**3. The methods for prediction of postoperative lung function**

The surgical procedure depends on the stage of lung cancer and on the cardiopulmonary reserve of the patients. A prospective randomized trial comparing limited resection to lobectomy in patients with peripheral stage I lung cancers reported that the patients treated with limited resection had a three-fold increase in local recurrence, a 75% increase in combined local and distant recurrence, and a 50% increase in death with cancer (Ginsberg & Rubinstein 1995). Therefore, anatomic resection such as lobectomy or pneumonectomy should be done if physiologically feasible. Lung sparing anatomic resection like sleeve lobectomy is preferred over pneumonectomy, if anatomically appropriate and if marginnegative resection can be achieved. A study compared clinical outcomes of the elderly patients undergoing sleeve lobectomy or pneumonectomy due to non-small-cell lung cancer (Bolukbas, et al. 2011). The loss of FEV1 was 12.0% vs. 27.3% (p = 0.001), and 5 year survival rate was 59% vs. 0% favoring sleeve lobectomy although there was no statistical difference in postoperative mortality (6.5% in sleeve lobectomy vs. 10.3% in pneumonectomy). Sublobar resection, either segmentectomy (preferred) or wedge resection, is appropriate in selected patients if margin-negative resection can be achieved. The limited resection is appropriate especially for the patients with poor pulmonary reserve or other major comorbidity that contraindicates lobectomy.

For the prediction of postoperative remnant lung function, the anatomy of the lung should be understood. The right lung consists of three lobes; upper, middle and lower lobe. The right upper lobe consists of 3 segments, the right middle lobe 2 segments, and

Prediction of Postoperative Lung Function 255

Quantitative computed tomography scanning has been studied as a technique to estimate postresection lung function. The basic concept is similar to radionuclide perfusion scanning method. This measures the split lung function using the CT attenuation density instead of radionuclide signal intensity. The volume of lung with attenuation between -500 and -910 Hounsfield units was used to estimate functional lung volume. The portion of the lung remaining postresection was predicted by calculating lung volume in the area to be resected as a portion of total lung volume. With this, predicted postoperative function correlated as well as the method using radionuclide quantitative perfusion imaging (Wu, et al. 2002).

A study showed that magnetic resonance (MR) perfusion imaging had almost the same sensitivity and specificity for diagnosis of pulmonary perfusion defects as conventional perfusion scintigraphy (Berthezene, et al. 1999). The regional lung function is calculated from the subtraction images for normal lung parenchyma using image analysis software. The accuracy of MR perfusion for the prediction of ppoFEV1 was validated in a study (Iwasawa, et al. 2002). This study demonstrated that the correlation between perfusion ratios derived from MR perfusion image and radionuclide perfusion scanning was excellent (R = 0.92). The correlation between ppoFEV1 and actual postoperative FEV1 was similar when the two methods were compared (R = 0.682 in MR perfusion and R = 0.667 in radionuclide perfusion).

**4. Clinical parameters affecting prediction accuracy of postoperative lung** 

Prediction accuracy of postoperative lung function is affected not only by the calculation technique, but also the clinical factors associated with the actual postoperative lung function. The actual lung function closely relates to the physiology of lung volume reduction, reversibility of airway obstruction, pharmacotherapy, and postoperative

Pulmonary function is affected by lung resection and the decline in lung function varies with the extent of the resection. The degree of functional loss appears to be less in individuals with poor baseline lung function uniformly across the studies (Bobbio, et al. 2005; Boushy, et al. 1971; Edwards, et al. 2001). In patients with severe emphysema, surgery performed to remove the most emphysematous portion of the lung may lead to improvements in lung function (Fishman, et al. 2003). A case-matched study demonstrated that the patients with COPD had a three-fold higher rate of cardiopulmonary morbidity (28% versus 10%, p=0.04), but lower reduction in FEV1 (6% versus 13%, p=0.0002) compared with non-COPD patients after lobectomy for lung cancer (Pompili, et al. 2010). Importantly, although the postoperative quality of life in both groups was reduced, there were no significant differences in quality of life between the groups. This suggests that the patients with lung cancer and COPD may be unexpectedly tolerable with the curative lung resection if the candidates are carefully selected. This is attributed to lung volume reduction effect which takes place very early. The risk-benefit should be balanced based on the negative physiologic effects of thoracotomy versus positive effects of lung volume reduction in the

**3.4 Dynamic perfusion magnetic resonance imaging (MRI)** 

**3.3 Quantitative CT scanning** 

**function** 

respiratory rehabilitation.

**4.1 Chronic obstructive pulmonary sisease** 

the right lower lobe 5 segments. The left lung consists of two lobes; upper and lower lobe. The left upper lobe consists of 4 segments, the left lower lobe 4 segments. The anteromediobasal segment of the left lower lobe is the counterpart of anterobasal segment and mediobasal segment of the right lower lobe, but it has a common bronchial orifice. Although it is anatomically one segment, it contains two segments in volume. Therefore, the left lung is regarded to have 9 segments in terms of volume; 4 segments in the left upper lobe and 5 segments in the left lower lobe. Each segment is assumed to have same volume (1/19 of the lung function). The four validated methods to predict postoperative lung function are: 1) anatomic calculation, 2) split radionuclide perfusion scanning, 3) quantitative computed tomography (CT) scanning, and 4) dynamic perfusion magnetic resonance imaging (MRI). Using these techniques, the actual lung function was consistently underestimated, particularly if the starting value was lower (Giordano, et al. 1997; Zeiher, et al. 1995). The accuracy of prediction of anatomic calculation is slightly lower than the other methods, but the other methods effectively predict postoperative FEV1 with similar accuracy. Medical cost, local expertise and availability of equipments should be considered for the choice.

#### **3.1 Anatomic calculation**

Estimation of predicted postoperative lung function based on anatomical calculation is the simplest, but the accuracy is slightly lower than other methods. Predicted postoperative FEV1 can be calculated by simple subtraction of the FEV1 proportion of resected lung segments from the total preoperative FEV1 (Juhl & Frost 1975; Zeiher, et al. 1995).

ppoFEV1 = preoperative FEV1 x [1 - (number of segments to be resected 19)]

For example, if a patient will undergo right upper lobectomy, the predicted postoperative FEV1 will be calculated to remain 16/19 (84.2%) of preoperative FEV1. This method is also applied to the other methods as a basic concept.

#### **3.2 Split radionuclide perfusion scanning**

In most clinical cases, the extent of pulmonary disease may be different in each side of the lung, which is leading cause of different regional lung function according to the disease status. This is a major violation of the assumption that each lung segment represents the same lung function. The general principle of the radionuclide scanning technique is same for the anatomic calculation method. The prediction of postoperative lung function can be calculated by the following two steps. The first step is determining the functional contribution of resected (right or left) lung by quantitative perfusion lung scan, and the second step is applying the principle of anatomical calculation to the resected lung. Postoperative lung function is then estimated to be the product of the preoperative function and the portion of lung function that will remain after resection as estimated by the scan.

ppoFEV1 = preoperative FEV1 x [1 – functional fraction of the resected lung

(number of segments to be resected/total segments of that lung)]

The accuracy of these techniques has been questioned. A recent study using technetium scanning calculated values of imprecision from 18%-21% despite showing reasonable correlation (Giordano, et al. 1997).

#### **3.3 Quantitative CT scanning**

254 Topics in Cancer Survivorship

the right lower lobe 5 segments. The left lung consists of two lobes; upper and lower lobe. The left upper lobe consists of 4 segments, the left lower lobe 4 segments. The anteromediobasal segment of the left lower lobe is the counterpart of anterobasal segment and mediobasal segment of the right lower lobe, but it has a common bronchial orifice. Although it is anatomically one segment, it contains two segments in volume. Therefore, the left lung is regarded to have 9 segments in terms of volume; 4 segments in the left upper lobe and 5 segments in the left lower lobe. Each segment is assumed to have same volume (1/19 of the lung function). The four validated methods to predict postoperative lung function are: 1) anatomic calculation, 2) split radionuclide perfusion scanning, 3) quantitative computed tomography (CT) scanning, and 4) dynamic perfusion magnetic resonance imaging (MRI). Using these techniques, the actual lung function was consistently underestimated, particularly if the starting value was lower (Giordano, et al. 1997; Zeiher, et al. 1995). The accuracy of prediction of anatomic calculation is slightly lower than the other methods, but the other methods effectively predict postoperative FEV1 with similar accuracy. Medical cost, local expertise and availability of equipments

Estimation of predicted postoperative lung function based on anatomical calculation is the simplest, but the accuracy is slightly lower than other methods. Predicted postoperative FEV1 can be calculated by simple subtraction of the FEV1 proportion of resected lung

ppoFEV1 = preoperative FEV1 x [1 - (number of segments to be resected 19)] For example, if a patient will undergo right upper lobectomy, the predicted postoperative FEV1 will be calculated to remain 16/19 (84.2%) of preoperative FEV1. This method is also

In most clinical cases, the extent of pulmonary disease may be different in each side of the lung, which is leading cause of different regional lung function according to the disease status. This is a major violation of the assumption that each lung segment represents the same lung function. The general principle of the radionuclide scanning technique is same for the anatomic calculation method. The prediction of postoperative lung function can be calculated by the following two steps. The first step is determining the functional contribution of resected (right or left) lung by quantitative perfusion lung scan, and the second step is applying the principle of anatomical calculation to the resected lung. Postoperative lung function is then estimated to be the product of the preoperative function and the portion of lung function that will remain after resection as estimated by the scan.

ppoFEV1 = preoperative FEV1 x [1 – functional fraction of the resected lung

 (number of segments to be resected/total segments of that lung)] The accuracy of these techniques has been questioned. A recent study using technetium scanning calculated values of imprecision from 18%-21% despite showing reasonable

segments from the total preoperative FEV1 (Juhl & Frost 1975; Zeiher, et al. 1995).

should be considered for the choice.

applied to the other methods as a basic concept.

**3.2 Split radionuclide perfusion scanning** 

correlation (Giordano, et al. 1997).

**3.1 Anatomic calculation** 

Quantitative computed tomography scanning has been studied as a technique to estimate postresection lung function. The basic concept is similar to radionuclide perfusion scanning method. This measures the split lung function using the CT attenuation density instead of radionuclide signal intensity. The volume of lung with attenuation between -500 and -910 Hounsfield units was used to estimate functional lung volume. The portion of the lung remaining postresection was predicted by calculating lung volume in the area to be resected as a portion of total lung volume. With this, predicted postoperative function correlated as well as the method using radionuclide quantitative perfusion imaging (Wu, et al. 2002).

#### **3.4 Dynamic perfusion magnetic resonance imaging (MRI)**

A study showed that magnetic resonance (MR) perfusion imaging had almost the same sensitivity and specificity for diagnosis of pulmonary perfusion defects as conventional perfusion scintigraphy (Berthezene, et al. 1999). The regional lung function is calculated from the subtraction images for normal lung parenchyma using image analysis software. The accuracy of MR perfusion for the prediction of ppoFEV1 was validated in a study (Iwasawa, et al. 2002). This study demonstrated that the correlation between perfusion ratios derived from MR perfusion image and radionuclide perfusion scanning was excellent (R = 0.92). The correlation between ppoFEV1 and actual postoperative FEV1 was similar when the two methods were compared (R = 0.682 in MR perfusion and R = 0.667 in radionuclide perfusion).
