To explore the predictive value of diffusion-weighted imaging (DWI) in evaluating the short-term efficacy of concurrent chemoradiotherapy (CCRT) in the treatment of patients with non-small cell lung cancer (NSCLC).
A total of 192 patients with NSCLC were selected and treated with CCRT. Dynamic contrast-enhanced magnetic resonance imaging combined with DWI was performed on all patients before and after CCRT treatment. Correspondingly, apparent diffusion coefficient (ADC) values were recorded before treatment (ADCpre), during treatment (ADCmid), and after treatment (ADCpost). Tumor response was evaluated as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic power of quantitative DWI parameters in predicting the short-term efficacy of CCRT for patients with NSCLC.
There were 21 patients with CR, 82 with PR, 77 with SD, and 12 with PD. The ADCpre was negatively correlated with tumor regression rate, whereas ADCmid, ADCpost, and their respective change rates ∆ADCmid and ∆ADCpostwere positively related to tumor regression rate. The ROC curve analysis suggested ADCpre= 1.38 × 10−3 mm2/s, ∆ADCmid= 14.14%, and ∆ADCpost= 20.39% as thresholds to predict the short-term efficacy of CCRT, with corresponding areas under the curve of 0.637, 0.743, and 0.752, respectively.
These findings indicate that DWI provides promising predictive value in evaluating the short-term efficacy of CCRT in the treatment of patients with NSCLC.
Post author correction
Article Type: ORIGINAL RESEARCH ARTICLE
AuthorsHai-Dong Xu, Yu-Qin Zhang, Wei-Yu Shen, Zheng-Chun Mao
- • Accepted on 25/01/2017
- • Available online on 27/03/2017
This article is available as full text PDF.
Non-small cell lung cancer (NSCLC) is the most common histologic type of lung cancer, which accounts for approximately 80% to 85% of cases (1). Since the 1960s, lung cancer has become the most important cause of cancer mortality among men, and since the 1990s, lung cancer has equaled breast cancer as a cause of mortality in women (2). The increase in NSCLC is mostly attributed to increased numbers of smokers. The smoking rate is increasing in developing countries, although it is decreasing in developed countries. In addition to smoking, sex, nationality, and histology also account for NSCLC (3-4-5). Despite advances in standard treatment and early detection, NSCLC is often diagnosed at an advanced stage. Patients typically have poor prognosis, with a 5-year survival rate of less than 15% (6). Concurrent chemoradiotherapy (CCRT) is a new model developed in recent decades. A clinical study reported that the 5-year survival rate of patients with stage III NSCLC receiving CCRT was 15%-25%. This response was clearly better than other treatments, and CCRT was established as the standard treatment for stage III NSCLC (7).
Diffusion-weighted imaging (DWI) is a type of functional imaging modality involving magnetic resonance imaging (MRI) (8). Diffusion-weighted imaging characterizes the restriction of the random thermal motion of water molecules within biological tissues and has been applied to the detection and characterization of focal pulmonary lesions (9-10-11-12). Diffusion-weighted imaging provides quantitative and qualitative functional information about water microscopic movement at the cellular level (13). The apparent diffusion coefficient (ADC) is proposed to qualify differences in water mobility, which reflects the signal loss on DWI (14). The ADC value is inversely associated with tumor cellularity and can potentially be used to differentiate malignant and benign lesions and to predict the tumor subtypes and their response to therapy (15-16-17). Tumors are quantitatively measured by DWI and assessed by the ADC value, which is correlated to the proportion of extracellular and intracellular components (18). Diffusion-weighted imaging was initially applied to the central nervous system. However, with the development of equipment and rapid imaging technology, DWI has been increasingly used in the diagnosis and treatment of multiple organ tumors (19). A previous study demonstrated that DWI has great clinical application value and potential in the differential diagnosis of lung cancer (20). Shen et al (21) suggested that DWI is beneficial in the nodal assessment of patients with lung cancer. Tumor cellular densities increase due to the increase in DWI signals, and the signal intensity and distribution of DWI can represent the number of cancer cells (22). Therefore, DWI has potential value for the diagnosis and assessment of treatment efficacy in lung cancer. In the present study, we aimed to explore the predictive value of DWI in evaluating the short-term efficacy of CCRT in the treatment of patients with NSCLC.
This study was approved by the Ethics Committee of Ningbo Medical Center Lihuili Eastern Hospital. All patients were informed and signed the informed consent.
From February 2012 to March 2015, 192 patients with NSCLC who received CCRT were selected, including 135 male and 57 female participants. The mean age of patients was 59.5 ± 7.7 years (range 41-80 years). There were 17 cases of large cell carcinoma, 72 cases of adenocarcinoma, 93 cases of squamous cell carcinoma, and 10 cases of adenosquamous carcinoma. Inclusion criteria included the following: (1) patients with locally advanced stage III cancer were pathologically confirmed by computed tomography-guided fine-needle aspiration biopsy; (2) patients were willing to accept CCRT treatment and clinical long-term follow-up; and (3) patients had at least one tumor lesion with a well-defined border in the lung. Exclusion criteria were as follows: (1) patients had a cardiac pacemaker, metal artery clamp, or steel plate and bracket in vivo or had MRI scanning contraindications, such as claustrophobia; (2) patients had poor physical condition and failed to complete imaging examinations (single breath-hold time less than 15 seconds); (3) patients received surgery, radiotherapy, or chemotherapy before CCRT; (4) patients were allergic to enhanced contrast agent or of allergic constitution; and (5) patients were unable to tolerate the CCRT treatment and quit during the course of treatment.
A Discovery MR750 3.0T MRI scanner (GE Medical Systems, Milwaukee, WI, USA) with an 8-channel phased array combined coil was used. Plain scanning and dynamic contrast-enhanced MRI combined with DWI were performed on all patients before treatment, at the dose of 20 Gy, and 1 month after CCRT treatment. Sequences and parameters of the imaging examinations obtained 3 times were consistent.
Plain scanning was conducted using the fast recovery sequence of first spin echo (FSE). The cross-section of the T1-weighted scan and the FSE T2-weighted scan (respiratory gating) was conducted as follows: cross-section T1 repetition time (TR) sequence of 600-900 ms, echo time (TE) of 5.9 ms, stratum depth of 7.0 mm, interlayer spacing of 0.5 mm, number of excitations (NEX) of 2, field of view (FOV) of 39 cm, and matrix array of 256 × 192; FSE T2 fat inhibition TR sequence of 6,000-8,000 ms, TE of 86 ms, stratum depth of 6.0 mm, interlayer spacing of 1 mm, NEX of 2, FOV of 39 cm, and matrix array of 256 × 192.
The DWI scan was conducted through the background suppression technique for inversion recovery of planar echo DWI and the DWI combined scan through the array spatial sensitivity encoding technique with a stratum depth of 6.0 mm, interlayer spacing of 1 mm, NEX of 4, FOV of 38 cm, matrix array of 128 × 128, b of 0, and scanning speed of 900 s/mm2.
Enhanced scan sequences included whole lung axial scanning, 3D axial scanning of liver acquisition with volume acceleration, 3D fast phase gradient echo scanning, and coronal image scanning.
All DWI data were input into a GEADW4.4 workstation (GE Healthcare, Shanghai, China). The ADC values were measured and analyzed via the common use of FuncTool software by 2 senior radiologists to determine unknown clinical data and curative effects. The maximum tumor level was selected as the region of interest (ROI), and ROI outline scope was as large as possible with avoidance of necrotic areas, obstructive pulmonary expansion, pneumonia, and visible vascular areas. The average ADC value and corresponding change rate before treatment (ADCpre), during treatment (ADCmid, ∆ADCmid), and after treatment (ADCpost, ∆ADCpost) were recorded. The formulas for these calculations were ∆ADCmid= (ADCmid− ADCpre)/ADCpre; ∆ADCpost= (ADCpost− ADCpre)/ADCpre.
All patients received CCRT and were treated by intensity modulated radiation therapy with 6 MV X-ray application of linear accelerator at 2 Gy daily doses 5 times per week with a 60 Gy total radiotherapy dose. In addition, patients received cisplatin combined with etoposide (produced by Qilu Pharmaceutical Co., Ltd., Shandong Sheng, China; batch number: 04080213) chemotherapy in which the cisplatin dose was 50 mg/m2 (patients were administered on the first to fifth day in each cycle). Every 28 days was considered a chemotherapy cycle. A total of 2 chemotherapy cycles were administered to patients. During the course of CCRT, any adverse reactions to radiation were immediately addressed.
Short-term efficacy evaluation
Tumor response was evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST 1.1) as follows (23): (1) complete response (CR) was defined as the loss of all lesions; (2) partial response (PR) was defined by ≥30% reduction in the sum of maximum diameters of the target lesion; (3) stable disease (SD) was defined as a maximum diameter change of the target lesion between PR and progressive disease (PD); (4) PD was defined by ≥20% increase in the sum of maximum diameters of the target lesion as well as the appearance of new lesions. Patients with CR and PR were considered the effective group, whereas patients with SD and PD were treated as the ineffective group.
The tumor regression rate was calculated based on tumor size from a CT scan conducted before and after CCRT treatment. The formula for this calculation was tumor regression rate = (long diameter before chemoradiotherapy - long diameter 1 month after chemoradiotherapy)/long diameter before chemoradiotherapy ×100%.
SPSS (Chicago, IL) 22.0 software was used for statistical analysis. The ADC values of each group were in accordance with the normal distribution based on the Kolmogorov-Smirnov test. The
Comparisons of the baseline characteristics between the effective and ineffective groups
Among 192 patients with NSCLC, there were 21 cases (10.94%) with CR, 82 cases (42.71%) with PR, 77 cases (40.10%) with SD, and 12 cases (6.25%) with PD. There were 103 patients in the effective group (CR + PR) and 89 patients in the ineffective group (SD + PD). Before CCRT treatment, there was no significant difference in age, sex, pathologic type, or tumor size between the effective and ineffective groups (all p>0.05). However, significant differences were identified in the mean ADCpre values and tumor node metastasis (TNM) stage between the effective and ineffective groups (both p<0.05) (
Comparisons of baseline characteristics between the effective and the ineffective groups
|Characteristics||Effective group (n = 103)||Ineffective group (n = 89)||p value|
|ADC = apparent diffusion coefficient; ADCpre = average apparent diffusion coefficient value before treatment; TNM = tumor node metastasis.|
|Age, y, mean ± SD||59.00 ± 7.30||60.20 ± 8.20||0.285|
|Pathologic type, n (%)||0.780|
|Large cell carcinoma||9 (8.74)||8 (8.99)|
|Adenocarcinoma||38 (36.89)||34 (38.20)|
|Squamous cell carcinoma||50 (48.54)||43 (48.31)|
|Adenosquamous carcinoma||6 (5.83)||4 (4.49)|
|TNM stages, n (%)||<0.001|
|IIIA||62 (60.19)||28 (31.46)|
|IIIB||41 (39.81)||61 (68.54)|
|Maximum tumor diameter before treatment, cm, mean ± SD||5.60 ± 1.10||5.90 ± 1.30||0.083|
|ADCpre, 10−3 mm2/s, mean ± SD||1.36 ± 0.32||1.54 ± 0.35||<0.001|
|Tumor regression rate, %, mean ± SD||65.40 ± 8.50||27.60 ± 4.80||<0.001|
Analysis of the contrast-enhanced scanning, DWI, and ADC images of patients with NSCLC in the effective and ineffective groups before, during, and 1 month after CCRT treatment
An evaluation of a patient with NSCLC was performed based on contrast-enhanced scanning images, DWI, and ADC images before treatment, at the dose of 20 Gy, and 1 month after CCRT treatment. As shown in
Analysis of the contrast-enhanced scanning, diffusion-weighted imaging, and apparent diffusion coefficient images of a patient with non-small cell lung cancer in the effective group before (
Analysis of the contrast-enhanced scanning, diffusion-weighted imaging, and apparent diffusion coefficient images of a patient with non-small cell lung cancer in the ineffective group before (
Correlations between quantitative DWI parameters and tumor regression rates of patients with NSCLC
Significant differences in ADCpre, ADCmid, ADCpost, ∆ADCmid, and ∆ADCpost values were noted between the effective and ineffective groups (all p<0.05). Compared with the ineffective group, the effective group had a lower ADCpre value but higher ADCmid, ADCpost, ∆ADCmid, and ∆ADCpostvalues (all p<0.05). The results of Spearman rank analysis indicated that ADCpre value was negatively correlated with the tumor regression rate (p<0.05), whereas ∆ADCmid, ∆ADCpost, ADCmid, and ADCpost values were positively related with tumor regression rate (p<0.05) (
Correlations between quantitative DWI parameters and tumor regression rate of patients with NSCLC
|Parameter||Effective group (n = 103)||Ineffective group (n = 89)||Tumor regression rate|
|∆ADCmid = average apparent diffusion coefficient change rate during treatment; ∆ADCpost = average apparent diffusion coefficient change rate after treatment; ADC = apparent diffusion coefficient; ADCmid = average apparent diffusion coefficient value during treatment; ADCpre = average apparent diffusion coefficient value before treatment; ADCpost = average apparent diffusion coefficient value after treatment; DWI = diffusion-weighted imaging; NSCLC = non-small cell lung cancer.|
|a p<0.05 Compared with the ineffective group.|
|Values are mean ± SD.|
|ADCpre, 10-3 mm2/s||1.36 ± 0.32a||1.54 ± 0.35||-0.261||<0.001|
|ADCmid, 10-3 mm2/s||1.83 ± 0.31a||1.62 ± 0.23||0.312||<0.001|
|ADCpost, 10-3 mm2/s||2.11 ± 0.43a||1.75 ± 0.36||0.368||<0.001|
|∆ADCmid, %||43.72 ± 54.02a||10.64 ± 31.33||0.322||<0.001|
|∆ADCpost, %||68.63 ± 76.73a||19.90 ± 38.62||0.343||<0.001|
Comparisons of DWI parameters in patients with different pathologic types of NSCLC
No significant differences were noted in the average ADC values and changing rate among patients with large cell carcinoma, adenocarcinoma, squamous cell carcinoma, and adenosquamous carcinoma before treatment, during treatment, and 1 month after treatment (all p>0.05) (
Comparisons of average ADC value and its change rate in patients with different pathologic types of NSCLC
|ADC value||Large cell carcinoma (n = 17)||Adenocarcinoma (n = 72)||Squamous cell carcinoma (n = 93)||Adenosquamous carcinoma (n = 10)||p value|
|∆ADCmid = average apparent diffusion coefficient changing rate during treatment; ∆ADCpost = average apparent diffusion coefficient changing rate after treatment; ADC = apparent diffusion coefficient; ADCmid = average apparent diffusion coefficient value during treatment; ADCpre = average apparent diffusion coefficient value before treatment; ADCpost = average apparent diffusion coefficient value after treatment; NSCLC = non-small cell lung cancer.|
|Values are mean ± SD.|
|ADCpre, 10-3 mm2/s||1.60 ± 0.43||1.43 ± 0.34||1.43 ± 0.35||1.36 ± 0.16||0.252|
|ADCmid, 10-3 mm2/s||1.80 ± 0.31||1.75 ± 0.32||1.70 ± 0.27||1.74 ± 0.26||0.581|
|ADCpost, 10-3 mm2/s||1.95 ± 0.44||1.91 ± 0.41||1.98 ± 0.47||1.82 ± 0.28||0.572|
|∆ADCmid, %||20.88 ± 45.32||29.63 ± 41.16||28.68 ± 54.83||29.38 ± 24.60||0.926|
|∆ADCpost, %||34.60 ± 60.41||41.52 ± 49.72||52.65 ± 80.23||36.63 ± 34.10||0.586|
Logistic regression analysis of the short-term efficacy of CCRT in the treatment of patients with NSCLC
Logistic regression analysis was performed using the efficacy of CCRT as the dependent variable and TNM stages (IIIA, IIIB), ADCpre, ADCmid, ADCpost, ∆ADCmid, and ∆ADCpost as covariates. The results indicated that TNM stages (III A, III B) and values of ADCpre, ADCpost, ∆ADCmid, and ∆ADCpostwere independent risk factors that affected the efficacy of CCRT in patients with NSCLC (all p<0.05) (
Logistic regression analysis of short-term efficacy of CCRT in treatment of NSCLC
|Variables analysis||p value||OR value||95% CI||Wald value|
|∆ADCmid = average apparent diffusion coefficient changing rate during treatment; ∆ADCpost = average apparent diffusion coefficient changing rate after treatment; ADC = apparent diffusion coefficient; ADCmid = average apparent diffusion coefficient value during treatment; ADCpre = average apparent diffusion coefficient value before treatment; ADCpost = average apparent diffusion coefficient value after treatment; CCRT = current chemoradiotherapy; CI = confidence interval; OR = odds ratio; NSCLC = non-small cell lung cancer.|
|ADCpre, 10-3 mm2/s||0.002||41.986||3.862-456.496||9.423|
|ADCmid, 10-3 mm2/s||0.785||1.405||0.123-16.114||0.075|
|ADCpost, 10-3 mm2/s||0.002||40.302||3.872-419.494||9.564|
Receiver operating characteristic curve analysis of quantitative DWI parameters for predicting the short-term efficacy of CCRT
Receiver operating characteristic curve analysis indicated that the sensitivity, specificity, and accuracy were 71.9%, 52.4%, and 64.7%, respectively, when the threshold value was set at ADCpre= 1.38 × 10−3mm2/s for predicting the short-term efficacy of CCRT in patients with NSCLC; the corresponding area under the curve (AUC) was 0.637 (
In the last few years, the use of DWI has been extended from predicting and monitoring brain tumors to a diverse variety of diseases, such as malignant hepatic tumors and renal cortical tumors (24, 25). However, few studies have been performed to date to explore its possible role in NSCLC. This study was thus established to determine whether this technique could be applied to predict the efficacy of CCRT on patients with NSCLC.
Our study revealed that quantitative DWI parameters can serve as prognostic markers of tumors for short-term efficacy of CCRT in the treatment of patients with NSCLC by reflecting a clear state of scanning area and tumor proliferation. Previous studies demonstrated that CCRT played an important role in the treatment of stage IIIA NSCLC and provided statistically significant favorable prognostic value in local recurrence-free survival in patients with completely resected NSCLC (26, 27). Diffusion-weighted imaging provides information related to tumor cellularity and cell membrane integrity by measuring ADC, a quantitative measure of the diffusivity of water (20). The technical principle of DWI is based on the motion of water molecules, which is considered unrestricted in a homogenous liquid but impeded by microstructural barriers, such as cell membranes, intracellular organelles, macromolecules, and other tissue compartments, within biological tissues (28). In qualitative assessment, the visual assessment of relative tissue signal attenuation on DWI is used for tumor detection and characterization (29). In addition, studies have found that the signal intensity of pulmonary nodules/masses on DWI may be helpful for malignant and benign tumor differentiation (11, 30, 31). Nasu et al (32) also demonstrated that DWI was effective in predicting a wide range of tumors, including malignant hepatic tumors, which is consistent with our discovery. The possible mechanism of these results could be that during the apoptotic process triggered by CCRT, initial cellular swelling was replaced by a reduction in cellular volume due to membrane blebbing and cell lysis, which was followed by enhanced movement of water between the intracellular and extracellular compartments, causing a change in DWI (33).
Furthermore, our study found that the ADCpre value of patients with NSCLC exhibited a negative correlation to tumor regression rate, whereas ADCmid and ADCpost values were positively correlated to tumor regression rate, indicating that a higher ADCpre value suggested a worse outcome and a higher ADCmid value could predict a better prognosis. A previous study demonstrated that the ADC value can help differentiate viable tumor tissues from necrotic tissues. Intact cell membranes in viable tumor tissues exhibit restricted water diffusion, thus presenting a correspondingly low ADC, while disrupted cellular integrity in necrotic tissues results in increased water diffusion, thus exhibiting a corresponding increase in ADC value (34). The finding that necrotic areas in tumors with high ADCpre values are likely to reduce the delivery of CCRT agents to the tumor is likely due to poor perfusion coupled by a hypoxic environment and reduced metabolism, thus causing the area to be less sensitive to CCRT and to exhibit a smaller regression rate (33). However, when CCRT was initiated gradually, tumor necrosis and apoptosis followed by reduced tumor cellularity and expanded extracellular space resulted in the disintegration of intact cell membranes and more unrestricted movement of water molecules, thus causing an increase in ADCmid and ADCpost value together with a correspondingly proportional reduction in tumor size (28, 35). Studies of a few carcinomas have demonstrated that tumor cells with low baseline ADC values respond better to chemotherapy or radiation treatment than tumors with high pretreatment ADC values (16, 36, 37). Cai et al (38) arrived at a similar conclusion, finding that a strong negative correlation existed between average ADCpre and tumor regression as well as a significant increase in average ADC after CCRT.
In conclusion, our results provide strong evidence that DWI provides promising predictive value in evaluating the short-term efficacy of CCRT in the treatment of patients with NSCLC. However, our study failed to utilize the topotecan + etoposide + cisplatin regimen and to observe the inflammatory phenomena for a longer time period. Furthermore, because no patients received a positron emission tomography (PET) examination before and after treatment, we did not study tumor response using PET combined with DWI. Therefore, we are carrying out further studies to obtain more convincing evidence and to perform more case reviews before verifying DWI’s predictive role in patients with NSCLC treated with CCRT.
Kasper B Gruenwald V Reichardt P Bauer S Hohenberger P Haller F Correlation of CTNNB1 mutation status with progression arrest rate in recist progressive desmoid-type fibromatosis treated with imatinib: Translational research results from a phase 2 study of the german interdisciplinary sarcoma group (GISG-01). 2016 23 6 1924 1927
- Xu, Hai-Dong [PubMed] [Google Scholar] 1, * Corresponding Author (email@example.com)
- Zhang, Yu-Qin [PubMed] [Google Scholar] 1
- Shen, Wei-Yu [PubMed] [Google Scholar] 2
- Mao, Zheng-Chun [PubMed] [Google Scholar] 2
Department of Radiology, Ningbo Medical Center Lihuili Eastern Hospital, Ningbo, Zhejiang - PR China
Department of Thoracic Surgery, Ningbo Medical Center Lihuili Eastern Hospital, Ningbo, Zhejiang - PR China