Postoperative Lung Function in Lung Cancer Patients: Comparative Analysis of Predictive Capability of MRI, CT, and SPECT
Abstract
OBJECTIVE. The purpose of this study was to prospectively compare the utility of dynamic contrast-enhanced perfusion MRI in the prediction of postoperative lung function in patients with lung cancer with the utility of quantitative and qualitative assessment of CT and perfusion SPECT.
SUBJECTS AND METHODS. One hundred fifty lung cancer patients (87 men, 63 women) underwent dynamic perfusion MRI, MDCT, perfusion SPECT, and measurement of preoperative and postoperative forced expiratory volume in the first second of expiration (FEV1) expressed as percentage of predicted value. Postoperative FEV1 was predicted with dynamic perfusion MRI by semiquantitative assessment of the perfusion of whole lungs and resected segments of lungs, with quantitative assessment of functional lung volume on CT with commercially available software, with qualitative assessment of CT on the basis of the number of segments of total and resected lung, and with perfusion SPECT by assessment of uptake of microaggregated albumin particles in whole lungs and resected segments of lungs. Correlation and limits of agreement between actual and predicted postoperative FEV1 values were statistically evaluated.
RESULTS. Actual postoperative FEV1 had stronger correlation with postoperative FEV1 predicted from perfusion MRI (r = 0.87, p < 0.0001) and quantitative CT (r = 0.88, p < 0.0001) than with postoperative FEV1 predicted from qualitative CT (r = 0.83, p < 0.0001) and perfusion SPECT (r = 0.83, p < 0.0001). The limits of agreement between the actual postoperative FEV1 and postoperative FEV1 predicted from perfusion MRI (5.3% ± 11.8% [mean ± 2 SD]) were smaller than the values for postoperative FEV1 predicted from qualitative CT (6.8% ± 14.4%) and perfusion SPECT (5.1% ± 14.0%) and was almost equal to the value for postoperative FEV1 predicted from quantitative CT (5.0% ± 11.6%).
CONCLUSION. Dynamic perfusion MRI is more accurate in prediction of the postoperative lung function of patients with lung cancer than are qualitative CT and perfusion SPECT and may be at least as accurate as quantitative CT.
Introduction
Lung cancer is the most common cause of cancer-related death among both men and women [1]. Despite advances in radiation therapy and chemotherapy, surgical resection remains the choice of treatment of patients with resectable non-small cell lung cancer. Nevertheless, it is estimated that only 20-25% of patients with non-small cell lung cancer undergo resection [2]. One reason for this low resection rate is that most patients with lung cancer have a history of cigarette smoking, which often engenders other conditions, such as chronic obstructive pulmonary disease and coronary artery disease, that can increase operative risk. Because clinicians are frequently asked to evaluate the risks and feasibility of lung resection for patients with multiple conditions, exercise testing and prediction of postoperative lung function have become increasingly important in the evaluation for lung resection [3, 4]. An algorithm for the functional assessment of candidates for lung resection has been proposed by Wyser et al. [5].
In current medical practice, ventilation-perfusion lung scintigraphy combined with spirometry is the most widely used radiologic examination when spirometric findings alone indicate pulmonary function may not be sufficient to tolerate resection [5]. The reported correlation coefficients of predicted and actual postoperative lung function measured with lung scintigraphy vary between 0.51 and 0.92 [5-11]; however, poor spatial resolution, especially troublesome for differentiating lobes and segments, is a major limitation of this method. A few investigators [12, 13] have suggested perfusion SPECT may be more effective than perfusion scintigraphy for the prediction of postoperative lung function. Another approach to evaluating surgical risk for lung cancer patients is quantitative or qualitative evaluation based on lung attenuation or anatomic findings on CT [4, 5, 9-12]. Although the correlation between postoperative lung function predicted with quantitative assessment of CT and actual postoperative lung function has been excellent, qualitative assessment (i.e., simple calculation based on the number of segments of total and resected lung) is used by many clinicians.
It has been found that 3D dynamic contrast-enhanced perfusion MRI is useful for evaluation of regional pulmonary perfusion and assessment of physiologic and pathophysiologic conditions in healthy volunteers, animal models, and patients with pulmonary vascular disease [11, 14-19]. A few investigators compared the capability of this technique with that of perfusion scintigraphy in assessment of regional perfusion and prediction of outcome among lung cancer patients [11, 17]. The number of patients in those studies was limited, however, and the capability of perfusion MRI in prediction of postoperative lung function was not compared with the predictive capability of quantitative and qualitative assessment of CT (quantitative and qualitative CT). For our study, we planned prospective recruitment of a large cohort of patients with lung cancer. The purpose of the study was to prospectively determine the capability of dynamic contrast-enhanced perfusion MRI in prediction of post-operative lung function in direct comparison with the capability of quantitative and qualitative CT and perfusion SPECT.
Subjects and Methods
Subjects
The study cohort comprised 150 consecutively registered lung cancer patients (87 men, 63 women; age range, 43-85 years; mean, 66 years) considered candidates for lung resection. All of them underwent preoperative contrast-enhanced MDCT, dynamic perfusion MRI, perfusion SPECT, and measurement of preoperative and postoperative forced expiratory volume in the first second of expiration (FEV1) expressed as percentage of predicted value. All preoperative radiologic examinations were performed in random order and less than 1 week before or after MRI (range, 1-6 days; mean, 3.2 days). Of the 150 patients, 87 had adenocarcinoma other than bronchioalveolar carcinoma, 47 had bronchioalveolar carcinoma, nine had squamous cell carcinoma, six had large cell carcinoma, and one had small cell carcinoma. The final diagnosis in all cases was confirmed with pathologic examination of resected specimens. Our institutional review board approved this study, and written informed consent was obtained from the subjects before they joined the study. Patient characteristics are summarized in Table 1.
| Characteristic | Value |
|---|---|
| Age (y) | |
| Mean | 66 |
| Range | 43-85 |
| Sex (no.) | |
| Men | 87 |
| Women | 63 |
| Histologic subtype (no.) | |
| Bronchioalveolar carcinoma | 47 |
| Adenocarcinoma other than bronchioalveolar carcinoma | 87 |
| Squamous cell carcinoma | 9 |
| Large cell carcinoma | 6 |
| Small cell carcinoma | 1 |
| Operation (no.) | |
| Lobectomy | 97 |
| Bilobectomy | 10 |
| Pneumonectomy | 5 |
| Segmentectomy | 38 |
| Predicted preoperative FEV1 (percentage of predicted value) | |
| Mean ± SD | 85.1 ± 15.0 |
| Range | 45.0-120.0 |
| Actual postoperative FEV1 (percentage of predicted value) | |
| Mean ± SD | 74.0 ± 11.9 |
| Range | 49.0-99.8 |
Dynamic Contrast-Enhanced Perfusion MRI
All MRI studies were performed on a 1.5-T MRI unit (Gyroscan Intera, Philips Medical Systems) with a phased-array coil. Dynamic perfusion MRI (TR/TE, 2.7/0.6; flip angle, 40°; matrix size, 128 × 96; reconstructed matrix size, 256 × 192; rectangular field of view, 450-530 × 315-371 mm) were acquired with a 3D radiofrequency spoiled gradient-echo sequence. A 3D slab thickness of 100 mm was used with 10 partitions and an overlapping slice in the coronal plane in a left-to-right phase-encoded direction. The result was an effective partition thickness of 10 mm and five-step real-phase encoding in the slice direction. The temporal resolution was 1.0 second for each 3D data set. All patients received 3-5 mL of gadopentetate dimeglumine (Magnevist, Schering) in a bolus administered through an antecubital vein with an automatic infusion system (Sonic shot, Nemoto) at a rate of 3-5 mL/s followed by 20 mL of saline solution at the same rate. The basic theory and application of dynamic perfusion MRI have been documented in the past literature [11, 16, 19]. After careful instruction, patients practiced the breath-hold technique to reproduce precisely the same degree of inspiration for each MR image series. In each acquisition, 25 images were obtained during a 25-second breath-hold at end inspiration. All 150 dynamic contrast-enhanced perfusion MRI examinations were completed successfully without adverse effects.
Image and Data Analysis of Dynamic Contrast-Enhanced Perfusion MRI
Signal intensity-time course curves after administration of gadopentetate dimeglumine were generated by measurement of the signal intensity in regions of interest (ROIs) delineated with our proprietary software in the right upper, right middle, right lower, left upper, left middle, and left lower lung fields in every slice of every subject (60 ROIs per patient). Large vessels and pulmonary arteries were excluded from the ROIs. From each ROI, data were transferred to a PC (FMV-900, Fujitsu) and were analyzed with Excel 2003 software (Microsoft) by a chest radiologist with 13 years of experience.
For extraction of quantitative indexes, the signal intensity-time course curves were fitted to a gamma variate function with the following equation [11, 16]:
\[S_{(t)}=S_{\mathit{\mathrm{peak}}}\left(\frac{\mathrm{e}}{{\alpha}{\beta}}\right)^{{\alpha}}(t-T_{a})^{{\alpha}}\mathrm{e}^{[-(t-T_{a}){/}{\beta}]}+S_{0}\]
where t is the time, S(t) is the measured signal intensity as a function of time, S0 and Speak are the base-line and peak signal intensities, Ta is the arrival time of the contrast bolus, and α and β are fitting parameters of the gamma variate function. Because of the small dose of injected contrast agent used, it was assumed that a linear relation exists between first-pass MRI signal intensity and contrast medium concentration in the ROI.
From the gamma variate function, the following equation [11, 16] was used to calculate the apparent mean transit time as the first moment on the MRI signal intensity-time course curve:
\[\mathrm{Mean\ transit\ time}=\frac{{{\int}}t{\times}(S_{(t)}-S_{0})dt}{{{\int}}(S_{(t)}-S_{0})dt}\]
Regional pulmonary blood volume was calculated directly from the area of the MRI signal intensity-time course curve for a given ROI. According to the central volume principle, regional blood flow in each ROI (QROI) was determined by dividing pulmonary blood volume by mean transit time [11, 16, 19]. The QROI value was then normalized to the integrated arterial input function from the main trunk of the pulmonary artery [11, 16]. The approximate time for an ROI measurement was 1 minute.
To determine regional perfusion in a lung field on dynamic contrast-enhanced perfusion MRI for prediction of postoperative lung function, blood flow in an ROI evaluated with dynamic perfusion MRI (QMRI), expressed as a percentage, was calculated as follows:
\[Q_{\mathrm{\mathit{MRI}}}=\frac{{{\sum}_{n=1}^{10}}Q_{\mathrm{\mathit{ROI}}}(n)}{{{\sum}_{n=1}^{10}}(Q_{\mathrm{\mathit{RUL}}(n)}+Q_{\mathrm{\mathit{RML}}(n)}+Q_{\mathrm{\mathit{RLL}}(n)}+Q_{\mathrm{\mathit{LUL}}(n)}+Q_{\mathrm{\mathit{LML}}(n)}+Q_{\mathrm{\mathit{LLL}}(n)})}{\times}100\]
where n is the slice number, QROI(n) is the blood flow of the ROI, QRUL(n) is the blood flow of the right upper lung field, QRML(n) is the blood flow of the right middle lung field, QRLL(n) is the blood flow of the right lower lung field, QLUL(n) is the blood flow of the left upper lung field, QLML(n) is the blood flow of left middle lung field, and QLLL(n) is the blood flow of the left lower lung field, all on slice n.
To evaluate capability in prediction of postoperative lung function, postoperative FEV1 predicted with dynamic perfusion MRI was calculated as follows:
\[\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\ }\mathrm{from\ perfusion}{\ }\mathrm{MRI}=\mathrm{FEV}_{1}{\times}\left(1-\frac{Q_{\mathrm{\mathit{MRI}}}{\ }\mathrm{of\ resected\ lung\ or\ lobe}}{100}\right)\]
The details of calculation of regional blood flow and postoperative FEV1 predicted from perfusion MRI have been described in the literature [11].
CT Examination
All CT examinations were performed with an MDCT scanner (Somatom Plus 4 Volume Zoom, Siemens Medical Solutions). The scans were obtained from the lung apex to the diaphragm (collimation, 4 × 1 mm; pitch, 6:1; field of view, 300-350 mm; matrix size, 512 × 512; 330 mA at 140 kV) and reconstructed as 5-mm-thick slices. Before CT, patients practiced breathing to produce full and consistent inspiration. CT was performed during a breath-hold at the end of full inspiration. Contrast medium (iopamidol, Iopamiron 300, Schering Japan) was administered IV with a power injector (Auto Enhance-50, Nemoto) at 2-3 mL/s through an antecubital vein with an empiric scan delay of 20 seconds to delineate the boundaries between tumor and mediastinal structures.
Image and Data Analysis of CT
Quantitative prediction of postoperative lung function—For quantitative prediction of postoperative lung function from functional lung volume, we used an assessment method described in the literature [10, 11]. After applications of dual thresholds of -500 and -910 H, total functional lung volume and regional functional lung volume of the lung or lobe to be resected were calculated by multiplying the area of each functionally relevant lung tissue by the slice thickness. The area of associated emphysema was excluded by the lower threshold value (-910 H). Areas of tumor-related air space loss, such as the tumor itself and postobstructive atelectasis, and areas of air space loss not related to the tumor, such as fibrosis and atelectasis due to previous tuberculosis, were satisfactorily excluded during visual inspection of the functional lung volume map. A chest radiologist with 15 years of experience used commercially available software (Pulmo, Siemens Medical Solutions) to perform all quantitative assessments of functional lung volume.
Predicted postoperative FEV1 evaluated by quantitative assessment of CT scans was calculated from total functional lung volume and regional functional lung volume with the following equation [10, 11, 20]:
\[\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\ }\mathrm{from\ quantitative}{\ }\mathrm{CT}=\mathrm{Preoperative}{\ }\mathrm{FEV}_{1}{\times}\left(1-\frac{\mathrm{Regional\ funtional\ volume\ of\ resected\ lung\ or\ lobe}}{\mathrm{Total\ functional\ lung\ volume}}\right)\]
The regional functional lung volume of resected lung or lobe was determined for each slice as the sum of regional functional lung volumes calculated from ROIs placed on the resected lobe or lung.
Qualitative prediction of postoperative lung function—For qualitative prediction of postoperative lung function, postoperative FEV1 was obtained from preoperative pulmonary function test data and information on the number of bronchopulmonary segments removed obtained according to a method used in many surgical institutions [3, 20-23]. For determination of the number of bronchopulmonary segments removed, all studies were interpreted by a chest radiologist with 7 years of experience and by a pulmonary surgeon with 24 years of experience; final assessments were made in consensus. The postoperative FEV1 predicted with qualitative assessment of CT scans was estimated with the following formula [3, 20-23]:
\[\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\ }\mathrm{qualitative}{\ }\mathrm{CT}=\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\times}(1-S{\times}0.0526)\]
where S is the number of bronchopulmonary segments removed by lung resection. Each segment is considered to represent 1/19 of lung function (1/19 = 0.0526). The lower lobes were considered to have five pulmonary segments each, the upper lobe of the right lung to have three segments, the middle lobe of the right lung to have two segments, and the upper lobe of the left lung to have four segments [3, 20-23].
Perfusion SPECT Examination
All perfusion SPECT data were obtained with a SPECT system (e-CAM, Siemens Medical Solutions) equipped with a medium-energy all-purpose collimator and with the subject in the supine position. All perfusion SPECT examinations were performed without respiratory gating and with IV administration of 185 MBq of 99mTc-microaggregated albumin. Images were acquired in 60 projections over 360° by the step-and-shoot method. The integration time per image was 20 seconds, and the matrix size was 64 × 64. A total of 38-49 6.4-mm-thick transaxial sections covering both whole lungs were reconstructed with a Butterworth filter (order, 8; cutoff frequency, 0.34 cycle/cm) and a ramp back-projection filter. The energy window of 99mTc was 140 keV ± 10%. The lung contour was drawn at a threshold of 20% of the maximum radioactivity of the lung.
Image and Data Analysis of Perfusion SPECT
For prediction of postoperative lung function with the aid of SPECT with 99mTc-microaggregated albumin, ROIs were placed over the lobe that was later resected and over both whole lungs by a chest radiologist with 5 years of experience. All ROI measurements were performed on the same PC with commercially available software (iNRT, Nihon Medi-Physics, Nisihinomiya, Japan). Each prediction of postoperative lung function (FEV1) assessed with SPECT was estimated with the following formula [24]:
\[\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\ }\mathrm{from\ perfusion}{\ }\mathrm{SPECT}=\mathrm{Postoperative}{\ }\mathrm{FEV}_{1}{\times}\left(1-\frac{\mathrm{Summed\ radioactivity\ within}{\ }\mathrm{ROIs}{\ }\mathrm{placed\ over\ resected\ lobe}}{\mathrm{Total\ lung\ volume\ activity}}\right)\]
Physiologic Index and Outcome Measures
Pulmonary function testing was performed according to American Thoracic Society standards with an automatic spirometer (System 9, Minato Ik-agaku). All subjects underwent preoperative and postoperative pulmonary function testing. All preoperative pulmonary function spirometric tests were performed within 2 weeks (mean, 6.4 days) before MRI. All postoperative pulmonary function spirometric tests were performed within 24-48 weeks (mean, 32 weeks) after surgery. Pulmonary function testing was performed according to American Thoracic Society standards [25, 26].
Statistical Analysis
To determine the utility of dynamic perfusion MRI in prediction of postoperative lung function, the correlation and limits of agreement between each version of predicted postoperative FEV1 and the actual postoperative FEV1 were statistically evaluated. Limits of agreement between actual and predicted postoperative FEV1 were analyzed by means of Bland-Altman analysis. Statistical significance was p < 0.05 for all analyses. The basic theory and application of the limits of agreement have been documented in the literature [27].
Results
A representative case is shown in Figure 1A, 1B, 1C, 1D, 1E. Correlation between the predicted postoperative FEV1 value for each method of measurement and the corresponding actual postoperative FEV1 (FEV1 values expressed as percentage of predicted value) is shown in Figure 2A, 2B, 2C, 2D. The postoperative FEV1 values predicted from perfusion MRI (r =0.87, r2 = 0.76, p < 0.0001) and quantitative CT assessment (r = 0.88, r2 = 0.77, p < 0.0001) had better correlation with actual postoperative FEV1 than did postoperative FEV1 predicted from qualitative CT assessment (r = 0.83, r2 = 0.69, p < 0.0001) and perfusion SPECT (r = 0.83, r2 = 0.69, p < 0.0001). Despite the differences in results between methods, all four versions of predicted postoperative FEV1 correlated well with actual postoperative FEV1.
The mean of difference and the limits of agreement between actual postoperative FEV1 and each predicted postoperative FEV1 are shown in Figure 3A, 3B, 3C, 3D. For postoperative FEV1 predicted from perfusion MRI, the mean and standard error were 5.3% ± 0.5% and the limits of agreement were between -6.5% and 17.1%. For postoperative FEV1 predicted from quantitative CT, the corresponding values were 5.0% ± 0.5%, and between -6.6% and 16.6%. For postoperative FEV1 predicted from qualitative CT, they were 6.8% ± 0.6%, and between -7.6% and 21.2%; and for postoperative FEV1 predicted from perfusion SPECT, 5.1% ± 0.6%, and between -8.9% and 19.1%.
Discussion
Our results show the utility of dynamic perfusion MRI in prediction of postoperative lung function after pulmonary resection in a large prospective cohort of patients with lung cancer. With this technique, regional pulmonary function can be assessed on the basis of pulmonary perfusion without radiation exposure [11, 14-19]. In addition, to the best of our knowledge, this study is the first in which the capability of dynamic perfusion MRI is compared directly with that of quantitative and qualitative CT and perfusion SPECT.
The comparison of actual and predicted postoperative FEV1 values showed that the correlation and limits of agreement between postoperative FEV1 predicted from perfusion MRI and the actual value were superior to those of postoperative FEV1 predicted from qualitative CT and perfusion SPECT and similar to those of postoperative FEV1 predicted from quantitative CT. Because the limits of agreement of dynamic contrast-enhanced perfusion MRI are considered small enough for clinical purposes, dynamic perfusion MRI can be substituted for perfusion SPECT and qualitative CT in prediction of postoperative lung function and can be considered as at least as useful as quantitative CT.
For prediction of postoperative lung function with dynamic perfusion MRI, regional and total functional lung volumes were assessed on the basis of regional perfusion assessed in a manner similar to the procedure for perfusion SPECT. Our results suggest that dynamic perfusion MRI is better than is perfusion SPECT for assessment of regional functional lung volume. This finding is based mainly on the following characteristics of dynamic perfusion MRI: higher spatial resolution for assessment of resected segments and lobes not including large vessels; less gravitational influence on regional perfusion parameters, because of the small molecular size and weight of gadopentetate dimeglumine; and greater accuracy of calculation of regional and total lung volumes, because of the use of regional pulmonary perfusion data obtained from the signal intensity-time course curve. Previous studies [11, 17] have shown that regional perfusion parameters semiquantitatively or quantitatively assessed with dynamic perfusion MRI had good or excellent correlation with those assessed with perfusion scintigraphy. It has been suggested [11, 17] that dynamic perfusion MRI has similar or slightly better utility than perfusion scintigraphy for prediction of postoperative lung function. Our results are compatible with the previous results, even when perfusion SPECT is substituted for perfusion scintigraphy. Perfusion scintigraphy with or without SPECT has been proposed as the reference standard for estimating postoperative lung function after lung resection in the treatment of lung cancer patients. However, perfusion scintigraphy may well be replaced by dynamic perfusion MRI in the near future.
For comparison with quantitative CT assessment, predicted postoperative FEV1 was calculated from regional and total functional lung volumes. The latter was determined by subtracting from the entire lung volume the nonfunctional lung volume resulting from pulmonary emphysema, tumor, atelectasis, and fibrosis identified by segmentation based on attenuation in the lung. Our results and those reported by others [9, 10, 20] indicate that this method can be considered the most accurate for prediction of postoperative lung function. The aforementioned pathologic conditions are assessed according to the underlying physiopathologic features by use of semiquantitatively or quantitatively assessed regional pulmonary perfusion parameters on dynamic perfusion MRI. Therefore, our results suggest that dynamic perfusion MRI is at least as valuable as quantitative CT for prediction of postoperative lung function.
In contrast to quantitative CT assessment, qualitative assessment of CT scans on the basis of number of bronchopulmonary segments is performed with a simple calculation method, and IV administration of contrast medium is not necessary. Many surgical institutions claim that this simple calculation provides a satisfactory estimate of operative risk. This claim is true for surgical candidates who are at low risk and have good results of preoperative pulmonary function tests, but we found that the correlation coefficient was lower and limits of agreement greater for qualitative CT assessment than for the other methods of prediction. These findings suggest that the greater overestimation and underestimation of postoperative FEV1 with this simple calculation than with other methods may be due to omission of evaluation of associated pulmonary emphysema, functional loss due to obstructive atelectasis and presence of nonfunctional lung area surrounding lung cancer, ventilation-perfusion mismatch due to invasion of the pulmonary vasculature by central lung cancer, the presence of hilar lymph node metastasis, and differences in size among various segments. When pulmonary resection is being considered in the treatment of high-risk surgical candidates, dynamic perfusion MRI or quantitative CT even with perfusion SPECT may be helpful for more accurate prediction of postoperative FEV1 in patients with non-small cell lung cancer.
There were limitations to this study. First, although all dynamic perfusion MRI examinations were successfully completed without adverse effects and regional perfusion was completely calculated from signal intensity-time course curves, 12 lung cancer patients with severe chronic obstructive pulmonary disease needed to breathe shallowly for data acquisition, resulting in slight deterioration of image quality. In lung cancer patients with poor pulmonary function, poor breath-hold capability can result in underestimation of regional perfusion and regional pulmonary function. However, because dynamic perfusion MRI is a new technique for estimation of postoperative lung function, advances leading to faster imaging time may make this technique more practical than earlier techniques.
A second limitation was that measurement of regional pulmonary perfusion parameters with the indicator dilution method is a semiquantitative assessment conducted with gadolinium contrast medium. Although indicator dilution theories are frequently used to determine regional blood volume and regional blood flow by means of various perfusion MRI techniques, direct application of these principles to contrast-enhanced, first-pass dynamic perfusion MRI experiments is difficult. Although regional blood volume can be determined by direct calculation of the area under the observed tissue-concentration curve, determination of calculated regional blood flow and mean transit time is less straightforward. Weisskoff et al. [28] pointed out that the use of the central volume principle to calculate mean transit time for locations within a tissue volume is inappropriate because changes in MRI signal intensity for specific tissue ROIs reflect the tracer concentration remaining in the tissue rather than that leaving the tissue. In addition, the observed signal intensity is assumed to be linearly related to the concentration of contrast medium within blood. Theoretically, this relation is not linear, but over a limited range of contrast concentrations it appears to be sufficiently linear to make the method valid [29]. Moreover, with use of the model it is assumed that the contrast agent remains within the intravascular space [30]. To the extent that the contrast agent acts as a purely intravascular marker, the volume of its distribution reflects blood volume. Therefore, despite the limitations, indicator dilution techniques with gadolinium contrast media can provide valid measures of regional pulmonary perfusion parameters.
A third limitation was associated with the fact that lung resection can improve the postoperative pulmonary function of patients with severe emphysema in the same way as does lung volume reduction surgery. The resultant beneficial effects on elastic recoil of the lung and chest wall mechanics make it difficult to accurately predict postoperative lung function [31, 32]. It is therefore necessary for lung cancer patients with severe emphysema to undergo exercise lung function tests before lung resection for evaluation of their cardiopulmonary reserve, even though dynamic perfusion MRI and other radiologic methods are more accurate in prediction of postoperative lung function.
A fourth limitation was that the physiologic index and outcome were assessed 24-48 weeks postoperatively, not at a specified time. This factor affected the determination of actual postoperative lung function and the statistical comparison of predicted and actual postoperative lung function. Fifth, for quantitative assessment of functional lung from chest CT images of the entire lung, we had to use 5-mm slices because of the performance limitations of the computer. It would have been better to use images obtained in thinner sections for assessment of pulmonary emphysema. This disadvantage is a potential limitation for preoperative assessment of regional and total functional lung volume and estimation of postoperative lung function by means of quantitative CT assessment.
A sixth limitation was that we used commercially available software for prediction of postoperative lung function with quantitative CT and perfusion SPECT but used proprietary software for calculation of regional perfusion with dynamic perfusion MRI. Several proprietary software packages for quantitative assessment of regional perfusion parameters have been described [11, 15, 16, 19, 33, 34]. In this study, we did not compare the prediction error of our software with that of other software packages in analysis of data from the same patients. Further investigations are needed to determine the real statistical significance of our software for prediction of postoperative lung function.
A seventh limitation of this study was that we compared capabilities for prediction of postoperative lung function among dynamic perfusion MRI, quantitative and qualitative CT, and perfusion SPECT. We did not compare capabilities for prediction of other pulmonary functional parameters, such as static lung volume, diffusing capacity for oxygen, survival, quality of life, and cost-effectiveness. A prospective comparative study with a larger number of subjects may be warranted to determine the real importance in routine clinical practice of dynamic perfusion MRI as a substitute for quantitative and qualitative CT and perfusion SPECT or integrated SPECT and CT for prediction of the postoperative lung function of lung cancer patients.
In conclusion, dynamic perfusion MRI can be used to predict the postoperative lung function of lung cancer patients more accurately than qualitative CT and perfusion SPECT and at least as accurately as quantitative CT.
Acknowledgments
We thank Yoshiyuki Ohno, Professor Emeritus, Nagoya University (Department of Preventive Medicine, Graduate School of Medicine) for his advice on the statistical component of this study. We also thank Yoshimasa Maniwa (Division of Cardiovascular, Thoracic and Pediatric Surgery, Kobe University Graduate School of Medicine), and Yoshihiro Nishimura (Division of Cardiovascular and Respiratory Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine) for their contributions to this study.
Footnotes
Address correspondence to Y. Ohno ([email protected]).
Presented at the 2006 annual meeting of the Radiological Society of North America, Chicago, Illinois.
Partially supported by the Knowledge Cluster Initiative of the Ministry of Education, Culture, Sports, Science and Technology of Japan; Philips Medical Systems; and Schering AG.
References
1.
Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancer statistics, 2001. CA Cancer J Clin 2001; 51:15-36
2.
Shields TW. Surgical therapy for carcinoma of the lung. Clin Chest Med 1993; 14:121-147
3.
Pierce RJ, Copland JM, Sharpe K, Barter CE. Preoperative risk evaluation for lung cancer resection: predicted postoperative product as a predictor of surgical mortality. Am J Respir Crit Care Med 1994; 150:947-955
4.
Bolliger CT, Jordan P, Soler M, et al. Exercise capacity as a predictor of postoperative complications in lung resection candidates. Am J Respir Crit Care Med 1995; 151:1472-1480
5.
Wyser C, Stulz P, Soler M, et al. Prospective evaluation of an algorithm for the functional assessment of lung resection candidates. Am J Respir Crit Care Med 1999; 159:1450-1456
6.
Olsen GN, Block AJ, Tobias JA. Prediction of post-pneumonectomy pulmonary function using quantitative macroaggregate lung scanning. Chest 1974; 66:13-16
7.
Markos J, Mullan BP, Hillman DR, et al. Preoperative assessment as a predictor of mortality and morbidity after lung resection. Am Rev Respir Dis 1989; 139:902-910
8.
Giordano A, Calcagni ML, Meduri G, Valente S, Galli G. Perfusion lung scintigraphy for the prediction of postlobectomy residual pulmonary function. Chest 1997; 111:1542-1547
9.
Bolliger CT, Guckel C, Engel H, et al. Prediction of functional reserves after lung resection: comparison between quantitative computed tomography, scintigraphy, and anatomy. Respiration 2002; 69:482-489
10.
Wu MT, Pan HB, Chiang AA, et al. Prediction of postoperative lung function in patients with lung cancer: comparison of quantitative CT with perfusion scintigraphy. AJR 2002; 178:667-672
11.
Ohno Y, Hatabu H, Higashino T, et al. Dynamic perfusion MRI versus perfusion scintigraphy: prediction of postoperative lung function in patients with lung cancer. AJR 2004; 182:73-78
12.
Imaeda T, Kanematsu M, Asada S, et al. Prediction of pulmonary function after resection of primary lung cancer: utility of inhalation-perfusion SPECT imaging. Clin Nucl Med 1995; 20:792-799
13.
Piai DB, Quagliatto R Jr, Toro I, Cunha Neto C, Etchbehere E, Camargo E. The use of SPECT in preoperative assessment of patients with lung cancer. Eur Respir J 2004; 24:258-262
14.
Hatabu H, Gaa J, Kim D, Li W, Prasad PV, Edelman RR. Pulmonary perfusion: qualitative assessment with dynamic contrast-enhanced MRI using ultra-short TE and inversion recovery turbo FLASH. Magn Reson Med 1996; 36:503-508
15.
Hatabu H, Tadamura E, Levin DL, et al. Quantitative assessment of pulmonary perfusion with dynamic contrast-enhanced MRI. Magn Reson Med 1999; 42:1033-1038
16.
Levin DL, Chen Q, Zhang M, Edelman RR, Hatabu H. Evaluation of regional pulmonary perfusion using ultrafast magnetic resonance imaging. Magn Reson Med 2001; 46:166-171
17.
Iwasawa T, Saito K, Ogawa N, Ishiwa N, Kurihara H. Prediction of postoperative pulmonary function using perfusion magnetic resonance imaging of the lung. J Magn Reson Imaging 2002; 15:685-692
18.
Fink C, Puderbach M, Bock M, et al. Regional lung perfusion: assessment with partially parallel three-dimensional MR imaging. Radiology 2004; 231:175-184
19.
Ohno Y, Hatabu H, Murase K, et al. Quantitative assessment of regional pulmonary perfusion in the entire lung using three-dimensional ultrafast dynamic contrast-enhanced magnetic resonance imaging: preliminary experience in 40 subjects. J Magn Reson Imaging 2004; 20:353-365
20.
Ohno Y, Hatabu H, Higashino T, et al. Oxygen-enhanced MR imaging: correlation with postsurgical lung function in patients with lung cancer. Radiology 2005; 236:704-711
21.
Beccaria M, Corsico A, Fulgoni P, et al. Lung cancer resection: the prediction of postsurgical outcomes should include long-term functional results. Chest 2001; 120:37-42
22.
Zeiher BG, Gross TJ, Kern JA, Lanza LA, Peterson MW. Predicting postoperative pulmonary function in patients undergoing lung resection. Chest 1995; 108:68-72
23.
Kearney DJ, Lee TH, Reilly JJ, DeCamp MM, Sugarbaker DJ. Assessment of operative risk in patients undergoing lung resection: importance of predicted pulmonary function. Chest 1994; 105:753-759
24.
Suga K, Kawakami Y, Zaki M, Yamashita T, Shimizu K, Matsunaga N. Clinical utility of co-registered respiratory-gated(99m)Tc-Techne-gas/MAA SPECT-CT images in the assessment of regional lung functional impairment in patients with lung cancer. Eur J Nucl Med Mol Imaging 2004; 31:1280-1290 25.
25.
[No authors listed]. Standardization of spirometry: 1987 update—statement of the American Thoracic Society. Am Rev Respir Dis 1987; 136:1285-1298
26.
[No authors listed]. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis 1991; 144:1202-1218
27.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307-310
28.
Weisskoff RM, Chesler D, Boxerman JL, Rosen BR. Pitfalls in MR measurement of tissue blood flow with intravascular tracers: which mean transit time? Magn Reson Med 1993; 29:553-558
29.
Jerosch-Herold M, Wilke N, Stillman AE. Magnetic resonance quantification of the myocardial perfusion reserve with a Fermi function model for constrained deconvolution. Med Phys 1998; 25:73-84
30.
Donahue KM, Weisskoff RM, Burstein D. Water diffusion and exchange as they influence contrast enhancement. J Magn Reson Imaging 1997; 7:102-110
31.
Korst RJ, Ginsberg RJ, Ailawadi M, et al. Lobectomy improves ventilatory function in selected patients with severe COPD. Ann Thorac Surg 1998; 66:898-902
32.
Sinjan EA, Van Schil PE, Ortmanns P, Van den Brande F, Hendriks JM, Eyskens E. Improved ventilatory function after combined operation for pulmonary emphysema and lung cancer. Int Surg 1999; 84:185-189
33.
Nikolaou K, Schoenberg SO, Brix G, et al. Quantification of pulmonary blood flow and volume in healthy volunteers by dynamic contrast-enhanced magnetic resonance imaging using a parallel imaging technique. Invest Radiol 2004; 39:537-545
34.
Risse F, Semmler W, Kauczor HU, Fink C. Dualbolus approach to quantitative measurement of pulmonary perfusion by contrast-enhanced MRI. J Magn Reson Imaging 2006; 24:1284-1290
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: December 5, 2006
Accepted: February 28, 2007
First published: November 23, 2012
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.