologic processes, such as inflammation and neoplasm. Various groups have investigated sonoelastography in cancer detection and differential diagnosis in several organs, such as breast, liver, thyroid, and prostate [
]. In these cases, the malignant tissues are shown less compressible, resulting in less strain than the normal tissues under uniform stress. Recently, sonoelastography imaging was used to study the normal and abnormal cervix [
]. Several studies have shown the potential of sonoelastography in monitoring and predicting the therapeutic response such as the responding and non-responding malignant tissues in patients following CCRT [
The purpose of this study is to investigate the utility of strain elastography in monitoring and early predicting therapy response to CCRT in patients with locally advanced cervical cancer. To the best of our knowledge, only one study with four cases had investigated the role of strain elastography as imaging biomarkers for prediction of therapeutic response in cervical cancers [
]. Therefore, our study aimed to investigate whether the strain ratio can be used as imaging biomarker in evaluating early response in 1 or 2 weeks after the initiation of treatment in patients with locally advanced cervical cancer.
Patients and tumor characteristics
This study was approved by our institutional review board, and written informed consents were obtained from all patients. Between January 2015 and June 2016, we prospectively enrolled 47 consecutive patients with histologically confirmed cervical cancer who were scheduled to receive CCRT at our hospital. Each patient was staged according to the criteria of the International Federation of Gynecology and Obstetrics (FIGO). The inclusion criteria were as follows: (1) age over 21, (2) FIGO Stage IB to IV, and (3) no history of chemotherapy or radiotherapy.
All patients were treated with a combination of radiotherapy and chemotherapy. Radiotherapy consisted of external beam radiotherapy (EBRT) and intracavitary high-dose-rate brachytherapy (ICR). EBRT was delivered to the whole pelvis with a total dose of 50 Gy (a daily dose of 2 Gy and 5 times per week). EBRT was accompanied by concurrent chemotherapy, as follows: three-four cycles of every 2 weeks Nedaplatin (40-60 mg/m2) plus Paclitaxel (80 mg/m2) in 32 patients. ICR was initiated after an EBRT dose of 50 Gy. ICR was delivered twice a week with a dose of 5 Gy at point A (6 times, total dose 30Gy). The definition of point A follows the American Brachytherapy Society recommendation [
]. The entire CCRT for each patient was completed within 8 weeks.
Treatment response evaluation
The treatment response of the CCRT was determined by the shrinkage of the longest diameter of the cervical cancer with MRI prior to and right after therapy completion. All MRI was performed with a 3.0-T MRI scanner (Achieva 3.0 T, Philips Healthcare, Best, the Netherlands) with a 16-channel torso phased-array body coil at the time of diagnosis and immediately after therapy completion. Two radiologists independently evaluated longest tumor diameter based on T2-weighted images according to the corresponding diffusion-weighted as well as contrast enhanced images with the maximal magnification and compared them in consensus. The change of tumor size was calculated according to the following equation: change in tumor size % = (pre-longest diameter-post-longest diameter)/pre-longest diameter × 100%. The clinical responses were classified using the Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria [
] in the following 4 categories: complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). CR is defined as no residual cancer, PR as at least a 30% decrease in the sum of diameters of the cancer, PD as at least 20% increase and SD as no sufficient shrinkage to qualify for PR or sufficient increase to qualify for PD.
Strain elastography imaging and analysis
All patients underwent ultrasound examinations at the following 4 time points: prior to CCRT, at week 1 and week 2 during CCRT, as well as within 1 week post CCRT. Ultrasound data were acquired using GE Voluson E8 ultrasound machine (GE Medical Systems, USA) with a 3D/4D endocavitary convex array transducer (GE RIC5–9, bandwidth 5–9 MHz). All the examinations were performed by a single ultrasonographer with seven-year experience, who was blinded to the treatment outcome and MRI results.
The ultrasound examination was performed with the patient in the lithotomy position (with an empty bladder). A disposable condom with coupling gel was used to cover the endocavitary probe which was gently inserted in the anterior vaginal fornix. All ultrasound data were acquired with the same settings: 5.0 cm depth, 1 focal zone, 7 Gy map, −15 gain, 121°angle, 3 persist, 2 enhance, 20 reject and 7 dynamic control. This setting was kept consistent throughout the study to ensure quantitative ultrasound comparison. Each patient first received a B-mode transvaginal ultrasound examination and the cervical tumoral tissue was identified as a mass with heterogeneous echogenicity and irregular borders with disruption of the cervical canal. The optimal gray-scale image of the sagittal view along the longest diameter of the cancer was obtained. B-mode scanning was followed by dual-mode scanning for real-time elastography with color-encoded superimposition of the information. Elastographic images were generated by soft and rhythmic compression of the cervix using the ultrasound transducer. On the monitor, the two panel image was displayed with the conventional B-mode image on the left and the elastography image on the right. The elasticity information is presented in color, with blue indicating stiffer tissue, red indicating softer tissue and green as intermediate stiffness.
All elastography images were analyzed by 2 experienced ultrasonographers (X.X., X.X.X.)with 12 and 15 years’ experiences in gynecology. Strain ratio was employed to evaluate the strain difference between the cervical cancer and the normal parametrial tissue quantitatively. A press indicator on the left upper side of screen was used to evaluate the condition of compression in the region of interest (ROI) from the minimum to the maximum (level 1–6). Compression and relaxation waveforms were shown on the lower right side screen. While the compression indicator was green color at the value of 5 or 6 and the pressure waveform was simultaneously at the peak, the images of measurements and examination techniques were stored digitally. To compute the strain ratio, we first circled the normal parametrial tissue at the same depth of the cervical cancer as A,then manually contoured entire cervical cancer as B, and strain ratio was computed as A/B. We performed this three times for each patient and computed the mean value of strain ratio.
The clinical characteristics of the patients and tumors were expressed in mean and standard deviation (SD). Mann-Whitney test was used to compare the difference in FIGO stage, histological grade and lymphatic metastasis between the CR and PR groups. The comparison of mean age and the maximum tumor diameter between two groups were performed using Student’s unpaired t-test. Repeated measures analysis of variance (ANOVA) and Student’s unpaired t-test were constructed to the multiple comparisons in strain ratios for the CR group and PR group at each time point. A two-sided
p < 0.05 was considered to be statistically significant. The inter-observer and intra-observer variability of measurements were assessed using intra-class correlation coefficients (ICCs) with a 95% confidence internal (CI). Statistical analysis was performed using SPSS software version 19.0 (SPSS Inc., Chicago, IL,USA).
This study provided initial evidence that strain elastography may be used to monitor and predict therapeutic response of CCRT in patients with cervical cancer. It is well-known that tumor formation and its degeneration in response to CCRT may exhibit corresponding changes such as inflammation and fibrosis with stromal cells [
], which substantially alter the biomechanical properties of tumor tissues. After therapy completion, all patients were classed by MRI as complete or partial responders. The cervical cancer appeared a decrease in the strain ratio, thus indicating an increase in the tumor strain and a decrease in stiffness. The strain ratios revealed significant changes between time points in the CR group (
< 0.001) and PR group (
< 0.001), and the difference between CR and PR groups showed significant (
= 0.011). Strain ratios between CR group and PR group were significant from week 1 during CCRT to therapy completion (all
< 0.05). Moreover, strain ratio exhibited a significant decrease at week 1 during CCRT for the CR group (
< 0.001) and at week 2 for the PR group (
In our study, the strain ratio of the pre-treatment cervical cancer (mean 3.99) is in line with the published results (mean 3.8) and significantly different from normal cervical tissues (mean 1.2) [
]. The high strain ratio of cervical cancer is due to the increased cellular proliferation, which is an important factor that influences the strain in tumor tissue. In a study of 55 breast-cancer patients, Hayashi et al. [
] investigated the tumor stiffness post neoadjuvant chemotherapy and found that relatively soft tumors were highly responsive to neoadjuvant chemotherapy and more frequently displayed pathologic complete response as compared with hard tumors. However, in our study, the baseline strain ratios were not significantly different between the CR and PR groups, which were consistent with the study reported by Rafaelsen et al. [
] and the reason may be relevant to the identical histological type. Additionally, there were no significant differences in the age, tumor size, FIGO stage, histological grades, and lymphatic metastasis between these two groups. Therefore, the therapy response doesn’t seem to correlate with the baseline strain ratio, tumor size, patient age, FIGO stage, histological grades, or lymphatic metastasis.
Due to tumor heterogeneity of radioresponsiveness, the timing of optimal evaluation for therapeutic response to CCRT remains controversial. Thus, how to identify patients at risk of treatment failure in the early stage is very important for clinicians. Recently, several studies have focused on the early detection of response to chemoradiation in cervical cancer using DW and DCE-MRI. Harry et al. [
] reported that tumor apparent diffusion coefficients (ADCs) after 2 weeks of CCRT were significantly correlated with final tumor response, while Liu et al. [
] conflicted with this results, showing significant correlation with final therapeutic responses at 4 weeks after initiation of CCRT. Moreover, a recent study of Park et al. [
] demonstrated that the mean ADC and tumor volume transfer constant K(trans) and extravascular extracellular volume fraction (ve) of cervical cancer increased even 1 week after initiating CCRT. To date, the role of sonoelastography as a predictor of therapeutic response has been demonstrated in some tumors. Rafaelsen et al. showed elastography after two weeks of chemo-radiation seems to hold early predictive treatment response information for rectal cancers [
]. However, a study conducted by Falou et al. [
] using ultrasound elastography suggested that there was no significance at 1 week after the start of neoadjuvant chemotherapy with breast cancer patients, while non-responders and responders were found to be highly significantly different 4 weeks after treatment initiation for averaged strain ratios. In our series, strain ratios for CR group and PR group were significant at 1 week after treatment initiation (
< 0.05). Moreover,strain ratios showed a significant decrease after 1 week of CCRT in CR group. These results might be explained that with the effective treatment, the tumor begins to become less stiff as a result of a decrease in cellular proliferation and moderate induction of apoptosis which changes its structure and biomechanical properties, leading to an increase of tumor strain and a decrease of strain ratio. Taken in aggregate, these findings suggest that the potential of strain elastography as a surrogate biomarker to evaluate an early therapeutic response.
With successful therapy, treatment responses are tumor cell necrosis, apoptosis and lysis, resulting in a decrease in tumor stiffness. After therapy completion, our results showed that strain ratios of CR and PR groups both significantly decreased (both
< 0.001), which were strongly correlated with clinical assessment. These findings pave the way for clinical application of strain elastography imaging in monitoring CCRT of cervical cancer. It is clear that if any quantitative features are used to assess tumor response, reproducibility is a prerequisite. Our study further demonstrated the inter-observer and intra-observer reliability of strain ratio evaluations, which were similar to the findings reported by Sun et al. [
]. Future studies will be designed to further investigate the correlation between change in stiffness and treatment response in non-responders.
Tumor size is one of the most important prognostic factors for cervical cancers [
]. With the superior soft tissue contrast resolution, MRI displays high accuracy (70%) for determination of tumor size [
] and has been considered as the reliable method strongly recommended by RECIST 1.1 for follow-up in tumor size delineation after non-surgery treatment [
]. Vincens et al. demonstrated that the sensitivity of MRI in evaluating residual tumor were 80% in stage IB2/II cervical carcinoma after CCRT [
]. A prospective study on cervical cancer showed that tumor regression rate obtained during mid-radiation therapy had the best outcome prediction rate for local control (84%) and disease-free survival (63%) [
]. In this study, as ultrasound was not recommended to measure tumor lesions during therapy [
], tumor sizes were measured by MRI.
Several limitations exist in this study. First, this study was performed with small sample size and a short follow-up period. Second, our study observed morphologic changes before and after treatment based on imaging by MRI as a reference of therapeutic response and lacked pathologic results and clinical outcome, such as progression-free survival and overall survival. Third, we did not investigate the effect of CCRT on the stiffness of normal parametrial tissue in the current study. Fourth, only cervical squamous carcinoma was included in this study, so we are uncertain whether other types of cervical cancers will show similar trends in the treatment response monitoring by strain elastography.
This work was supported by National Natural Science Foundation of China (81371516, 81501441, 81671751), Foundation of National Health and Family Planning Commission of China (W201306), Social Development Foundation of Jiangsu Province (BE2015605), Natural Science Foundation of Jiangsu Province (the Youth Foundation, BK20150109 and BK20150102). The funding sources had no role in the study design, data collection, data analysis, or interpretation of the findings.
Availability of data and materials
The analyzed data sets generated during the study are available from the corresponding author on reasonable request.
YX and LJZ made substantial contributions to data analysis and drafting the manuscript; BRL provided all oncological support; HHW and SL had significant roles in the acquisition data and interpretation of data; TR and JH carried out the quality control of ultrasound & MR examinations and data analysis; XFY and ZYZ made substantial contributions to conception and design. TL had significant roles in revising the manuscript. All authors have read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Written informed consent for publication of their clinical details and any accompanying images was obtained from all the patients.
Ethics approval and consent to participate
This study was approved by the institutional review board of Nanjing Drum Tower hospital, and all the patients offered the written informed consents.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.