Background: In the Mayo Lung Project Screening Trial, there were more carcinomas identified in the screened group compared with the control group. The screened group had better survival, but there was no difference in lung carcinoma mortality between the screened group and the control group. The purpose of this study was to review all available original pathology from the trial to determine whether overdiagnosis (carcinomas that do not result in the death of the patient) or misdiagnosis of lung carcinoma may explain this discrepancy.
Methods: All available lung pathology slides from patients who underwent surgery at the Mayo Clinic were reviewed independently by three blinded lung pathologists. Tumors were classified according to the 1999 World Health Organization criteria. In addition, agreement among the pathologists was assessed.
Results: Among 106 patients who underwent surgery at the Mayo Clinic, slides were available for review from 105 patients, including 77 slides from the screened group and 28 slides from the control group. The original diagnosis of carcinoma was confirmed in all patients. In 7 patients (6.7%), there was unanimous agreement that the lesion was preinvasive (carcinoma in situ), and these lesions all were from the screened group. In 90 patients (85.5%), there was unanimous agreement that the tumors were invasive. In 8 patients (7.8%), there was some disagreement between the observers about whether lesions were invasive or preinvasive; 7 of these 8 lesions were from the screened group. The level of agreement among pathologists for invasive carcinomas was > 94% for all comparisons, and the kappa statistic ranged from 0.67 (substantial agreement) to 0.84 (almost perfect agreement). There was good agreement among the pathologists about tumor cell type with the kappa statistic >/= 0.65.
Conclusions: The histologic diagnosis of carcinoma was confirmed for all 105 slides that were reviewed. The results of this study indicate that misdiagnosis does not explain the increased numbers of carcinomas identified in the screened group. The increased numbers of in situ carcinomas in the screened group resulted in increased numbers of squamous carcinomas in the screened group compared with the control group and may have contributed to the better survival. It is possible that carcinoma in situ accounted for some instances of overdiagnosis.
Copyright 2002 American Cancer Society.DOI 10.1002/cncr.10930