Ductal lavage for detection of cellular atypia in women at high risk for breast cancer.

BACKGROUND: Breast cancer originates in breast epithelium and is associated with progressive molecular and morphologic changes. Women with atypical breast ductal epithelial cells have an increased relative risk of breast cancer. In this study, ductal lavage, a new procedure for collecting ductal cells with a microcatheter, was compared with nipple aspiration with regard to safety, tolerability, and the ability to detect abnormal breast epithelial cells. METHODS: Women at high risk for breast cancer who had nonsuspicious mammograms and clinical breast examinations underwent nipple aspiration followed by lavage of fluid-yielding ducts. All statistical tests were two-sided. RESULTS: The 507 women enrolled included 291 (57%) with a history of breast cancer and 199 (39%) with a 5-year Gail risk for breast cancer of 1.7% or more. Nipple aspirate fluid (NAF) samples were evaluated cytologically for 417 women, and ductal lavage samples were evaluated for 383 women. Adequate samples for diagnosis were collected from 111 (27%) and 299 (78%) women, respectively. A median of 13,500 epithelial cells per duct (range, 43-492,000 cells) was collected by ductal lavage compared with a median of 120 epithelial cells per breast (range, 10-74,300) collected by nipple aspiration. For ductal lavage, 92 (24%) subjects had abnormal cells that were mildly (17%) or markedly (6%) atypical or malignant (<1%). For NAF, corresponding percentages were 6%, 3%, and fewer than 1%. Ductal lavage detected abnormal intraductal breast cells 3.2 times more often than nipple aspiration (79 versus 25 breasts; McNemar's test, P<.001). No serious procedure-related adverse events were reported. CONCLUSIONS: Large numbers of ductal cells can be collected by ductal lavage to detect atypical cellular changes within the breast. Ductal lavage is a safe and well-tolerated procedure and is a more sensitive method of detecting cellular atypia than nipple aspiration.

Effect of cholecystokinin on human cholangiocarcinoma xenografted into nude mice.

Gastrointestinal polypeptide hormones regulate growth of various normal gastrointestinal tissues as well as certain visceral cancers. Since cholecystokinin (CCK) promotes growth of normal biliary tract, we sought to determine whether CCK affects the growth and metabolism of human cholangiocarcinoma line SLU 132. Twenty-six nude mice with s.c. xenografts of this cancer received either CCK octapeptide (50 micrograms/kg/dose) or 0.9% NaCl solution (saline) twice a day i.p. for 14 days. Tumor volume was calculated from Vernier caliper measurements. At sacrifice on Day 15, tumors were excised, weighed, and examined histologically. DNA, RNA, and protein were measured in the xenografted carcinomas. Because this cholangiocarcinoma produces carcinoembryonic antigen (CEA), we obtained serum at sacrifice for CEA radioimmunoassay and also tumor tissue for CEA immunolabeling with murine anti-CEA monoclonal antibody. Serum CEA levels were 90% higher in the CCK-treated group. Tumor tissue in the CCK-treated group also contained more CEA than did the controls. Mean tumor volume increased significantly in the saline group during the 14-day treatment period, whereas mean tumor volume did not increase significantly in the CCK group. Exogenous high-dose CCK thus appears to increase production and release of CEA from SLU-132; it also appears to retard growth of this tumor line in the nude mouse.

Loss of expression and aberrant methylation of the CDH13 (H-cadherin) gene in breast and lung carcinomas.

Expression of some members of the cadherin family is reduced in several human tumors, and CDH13 (H-cadherin), located on chromosome 16q24.2-3, may function as a tumor suppressor gene. In human tumors, loss of expression of many tumor suppressor genes occurs by aberrant promoter region methylation. We examined the methylation status of the CDH13 promoter in breast and lung cancers and correlated it with mRNA expression using methylation-specific PCR and reverse transcription-PCR. Methylation was frequent in primary breast tumors (18 of 55, 33%) and cell lines (7 of 20, 35%). In lung cancers, methylation was present more frequently in non-small cell lung cancer tumors (18 of 42, 43%) and cell lines (15 of 30, 50%) than in small cell lung cancer cell lines (6 of 30, 20%; P = 0.03). Only the methylated or unmethylated forms of the gene were present in most (73 of 80, 91%) tumor cell lines. CDH13 expression was present in 24 of 30 (80%) of nonmethylated tumor lines. All 18 methylated lines tested lacked expression irrespective of whether the unmethylated form was present, confirming biallelic inactivation in methylated lines. Gene expression was restored in all five methylated cell lines tested after treatment with the demethylating agent 5'-aza-2-deoxycytidine. Our results demonstrate frequent aberrant methylation of CDH13 in breast and lung cancers accompanied by loss of gene expression, although expression may occasionally be lost by other mechanisms.

Aberrant methylation of the adenomatous polyposis coli (APC) gene promoter 1A in breast and lung carcinomas.

The adenomatous polyposis coli (APC) gene is a tumor suppressor gene associated with both familial and sporadic cancer. Despite high rates of allelic loss in lung and breast cancers, point mutations of the APC gene are infrequent in these cancer types. Aberrant methylation of the APC promoter 1A occurs in some colorectal and gastric malignancies, and we investigated whether the same mechanism occurs in lung and breast cancers. The methylation status of the APC gene promoter 1A was analyzed in 77 breast, 50 small cell (SCLC), and 106 non-small cell (NSCLC) lung cancer tumors and cell lines and in 68 nonmalignant tissues by methylation-specific PCR. Expression of the APC promoter 1A transcript was examined in a subset of cell lines by reverse transcription-PCR, and loss of heterozygosity at the gene locus was analyzed by the use of 12 microsatellite and polymorphic markers. Statistical tests were two-sided. Promoter 1A was methylated in 34 of 77 breast cancer tumors and cell lines (44%), in 56 of 106 NSCLC tumors and cell lines (53%), in 13 of 50 SCLC cell lines (26%), and in 3 of 68 nonmalignant samples (4%). Most cell lines tested contained the unmethylated or methylated form exclusively. In 27 cell lines tested, there was complete concordance between promoter methylation and silencing of its transcript. Demethylation with 5-aza-2'-deoxycytidine treatment restored transcript 1A expression in all eight methylated cell lines tested. Loss of heterozygosity at the APC locus was observed in 85% of SCLCs, 83% of NSCLCs, and 63% of breast cancer cell lines. The frequency of methylation in breast cancers increased with tumor stage and size. In summary, aberrant methylation of the 1A promoter of the APC gene and loss of its specific transcript is frequently present in breast and NSCLC cancers and cell lines and, to a lesser extent, in SCLC cell lines. Our findings may be of biological and clinical importance.

Surgical resection for metastatic melanoma to the lung.

From 1971 through December 1986, the courses of 47 patients who underwent thoracotomy for pulmonary metastases from melanoma were retrospectively reviewed to determine the efficacy of this approach in the management of selected patients with melanoma. The overall five-year survival rate was 25% (median survival, 19 months). Thirty-eight patients were free of disease following thoracotomy. These patients fared significantly better than those who had residual disease following thoracotomy, with a five-year survival rate of 31% (median survival, 24 months) compared with 0% (median survival, six months). Survival was not influenced by the addition of adjuvant therapy or duration of time before the development of metastases (less than 12 months vs greater than or equal to 12 months). In selected patients with melanoma metastatic to the lung, thoracotomy with complete excision of the metastatic deposits results in improved survival and should be considered the treatment of choice.

Influence of parenteral progesterones on the prevalence and severity of mastalgia in premenopausal women: a multi-institutional cross-sectional study.

BACKGROUND: More than 30 percent of women seen in surgical breast clinics suffer from mastalgia. Although the causes of mastalgia are poorly understood, imbalances in estrogen and progesterone effects in the breast have been implicated. Progesterones have been proposed as a treatment for mastalgia, but the literature supporting their use is conflicting and currently inconclusive. STUDY DESIGN: The prevalence and severity of mastalgia in women receiving parenteral progesterones for contraception was compared to that of a randomly selected, age-matched control group using a validated survey instrument. Surveys were completed by 671 case subjects and 1,433 randomly selected, age-matched control subjects. RESULTS: Nine percent of women using medroxyprogesterone acetate (Depo-Provera, Upjohn, Kalamazoo, Mich) reported frequent breast pain compared to 21 percent of control subjects (OR 0.220, p < 0.001). The prevalence of clinically significant breast pain was 2.3 percent in the progesterone group compared to 4.9 percent in the control group (p < 0.02). Focal, noncyclic mastalgia predominated in the progesterone group with continued breast pain (78.8 percent), while diffuse, cyclic breast pain was more common in the control group (67.7 percent, p < 0.001). CONCLUSIONS: Medroxyprogesterone acetate effectively suppresses cyclic mastalgia in reproductive-age women and warrants additional study as a primary therapy.

Laparoscopic training on bench models: better and more cost effective than operating room experience?

BACKGROUND: Developing technical skill is essential to surgical training, but using the operating room for basic skill acquisition may be inefficient and expensive, especially for laparoscopic operations. This study determines if laparoscopic skills training using simulated tasks on a video-trainer improves the operative performance of surgery residents. STUDY DESIGN: Second- and third-year residents (n= 27) were prospectively randomized to receive formal laparoscopic skills training or to a control group. At baseline, residents had a validated global assessment of their ability to perform a laparoscopic cholecystectomy based on direct observation by three evaluators who were blinded to the residents' randomization status. Residents were also tested on five standardized video-trainer tasks. The training group practiced the video-trainer tasks as a group for 30 minutes daily for 10 days. The control group received no formal training. All residents repeated the video-trainer test and underwent a second global assessment by the same three blinded evaluators at the end of the 1-month rotation. Within-person improvement was determined; improvement was adjusted for differences in baseline performance. RESULTS: Five residents were unable to participate because of scheduling problems; 9 residents in the training group and 13 residents in the control group completed the study. Baseline laparoscopic experience, video-trainer scores, and global assessments were not significantly different between the two groups. The training group on average practiced the video-trainer tasks 138 times (range 94 to 171 times); the control group did not practice any task. The trained group achieved significantly greater adjusted improvement in video-trainer scores (five of five tasks) and global assessments (four of eight criteria) over the course of the four-week curriculum, compared with controls. CONCLUSIONS: Intense training improves video-eye-hand skills and translates into improved operative performance for junior surgery residents. Surgical curricula should contain laparoscopic skills training.

Sentinel lymph node biopsy, an alternative to elective axillary dissection for breast cancer.

BACKGROUND: Axillary metastases remain an important prognostic indicator in breast cancer. Axillary lymphadenectomy (ALND) carries significant morbidity and is unnecessary in most patients with early breast cancer; thus, sentinel lymph node (SLN) biopsy has been advocated for axillary staging. We studied the SLN identification rate and its accuracy in predicting axillary metastases. METHODS: One hundred nineteen women with breast carcinoma underwent SLN and ALND. Lymphoscintigraphy was performed using Technetium99 sulfur colloid supplemented by Isosulfan blue dye. Hematoxylin/eosin-stained lymph node sections were examined by light microscopy. RESULTS: The SLN identification rate was 81%. One SLN was negative (1%) in a patient with axillary disease. SLN histology correctly predicted the absence of axillary disease in 98.6%. Sensitivity, specificity, and positive and negative predictive values were 96%, 100%, 100%, and 99%, respectively. CONCLUSIONS: Sentinel lymph node biopsy accurately predicts total axillary status and is valuable in the surgical staging of breast cancer.

Management of the acute abdomen complicating oral anticoagulation therapy.

Acute abdominal pain in the patient receiving oral anticoagulants poses a difficult diagnostic and therapeutic challenge. We describe two cases of peritonitis requiring laparotomy in anticoagulated patients, and review 49 similar case reports from the world literature. These patients were usually explored for signs of bowel obstruction. At operation, the intestine often appeared infarcted, but pathologic examination commonly revealed intramural hematomata. In contrast, we present microscopic evidence of hemorrhagic cecal infarction complicating oral anticoagulation therapy in one patient. Intramural intestinal hemorrhage is the most common cause of acute abdominal pain in the anticoagulated patient who undergoes laparotomy. In addition to intramural hemorrhage, 14 per cent of patients had coexistent volvulus, appendicitis, intestinal wall disruption or intestinal infarction. We conclude that anticoagulated patients with suspected intramural intestinal hemorrhage may have severe intraabdominal pathology requiring operation. Therefore, operation is mandatory for patients who fail to improve after a short course of expectant management.

Improved colorectal cancer survival in an army community hospital.

In the later portion of 1988, surgeons at Darnall Army Community Hospital abruptly liberalized the use of colonoscopy in the evaluation of colorectal symptoms. During this period, colonoscopy rates jumped from a median of 31.6 examinations/100,000/year to 217.3 examinations/100,000/year. This study details the changes in outcome from colorectal cancer that followed this change. Before 1989, 74.5% of colorectal cancer patients presented with stage III or IV disease, and median survival was only 25 months. After 1988, the proportion of early cancers (stage 0, I, and II) increased dramatically to 56.4% (p = 0.002). Five-year survival increased concomitantly from 34.0 to 53.5% (p = 0.042). The liberalization of colonoscopy is judged to have had a beneficial effect on the health of the population served by this community hospital.

Understanding mathematical models for breast cancer risk assessment and counseling.

Chemoprevention and prophylactic surgery are effective interventions for lowering breast cancer incidence. However, these approaches are associated with risks of their own. Accurate individualized breast cancer risk assessment is an essential component of the risk/benefit analysis that must take place prior to implementing either of these strategies. Several mathematical models for estimating individual breast cancer risk have been proposed over the last decade. The Gail model is the most generally applicable model; however, it neglects family history information in second-degree relatives, treats pre- and postmenopausal breast cancer the same, and ignores personal histories of lobular neoplasia. The Claus model is a better family history model, but it does not assign any special relevance to histories of bilateral breast cancer or ovarian cancer, and neglects all of the nonfamily history information accounted for by the Gail model. BRCAPRO is a Bayesian family history model that calculates individual breast cancer probabilities based on the probability that a family carries a mutation in one of the BRCA genes. Though its treatment of family history information is more thorough than the other models, it neglects the nonfamily history risk factors accounted for by the Gail model and may not appreciate familial clustering unrelated to BRCA gene mutation. A thorough understanding of the principles of risk analysis and the available mathematical models is essential for anyone wishing to perform intervention counseling. This review describes the basic components of risk analysis, explains how the mathematical models work and compares the strengths and weaknesses of the various models. CancerGene is a software tool for running all of these models. It may be obtained without charge at http://www.swmed.edu/home_pages/cancergene.

Appraisal of anti-idiotypic antibodies in the treatment of solid tumors in humans.

Anti-idiotypic antibodies may be valuable in the induction of antitumor immunity in two ways--they can serve as a ready source of antigen when the appropriate TAA is difficult or impossible to purify. More importantly, regulatory anti-idiotypic antibodies can activate specific T-helper cells, bringing all the components of cellular immunity to bear on neoplastic process. Although the results in studies in animals have demonstrated resistance to tumor challenge after immunization with anti-idiotypic antibodies, studies of humans with advanced malignancies have failed to produce substantial clinical results. Nevertheless, immunization with anti-idiotypic antibodies has influenced some tumors. These studies represent important, initial steps toward understanding the immune network and modulating it in favor of the host, against human tumors. As a continued understanding of the immune network evolves and strategies for activating and suppressing specific immune responses are developed, it should be possible to design vaccines for specific uses. Although current anti-idiotypic vaccines do not seem promising for the treatment of established solid tumors in humans, we can look with anticipation to studies of polyvalent vaccines for the prevention of carcinoma in high risk groups.

Cyclosporin A does not affect adenoviral-mediated transgene expression of interleukin-2.

The adenoviral vector demonstrates efficient gene transfer and high transient gene expression making it an attractive vehicle for human gene therapy trials. Unfortunately, the virus stimulates a potent inflammatory immune response that limits transgene expression and makes repeat viral dosing ineffective. Transient immunosuppression has emerged as one technique to prolong adenoviral-mediated transgene expression and enable readministration of the viral vector. Cyclosporin A (CsA) causes immunosuppression by blocking the promotor/enhancer region of the gene for interleukin-2 (IL-2). The affect of CsA on transgene IL-2 expression was examined. Viral-mediated gene transfer was optimized in a human cell line using a type five adenoviral vector (Ad5) containing the gene for human IL-2 or bacterial beta-galactosidase (lac-Z) driven by a cytomegalovirus (CMV) promoter. CsA at various concentrations had no affect on IL-2 or lac-Z transgene expression. The immunosuppressive drug CsA is known to block native IL-2 transcription but has no affect on the adenoviral-mediated IL-2 or lac-Z transgene.

Characterization of a 90-100 kDa tumor-associated antigen in the sera of melanoma patients.

Using allogeneic antibody, we previously described a high-molecular-weight glycoprotein in the urine of 68% of melanoma patients. This glycoprotein has been termed urinary-tumor-associated antigen (U-TAA). A murine monoclonal antibody (MAb) specific for U-TAA (ADI-40F4) has been developed. By the use of ADI-40F4, U-TAA was detected in serum samples from 63% (33/52) of stage II and stage III melanoma patients, but from only 5% (1/20) of normal controls. This report describes the physical and immunochemical properties of U-TAA in the serum. The antigen elutes from a DEAE-Sephacel column in association with IgG in the void volume and as free antigen in a second peak. The molecular mass of the free antigen is 590-620 kDa and it sediments in the region of 28-29% sucrose by density gradient ultracentrifugation. Free antigen has an isoelectric point of 6.1. This high molecular weight antigen is composed of smaller subunits linked by reducible bonds. The ADI-40F4 reactive epitope resides on a 90-100 kDa subunit. These results provide evidence that U-TAA which is produced by melanoma cells in vitro is present in the circulation of melanoma patients.

Detection of a tumor-associated glycoprotein antigen in serum and urine of melanoma patients by murine monoclonal antibody (AD1-40F4) in enzyme immunoassay.

The urine of 68% of melanoma patients contains a high molecular weight glycoprotein which is expressed by melanoma cells and reacts with autologous antibody. Since high levels of this antigen in urine correlate with disease recurrence in surgically treated melanoma patients, it has been termed urinary tumor-associated antigen (U-TAA). We report the development of a murine monoclonal IgM antibody (AD1-40F4), which is specific for U-TAA. AD1-40F4 showed the same pattern of reactivity as the allo-antibodies previously used for the detection of U-TAA. The antigen recognized by AD1-40F4 has a high molecular weight (590-620 kilodaltons [kd]) and is heat stable. The AD1-40F4-reactive epitope is a protein. When AD1-40F4 was applied in an enzyme immunoassay, it allowed for the detection of U-TAA in the serum of 64% (33/52) of melanoma patients as opposed to only 5% (1/20) of normal controls. Thus, the murine monoclonal antibody AD1-40F4, which has been specifically developed against an allogeneic antibody defined antigen, U-TAA, appears to be important for immuno-prognosis of human melanoma.

Induction of antibodies to a tumor-associated antigen by immunization with a whole melanoma cell vaccine.

Urinary-tumor-associated antigen (U-TAA) is a glycoprotein present in the urine of melanoma patients. Previous studies have addressed the role of U-TAA in immunoprognosis. The present investigation was undertaken to determine whether the administration of whole melanoma cell vaccine (MCV) could induce the formation of anti-(U-TAA) antibodies in melanoma patients. The subjects of this study were stage II and III melanoma patients receiving MCV alone or in conjunction with cyclophosphamide. Anti-(U-TAA) IgM and IgG antibody levels were determined by enzyme immunoassay in sequential serum samples from 15 stage II and III melanoma patients receiving MCV. U-TAA purified from the urine of a melanoma patient was used as a target in this assay. The mean anti-(U-TAA) IgM titer prior to vaccination was similar to that of a non-vaccinated melanoma control group (1:1138 +/- 214, n = 15 vs 1:1334 +/- 254, n = 7; P = 0.375) but prevaccination IgG levels were generally higher than in the control group (1:3984 +/- 602 vs 1:2595 +/- 423; 0.1 greater than P greater than 0.05). While only 6 of the 15 patients demonstrated a rise in levels of IgG antibodies (mean 1:2964 +/- 1047 pre-MCV to 1:9958 +/- 2677 post MCV, P less than 0.01), 11 of the 15 patients demonstrated a greater than twofold rise in their anti-(U-TAA) IgM titers following vaccination (1:1051 +/- 259 pre-MCV to 1:2518 +/- 576 post-MCV; P less than 0.005). In addition, patients with visceral metastases consistently elicited anti-(U-TAA) responses equivalent to those with more limited disease. Concomitant administration of cyclophosphamide did not affect the response rates of peak antibody levels. The possibility that these antibody responses were actually against histocompatibility locus antigens (HLA) (contaminating our U-TAA preparation) was ruled out because the target antigen (U-TAA) was devoid of HLA, and because the induction of anti-(U-TAA) antibodies did not correlate with the induction of anti-HLA antibodies. These results demonstrate augmentation of anti-(U-TAA) IgM and IgG antibodies by immunization with the MCV.

Association between allo-immunoreactive and xeno-immunoreactive subunits of a glycoprotein tumor-associated antigen.

Using allogeneic antibody, we previously described a tumor-associated antigen (TAA) in the urine of 68% of melanoma patients. The TAA was purified from urine of a melanoma patient and used as immunogen to develop a murine monoclonal antibody (AD1-40F4) and xenopolyclonal antibodies in a baboon. Sera from melanoma patients treated with whole melanoma cell vaccine were used as the source of human antibody to the glycoprotein antigen. Treatment with 2-mercaptoethanol and separation by sodium dodecyl sulfate/polyacrylamide gel electrophoresis resolved the high-molecular-mass glycoprotein TAA into smaller subunits. Immunoblot analysis indicates that the murine monoclonal antibody (AD1-40F4) recognized a 90-100-kDa subunit of the antigen while human anti-TAA antibodies primarily recognized a 65-kDa subunit in addition to the 90-100-kDa subunit. Baboon polyclonal antibodies recognized the same subunits plus a 120-kDa subunit. Blocking studies indicated that the murine monoclonal and baboon polyclonal antibodies recognized the closely related epitopes on the 90-100-kDa subunit, while human antibodies recognized an epitope entirely distinct from that recognized by the mouse antibody. These results demonstrate the epitope complexity associated with the high-molecular-mass glycoprotein TAA.

Expansion of CD3+CD56+ lymphocytes correlates with induction of cytotoxicity by interleukin-2 gene transfer in human breast tumor cultures.

BACKGROUND: Mice immunized with murine mammary carcinoma cells genetically engineered to secrete interleukin-2 (IL-2) are rendered resistant to subsequent challenge with unmodified tumor cells, and in the case of mice bearing established tumors, the rate of development of pulmonary metastases is reduced. Despite these encouraging animal results, little is known about the induction of antitumor immunity by IL-2 gene transfer in human breast cancer. METHODS: Adenovirally mediated IL-2 gene transfer was performed in 12 tumor fragment cultures established from seven primary breast cancers. Autologous tumor infiltrating lymphocytes (TILs) or peripheral blood mononuclear cells (PBMCs) were cocultured with transduced tumor fragments, and changes in phenotype and cytotoxicity were measured. RESULTS: IL-2 was never detectable in the untransduced cultures, but it peaked at 5.0-1,324.8 ng/ml in the transduced cultures. Lymphocyte counts declined in all untransduced cultures, but they increased two- to sevenfold in four transduced cultures. CD4:CD8 ratios decreased from a mean of 2.11 at baseline to 1.27 after stimulation in coculture (p = 0.03). Expansion of lymphocytes expressing the natural killer cell phenotype (CD3-CD56+) occurred in only one culture, but the CD3+CD56+ population increased in four of six cultures. Lymphocytes from four of 10 cocultures generated significant cytotoxicity against allogeneic breast cancer cells. Induction of cytotoxicity correlated with expansion of the CD3+CD56+ phenotype (R2 = 0.805, p = 0.02). CONCLUSIONS: IL-2 gene expression by human breast cancer causes expansion of CD3+CD56+ cytotoxic-lymphocytes. This phenotype is consistent with that of a non-major histocompatibility complex (MHC)-restricted cytokine induced killer cell population previously described.

Tumor measurement in the nude mouse.

In this study, the pitfalls of tumor measurement in the nude mouse were evaluated. Regarding intermethod variation, diameters of subcutaneous tumors in nude mice were expressed as length, area, and volume; tumor weights were also recorded. These measurements were all compared to a reference standard: water displacement volume. Estimates of area and volume derived from caliper measurements correlated well with water displacement volume (r = 0.97 and 0.98, respectively). At necropsy, tumor weight was the most consistent and reproducible reflection of tumor volume (r = 1.0000). Regarding interobserver variation, mean absolute difference among volumes determined by several investigators who measured the same tumors in living animals was determined. This averaged 15% of the mean calculated volume. Regarding intraobserver variation, observers measured four separate masses in nude mice eight times each. The observers were prevented from realizing that the same animals were being repeatedly evaluated. Volumes were compared in order to quantify the average variation that occurs when a single investigator repeatedly measures the same mass. When large masses were measured, this error was 7%; when small masses were measured, the error was 27%. Recommendations are made for future work employing tumor measurement.

Loss of heterozygosity at 3p in benign lesions preceding invasive breast cancer.

BACKGROUND: Loss of heterozygosity (LOH) at chromosome 3p is one of the most common genetic abnormalities identified in human cancers and has occasionally been noted in benign proliferative lesions predisposing to breast cancer. If the frequency of LOH at 3p in benign proliferative lesions correlates with the subsequent development of breast cancer, it may be possible to develop powerful tools for molecular risk assessment based on this technology. MATERIALS AND METHODS: Archival paraffin-embedded tissues from benign breast biopsies in five women who have developed breast cancer and three women who have not developed breast cancer were microdissected and allelotyped at 3p using six microsatellite markers. RESULTS: No LOH was detected in the biopsies from women who have not developed breast cancer. For women developing breast cancer, the proportion of informative loci showing LOH in the benign proliferative lesions was 0.47 as compared to 0.57 for the associated breast cancers. There was no LOH detected in epithelial DNA from a fibroadenoma. Of 15 informative loci, 4 (27%) showed LOH in both the benign proliferative lesion and the associated cancer; however, the actual parental allele lost was different in three of these four cases. CONCLUSIONS: These results suggest that there are specific patterns of genetic instability common to preneoplastic lesions and the breast cancers that subsequently develop even when the paired lesions are not clonally related. LOH analysis of benign breast epithelium may provide a tool for molecular risk assessment and a surrogate endpoint for breast cancer chemoprevention trials.