Screening mammograms comprising of 32 first round, 10 interval and 32 second round detected cancers and 46 normal were examined by an expert screener, a screening radiologist, a clinical radiologist and a computer-assisted diagnosis (CAD) system. The expert screener, screening radiologist, clinical radiologist and the CAD detected 44, 41, 34 and 37 cancers, respectively, while their respective specificities were 80%, 83%, 100% and 22%. Later, with CAD prompting, the screening and the clinical radiologist detected 1 and 3 additional cancers each with unchanged specificities.
Screening mammograms comprising 35 first round, 12 interval and 14 second round detected cancers and 89 normal findings were examined without and with previous mammograms by experienced screeners. Without previous mammograms, the screeners detected 40.3 cancers with a specificity of 87%. With previous mammograms, 37.7 cancers were detected with a 96% specificity. The decrease in sensitivity was not significant but the screeners showed significant increase in specificity.
Local recurrences in 303 nonpalpable breast cancers with preoperative localizations and breast conservation therapy were evaluated for needle-caused implant metastasis. A total of 214 percutaneous biopsies were performed. There were 33 local recurrences. Needle-caused seeding or implantation as based on the location of the recurrence in comparison to the needle path in the mammograms was suspected in 3/44 (7%) invasive cancers without radiotherapy.
The mammographic characteristics of 317 nonpalpable breast cancers were categorized. Logistic regression showed that the risk ratios for a spiculated mass without calcifications and calcifications alone were 12 and 19 for invasive cancer and ductal cancer in situ (DCIS), respectively. Invasive ductal grade 1, ductal grade 2, lobular and ductal grade 3, had a risk ratio (RR) of 28, 17, 11 and 4.6, respectively, for a spiculated mass without calcifications. DCIS nuclear grade 3 and invasive ductal grade 3 had an RR of 17 and 9.7, respectively, for sole casting calcifications.
The eight-year survival of 96 1-9 mm invasive breast cancers were investigated in relation to their mammographic appearance, node status and histologic grade. After a median follow-up of 7 years, 6/96 died from breast cancer: 3/14 had calcifications alone, 2/56 had spiculated masses, 1/12 had rounded mass, 5/78 were node-negative and 1/4 was node-positive. The survival rate was 93%: 77% for the calcifications alone, 95% for spiculated masses, 91% for rounded masses, 92% for node-negative and 75% for node-positive. Calcifications alone and node positivity, each, carried a significantly higher risk of death.
Uppsala: Acta Universitatis Upsaliensis , 2001. , 22 p.