Colorado-Massachusetts trial

Table 9.2. Major reasons given for the difference in interpretation for the 24 cancers detected only on one modality, either screen-fihn mammography or digital mammography, in the Colorado-Massachusetts trial (Lewin et al. 2002)...

A calculation of statistical power determined that 50,000 subjects would need to be enrolled to have a chance of detecting the small difference predicted by the Colorado-Massachusetts trial. Even then, a secondary reader study with multiple readers and a cancer-enriched subset of cases would most likely be needed to measure a difference with statistical significance. The plan was to enlist 20 sites to enroll 2,500 subjects each, equally distributed among four makers of digital mammography units GE, Lorad/ Trex, Fischer Imaging, and Fuji. At the time of study initiation, only the GE and Fischer units were approved by the US. Food and Drug Administration (FDA) for clinical use in the U.S. 

The four machines represented four distinct technologies. The GE machine, the commercial version of the prototype used in the Colorado-Massachusetts trial, used an 18 x 23 cm area ( flatpanel ) detector consisting of a cesium-iodide (Csl) phosphor bonded to an amorphous silicon substrate containing a rectangular photodiode array. The Csl phosphor converts the incident x-rays to light, which are then detected by the photodiode array. A thin-film transistor array deep to the amorphous silicon layer processes and transmits the electronic signal to the external electronics. The pitch (size) of a detector element in this system was 100 pm. 

As in the Colorado-Massachusetts trial, each subject was imaged twice, once on screen-film mammography and once on digital mammography. The order of the acquisitions was randomized so that half the time the screen-film exam was obtained first and half the time the digital exam was obtained first. In general, b oth exams were performed by the same technologist on the same day. Technical factors were recorded. 

The Oslo 1 study and the US Colorado-Massachusetts (Co-Ma) study (Lewin et al. 2001, 2002) were the pioneer studies comparing SFM and FFDM in mammographyscreening. The prospective Oslo 1 trial was carried out from January to June, 2000 (Skaane et al. 2003). A follow-up analysis of cancers missed at FFDM and subsequent cancers with a true positive... 

The NIH and the pharmaceutical industry have formed a new partnership to overcome barriers to early-phase clinical trials. The Secretary of Health and Human Services, Tommy Thompson, announced on July 9, 2003 grant awards for six cancer centers involved in a unique public-private partnership. Five pharmaceutical companies together with NCI put forth a total of 5.7 million for this partnership. Institutions receiving the funding include Massachusetts General Hospital University of Colorado Health Sciences Center Washington University, St. Louis University of Pittsburgh Cancer Institute University of California, Davis Cancer Center and Ohio State University Comprehensive Cancer Center (http //WWW. bms.com/news/other/data/pf other.news.385 5. html). 

---