Diagnosed chest lesion on diffusion-weighted magnetic resonance images using apparent diffusion coefficients

Abstract

PURPOSE:

A novel diagnostic method using the standard deviation (SD) value of apparent diffusion coefficient (ADC) by diffusion-weighted (DWI) magnetic resonance imaging (MRI) is applied for differential diagnosis of primary chest cancers, metastatic tumors and benign tumors.

MATERIALS AND METHODS:

This retrospective study enrolled 27 patients (20 males, 7 female; age, 15–85; mean age, 68) who had thoracic mass lesions in the last three years and underwent an MRI chest examination at our institution. In total, 29 mass lesions were analyzed using SD of ADC and DWI. Lesions were divided into five groups: Primary lung cancers (N = 10); esophageal cancers (N = 5); metastatic tumors (N = 8); benign tumors (N = 3); and inflammatory lesions (N = 3). Quantitative assessment of MRI parameters of mass lesions was performed. The ADC value was acquired based on the average of the entire tumor area. The error-plot, t-test and the area under receiver operating characteristic (AUC) were applied for statistical analysis.

RESULTS:

The SD of ADC value (mean±SD) was (4.867±1.359)×10^–4 mm²/sec in primary lung cancers, and (3.598±0.350)×10^–4 mm²/sec in metastatic tumors. The SD of ADC values of primary lung cancers and metastatic tumors (P < 0.05) were significantly different and the AUC was 0.800 (P < 0.05). The means of SD of ADC values was 4.532±1.406×10^–4 mm²/sec and 2.973±0.364×10^–4 mm²/sec for malignant tumors (including primary lung cancers, esophageal cancers) and benign tumors with respectively. The mean of SD of ADC values between malignant chest tumors and benign chest tumors was shown significant difference (P < 0.01). The values of AUC was 0.967 between malignant chest tumors and benign chest tumors (P < 0.05). The ADC values for primary lung cancers, metastatic tumors and benign tumors were not significantly difference (P > 0.05).

CONCLUSIONS:

The mean of SD of ADC value by DWI can be used for differential diagnosis of chest lesions.

Keywords

Diffusion-weighted imaging SD of ADC AUC

1 Introduction

Diffusion-weighted imaging (DWI) is a magnetic resonance imaging (MRI) technique that is sensitive to the Brownian molecular motion of spins. Molecular motion is related to the thermal kinetic energy of molecules, which is proportional to temperature of tissues. The DWI, which is based on the diffusivity of water molecules within tissues, can be utilized to analyze important characteristics of tissue. Meanwhile, the DWI could be used to detect of acute cerebral ischemia in the central nervous system. The DWI has been increasingly applied to image for mediastinum, pancreas, liver, and etc. Recently, the effectiveness of echo-planar imaging (EPI) for chest DWI is more suitable for chest applications due to innovate and improve in hardware and acquisition techniques. For example, reduction of eddy-current effects and less geometric distortions were generated by improving gradient systems [1 –5].

It is difficult that accurate pre-operative diagnosis of chest neoplasms and inflammatory nodules used computed tomography (CT), MRI and positron emission tomography-computed tomography (PET-CT). The most sensitive diagnostic method for chest neoplasms and inflammatory nodules is applied surgical biopsy. However, DWI provided an alternative method to diagnose chest neoplasms and inflammatory nodules. The advantages of DWI are noninvasive, none ionizing radiation, none administration of exogenous contrast medium and does, and comfortable procedure.

Directly interpretation of signal intensity on DWI images for nonsolid neoplasms and active inflammatory pulmonary nodules was difficulty. The main difficulty of interpretation on DWI for diagnostic pulmonary nodules is due to artifacts appeared on images. Moreover, the signal intensity of T2-weighted MRI images could not be used to observe and distinguish between malignant and benign pulmonary nodules [6, 7].

The apparent diffusion coefficient (ADC) of DWI has the ability to discriminate the malignant and benign pulmonary nodules [8 –10]. The ADC value was dependent different gradients to be helpful diagnostic between benign and malignant lesions [10]. On the other side, ADC value of lung carcinomas was hardly to evaluate and diagnose between histologic subtypes of lung carcinoma. The ADC values were not useful to diagnose pulmonary nodule due to inhomogeneous magnetic field under air-containing areas. However, ADC values may be the one of effective parameters for identifying the degree of lung adenocarcinoma [11, 12]. The lesion-to-spinal-cord ratio (LSR) is more effective than ADC for differentiating between lung cancer and benign lesions [13]. The diagnostic performance of DWI using the LSR was similar to that of 2-deoxy-2-(18F)fluoro-D-glucose (FDG) PET, which reportedly increases diagnostic accuracy when differentiating between benign and malignant lesions [14].

In this study, the standard deviation (SD) of ADC value of DWI images was applied for differential diagnosis of chest lesions, including primary chest cancers, metastatic tumors and benign tumors.

2 Materials and methods

2.1 Patients

This retrospective study enrolled patients with thoracic mass lesions within the last three years who underwent an MRI examination of the chest at E-DA Hospital. This retrospective experimental study was also approved by the internal review board of E-DA Hospital (Approval number: EMRP-099-111). In total, 27 patients with 29 mass lesions were enrolled (20 males, 7 females; age, 15–85; mean age, 68). The lesions were divided into five groups: Primary lung cancers (N = 10); esophageal cancers (N = 5); metastatic tumors (N = 8); benign tumors (N = 3); and inflammatory lesions (N = 3). Pathologic confirmation was based on surgical resection findings, endoscopic or percutaneous biopsy findings. To avoid hemorrhage-related distortions, each MRI was performed before the biopsy. Proven diagnosis of mass lesions depends on either histopathological data or clinical, radiological and laboratory data.

2.2 MR imaging

All MRI examinations were performed with a 1.5-T imager (Signa Excite HD; GE Healthcare, Waukesha, WI, USA) with a phased-array body coil (Eight-Channel Body Array Coil; GE Healthcare). Patients were imaged in the supine position. Prior to DWI, T1-and T2-weighted images were obtained for each patient. The images were obtained during quiet breathing with a motion-probing gradient (MPG) ofb values of 1000 and 0 sec/mm² in all of the x, y, and z directions. The following parameters were used: Spin-echo-based echo-planar imaging; 4100–5100/49.8; receiver bandwidth, 250 kHz; section thickness, 5mm (gapless); field of view, 40–48 cm; echo space, 396–420μsec; and real spatial resolution in the phase-encoding direction, 2.08–2.50 mm. Imaging time was 164–204 seconds for 40–50 sections. B values for all patients were fixed as 0 and 1000 in DWI MRI. The TR was fixed as 1000ms. The TE was used minimum full with range from 60 to 80ms [13].

2.3 Image analysis

The 29 mass lesions were analyzed using mean and SD of ADC values and DWI by manually drawn proper size of region of interest (ROI). Quantitative measurement of MRI parameters of mass lesions was performed by ROI base estimation. The size of ROI in ADC and DWI images for all patients are not the same due to the different size of tumor. Hence, the proper size of ROI is related to the size of tumor. The slice with the largest diameter of the chest lesion(s) was selected for measurement of ADC and DWI parameters in this study. The analysis of flow chart was shown in Fig. 1.

2.4 Statistical analysis

The mean and SD of ADC values and DWI were acquired based on the entire tumor area (i.e., ROI). The error-plot was adopted to check dispersion of SD values between ADC and DWI. The area under receiver operating characteristic (AUC) were used to identify the sensitivity and specificity of SD values between ADC and DWI. The t-test was applied to exam the significant difference between groups. The p < 0.05 was considered as statistically significance.

3 Results

The mean age of patients was 68±21 years (range, 15–85 years). Final diagnosis of 10 primary lung cancers, 5 esophageal cancers, 8 metastatic tumors and 3 benign tumors was confirmed by histopathological analysis based on percutaneous or endoscopic biopsy. Diagnosis was confirmed with laboratory, radiological and clinical data for 3 chronic inflammatory changes.

The SD of ADC values of primary lung cancers and metastatic tumors differed significantly (P < 0.05), and the AUC was 0.800 (P < 0.05) (Table 1). The SD of ADC value (mean±SD) was 4.532±1.406×10^–4 mm²/sec for all malignant chest tumors, including primary lung cancers and esophageal cancers (Table 2). The SD of ADC value was 2.973±0.364×10^–4 mm²/sec for benign tumors. A significant difference existed between SD of ADC values (P < 0.01) for malignant chest tumors and benign chest tumors, and the AUC was 0.967 (P < 0.05) (Table 2). No significant difference existed among ADC values for primary lung cancers, metastatic tumors and benign tumors.

The SD of ADC value was (4.867±1.359) ×10^–4 mm²/sec for primary lung cancers, and (3.598±0.350) ×10^–4 mm²/sec for metastatic tumors (Table 1); these values differed significantly (P < 0.05) (Fig. 2). Figures 3 and 4 present case illustrations of MRI findings and ADC values for primary lung cancers and metastatic tumors.

The SD of ADC value was significantly different from that for malignant chest tumors (P < 0.01). Meanwhile, the box plot of SD of ADC values for benign and malignant chest tumors was shown separable (Fig. 5). Figures 6 and 7 show the MRI findings and SD of ADC values for malignant tumors and benign tumors. The ADC values for primary lung cancers, metastatic tumors and benign tumors did not differ significantly.

4 Discussion

Characterizing pulmonary nodules remains a difficult problem for clinical diagnosis. The probability of malignancy increases with a nodule’s size. That is, only 20% of nodules≥20 mm are benign, and the prevalence of malignancy for nodules≥20 mm is 64–82% [15]. To avoid unnecessary surgical resection of benign nodules, obtaining an image feature as precise as possible is important. Although CT, MRI and PET-CT are typically used for differential diagnosis of benign and malignant tumors because of their high specificity and accuracy, they have many limitations. Recent innovations in hardware and acquisition techniques have markedly improved the diagnosis of chest lesions. The DWI technique has been reported as an in vivo biomarker of tumor grade and for differentiating oncologic lesions in other organs because more aggressive lesions are more hypercellular than well differentiated lesions [16].

Quantitative DWI has also been used for tissue characterization of lung cancers. Matoba et al. [11] reported that the mean ADC value of adenocarcinoma was significantly higher than that of squamous cell carcinoma or large cell carcinoma. Previous investigators have reported that no significant difference existed between ADC values for lung cancers and benign lesions [13]. Thoracic MRI is problematic because of the low signal-to-noise ratio (SNR) of the inherently low lung proton density, cardiac and respiratory motion, and magnetic susceptibility effects of air-filled lung tissue subjected to large magnetic field gradients. Whether the ADC value can be used as a reliable indicator for differentiating malignancy from other anomalies, such as those in other organs, is therefore uncertain. This report is consistent with findings acquired by our study. The DWI technique does not generate a simple map of diffusion constants. Generally, signal intensity on DWI images is significantly affected by the T2 value of the lesions. Since the signal is generated by a spin-echo pulse sequence with a long repetition and echo time, overall signal intensity on DWI MRI images reflects not only diffusion conspicuity, but also the spin density and T2 value of image voxels [17]. Most malignant tumors, including lung cancers, are characterized by elongated T2 values; therefore, it works favorably on lesion conspicuity of malignant tumors via DWI. The ADC value should be used as the standard. If the imaging purpose is to visualize malignant tumors, deriving the ADC value is not essential.

The SD of ADC value of primary lung cancers was significantly higher than that of metastatic tumors (Fig. 1). The SD of ADC value can be a useful method for differential diagnosis of primary lung cancers and metastatic tumors [18 –21]. The primary lung cancers were more heterogeneous than metastatic tumors. Figure 2 exhibits MRI and DWI findings of the primary lung cancer. A 62-year-old male presented with a malignant tumor in his left lower lung; the SD of ADC value was 5.03×10^–4 mm²/sec. The axial T1-weighted MRI image shows the mass adjacent to the pericardium. The mass reveals heterogeneous enhancement on axial T1-weighted MRI image with fat suppression post gadolinium administration. Figure 3 shows MRI and DWI findings of metastatic tumor. A 47-year-old male presented with metastatic meningioma in the right lower lung; the SD of ADC value was 3.62×10^–4 mm²/sec. The axialT1-weighted MRI image shows the mass adjacent to the paraesophageal region. The axial T1-weighted MR image with fat suppression post gadolinium administration reveals homogeneous enhancement of the tumor. El-Badrawy et al. [22] used DWI to differentiate between different malignant chest wall tumors. The P-value for the malignant group in their study was <0.012. Differences in ADC values may reflect differences in histopathological features: Metastases and non-Hodgkin lymphomas (NHL) generally had enlarged cells and tumor cellularity was relatively high, so its ADC values were typically lower in their study.

The SD of ADC value (P < 0.01) of malignant chest tumors was significantly higher than that of benign chest tumors (Fig. 4). Figure 5 shows the MRI and DWI findings of the benign tumor. A 34-year-old male presented with schwannoma in the right paraspinal region of the upper mediastinum; the SD of ADC value was 2.72×10^–4 mm²/sec. On axial T1-weighted image, the mass shows iso-signal intensity similar to that of muscle in the right upper mediastinum. The axial T1-weighted MR image with fat suppression post gadolinium administration reveals homogeneous enhancement of the tumor. Figure 6 displays MRI and DWI findings of malignant tumor. A 70-year-old female presented with breast cancer metastasis involving the sternum; the SD of ADC value was 3.37×10^–4 mm²/sec. The axial T1-weighted MRI image shows the metastatic tumor involving the sternum of the anterior chest wall. The axialT1-weighted MRI image with fat suppression post gadolinium administration reveals minor heterogeneous enhancement of the metastatic tumor. Most malignant lesions are hyperintense relative to chest wall muscles or the spinal cord on DWI images and hypointense on ADC maps. Quantitative DWI analysis depends on ADC measurement. A circular or elliptical region of interest (ROI) is drawn on the section where the target lesion was detected on the T2 weighted on contrast-enhanced images. In heterogeneous tumors, care should be taken to locate the ROI in the solid portion of the mass, excluding cystic or necrotic areas. Türkbey B. used DWI to characterize lung lesions [23].

The range of SD of ADC values was large due to the small number of cases of inflammatory lesions. We speculate that different degrees of inflammation are related to different degrees of diffusion of interstitial restricted water. Inflammatory cells, such as white blood cells, or necrotic tissues may accumulate densely in lesions. Consequently, diffusion may be restricted in high-viscosity lesions; however, this has not yet been seen in the lung. The first intraalveolar stage in the process of developing pneumonia is characterized by formation of fibrinoid inflammatory cell clusters [24]. Therefore, no significant difference existed among primary lung cancers, metastatic tumors and inflammatory lesions by mean ADC value or SD of ADC value. The ADC values among primary lung cancers, metastatic tumors and benign tumors are similar because the mean of intensities among ROIs were closed to each other’s. However, the standard deviations of intensities among ROIs were significant difference in this study.

This study had some limitations. First, the patient population especially that reflected in the benign tumor subgroup was relatively small, possibly compromising the accuracy of analytical results. Second, the use of DWI for the chest was hindered by certain limitations such as physiologic motion artifacts (respiration and cardiac motion), low SNR of the low lung proton density and the susceptibility artifacts caused by air-tissue interfaces [25]. Additionally, DWI images were acquired using a breath-hold echo-planar sequence with Sensitivity-encoded (SENSE), rendering measurements vulnerable to susceptibility effects. In measuring the ROI of ADC, manual discrimination was used, which may reduce the accuracy of the ADC measurement.

5 Summary

Although technically complex, DWI of the chest is technically feasible with recently developed software and hardware. Analytical results demonstrate that the SD of ADC value of primary lung cancers was significantly higher than those of metastatic tumors and benign tumors. Subsequent analyses include analyses of 3D volume data of DWI and further study to confirm findings using a larger patient population. To differentiate primary lung cancers from metastatic tumors and benign tumors, the DWI SD of ADC value is effective.

Footnotes

Acknowledgments

The authors would like to thank the Ministry of Science and Technology of the Republic of China, Taiwan, for partially financially supporting this research under Contract No. MOST103-2320-B-214-009. Ted Knoy is appreciated for his editorial assistance.

References

Kauczor

H.U.

and Ley

, Thoracic magnetic resonance imaging 1985 to 2010, J Thorac Imaging 25(1) (2010), 34–38.

Dietrich

, Diffusion-weighted imaging and diffusion tensor imaging, In: Reiser

M.F.

, Semmler

, Hricak

, editors. Magnetic resonance tomography, Berlin: Springer, 2008, pp. 130–152.

Gümüştaş

, Inan

, Akansel

, Ciftçi

, Demirci

and Ozkara

S.K.

, Differentiation of malignant and benign lung lesions with diffusion-weighted MR imaging, Radiol Oncol 46(2) (2012), 106–113.

Luna

, Sánchez-Gonzalez

and Caro

, Diffusion-weighted imaging of the chest, Magn Reson Imaging Clin N Am 19(1) (2011), 69–94.

Poser

B.A.

, Barth

, Goa

P.E.

, Deng

and Stenger

V.A.

, Single-shot echo-planar imaging with Nyquist ghost compensation: Interleaved dual echo with acceleration (IDEA) echo-planar imaging (EPI), Magn Reson Med 69(1) (2013), 37–47.

Satoh

, Kitazume

, Ohdama

, Kimula

, Taura

and Endo

, Can malignant and benign pulmonary nodules be differentiated with diffusion-weighted MRI?, AJR Am J Roentgenol 191(2) (2008), 464–470.

Sommer

, Tremper

, Koenigkam-Santos

, Delorme

, Becker

, Biederer

, Kauczor

H.U.

, Heussel

C.P.

, Schlemmer

H.P.

and Puderbach

, Lung nodule detection in a high-risk population: Comparison of magnetic resonance imaging and low-dose computed tomography, Eur J Radiol 83(3) (2014), 600–605.

Bozgeyik

, Onur

M.R.

and Poyraz

A.K.

, The role of diffusion weighted magnetic resonance imaging in oncologic settings, Quant Imaging Med Surg 3(5) (2013), 269–278.

Vermoolen

M.A.

, Kwee

T.C.

and Nievelstein

R.A.

, Apparent diffusion coefficient measurements in the differentiation between benign and malignant lesions: A systematic review, Insights Imaging 3(4) (2012), 395–409.

10.

Onur

M.R.

, Çiçekçi

, Kayali

, Poyraz

A.K.

and Kocakoç

, The role of ADC measurement in differential diagnosis of focal hepatic lesions, Eur J Radiol 81(3) (2012), e171–e176.

11.

Matoba

, Tonami

, Kondou

, Yokota

, Higashi

, Toga

and Sakuma

, Lung carcinoma: Diffusion-weighted mr imaging–preliminary evaluation with apparent diffusion coefficient, Radiology 243(2) (2007), 570–577.

12.

Lee

H.Y.

, Jeong

J.Y.

, Lee

K.S.

, Yi

C.A.

, Kim

B.T.

, Kang

, Kwon

O.J.

, Shim

Y.M.

and Han

, Histopathology of lung adenocarcinoma based on new IASLC/ATS/ERS classification: Prognostic stratification with functional and metabolic imaging biomarkers, J Magn Reson Imaging 38(4) (2013), 905–913.

13.

Uto

, Takehara

, Nakamura

, Naito

, Hashimoto

, Inui

, Suda

, Nakamura

and Chida

, Higher sensitivity and specificity for diffusion-weighted imaging of malignant lung lesions without apparent diffusion coefficient quantification, Radiology 252(1) (2009), 247–254.

14.

Gong

, Cao

, Zhang

, Deng

, Kang

, Zhu

, Liu

and Zhou

, Diagnostic efficacy of whole-body diffusion-weighted imaging in the detection of tumour recurrence and metastasis by comparison with 18F-2-fluoro-2-deoxy-D-glucose positron emission tomography or computed tomography in patients with gastrointestinal cancer, Gastroenterol Rep (Oxf) 3(2) (2015), 128–135.

15.

Wahidi

M.M.

, Govert

J.A.

, Goudar

R.K.

, et al., Evidence for the treatment of patients with pulmonary nodules: When is it lung cancer? ACCP evidence-based clinical practice guidelines (2nd edition), Chest 132(Suppl 3) (2007), 94Se107S.

16.

Padhani

A.R.

, Liu

, Koh

D.M.

, et al., Diffusion-weighted magnetic resonance imaging as a cancer biomarker: Consensus and recommendations, Neoplasia 11(2) (2009), 102e25.

17.

Stejskal

E.O.

and Tanner

J.E.

, Spin diffusion measurements: Spin-echoes in the presence of a time-dependent field gradient, J Chem Phys 42 (1965), 288–292.

18.

Herneth

A.M.

, Guccione

and Bednarski

, Apparent diffusion coefficient: A quantitative parameter for in vivo tumor characterization, Eur J Radiol 45 (2003), 208–213.

19.

Kim

, Murakami

, Takahashi

, Hori

, Tsuda

and Nakamura

, Diffusion-weighted single-shot echo planar MR imaging for liver disease, AJR Am J Roentgenol 173 (1999), 393–398.

20.

Lee

S.S.

, Byun

J.H.

, Park

B.J.

, Park

S.H.

, Kim

, Park

, Kim

J.K.

and Lee

M.G.

, Quantitative analysis of diffusion-weighted magnetic resonance imaging of the pancreas: Usefulness in characterizing solid pancreatic masses, J Magn Reson Imaging 28 (2008), 928–936.

21.

Fujii

, Matsusue

, Kigawa

, Sato

, Kanasaki

, Nakanishi

, Sugihara

, Kaminou

, Terakawa

and Ogawa

, Diagnostic accuracy of the apparent diffusion coefficient in differentiating benign from malignant uterine endometrial cavity lesions: Initial results, Eur Radiol 18 (2008), 384–389.

22.

Adel

, Maha

, Tamer

F.Y.

and Mohammed

K.E.

, Role of diffusion-weighted MR imaging in chest wall masses, The Egyptian Journal of Radiology and Nuclear Medicine 42(2) (2011), 147–51.

23.

Türkbey

, Aras

Ö.

, Karabulut

, Turgut

A.T.

, Akpinar

, Alibek

, Pang

, Ertürk

Ş.M.

, El Khouli

R.H.

, Bluemke

D.A.

and Choyke

P.L.

, Diffusion-weighted MRI for detecting and monitoring cancer: A review of current applications in body imaging, Diagn Interv Radiol 18(1) (2012), 46–59.

24.

Nazer

, Alnajjar

, Salah

, Khzouz

, Alfaqeer

and Qandeel

, Fatal case of cryptogenic organizing pneumonia associated with Everolimus, Ann Saudi Med 34(5) (2014), 437–439.

25.

Mürtz

, Flacke

, Träber

, van den Brink

J.S.

, Gieseke

and Schild

H.H.

, Abdomen: Diffusion weighted MR imaging with pulse-triggered single shot sequences, Radiology 224(1) (2002), 258–264.