| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
Interact CardioVasc Thorac Surg 2008;7:1019-1023. doi:10.1510/icvts.2008.176420 © 2008 European Association of Cardio-Thoracic Surgery
Clinical parameters affecting prediction accuracy of postoperative lung function in non-small cell lung cancer
a Division of Pulmonary, Allergy and Critical Care Medicine of Hallym University Sacred-Heart Hospital, 896 Pyungchon-dong, Anyang, Gyeonggi-do 431-070, Republic of Korea Received 25 January 2008; received in revised form 9 June 2008; accepted 20 July 2008
Corresponding author. Tel.: +82-31-380-3718; fax: +82-31-380-3973.
Despite significant development in chemotherapy and radiotherapy, surgery is still the cornerstone for curative lung cancer treatment. Accurate prediction of postoperative lung function is mandatory. The goal of this study was to identify important clinical factors affecting prediction accuracy of postoperative lung function for more careful patient selection. The medical records of non-small cell lung cancer patients undergoing pulmonary resection were reviewed. An accuracy index, apo/ppoFEV1 was defined as the ratio of actual postoperative FEV1 [apoFEV1] to predicted postoperative FEV1 [ppoFEV1]. We used multivariate analysis to inspect the relationship between the accuracy index and seven tentative clinical factors: age, gender, preoperative FEV1, time interval between operation and the first postoperative FEV1, bronchodilator response (%), resected lung portion, and the number of resected lung segments. A total of 82 patients were analyzed. Accuracy index of quantitative perfusion lung scan-based prediction was better than that of simple calculation. Multivariate analysis identified the number of resected lung segments and preoperative FEV1 as the significant clinical factors affecting the accuracy index (P=0.026 and 0.002, respectively). Preoperative FEV1 and the number of resected lung segments are significant clinical factors affecting prediction accuracy of postoperative lung function.
Key Words: Lung function; Prediction accuracy; Lung cancer
Despite significant development in chemotherapy and radiotherapy, surgery is still the cornerstone for curative lung cancer treatment. However, since many lung cancer patients have cardiopulmonary comorbidities, precise prediction of postoperative residual lung function is mandatory for a safe operation [1, 2]. Predicted postoperative forced expiratory volume in 1 s (ppoFEV1) has been regarded as the most helpful indicator for this purpose [1–4]. To prevent respiratory crippling after operation, postoperative residual FEV1 should be over 0.8 l or more than 40% of the predicted value [5, 6]. Three methods are being used for calculation of ppoFEV1: quantitative perfusion lung scan, simple calculation, and CT-based prediction. Among them, perfusion lung scan-based prediction has been the gold standard for many years irrespective of the extent of resection [1, 2, 7]. The goal of this study was to identify clinical factors affecting prediction accuracy of postoperative lung function for more careful selection of operable lung cancer patients.
2.1. Patients We reviewed the medical records of non-small cell lung cancer patients who underwent radical lung resection for curative intent between January 1999 and June 2005 at Hallym University Sacred Heart Hospital. Preoperative FEV1, postoperative FEV1 and quantitative perfusion lung scans were requisite for the case selection. All the patients were routinely encouraged to do incentive spirometry from one week before to two weeks after the operation. The cases having neoadjuvant or adjuvant chemotherapy or radiotherapy, bulla larger than 3 cm on chest CT, and benign bronchial strictures proximal to the segmental bronchus were excluded. Lung function tests were performed using the SensorMedics V-Max 22 (SensorMedics Corp., Yorba Linda, CA, USA). Tests were performed at least two times, and the best values were selected for analysis. The prebronchodilator lung function values were adopted for data uniformity. The preoperative lung function test was performed within two weeks before surgery, and the follow-up test was executed between 14 and 40 days after the operation. If the third lung function test at 90 days or later after surgery was available, we considered it as plateau lung function because postoperative lung functions become clinically stable and reach plateau level at this period. 2.3. Quantitative perfusion lung scan A dual-head gamma camera (E. Cam, Siemens, Germany) coupled with a low-energy general purpose collimator was used. Five minutes after intravenous injection of (99m) Tc-macroaggregated albumin in a 0.5 ml volume, multiple static images in the anterior, posterior, right anterior oblique, left anterior oblique, right posterior oblique, left posterior oblique, right lateral, and left lateral positions were acquired. Quantitative analysis was performed computing the left-to-right ratio from the averages between anterior and posterior imaging.2.4. Calculation of predicted postoperative FEV1 Predicted postoperative FEV1 was calculated by two methods: simple calculation and perfusion lung scan-based calculation. The simple calculation method introduced by Juhl et al. [8] assumes that the right lung is composed of 10 segments (3 segments in the upper lobe, 2 segments in the middle lobe, and 5 segments in the lower lobe), the left lung of 9 segments (4 segments in the upper lobe, 5 segments in the lower lobe), and that all the segments contribute equally to lung function. Thus, ppoFEV1 can be calculated by simple subtraction of the FEV1 proportion of resected lung segments from the preoperative FEV1: ppoFEV1=preoperative FEV1x[1–(number of resected segments/19)]. Perfusion lung scan-based prediction follows two steps. The first step determines the functional contribution of the resected lung by scan, and the second step applies the principle of simple calculation: ppoFEV1=preoperative FEV1x[1–functional fraction of the resected lungx(number of resected segments/total segments of that lung)].We adopted an accuracy index, apo/ppoFEV1 (the ratio of actual postoperative FEV1 [apoFEV1] to ppoFEV1) for comparison of each method and identification of clinical parameters affecting prediction accuracy of postoperative lung function. The closer the index is to one, the more accurate the lung function prediction is. 2.5. Identification of clinical parameters affecting prediction accuracy of postoperative lung function The accuracy index was investigated in terms of seven tentative clinical factors: age, gender, preoperative FEV1, time interval between operation and the date of postoperative FEV1, bronchodilator response (%), resected lung portion (upper or lower), and the number of resected lung segments. Preoperative bronchodilator response (%) was analyzed only if available. The patients were divided into two groups according to resected lung portion. The upper portion resection group included all cases of upper, middle lobectomy and upper-middle bilobectomy. The lower portion resection group included all cases of lower lobectomy or middle-lower bilobectomy [9]. To evaluate the effects of airway obstruction, patients were classified as COPD group and non-COPD group based on the criteria of FEV1/FVC <70% and FEV1 <80%. An independent sample t-test was used to compare means of continuous variables. If the case number was <30, the Mann–Whitney U-test was used. The correlation between each clinical parameter and accuracy index was screened on bivariate analysis of Pearson's correlation coefficient, and significant candidate factors were confirmed by multivariate linear regression analysis. A P-value <0.05 was considered to be significant in all statistical analyses. Statistical analysis used the SPSS for Windows statistical software package (SPSS Inc., version 13.0, Chicago, IL).
3.1. Patient characteristics The baseline characteristics of the patients are summarized in Table 1. A total of 82 were eligible for analysis. Fifty-two were men and 30 were women. The mean age was 61.3±10.5 years. Lung functions were tested at 9.2±5.4 days before operation and at 24.2±7.1 days after operation. The third lung function test for the plateau was available in 46 patients, and these were performed at 105.7±30.2 days after operation. The mean preoperative FEV1 was 91.6±20.8% of predicted. Twenty-four patients (29.3%) were in the COPD group and 58 patients (70.7%) were in the non-COPD group.
3.2. Prediction accuracy of postoperative lung function Perfusion lung scan-based prediction was more accurate than simple calculation (apo/ppoFEV1=1.00±0.19 vs. 1.07±0.23, P<0.001). Simple calculation tended to underestimate ppoFEV1 by 7% compared to apoFEV1. As expected, plateau FEV1 (plateau-apoFEV1) measured at three months or later after surgery was increased by 13% compared to apoFEV1. In predicting plateau lung function, perfusion lung scan-based prediction also proved superior to simple calculation (1.11±0.24 vs. 1.18±0.30, P<0.001). These also indicate that the known prediction methods are fitter for short-term follow-up than long-term follow-up. However, in nine patients (19%) among 46 patients whose postoperative FEV1 at three months or later was available, the plateau-apoFEV1 was lower than the apoFEV1. Their preoperative bronchodilator response values (%) were higher than those of the others although it did not reach statistical significance (11.2±8.40% vs. 7.0±6.8%, P= 0.11). 3.3. Clinical parameters affecting prediction accuracy of postoperative lung function Of the seven tentative clinical factors mentioned above, bivariate analysis identified gender, preoperative FEV1, and the number of resected lung segments to be related with the accuracy index (Figs. 1 and 2). Multivariate analysis confirmed preoperative FEV1 and the number of resected lung segments as the significant clinical parameters affecting accuracy index (Table 2). It was notable that the corresponding means of apo/ppoFEV1 for each number of resected lung segments gathered on a straight line of a constant slope (Fig. 2). However, resected lung portion (upper or lower portion resection) was not related to the accuracy index (P=0.10) though lower portion resection was significantly higher than upper portion resection in the number of resected lung segments (5.6±1.1 vs. 3.7±1.0, P<0.001).
3.4. Comparison of the accuracy index according to airway obstruction Preoperative FEV1 was significantly lower in the COPD group than in the non-COPD group (67.1%±8.3% and 101.6%±15.2%, respectively, P<0.001). Actual postoperative FEV1 was about 14% larger than ppoFEV1 in the COPD group while apoFEV1 was very similar to ppoFEV1 in the non-COPD group (apo/ppoFEV1=1.14±0.2 vs. 0.99±0.1, P<0.001). These results indicate that the prediction of postoperative lung function is more accurate in the non-COPD group than in the COPD group, and postoperative lung function of the COPD group may be less deteriorated than that of the non-COPD group. Three were no clinical factors related to prediction accuracy in the COPD group. However, in the non-COPD group, the number of resected lung segments and bronchodilator response (%) were significantly related to the accuracy index (P=0.014 and 0.047, respectively).
Accurate prediction of postoperative residual lung function is mandatory to minimize postoperative morbidity and mortality. As expected, perfusion lung scan-based prediction was a more accurate method than simple calculation for both early and plateau postoperative FEV1, even though it was more accurate for early FEV1. This is consistent with previous data that perfusion scintigraphy-based postoperative lung function correlated best with actual postoperative lung function irrespective of the extent of resection [1, 3, 8, 10]. The relative inaccuracy of predicting plateau lung function may be due to the continuous increase of postoperative lung function up to 3–6 months [4, 11]. However, in some patients, long-term follow-up plateau FEV1 was somewhat decreased. Intriguingly, their preoperative bronchodilator response values were higher than those of the others. This suggests that adequate perioperative bronchodilator therapy is necessary for those patients who have marginal lung function with high bronchodilator response. Recently, Dragan et al. [12] also demonstrated similar findings that postoperative FEV1 and small airways function significantly improved after bronchodilator therapy in 35 patients with COPD undergoing pulmonary resection. Even if quantitative perfusion lung scan may provide useful information, clinical factors should be considered for more accurate prediction. In our study, preoperative FEV1 and the number of resected lung segments were identified as the significant clinical factors to affect prediction accuracy. However, the prediction accuracy was not related to resected lung portion. This was somewhat contrary to the published data that accurate prediction of postoperative FEV1 was influenced by resected lung portion i.e. lower portion lobectomy was an independent factor for the minimal deterioration of postoperative FEV1. The authors explained the phenomenon as anatomic repositioning following upper lung portion resection, which induce narrowing of lower and middle lobe bronchus and resultant decrement of apo/ppoFEV1 [9]. The discordance between these two studies may originate from the fact that the number of resected lung segments was significantly higher in lower portion lobectomy than in upper portion lobectomy. The prediction method was more reliable in the non-COPD group than in the COPD group. This could be explained by the concept of volume reduction surgery as increments of postoperative lung function by resection of relatively functionless emphysematous lungs in the COPD group [11–15]. This study has several limitations. The patients were enrolled retrospectively, so we could not know the precise nature of their general physiologic conditions, and we only analyzed patients who had been discharged after surgery. Thus, we could not include cases with early postoperative mortality. Moreover, many patients lacked long-term follow-up lung functions, which weakened statistical reliability. In conclusion, actual postoperative FEV1 can be more accurately predicted by performing perfusion lung scan and should be modified according to preoperative FEV1 and the number of resected lung segments. The actual postoperative FEV1 tended to be larger than expected in patients with smaller preoperative FEV1 or larger lung resection. Additionally, adequate postoperative pulmonary care such as continuous bronchodilator therapy may be important to maintain lung function, especially in patients with marginal lung function and large bronchodilator response.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
Top
Abstract
1. Introduction
