Therapy response in malignant pleural mesothelioma-role of MRI using RECIST, modified RECIST and volumetric approaches in comparison with CT

Transcription

1 Eur Radiol (2008) 18: DOI /s CHEST Christian Plathow Michael Klopp Christian Thieke Felix Herth Andreas Thomas Astrid Schmaehl Ivan Zuna Hans-Ulrich Kauczor Therapy response in malignant pleural mesothelioma-role of MRI using RECIST, modified RECIST and volumetric approaches in comparison with CT Received: 30 October 2007 Accepted: 2 February 2008 Published online: 28 March 2008 # European Society of Radiology 2008 C. Plathow (*) Department of Nuclearmedicine, University of Freiburg, Hugstetter 55, Freiburg, Germany christian.plathow@uniklinikfreiburg.de Tel.: C. Plathow. I. Zuna. H.-U. Kauczor Department of Radiology, German Cancer Research Center Heidelberg, M. Klopp Department of Thoracic Surgery, Clinic for Thoracic Diseases, C. Thieke Department of Radiation Oncology, German Cancer Research Center Heidelberg, F. Herth Department of Pneumology, Clinic for Thoracic Diseases, A. Thomas Department of Oncology, Clinic for Thoracic Diseases, A. Schmaehl Department of Radiology, Clinic for Thoracic Diseases, Abstract To evaluate and compare early therapy response according to RECIST (response evaluation criteria in solid tumours) and modified RECIST criteria using MRI techniques in patients with malignant pleural mesothelioma (MPM) in comparison with CT. Fifty patients with MPM (32 male/18 female) were included in this study. Early therapy response was evaluated after 9 weeks [three of six chemotherapy (CHT)] cycles. Additionally patients were examined before chemotherapy, 4 weeks after early therapy response evaluation and after six cycles to evaluate diagnostic follow-up. RECIST and modified RECIST criteria were applied using CT and MRI (HASTE, VIBE, T2-TSE sequences). In MRI additionally a volumetric approach measuring tumour weight (overall segmented tumour volume) was applied. Additionally vital capacity (VC) was measured for correlation. Image interpretation was performed by three independent readers independently and in consensus. The gold standard was follow-up examination. Twenty-eight patients showed partial response, 12 patients stable disease and 10 patients progressive disease at early therapy response evaluation. In the follow-up these results remained. For MRI, in 46 cases patients were identically classified using RECIST and modified RECIST criteria. Modified RECIST criteria were identically classified as gold standards in all cases, whereas using RECIST criteria in four cases there was a mismatch (partial response vs. stable disease). Modified RECIST kappa values showed better interobserver variability compared with RECIST criteria (κ= vs ). For CT, in 44 cases patients were identically classified using RECIST and modified RECIST criteria. Modified RECIST criteria were identically classified as in gold standards in 48 out of 50 patients, whereas using RECIST criteria in 6 cases there was a mismatch (partial response vs. stable disease). Modified RECIST kappa values showed better interobserver variability compared with RECIST criteria (κ= vs ). Modified RECIST criteria especially in combination with high-resolution MRI is a very accurate and reproducible technique to correctly evaluate early therapy response in MPM. Keywords Mesothelioma. RECIST. MRI. Tumour volumetry. Validation

2 1636 Introduction Despite international efforts, occupational exposure to asbestos in industrialized countries remains a major cause of morbidity and mortality from both lung cancer and malignant pleural mesothelioma (MPM) [1]. Incidence of MPM currently ranges from about 7 to 40 per 1,000,000 in industrialized Western nations, depending on the amount of asbestos exposure of the populations during the past several decades [2]. The incidence is increasing, and prognosis for the disease is uniformely dismal. The major problem encountered in the treatment of MPM is poor local control [3]. To determine disease control an early reproducible and correct measurement of tumour response to treatment is vital, especially for the development of new drugs and therapeutic combinations [4]. Conventional response criteria are CT based and have always been difficult to apply to MPM due to its unique pattern of growth. MPM most commonly grows as a rind around the pleural surface and does not produce spherical lesions with bidimensionally measurable diameters. Thus, WHO criteria are poorly suited to evaluating response in MPM, as they were developed principally to assess bidimensionally measurable disease. The RECIST (response evaluation criteria in solid tumours [5]) criteria are more suited to tumour size of MPM, as they specify the use of unidimensional measurements [6]. However the selection of measurement sites in MPM is difficult [7]. Thus, in CT modified RECIST criteria that are specifically designed to address the unique growth pattern of MPM have been introduced [8]. It could be shown that using these criteria, difficult and ambiguous situations that arise in response interpretation can be avoided. Volumetric approaches measuring tumour weight (segmented overall tumour volume) have also been proposed to monitor tumour volume and therapy response, but still these semiautomatic techniques are time consuming and therefore not applied in clinic. Thus, recently a high precision, high reproducibility and a significant correlation of MR-base lung volumetry with vital capacity (VC) in patients with MPM also in the follow-up could be proven [9]. Even though MRI has been frequently shown to have certain advantages compared to CT [10, 11] in correctly delineating MPM, CT still is the most frequently applied method in clinic [12]. With the improvement of MR high resolution and dynamic imaging techniques together with new post-processing techniques, the potential of MRI has further increased [13 15]. However, so far there is no study that evaluated response criteria in MPM using MRI. In this study RECIST and modified RECIST criteria using CT and MRI techniques in MPM were compared at an early time point after the start of chemotherapy (CHT). Follow-up controls were performed for verification; also VC was measured and compared with a volumetric approach using multireader analysis. Materials and methods Fifty patients with a histologically proven MPM and a confirmed long-term asbestos exposure (32 males; 18 females; mean age 59 years; range 45 to 73 years, Karnofsky Index >80) were included in this study. After the nature of the procedure had been fully explained, informed consent was signed by all participants under an institutionally approved subject research protocol approved by the local ethical committee. After histological verification of the tumour, no treatment had been performed before CHT (combination of pemetrexed and cisplatin [16], there was no change of CHT during our investigations) in all patients. Early therapy response was evaluated after three out of six CHT cycles (9 weeks). All patients had at least four MRI and CT examinations: before chemotherapy (I), after three out of six cycles (9 weeks) (II), 4 weeks later (III) to reevaluate response criteria and after six cycles (follow-up control) (IV). Therapy response and the comparison between CT and MRI using RECIST and modified RECIST criteria were evaluated after three out of six cycles (time point II). MRI and CT All examinations were performed using a clinical 1.5-T whole-body MR system (Magnetom Symphony, Siemens Medical Solutions, Erlangen, Germany) equipped with eight receiver channels, a six channel coil and a high performance gradient system (30 mt/m) using a six-channel coil. For morphologic examination the following MRI sequences were applied (duration about 25 min): HASTE (half-fourier single-shot fast spin-echo sequence) was acquired in coronal and and transversal planes [TR/TE: 554/28 ms; flip angle: 180 ; receiver bandwidth: 610 Hz/ pixel; GRAPPA acceleration factor: 2; field of view (FOV): ; slice thickness: 6 mm]. VIBE [3-dimensional (3D) gradient-recalled echo (GRE) volumetric interpolated breath-hold examination] was acquired before and after contrast media application in coronal and transversal planes [TR/TE: 6.49/2.63 ms; flip angle: 10 ; receiver bandwidth: 300 Hz/pixel; field of view (FOV): ; slice thickness: 4 mm]. T2-TSE (turbo-spin echo) with respiratory gating was acquired in coronal planes [TR/TE: 775/18 ms; flip angle: 160 ; receiver bandwidth: 257 Hz/pixel; GRAPPA acceleration factor: 2; field of view (FOV): ; slice thickness: 4 mm] [17, 18]. Altogether patients had to hold breath for about 10 to 15 s at each sequence. This was tested and trained in each patient before MRI examination. For RECIST evaluation and volumetric segmentation HASTE sequence was used. CT imaging of the thorax was performed using a standard CT system (Somatom Plus 16 Volume Zoom, Siemens Medical Solutions, Erlangen, Germany). Highresolution (1-mm slice thickness, 137 kv, 110 mas) and

3 1637 contrast-enhanced CT images were acquired (5-mm slice thickness, 137 kv, 180 mas, 80 ml contrast medium iomeprol; Imeron 300, BRACCO-Altana, Konstanz, Germany). Native, arterial, portal venous and late phases were performed. The time interval between CT and MRI did not exceed 1 week. Image interpretation was performed by three independent experienced readers assessing tumour response according to the tumour response criteria (see below). All three readers were residents in a department focused on chest diseases (CT and MRI experience >2 years). The presets for CT images were as follows: soft tissue window, width + 360, level + 40; lung parenchyma/ pleural window, width , level 300. All images were jointly randomised and showed the readers in order of randomisation. All readers were blinded to clinical data as well as the interpretations of other readers. After individual scoring a consensus scoring was performed. Analysis Tumour measurement Tumour measurements were performed on transverse cuts on thoracic images at three separate anatomically reproducible levels in MRI and CT, chosen by the three readers in consensus, and at the same levels on subsequent following images. This technique has been recently proposed for image interpretation in MPM by Byrne and Nowak [9]. RECIST criteria For unidimensionally measurable lesions, the dimensions of the pleural tumour were measured at two separate sites on each of the three levels, and the six measurements were summated to produce a total measurement. A pleural effusion was not considered a measurable lesion. Tumour response was defined as: complete response (CR) disappearance of all visible disease; partial response (PR) a >=30% decrease in the sum of linear tumour measurements on two observations not less than 4 weeks apart; no change (NC) a decrease in the sum of unidimensional measurements of <30% or an increase of <20% in the size of the tumour; progressive disease (PD) an increase of at least 20% in the total tumour measurement, or appearance of one or more new lesions (Fig. 1) Fig. 1 Example of measurement of a single HASTE slice of a patient with left-sided MPM using 2D and 3D technique. a Lines represent possible interpretation longest tumour diameter according to current RECIST criteria. b Lines represent measurement sites perpendicular to fixed structures, chest wall, sternum and vertebral column, according to modified RECIST criteria. c Volumetric approach using HASTE sequence. Normal lung tissue (bright grey) and tumour 2D-technique 3D-technique Modified RECIST criteria For details see [9]. With the widespread adoption of the RECIST criteria in other malignancies, the need to use both unidimensional and bidimensional lesions to assess response in MPM became less important. The Modified RECIST criteria use unidimensional measurement of tumour thickness perpena b c

4 1638 dicular to the chest wall or mediastinum, measured in two sites at three different levels on CT. Transverse images used for measurement must be at least 1 cm apart and related to anatomical landmarks in the thorax, preferably above the level of the division of the main bronchi. At reassessment, pleural thickness must be measured at the same position and level. Extensive documentation of nonmeasurable lesions is not required; however, nodal, subcutaneous, and other bi-dimensionally measurable lesions are measured uni-dimensionally as per the RECIST criteria. Uni-dimensional measurements are added to produce a total tumour measurement. Definition of complete and partial response is concordant with the RECIST criteria (Fig. 1). Volumetric approach in MRI The technique of segmentation and tumour volume quantification recently has been published in MPM [10]. Using MRI an automatic segmentation of the tumour volumes failed due to intensity inhomogeneities and motion artefacts. As correlation between tumour segmentation and real tumour volume using this technique proved to be very precise, segmentation could be performed by one reader [19]. Each tumour slice was segmented semi-automatically using an interactive region-growing technique as provided by the segmentation plug-in of the teleradiological workstation CHILI (CHILI GmbH, [20, 21]). Manual interaction could be reduced by segmenting only every second slice and interpolating the intermediate image using shape-based interpolation [22]. Segmentation inconsistencies between neighbouring slices were diminished by means of morphological image processing [23]. Segmentation of one HASTE data set took about 30 s, thus for each patient about 10 min were necessary for segmentation. For each data set, the 2D segmentation results have been recombined to obtain a voxel-based 3D description of the current breathing state. The number of voxels was counted and multiplied by the volume of a single voxel to receive the total volume during the breathing cycle as provided by the Visualization Toolkit VTK (Kitware Inc, Clifton Park, NY) [24]. This process took about 5 min in every patient. Tumour response was categorized as described before. Volumetry was performed before CHT (time point I), at time point II and after CHT (time point IV) (Figs. 1 and 2). Response status was assessed and compared using RECIST and modified RECIST. Spirometry It has been shown that vital capacity (VC) has a significant correlation with therapy response [9]. Procedures and quality criteria of the European Respiratory Society were used for all measurements [25]. An improvement of >=30% VC was defined as PR, an improvement of <=30% to a deterioration of <=20% was defined as NC, and a deterioration of >=20% was defined as PD [7]. Statistical analysis Response status was assessed and compared using RECIST and modified RECIST. To compare the value of both techniques using MRI and CT, follow-up controls served as gold standard. Additionally the volumetric approach was performed, as recently has been proposed [13]. Additionally VC was measured. Fig. 2 Example of a patient with right-sided MPM malignant pleural mesothelioma. On the basis of the contrast-enhanced CT images (a), pleural effusion could only barely be detected; also mediastinal infiltration could not be excluded. Using T2w (b, upper image) and T1w contrast-enhanced MRI (b, image below), these aspects could be exactly evaluated. Here a substantial mediastinal infiltration and pleural effusion could be delineated, which has to be taken care of in the RECIST categorisation. On the basis of CT imaging, a stable disease was evaluated in the follow-up; on the basis of MRI a partial response could be proven with an increasing pleural effusion

5 1639 Tumorweight [%] ** ** ns Results Volumetric changes revealed a partial response in 28 patients, stable disease in 12 patients and progressive disease in 10 patients. In the two follow-up examinations, the same 28 patients showed also partial response, the same 12 patients showed stable disease and 10 patients showed progressive disease. Using the volumetric approach, there was an overall decrease in relative tumour volume from 100% to 50±30% and 60±25% (Fig. 3). There were no relevant motion or breathing artefacts using MRI. 0 Before CHT Early therapy After CHT response during CHT Fig. 3 Relative volumetric tumour weight according to segmented Haste sequences (see also Fig. 1). There was a significant overall decrease of tumour weight according to chemotherapy (CHT = chemotherapy, **p<0.01, ns = non-significant) Data entry procedures and statistical analysis were performed with a statistical analysis software system (SAS for Windows, SAS Institute, Cary, NC). For statistical test comparisons, non-parametric methods were applied. For comparison of observer performance, a κ statistic was used. Inter-observer agreement was considered as slight, κ 0.2; fair, κ= ; moderate, κ= ; substantial, κ= ; or almost complete, κ= [26]. Also consensus reading was performed. To obtain simultaneous tests for mean comparisons, the Duncan s multiple-range test of ANOVA on the significance level of p=0.05 was used. The mean and standard deviation were calculated. In addition, a t-test for paired samples was performed. P-values less than 0.05 were considered statistically significant; P-values less than 0.01 were considered statistically highly significant. RECIST vs. modified RECIST during CHT (time point II) MRI HASTE sequence proved to be most helpful for all readers in the correct delineation of the tumour. In 46 cases patients were identically classified using RECIST and modified RECIST criteria. Criteria were identically classified by modified RECIST as by volumetric approaches in all cases, whereas using RECIST criteria there was a mismatch in four cases (partial response vs. stable disease). These four cases showed a partial response after six cycles using RECIST, VC and modified RECIST (Table 1). Modified RECIST kappa values showed substantially better interobserver variability compared with RECIST criteria (κ= vs , Table 2). CT In 44 cases patients were identically classified using RECIST and modified RECIST criteria. Modified RECIST criteria were identically classified as in volumetric approaches in 48 out of 50 patients, whereas using RECIST criteria in 6 cases there was a mismatch (partial response vs. stable disease, Table 3). CT showed a tendency to underclassify therapy response. Modified RECIST kappa values showed substantial better interobserver variability compared with RECIST criteria (κ= vs , Table 4). Table 1 Comparison of response at individual time point observations between original response criteria and modified RECIST criteria using MRI (between modified RECIST criteria and volumetric approach was not distinguished, as there was no numerical difference Original RECIST Modified RECIST/volumetric approach CR PR SD PD Total CR PR SD PD Total Next follow-up controls CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; for details see text). Response in the next followup control is shown

6 1640 Table 2 Kappa values for the interobserver agreement Original RECIST/modified RECIST/ volumetric Approach Reader 1 vs. 2 Reader 1 vs. 3 Reader 2 vs. 3 Median CR 1.0/1.0/ /1.0/ /1.0/ /1.0/1.0 PR 0.7/0.9/ /0.9/ /1.0/ /0.9/1.0 SD 0.7/0.9/ /0.9/ /0.8/ /0.9/1.0 PD 0.95/1.0/ /1.0/ /1.0/ /1.0/1.0 Median 0.82/0.95/ /0.95/ /1.0/1.0 CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; for details see text) Changes of the VC in different therapy response groups of patients are shown in Table 5. In patients with a partial response, VC improved by 52±14%, in patients with stable disease VC improved by 7±9% (p<0.01 to patients with partial response), in patients with a progressive disease VC changed by 41±9% (p<0.05 to patients with stable disease). Discussion In this study, MRI has been applied to classify treatment response in MPM in comparison with CT. RECIST and Modified RECIST have been applied and compared as these techniques are the most frequently clinically applied techniques to evaluate therapy response. Follow-up control served as the gold standard. Also a volumetric approach was performed, which might be the technique of choice in the future, but still this technique is clinically too time consuming. Modified RECIST and volumetric approaches proved to be the most adequate scores to correctly classify tumour response. CT and conventional RECIST proved to be less precise. MRI has several advantages compared with CT. First, it avoids ionising radiation, even if this is not an issue for these patients. Secondly, MRI allows direct acquisition of images in multiple different planes. Although this is not a real limitation of CT using modern CT techniques, it is important especially in the evaluation of irregular growing tumours such as MPM. In comparison to chest radiography and CT, MRI has additionally a superior soft tissue contrast even without intravenous contrast media [27]. Therefore, several studies recommended MRI to better delineate local extent of MPM, e.g., T2-weighted HASTE images have been proven to better image pleural fluid [28, 29] in comparison with CT. Also, in a recent study it was shown that high-resolution MRI achieves better interobserver agreement in assessing pleural thickening, pleural effusion and MPM [9, 19]. Volumetric approaches also showed that MRI techniques in combination with sophisticated segmentation techniques are able to correctly quantify lung and lung tumour volumes with a good correlation of therapy response and spirometry [18, 19]. This was the basis to evaluate new MRI techniques in MPM using different established scoring techniques. It must emphasised that presently the speed of CT imaging and thus the certain independence from the patient compliance still is an advantage of CT. As CT also is the most commonly applied technique in MPM and still the gold standard, CT was also evaluated. But, as in recently published studies, the accuracy of CT measurements was less than with MRI [8]. RECIST criteria have been developed and become widely accepted and used in clinical trials. The application of conventional RECIST criteria could, however, be variably interpreted by different investigators in MPM, and this may lead to unsatisfactory results [5]. In MPM, there are numerous possible interpretations of the RECIST criteria, and more guidelines for applying these criteria are Table 3 Comparison of response at individual time point observations between original response criteria and modified RECIST criteria using CT (CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; for details see text) Original RECIST Modified RECIST/ volumetric approach CR PR SD PD Total CR PR SD PD Total Next follow-up controls

7 1641 Table 4 Kappa values for the interobserver agreement using CT (CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; for details see text) Original RECIST/modified RECIST Reader 1 vs. 2 Reader 1 vs. 3 Reader 2 vs. 3 Median CR 0.9/ / / /1.0 PR 0.6/ / / /0.9 SD 0.7/ / / /0.9 PD 0.9/ / / /1.0 Median 0.8/ / /1.0 needed if investigators from different sites and countries are to be consistent in their definition of response. Consistency is essential for interpretation of results, particularly in phase II clinical trials. For example, what constitutes the longest diameter of the tumour in MPM? The longest diameter commonly follows the curve of the chest wall; however, defining the limits of such lesions is often difficult or impossible. While tumour thickness may change in response to successful treatment, the extent of tumour along the chest wall, and thus this putative longest diameter, may change little. Furthermore, the longest diameter may be between two fixed structures, such as the carina and thoracic vertebra, and should not be used to assess response. More guidance is needed to choose the anatomical level of tumour measurement, how many levels should be assessed, how to choose five lesions from the hemithorax, and what lesions should be recorded, if any, as non-target lesions. Van Klaveren et al. reported a series of tumour measurements comparing the use of RECIST and WHO criteria for assessment of response in MPM [30]. They found a discrepancy in response as defined by the two sets of criteria in 9 of 34 cases (27%). Eight of 34 cases demonstrated progression on the WHO criteria, which was missed by RECIST at some time point. These results were no better in the subset of patients with bi-dimensonally measurable lesions. However, problems with both the WHO criteria and the unmodified RECIST criteria suggest that neither should be considered a gold standard for tumour measurement in this disease. This was confirmed in our study. Besides lower kappa values in comparison with the two other scoring techniques, 4 patients out of 50 were categorized incorrectly using MRI and even 6 using CT. As shown by Nowak et al. Table 5 Vital capacity changes VC CR PR SD PD Patients [n] Changes [%] 0 52±14 7±9 41±9 [9] using modified RECIST criteria, there was a benefit to patients in terms of survival and lung function, and therefore a valid clinical trial endpoint. There was a significant difference in survival between the two patient groups, with a median survival of 15.1 months for responding patients and 8.9 months for non-responding patients. Lung function as measured by forced vital capacity improved significantly over the course of treatment for responding patients as compared with nonresponding, and there was a significant correlation between lung function and change in linear tumour measurement. Using modified RECIST citeria, response of all patients (in CT in all but one) was categorized correctly with very high kappa values. Thus, difficult and ambiguous situations could be avoided using modified RECIST criteria. The overall percentage of different tumour response categorisation was similar to a recently published study [6]. In our study data were compared with 3D volumetric approaches, but also VC and follow-up were applied as recently has been proposed [8]. The high precision of MRIbased volumetry measurement recently has been published [10, 19], but still this semiautomatic approach is too time consuming to generally recommend for tumour response evaluation. Another important information of this study is that therapy response can be evaluated at a very early stage of CHT. Follow-up controls showed the same therapy response, which might imply that therapy responders and non-responders might be selected at a very early stage. Therapy changes and optimization might be possible at an early stage, and useless CHT and thus side effects might be preventable. PET/CT recently has been proposed to be the technique of choice in the correct staging and monitoring of patients with MPM. It has been shown that FDG-PET is useful for the differentiation of benign from malignant lesions, for staging and monitoring metabolic response to therapy against MPM, and that it has prognostic value [31]. An initial report on PET/CT imaging of MPM has shown increased accuracy of overall staging, improving the assessment of tumour respectability. But still PET/CT is not the clinical standard, and further research will have to show the potential additional benefit of PET/CT to MRI and CT [32, 33].

8 1642 On the basis of this study modified RECIST criteria in combination with MRI have to be recommended to evaluate early therapy response. Modified RECIST criteria in combination with CT deliver certain less precise results, but also can be recommended. Especially in patients with limited physical status and breathing problems, the speed of CT might render it the technique of choice. Conclusion Modified RECIST criteria in combination with MRI is a very accurate technique to correctly evaluate therapy response in MPM. Volumetric approaches might be the technique of choice in the future, but still this technique is too time consuming for clinical routine. Conflict of interest statement There is no conflict of interest. References 1. Tossavainen A (2000) International expert meeting on new advances in the radiology and screening of asbestosrelated diseases. Scand J Work Environ Health 26: Robinson BWS, Lake RA (2005) Advances in malignant mesothelioma. Eng J Med 353: Rusch VW (1993) Pleurectomy/ decortication and adjuvant therapy for malignant mesothelioma. Chest 103: Roach HD, Davies GJ, Attanoos R et al (2002) Asbestos: When the dust settlesan imaging review of asbestos-related disease. Radiographics 22: Therasse P, Arbuck SG, Eisenhauer EA et al (2000) New guidlines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 92: Ceresoli GL, Chiti A, Zucali PA et al (2007) Assessment of tumor response in malignant pleural mesothelioma. Cancer Treat Rev 33: Van Klaveren RJ, Aerts JGJV, de Bruin HG et al (2002) Inadequacy of the RECIST criteria for the evaluation of response in patients with malignant pleural mesothelioma (MPM). Proc Soc Clin Oncol 21:310a 8. Byrne MJ, Nowak AK (2004) Modified RECIST criteria for assessment of response in malignant pleural mesothelioma. Ann Oncol 15: Plathow C, Klopp M, Schoebinger M et al (2006) Monitoring of lung motion in patients with malignant pleural mesothelioma using two-dimensional and three-dimensional dynamic magnetic resonance imaging: comparison with spirometry. Invest Radiol 41: Weber MA, Bock M, Plathow C et al (2004) Asbestos-related pleural disease Value of dedicated magnetic resonance imaging techniques. Invest Radiol 39: Yamamuro M, Gerbaudo VH, Gill RR et al (2007) Morphologic and functional imaging of malignant pleural mesothelioma. Eur J Radiol 64(3): Heelan RT, Rusch VW, Gebb CB et al (1999) Staging of malignant pleural mesothelioma: Comparison of CT and MR imaging. AJR Am J Roentgenol 172: Cluzel P, Similowski T, Chartrand- Lefebvre C et al (2000) Diaphragm and chest wall: assessment of the inspiratory pump with MR imagingpreliminary observations. Radiology 215: Plathow C, Fink C, Ley S et al (2004) Measurement of diaphragmatic length during the breathing cycle by dynamic MRI: comparison between healthy adults and patients with an intrathoracic tumor. Eur Radiol 14: Plathow C, Ley S, Fink C et al (2004) Evaluation of chest motion and volumetry during the breathing cycle by dynamic MRI in healthy subjects: comparison with pulmonary function tests. Invest Radiol 39: Vogelzang NJ, Rusthoven JJ, Symanowski J et al (2003) Phase III study of pemetrexed in combination with cisplatin versus cisplatin alone in patients with malignant pleural mesothelioma. J Clin Oncol 21: Schafer JF, Vollmar J, Schick F et al (2005) Detection of pulmonary nodules with breath-hold magnetic resonance imaging in comparison with computed tomography. Rofo 177: Both M, Schultze J, Reuter M et al (2005) Fast T1- and T2-weighted pulmonary MR-imaging in patients with bronchial carcinoma. Eur J Radiol 53: Plathow C, Schoebinger M, Fink C et al (2006) Quantification of lung tumor volume and rotation at 3D dynamic parallel MR imaging with view sharing: preliminary results. Radiology 240: Kunert T, Heimann T, Schröter A, et al (2004) An Interactive System for Volume Segmentation in Computer- Assisted Surgery. Proc. SPIE Medical Imaging Vol / Engelmann U, Schroeter A, Schwab M et al (1999) Borderless teleradiology with CHILI. J Med Internet Res 1:E8 22. Herman GT, Zheng J, Bucholts CA (1992) Shape-based interpolation. IEEE Comp Graph Appl 12: Soille P (2002) Morphological image analysis, 2nd edn. Springer, Berlin, Heidelberg, New York 24. Schroeder W, Martin K, Lorensen B (2003) The visualization toolkit: An object-oriented approach to 3D Graphics, 3rd edn., Kitware, Inc. 25. Quanjer PHH, Tammeling GJ, Cotes J et al (1993) Lung volumes and forced ventilatory flows. Eur Respir J 6: Svanholm H, Starklint H, Gundersen HJ et al (1989) Reproducibility of histomorphologic diagnoses with special reference to the kappa statistic. APMIS 97: Wang ZJ, Reddy GP, Gotway MB et al (2004) Malignant pleural mesothelioma: evaluation with CT, MR imaging, and PET. Radiographics 24: Knuutila A, Kivisaari L, Kivisaari A et al (2001) Evaluation of leural disease using MR and CT. With special reference to malignant pleural mesothelioma. Acta Radiol 42: Eibel R, Tuengerthal S, Schoenberg SO (2003) The role of new imaging techniques in diagnosis and staging of malignant pleural mesothelioma. Curr Op in Oncol 15:

9 Van Klaveren RJ, Aerts JGJV, de Bruin HG et al (2006) Inadequacy of the RECIST criteria for the evaluation of response in patients with malignant pleural mesothelioma. Lung Cancer 43: Yamamuro M, Gerbaudo VH, Gill RR et al (2007) Morphologic and functional imaging of malignant pleural mesothelioma. Eur J Radiol 64(3): Armato SG, Entwise J, Truong MT, et al. (2007) Current state and future directions of pleural mesothelioma imaging. Lung Cancer 2007 Nov 29, Epub ahead of print. DOI /j. lungcan Plathow C, Walz M, Lichy MP, et al. (2008) Cost considerations of wholebody MRI and PET/CT as part of oncologic staging. Radiologe Epub ahead of print /s z