Marc J. van de Vijver

Simultaneous loss of E-cadherin and catenins in invasive lobular breast cancer and lobular carcinoma in situ

Loss of expression of the intercellular adhesion molecule E‐cadherin frequently occurs in invasive lobular breast carcinomas as a result of mutational inactivation. Expression patterns of E‐cadherin and the molecules comprising the cytoplasmic complex of adherens junctions, α‐, β‐ and γ‐catenin, were studied in a series of 38 lobular breast carcinomas with known E‐cadherin mutation status. The effect of loss of E‐cadherin by mutational inactivation (or other mechanisms) on the expression of catenins was investigated. Complete loss of plasma membrane‐associated E‐cadherin expression was observed in 32 out of 38 invasive lobular carcinomas, for which in 21 cases a mutation was found in the extracellular domain of E‐cadherin. In total, 15 frameshift mutations of small deletions or insertions, ranging from 1 to 41u2009bp, three non‐sense mutations, and three splice mutations were identified. Mutations were scattered over the whole coding region and no hot spots could be detected. In all cases, simultaneous loss of E‐cadherin and α‐ and β‐catenin expression was found; in 50 per cent of these cases, additional loss of γ‐catenin was observed. In six invasive lobular carcinomas, expression of both E‐cadherin and catenins was retained. In none of these carcinomas was an E‐cadherin mutation detected. Lobular carcinoma in situ adjacent to invasive lobular carcinoma showed simultaneous loss of E‐cadherin and catenins in all the cases studied—remarkably, also, in four cases positive for E‐cadherin and catenin expression in the invasive component. These results indicate that simultaneous loss of E‐cadherin and α‐, β‐ and γ‐catenin may be an important step in the formation of lobular carcinoma in situ, as a precursor of invasive lobular breast cancer. Events additional to E‐cadherin inactivation must be involved in the transition of lobular carcinoma in situ to invasive lobular carcinoma.

Local Recurrence After Breast-Conserving Therapy for Invasive Breast Cancer: High Incidence in Young Patients and Association with Poor Survival

PURPOSEnTo study risk factors for local recurrence (LR) after breast-conserving therapy (BCT) for invasive breast cancer and, for patients with an LR, the mode of detection, location, treatment, influence of radiation therapy, and impact on survival.nnnMETHODS AND MATERIALSn1360 patients (median age 52 years; range 24-88) with a total of 1393 pT1-2 N0-1 tumors treated with BCT between 1980-1994 were studied (median follow-up 52 months). The adequacy of radiation treatment of the patients developing LR was studied in a quality control study. The impact of LR on overall survival and distant metastasis was studied in a Cox regression model with LR as a time-dependent covariate.nnnRESULTSnA total of 88 LR occurred with a 5- and 10-year LR risk of 8 and 12%. Age was the only significant risk factor. Compared to patients > 65 years old, patients < 45 years old and patients 45-65 years old had a relative risk (RR) of 4.09 and 2.41, respectively, of developing LR. Risk on LR was found to increase gradually with younger age. Radiation therapy was considered adequate and did not play a role in influencing the LR rate. Almost 65% of the LR were true or marginal recurrences. Of all LR, 80% appeared during the first 5 years and were detected with equal frequency by the patient herself, the physician, and annual mammography. LR was a major predictor for distant metastasis (RR: 4.90; 3.15-7.62) and death (RR: 4.29; 2.93-6.28).nnnCONCLUSIONnYoung age is a major risk factor for LR and there is a significant gradual increase in LR with decreasing age. LR is associated with a higher risk of distant metastasis and death. Whether LR is the cause of or a marker for distant metastasis remains unresolved.

Genetic alterations during the progression of squamous cell carcinomas of the uterine cervix

Most cervical carcinomas appear to arise from cervical intraepithelial neoplasia (CIN) lesions. In addition to infection with high‐risk human papilloma viruses, which is indicative of an increased risk of progression, alterations of oncogenes and tumor suppressor genes play a role. Genetic studies of CIN lesions, primary cervical carcinoma, and metastases may shed light on the relative importance of various genetic alterations involved in the progression of CIN to invasive carcinoma. We examined tumor material from 10 patients with squamous cell carcinoma of the uterine cervix and synchronous CIN lesions and lymph node metastases. The CIN component, invasive carcinoma, and lymph node metastases were analyzed separately for loss of heterozygosity (LOH) on the following loci: VHL (3p21), HLA region (6p22–23), PGL (11q 22–24), E6 associated protein (15q11–13), TP53 (17p13), DCC (18q21.1), and chromosomes 1, 2, 4, 9, 20, and X. Using immunohistochemistry, the expression of the EGF receptor, ERBB2, and TP53 was determined. In CIN lesions, frequent LOH was found at chromosome arms 3p, 6p, and 11q. Primary invasive carcinoma showed additional LOH at chromosome arms 6q, 17p, and 18q. In lymph node metastases, an additional locus on the X chromosome displayed LOH. All carcinomas and synchronous lesions but one showed high expression levels of the EGF receptor. TP53 staining, when present, was found in all synchronous lesions. Focal staining of ERBB2 was found in one CIN lesion, two invasive carcinomas, and four metastases. The molecular alterations accumulated in a fashion that paralleled the progression of the tumors. These results indicate that cervical tumorigenesis occurs in a stepwise fashion, including infection and integration of oncogenic HPV and several specific genetic alterations. Genes Chromosomes Cancer 26:346–354, 1999.

Risk factors for local recurrence after breast-conserving therapy for invasive carcinomas: a case-control study of histological factors and alterations in oncogene expression.

Abstract Purpose: Many studies have focused on histological risk factors for local recurrence (LR) after breast-conserving therapy (BCT). In addition to histological factors, we studied alterations in the expression of various proteins in relation to LR using a case-control approach. Methods and Materials: Ninety-nine LR occurred in a patient cohort of 1,481 tumors treated with BCT. These patients were randomly matched, each with two controls. Matching was performed for age group (≤ 50 and > 50 years), pN stage, and follow-up time. Histology slides were reviewed. Immunohistochemical staining was performed for the following proteins: bcl-2, CD31, cyclin D1, E-cadherin, EGF receptor, ER, PR, Ki-67, c- erb B2/ neu , and p53. Statistical analyses were performed using conditional logistic regression. Results: Sixty-six cases and 139 controls with invasive carcinoma remained for analysis. The following variables were significant risk factors for LR: young age ( p = 0.006), high nuclear grade ( p = 0.04), high mitotic count ( p = 0.03), extensive DCIS around the tumor ( p = 0.02) but not within the tumor, poorly differentiated type of DCIS ( p = 0.03), > 20% ki-67 positive cells ( p = 0.006), and PR negativity ( p = 0.03). When the analysis was performed for patients ≤ and > 50 years, these risk factors were found in the older patients, but not in the younger patients. Conclusion: High mitotic count and Ki-67 positivity are risk factors for LR. EDCIS surrounding the invasive tumor is a risk factor for LR, especially when of poorly differentiated type. Age is an important risk factor for LR independent of other risk factors, including alterations in oncogene expression.

Isolated late local recurrences with high mitotic count and early local recurrences following breast-conserving therapy are associated with increased risk on distant metastasis

PURPOSEnLocal recurrence (LR) after breast-conserving therapy (BCT) is associated with an increased risk for the development of distant metastasis. We studied risk factors for distant metastasis risk (DMR) and poor prognosis within a group of patients with LR as first event.nnnPATIENTS AND METHODSnFrom a cohort of 1481 breast carcinomas treated with BCT in the period 1980-1994, a total of 68 pT1-3 N0-1 patients developed LR as first event. We have studied risk factors for the development of distant metastasis within this group of patients with LR. In addition to clinical factors (age at BCT and LR, mode of detection, location of LR, and treatment of LR), the histology slides of the primary and the recurrent tumor were reviewed. Immunohistochemical staining was performed for the following proteins: bcl-2, cyclin D1, E-cadherin, EGF receptor, ER, PR, Ki-67, c-erbB-2/neu, and p53. Statistical analyses were performed using conditional logistic regression.nnnRESULTSnAt a median follow-up after LR of 5.6 years, the 5-year DMR was 53%. In univariate analysis, none of the factors of the primary tumor was found to be associated with DMR after LR. Of the recurrent tumor the following factors were found to be risk factors for high DMR after LR: interval between treatment of the primary tumor and LR at 2 years or less (relative risk, 2.38; 95% confidence interval, 1.22-4.76; p = 0.008) and high mitotic count (relative risk, 2.51; 95% confidence interval, 1.03-6.15; p = 0.04). All patients with noninvasive recurrent tumor were alive at the time of analysis. Patients with an interval of greater than 2 years and a recurrent tumor with high mitotic count were found to have an equally poor prognosis compared to patients with LRs detected after a short interval.nnnCONCLUSIONnLR after BCT is associated with higher DMR and poor prognosis. Patients with LR within 2 years after BCT are especially at high risk. Late recurrences with high mitotic count have the same poor prognosis as early recurrences. For these patients, systemic treatment at time of the detection of LR should be considered.

A distinct phenotype characterizes tumors from a putative genetic trait involving chondrosarcoma and breast cancer occurring in the same patient

Recently, we documented an increased risk for the occurrence of breast- and cartilaginous tumors in the same patient, statistically pointing towards a potential genetic trait. This trait is most probably not associated with mutations in the two major hereditary breast cancer genes since no cases of enchondroma or chondrosarcoma were found in Dutch BRCA1 and BRCA2 families. We were able to collect and review the tumor tissue samples from 34 patients with both breast- and cartilaginous tumors and compared histopathological and immunohistochemical features of these tumors with controls. Breast cancer controls were available from literature data generated to compare familial breast cancers with nonselected cases. Clinical markers for chondrosarcoma controls were collected from the Netherlands Committee of Bone Tumors. Immunohistochemical data on chondro-tumor controls were available from our own files. Breast tumors of patients with cartilaginous sarcomas showed a significantly higher mitotic count (P=0.001), contained less lymphocyte infiltrate (P=0.025) and less nuclear pleomorphism. Remarkably, all cartilaginous tumors are of one common histological category originating centrally (P=0.014). Estrogen receptor and p53 expression were significantly higher (P<0.001) in breast cancer associated with chondro-tumors. p21 staining was more often negative in chondro-tumors associated with breast cancer. In seven cases of breast cancer, we found a slight decrease in CHEK2 expression. However, we could not identify the CHEK2 1100delC mutation in these cases nor in cases with normal CHEK2 expression. Hierarchical cluster analysis of all parameters within chondro-tumor-associated breast cancer specimens revealed two different subgroups, the largest one associated with estrogen receptor-positive breast cancer, which may distinguish sporadic cases from those belonging to the potential genetic trait. These distinct phenotypic findings support the existence of a new hitherto unrecognized syndrome, characterized by an increased risk to develop both breast cancer and centrally originating cartilaginous tumors.

Comparison of the genetic alterations in two epithelial collision tumors of the uterine cervix. A report of two cases.

In a minority of cervical carcinomas, a distinct adenocarcinoma and squamous cell carcinoma component can be recognized. These tumors are considered collision tumors; the differential diagnosis is adenosquamous carcinoma. To investigate whether the squamous and adenocarcinoma component are of multiclonal or monoclonal origin, we used loss of heterozygosity (LOH) as a method to establish clonality. Each tumor component of two tumors with a distinct adenocarcinoma and squamous cell carcinoma component were microdissected and the presence of LOH was studied for nine chromosomes, i.e., 1, 2, 3, 6, 11, 15, 17, 18, and X, which are known to contain frequent LOH in cervical cancer. The tumor of patient AK13 showed identical LOH in both the adenocarcinoma and squamous cell carcinoma tissue with various microsatellite markers on chromosomes 1, 2, 6, 18, and X. For markers on chromosomes 3 and 15, different LOH patterns were found in both components. The squamous epithelium showed LOH on chromosome 3, whereas the adenocarcinoma component had LOH on chromosome 15. For patient AK18 the LOH pattern on chromosomes 6p and 17 was the same in the adenocarcinoma and the squamous cell carcinoma component. The adenocarcinoma component showed additional LOH on chromosomes 6q and chromosome 11q. The tumor of patient AK18 showed common boundaries of LOH in both components on chromosome 17q, between markers D17S578 and D17S250. In conclusion, the squamous cell carcinoma and adenocarcinoma components in both tumors most likely have one cell of origin because many genetic alterations are the same in each component. The presence of genetic changes uniquely associated with one of the tumors favors a diversion of developmental pathways.