group during the screening phase (95 and 73 stage III and IV cancers in the intervention and control groups, respectively). An explanation for the lack of a mortality reduction may be that the screening tools, threshold criteria, or both were nonspecific and lacked sensitivity to detect cancers at an earlier and potentially curable stage. We agree that obtaining serial CA-125 measurements and applying an interpretative model may enhance the utility of CA-125 as a screening tool. The Risk of Ovarian Cancer Algorithm had not been developed when the PLCO trial was initiated in 1993. The trial used the best available evidence at that time to establish the screening protocol. In North America, in contrast to the United Kingdom, care of participants following an abnormal screening test in a randomized trial remains the responsibility of the patient and her physician. Although the PLCO trial did not specify a diagnostic workup, 87% of women who had abnormal or suspicious screening results for cancer (n=3358) underwent further evaluation. However, we agree that time to intervention is critical in a disease that is believed to progress rapidly. In women diagnosed with screen-detected cancer, the median time from screening to initial treatment was 2 months (interquartile range, 1-3.3 months). The PLCO findings thus represent the effectiveness of simultaneous CA125 and transvaginal ultrasound screening in the general US population. Contrary to the comment by Menon et al, the PLCO trial recognized the importance of a systematic process for quality markers and did develop and implement an extensive protocol (available on request) for the transvaginal ultrasound examination procedures, examiner training and certification, equipment specifications, examination standardization, and quality assurance procedures. As with any study, the PLCO trial has its limitations. Despite this, we believe that the findings remain clear: there was no mortality benefit in women screened for ovarian cancer with simultaneous CA-125 at a threshold of 35 IU/mL and transvaginal ultrasound. We are interested in the results of other ongoing clinical trials. However, even if a small benefit is ultimately detected, it must be weighed against the clear harms that occur from screening and subsequent surgery, including the possibility of overdiagnosis that appears to be emerging in our trial.
Download Full PDF Version (Non-Commercial Use)