A review of the literature and British Thoracic Society guidelines state that an FEV1 ≥2L is a safe lower limit for pneumonectomy in patients with lung cancer. Others have utilized FEV1 based on % of predicted normal ranging between ≥60% and ≥80% to indicate a patient suitable for pneumonectomy. There is an apparent lack of congruency between various methods for preoperative patient selection.
To develop a more reliable algorithm for the preoperative selection criteria we compared FEV1 in absolute values(L) and as % of predicted normal to quantitative radionuclide estimates of predicted postoperative FEV1(ppo%FEV1) for all patients with unilateral lung cancer referred to our pulmonary laboratory for preoperative evaluation between January 2002 and May 2005.
A total of 1,334 patients (M=774,F=560), mean age=64yrs±12 (range=19-93) were studied. Patients’ mean FEV1=2.04L±0.72(70.4%±19.7). Six hundred and thirty one patients (47%) had an FEV1 ≥2L (mean=2.64L±0.54). Of these, 309(49%) had an FEV1<80%. Among patients with a baseline FEV1≥2L, the ppo%FEV1was<40% in 189(30%) and it was <35% in 84(13%). Four hundred and seventeen patients(31%) had a baseline FEV1≥80%. Of these, 95(23%) had an FEV1<2L. For patients with a baseline FEV1≥80%, the ppo%FEV1 was <40% in 47(11%) and it was <35% in 9(2%).There were 552(41%) patients with FEV1≥60% and ≤79%. For these, the ppo%FEV1was<40% in 227(41%) and it was <35% in 141(26%).
There is a significant discrepancy between preoperative selection criteria that use FEV1 as absolute value (L) and as % of predicted. Almost one-half of patients with an FEV1 ≥2L had an FEV1 <80% of predicted. Approximately one-third of patients an FEV1 ≥2L would have been deemed inoperable based on ppo%FEV1 of <40%.
Algorithms for preoperative functional assessment for patients with lung cancer that use values of FEV1 ≥80% for pneumonectomy or ppo%FEV1≥35% are better than predictions for resection using absolute values of FEV1 (L). This approach is more reliable and reduces bias caused by variations in patients’ age, gender, race, and height.
M. Ersoy, None.