Oncotype-DX recurrence score distribution in breast cancer patients with BRCA1/2 mutations.

Oncotype-DX assay has never been validated for BRCA mutation carriers. This study compares the recurrence score (RS) distribution in BRCA-positive breast cancer patients with that of a general population (GP) of patients and reports their outcomes. Eligible patients were BRCA carriers who performed the Oncotype-DX assay. Two sets of databases were cross-linked: BRCA carriers at Rabin Medical Center and Sheba Medical Center with Oncotype-DX tests performed through Clalit Health Services HMO, from 2003 to 2015. Fifty-eight BRCA patients were included (20 BRCA1, 38 BRCA2). The GP included 1020 patients. Compared to the GP, BRCA1 patients were younger, had higher rate of grade three tumors, and higher Ki67. BRCA2 patients had lower PR index, higher rate of grade three tumors, and higher Ki67. Among the GP, 52.9, 37.9, and 9.1 % had low, intermediate, and high risk RS, respectively. Corresponding rates were 15, 35, and 50 % in BRCA1 patients, and 18.4, 52.6, and 29 % in BRCA2 patients. Subgroup analysis revealed a similar RS distribution pattern regardless of the nodal status. Median follow-up was 45 months. Four BRCA patients (7 %) developed disease recurrence. RS of these patients were in the intermediate and low range. All recurrences occurred in chemo-naïve patients who had not undergone bilateral oophorectomy. This study revealed significantly different RS distributions between BRCA patients and the GP. RS values shifted toward high and intermediate risk categories. This pattern held regardless of the nodal status and was more pronounced in the BRCA1 group.

Colorectal cancer in young patients in Israel: a distinct clinicopathological entity?

PURPOSE: This study was designed to characterize the entity of colorectal cancer (CRC) in young patients and to evaluate whether it has any unique epidemiological or clinicopathological features. METHODS: The study population consisted of all consecutive young (≤50 years old at diagnosis) patients with CRC who were diagnosed during the years 1997-2007 and were treated at our institution, and a matching group of patients (>50 years at diagnosis). The medical files of these patients were reviewed, and the epidemiological, clinical, and pathological features of both groups were compared. RESULTS: There were 406 patients: 203 in each group. The features of the older group were typical for patients with CRC, but the younger group showed female predominance, different ethnic composition, prevalence of family history of cancer and hereditary CRC syndromes, and lower incidence of polyps. The incidence of left-sided tumors and advanced stages (III-IV) at diagnosis was higher in the younger patients. Mucinous/signet ring histology, grade, stage, lymphatic and vascular invasion were all predictive of survival, whereas age was not. CONCLUSIONS: Colorectal cancer in young patients was found to display a cluster of unique characteristics but fewer than previously reported and young age by itself was not found to impact patient outcome.

Decline in pulmonary function in patients with breast cancer receiving dose-dense chemotherapy: a prospective study.

BACKGROUND: Prompted by complaints of dyspnea in breast cancer patients receiving adjuvant dose-dense chemotherapy (DDC), we sought to evaluate the possible association of DDC with pulmonary dysfunction. PATIENTS AND METHODS: A total of 34 consecutive patients receiving adjuvant DDC were enrolled. The chemotherapy regimen consisted of i.v. doxorubicin 60 mg/m(2) and cyclophosphamide 600 mg/m(2) (AC) every 14 days x4 with growth factor support followed by weekly i.v. paclitaxel 80 mg/m(2) x12. The following parameters were prospectively measured before and after the AC protocol (P1, P2) and at completion of paclitaxel treatment (P3): presence of dyspnea, blood pressure, pulse rate, hemoglobin, erythrocyte sedimentation rate, C-reactive protein level, cardiac ejection fraction, and pulmonary function. Repeated measures analysis was used to evaluate differences among the time points, and paired t-test was used to evaluate differences between consecutive time points. RESULTS: Although only five patients (15%) complained of dyspnea, there was a significant decrease in mean carbon monoxide diffusing capacity (DLCO), in all patients from P1 (22.09 ml/min/mmHg) to P3 (15 ml/min/mmHg) and in 29 of 32 patients (90.6%) from P1 to P2 (15.96 ml/min/mmHg) (P<0.001). CONCLUSIONS: DDC is associated with a statistical significant reduction in DLCO. Awareness of this potential toxicity may be important in women with preexisting lung disease.

High incidence of non-upper aerodigestive primary tumors in patients with esophageal cancer.

Earlier reports have described an association between esophageal cancer (EC) and high incidence of other primary tumors (OPTs) of the upper aerodigestive tract and breast cancer. We evaluated the incidence of non-upper aerodigestive OPTs among Israeli EC patients; 2328 EC patients were retrieved from the Israeli National Cancer Registry between 1980 and 2004. The relative risk of OPTs for EC patients was measured using standardized incidence ratio (SIR). Two cohorts, Israeli National Cancer Registry registered colorectal cancer (CRC) patients and the general Israeli population, were used for reference; 297 EC patients (12.7%) had OPTs, including breast (18.9%), CRC (16.2%), prostate (8.8%), and bladder (8.4%) cancers. Upper aerodigestive OPTs were less common. Most OPTs were identified before (74.4%) or simultaneously with (13.8%) EC diagnosis. The median time interval between OPTs diagnoses and EC development was 6.0 years. The incidence of OPTs was significantly higher among EC patients compared with CRC patients (SIR: 2.05, P < 0.01) or the general Israeli population (SIR: 3.90, 95% CI: 3.46-4.34, P < 0.01) regardless of gender or tumor histology. Patients with EC have high incidence of non-upper aerodigestive malignancies. Unlike previous reports, the distribution of OPTs in EC seems to represent the relative incidences of these cancers in the western populations.

Radiation treatment for ductal carcinoma in situ (DCIS): is a boost to the tumor bed necessary?

The aim of the presented study was to evaluate the long-term outcome of breast-conserving surgery and radiation for the treatment of ductal carcinoma in situ (DCIS) and the role of the radiation boost to the tumor bed. The files of 75 women with DCIS treated by breast-conserving surgery followed by definitive radiation from 1988 to 1997 were reviewed for demographic data, prognostic variables, radiation dose, radiation boost, recurrence, and outcome. Total radiation dose was 5000 cGy delivered in 25 fractions. Twenty patients (26.7%) received an additional boost to the tumor bed of 1000 cGy in 5 fractions. Median follow-up time was 81.5 months (range, 22-145). Pearson correlation coefficient and its significance was calculated between the variables. Log rank test was used to analyze differences in local recurrence rates between patients who did or did not receive a boost, and a Cox regression model was fitted to the data to predict recurrence. Ten patients (13%) had local recurrence; one patient showed lymphatic spread. Histopathologic examination revealed DCIS in 6 cases (60%) and invasive duct carcinoma in 4 (40%)(one minimally invasive). The recurrence group included 3 of the 20 patients who received a radiation boost (15%) and 7 of the 55 who did not (12.7%) (p=0.7). Correlation analysis of patient characteristics, prognostic factors, and treatment was significant only between mastitis as the presenting symptom (n=4) and longer time to recurrence (p=0.02). The recurrence rate in the present study was similar to other series of conservative treatment for DCIS of the breast. No additional value was found for the radiation boost. Larger controlled randomized studies are needed to confirm these findings.

Tritiated thymidine labeling index and response in human breast cancer.

Tumor cell uptake of tritiated thymidine, expressed as the labeling index (Ll), was determined prior to treatment of 25 patients with disseminated breast carcinoma. All patients subsequently received combination chemotherapy with adriamycin, cyclophosphamide, and an antimetabolite (usually 5-fluorouracil), with or without vincristine. The Ll was significantly higher in responders to chemotherapy than in nonresponders (mean, 15 vs. 7.1; P less than 0.01). Other pretreatment variables examined did not show a significant association with likelihood of response. Measurements of the tumor cell Ll in patients with accessible tumor may be of benefit in selection of treatment; 0 of 9 patients with an Ll less than 9 had a response in our series in contrast to 11 of 16 patients with an Ll greater than 9 who had a response (P = 0.001).

Phase I and clinical pharmacological evaluation of aphidicolin glycinate.

The toxicity profile and the pharmacokinetics of aphidicolin glycinate, a water-soluble analogue of aphidicolin, have been evaluated in two consecutive phase I clinical studies. In the first study, aphidicolin glycinate was given by 1-hour infusion for 5 consecutive days, every 3 weeks (daily x 5 study); in the second study, which was planned on the basis of the pharmacokinetic information obtained in the previous study, the drug was given by 24-hour continuous infusion. Treatment was repeated every 3 weeks. In the daily x 5 study, the daily dose was escalated from 12 mg/m2 to the maximum tolerated dose of 2250 mg/m2. Local toxicity was dose limiting. Elimination half-life was 2 +/- 0.2 hours (mean +/- SE) with aphidicolin being undetectable 6-8 hours after the end of the infusion. In the 24-hour continuous-infusion study, the dose was escalated from 435 mg/m2 to the maximum tolerated dose of 4500 mg/m2. Local toxicity was dose limiting, while other toxic effects were absent. The experimentally determined concentrations at the steady state were in agreement with those predicted on the basis of the available pharmacokinetic data. The targeted concentration at the steady state of 3 micrograms/mL was achieved at doses greater than or equal to 3000 mg/m2. Twenty-four-hour continuous infusion is the recommended schedule for clinical evaluations of aphidicolin glycinate as the synchronizing agent or in combination with cisplatin.

Phase I and clinical pharmacology study of intravenous flavone acetic acid (NSC 347512).

We have conducted a Phase I and pharmacological study of flavone acetic acid, one of a series of novel flavonoids. The drug was administered i.v. weekly for 4 weeks, with a 2-week rest and then repeated. Flavone acetic acid was given initially in a 1-h infusion, but at the 3900-mg/m2 dose level, the infusion time was lengthened to 3 h. A total of 31 patients were treated with 9 different dose levels, ranging from 330 to 6400 mg/m2. Dose-limiting toxicity was acute hypotension that began after about one-third of each drug dose had been infused and rarely lasted more than a few minutes after the infusion was discontinued. In addition, subjective fatigue and asthenia causing unacceptable patient discomfort was dose limiting. A significant side effect noted that was not dose limiting was diarrhea during the infusion. This drug exhibited nonlinear pharmacokinetic behavior. Plasma levels exceeded 300 micrograms/ml during the infusion at the maximally tolerated dose. After the infusion ended the principal half-life was about 2 h. In 24-h urine collections 27% of the flavone acetic acid dose was recovered as intact drug and an additional 37% was recovered as a metabolite. The maximally tolerated dose determined in this study is 6400 mg/m2 given i.v. over 3 h.

Uracil-ftorafur: an oral fluoropyrimidine active in colorectal cancer.

PURPOSE AND DESIGN: This review describes the early clinical development of uracil-ftorafur (UFT), an oral fluoropyrimidine, designed in 1978 by adding uracil to ftorafur. The review focuses on the treatment of colorectal cancer and summarizes the Japanese experience and the phase I and II trials performed in the United States and Europe. RESULTS: Clinical trials of UFT published in the Western world have included 581 patients with colorectal cancer. UFT has been administered in these trials as a single agent or biomodulated by leucovorin (LV). UFT was administered daily in split doses for periods that ranged from 14 to 28 days. The activity of oral UFT in large-bowel cancer when administered with oral LV (approximately 50 mg/dose) has resulted in objective response rates of approximately 40%. Response rates of approximately 25% (range, 17% to 39%) were reported when UFT was administered as a single agent or with lower doses of LV. The highest dose-intensities of UFT are achieved with 28-day schedules of administration. The maximum-tolerated dose (MTD) of UFT with this schedule, when administered concomitantly with oral LV 150 mg daily, is 300 mg/m2 daily. The dose-limiting toxicity (DLT) of UFT has generally been diarrhea. Other commonly described toxicities include nausea and vomiting, fatigue, and stomatitis. Myelosuppression occurs infrequently. Typically, hand-foot syndrome and neurologic toxicity are lacking. CONCLUSION: UFT is a fluoropyrimidine active in colorectal cancer. The oral route of administration and improved safety profile represent important advantages over both conventional and infusional fluorouracil (5-FU) regimens.

Response to mesna, doxorubicin, ifosfamide, and dacarbazine in 108 patients with metastatic or unresectable sarcoma and no prior chemotherapy.

In this phase II trial, 105 eligible patients with no prior chemotherapy and advanced sarcoma received doxorubicin, ifosfamide, and dacarbazine (DTIC) with mesna uroprotection (MAID). Starting doses of these drugs were 60, 7,500, and 900 mg/m2 divided over 72 hours by continuous infusion, respectively. Mesna was given for 84 to 96 hours at 2,500 mg/m2/d. Myelosuppression was dose limiting, causing the only toxic death (sepsis). Nonhematologic toxicity consisted predominantly of anorexia and vomiting. Severe mucositis, macroscopic hematuria, renal tubular acidosis, renal failure, and CNS toxicity occurred in less than 5% of cycles. No cardiotoxicity was detected. The overall response rate (10% complete response [CR]) was 47% (95% confidence intervals, 5% to 18% and 37% to 57%, respectively). Most responses (approximately 70%) were observed within two cycles. Median times to progression were 10 and 9 months, respectively. Histologic high tumor grade, lesions less than 5 cm, and less than 1 year from diagnosis to study entry correlated with the probability of response. The median survival was 16 months. Time from diagnosis to study entry, performance status, and extent of disease, but not histologic grade, correlated with survival. Following CR, two patients remain disease-free at 32 and 16 months. Of the 15 additional patients rendered disease-free with surgery, two remain disease-free at 30 and 18 months with no further therapy. While most relapses occurred in sites of prior involvement, death from CNS metastases occurred in 11 of the 80 patients with high-grade sarcomas, of whom seven were still responding systematically (three complete responders). Because of its substantial response in this phase II trial, the MAID regimen is being compared with doxorubicin and DTIC alone in advanced sarcomas and to observation in the adjuvant treatment of high-grade sarcomas in randomized trials.

Successful chemotherapy of recurrent intracranial germinoma with spinal metastases.

A recurrent CNS germinoma with subependymal and spinal metastases was treated by combination of cis-platinum, bleomycin, and vinblastine, resulting in disappearance of intracranial and spinal tumor. Second intracranial relapse responded again to the same combination chemotherapy. Response of spinal metastases to this combination chemotherapy has not been reported previously.

Phase II studies of docetaxel in the treatment of various solid tumours. EORTC Early Clinical Trials Group and the EORTC Soft Tissue and Bone Sarcoma Group.

Docetaxel has been evaluated in six tumour types in a total of 189 patients entered into phase II studies. Treatment consisted of a 1 h intravenous infusion of docetaxel 100 mg/m2 repeated every 3 weeks. No premedication was administered for possible hypersensitivity reactions. Docetaxel was found to be effective as first-line chemotherapy for head and neck cancer (response rate 44%) gastric cancer (23%) and melanoma (14%) and as second-line chemotherapy for soft tissue sarcomas (21%; 95% confidence interval: 7.5%-43.7%). The results in colorectal and renal cancer were disappointing, with response rates of less than 10%. The most frequent adverse effects were alopecia (81%), grade III-IV leukocytopenia of short duration (66%) and skin reactions (52%). Hypersensitivity reactions were mild and occurred in 26% of patients. Docetaxel is an important new drug in the treatment of solid tumours.

Patient preference and pharmacokinetics of oral modulated UFT versus intravenous fluorouracil and leucovorin: a randomised crossover trial in advanced colorectal cancer.

The aim of this study was to determine the patient's preference for oral UFT/leucovorin (LV) or intravenous (i.v.) 5-fluorouracil (5-FU)/LV chemotherapy in metastatic colorectal cancer and to compare 5-FU exposure with these two treatment options. A total of 37 previously untreated patients with advanced colorectal cancer were randomised to start treatment with either oral UFT 300 mg/m2/day plus oral LV 90 mg/day for 28 days every 5 weeks or i.v. 5-FU 425 mg/m2/day plus LV 20 mg/m2/day for 5 days every 4 weeks. For the second treatment cycle, patients were crossed-over to the alternative treatment regimen. Prior to the first and after the second therapy cycle, patients were required to complete a therapy preference questionnaire (TPQ). The pharmacokinetics of 5-FU were determined by taking blood samples on days 8, 15 or 22 and 28 for UFT and on days 1 and 5 for i.v. 5-FU. 36 patients were eligible. 84% of the patients preferred oral UFT over i.v. 5-FU. After having experienced both treatment modalities, patients indicated taking the medication at home, less stomatitis and diarrhoea, and pill over injection as the most important reasons for their preference. The area under the plasma concentration curve (AUC) for 5-FU after UFT administration was 113 microM x min on day 8, 114 on day 15 and 98 on day 28; the peak levels (Cmax) were 1.2, 1.3 and 1.0 microM, respectively. The AUC for the 5-FU/LV courses was 3083 microM x min for day 1 and 3809 for day 5 (P=0.002). The Cmax was 170.1 and 196.2 microM (P=0.06) and the clearance 2.6 and 1.9 l/min, respectively (P=0.002). Patients with metastatic colorectal cancer clearly preferred oral over i.v. chemotherapy treatment. This choice was most importantly influenced by convenience and toxicity considerations. Although i.v. bolus 5-FU leads to higher peak 5-FU concentrations and AUC values compared with oral UFT, this pharmacokinetic advantage of i.v. 5-FU seems to translate mainly into higher toxicity as seen in large randomised studies comparing oral UFT/LV with i.v. 5-FU/LV. Oral UFT/LV compares favourably with i.v. 5-FU/LV in terms of toxicity and patient's preference and leads to prolonged 5-FU exposure, which is comparable to continuous i.v. 5-FU treatment.

Prognostic factors for local control of early glottic cancer: the Rabin Medical Center retrospective study on 207 patients.

PURPOSE: Different radiation therapy schedules and devices have been used over the last 20 years at Rabin Medical Center in patients with early glottic cancer. The aim of the present retrospective analysis was to identify the subgroup of patients at high risk of failure of radiation treatment. MATERIALS AND METHODS: Between 1974 and 1994, 207 patients with squamous cell carcinoma of the glottis, 182 Stage T1 and 25 Stage T2, underwent definitive radiation therapy. During this period, treatment was administered with different radiation devices (60Co or 6-MV X ray), using different dose/fraction protocols (1.8 or 2 Gy per day, 5 or 6 fractions per week), total doses (42-77.4 Gy), overall radiation times, and delays. These treatment variables, in addition to certain patient and tumor characteristics, were correlated with local control at a median follow-up of 57 months (range 18-265 months). RESULTS: The 5-year local control rates for T1 and T2 tumors were 88% and 73%, respectively. Univariate analysis showed that smoking, diabetes mellitus, anterior commissure involvement, T stage, and extension of tumor to one third or more of the vocal cord were highly significantly correlated with decreased local control. None of the treatment variables, including dosage at which complete tumor regression was noted, were found to be predictive. By multivariate analysis, only anterior commissure involvement was found to be highly significant (risk ratio 1.9, 95% CI 1.2-3.0, p = 0.027), and T stage was borderline significant (risk ratio 1.6, 95% CI 1.0-2.5, p = 0.054). CONCLUSION: This study suggests that only two tumor characteristics are predictive of local failure of early glottic cancer: anterior commissure involvement and T stage. Treatment variables apparently do not influence local control.

Combined modality treatment for stage III ovarian carcinoma.

Thirty-eight Stage III ovarian carcinoma patients were treated with a combined modality protocol consisting of sequential initial surgery with a maximal tumor reduction, CHAD combination chemotherapy, second look reductive surgery and whole abdominal irradiation. Sixteen patients (42%) had minimal residual tumors (less than 2 cm) after initial surgery (Stage IIIA) and 22 (58%) had large residual tumors (greater than 2 cm) (Stage IIIB). The patients received 3-14 courses of CHAD combination chemotherapy, with a response rate (CR + PR) in the evaluable (Stage IIIB) patients of 91%. Twenty-eight patients had a second attempt of cytoreductive operation (10 Stage IIIA patients and 18 Stage IIIB patients). In 10 patients no residual tumor was found. In another 12 patients residual tumor (less than 2 cm) was found and completely resected, whereas in six patients a complete resection of large residual tumors (greater than 2 cm) was not possible. Twenty-one of the patients also completed a course of whole abdominal radiotherapy. Radiation was well-tolerated with the usual expected amounts of nausea, vomiting, diarrhea and transient leukopenia and thrombocytopenia. 11/21 (52%) of the patients relapsed within the first 18 months after completion of radiotherapy. The actuarial relapse-free survival at 36 months from completions of radiotherapy was 44%. The actuarial survival for the whole group from diagnosis was 43% at 3 years (70% for Stage IIIA and 41% for Stage IIIB). The data indicated that this combined modality protocol is both feasible and well-tolerated but its curative potential for patients with advanced ovarian carcinoma is as yet unknown.

Does mca contribute to the follow-up of breast-cancer patients by ca-15-3.

Mucin-like carcinoma-associated antigen (MCA) and CA 15-3 tumor markers were randomly assayed in 234 consecutive breast cancer patients. It was found that 45 patients (19.2%) had elevated MCA levels (cut-off level >14 U/ml) and normal CA 15-3 levels (cut off level >30 U/ml). In 14 of these 45 patients (31.1%), overt metastases were detected, although five had started their follow-up with no evidence of disease. In these five patients, the median lead time was nine months. In our limited experience, it was found that measuring MCA levels in the serum in the presence of normal CA15-3 levels contributes to early detection and monitoring of recurrences in follow-up of breast cancer patients.

Tailoring treatment for classical Kaposi's sarcoma: comprehensive clinical guidelines.

Classical Kaposi's sarcoma (CKS) is a rare indolent proliferative disease which is particularly prevalent among Jews of Ashkenazi and Mediterranean origin. To define guidelines for its comprehensive management, we conducted a retrospective analysis of 123 patients, focusing mainly on treatment modalities. The CKS-related mortality was 4% (5 patients). Of the 39 patients for whom observation only was the primary approach, 15 (38%) remained progression-free for 1-83 months (median, 4 months). Twenty-nine of the 52 (56%) patients who underwent surgery as the primary approach remained recurrence-free for 1-162 months (median, 15 months). Radiotherapy achieved an objective response in 74 courses (85%), including 50 (58%) complete responses. Symptomatic relief was reported in 95% of the patients. Vinblastine (27 series) achieved an objective response in 73% of series, including 22% complete responses. Multivariate analysis of time to progression with observation alone identified immunosuppression as the only significant independent factor that predicted disease progression. Our study suggests that observation alone may be sufficient for immunocompetent asymptomatic patients; symptomatic resectable lesions are suitable for simple excision; and more advanced disease or unresectable lesions require radiotherapy. If disease is extensive or the other approaches fail, chemotherapy is appropriate. Tailoring the treatment for CKS is an integrative process, requiring good understanding of the role of each available modality in the different clinical disease settings.

Adjuvant treatment of high-risk stage II breast cancer with doxorubicin followed by high-dose chemotherapy and autologous stem-cell transplantation: a single-institution experience with 132 consecutive patients.

Several studies have shown conflicting results with the use of intensive consolidation chemotherapy for breast cancer. The aim of the present study was to investigate the efficacy, feasibility and toxicity of high-dose chemotherapy with stem cell support in patients with high-risk stage II breast cancer. From February 1994 to November 1998, 132 consecutive patients with multinode positive breast cancer were entered to the study. In total, 86 patients had >or=10 positive axillary lymph nodes, and 46 had 4-9 positive axillary lymph nodes with at least two additional predetermined risk factors at diagnosis. All patients were offered adjuvant chemotherapy (doxorubicin, 75 mg/m(2) x 4) followed by high-dose chemotherapy (cyclophosphamide 6000 mg/m(2), carboplatin 800 mg/m(2) and thio-tepa 500 mg/m(2)) and autologous stem cell support with growth factor. In all, 131 patients also received local radiation therapy and tamoxifen based on receptor status. After a median follow-up of 51 months (range 27-87), the disease-free and overall survival rates were 72 and 81%, respectively. There was no difference in the outcome for high-risk patients with > or < than 10 positive axillary lymph nodes. On Cox regression analysis only progesterone receptor status was predictive of disease-free, but not overall survival. There were no treatment-related deaths; grades III-IV toxicity was relatively low. This combined approach of doxorubicin followed by high-dose chemotherapy and stem-cell support, followed by locoregional radiotherapy, was safe and seems to be effective in patients with multinode positive stage II breast cancer. In previous trials of adjuvant high-dose therapy in this patient population, treatment-related morbidity and mortality markedly influenced the outcome. For this high-risk patient population, further testing of intensive chemotherapy regimens with a lower toxicity profile is warranted.

HLA-DR and beta 2 microglobulin expression in medullary and atypical medullary carcinoma of the breast: histopathologically similar but biologically distinct entities.

AIMS: To examine the expression of HLA-DR and beta 2 microglobulin in medullary carcinoma and atypical medullary carcinoma of the breast to determine if the effective presentation of tumour antigens to the immune system can differentiate between these two histopathologically similar entities. METHODS: Expression of HLA-DR and beta 2 microglobulin was examined by immunohistochemical methods in five samples of medullary carcinoma of the breast, which has a relatively favourable prognosis, six samples of atypical medullary carcinoma of the breast, which has a prognosis closer to that of regular invasive duct carcinoma, and 20 samples of invasive duct carcinomas, 10 with an accompanying lymphocytic infiltrate. RESULTS: A positive and significant correlation was found between tumour type and both HLA-DR and beta 2 microglobulin expression. Expression was most prominent in medullary carcinoma, followed by atypical medullary carcinoma and invasive duct carcinoma with and without lymphocytic infiltrates. The mean intensity and percentage of HLA-DR tumour immunostaining were significantly higher in medullary carcinoma than in the other three tumour groups, as was the mean intensity of beta 2 microglobulin immunostaining. Mean percentage of beta 2 microglobulin immunostaining was significantly higher in medullary carcinoma than in invasive duct carcinoma without lymphocytic infiltrates, and showed a trend to increase from invasive duct carcinoma with lymphocytic infiltrates to atypical medullary carcinoma and medullary carcinoma. CONCLUSIONS: Medullary carcinoma and atypical medullary carcinoma of the breast differ in their expression of HLA-DR and beta 2 microglobulin. The relatively favourable prognosis of medullary carcinoma of the breast may be related to effective tumour antigen presentation to the immune system through MHC-I and MHC-II expression. Immunotherapy aimed at MHC-I and MHC-II induction might have a beneficial effect in breast cancer.