Patients with bronchogenic carcinoma frequently have impaired pulmonary function, usually secondary to chronic airway obstruction [1]. If these patients are referred for possible curative surgery, they are at increased risk of developing postoperative complications, and some have such poor respiratory reserve that a pneumonectomy may result in an unacceptable quality of life. Numerous techniques have been used to evaluate the postsurgical risk. These include pulmonary function tests (PFTs), exercise tests, and quantitative ventilation-perfusion (V/Q) scintigraphy.

Recent American College of Chest Physicians (ACCP) guidelines [2], as well as those from equivalent European organizations such as the British Thoracic Society (BTS) [3], recommend using quantitative Q scintigraphy in patients with borderline lung function (forced expiratory volume at 1 second [FEV₁] < 2.0 L) to predict the postoperative lung function in lung cancer patients who are considering undergoing pneumonectomy. The evidence base for the ACCP and others' guidelines comes from previous V/Q scintigraphy data, much of which were obtained between 10 and 30 years ago from studies that used small numbers of participants (usually fewer than 20), with populations significantly younger and with a male predominance compared with the patients who presented for curative surgery in 2005.

The correlation between the actual and predicted postoperative FEV₁ using quantitative V/Q scintigraphy has been variable, with correlative figures quoted between r = 0.67 and r = 0.9 [4-11]. Many of these investigators, however, found that the scintigraphically predicted postoperative FEV₁ was lower than the actual postoperative FEV₁ [4, 8, 9]. Also in the literature, we found inconsistencies in terminology and in the choice of the most useful postoperative measurement. Some investigators use scintigraphy to predict the FEV₁ after pneumonectomy [8, 9], whereas others quote the predicted percentage of FEV₁ (the FEV₁ value predicted for a patient of a given height, sex, and age) [12-16]. Moreover, we found no clear evidence as to whether it is better to use a V scan or a Q scan alone or in combination. We therefore revisited the use of V/Q scintigraphy in pneumonectomy patients for predicting postoperative pulmonary function and the appropriateness of present clinical guidelines.

Over 24 months, we prospectively performed pre- and postsurgical PFTs and quantitative planar V/Q scintigraphy on 32 patients (21 men and 11 women; mean age, 68.4 years; range, 42-85 years) who were undergoing pneumonectomy for non-small cell lung cancer. All patients were former smokers and had preoperative peak oxygen consumption (VO_{2 peak}) greater than 10 mL/kg/min. Twelve out of 32 had FEV₁ less than 2.0 L, connoting borderline lung function. Twenty-one patients underwent left pneumonectomy, and 11 underwent right pneumonectomy. All pneumonectomies were performed using a standard posterolateral thoracotomy technique by one of three designated cardiothoracic surgeons. All patients gave written consent, and the study had local ethical committee approval.

The mean preoperative FEV₁ was 2.13 L (SD, 0.38 L; range, 1.2-3.3 L), and the mean actual postoperative FEV₁ was 1.36 L (SD, 0.36 L; range, 0.8-2.5 L). The mean preoperative percentage of predicted FEV₁ was 85% (SD, 16%; range, 42-119%), and the mean actual postoperative percentage of predicted FEV₁ was 53% (SD, 13%; range, 28-90%).

One death (secondary to tumor reoccurrence) occurred 6 months after surgery. None of the 31 surviving patients suffered from respiratory insufficiency during this period.

The presence of bronchogenic carcinoma causes a decrease in both ventilation and perfusion of the involved lung, especially if the lesions are in the hilar region. The lung function will further diminish with loss of lung parenchyma after a pneumonectomy procedure. Radionuclide lung quantification offers a simple method of predicting the loss in lung air capacity with quantitative measurement of differential ventilation or perfusion [13, 18]. Our study contained an older population with more female patients compared with previously reported studies and, therefore, reflects current practice. Our study population was considerably larger than that used in most previous studies [4-9] (Table 5) and also contained patients with lower FEV₁ values than has been investigated previously, reflecting the impact of modern surgical techniques.

The correlation between actual and predicted values was significant for both FEV₁ and percentage of predicted FEV₁ values. The correlation was not as good as that previously reported [4-9] (Table 5). The highest reported correlation was obtained from V/Q SPECT quantification [10]. Although the American College of Chest Physicians [2] and the British Thoracic Society [3] recommend using perfusion scintigraphy only, we found that the ventilation scan offered the best correlation, a finding that has been previously reported [17]. Most likely the guidelines favor perfusion scintigraphy because more reports have been published on perfusion scintigraphy [6, 7, 9, 19, 20] than on ventilation scintigraphy. The use of ventilation scintigraphy alone has the advantages of requiring a lower radiation dose and avoiding the use of needles. We also found that no added predictive advantage was gained from combining both ventilation and perfusion studies, a finding that also has been previously noted [17]. Although ventilation scintigraphy offered better correlation than perfusion scintigraphy (r = 0.7 vs 0.58), a statistically significant difference between the two tests was not seen. Therefore, it would still be acceptable to perform perfusion scintigraphy alone as is done in traditional practice.

In keeping with previous findings, we showed that scintigraphy usually predicted lower postoperative FEV₁ values than what actually occurred. This tendency also has been reported by other investigators [4, 8]. This finding has potentially serious implications in deciding patients' fitness for lung cancer surgery, because underestimating patients' postoperative lung function may unnecessarily preclude them from curative surgery. Our experience suggests that this underestimation occurs both when scintigraphy is used to predict the postoperative FEV₁ and when it is used to determine the percentage of predicted FEV₁, although we found the latter marginally more accurate.

Improvements in the correlation between scintigraphy and pulmonary function tests may have been found by using more sophisticated techniques such as SPECT. This is a tomographic technique (similar to the principles of CT) where the gamma camera rotates around the patient's body, acquiring images in different positions. SPECT gives better target-to-nontarget ratios than do planar images. This, in turn, affords improved resolution in any given plane, better definition of shape, and accurate quantification. However, because multiple camera projections are required, imaging time is much longer than for a single planar image. Unfortunately, many institutions do not have enough camera time capacity to perform this on all such patients.

Our technique could have been further enhanced by the use of ventilation gases (e.g., ^99mTc gas) rather than aerosol. Aerosols, being larger and heavier particles than gases, are more likely to undergo central airway clumping. This is especially so in patients with obstructive airway disease. Technetium-99m aerosol is used because it is a cheap and readily available tracer and does not require complicated delivery systems. It is reassuring to note that in our series, only six patients had criteria for airway obstruction and only one had evidence of clumping. Technetium-99m gas is not only more expensive, but also requires more sophisticated delivery systems because very high concentrations of tracers are needed in tiny volumes. For these reasons, our institution has greater expertise in and prefers to use aerosols. Nonetheless, our use of aerosols may, in part, have explained why our correlation figures were not as high as previously reported (Table 5). Probably the best ventilation agent would have been krypton-81m (^81mKr) [10], but the availability of this agent remains limited.

In conclusion, in certain lung cancer patients being considered for pneumonectomy, and in contradiction to current American College of Chest Physician and British Thoracic Society guidelines, we recommend the use of ventilation scintigraphy alone to predict postoperative FEV₁. However, it is important to note that ventilation scintigraphy tends to underestimate actual postoperative lung function and, if this underestimation is not fully appreciated, some patients may unnecessarily be deprived of curative surgery.

We thank Linda Sharples, PhD, Cambridge MRC, for her invaluable statistical input and Prof. Adrian Dixon, Radiology Department, University of Cambridge, for his input.

The study received funding from a National Health Service Research and Development Grant.

Address correspondence to T. Win ([email protected]).