This article is mentioned in:

Introduction

Nasopharyngeal carcinoma (NPC) occurs commonly in the Asian population, particularly in Southern and Southeast China (1). Due to the peculiar characteristics in its epidemiology, pathology, clinical behavior and response to treatment, NPC is different from other head and neck squamous cell cancer and has a relatively high overall survival rate with the integration of chemotherapy into radiotherapy (2,3). However, the majority of NPC patients present with a late stage disease accompanying neck nodal metastases when diagnosed, and the cure rate for those advanced NPCs remains unsatisfactory (4). In addition, numerous NPC survivors are often affected by moderate to severe late complications, resulting from the impact of radiation on the organs that are adjacent to the nasopharynx and neck nodes, and chemotherapy in advanced cases further exacerbates these side-effects (5). Therefore, exploring novel therapeutic regimens and improvements in disease monitoring is required. If the therapeutic effect can be predicted prior to or at the early course of the treatment, it is possible to modify the therapeutic strategies for the remaining treatment. Novel therapeutic alternatives, such as anti-epidermal growth factor receptor monoclonal antibodies, including Cetuximab and Nimotuzumab, could be advised as a combination for patients likely to be resistant to conventional treatment.

Functional images, mainly including dynamic contrast-enhanced-computed tomography (DCE-CT), positron emission tomography (PET-CT), magnetic resonance spectrometry, diffusion-weighted images (DWI) and dynamic contrast-enhanced-magnetic resonance imaging (DCE-MRI), play an important role in assessing the treatment effect on the solid tumor. Considering the apparent diffusion coefficient value of DWI and the perfusion parameter, Ktrans, of DCE-MRI as examples, the aforementioned parameters may allow the possibility to predict an early treatment response and prognosis for chemotherapy and radiotherapy (6,7). The aforementioned image modalities are mainly dependent on contrast agents that may cause an allergic reaction. DCE-CT and PET use ionizing radiation, which has known risks. Contrast-enhanced ultrasound (CEUS) is another type of functional image. The medium of CEUS is SonoVue, which contains micrometer-sized (1–10 μm in diameter) bubbles of sulfur hexafluoride with a stabilizing shell. The introduction of exogenous microbubbles into the vasculature causes enhancement of the backscattered intensity of the blood and can be used to assess tissue blood flow. Relevant analysis software has been developed to analyze ultrasound signal intensity (SI) patterns obtained by imaging continuously prior or subsequent to treatment, and information regarding tissue blood flow and vascular integrity, branching patterns and density will be assessed (8). As reported previously, the information extracted from the CEUS data was similar to that obtained from DCE-MRI (9), and CEUS has numerous benefits, including a sufficient high safety profile that is acceptable for patients with renal failure or an iodine allergy, absence of radiation, easy reproducibility and high temporal resolution (10–13).

Thus, the present study was performed to evaluate the potential utility of CEUS-derived parameters from the metastatic cervical lymph nodes, prior or subsequent to the early course of the treatment, in predicting the nodal treatment response to radiation-based therapy in patients with NPC.

Patients and methods

Patient selection

Sixty-seven NPC with cervical lymph node metastases patients who were treated at The Second Affiliated Hospital (Zhejiang University School of Medicine, Hangzhou, China) between December 2011 and February 2013, were enrolled in the study. The study was approved by the Institutional review board, and written informed consent was obtained from each participant prior to the CEUS examination. All the subjects qualified for the following criteria: i) NPC with metastatic lymph nodes proven by pathology; ii) Eastern Cooperative Oncology Group performance status score, ≤2; iii) adequate organ function; and iv) no concomitant malignancy. The stage of disease was classified according to the 7th edition of the Union for International Cancer Control (UICC) and the American Joint Committee on Cancer (AJCC) staging system (14). All the clinical characteristics of the patient are listed in Table I.

Table I Clinical data according to the therapeutic response for nasopharyngeal carcinoma patients with lymph node metastases, n=67.

Table I

Clinical data according to the therapeutic response for nasopharyngeal carcinoma patients with lymph node metastases, n=67.

	Therapeutic response
Characteristics	CR	PR	P-value
Patients, n	48	19
Age, years	55.83±10.38	56.08±10.94	0.912
Gender, n			0.782
Male	26	11
Female	22	8
PS, n			0.492
0	39	14
1	9	5
Treatment, n			0.023
RT	6	7
RT+CT/T	42	12
Differentiation, n			0.310
Well/moderate	4	2
Poor	44	17
T stage, n			0.686
T1	5	3
T2	28	9
T3	9	6
T4	6	1
N stage, n			0.021
N1	27	4
N2	19	12
N3	2	3
Lymph node	3.24±0.92	3.83±0.93	0.005
Max. D, cm

[i] CR, complete response; PR, partial response; PS, performance status; RT, radiotherapy; CT, chemotherapy; T, targeted therapy; Max. D, maximum diameter.

Treatment protocol

The first CEUS examinations were performed prior to any treatment, and the second CEUS examinations were arranged at the 5th fraction of radiotherapy (using the CEUS methodology). All the 67 patients underwent intensity-modulated radiotherapy (IMRT) and among them, 52 patients received platinum-based concomitant chemotherapy and 10 received weekly Nimotuzumab-targeted therapy at a dose of 200 mg 1 week before and during the course of radiotherapy and the other 5 cases received only IMRT. Nimotuzumab, a monoclonal antibody against epidermal growth factor receptor, has been officially approved by the State Food and Drug Administration of China to treat advanced NPC (15). The target volume contouring was made following the guideline of the Radiation Therapy Oncology Group Contouring Atlas (http://www.rtog.org/CoreLab/ContouringAtlases/HNAtlases.aspx). The planning target volume of the metastatic lymph nodes was defined as PCTVnd. IMRT was performed using 6-MV photon beams and the IMRT plan was normalized such that 95% of the PCTVnd was covered with the prescription dose [60–70 Gy/30-32 fractions], and all the patients underwent IMRT once daily and 5 fractions a week.

CEUS methodology

The first section focused exclusively on the pre-treatment CEUS examination. All the ultrasound investigations were performed using the Sequoia 512 Acuson sonographic system (Siemens Healthcare, Erlangen, German) equipped with CadenceTM contrast pulse-sequencing visualization technology and a high-resolution broadband ultrasound transducer (8L5; 5–8 MHz). Each dose of the intravenous contrast medium of microbubbles (SonoVue; Bracco, Milan, Italy) was dissolved in 5 ml of saline and a 2.4 ml bolus was injected into the superficial elbow vein of the patient at the rate of 1 ml/sec, followed by a 5.0 ml saline flush (16). The process of CEUS was performed by the same ultrasound investigator with >5 years of experience in CEUS. The process of CEUS should include the following: i) Scan parameters (depth, focus, pulse repletion frequency, mechanical index and depth-gain compensation) were optimized for a clear, artifact-free depiction; ii) the probe was manually stabilized at the largest diameter of the target lymph node; and iii) the duration of the video was ~90 sec for analysis (17). The video was stored as a digital archive (Audio Video Interleave) in the hard disc and was transferred to a personal computer for off-line parametric analysis.

The second section was in-treatment CEUS examination, and each of the 67 patients underwent the aforementioned CEUS examination at the 5th fraction radiotherapy once again. To assure agreement of the lymph nodes examination with the first examination, the ultrasound investigator reviewed the previous CEUS video and subsequently performed the second examination. All the patients underwent the CEUS examinations twice; the former digital archive was the baseline as a control, whereas the later digital archive was under the treatment as a comparison.

Off-line parametric CEUS analysis

The aforementioned digital archives were processed with the use of contrast-enhanced computer-assisted perfusion analysis of the metastatic nodes using the QontrastTM analysis software (Qontrast 4.0; Bracco SpA, Milan, Italy), which is a post-processing computational tool and can be used to obtain objective and quantitative parameters of the microvessels in various organs, including the lymph node (6). A region of interest (ROI) encompassing the whole area of the lymph node was manually drawn, and subsequently the software automatically processed and a time-intensity curve and parametric graphs were produced. The following parameters were automatically generated (Fig. 1): i) Peak intensity [PI; including pre-treatment (PIpre) and in-treatment (PIin) as percentages] defined as the increase in signal intensity (SI) from baseline SI to the maximal SI measured in the selected ROI; and ii) time to peak (TTP; including TTPpre and TTPin, in sec) defined as the time period from the onset of the lymph node enhancement to the moment the maximal SI is reached. The aforementioned analyses were all performed by the same investigator who was well experienced in the Qontrast software, and was blinded to the clinical data. To maintain intra-investigator agreement, >3 repeats of the aforementioned analysis were carried out by the same investigator. Concordant measurements were those that differed by no more than ±1 sec. Subsequently, PIΔ and PIratio were calculated by the algorithm that represented the change of the contrast agent perfusion via treatment: PIΔ = PIpre-PIin and PIratio = PIin/PIpre.

Figure 1 Schematic diagram showing the parameters extrapolated by the Qontrast 4.0 analysis software of the time-intensity curve in the lymph node. SI, signal intensity; PI, peak intensity; TTP, time to peak.

Notably, the Qontrast software can compensate for minor changes in the imaging plane (such as from extremely shallow breathing). In the case of more pronounced changes in the imaging plane, frame-by-frame editing can be performed, and the respective frames can be manually selected and characterized as ‘wrong’ (18). By contrast, Qontrast-assisted CEUS parameters exhibited high inter-investigator reproducibility (11).

Evaluation of the therapeutic response

CE-MRI was recommended as the tool for evaluating the treatment effect of the metastatic cervical lymph nodes. Each patient underwent the CE-MRI examination twice, the former arranged prior to any treatment and the latter scheduled 1 month after the completion of all the radiation fractions. MR images were interpreted by the same experienced radiation oncologist with >10 years of clinical experience in NPC, who was informed of the target lymph nodes in advance. The Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria was referred to for assessment of the therapeutic response (19).

Statistical analysis

All the statistical analyses were carried out using SPSS 16.0 software (SPSS, Inc., Chicago, IL, USA). The descriptive statistics were produced for the continuous variables and the results are presented as mean ± standard deviation (SD). The χ2 or Fisher’s exact tests were used to determine the significance of the associations between the therapeutic effect and categorical variables, whereas the correlation between the continuous variables and the therapeutic effect was assessed by the analysis of variance. The statistical significance of the changes in PI and TTP was evaluated with the paired t-test. All the tests were two-sided and P<0.05 was considered to indicate a statistical significance difference. The Spearman’s correlation coefficient between the situation of the lymph nodes and changes in the perfusion parameters was calculated. The correlations were interpreted according to Cohen’s standard, in which absolute correlations of <0.3 were considered weak, 0.3–0.5 were moderate and 0.5–1.0 were strong. A receiver-operating characteristic (ROC) curve was constructed to assess the accuracy of the parameters for the prediction of the therapeutic responses. The ROC was constructed to illustrate the predicted probability of the parameters.

Results

Clinical data and treatment response

Of the 67 patients, there were 37 males and 30 females, with a mean age of 55.93±10.55 years (range, 24–88 years). The baseline patient and tumor characteristics are summarized in Table I. All the patients completed the entire radiation-therapy plan. The response assessment at the lymph nodes revealed a complete response (CR) in 48 patients and partial response (PR) in 19 patients. No patients showed stable or progressive disease at the lymph nodes or at the primary site. Age, gender, performance status score, T stage or pathological differentiation status were not associated with the treatment response (P>0.05). However, N stage, the size of the lymph node and the treatment modalities were significantly different between the CR and PR groups, respectively (P<0.05). Notably, radiation therapy in combination with chemotherapy/targeted therapy was superior to radiotherapy alone, in regards to therapeutic response (Table I).

Data of CEUS parameters and therapeutic response

For all the 67 cases investigated, PIpre ranged from 33.9–50.4% (40.9±3.4%), and the PI following 5 fractions of radiation (PIin) ranged from 20.4–42.5% (30.7±5.3%). There was a significant difference in PIpre between the patients who showed a CR or PR of the metastatic lymph nodes, and the mean values of PIpre were higher in patients with CR nodes compared to the patients with PR nodes (41.90±3.62 vs. 39.39±2.48%; P=0.002). Following 5 fractions of radiation, a decrement in PI was observed in all the patients without any exception (Fig. 2A). The mean PIin value of the lymph nodes that achieved CR was 34.24±3.78%, which was significantly higher compared to the PIin value for the PR, 25.62±2.30% (P<0.001). To further standardize the data, a PI-quotient known as PIratio, was calculated by dividing the PIin by the corresponding PIpre for the same lymph node. A higher PIratio was also observed in the CR lymph nodes (0.81±0.01 vs. 0.66±0.01; P=0.001). As shown in Table II, the mean change in PI (PIΔ; PIΔ = PIpre-PIin) was smaller in the patients with CR nodes compared to the patients with PR nodes (7.79±3.28 vs. 13.77±1.90%; P<0.001).

Figure 2 Box and whisker plots showing the statistically significant changes observed in (A) peak intensity (PI) and (B) time to peak (TTP) (the upper and lower end of each box corresponds to the maximum and minimum value, respectively, and the horizontal line inside the box corresponds to the mean value).

Table II CEUS parameters with different therapeutic responses for lymph node metastases of nasopharyngeal carcinoma patients, n=67.

Table II

CEUS parameters with different therapeutic responses for lymph node metastases of nasopharyngeal carcinoma patients, n=67.

Parameters	CR (n=48)	PR (n=19)	P-value
PIpre, %	41.90±3.62	39.39±2.48	0.002
PIin, %	34.24±3.78	25.62±2.30	0.001
TTPpre, sec	10.92±0.26	11.07±0.61	0.334
TTPin, sec	12.42±1.49	12.13±1.40	0.356
PIΔ, %	7.79±3.28	13.77±1.90	0.001
PIratio	0.81±0.01	0.66±0.01	0.001

[i] CEUS, contrast-enhanced ultrasound; CR, complete response; PR, partial response; PI, peak intensity; pre, pre-treatment; in, in-treatment.

In the present study, TTPpre ranged from 9.36 to 13.21 sec (10.98±0.69 sec), and TTPin ranged from 10.09 to 14.43 sec (12.30±1.46 sec), showing that the value of TTP had a tendency to increase following treatment and the change in TTP was statistically significant (P<0.05) (Fig. 2B). However, the mean TTPpre was 10.92±0.26 sec in the CR nodes and 11.07±0.61 sec in the PR nodes. No significant difference was observed between the two groups (P=0.334). Similarly, there was no significant difference in TTPin between patients who showed a CR or PR of the metastatic lymph nodes (12.42±1.49 vs. 12.13±1.40; P=0.356), as shown in Table II.

Correlation between the CEUS parameters and therapeutic response

The Spearman’s correlation coefficient between the lymph nodes therapeutic response and changes in CEUS perfusion parameters was calculated. There was a strong-positive correlation between the PIin, PIratio and therapeutic response (ρ=0.81, ρ=0.734), a moderate-positive correlation between the PIpre and therapeutic response (ρ=0.368) and a strong-negative correlation between the therapeutic response and PIΔ (ρ=−0.777) (Table III). Logistic regression analysis of the CEUS parameters indicated that PIin and PIratio were the significant predictors of the therapeutic response (P<0.01).

Table III Spearman’s rank correlation coefficients for the CEUS parameters and therapeutic response.

Table III

Spearman’s rank correlation coefficients for the CEUS parameters and therapeutic response.

	Therapeutic response	PIpre	PIin	TTPpre	TTPin	PIΔ	PIratio
Therapeutic response	1
PIpre	0.368a	1
PIin	0.810a	0.681a	1
TTPpre	−0.141	−0.159	−0.156	1
TTPin	0.121	0.009	0.129	0.164	1
PI	−0.777a	0.120	−0.785a	0.066	−0.188	1
PIratio	0.734b	−0.263a	0.868b	0.070	−0.204	−0.956b	1

a P<0.05;

b P<0.01.

{ label (or @symbol) needed for fn[@id='tfn5-mco-02-05-0666'] } CEUS, contrast-enhanced ultrasound; PI, peak intensity; pre, pre-treatment; in, in-treatment.

An ROC curve was constructed to assess the accuracy of the parameters for the prediction of the therapeutic responses. The ROC curve showed that the therapeutic response could be well predicted by the parameters of PIin and PIratio compared to PIpre and PIΔ. The PIin area under the ROC curve (AUC) was 0.936 (95% confidence interval, 0.877–0.988), and the AUC of the PIratio was 0.931 (95% confidence interval, 0.877–0.985) (Fig. 3). When the cut-off value of PIin was set at 29.4%, sensitivity and specificity in predicting the lymph node therapeutic response was 94.3 and 88.2%. The best cut-off value of the PIratio was ≥0.69 by coordinating the points of the ROC curve, and the predicted sensitivity and specificity of the PIratio was 92.5 and 83.8%, respectively (Table IV).

Figure 3 Receiver operating characteristic (ROC) curve constructed to depict predictive probabilities of the logistic regression model for the imaging parameters obtained [peak intensity (PI)in and PIratio are significant predictors compared to PIpre and PIΔ]. AUC, area under curve.

Table IV Sensitivity and specificity for the CEUS parameters, PIpre, PIin and PIratio.

Table IV

Sensitivity and specificity for the CEUS parameters, PIpre, PIin and PIratio.

Parameters	Cut-off value	Sensitivity, %	Specificity, %
PIpre	≥39.65%	72.0	52.0
PIin	≥29.40%	94.3	88.2
PIratio	≥0.69	92.5	83.8

[i] CEUS, contrast-enhanced ultrasound; PI, peak intensity; pre, pre-treatment; in, in-treatment.

Discussion

To the best of our knowledge, the present study is the first to assess the early predictive value of parametric CEUS for the therapeutic response in the metastatic cervical lymph nodes of NPC patients treated with radiation-based therapy. The aim was to evaluate whether any of these CEUS parameters are suitable as a predictor for the nodal treatment response to radiation-based therapy in NPC patients with cervical lymph nodes metastases.

In the present study, a higher mean value of PIpre was observed in the lymph nodes that achieved CR compared to PR, and the difference was statistically significant. Functional images, including DCE-MRI, have been employed for the prediction of the early treatment response and prognosis for head and neck cancers (18,20,21). DCE-MRI provides a perfusion parameter, Ktrans, which reflects a combination of the tumor blood flow and microvascular permeability. In the study reported by Chawla et al (18), patients with head and neck squamous cell carcinoma that was responsive to chemoradiation therapy had significantly higher pre-treatment Ktrans values from nodal masses than patients with a PR. In another cohort of 33 patients with head and neck squamous cell carcinoma who were treated with chemoradiotherapy, the average pre-treatment Ktrans value of the CR group was found to be significantly higher (P=0.001) than that of the PR group (21). The result of the present study is partially consistent with the aforementioned DCE-MRI studies. However, the majority of the PIpre values fell in a wide region and there was a large range of PIpre values from the CR and PR groups that overlapped with each other. The predicted sensitivity and specificity of PIpre was relatively low, and therefore the PIpre value alone is of less interest for predicting the therapeutic response of the lymph node metastases from NPC.

We hypothesized that alternations in the metastatic nodal perfusion parameters during the early course of treatment for NPC may improve the predictive value for the therapeutic response than pre-treatment measurements alone. Thus, the second CEUS examinations were arranged at the 5th fraction of radiotherapy for all the 67 patients. A decrement in PI was observed in all the investigated nodes regarding the PIpre. There was a significant difference in the PIin between the patients who showed CR or PR, and the values of PIin were much higher in the patients with CR nodes (34.24±3.78%) compared to those with PR nodes (25.62±2.30%). For each individual lymph node, the PIratio was calculated by dividing the PIin by the PIpre, and a higher PIratio was also observed in the lymph nodes that achieved a CR (0.81±0.01 vs. 0.66±0.01; P=0.001). The PIΔ was found to be smaller in the CR lymph node. These findings suggest that an improved blood supply and potentially improved oxygenation during the early course of treatment may be a positive indicator for the therapeutic response at the metastatic lymph node of NPC. Based on the ROC curve, a cut-off for the PIin and PIratio were established, which predicted the response with a specificity of 88.2 and 83.8%, and a sensitivity of 94.3 and 92.5%, respectively. To the best of our knowledge, the present study describes for the first time the CEUS parameters during the early course of chemo-radiotherapy that reliably predict the metastatic lymph node response. These parameters may help to optimize the patient selection, thereby individualizing treatment and preventing non-responders from undesirable side-effects.

TTP represents the arrival time of the contrast agent to reach its maximum. In the present study, TTP had a tendency to increase 1 week after the initiation of the radiation-based treatment. However, TTPpre and TTPin were found to have no significant difference regarding the CR or PR lymph node response status. In a study by Knieling et al (22), hepatocellular carcinoma patients were treated with sorafenib and the TTP increased as early as 1 month after the initiation of the treatment in the responder group compared to the non-responder group. In the study by Schirin-Sokhan et al (23), non-primary resectable liver metastases from colorectal cancer were treated with bevacizumab-based chemotherapy and it was demonstrated that the baseline TTP was significantly lower in the responder group compared to the non-responders, suggesting that low baseline TTP significantly correlates with tumor response according to RECIST. Furthermore, correlating to the antiangiogenic effect of bevacizumab, a strong increase in TTP was observed during chemotherapy, which was restricted to the responder group. The data concerning TTP and radiation-therapeutic response are sparse and should be investigated further in a larger patient population with more CEUS examinations during the course of the radiation-based therapy.

The patients treated with radiotherapy combined with chemotherapy and/or targeted therapy had a higher CR rate of lymph node metastases compared to those who received radiotherapy alone in the present study. A conventional radiotherapy course is usually 6–7 weeks, and may be followed by adjuvant chemotherapy and targeted therapy. If the early changes in nodal perfusion could help to predict the therapeutic response, it will be possible to alter the intensity of the treatment regimen, thus individualizing the remaining treatment.

The present study is, to the best of our knowledge, the first to investigate CEUS as a predictor for the therapeutic response in NPC cervical lymph nodes metastases. The data suggests that the CEUS parameters during the early course of chemo-radiotherapy, PIin and PIratio, are associated with the therapeutic response of the lymph node metastases from NPC, thus yielding conceivable predictors with the potential to modify and individualize treatment.

Acknowledgements

The authors would like to thank all the colleagues in the Department of Radiation Oncology and Ultrasound for their good cooperation. The present study was supported by the National Natural Science Foundation of China (grant no. 81071823 and 81201811) and the Zhejiang University Research Foundation.

References