Original Research n Thoracic
Imaging
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Malignant Pleural Mesothelioma: Visual Assessment by Using Pleural Pointillism at Diffusion-weighted MR Imaging1 Johan Coolen, MD Frederik De Keyzer, MSc Philippe Nafteux, MD Walter De Wever, MD, PhD Christophe Dooms, MD, PhD Johan Vansteenkiste, MD, PhD Aurélie Derweduwen, MD Ilse Roebben, MSc Eric Verbeken, MD, PhD Paul De Leyn, MD, PhD Dirk Van Raemdonck, MD, PhD Kristiaan Nackaerts, MD, PhD Steven Dymarkowski, MD, PhD Johny Verschakelen, MD, PhD
1
From the Departments of Radiology (J.C., F.D.K., W.D.W., I.R., S.D., J. Verschakelen), Thoracic Surgery (P.N., P.D.L., D.V.R.), Pneumology (C.D., J. Vansteenkiste, A.D., K.N.), and Pathology (E.V.), University Hospitals Leuven, Herestraat 49, B-3000 Leuven, Belgium. Received September 16, 2013; revision requested November 20; final revision received May 16, 2014; accepted June 26; final version accepted July 18. Address correspondence to J.C. (e-mail: Johan. [email protected]).
Purpose:
To evaluate the diagnostic accuracy of the visual assessment of malignant pleural mesothelioma (MPM) on magnetic resonance (MR) images by using two known visual markers (mediastinal pleural thickness and shrinking of the lung) and a newly introduced one (pleural pointillism).
Materials and Methods:
With the approval of the local ethics committee, 100 consecutive patients (mean age, 61.4 years; age range, 18–87 years; 75 men, 25 women) suspected of having MPM pleural abnormalities underwent positron emission tomography/computed tomography and MR imaging, including diffusion-weighted (DW) MR imaging, followed by explorative thoracoscopy or guided biopsy with histopathologic confirmation. Because visual assessment is still the preferred method of image interpretation, the diagnostic accuracy of mediastinal pleural thickening, shrinking lung (hemithorax volume decrease due to fibrosis), and pleural pointillism were examined. Pleural pointillism was denoted by the presence of multiple, hyperintense pleural spots on high-b-value DW images. Histopathologic findings in the surgical specimen served as the reference standard. McNemar tests with Bonferroni correction were used to assess differences in accuracy among the three examined markers.
Results:
Of 100 patients, 33 had benign pleural alterations, and 67 had malignant pleural diseases (MPDs); 57 of 67 had MPM. A total of 78 patients received a correct diagnosis (benign vs malignant) on the basis of mediastinal pleural thickening (sensitivity, 81%; specificity, 73%; accuracy, 78%); and 66 patients, on the basis of shrinking lung (sensitivity, 60%; specificity, 79%; accuracy, 66%). The correct diagnosis was indicated on the basis of pleural pointillism in 88 patients (sensitivity, 93%; specificity, 79%; accuracy, 88%).
Conclusion:
Visual assessment of pleural pointillism on high-b-value DW images is useful to differentiate MPD from benign alterations, performing substantially better than mediastinal pleural thickness and shrinking lung, and might obviate unnecessary invasive procedures for MPM. q RSNA, 2014
RSNA, 2014
q
576
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
M
alignant pleural mesothelioma (MPM) is an uncommon tumor of the pleura that can present a diagnostic challenge. Currently, the definitive diagnosis of MPM always is based on the results obtained from an adequate biopsy sample in the context of appropriate clinical, radiologic, and surgical findings (1). Computed tomography (CT) has been used widely as the primary imaging modality for evaluation of MPM (2–6). Although there are no pathognomonic features, circumferential pleural thickening with or without pleural effusion, nodular pleural thickening, involvement of the interlobar fissures, mediastinal pleural involvement, contraction of the pleura, and mediastinal lymph node invasion (6,7) are considered strongly suggestive of malignant pleural disease (MPD). Whereas the positive predictive value (PPV) of these features is high, their absence does not exclude diagnosis of MPM. Sensitivity varies between 38% and 70%; and specificity, between 64% and 100% (2). At present, integrated fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT
Advances in Knowledge nn Visual evaluation of pleural pointillism, consisting of multiple, hyperintense pleural spots on high-b-value diffusion-weighted (DW) MR images, enables differentiation between benign and malignant pleural lesions with a sensitivity of 93% (62 of 67) and a specificity of 79% (26 of 33). nn Malignant pleural disease is likely to be present if the thickened unilateral pleura manifests as pleural pointillism on high-bvalue DW MR images (positive predictive value, 90% [62 of 69]). nn The diagnostic accuracy of pleural pointillism (88% [88 of 100]) is substantially better than that of mediastinal pleural thickness (78% [78 of 100], P = .23) and of shrinking lung (66% [66 of 100], P , .001).
is considered the standard technique for staging MPM (8,9) because it combines metabolic and anatomic information (9). Several authors (4,10–13) evaluated pleural disease that was suspected of being MPM by using FDG PET with visual analysis and reached a sensitivity of 92%, with a specificity of 75% and an accuracy of 89%. However, FDG PET is not completely tumor specific, and tracer uptake can be seen in benign inflammatory lesions or talc pleurodesis as well, occasionally making accurate staging of MPM difficult (14). Although magnetic resonance (MR) imaging is not used widely for these applications, it has advantages over FDG PET/CT because of the excellent softtissue contrast and the absence of ionizing radiation (15). Moreover, functional MR imaging techniques such as diffusion-weighted (DW) MR imaging and dynamic contrast material–enhanced MR imaging already have proved their worth in the differentiation of malignant from benign pleural disease and in the assessment of chest wall and diaphragmatic involvement (16–20). Especially relatively late in disease progression, when pleural lesions are overt, differentiation between normal tissue and malignant thoracic cage invasion becomes crucial to distinguish between patients who are most likely to benefit from an aggressive multimodality regimen and those who would benefit from systemic treatment before performance status decreases. In these situations, CT tends to lead to underestimation of early chest wall invasion (21). In this prospective study, we aimed to evaluate the diagnostic accuracy of the visual assessment of MPM at MR imaging by using two known visual markers (mediastinal pleural thickness and shrinking of
Implication for Patient Care nn Visual assessment of pleural pointillism could allow accurate diagnosis of malignant pleural mesothelioma and could help in selection of an appropriate biopsy location.
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
Coolen et al
the lung) and a newly introduced one (pleural pointillism)
Materials and Methods Subjects This prospective study was approved by the local ethics committee of University Hospitals Leuven (Leuven, Belgium), and written informed consent was obtained from all patients. Between November 2009 and December 2012, we included 109 consecutive patients who had pleural disease and were suspected of having MPM; all were scheduled for histopathologic examination within the first weeks after CT and MR imaging. Nine patients were excluded for the following reasons: three patients refused to sign the informed consent form, two patients had claustrophobia, two patients had a pacemaker, one patient wore a nerve stimulator, and one patient was unable to be placed in the dorsal decubitus position because of pain. The remaining 100 patients (mean age, 61.4 years; range, 18–87 years; 25 women [mean age, 58.4 years; range, 32–79 years] and
Published online before print 10.1148/radiol.14132111 Content codes: Radiology 2015; 274:576–584 Abbreviations: DW = diffusion weighted FDG = fluorine 18 fluorodeoxyglucose H-E = hematoxylin-eosin MPD = malignant pleural disease MPM = malignant pleural mesothelioma NPV = negative predictive value PPV = positive predictive value Author contributions: Guarantors of integrity of entire study, J.C., C.D., E.V., K.N., J. Verschakelen; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, J.C., W.D.W., E.V., J. Verschakelen; clinical studies, J.C., F.D.K., P.N., W.D.W., J. Vansteenkiste, A.D., I.R., E.V., P.D.L., D.V.R., K.N., S.D.; statistical analysis, J.C., F.D.K., P.N., W.D.W.; and manuscript editing, J.C., F.D.K., P.N., W.D.W., C.D., J. Vansteenkiste, E.V., P.D.L., K.N., S.D., J. Verschakelen Conflicts of interest are listed at the end of this article.
577
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 1
Figure 1: Data obtained in a 69-year-old man with low back pain. A, B, CT images used to guide biopsy of a tumor-involved right diaphragmatic crus (arrow on A); biopsy results indicated an epithelioid MPM (not shown), but pleuroscopic examination results could not be used to confirm this diagnosis initially. E, F, I, J, Additional PET/CT images show an FDG-avid lesion in the right diaphragmatic crus (arrow on I, J), and the pleural activity (arrow on E, F) on the right was interpreted as postoperative pleuritis. C, D, DW images show the diaphragmatic lesion (arrowhead on D) and two hyperintense pleural spots (arrows on C). These anatomic lesions (arrows on G, arrowhead on H) were confirmed at thoracotomy. All images were obtained in the axial plane. K, L, Histopathologic specimen of the pleural lesion shows a tumoral pleural node at hematoxylin-eosin (H-E) staining (K) and is positive at additional immunohistochemical calretinin staining (L). (Original magnification, 312.5.)
75 men [mean age, 62.4 years; range, 18–87 years]) were scheduled for MR imaging within 24 hours before the invasive diagnostic procedure. Of these 100 patients, 69 (69%) were assigned to a therapy planning program for pleural disease, including additional PET/ CT within 24 hours before the invasive diagnostic procedure, which was thoracotomy for 39 (39%) (Fig 1) and videoassisted thoracic surgery with biopsy for 30 (30%). In the other 31 patients (31%), pleural tissue sampling occurred during pleuroscopy (24 patients) (Fig 2) or with transthoracic core biopsy puncture with CT guidance (seven patients) (Fig 3). Diagnostic CT was performed an average of 7 days (range, 0–14 days) before MR imaging. 578
Thoracic Imaging CT images were obtained by using multidetector CT scanners (Somatom Sensation 64 or 16 or Volume Zoom; Siemens Healthcare, Erlangen, Germany). The examinations were performed at inspiration after intravenous administration of 1.5 mL of iobitridol (Xenetix 350; Guerbet, Roissy, France) per kilogram of body weight at the rate of 2.5 mL/sec with a power injector (Dual Shot Alpha; Nemoto Kyorindo, Tokyo, Japan), extending from the lung apices to the costodiaphragmatic sulci. Technical parameters of the scanning protocol included 120 kVp, 120–250 mAs (automatic dose modulation), pitch of 1.2 mm, and collimation of 0.75–1.5
mm, from which 3-mm-thick axial and coronal images (in-plane resolution, 0.7 3 0.7 mm) were reconstructed. PET/CT was performed on an integrated PET/CT scanner (Biograph TruePoint V; Siemens Healthcare) including a 40-section CT scanner. Patients received an intravenous injection of FDG, which was produced in-house, and then rested for 60 minutes before imaging. The activity was based on weight according to the following formula: A = (W · 4) + 20, where A is activity in megabecquerels and W is weight in kilograms. The standard interval between injection and scanning was 75 minutes. Contrast-enhanced CT and emission PET acquisitions were obtained in the
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 2
Figure 2: Data obtained in a 76-year-old man with dysphagia with solid intake. A, Axial chest CT image shows tumoral encasement of the esophagus and a moderate amount of bilateral pleural effusion. This finding is confirmed with, B, a T2-weighted fat-saturated axial MR image, whereas, C, an axial DW image with a b value of 1000 sec/mm2 demonstrates the restrictive lesion around the distal part of the esophagus, as well as hyperintense spots (arrows) on the right pleura that are suggestive of MPD. C, D, E, F, Increasing the b value in the axial DW images allows the hyperintense spots (arrows on C, E, F), probably reflecting tumoral tissue, to become more visible. G, Pleuroscopic image illustrates the tapiocalike pleural lesions. H, H-E staining shows invasive proliferation of polygonal cells with large eosinophilic cytoplasm, suggestive of malignant pleural tumor. I, Epithelial membrane antigen staining shows a diffuse, strong positivity. (Original magnification, 3200.)
identical transverse plane, and both independent and coregistered images were available for readout, including CT-based attenuation-corrected images and nonattenuation-corrected images. Neither respiratory gating nor breath-hold modes were used. MR images were acquired with a 3-T whole-body system (Achieva; Philips Healthcare, Best, the Netherlands) with the manufacturer’s 16-channel phased-array torso coil (Sense XL Torso; Philips Healthcare) for signal reception. The MR imaging protocol consisted of nonenhanced
T2-weighted single-shot turbo spinecho (repetition time msec/echo time msec, 828/70; field of view, 302 3 375 mm; matrix, 187 3 288; and fat suppression by means of spectral selection attenuated inversion recovery, or SPAIR) and DW imaging sequences. DW imaging was performed by using a spin-echo echo-planar imaging sequence (6481/60; number of transverse sections, 38; field of view, 420 3 323 mm; section thickness, 5 mm; and matrix, 104 3 80, yielding a voxel resolution of 4.0 3 4.0 3 5.0 mm; intersection gap, 0.7 mm; echo-planar
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
imaging factor, 43; sensitivity encoding parallel imaging factor, two; short tau inversion recovery fat suppression with an inversion time of 260 msec; and number of signals acquired, two) for an imaging time of 12 minutes 19 seconds during free breathing. Diffusion sensitization was performed with b values of 0, 50, 100, 500, 750, and 1000 sec/mm2.
Image Interpretation CT, PET/CT, and MR images were evaluated separately, with observers blinded to all information regarding 579
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 3
Figure 3: Data obtained in a 52-year-old man with recurrent pain in the left side of his chest after left extrapleural pneumonectomy for MPM. A, B, Axial PET/CT images obtained after extrapleural pneumonectomy indicate difficult differentiation of the FDG-avid left thoracic cage lesion (recurrence vs postsurgical status). After axial MR imaging (D), including DW imaging, transthoracic CT-guided biopsy was performed; positioning is shown on a fused DW image/CT image (E). Histopathologic findings were used to confirm epithelioid MPM, with positive H-E staining (C) and prekeratin staining (F); at original magnification (312.5), a pointillistic growth pattern is patent (obvious).
image results of the other modality and histopathologic results. CT images were assessed by observer 1 (W.D.W., a radiologist with 28, 12, and 3 years of experience in thorax CT, PET/CT, and MR imaging, respectively) for the presence or absence of two radiologic signs. First, maximal thickness of mediastinal pleura (in millimeters) was measured, avoiding possible adenopathies; this maximal thickness then was categorized as either absent (thickness 1 mm) or present (thickness . 1 mm). Second, the involved hemithorax volume loss due to fibrotic changes was evaluated. This shrinkage was assessed visually in comparison with the nonaffected hemithorax on a binary scale (present or absent), with visual assessment of lung volume and interpretation of the intercostal spaces of both hemithoraces as evaluating elements. The other pleura signs, such as irregularity and nodularity of the pleural contour, 580
were not interpreted because absence of these signs does not allow the clinician to rule out disease (2). Observer 2 (J.C., a radiologist with 32, 6, and 12 years of experience in thorax CT, PET/CT, and MR imaging, respectively) assessed the MR images visually by examining pleural hyperintense areas on the images acquired by using a b value of 1000 sec/ mm2. When multiple (minimum of two) (Fig 1) hyperintense areas visible by using a b value of 1000 were present in the pleura but were not identifiable as such with lower b values (Fig 2), it was described as pleural pointillism, and the presence or absence of this visual marker was evaluated; these results were later correlated with the histopathologic analysis results. The fat-suppressed T2-weighted MR images were used to correlate the locations of the lesions on the DW images. After 3 months, each chest radiologist assessed the images from
the other modality so we could assess interobserver variability.
Histopathologic Analysis Three thoracic surgeons (P.N., P.D.L., and D.V.R., each with more than 12 years of experience) performed all surgical procedures. The resected tissues were fixed in a 10% formalin solution, embedded in paraffin, and stained with H-E, complemented by immunohistochemical staining (1). One pathologist (E.V., with 27 years of experience) examined all sections with light microscopy (magnification range of 312 to 3400). Using the updated classification of the International Mesothelioma Interest Group of the 2004 World Health Organization, he coded the histopathologic results (22). The distribution of the three examined diagnostic markers (mediastinal thickening, shrinking lung, and pleural pointillism) according to the histopathologic diagnosis was evaluated.
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 4
Figure 4: Data obtained in a 70-year-old woman with shrinkage of the right hemithorax after left-sided mastectomy 10 years previously. A, B, Axial PET/CT images show circumferential pleural thickening on the right side with lung volume loss. This finding was confirmed at axial MR imaging (D), and clear hyperintense points (arrows on E) were visible at DW imaging. These were confirmed histologically as MPM by means of H-E staining (C) and cytokeratin 5.6 staining (F). (Original magnification, 312.)
Statistical Analysis Statistical analysis was performed by using a software package (Analyse-it, version 2.12; Analyse-it, Leeds, England). Sensitivity, specificity, PPV, negative predictive value (NPV), and accuracy were calculated for each of the examined parameters (mediastinal pleural thickness, lung shrinkage, and pleural pointillism) in correlation with histopathologic findings; 95% confidence intervals obtained by using the Wilson score method are provided. The diagnostic accuracies of the different parameters examined were compared by using two-tailed McNemar tests, with Bonferroni correction for multiple testing. Interobserver agreement was evaluated by using the k test, with P values calculated according to the Fleiss method. A P value less than .05 was considered to indicate a significant difference. Results All CT and MR examinations were completed successfully, and histopathologic
verification was available for each patient. Of this cohort, 67 (67%) of 100 received a diagnosis of MPD: 57 patients had mesothelioma, of which 46 were epithelioid, six were sarcomatoid, and three were biphasic, and there were two patients with desmoplastic pleural disease (1,11,23). Ten patients had metastatic diseases—for example, adenocarcinoma, squamous lung carcinoma, invasive thymoma, and epithelioid malignant peripheral nerve sheath tumor. The remaining 33 (33%) of 100 patients had benign tumors or alterations and inflammatory disorders, such as a solitary fibrous tumor, pleural plaques, talc pleurodesis, or pleuritis. All patients had some form of pleural thickening. Examined CT features suggestive of malignancy were the thickness of the mediastinal pleural involvement, as well as circumferential pleural thickening leading to apparent shrinking of the lung (Fig 4). When we examined the mediastinal pleural thickness, we found a range of
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
0–50 mm. Subdividing the mediastinal pleural thickness into two groups, namely absent (thickness 1 mm) or present (thickness . 1 mm) pleural thickening, led to 78 patients with a correct diagnosis (sensitivity of 81% [54 of 67], specificity of 73% [24 of 33], PPV of 86% [54 of 63], NPV of 65% [24 of 37], and accuracy of 78% [78 of 100]). The presence of circumferential pleural thickening leading to shrinking lung correctly helped in the prediction of pleural disease in 66 patients (sensitivity of 60% [40 of 67], specificity of 79% [26 of 33], PPV of 85% [40 of 47], NPV of 49% [26 of 53], and accuracy of 66% [66 of 100]). However, using the occurrence of pleural pointillism as a diagnostic marker resulted in 88 correct diagnoses (sensitivity of 93% [62 of 67], specificity of 79% [26 of 33], PPV of 90% [62 of 69], NPV of 84% [26 of 31], and accuracy of 88% [88 of 100]). The diagnostic accuracies and confidence intervals are summarized in Table 1, and the distribution of the histopathologic diagnosis versus the three examined diagnostic markers is given in Table 2. 581
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Table 1 Diagnostic Accuracies and 95% Confidence Intervals of the Three Examined Visual Signs Mediastinal Thickening Measure Sensitivity Specificity PPV NPV Accuracy
Shrinking Lung
Pleural Pointillism
Percentage*
95% Confidence Interval (%)
Percentage*
95% Confidence Interval (%)
Percentage*
95% Confidence Interval (%)
80.6 (54/67) 72.7 (24/33) 85.7 (54/63) 64.9 (24/37) 78.0 (78/100)
69.6, 88.3 55.8, 84.9 75.0, 92.3 48.8, 78.2 68.9, 85.0
59.7 (40/67) 78.8 (26/33) 85.1 (40/47) 49.1 (26/53) 66.0 (66/100)
47.7, 70.6 62.2, 89.3 72.3, 92.6 36.1, 62.1 56.3, 74.5
92.5 (62/67) 78.8 (26/33) 89.9 (62/69) 83.9 (26/31) 88.0 (88/100)
83.7, 96.8 62.2, 89.3 80.5, 95.0 67.4, 92.9 80.2, 93.0
* Numbers in parentheses were used to calculate the percentages; percentages were rounded.
Table 2 Distribution of Histopathologic Diagnosis versus the Presence of the Three Examined Visual Signs Diagnosis MPM Metastasis Benign Inflammation
Mediastinal Thickening
Shrinking Lung
Pleural Pointillism
82.5 (47/57) 70.0 (7/10) 25.0 (5/20) 30.8 (4/13)
66.7 (38/57) 20.0 (2/10) 20.0 (4/20) 23.1 (3/13)
96.5 (55/57) 70.0 (7/10) 25.0 (5/20) 15.4 (2/13)
Note.—Data are percentages; numbers in parentheses were used to calculate the percentages, which were rounded.
Although McNemar test results showed a significant difference after Bonferroni correction between the accuracy of lung shrinkage compared with the presence of pleural pointillism (P , .001), the differences between mediastinal thickening and lung shrinkage (P = .13) and between mediastinal thickening and pleural pointillism (P = .23) did not reach significance. The k values for interobserver agreement between both readers were 0.71, 0.48, and 0.53 for mediastinal thickening, lung shrinkage, and pleural pointillism, respectively, all P , .001 according to the Fleiss method.
Discussion In this study, we compared visual assessments of different imaging parameters that could be used to characterize primary MPD in a cohort of patients who were suspected of having MPM. We found a good accuracy of 88% for pleural pointillism and lower accuracies for mediastinal pleural thickness (accuracy, 78%) and 582
apparent lung shrinkage due to circumferential fibrotic changes (accuracy, 66%). We used the term pleural pointillism to describe the presence of multiple hyperintense spots at high-bvalue DW imaging because it is visually reminiscent of the Postimpressionistic painting technique known as Pointillism (24). In the last decade, chest CT has become the standard imaging method for diagnosis of MPD. The diagnosis is favored by the presence of parietal pleural thickening greater than 1 mm, circumferential pleural thickening, nodular pleural thickening, and mediastinal pleural thickening. In a study by Leung and coworkers (2), sensitivities of these findings were 47%, 54%, 38%, and 70%, and specificities of these findings were 94%, 100%, 96%, and 88%, respectively. The visual assessment of the mediastinal pleural thickening in our study has a slightly higher sensitivity but a lower specificity (81% and 73%, respectively). In a study by Luo et al (17), overall sensitivity of 90% and overall specificity
of 60% for pleural malignancy were achieved by using CT parameters. Leung et al (2) and Scott et al (25) described contraction of the affected hemithorax (ie, reduced size of the hemithorax involved) with narrowed intercostal spaces and elevation of the ipsilateral hemidiaphragm, possibly associated with ipsilateral mediastinal shift, as a frequently occurring phenomenon in patients with MPM. Kawashima and Libshitz (7) also described this volume loss of the involved hemithorax as a CT feature of MPM, which occurred in 42% of their cohort. Sahin et al (6) reported volume contraction in 73%, and Hierholzer et al (16) used the term shrinking lung for this phenomenon. However, it also is observed in patients with systemic sclerosis, rounded atelectasis, systemic lupus erythematosus, and primary Sjögren syndrome (7), as well as in patients with chronic empyema and hemothorax. In the patient population in our study, shrinking lung showed sensitivity and specificity of 60% and 79%, respectively, and a PPV of 85%. There are also other signs hinting at the diagnosis of MPM, such as direct extension of the lesion into vascular structures and mediastinal organs including the heart, esophagus, trachea, and aorta. PET/CT has become routine clinical practice for assessing stage (detecting N3 and M1 disease) and tumor progression in mesothelioma (26). Gill et al (11) found sensitivity, specificity, and accuracy of FDG imaging for diagnosis of MPM of 97%, 80%, and 94%,
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
respectively, compared with 83%, 80%, and 82% for diagnostic CT. They mentioned false-positive findings only in patients treated with talc pleurodesis. Other research groups (12,13) reported benign inflammatory pleuritis, benign asbestos-related plaques, parapneumonic effusions, uremic pleuritis, and tuberculous pleuritis as causes for false-positive findings. False-negative findings due to mild, or lack of, tracer uptake were reported in patients with mesothelioma of the epithelial subtype and slow-growing fibrous tumors such as low-grade lymphoma or prostate metastasis (17,27). Currently, the use of MR imaging to detect MPD is increasing because of the high soft-tissue contrast, intrinsic flow sensitivity, and absence of ionizing radiation (15). With MR imaging, the combination of morphologic and functional imaging sequences has been reported to have high sensitivity and specificity in detecting pleural malignancy (19). By using a combined approach of DW imaging and dynamic contrast-enhanced MR imaging, Coolen et al (19) reached a sensitivity of 71%, a specificity of 94%, and an accuracy of 93% for diagnosis of MPD. The fast and easy visual assessment of pleural pointillism in the current study had an only slightly lower accuracy of 88%. A definitive diagnosis of MPM normally requires histologic sampling with H-E staining and often additional histochemical stains. In addition, a correlation between multimodality imaging and surgical investigations usually is required. Results from a 2005 report showed that extensive surgery combined with chemotherapy and radiation therapy (trimodality treatment) can prolong survival in selected patients with early-stage disease (4). This finding emphasizes the need for accurate staging procedures at diagnosis. In that respect, pleural pointillism has a number of advantages with clinical relevance. First, pointillism already occurs early in the stages of MPD development, probably representing small clusters of malignancy that, because of partial volume and movement effects in
the background of high-mobility pleural fluid, do not yet lead to the low apparent diffusion coefficient values indicative of malignancy. Second, because pleural pointillism is a straightforward and easy visual assessment of a single set of images without any required calculations, the diagnostic throughput is high. Third, the pleural pointillism locations can be mapped easily and used as an aid to perform biopsy sampling of the most appropriate areas, which is important because biopsy sampling can lead to tumor seeding along the track (reported in up to 22% of MPM patients) and postprocedural pneumothorax rates as high as 9.5% (3). Finally, pleural pointillism can help to define the extent of malignancy better without exhausting biopsy sampling of all possibly malignant regions. Limitations of this study are the relatively small number of patients included because of the relative rarity of the condition. Also, because only patients who were suspected of having MPM were included, the results might be biased toward this group composition. We expect, therefore, that pleural pointillism will not be as specific as reported in this article when used in a more general patient population. However, we still would advocate that the presence of pointillism should alert the investigator because the importance of pleural pointillism might be greatest when clinical symptoms are still absent or nonspecific. Also, although interobserver agreement was acceptable, a relatively low k value was found for the DW imaging pointillism assessment, which was probably due to a lower level of experience in visual DW image interpretations of the radiologist performing the second DW image reading. However, the reproducibility was still similar to that of the evaluations of the other examined parameters in this study. In summary, visual assessment of pleural pointillism on high-b-value DW images seems to be a reasonably accurate tool for diagnosing MPD. Pleural pointillism might be used to provide guidance for biopsy or thoracoscopic evaluation and eventually could help obviate invasive procedures, thereby
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
Coolen et al
minimizing patient discomfort. Further larger studies are necessary to validate the clinical applicability and specificity of pleural pointillism. Disclosures of Conflicts of Interest: J.C. disclosed no relevant relationships. F.D.K. disclosed no relevant relationships. P.N. disclosed no relevant relationships. W.D.W. disclosed no relevant relationships. C.D. disclosed no relevant relationships. J. Vansteenkiste disclosed no relevant relationships. A.D. disclosed no relevant relationships. I.R. disclosed no relevant relationships. E.V. disclosed no relevant relationships. P.D.L. disclosed no relevant relationships. D.V.R. disclosed no relevant relationships. K.N. disclosed no relevant relationships. S.D. disclosed no relevant relationships. J. Verschakelen disclosed no relevant relationships.
References 1. Husain AN, Colby TV, Ordóñez NG, et al. Guidelines for pathologic diagnosis of malignant mesothelioma: a consensus statement from the International Mesothelioma Interest Group. Arch Pathol Lab Med 2009;133(8):1317–1331. 2. Leung AN, Müller NL, Miller RR. CT in differential diagnosis of diffuse pleural disease. AJR Am J Roentgenol 1990;154(3):487–492. 3. Wang ZJ, Reddy GP, Gotway MB, et al. Malignant pleural mesothelioma: evaluation with CT, MR imaging, and PET. RadioGraphics 2004;24(1):105–119. 4. Benamore RE, O’Doherty MJ, Entwisle JJ. Use of imaging in the management of malignant pleural mesothelioma. Clin Radiol 2005;60(12):1237–1247. 5. Traill ZC, Davies RJ, Gleeson FV. Thoracic computed tomography in patients with suspected malignant pleural effusions. Clin Radiol 2001;56(3):193–196. 6. Sahin AA, Cöplü L, Selçuk ZT, et al. Malignant pleural mesothelioma caused by environmental exposure to asbestos or erionite in rural Turkey: CT findings in 84 patients. AJR Am J Roentgenol 1993;161(3):533–537. 7. Kawashima A, Libshitz HI. Malignant pleural mesothelioma: CT manifestations in 50 cases. AJR Am J Roentgenol 1990;155(5):965–969. 8. Beyer T, Townsend DW, Brun T, et al. A combined PET/CT scanner for clinical oncology. J Nucl Med 2000;41(8):1369–1379. 9. Erasmus JJ, Truong MT, Smythe WR, et al. Integrated computed tomography-positron emission tomography in patients with potentially resectable malignant pleural mesothelioma: staging implications. J Thorac Cardiovasc Surg 2005;129(6):1364–1370.
583
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
10. Bénard F, Sterman D, Smith RJ, Kaiser LR, Albelda SM, Alavi A. Metabolic imaging of malignant pleural mesothelioma with fluorodeoxyglucose positron emission tomography. Chest 1998;114(3):713–722.
17. Luo L, Hierholzer J, Bittner RC, Chen J, Huang L. Magnetic resonance imaging in distinguishing malignant from benign pleural disease. Chin Med J (Engl) 2001;114(6):645–649.
11. Gill RR, Gerbaudo VH, Jacobson FL, et al. MR imaging of benign and malignant pleural disease. Magn Reson Imaging Clin N Am 2008;16(2):319–339, x.
18. Luna A, Sánchez-Gonzalez J, Caro P. Diffusionweighted imaging of the chest. Magn Reson Imaging Clin N Am 2011;19(1):69–94.
12. Duysinx B, Nguyen D, Louis R, et al. Evaluation of pleural disease with 18-fluorodeoxyglucose positron emission tomography imaging. Chest 2004;125(2):489–493. 13. Kramer H, Pieterman RM, Slebos DJ, et al. PET for the evaluation of pleural thickening observed on CT. J Nucl Med 2004;45(6): 995–998. 14. Jaruskova M, Belohlavek O. Role of FDG PET and PET/CT in the diagnosis of prolonged febrile states. Eur J Nucl Med Mol Imaging 2006;33(8):913–918. 15. Fujimoto K. Usefulness of contrast-enhanced magnetic resonance imaging for evaluating solitary pulmonary nodules. Cancer Imaging 2008;8:36–44. 16. Hierholzer J, Luo L, Bittner RC, et al. MRI and CT in the differential diagnosis of pleural disease. Chest 2000;118(3):604–609.
584
19. Coolen J, De Keyzer F, Nafteux P, et al. Malignant pleural disease: diagnosis by using diffusion-weighted and dynamic contrastenhanced MR imaging—initial experience. Radiology 2012;263(3):884–892. 20. Gill RR, Umeoka S, Mamata H, et al. Diffusion-weighted MRI of malignant pleural mesothelioma: preliminary assessment of apparent diffusion coefficient in histologic subtypes. AJR Am J Roentgenol 2010;195(2):W125–W130. 21. Rusch VW, Piantadosi S, Holmes EC. The role of extrapleural pneumonectomy in malignant pleural mesothelioma: a Lung Cancer Study Group trial. J Thorac Cardiovasc Surg 1991;102(1):1–9.
cer staging manual. 7th ed. New York, NY: Springer, 2007; 271–277. 23. Attanoos RL, Gibbs AR. Morphologic variants and their mimics. In: Attanoos RL, Allen TC, eds. Advances in surgical pathology: mesothelioma. Philadelphia, Pa: Lippincott Williams & Wilkins, 2014; 67–90. 24. Sanna A. Pointillism. In: Impressionism. Florence, Italy: Scala Group, 2011; 441– 496. 25. Scott IR, Müller NL, Miller RR, Evans KG, Nelems B. Resectable stage III lung cancer: CT, surgical, and pathologic correlation. Radiology 1988;166(1 pt 1):75–79. 26. Yi CA, Jeon TY, Lee KS, et al. 3-T MRI: usefulness for evaluating primary lung cancer and small nodules in lobes not containing primary tumors. AJR Am J Roentgenol 2007;189(2):386–392. 27. Pass HI. Malignant pleural mesothelioma: surgical roles and novel therapies. Clin Lung Cancer 2001;3(2):102–117.
22. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. Pleural mesothelioma. In: Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC can-
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
Imaging
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Malignant Pleural Mesothelioma: Visual Assessment by Using Pleural Pointillism at Diffusion-weighted MR Imaging1 Johan Coolen, MD Frederik De Keyzer, MSc Philippe Nafteux, MD Walter De Wever, MD, PhD Christophe Dooms, MD, PhD Johan Vansteenkiste, MD, PhD Aurélie Derweduwen, MD Ilse Roebben, MSc Eric Verbeken, MD, PhD Paul De Leyn, MD, PhD Dirk Van Raemdonck, MD, PhD Kristiaan Nackaerts, MD, PhD Steven Dymarkowski, MD, PhD Johny Verschakelen, MD, PhD
1
From the Departments of Radiology (J.C., F.D.K., W.D.W., I.R., S.D., J. Verschakelen), Thoracic Surgery (P.N., P.D.L., D.V.R.), Pneumology (C.D., J. Vansteenkiste, A.D., K.N.), and Pathology (E.V.), University Hospitals Leuven, Herestraat 49, B-3000 Leuven, Belgium. Received September 16, 2013; revision requested November 20; final revision received May 16, 2014; accepted June 26; final version accepted July 18. Address correspondence to J.C. (e-mail: Johan. [email protected]).
Purpose:
To evaluate the diagnostic accuracy of the visual assessment of malignant pleural mesothelioma (MPM) on magnetic resonance (MR) images by using two known visual markers (mediastinal pleural thickness and shrinking of the lung) and a newly introduced one (pleural pointillism).
Materials and Methods:
With the approval of the local ethics committee, 100 consecutive patients (mean age, 61.4 years; age range, 18–87 years; 75 men, 25 women) suspected of having MPM pleural abnormalities underwent positron emission tomography/computed tomography and MR imaging, including diffusion-weighted (DW) MR imaging, followed by explorative thoracoscopy or guided biopsy with histopathologic confirmation. Because visual assessment is still the preferred method of image interpretation, the diagnostic accuracy of mediastinal pleural thickening, shrinking lung (hemithorax volume decrease due to fibrosis), and pleural pointillism were examined. Pleural pointillism was denoted by the presence of multiple, hyperintense pleural spots on high-b-value DW images. Histopathologic findings in the surgical specimen served as the reference standard. McNemar tests with Bonferroni correction were used to assess differences in accuracy among the three examined markers.
Results:
Of 100 patients, 33 had benign pleural alterations, and 67 had malignant pleural diseases (MPDs); 57 of 67 had MPM. A total of 78 patients received a correct diagnosis (benign vs malignant) on the basis of mediastinal pleural thickening (sensitivity, 81%; specificity, 73%; accuracy, 78%); and 66 patients, on the basis of shrinking lung (sensitivity, 60%; specificity, 79%; accuracy, 66%). The correct diagnosis was indicated on the basis of pleural pointillism in 88 patients (sensitivity, 93%; specificity, 79%; accuracy, 88%).
Conclusion:
Visual assessment of pleural pointillism on high-b-value DW images is useful to differentiate MPD from benign alterations, performing substantially better than mediastinal pleural thickness and shrinking lung, and might obviate unnecessary invasive procedures for MPM. q RSNA, 2014
RSNA, 2014
q
576
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
M
alignant pleural mesothelioma (MPM) is an uncommon tumor of the pleura that can present a diagnostic challenge. Currently, the definitive diagnosis of MPM always is based on the results obtained from an adequate biopsy sample in the context of appropriate clinical, radiologic, and surgical findings (1). Computed tomography (CT) has been used widely as the primary imaging modality for evaluation of MPM (2–6). Although there are no pathognomonic features, circumferential pleural thickening with or without pleural effusion, nodular pleural thickening, involvement of the interlobar fissures, mediastinal pleural involvement, contraction of the pleura, and mediastinal lymph node invasion (6,7) are considered strongly suggestive of malignant pleural disease (MPD). Whereas the positive predictive value (PPV) of these features is high, their absence does not exclude diagnosis of MPM. Sensitivity varies between 38% and 70%; and specificity, between 64% and 100% (2). At present, integrated fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT
Advances in Knowledge nn Visual evaluation of pleural pointillism, consisting of multiple, hyperintense pleural spots on high-b-value diffusion-weighted (DW) MR images, enables differentiation between benign and malignant pleural lesions with a sensitivity of 93% (62 of 67) and a specificity of 79% (26 of 33). nn Malignant pleural disease is likely to be present if the thickened unilateral pleura manifests as pleural pointillism on high-bvalue DW MR images (positive predictive value, 90% [62 of 69]). nn The diagnostic accuracy of pleural pointillism (88% [88 of 100]) is substantially better than that of mediastinal pleural thickness (78% [78 of 100], P = .23) and of shrinking lung (66% [66 of 100], P , .001).
is considered the standard technique for staging MPM (8,9) because it combines metabolic and anatomic information (9). Several authors (4,10–13) evaluated pleural disease that was suspected of being MPM by using FDG PET with visual analysis and reached a sensitivity of 92%, with a specificity of 75% and an accuracy of 89%. However, FDG PET is not completely tumor specific, and tracer uptake can be seen in benign inflammatory lesions or talc pleurodesis as well, occasionally making accurate staging of MPM difficult (14). Although magnetic resonance (MR) imaging is not used widely for these applications, it has advantages over FDG PET/CT because of the excellent softtissue contrast and the absence of ionizing radiation (15). Moreover, functional MR imaging techniques such as diffusion-weighted (DW) MR imaging and dynamic contrast material–enhanced MR imaging already have proved their worth in the differentiation of malignant from benign pleural disease and in the assessment of chest wall and diaphragmatic involvement (16–20). Especially relatively late in disease progression, when pleural lesions are overt, differentiation between normal tissue and malignant thoracic cage invasion becomes crucial to distinguish between patients who are most likely to benefit from an aggressive multimodality regimen and those who would benefit from systemic treatment before performance status decreases. In these situations, CT tends to lead to underestimation of early chest wall invasion (21). In this prospective study, we aimed to evaluate the diagnostic accuracy of the visual assessment of MPM at MR imaging by using two known visual markers (mediastinal pleural thickness and shrinking of
Implication for Patient Care nn Visual assessment of pleural pointillism could allow accurate diagnosis of malignant pleural mesothelioma and could help in selection of an appropriate biopsy location.
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
Coolen et al
the lung) and a newly introduced one (pleural pointillism)
Materials and Methods Subjects This prospective study was approved by the local ethics committee of University Hospitals Leuven (Leuven, Belgium), and written informed consent was obtained from all patients. Between November 2009 and December 2012, we included 109 consecutive patients who had pleural disease and were suspected of having MPM; all were scheduled for histopathologic examination within the first weeks after CT and MR imaging. Nine patients were excluded for the following reasons: three patients refused to sign the informed consent form, two patients had claustrophobia, two patients had a pacemaker, one patient wore a nerve stimulator, and one patient was unable to be placed in the dorsal decubitus position because of pain. The remaining 100 patients (mean age, 61.4 years; range, 18–87 years; 25 women [mean age, 58.4 years; range, 32–79 years] and
Published online before print 10.1148/radiol.14132111 Content codes: Radiology 2015; 274:576–584 Abbreviations: DW = diffusion weighted FDG = fluorine 18 fluorodeoxyglucose H-E = hematoxylin-eosin MPD = malignant pleural disease MPM = malignant pleural mesothelioma NPV = negative predictive value PPV = positive predictive value Author contributions: Guarantors of integrity of entire study, J.C., C.D., E.V., K.N., J. Verschakelen; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, J.C., W.D.W., E.V., J. Verschakelen; clinical studies, J.C., F.D.K., P.N., W.D.W., J. Vansteenkiste, A.D., I.R., E.V., P.D.L., D.V.R., K.N., S.D.; statistical analysis, J.C., F.D.K., P.N., W.D.W.; and manuscript editing, J.C., F.D.K., P.N., W.D.W., C.D., J. Vansteenkiste, E.V., P.D.L., K.N., S.D., J. Verschakelen Conflicts of interest are listed at the end of this article.
577
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 1
Figure 1: Data obtained in a 69-year-old man with low back pain. A, B, CT images used to guide biopsy of a tumor-involved right diaphragmatic crus (arrow on A); biopsy results indicated an epithelioid MPM (not shown), but pleuroscopic examination results could not be used to confirm this diagnosis initially. E, F, I, J, Additional PET/CT images show an FDG-avid lesion in the right diaphragmatic crus (arrow on I, J), and the pleural activity (arrow on E, F) on the right was interpreted as postoperative pleuritis. C, D, DW images show the diaphragmatic lesion (arrowhead on D) and two hyperintense pleural spots (arrows on C). These anatomic lesions (arrows on G, arrowhead on H) were confirmed at thoracotomy. All images were obtained in the axial plane. K, L, Histopathologic specimen of the pleural lesion shows a tumoral pleural node at hematoxylin-eosin (H-E) staining (K) and is positive at additional immunohistochemical calretinin staining (L). (Original magnification, 312.5.)
75 men [mean age, 62.4 years; range, 18–87 years]) were scheduled for MR imaging within 24 hours before the invasive diagnostic procedure. Of these 100 patients, 69 (69%) were assigned to a therapy planning program for pleural disease, including additional PET/ CT within 24 hours before the invasive diagnostic procedure, which was thoracotomy for 39 (39%) (Fig 1) and videoassisted thoracic surgery with biopsy for 30 (30%). In the other 31 patients (31%), pleural tissue sampling occurred during pleuroscopy (24 patients) (Fig 2) or with transthoracic core biopsy puncture with CT guidance (seven patients) (Fig 3). Diagnostic CT was performed an average of 7 days (range, 0–14 days) before MR imaging. 578
Thoracic Imaging CT images were obtained by using multidetector CT scanners (Somatom Sensation 64 or 16 or Volume Zoom; Siemens Healthcare, Erlangen, Germany). The examinations were performed at inspiration after intravenous administration of 1.5 mL of iobitridol (Xenetix 350; Guerbet, Roissy, France) per kilogram of body weight at the rate of 2.5 mL/sec with a power injector (Dual Shot Alpha; Nemoto Kyorindo, Tokyo, Japan), extending from the lung apices to the costodiaphragmatic sulci. Technical parameters of the scanning protocol included 120 kVp, 120–250 mAs (automatic dose modulation), pitch of 1.2 mm, and collimation of 0.75–1.5
mm, from which 3-mm-thick axial and coronal images (in-plane resolution, 0.7 3 0.7 mm) were reconstructed. PET/CT was performed on an integrated PET/CT scanner (Biograph TruePoint V; Siemens Healthcare) including a 40-section CT scanner. Patients received an intravenous injection of FDG, which was produced in-house, and then rested for 60 minutes before imaging. The activity was based on weight according to the following formula: A = (W · 4) + 20, where A is activity in megabecquerels and W is weight in kilograms. The standard interval between injection and scanning was 75 minutes. Contrast-enhanced CT and emission PET acquisitions were obtained in the
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 2
Figure 2: Data obtained in a 76-year-old man with dysphagia with solid intake. A, Axial chest CT image shows tumoral encasement of the esophagus and a moderate amount of bilateral pleural effusion. This finding is confirmed with, B, a T2-weighted fat-saturated axial MR image, whereas, C, an axial DW image with a b value of 1000 sec/mm2 demonstrates the restrictive lesion around the distal part of the esophagus, as well as hyperintense spots (arrows) on the right pleura that are suggestive of MPD. C, D, E, F, Increasing the b value in the axial DW images allows the hyperintense spots (arrows on C, E, F), probably reflecting tumoral tissue, to become more visible. G, Pleuroscopic image illustrates the tapiocalike pleural lesions. H, H-E staining shows invasive proliferation of polygonal cells with large eosinophilic cytoplasm, suggestive of malignant pleural tumor. I, Epithelial membrane antigen staining shows a diffuse, strong positivity. (Original magnification, 3200.)
identical transverse plane, and both independent and coregistered images were available for readout, including CT-based attenuation-corrected images and nonattenuation-corrected images. Neither respiratory gating nor breath-hold modes were used. MR images were acquired with a 3-T whole-body system (Achieva; Philips Healthcare, Best, the Netherlands) with the manufacturer’s 16-channel phased-array torso coil (Sense XL Torso; Philips Healthcare) for signal reception. The MR imaging protocol consisted of nonenhanced
T2-weighted single-shot turbo spinecho (repetition time msec/echo time msec, 828/70; field of view, 302 3 375 mm; matrix, 187 3 288; and fat suppression by means of spectral selection attenuated inversion recovery, or SPAIR) and DW imaging sequences. DW imaging was performed by using a spin-echo echo-planar imaging sequence (6481/60; number of transverse sections, 38; field of view, 420 3 323 mm; section thickness, 5 mm; and matrix, 104 3 80, yielding a voxel resolution of 4.0 3 4.0 3 5.0 mm; intersection gap, 0.7 mm; echo-planar
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
imaging factor, 43; sensitivity encoding parallel imaging factor, two; short tau inversion recovery fat suppression with an inversion time of 260 msec; and number of signals acquired, two) for an imaging time of 12 minutes 19 seconds during free breathing. Diffusion sensitization was performed with b values of 0, 50, 100, 500, 750, and 1000 sec/mm2.
Image Interpretation CT, PET/CT, and MR images were evaluated separately, with observers blinded to all information regarding 579
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 3
Figure 3: Data obtained in a 52-year-old man with recurrent pain in the left side of his chest after left extrapleural pneumonectomy for MPM. A, B, Axial PET/CT images obtained after extrapleural pneumonectomy indicate difficult differentiation of the FDG-avid left thoracic cage lesion (recurrence vs postsurgical status). After axial MR imaging (D), including DW imaging, transthoracic CT-guided biopsy was performed; positioning is shown on a fused DW image/CT image (E). Histopathologic findings were used to confirm epithelioid MPM, with positive H-E staining (C) and prekeratin staining (F); at original magnification (312.5), a pointillistic growth pattern is patent (obvious).
image results of the other modality and histopathologic results. CT images were assessed by observer 1 (W.D.W., a radiologist with 28, 12, and 3 years of experience in thorax CT, PET/CT, and MR imaging, respectively) for the presence or absence of two radiologic signs. First, maximal thickness of mediastinal pleura (in millimeters) was measured, avoiding possible adenopathies; this maximal thickness then was categorized as either absent (thickness 1 mm) or present (thickness . 1 mm). Second, the involved hemithorax volume loss due to fibrotic changes was evaluated. This shrinkage was assessed visually in comparison with the nonaffected hemithorax on a binary scale (present or absent), with visual assessment of lung volume and interpretation of the intercostal spaces of both hemithoraces as evaluating elements. The other pleura signs, such as irregularity and nodularity of the pleural contour, 580
were not interpreted because absence of these signs does not allow the clinician to rule out disease (2). Observer 2 (J.C., a radiologist with 32, 6, and 12 years of experience in thorax CT, PET/CT, and MR imaging, respectively) assessed the MR images visually by examining pleural hyperintense areas on the images acquired by using a b value of 1000 sec/ mm2. When multiple (minimum of two) (Fig 1) hyperintense areas visible by using a b value of 1000 were present in the pleura but were not identifiable as such with lower b values (Fig 2), it was described as pleural pointillism, and the presence or absence of this visual marker was evaluated; these results were later correlated with the histopathologic analysis results. The fat-suppressed T2-weighted MR images were used to correlate the locations of the lesions on the DW images. After 3 months, each chest radiologist assessed the images from
the other modality so we could assess interobserver variability.
Histopathologic Analysis Three thoracic surgeons (P.N., P.D.L., and D.V.R., each with more than 12 years of experience) performed all surgical procedures. The resected tissues were fixed in a 10% formalin solution, embedded in paraffin, and stained with H-E, complemented by immunohistochemical staining (1). One pathologist (E.V., with 27 years of experience) examined all sections with light microscopy (magnification range of 312 to 3400). Using the updated classification of the International Mesothelioma Interest Group of the 2004 World Health Organization, he coded the histopathologic results (22). The distribution of the three examined diagnostic markers (mediastinal thickening, shrinking lung, and pleural pointillism) according to the histopathologic diagnosis was evaluated.
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Figure 4
Figure 4: Data obtained in a 70-year-old woman with shrinkage of the right hemithorax after left-sided mastectomy 10 years previously. A, B, Axial PET/CT images show circumferential pleural thickening on the right side with lung volume loss. This finding was confirmed at axial MR imaging (D), and clear hyperintense points (arrows on E) were visible at DW imaging. These were confirmed histologically as MPM by means of H-E staining (C) and cytokeratin 5.6 staining (F). (Original magnification, 312.)
Statistical Analysis Statistical analysis was performed by using a software package (Analyse-it, version 2.12; Analyse-it, Leeds, England). Sensitivity, specificity, PPV, negative predictive value (NPV), and accuracy were calculated for each of the examined parameters (mediastinal pleural thickness, lung shrinkage, and pleural pointillism) in correlation with histopathologic findings; 95% confidence intervals obtained by using the Wilson score method are provided. The diagnostic accuracies of the different parameters examined were compared by using two-tailed McNemar tests, with Bonferroni correction for multiple testing. Interobserver agreement was evaluated by using the k test, with P values calculated according to the Fleiss method. A P value less than .05 was considered to indicate a significant difference. Results All CT and MR examinations were completed successfully, and histopathologic
verification was available for each patient. Of this cohort, 67 (67%) of 100 received a diagnosis of MPD: 57 patients had mesothelioma, of which 46 were epithelioid, six were sarcomatoid, and three were biphasic, and there were two patients with desmoplastic pleural disease (1,11,23). Ten patients had metastatic diseases—for example, adenocarcinoma, squamous lung carcinoma, invasive thymoma, and epithelioid malignant peripheral nerve sheath tumor. The remaining 33 (33%) of 100 patients had benign tumors or alterations and inflammatory disorders, such as a solitary fibrous tumor, pleural plaques, talc pleurodesis, or pleuritis. All patients had some form of pleural thickening. Examined CT features suggestive of malignancy were the thickness of the mediastinal pleural involvement, as well as circumferential pleural thickening leading to apparent shrinking of the lung (Fig 4). When we examined the mediastinal pleural thickness, we found a range of
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
0–50 mm. Subdividing the mediastinal pleural thickness into two groups, namely absent (thickness 1 mm) or present (thickness . 1 mm) pleural thickening, led to 78 patients with a correct diagnosis (sensitivity of 81% [54 of 67], specificity of 73% [24 of 33], PPV of 86% [54 of 63], NPV of 65% [24 of 37], and accuracy of 78% [78 of 100]). The presence of circumferential pleural thickening leading to shrinking lung correctly helped in the prediction of pleural disease in 66 patients (sensitivity of 60% [40 of 67], specificity of 79% [26 of 33], PPV of 85% [40 of 47], NPV of 49% [26 of 53], and accuracy of 66% [66 of 100]). However, using the occurrence of pleural pointillism as a diagnostic marker resulted in 88 correct diagnoses (sensitivity of 93% [62 of 67], specificity of 79% [26 of 33], PPV of 90% [62 of 69], NPV of 84% [26 of 31], and accuracy of 88% [88 of 100]). The diagnostic accuracies and confidence intervals are summarized in Table 1, and the distribution of the histopathologic diagnosis versus the three examined diagnostic markers is given in Table 2. 581
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
Table 1 Diagnostic Accuracies and 95% Confidence Intervals of the Three Examined Visual Signs Mediastinal Thickening Measure Sensitivity Specificity PPV NPV Accuracy
Shrinking Lung
Pleural Pointillism
Percentage*
95% Confidence Interval (%)
Percentage*
95% Confidence Interval (%)
Percentage*
95% Confidence Interval (%)
80.6 (54/67) 72.7 (24/33) 85.7 (54/63) 64.9 (24/37) 78.0 (78/100)
69.6, 88.3 55.8, 84.9 75.0, 92.3 48.8, 78.2 68.9, 85.0
59.7 (40/67) 78.8 (26/33) 85.1 (40/47) 49.1 (26/53) 66.0 (66/100)
47.7, 70.6 62.2, 89.3 72.3, 92.6 36.1, 62.1 56.3, 74.5
92.5 (62/67) 78.8 (26/33) 89.9 (62/69) 83.9 (26/31) 88.0 (88/100)
83.7, 96.8 62.2, 89.3 80.5, 95.0 67.4, 92.9 80.2, 93.0
* Numbers in parentheses were used to calculate the percentages; percentages were rounded.
Table 2 Distribution of Histopathologic Diagnosis versus the Presence of the Three Examined Visual Signs Diagnosis MPM Metastasis Benign Inflammation
Mediastinal Thickening
Shrinking Lung
Pleural Pointillism
82.5 (47/57) 70.0 (7/10) 25.0 (5/20) 30.8 (4/13)
66.7 (38/57) 20.0 (2/10) 20.0 (4/20) 23.1 (3/13)
96.5 (55/57) 70.0 (7/10) 25.0 (5/20) 15.4 (2/13)
Note.—Data are percentages; numbers in parentheses were used to calculate the percentages, which were rounded.
Although McNemar test results showed a significant difference after Bonferroni correction between the accuracy of lung shrinkage compared with the presence of pleural pointillism (P , .001), the differences between mediastinal thickening and lung shrinkage (P = .13) and between mediastinal thickening and pleural pointillism (P = .23) did not reach significance. The k values for interobserver agreement between both readers were 0.71, 0.48, and 0.53 for mediastinal thickening, lung shrinkage, and pleural pointillism, respectively, all P , .001 according to the Fleiss method.
Discussion In this study, we compared visual assessments of different imaging parameters that could be used to characterize primary MPD in a cohort of patients who were suspected of having MPM. We found a good accuracy of 88% for pleural pointillism and lower accuracies for mediastinal pleural thickness (accuracy, 78%) and 582
apparent lung shrinkage due to circumferential fibrotic changes (accuracy, 66%). We used the term pleural pointillism to describe the presence of multiple hyperintense spots at high-bvalue DW imaging because it is visually reminiscent of the Postimpressionistic painting technique known as Pointillism (24). In the last decade, chest CT has become the standard imaging method for diagnosis of MPD. The diagnosis is favored by the presence of parietal pleural thickening greater than 1 mm, circumferential pleural thickening, nodular pleural thickening, and mediastinal pleural thickening. In a study by Leung and coworkers (2), sensitivities of these findings were 47%, 54%, 38%, and 70%, and specificities of these findings were 94%, 100%, 96%, and 88%, respectively. The visual assessment of the mediastinal pleural thickening in our study has a slightly higher sensitivity but a lower specificity (81% and 73%, respectively). In a study by Luo et al (17), overall sensitivity of 90% and overall specificity
of 60% for pleural malignancy were achieved by using CT parameters. Leung et al (2) and Scott et al (25) described contraction of the affected hemithorax (ie, reduced size of the hemithorax involved) with narrowed intercostal spaces and elevation of the ipsilateral hemidiaphragm, possibly associated with ipsilateral mediastinal shift, as a frequently occurring phenomenon in patients with MPM. Kawashima and Libshitz (7) also described this volume loss of the involved hemithorax as a CT feature of MPM, which occurred in 42% of their cohort. Sahin et al (6) reported volume contraction in 73%, and Hierholzer et al (16) used the term shrinking lung for this phenomenon. However, it also is observed in patients with systemic sclerosis, rounded atelectasis, systemic lupus erythematosus, and primary Sjögren syndrome (7), as well as in patients with chronic empyema and hemothorax. In the patient population in our study, shrinking lung showed sensitivity and specificity of 60% and 79%, respectively, and a PPV of 85%. There are also other signs hinting at the diagnosis of MPM, such as direct extension of the lesion into vascular structures and mediastinal organs including the heart, esophagus, trachea, and aorta. PET/CT has become routine clinical practice for assessing stage (detecting N3 and M1 disease) and tumor progression in mesothelioma (26). Gill et al (11) found sensitivity, specificity, and accuracy of FDG imaging for diagnosis of MPM of 97%, 80%, and 94%,
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
respectively, compared with 83%, 80%, and 82% for diagnostic CT. They mentioned false-positive findings only in patients treated with talc pleurodesis. Other research groups (12,13) reported benign inflammatory pleuritis, benign asbestos-related plaques, parapneumonic effusions, uremic pleuritis, and tuberculous pleuritis as causes for false-positive findings. False-negative findings due to mild, or lack of, tracer uptake were reported in patients with mesothelioma of the epithelial subtype and slow-growing fibrous tumors such as low-grade lymphoma or prostate metastasis (17,27). Currently, the use of MR imaging to detect MPD is increasing because of the high soft-tissue contrast, intrinsic flow sensitivity, and absence of ionizing radiation (15). With MR imaging, the combination of morphologic and functional imaging sequences has been reported to have high sensitivity and specificity in detecting pleural malignancy (19). By using a combined approach of DW imaging and dynamic contrast-enhanced MR imaging, Coolen et al (19) reached a sensitivity of 71%, a specificity of 94%, and an accuracy of 93% for diagnosis of MPD. The fast and easy visual assessment of pleural pointillism in the current study had an only slightly lower accuracy of 88%. A definitive diagnosis of MPM normally requires histologic sampling with H-E staining and often additional histochemical stains. In addition, a correlation between multimodality imaging and surgical investigations usually is required. Results from a 2005 report showed that extensive surgery combined with chemotherapy and radiation therapy (trimodality treatment) can prolong survival in selected patients with early-stage disease (4). This finding emphasizes the need for accurate staging procedures at diagnosis. In that respect, pleural pointillism has a number of advantages with clinical relevance. First, pointillism already occurs early in the stages of MPD development, probably representing small clusters of malignancy that, because of partial volume and movement effects in
the background of high-mobility pleural fluid, do not yet lead to the low apparent diffusion coefficient values indicative of malignancy. Second, because pleural pointillism is a straightforward and easy visual assessment of a single set of images without any required calculations, the diagnostic throughput is high. Third, the pleural pointillism locations can be mapped easily and used as an aid to perform biopsy sampling of the most appropriate areas, which is important because biopsy sampling can lead to tumor seeding along the track (reported in up to 22% of MPM patients) and postprocedural pneumothorax rates as high as 9.5% (3). Finally, pleural pointillism can help to define the extent of malignancy better without exhausting biopsy sampling of all possibly malignant regions. Limitations of this study are the relatively small number of patients included because of the relative rarity of the condition. Also, because only patients who were suspected of having MPM were included, the results might be biased toward this group composition. We expect, therefore, that pleural pointillism will not be as specific as reported in this article when used in a more general patient population. However, we still would advocate that the presence of pointillism should alert the investigator because the importance of pleural pointillism might be greatest when clinical symptoms are still absent or nonspecific. Also, although interobserver agreement was acceptable, a relatively low k value was found for the DW imaging pointillism assessment, which was probably due to a lower level of experience in visual DW image interpretations of the radiologist performing the second DW image reading. However, the reproducibility was still similar to that of the evaluations of the other examined parameters in this study. In summary, visual assessment of pleural pointillism on high-b-value DW images seems to be a reasonably accurate tool for diagnosing MPD. Pleural pointillism might be used to provide guidance for biopsy or thoracoscopic evaluation and eventually could help obviate invasive procedures, thereby
Radiology: Volume 274: Number 2—February 2015 n radiology.rsna.org
Coolen et al
minimizing patient discomfort. Further larger studies are necessary to validate the clinical applicability and specificity of pleural pointillism. Disclosures of Conflicts of Interest: J.C. disclosed no relevant relationships. F.D.K. disclosed no relevant relationships. P.N. disclosed no relevant relationships. W.D.W. disclosed no relevant relationships. C.D. disclosed no relevant relationships. J. Vansteenkiste disclosed no relevant relationships. A.D. disclosed no relevant relationships. I.R. disclosed no relevant relationships. E.V. disclosed no relevant relationships. P.D.L. disclosed no relevant relationships. D.V.R. disclosed no relevant relationships. K.N. disclosed no relevant relationships. S.D. disclosed no relevant relationships. J. Verschakelen disclosed no relevant relationships.
References 1. Husain AN, Colby TV, Ordóñez NG, et al. Guidelines for pathologic diagnosis of malignant mesothelioma: a consensus statement from the International Mesothelioma Interest Group. Arch Pathol Lab Med 2009;133(8):1317–1331. 2. Leung AN, Müller NL, Miller RR. CT in differential diagnosis of diffuse pleural disease. AJR Am J Roentgenol 1990;154(3):487–492. 3. Wang ZJ, Reddy GP, Gotway MB, et al. Malignant pleural mesothelioma: evaluation with CT, MR imaging, and PET. RadioGraphics 2004;24(1):105–119. 4. Benamore RE, O’Doherty MJ, Entwisle JJ. Use of imaging in the management of malignant pleural mesothelioma. Clin Radiol 2005;60(12):1237–1247. 5. Traill ZC, Davies RJ, Gleeson FV. Thoracic computed tomography in patients with suspected malignant pleural effusions. Clin Radiol 2001;56(3):193–196. 6. Sahin AA, Cöplü L, Selçuk ZT, et al. Malignant pleural mesothelioma caused by environmental exposure to asbestos or erionite in rural Turkey: CT findings in 84 patients. AJR Am J Roentgenol 1993;161(3):533–537. 7. Kawashima A, Libshitz HI. Malignant pleural mesothelioma: CT manifestations in 50 cases. AJR Am J Roentgenol 1990;155(5):965–969. 8. Beyer T, Townsend DW, Brun T, et al. A combined PET/CT scanner for clinical oncology. J Nucl Med 2000;41(8):1369–1379. 9. Erasmus JJ, Truong MT, Smythe WR, et al. Integrated computed tomography-positron emission tomography in patients with potentially resectable malignant pleural mesothelioma: staging implications. J Thorac Cardiovasc Surg 2005;129(6):1364–1370.
583
THORACIC IMAGING: Malignant Pleural Mesothelioma MR Visual Assessment
Coolen et al
10. Bénard F, Sterman D, Smith RJ, Kaiser LR, Albelda SM, Alavi A. Metabolic imaging of malignant pleural mesothelioma with fluorodeoxyglucose positron emission tomography. Chest 1998;114(3):713–722.
17. Luo L, Hierholzer J, Bittner RC, Chen J, Huang L. Magnetic resonance imaging in distinguishing malignant from benign pleural disease. Chin Med J (Engl) 2001;114(6):645–649.
11. Gill RR, Gerbaudo VH, Jacobson FL, et al. MR imaging of benign and malignant pleural disease. Magn Reson Imaging Clin N Am 2008;16(2):319–339, x.
18. Luna A, Sánchez-Gonzalez J, Caro P. Diffusionweighted imaging of the chest. Magn Reson Imaging Clin N Am 2011;19(1):69–94.
12. Duysinx B, Nguyen D, Louis R, et al. Evaluation of pleural disease with 18-fluorodeoxyglucose positron emission tomography imaging. Chest 2004;125(2):489–493. 13. Kramer H, Pieterman RM, Slebos DJ, et al. PET for the evaluation of pleural thickening observed on CT. J Nucl Med 2004;45(6): 995–998. 14. Jaruskova M, Belohlavek O. Role of FDG PET and PET/CT in the diagnosis of prolonged febrile states. Eur J Nucl Med Mol Imaging 2006;33(8):913–918. 15. Fujimoto K. Usefulness of contrast-enhanced magnetic resonance imaging for evaluating solitary pulmonary nodules. Cancer Imaging 2008;8:36–44. 16. Hierholzer J, Luo L, Bittner RC, et al. MRI and CT in the differential diagnosis of pleural disease. Chest 2000;118(3):604–609.
584
19. Coolen J, De Keyzer F, Nafteux P, et al. Malignant pleural disease: diagnosis by using diffusion-weighted and dynamic contrastenhanced MR imaging—initial experience. Radiology 2012;263(3):884–892. 20. Gill RR, Umeoka S, Mamata H, et al. Diffusion-weighted MRI of malignant pleural mesothelioma: preliminary assessment of apparent diffusion coefficient in histologic subtypes. AJR Am J Roentgenol 2010;195(2):W125–W130. 21. Rusch VW, Piantadosi S, Holmes EC. The role of extrapleural pneumonectomy in malignant pleural mesothelioma: a Lung Cancer Study Group trial. J Thorac Cardiovasc Surg 1991;102(1):1–9.
cer staging manual. 7th ed. New York, NY: Springer, 2007; 271–277. 23. Attanoos RL, Gibbs AR. Morphologic variants and their mimics. In: Attanoos RL, Allen TC, eds. Advances in surgical pathology: mesothelioma. Philadelphia, Pa: Lippincott Williams & Wilkins, 2014; 67–90. 24. Sanna A. Pointillism. In: Impressionism. Florence, Italy: Scala Group, 2011; 441– 496. 25. Scott IR, Müller NL, Miller RR, Evans KG, Nelems B. Resectable stage III lung cancer: CT, surgical, and pathologic correlation. Radiology 1988;166(1 pt 1):75–79. 26. Yi CA, Jeon TY, Lee KS, et al. 3-T MRI: usefulness for evaluating primary lung cancer and small nodules in lobes not containing primary tumors. AJR Am J Roentgenol 2007;189(2):386–392. 27. Pass HI. Malignant pleural mesothelioma: surgical roles and novel therapies. Clin Lung Cancer 2001;3(2):102–117.
22. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. Pleural mesothelioma. In: Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC can-
radiology.rsna.org n Radiology: Volume 274: Number 2—February 2015
