A retrospective study of 536 needle-localization biopsies of nonpalpable breast lesions and 623 excisional biopsies of palpable breast lesions was performed. Carcinoma was present in 17.9% of needle-localization biopsy specimens and in 11.1% of excisional biopsy specimens. Noninvasive carcinoma constituted 50% of carcinomas detected by needle-localization biopsy and only 7.3% of carcinomas detected by excisional biopsy (p less than 0.005). Invasive carcinoma detected by needle-localization biopsy was associated with axillary lymph node metastasis in 9.8% of patients who had axillary dissection, compared with 37.7% of patients with invasive carcinoma detected by excisional biopsy of a palpable mass (p less than 0.005). Invasive carcinoma detected by needle-localization biopsy was less than 2 cm in size (T1) in 93.5% of biopsy specimens; in contrast, invasive carcinoma detected by excisional biopsy was less than 2 cm in size in only 54.7% of biopsy specimens (p less than 0.005). Nonpalpable breast lesions that proved to be invasive carcinoma were pathologic stage I in 82.9% of patients. Palpable breast lesions that proved to be invasive carcinoma were pathologic stage I in only 47.2% of patients. Survival benefits of mammographic screening and biopsy of nonpalpable lesions are likely the result of detection of invasive carcinoma at an early stage and detection of noninvasive carcinoma that may later develop into or mark increased risk of invasive carcinoma.