ORIGINAL RESEARCH
Evaluation of Apparent Diffusion Coefficient Associated With Pathological Grade of Lung Carcinoma, Before Therapy Haidong Liu, MD, Ying Liu, MD, Tielian Yu, MD,* Ning Ye, MB, and Qing Wang, MD Purpose: To investigate the feasibility and utility of apparent diffusion coefficient (ADC) in predicting the tumor cellular density and grades of lung cancers. Materials and Methods: Forty-one consecutive patients (26 men and 15 women; mean age, 59.9 years) with histologically proven lung cancers were enrolled in the study and underwent MR examination. ADC values and tumor cellular density of different histological grades were analyzed. The relationship of the ADC with tumor cellular density and grades were also evaluated. Results: The ADC values of lung cancer in grade III was significantly lower than those in grade I and grade II (P 5 0.008 and 0.011, respectively). The cellular density in grade III was significantly higher than other two grades (P 5 0.029 and 0.022, respectively). ADC value of lung cancer correlated negatively with grades and tumor cellular density (P 5 0.001 and P 5 0.001, respectively). According to the ROC analysis, the cutoff value of ADC was 1.175 3 1023 mm2/s with the optimal sensitivity (88.2%) and specificity (62.5%), respectively. Conclusion: ADC measurement of lung cancer was a helpful method to evaluate the pathological grade and tumor cellular density. The quantitative analysis of ADC in conjunction with conventional MR findings could provide more valuable information for the assessment of pulmonary tumor. J. MAGN. RESON. IMAGING 2015;42:595–601.
L
ung cancer is the most common malignant tumor in many countries and has become the main cause of cancer mortality worldwide.1 The accurate and reliable diagnosis and evaluation of lung cancer is very important. Several noninvasive procedures are widely used for this aim, including CT and PET. Chest MR imaging is gradually increasing in clinical practice.2–4 In addition, the recent developments in diffusion-weighted imaging (DWI) make it feasible in the chest and some studies have addressed its potential advantages and applications in the assessment of pulmonary tumor characterization.5–9 DWI exploits the random motion of water molecules (Brownian motion) in vivo biological tissues and the diffusibility of water molecule can be quantified by the apparent diffusion coefficient (ADC).10 Water diffusion which represents the water microenvironment, such as cellularity, integrity of cell membranes, extracellular space and macromolecules is
complicated and frequently alters with the development of various diseases.11 In malignant tumors, the diffusion of water molecules is restricted, which results in a lower ADC value and facilitates the differential diagnosis of tumors. Some studies suggested that ADC was useful to evaluate pulmonary nodules,6,12 distinguish benign and malignant pulmonary lesions,5,7 and assess mediastinal and hilar nodal staging.13,14 Recently, investigators reported that ADC values could demonstrate the histologic characteristics of lung cancers and had potential ability to distinguish the different subtypes of adenocarcinomas and degree of cell differentiation.5,8,15–17 Pathological grade of lung cancer is an important tumor-related prognostic factor. The methods for different grading tumors may differ. As most lung tumors are heterogeneous, the histological results obtained by biopsy are not very reliable or even false because of improper target selection for biopsy site. Hence, it is useful to predict the tumor
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24823 Received Oct 2, 2014, Accepted for publication Nov 20, 2014. *Address reprint requests to: T.Y., Department of Radiology, General Hospital of Tianjin Medical University, No. 154, Anshan Road, Heping, Tianjin, 300052, China. E-mail: [email protected] Department of Radiology, General Hospital of Tianjin Medical University, Tianjin, China.
C 2014 Wiley Periodicals, Inc. V 595
Journal of Magnetic Resonance Imaging
grade before biopsy. Prior studies have reported that ADC value correlated well with tumor cellular density in the gliomas, uterine cervix cancers, and renal cell carcinomas.18–20 Because lung cancer was a solid tumor, we hypothesized that ADC measurement ought to be helpful for histological grading characterization of pulmonary tumors. The purpose of this study was to further assess correlation between ADC and the pathologic grade of lung cancer, as well as the tumor cellular density.
MATERIALS AND METHODS Subjects This prospective study was approved by our institutional review board and informed consent from patients was obtained before his or her enrollment. Between June 2007 and March 2008, 62 consecutive patients suspected of having lung cancer were assessed for eligibility if they had not undergone any therapy. Forty-one patients (15 women, 26 men; mean age 59.9 years; age range, 38–78 years) histopathologically confirmed lung cancer by surgical resection or biopsy after chest MR examinations were enrolled in this study. The primary tumor diameters were 1.6–10.0 cm.
Chest MR Imaging Technique All MR examinations were performed by a 1.5 Tesla (T) clinical MR scanner (Twin-Speed Infinity with Excite II, GE, USA) and an eight-channel body phased-array coil. To keep the accuracy of comparison, both conventional MR images and DWI were acquired in the same positioning scan line. Patients were in the supine position throughout the examination. Before DWI, conventional T1- and T2weighted images were acquired in the transverse plane. T1-weighted images were obtained by the fast spoiled gradient-echo sequence with the following parameters: repetition time/echo time, 225 ms/ 4.2 ms; number of signals acquired, 1; matrix, 256 3 128; field of view, 36 cm; slice thickness, 6 mm; gap, 1 mm; flip angle, 70 . Respiratory gated T2-weighted fast recovery fast spin echo images were obtained using the following parameters: repetition time/echo time, 4000 ms/85 ms; echo train length, 17; number of signals acquired, 2; matrix, 320 3 224; field of view, slice thickness and gap were same to T1-weighted sequences. DW images were acquired using a single-shot echo-planar imaging sequence with the array spatial sensitivity encoding technique (ASSET) in the transverse plane. The components of the applied gradients for diffusion weighting were equal in three orthogonal directions. The detailed parameters were as follows: b values, 0 and 500 s/mm2; repetition time/echo time, 4000 ms/48.9 ms; field of view, 36 cm; matrix, 128 3 128; section thickness, 6 mm; gap, 1 mm; number of signals acquired, 4; R factor, 2. DWI images were acquired under free breathing conditions.7,8 To minimize the influence of respiratory movement to image quality, a means of respiratory training was taken. The data acquisition time of DWI was 64 s.
MR Image Analysis MR images were reviewed by two senior radiologists (T.L.Y. and N.Y., with 19 years and 18 years’ experience of clinical MR, respectively). Both observers were blinded to evaluate the DWI 596
data in a blinded manner. ADC maps were automatically processed on a built-in Advantage Windows Workstation (ADW4.0 version, GE) with Functool software after scan. After image reconstruction, a single ADC image with largest parts of tumor was selected and region of interest (ROI) was placed within the solid area of the lesion as large as possible. Necrotic regions identified by conventional MR sequences were avoided for ROI placement. The ROI was larger than 1 cm2.19,21
Pathological Grades and Tumor Cellular Density Analysis In our study, lung cancer was divided into three grades according to tumor histological characteristics, that was, grade I (well-differentiated), grade II (moderately-differentiated) and grade III (poorly-differentiated), respectively. Because small cell lung carcinoma belonged to high grade neuroendocrine tumor and demonstrated more aggressive tumor biologic characteristics,1 we categorized it into the grade III. Sarcomatoid carcinoma was defined as a group of poorly differentiated nonsmall cell carcinoma which contained a component of sarcoma or sarcoma-like elements, and had a worse prognosis than other nonsmall cell carcinomas.1 It was similarly categorized into the grade III in our research. All subjects histopathologically proved lung cancers were evaluated for tumor cellular density. All tumor specimens were examined by an experienced thoracic pathologist. Analysis of tumor cellular density was performed by the method similar to previous studies.5,22 The cellular density of specimen was measured with a colored multifunction imaging analysis system. Initially, histological specimens were sliced, stained with hematoxylin and eosin, and then analyzed by means of optical microscopy. Five representative microphotographs of pathological findings in different areas of carcinoma were taken at 200 times amplification, and then were captured into a computer system for the further analysis. To improve the accuracy of our analysis, the microphotographs should select the fields of solid parts in tumor and avoid the areas of nonneoplastic tissue, such as blood vessels, necrosis, calcification and inflammatory cells. Finally, tumor cellular density per microphotograph was defined as the total area of tumor cell nuclei divided by the total histological area in the same microscopic field, and expressed in percentage. The rates were calculated five times in each cancer and the average was considered as the tumor cellular density for analysis.
Statistical Analysis Statistical analysis was performed using statistical software (SPSS, version 18.0). The data was expressed as the mean 6 standard deviation. To indicate the relationship among different parameters, several tables and charts were drawn. One-way analysis of variance was used to compare ADC values and tumor cellular density among the three pathological grades. Independent samples t-test was used to evaluate the difference between ADC values of low grade and high grade lung cancers. Pearson correlation was performed to determine relationship between ADC values of lung cancer and tumor cellular density. In addition, Spearman rank correlation analysis was used to demonstrate the relationship between the grades of tumor and ADC values of lung cancer. The receiver operating characteristic (ROC) analysis was used to determine the cutoff value of the ADC differentiating the low grade from the Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma
TABLE 1. Relationship Between Pathological Diagnosis and Grade of Lung Cancer
Pathological diagnosis Squamous cell carcinoma
Adenocarcinoma
Adenosquamous carcinoma
Small cell carcinoma
Other typesa
Total
I
5
3
0
0
0
8
II
5
10
1
0
0
16
III
2
3
1
9
2
17
Total
12
16
2
9
2
41
grades
a
Other types included one pulmonary sarcomatoid carcinoma confirmed by surgical resection and one poorly-differentiated carcinoma diagnosed by biopsy. Both of them were categorized into grade III. high grade lung cancers. A P value of less than 0.05 was considered statistically significant.
1023mm2/s) were significantly lower than those of low grade ([1.348 6 0.378] 3 1023mm2/s) (P 5 0.003).
RESULTS
Comparison of Tumor Cellular Density Among Different Pathological Grades The results of tumor cellular density were also exhibited in Table 2. The tumor cellular density also demonstrated significant differences among different grades (P 5 0.029) (Fig. 3). Although the tumor cellular density of grade III was significantly greater than that of the other two types (P 5 0.029 and 0.022, respectively), there was no significant difference between grade I and grade II (P 5 0.744).
Findings of Pathological Diagnosis Among the pathological diagnosis of 41 subjects, 29 patients were confirmed by surgical resection and 12 patients were proved by biopsy. According to the pathological examination results, different histological types and grades were shown in Table 1 in details. Up to December 31, 2013, only 12 patients were alive and the rest were dead. Comparison of ADC Values Among Different Pathological Grades The results of ADC values were exhibited in Table 2. There were significant differences among the ADC values of three tumor grades (P 5 0.009) (Figs. 1 and 2). Furthermore, the multiple comparison displayed that the ADC values of grade III were significantly lower than the other two grades (P 5 0.008 and 0.011, respectively). However, there was no significant difference between ADC values of grade I and grade II (P 5 0.548). Because of no significant differences between the ADC values of grade I and grade II, both grades were clubbed as low grade (n 5 24) and grades III were clubbed as high grade (n 5 17) lung cancers. ADC values of high grade lung cancers ([1.030 6 0.179] 3
Relationship Between ADC Values and Tumor Pathological Grades Table 2 showed that the ADC values among different pathological grades had significant differences (P 5 0.009). The higher grade tumors tended to have lower ADC values and vice versa. There was a significant negative linear correlation between ADC values of lung cancer and pathological grades (Spearman rank coefficient, 20.507; P 5 0.001).
TABLE 2. ADC Values and Tumor Cellular Density of Three Pathological Grades
n
ADC value (31023mm2/s)
Tumor cellular density (%)
I
8
1.40460.352
14.97363.329
II
16
1.32160.398
15.75264.155
III
17
1.03060.179
20.29567.052
Total
41
1.21660.347
17.48365.842
Grades
September 2015
FIGURE 1: Comparison of ADC value among different grades of lung cancer. The ADC values had significant difference, but there were some overlaps among them.
597
Journal of Magnetic Resonance Imaging
FIGURE 2: A 66-year-old woman with a grade III adenocarcinoma. a: Contrast enhanced CT scan demonstrated a homogenous enhancing mass in the right inferior lobe. b: T2-weighted image showed a hyperintense mass adjacent to chest wall. c: On DWI, the lesion showed markedly restriction of diffusion with high signal intensity. d: On ADC map, example of ROI placement within the solid parts of tumor was depicted. The tumor revealed a hypointense area with the ADC value of 1.430 3 1023 mm2/s. e: The histological section showed a large number of tumor cells with obvious nuclear atypia and the tumor cellular density was 19.075% (hematoxylin and eosin stain, magnification 3200).
ROC Analysis ROC analysis was performed for ADC in differentiating low grade and high grade lung cancers. According to the ROC analysis, the cutoff value of 1.175 3 1023mm2/s for the ADC value generated the best combination of sensitivity (88.2%) and specificity (62.5%) (Fig. 4). Relationship Between ADC Values and Tumor Cellular Density The ADC values and tumor cellular density of 41 patients were (1.216 6 0.347) 3 1023 mm2/s and (17.483 6 5.842)%, 598
respectively. There was a significant negative linear correlation between both (Pearson coefficient, 20.652; P 5 0.001) (Fig. 5).
DISCUSSION Despite advances in treatment of lung cancer, some histological types still have poor prognosis.1 The tumor grade is related to the prognosis. Regimbeau et al 23 reported that tumor grade was one of the most predictive factors for early recurrence after resection of hepatocellular carcinoma. Hence, predicting tumor grade by a reliable imaging Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma
FIGURE 3: Comparison of tumor cellular density among different grades of lung cancer. The tumor cellular density of grade III was significantly greater than others, but there was no significant difference between grade I and grade II.
modality is helpful for evaluating prognosis. DWI can reflect the characterization of biological tissues noninvasively by measurement of properties of water diffusion and quantitative analysis of ADC.24 Nakanishi et al 25 suggested that DWI had a potential role for tumor grade and prediction of early hepatocellular carcinoma recurrence before treatment. Furthermore, other Investigators also considered that ADC value was a good predictor for tumor grade.26,27 Our data showed ADC value of lung cancer was negatively related to grade. In higher grade the ADC was lower and vice versa. The result was similar to studies by Lee et al 18 and Goyal et al.21 One possible explanation for this was that the pathological tumor grade systems were based on nuclear features. With grades of lung cancers increasing, the sizes of the nucleus, nucleolus, nuclear-cytoplasmic ratio and nuclear
FIGURE 4: The ROC of ADC value for differentiating the lowgrade lung cancers from the high-grade ones showed that the optimal cutoff value was 1.175 3 1023 mm2/s and the sensitivity and specificity were 88.2% and 62.5%, respectively. Area under curve: 0.801 (95% confidence interval: 0.665–0.938).
September 2015
FIGURE 5: Relationship between ADC value and tumor cellular density of lung cancer. There was a significant negative correlation between them (Pearson coefficient, 20.652; P 5 0.001).
membrane irregularity increased.21 These factors could lead to greater hindrance to the motion of water molecules both in the extracellular and intracellular space in higher grades. The diffusibility of water molecules in malignant tumors is very complicated and influenced by many factors, such as rapid cell proliferation, high cell density, large nuclei, high amounts of intracellular macromolecular proteins, high nuclear/cytoplasm ratio, and small extracellular space.28,29 Some studies revealed that the water diffusion in tumor was mainly affected by tumor cellular density and the motion of water molecule in the interstitium mainly contributed to ADC values.22 When tumor cellular density increased, that was, the number of cells in unit volume were larger and the extracellular space relatively decreased, water diffusion was further restricted and finally resulted in ADC value reducing obviously.19 Therefore, ADC could provide us a promising way to evaluate the histological characteristics of tumor. Some studies revealed that the changes of ADC values were significantly earlier than morphological changes to evaluate tumor therapy.30,31 DWI was helpful to predict and monitor the therapeutic efficacy at the time of early treatment of tumor.32 Although our data revealed there was a negative linear correlation between ADC values and tumor cellular density, the correlation coefficient was not perfect (only 20.652). The higher cellularity could lead to more cell membranes and smaller extracellular space.32 These increased the barriers of water molecules’ diffusion and led to lower ADC. However, the diffusion of water molecules and ADC values were affected by a multitude of factors including cellularity.32 Hence, a full linear relationship between ADC and the inverse cellular density had not been established by our existing experimental data yet. At the same time, we should also note that necrotic, cystic or calcified areas within tumors had potential influence to accurate ADC measurements. Because 599
Journal of Magnetic Resonance Imaging
these areas often observed in lung cancer had smaller cellular density and larger extracellular space, the ADC values were higher than solid parts of tumor. To accurately measure ADC values and avoid potentially increasing, we should refer to the T2-weighted images or enhanced CT images and attempt to draw on the solid part of the tumors when we determined the position of ROI on ADC maps. There were some limitations in our study. First, we could not confirm the establishment of a point-to-point correspondence between histological specimens and the sites of ADC measured. To achieve this aim, the CT-guided percutaneous lung biopsy might be a good method. However, there were only a minority confirmed by biopsy and it was inevitable for the potential mismatch between them. More precise point to point experimental studies between ADC and histological specimen were required to resolve these problems in the future. Second, there was some overlaps between ADC values of high grade and low grade lung cancers. Hence, ADC analysis for histological grade of lung cancer should be in conjunction with the conventional imaging findings. Third, we chose a relatively small b value to improve image quality. The b value was a critical and adjustable parameter in DWI sequence. Some previous studies had reported that ADC values tended to be higher in low b values because of influence by tissue perfusion and T2 time to ADC values.5,33 Furthermore, the ADC measurement in higher b value seemed more accurate than that in lower b value. However, image quality would be greatly diminished and susceptibility artifacts would be very obvious in much higher b value in our study. We also took some other measures to gain good image quality, such as phase array coil, parallel imaging technology. Fortunately, image quality of DWI could meet the needs of the diagnosis. Finally, the number of sample was relatively small and a powerful analysis with large sample should be performed in further studies. In conclusion, ADC measurement was feasible in evaluating grade of lung cancer. Our data illustrated that ADC value was affected by tumor cellular density and could differentiate pathological grades of lung cancers. ADC was a useful and promising parameter to predict tumor cellular density and grade of lung cancers. We considered ADC could aid conventional MR to accurately assess pulmonary tumor characterization.
4.
Ohno Y. Since the clinical introduction of magnetic resonance (MR) imaging, the lung has been one of the most challenging applications. J Thorac Imaging 2013;28:135–136.
5.
Matoba M, Tonami H, Kondou T, et al. Lung carcinoma: diffusionweighted MR imaging–preliminary evaluation with apparent diffusion coefficient. Radiology 2007;243:570–577.
6.
Satoh S, Kitazume Y, Ohdama S, et al. Can malignant and benign pulmonary nodules be differentiated with diffusion-weighted MRI? AJR Am J Roentgenol 2008;191:464–470.
7.
Liu H, Liu Y, Yu T, Ye N. Usefulness of diffusion-weighted MR imaging in the evaluation of pulmonary lesions. Eur Radiol 2010;20:807–815.
8.
Koyama H, Ohno Y, Aoyama N, et al. Comparison of STIR turbo SE imaging and diffusion-weighted imaging of the lung: capability for detection and subtype classification of pulmonary adenocarcinomas. Eur Radiol 2010;20:790–800.
9.
Chen L, Zhang J, Bao J, et al. Meta-analysis of diffusion-weighted MRI in the differential diagnosis of lung lesions. J Magn Reson Imaging 2013;37:1351–1358.
10.
Herneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol 2003;45:208–213.
11.
Padhani AR, Liu G, Koh DM, et al. Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 2009;11:102–125.
12.
Wu LM, Xu JR, Hua J, et al. Can diffusion-weighted imaging be used as a reliable sequence in the detection of malignant pulmonary nodules and masses? Magn Reson Imaging 2013;31:235–246.
13.
Kim YN, Yi CA, Lee KS, et al. A proposal for combined MRI and PET/ CT interpretation criteria for preoperative nodal staging in non-smallcell lung cancer. Eur Radiol 2012;22:1537–1546.
14.
Wu LM, Xu JR, Gu HY, et al. Preoperative mediastinal and hilar nodal staging with diffusion-weighted magnetic resonance imagingand fluorodeoxyglucose positron emission tomography/computed tomography in patients with non-small-cell lung cancer: which is better? J Surg Res 2012;178:304–314.
15.
Tanaka R, Horikoshi H, Yoshida T, Nakazato Y, Seki E, Goya T. Diffusion-weighted magnetic resonance imaging in differentiating the invasiveness of small lung adenocarcinoma. Acta Radiol 2011;52: 750–755.
16.
Tanaka R, Nakazato Y, Horikoshi H, et al. Diffusion-weighted imaging and positron emission tomography in various cytological subtypes of primary lung adenocarcinoma. Clin Imaging 2013;37:876–883.
17.
Tanaka R, Horikoshi H, Nakazato Y, et al. Magnetic resonance imaging in peripheral lung adenocarcinoma: correlation with histopathologic features. J Thorac Imaging 2009;24:4–9.
18.
Lee EJ, Lee SK, Agid R, Bae JM, Keller A, Terbrugge K. Preoperative grading of presumptive low-grade astrocytomas on MR imaging: diagnostic value of minimum apparent diffusion coefficient. AJNR Am J Neuroradiol 2008;29:1872–1877.
19.
Liu Y, Bai R, Sun H, Liu H, Wang D. Diffusion-weighted magnetic resonance imaging of uterine cervical cancer. J Comput Assist Tomogr 2009;33:858–862.
20.
Yu X, Lin M, Ouyang H, Zhou C, Zhang H. Application of ADC measurement in characterization of renal cell carcinomas with different pathological types and grades by 3.0T diffusion-weighted MRI. Eur J Radiol 2012;81:3061–3066.
21.
Goyal A, Sharma R, Bhalla AS, et al. Diffusion-weighted MRI in renal cell carcinoma: a surrogate marker for predicting nuclear grade and histologicalsubtype. Acta Radiol 2012;53:349–358.
22.
Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999;9:53–60.
23.
Regimbeau JM, Abdalla EK, Vauthey JN, et al. Risk factors for early death due to recurrence after liver resection for hepatocellular carcinoma: results of a multicenter study. J Surg Oncol 2004;85:36–41.
REFERENCES 1.
Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The new World Health Organization classification of lung tumours. Eur Respir J 2001;18:1059–1068.
2.
Ohno Y, Hatabu H, Takenaka D, Adachi S, Kono M, Sugimura K. Solitary pulmonary nodules: potential role of dynamic MR imaging in management initial experience. Radiology 2002;224:503–511.
3.
Puderbach M, Hintze C, Ley S, et al. MR imaging of the chest: a practical approach at 1.5T. Eur J Radiol 2007;64:345–355.
600
Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma 24.
Guo Y, Cai YQ, Cai ZL, et al. Differentiation of clinically benign and malignant breast lesions using diffusion-weighted imaging. J Magn Reson Imaging 2002;16:172–178.
29.
Ichikawa T, Erturk SM, Motosugi U, et al. High-B-value diffusionweighted MRI in colorectal cancer. AJR Am J Roentgenol 2006;187: 181–184.
25.
Nakanishi M, Chuma M, Hige S, et al. Relationship between diffusionweighted magnetic resonance imaging and histological tumor grading of hepatocellular carcinoma. Ann Surg Oncol 2012;19:1302–1309.
30.
26.
Higano S, Yun X, Kumabe T, et al. Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis. Radiology 2006;241:839–846.
Park SH, Moon WK, Cho N, et al. Comparison of diffusion-weighted MR imaging and FDG PET/CT to predict pathological complete response to neoadjuvant chemotherapy in patients with breast cancer. Eur Radiol 2012;22:18–25.
31.
Calvar JA, Meli FJ, Romero C, et al. Characterization of brain tumors by MRS, DWI and Ki-67 labeling index. J Neurooncol 2005;72:273– 280.
Shin HJ, Baek HM, Ahn JH, et al. Prediction of pathologic response to neoadjuvant chemotherapy in patients with breast cancer using diffusion-weighted imaging and MRS. NMR Biomed 2012;25:1349–1359.
32.
Chen L, Zhang J, Chen Y, et al. Relationship between apparent diffusion coefficient and tumour cellularity in lung cancer. PLoS One 2014;9:e99865.
33.
Le Bihan D, Breton E, Lallemand D, et al. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168:497–505.
27.
28.
Henzler T, Schmid-Bindert G, Schoenberg SO, Fink C. Diffusion and perfusion MRI of the lung and mediastinum. Eur J Radiol 2010;76: 329–336.
September 2015
601
Evaluation of Apparent Diffusion Coefficient Associated With Pathological Grade of Lung Carcinoma, Before Therapy Haidong Liu, MD, Ying Liu, MD, Tielian Yu, MD,* Ning Ye, MB, and Qing Wang, MD Purpose: To investigate the feasibility and utility of apparent diffusion coefficient (ADC) in predicting the tumor cellular density and grades of lung cancers. Materials and Methods: Forty-one consecutive patients (26 men and 15 women; mean age, 59.9 years) with histologically proven lung cancers were enrolled in the study and underwent MR examination. ADC values and tumor cellular density of different histological grades were analyzed. The relationship of the ADC with tumor cellular density and grades were also evaluated. Results: The ADC values of lung cancer in grade III was significantly lower than those in grade I and grade II (P 5 0.008 and 0.011, respectively). The cellular density in grade III was significantly higher than other two grades (P 5 0.029 and 0.022, respectively). ADC value of lung cancer correlated negatively with grades and tumor cellular density (P 5 0.001 and P 5 0.001, respectively). According to the ROC analysis, the cutoff value of ADC was 1.175 3 1023 mm2/s with the optimal sensitivity (88.2%) and specificity (62.5%), respectively. Conclusion: ADC measurement of lung cancer was a helpful method to evaluate the pathological grade and tumor cellular density. The quantitative analysis of ADC in conjunction with conventional MR findings could provide more valuable information for the assessment of pulmonary tumor. J. MAGN. RESON. IMAGING 2015;42:595–601.
L
ung cancer is the most common malignant tumor in many countries and has become the main cause of cancer mortality worldwide.1 The accurate and reliable diagnosis and evaluation of lung cancer is very important. Several noninvasive procedures are widely used for this aim, including CT and PET. Chest MR imaging is gradually increasing in clinical practice.2–4 In addition, the recent developments in diffusion-weighted imaging (DWI) make it feasible in the chest and some studies have addressed its potential advantages and applications in the assessment of pulmonary tumor characterization.5–9 DWI exploits the random motion of water molecules (Brownian motion) in vivo biological tissues and the diffusibility of water molecule can be quantified by the apparent diffusion coefficient (ADC).10 Water diffusion which represents the water microenvironment, such as cellularity, integrity of cell membranes, extracellular space and macromolecules is
complicated and frequently alters with the development of various diseases.11 In malignant tumors, the diffusion of water molecules is restricted, which results in a lower ADC value and facilitates the differential diagnosis of tumors. Some studies suggested that ADC was useful to evaluate pulmonary nodules,6,12 distinguish benign and malignant pulmonary lesions,5,7 and assess mediastinal and hilar nodal staging.13,14 Recently, investigators reported that ADC values could demonstrate the histologic characteristics of lung cancers and had potential ability to distinguish the different subtypes of adenocarcinomas and degree of cell differentiation.5,8,15–17 Pathological grade of lung cancer is an important tumor-related prognostic factor. The methods for different grading tumors may differ. As most lung tumors are heterogeneous, the histological results obtained by biopsy are not very reliable or even false because of improper target selection for biopsy site. Hence, it is useful to predict the tumor
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24823 Received Oct 2, 2014, Accepted for publication Nov 20, 2014. *Address reprint requests to: T.Y., Department of Radiology, General Hospital of Tianjin Medical University, No. 154, Anshan Road, Heping, Tianjin, 300052, China. E-mail: [email protected] Department of Radiology, General Hospital of Tianjin Medical University, Tianjin, China.
C 2014 Wiley Periodicals, Inc. V 595
Journal of Magnetic Resonance Imaging
grade before biopsy. Prior studies have reported that ADC value correlated well with tumor cellular density in the gliomas, uterine cervix cancers, and renal cell carcinomas.18–20 Because lung cancer was a solid tumor, we hypothesized that ADC measurement ought to be helpful for histological grading characterization of pulmonary tumors. The purpose of this study was to further assess correlation between ADC and the pathologic grade of lung cancer, as well as the tumor cellular density.
MATERIALS AND METHODS Subjects This prospective study was approved by our institutional review board and informed consent from patients was obtained before his or her enrollment. Between June 2007 and March 2008, 62 consecutive patients suspected of having lung cancer were assessed for eligibility if they had not undergone any therapy. Forty-one patients (15 women, 26 men; mean age 59.9 years; age range, 38–78 years) histopathologically confirmed lung cancer by surgical resection or biopsy after chest MR examinations were enrolled in this study. The primary tumor diameters were 1.6–10.0 cm.
Chest MR Imaging Technique All MR examinations were performed by a 1.5 Tesla (T) clinical MR scanner (Twin-Speed Infinity with Excite II, GE, USA) and an eight-channel body phased-array coil. To keep the accuracy of comparison, both conventional MR images and DWI were acquired in the same positioning scan line. Patients were in the supine position throughout the examination. Before DWI, conventional T1- and T2weighted images were acquired in the transverse plane. T1-weighted images were obtained by the fast spoiled gradient-echo sequence with the following parameters: repetition time/echo time, 225 ms/ 4.2 ms; number of signals acquired, 1; matrix, 256 3 128; field of view, 36 cm; slice thickness, 6 mm; gap, 1 mm; flip angle, 70 . Respiratory gated T2-weighted fast recovery fast spin echo images were obtained using the following parameters: repetition time/echo time, 4000 ms/85 ms; echo train length, 17; number of signals acquired, 2; matrix, 320 3 224; field of view, slice thickness and gap were same to T1-weighted sequences. DW images were acquired using a single-shot echo-planar imaging sequence with the array spatial sensitivity encoding technique (ASSET) in the transverse plane. The components of the applied gradients for diffusion weighting were equal in three orthogonal directions. The detailed parameters were as follows: b values, 0 and 500 s/mm2; repetition time/echo time, 4000 ms/48.9 ms; field of view, 36 cm; matrix, 128 3 128; section thickness, 6 mm; gap, 1 mm; number of signals acquired, 4; R factor, 2. DWI images were acquired under free breathing conditions.7,8 To minimize the influence of respiratory movement to image quality, a means of respiratory training was taken. The data acquisition time of DWI was 64 s.
MR Image Analysis MR images were reviewed by two senior radiologists (T.L.Y. and N.Y., with 19 years and 18 years’ experience of clinical MR, respectively). Both observers were blinded to evaluate the DWI 596
data in a blinded manner. ADC maps were automatically processed on a built-in Advantage Windows Workstation (ADW4.0 version, GE) with Functool software after scan. After image reconstruction, a single ADC image with largest parts of tumor was selected and region of interest (ROI) was placed within the solid area of the lesion as large as possible. Necrotic regions identified by conventional MR sequences were avoided for ROI placement. The ROI was larger than 1 cm2.19,21
Pathological Grades and Tumor Cellular Density Analysis In our study, lung cancer was divided into three grades according to tumor histological characteristics, that was, grade I (well-differentiated), grade II (moderately-differentiated) and grade III (poorly-differentiated), respectively. Because small cell lung carcinoma belonged to high grade neuroendocrine tumor and demonstrated more aggressive tumor biologic characteristics,1 we categorized it into the grade III. Sarcomatoid carcinoma was defined as a group of poorly differentiated nonsmall cell carcinoma which contained a component of sarcoma or sarcoma-like elements, and had a worse prognosis than other nonsmall cell carcinomas.1 It was similarly categorized into the grade III in our research. All subjects histopathologically proved lung cancers were evaluated for tumor cellular density. All tumor specimens were examined by an experienced thoracic pathologist. Analysis of tumor cellular density was performed by the method similar to previous studies.5,22 The cellular density of specimen was measured with a colored multifunction imaging analysis system. Initially, histological specimens were sliced, stained with hematoxylin and eosin, and then analyzed by means of optical microscopy. Five representative microphotographs of pathological findings in different areas of carcinoma were taken at 200 times amplification, and then were captured into a computer system for the further analysis. To improve the accuracy of our analysis, the microphotographs should select the fields of solid parts in tumor and avoid the areas of nonneoplastic tissue, such as blood vessels, necrosis, calcification and inflammatory cells. Finally, tumor cellular density per microphotograph was defined as the total area of tumor cell nuclei divided by the total histological area in the same microscopic field, and expressed in percentage. The rates were calculated five times in each cancer and the average was considered as the tumor cellular density for analysis.
Statistical Analysis Statistical analysis was performed using statistical software (SPSS, version 18.0). The data was expressed as the mean 6 standard deviation. To indicate the relationship among different parameters, several tables and charts were drawn. One-way analysis of variance was used to compare ADC values and tumor cellular density among the three pathological grades. Independent samples t-test was used to evaluate the difference between ADC values of low grade and high grade lung cancers. Pearson correlation was performed to determine relationship between ADC values of lung cancer and tumor cellular density. In addition, Spearman rank correlation analysis was used to demonstrate the relationship between the grades of tumor and ADC values of lung cancer. The receiver operating characteristic (ROC) analysis was used to determine the cutoff value of the ADC differentiating the low grade from the Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma
TABLE 1. Relationship Between Pathological Diagnosis and Grade of Lung Cancer
Pathological diagnosis Squamous cell carcinoma
Adenocarcinoma
Adenosquamous carcinoma
Small cell carcinoma
Other typesa
Total
I
5
3
0
0
0
8
II
5
10
1
0
0
16
III
2
3
1
9
2
17
Total
12
16
2
9
2
41
grades
a
Other types included one pulmonary sarcomatoid carcinoma confirmed by surgical resection and one poorly-differentiated carcinoma diagnosed by biopsy. Both of them were categorized into grade III. high grade lung cancers. A P value of less than 0.05 was considered statistically significant.
1023mm2/s) were significantly lower than those of low grade ([1.348 6 0.378] 3 1023mm2/s) (P 5 0.003).
RESULTS
Comparison of Tumor Cellular Density Among Different Pathological Grades The results of tumor cellular density were also exhibited in Table 2. The tumor cellular density also demonstrated significant differences among different grades (P 5 0.029) (Fig. 3). Although the tumor cellular density of grade III was significantly greater than that of the other two types (P 5 0.029 and 0.022, respectively), there was no significant difference between grade I and grade II (P 5 0.744).
Findings of Pathological Diagnosis Among the pathological diagnosis of 41 subjects, 29 patients were confirmed by surgical resection and 12 patients were proved by biopsy. According to the pathological examination results, different histological types and grades were shown in Table 1 in details. Up to December 31, 2013, only 12 patients were alive and the rest were dead. Comparison of ADC Values Among Different Pathological Grades The results of ADC values were exhibited in Table 2. There were significant differences among the ADC values of three tumor grades (P 5 0.009) (Figs. 1 and 2). Furthermore, the multiple comparison displayed that the ADC values of grade III were significantly lower than the other two grades (P 5 0.008 and 0.011, respectively). However, there was no significant difference between ADC values of grade I and grade II (P 5 0.548). Because of no significant differences between the ADC values of grade I and grade II, both grades were clubbed as low grade (n 5 24) and grades III were clubbed as high grade (n 5 17) lung cancers. ADC values of high grade lung cancers ([1.030 6 0.179] 3
Relationship Between ADC Values and Tumor Pathological Grades Table 2 showed that the ADC values among different pathological grades had significant differences (P 5 0.009). The higher grade tumors tended to have lower ADC values and vice versa. There was a significant negative linear correlation between ADC values of lung cancer and pathological grades (Spearman rank coefficient, 20.507; P 5 0.001).
TABLE 2. ADC Values and Tumor Cellular Density of Three Pathological Grades
n
ADC value (31023mm2/s)
Tumor cellular density (%)
I
8
1.40460.352
14.97363.329
II
16
1.32160.398
15.75264.155
III
17
1.03060.179
20.29567.052
Total
41
1.21660.347
17.48365.842
Grades
September 2015
FIGURE 1: Comparison of ADC value among different grades of lung cancer. The ADC values had significant difference, but there were some overlaps among them.
597
Journal of Magnetic Resonance Imaging
FIGURE 2: A 66-year-old woman with a grade III adenocarcinoma. a: Contrast enhanced CT scan demonstrated a homogenous enhancing mass in the right inferior lobe. b: T2-weighted image showed a hyperintense mass adjacent to chest wall. c: On DWI, the lesion showed markedly restriction of diffusion with high signal intensity. d: On ADC map, example of ROI placement within the solid parts of tumor was depicted. The tumor revealed a hypointense area with the ADC value of 1.430 3 1023 mm2/s. e: The histological section showed a large number of tumor cells with obvious nuclear atypia and the tumor cellular density was 19.075% (hematoxylin and eosin stain, magnification 3200).
ROC Analysis ROC analysis was performed for ADC in differentiating low grade and high grade lung cancers. According to the ROC analysis, the cutoff value of 1.175 3 1023mm2/s for the ADC value generated the best combination of sensitivity (88.2%) and specificity (62.5%) (Fig. 4). Relationship Between ADC Values and Tumor Cellular Density The ADC values and tumor cellular density of 41 patients were (1.216 6 0.347) 3 1023 mm2/s and (17.483 6 5.842)%, 598
respectively. There was a significant negative linear correlation between both (Pearson coefficient, 20.652; P 5 0.001) (Fig. 5).
DISCUSSION Despite advances in treatment of lung cancer, some histological types still have poor prognosis.1 The tumor grade is related to the prognosis. Regimbeau et al 23 reported that tumor grade was one of the most predictive factors for early recurrence after resection of hepatocellular carcinoma. Hence, predicting tumor grade by a reliable imaging Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma
FIGURE 3: Comparison of tumor cellular density among different grades of lung cancer. The tumor cellular density of grade III was significantly greater than others, but there was no significant difference between grade I and grade II.
modality is helpful for evaluating prognosis. DWI can reflect the characterization of biological tissues noninvasively by measurement of properties of water diffusion and quantitative analysis of ADC.24 Nakanishi et al 25 suggested that DWI had a potential role for tumor grade and prediction of early hepatocellular carcinoma recurrence before treatment. Furthermore, other Investigators also considered that ADC value was a good predictor for tumor grade.26,27 Our data showed ADC value of lung cancer was negatively related to grade. In higher grade the ADC was lower and vice versa. The result was similar to studies by Lee et al 18 and Goyal et al.21 One possible explanation for this was that the pathological tumor grade systems were based on nuclear features. With grades of lung cancers increasing, the sizes of the nucleus, nucleolus, nuclear-cytoplasmic ratio and nuclear
FIGURE 4: The ROC of ADC value for differentiating the lowgrade lung cancers from the high-grade ones showed that the optimal cutoff value was 1.175 3 1023 mm2/s and the sensitivity and specificity were 88.2% and 62.5%, respectively. Area under curve: 0.801 (95% confidence interval: 0.665–0.938).
September 2015
FIGURE 5: Relationship between ADC value and tumor cellular density of lung cancer. There was a significant negative correlation between them (Pearson coefficient, 20.652; P 5 0.001).
membrane irregularity increased.21 These factors could lead to greater hindrance to the motion of water molecules both in the extracellular and intracellular space in higher grades. The diffusibility of water molecules in malignant tumors is very complicated and influenced by many factors, such as rapid cell proliferation, high cell density, large nuclei, high amounts of intracellular macromolecular proteins, high nuclear/cytoplasm ratio, and small extracellular space.28,29 Some studies revealed that the water diffusion in tumor was mainly affected by tumor cellular density and the motion of water molecule in the interstitium mainly contributed to ADC values.22 When tumor cellular density increased, that was, the number of cells in unit volume were larger and the extracellular space relatively decreased, water diffusion was further restricted and finally resulted in ADC value reducing obviously.19 Therefore, ADC could provide us a promising way to evaluate the histological characteristics of tumor. Some studies revealed that the changes of ADC values were significantly earlier than morphological changes to evaluate tumor therapy.30,31 DWI was helpful to predict and monitor the therapeutic efficacy at the time of early treatment of tumor.32 Although our data revealed there was a negative linear correlation between ADC values and tumor cellular density, the correlation coefficient was not perfect (only 20.652). The higher cellularity could lead to more cell membranes and smaller extracellular space.32 These increased the barriers of water molecules’ diffusion and led to lower ADC. However, the diffusion of water molecules and ADC values were affected by a multitude of factors including cellularity.32 Hence, a full linear relationship between ADC and the inverse cellular density had not been established by our existing experimental data yet. At the same time, we should also note that necrotic, cystic or calcified areas within tumors had potential influence to accurate ADC measurements. Because 599
Journal of Magnetic Resonance Imaging
these areas often observed in lung cancer had smaller cellular density and larger extracellular space, the ADC values were higher than solid parts of tumor. To accurately measure ADC values and avoid potentially increasing, we should refer to the T2-weighted images or enhanced CT images and attempt to draw on the solid part of the tumors when we determined the position of ROI on ADC maps. There were some limitations in our study. First, we could not confirm the establishment of a point-to-point correspondence between histological specimens and the sites of ADC measured. To achieve this aim, the CT-guided percutaneous lung biopsy might be a good method. However, there were only a minority confirmed by biopsy and it was inevitable for the potential mismatch between them. More precise point to point experimental studies between ADC and histological specimen were required to resolve these problems in the future. Second, there was some overlaps between ADC values of high grade and low grade lung cancers. Hence, ADC analysis for histological grade of lung cancer should be in conjunction with the conventional imaging findings. Third, we chose a relatively small b value to improve image quality. The b value was a critical and adjustable parameter in DWI sequence. Some previous studies had reported that ADC values tended to be higher in low b values because of influence by tissue perfusion and T2 time to ADC values.5,33 Furthermore, the ADC measurement in higher b value seemed more accurate than that in lower b value. However, image quality would be greatly diminished and susceptibility artifacts would be very obvious in much higher b value in our study. We also took some other measures to gain good image quality, such as phase array coil, parallel imaging technology. Fortunately, image quality of DWI could meet the needs of the diagnosis. Finally, the number of sample was relatively small and a powerful analysis with large sample should be performed in further studies. In conclusion, ADC measurement was feasible in evaluating grade of lung cancer. Our data illustrated that ADC value was affected by tumor cellular density and could differentiate pathological grades of lung cancers. ADC was a useful and promising parameter to predict tumor cellular density and grade of lung cancers. We considered ADC could aid conventional MR to accurately assess pulmonary tumor characterization.
4.
Ohno Y. Since the clinical introduction of magnetic resonance (MR) imaging, the lung has been one of the most challenging applications. J Thorac Imaging 2013;28:135–136.
5.
Matoba M, Tonami H, Kondou T, et al. Lung carcinoma: diffusionweighted MR imaging–preliminary evaluation with apparent diffusion coefficient. Radiology 2007;243:570–577.
6.
Satoh S, Kitazume Y, Ohdama S, et al. Can malignant and benign pulmonary nodules be differentiated with diffusion-weighted MRI? AJR Am J Roentgenol 2008;191:464–470.
7.
Liu H, Liu Y, Yu T, Ye N. Usefulness of diffusion-weighted MR imaging in the evaluation of pulmonary lesions. Eur Radiol 2010;20:807–815.
8.
Koyama H, Ohno Y, Aoyama N, et al. Comparison of STIR turbo SE imaging and diffusion-weighted imaging of the lung: capability for detection and subtype classification of pulmonary adenocarcinomas. Eur Radiol 2010;20:790–800.
9.
Chen L, Zhang J, Bao J, et al. Meta-analysis of diffusion-weighted MRI in the differential diagnosis of lung lesions. J Magn Reson Imaging 2013;37:1351–1358.
10.
Herneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol 2003;45:208–213.
11.
Padhani AR, Liu G, Koh DM, et al. Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 2009;11:102–125.
12.
Wu LM, Xu JR, Hua J, et al. Can diffusion-weighted imaging be used as a reliable sequence in the detection of malignant pulmonary nodules and masses? Magn Reson Imaging 2013;31:235–246.
13.
Kim YN, Yi CA, Lee KS, et al. A proposal for combined MRI and PET/ CT interpretation criteria for preoperative nodal staging in non-smallcell lung cancer. Eur Radiol 2012;22:1537–1546.
14.
Wu LM, Xu JR, Gu HY, et al. Preoperative mediastinal and hilar nodal staging with diffusion-weighted magnetic resonance imagingand fluorodeoxyglucose positron emission tomography/computed tomography in patients with non-small-cell lung cancer: which is better? J Surg Res 2012;178:304–314.
15.
Tanaka R, Horikoshi H, Yoshida T, Nakazato Y, Seki E, Goya T. Diffusion-weighted magnetic resonance imaging in differentiating the invasiveness of small lung adenocarcinoma. Acta Radiol 2011;52: 750–755.
16.
Tanaka R, Nakazato Y, Horikoshi H, et al. Diffusion-weighted imaging and positron emission tomography in various cytological subtypes of primary lung adenocarcinoma. Clin Imaging 2013;37:876–883.
17.
Tanaka R, Horikoshi H, Nakazato Y, et al. Magnetic resonance imaging in peripheral lung adenocarcinoma: correlation with histopathologic features. J Thorac Imaging 2009;24:4–9.
18.
Lee EJ, Lee SK, Agid R, Bae JM, Keller A, Terbrugge K. Preoperative grading of presumptive low-grade astrocytomas on MR imaging: diagnostic value of minimum apparent diffusion coefficient. AJNR Am J Neuroradiol 2008;29:1872–1877.
19.
Liu Y, Bai R, Sun H, Liu H, Wang D. Diffusion-weighted magnetic resonance imaging of uterine cervical cancer. J Comput Assist Tomogr 2009;33:858–862.
20.
Yu X, Lin M, Ouyang H, Zhou C, Zhang H. Application of ADC measurement in characterization of renal cell carcinomas with different pathological types and grades by 3.0T diffusion-weighted MRI. Eur J Radiol 2012;81:3061–3066.
21.
Goyal A, Sharma R, Bhalla AS, et al. Diffusion-weighted MRI in renal cell carcinoma: a surrogate marker for predicting nuclear grade and histologicalsubtype. Acta Radiol 2012;53:349–358.
22.
Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999;9:53–60.
23.
Regimbeau JM, Abdalla EK, Vauthey JN, et al. Risk factors for early death due to recurrence after liver resection for hepatocellular carcinoma: results of a multicenter study. J Surg Oncol 2004;85:36–41.
REFERENCES 1.
Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The new World Health Organization classification of lung tumours. Eur Respir J 2001;18:1059–1068.
2.
Ohno Y, Hatabu H, Takenaka D, Adachi S, Kono M, Sugimura K. Solitary pulmonary nodules: potential role of dynamic MR imaging in management initial experience. Radiology 2002;224:503–511.
3.
Puderbach M, Hintze C, Ley S, et al. MR imaging of the chest: a practical approach at 1.5T. Eur J Radiol 2007;64:345–355.
600
Volume 42, No. 3
Liu et al.: ADC and Grade of Lung Carcinoma 24.
Guo Y, Cai YQ, Cai ZL, et al. Differentiation of clinically benign and malignant breast lesions using diffusion-weighted imaging. J Magn Reson Imaging 2002;16:172–178.
29.
Ichikawa T, Erturk SM, Motosugi U, et al. High-B-value diffusionweighted MRI in colorectal cancer. AJR Am J Roentgenol 2006;187: 181–184.
25.
Nakanishi M, Chuma M, Hige S, et al. Relationship between diffusionweighted magnetic resonance imaging and histological tumor grading of hepatocellular carcinoma. Ann Surg Oncol 2012;19:1302–1309.
30.
26.
Higano S, Yun X, Kumabe T, et al. Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis. Radiology 2006;241:839–846.
Park SH, Moon WK, Cho N, et al. Comparison of diffusion-weighted MR imaging and FDG PET/CT to predict pathological complete response to neoadjuvant chemotherapy in patients with breast cancer. Eur Radiol 2012;22:18–25.
31.
Calvar JA, Meli FJ, Romero C, et al. Characterization of brain tumors by MRS, DWI and Ki-67 labeling index. J Neurooncol 2005;72:273– 280.
Shin HJ, Baek HM, Ahn JH, et al. Prediction of pathologic response to neoadjuvant chemotherapy in patients with breast cancer using diffusion-weighted imaging and MRS. NMR Biomed 2012;25:1349–1359.
32.
Chen L, Zhang J, Chen Y, et al. Relationship between apparent diffusion coefficient and tumour cellularity in lung cancer. PLoS One 2014;9:e99865.
33.
Le Bihan D, Breton E, Lallemand D, et al. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168:497–505.
27.
28.
Henzler T, Schmid-Bindert G, Schoenberg SO, Fink C. Diffusion and perfusion MRI of the lung and mediastinum. Eur J Radiol 2010;76: 329–336.
September 2015
601
