The tissue polypeptide antigen serum test in the preoperative evaluation of non-small cell lung cancer. Diagnostic yield and comparison with conventional staging methods.

Tissue polypeptide antigen (TPA) is a protein produced and released by proliferating cells that has been shown to possess several characteristics for an ideal tumor marker. Our purpose was to determine the yield of TPA in the pretreatment assessment of non-small cell lung cancer (NSCLC), in comparison with a baseline clinical evaluation and multiorgan computed tomography (CT) assumed to be the gold standard for presurgical staging. One hundred four patients with NSCLC underwent thoracotomy, mediastinoscopy, or biopsy of suspected metastatic deposits, in addition to an extensive noninvasive evaluation of their stage of disease. We restaged retrospectively (UICC 1987 classification) these patients, on the basis of the following: (1) clinical history and physical examination, routine laboratory tests, bronchoscopy, chest radiographs, and any other examination as indicated by the prior baseline evaluation (BE stage); (2) the serum level of TPA (TPA stage); (3) the reading of a CT scan of brain, thorax, and abdomen obtained with no limitation to clinical information (CT stage); and (4) pathologic findings (RE stage). The TPA stage was calculated using 20 threshold values ranging from 45 U/L to 450 U/L. On the basis of the RE stage, sensitivity, specificity, accuracy, and predictive capabilities of BE, CT, and TPA were determined for stage I and II (full operability, FO), stage IIIa (possible operability, PO), and stage IIIb and IV (full inoperability, FI). The TPA thresholds were 110 U/L for detecting FO with the highest rate of success, and 160 U/L for detecting FI. Using these thresholds BE, CT, and TPA showed a diagnostic accuracy of, respectively, 75%, 79%, and 68% for FO; 87%, 69%, and 77% for PO; 87%, 77%, and 76% for FI. The accuracy of BE, CT, and TPA for both FO and FI was, respectively, 85%, 69%, and 69%. Of 74 patients classified operable by BE, 6 had a serum concentration of TPA less than 50 U/L and all 6 were confirmed in stage I or II at the subsequent thoracotomy; 15 others, out of 26 patients judged to have inoperable conditions by BE, had a TPA test result above 135 U/L and all 15 were pathologically classified in stage IIIb or IV. Using appropriate threshold values of TPA, it should be possible to predict NSCLC resectability with a diagnostic accuracy similar to that routinely achieved by CT.