Final analysis of the randomized trial on imatinib as an adjuvant in localized gastrointestinal stromal tumors (GIST) from the EORTC Soft Tissue and Bone Sarcoma Group (STBSG), the Australasian Gastro-Intestinal Trials Group (AGITG), UNICANCER, French Sarcoma Group (FSG), Italian Sarcoma Group (ISG), and Spanish Group for Research on Sarcomas (GEIS)☆.

BACKGROUND: In 2004, we started an intergroup randomized trial of adjuvant imatinib versus no further therapy after R0-R1 surgery in localized, high/intermediate-risk gastrointestinal stromal tumors (GIST) patients. Interim analysis results were published in 2015 upon recommendation from an independent data review committee. We report the final outcome of the study. PATIENTS AND METHODS: This was a randomized, open-label, multicenter phase III trial carried out at 112 hospitals in 12 countries. Patients were randomized to 2 years of imatinib, 400 mg daily, or no further therapy after surgery. The primary endpoint was imatinib failure-free survival (IFFS), while relapse-free survival (RFS), relapse-free interval (RFI), overall survival (OS) and toxicity were secondary endpoints. Adjusting for the interim analyses, results on IFFS were assessed on a 4.3% significance level; for the other endpoints, 5% was used. RESULTS: Nine hundred and eight patients were randomized between January 2005 and October 2008: 454 to imatinib and 454 to observation; 835 patients were eligible. With a median follow-up of 9.1 years, 5 (10)-year IFFS was 87% (75%) in the imatinib arm versus 83% (74%) in the control arm [hazard ratio (HR) = 0.87, 95.7% confidence interval (CI) (0.65; 1.15), P = 0.31]; RFS was 70% versus 63% at 5 years and 63% versus 61% at 10 years, [HR = 0.71, 95% CI (0.57; 0.89), P = 0.002]; OS was 93% versus 92% at 5 years and 80% versus 78% at 10 years [HR = 0.88, 95% CI (0.65; 1.21), P = 0.43]. Among 526 patients with high-risk GIST by local pathology, 10-year IFFS and RFS were 69% versus 61%, and 48% versus 43%, respectively. CONCLUSIONS: With 9.1 years of follow-up, a trend toward better long-term IFFS in imatinib-treated patients was observed in the high-risk subgroup. Although the difference was not statistically significant and the surrogacy value of such an endpoint is not validated, this may be seen as supporting the results reported by the Scandinavian/German trial, showing a sustained small but significant long-term OS benefit in high-risk GIST patients treated with 3 years of adjuvant imatinib.

Safety and efficacy of quavonlimab, a novel anti-CTLA-4 antibody (MK-1308), in combination with pembrolizumab in first-line advanced non-small-cell lung cancer.

BACKGROUND: Quavonlimab (MK-1308), a novel anti-CTLA-4 antibody, in combination with pembrolizumab was investigated in a phase I study. PATIENTS AND METHODS: Dose-escalation (DE) phase: patients with advanced/metastatic solid tumors received an initial flat dose of quavonlimab as monotherapy [25 mg (cohort 1), 75 mg (cohort 2), or 200 mg (cohort 3)] followed by four treatments of the same quavonlimab dose plus pembrolizumab every 3 weeks (Q3W). Dose-confirmation phase (DC): patients with stage IIIB/IV non-small-cell lung cancer (NSCLC) received first-line quavonlimab [25 mg Q3W (arm A), 25 mg Q6W (arm B), 75 mg Q6W (arm C), or 75 mg Q3W (arm E)] plus pembrolizumab. Primary objectives were safety and tolerability and establishment of the recommended phase II dose (RP2D) of quavonlimab when used with pembrolizumab. Objective response rate (ORR) was a secondary endpoint. Efficacy based on PD-L1 expression, tumor mutational burden (TMB), and changes in circulating CD4+/CD8+ cells were exploratory endpoints. RESULTS: Thirty-nine patients were enrolled in DE [n = 14 (cohort 1); n = 17 (cohort 2); n = 8 (cohort 3)] and 134 in DC [n = 40 (arm A); n = 40 (arm B); n = 40 (arm C); n = 14 (arm E)]. Maximum-tolerated dose was not reached. Grade 3-5 treatment-related adverse events (AEs; graded according to NCI CTCAE v4.03) occurred in 0%, 23.5%, and 75.0% of patients in DE cohorts 1, 2, and 3, respectively, and 35.0%, 30.0%, 35.0%, and 57.1% of patients in DC arms A, B, C, and E, respectively. Efficacy was observed at all dose levels/schedules in patients with NSCLC. ORRs were 40.0% [95% confidence interval (CI), 24.9-56.7; arm A], 37.5% (95% CI, 22.7-54.2; arm B), 27.5% (95% CI, 14.6-43.9; arm C), and 35.7% (95% CI, 12.8-64.9; arm E). PD-L1 expression and total number of circulating CD4+ cells correlated with ORR. CONCLUSIONS: Quavonlimab 25 mg Q6W plus pembrolizumab demonstrated similar efficacy and a better safety profile among all quavonlimab doses/schedules evaluated; this regimen was the chosen RP2D.

Early and late changes in markers of aortic stiffness with breast cancer therapy.

BACKGROUND: Anthracyclines and trastuzumab are well recognised to cause cardiac toxicity. Further to their effects on left ventricular (LV) function, anthracyclines in particular are considered to cause negative arterial remodelling. Whether these changes reverse is unknown. In addition, whether trastuzumab causes specific effects on arterial remodelling is yet undetermined. METHODS: Patients receiving these agents for treatment of breast cancer and healthy volunteers prospectively underwent clinical evaluation and cardiovascular magnetic resonance (CMR) imaging at baseline, 1, 4 and 14 months post-therapy, including functional assessment, measurement of aortic pulse wave velocity (PWV) using velocity encoded imaging and distensibility at ascending aorta (AA) and proximal descending aorta (PDA). RESULTS: Twenty-nine patients pretherapy and 12 volunteers demonstrated no differences in PWV, distensibility and LV function. Among cancer subjects, PWV increased acutely, P = 0.002 (4 months), then decreased by 14 months (P < 0.001). In addition, a decrease was observed in distensibility at the AA within 1 (P = 0.001) and 4 months (P < 0.001) of commencing therapy. At the PDA, only significant reduction was observed at 14 month distensibility when compared with baseline, P < 0.001. Patients with anthracycline exposure only had a greater reduction in aortic distensibility in the AA with time, P = 0.005 at 1 month, P < 0.001 at 4 months and P = 0.009 at 14 months. CONCLUSION: Acute changes are observed in PWV and distensibility at the AA following contemporary breast cancer chemotherapy and partially reverse a year after therapy is discontinued, with more severe effects seen with anthracyclines.

A phase I study of extended dosing with lomeguatrib with temozolomide in patients with advanced melanoma.

Lomeguatrib, an O(6)-methylguanine-DNA methyltransferase inactivator, was evaluated in an extended dosing regimen with temozolomide, designed according to pharmacodynamic data from previous studies. Patients with unresectable stage 3 or 4 cutaneous or unknown primary melanoma metastases were treated with lomeguatrib 40 mg, b.i.d. for 10 or 14 days and temozolomide 75-100 mg m(-2) on days 1-5. Drugs were administered orally with cycles repeated every 28 days, for up to six cycles. A total of 32 patients were recruited to the study. Lomeguatrib for 10 days with temozolomide 75 mg m(-2) was established as the optimal extended lomeguatrib dosing schedule, with haematological toxicity being dose limiting. There were two partial responses to treatment giving an overall response rate of 6.25%. Extending lomeguatrib administration beyond that of temozolomide requires a reduced dose of the latter agent. Only limited clinical activity was seen, suggesting no advantage for this regimen over conventional temozolomide administration in the treatment of melanoma.

An open-label, single-arm study assessing safety and efficacy of panitumumab in patients with metastatic colorectal cancer refractory to standard chemotherapy.

BACKGROUND: A phase 3 study demonstrated that panitumumab, a human monoclonal anti-epidermal growth factor receptor antibody, significantly prolonged progression-free survival versus best supportive care (BSC) in patients with chemorefractory metastatic colorectal cancer. PATIENTS AND METHODS: This open-label extension study evaluated panitumumab monotherapy in BSC patients with radiographically documented disease progression in the phase 3 study. Patients received panitumumab 6 mg/kg every 2 weeks. The primary end point was safety; efficacy was also evaluated. RESULTS: One hundred and seventy-six patients were randomly assigned to the BSC arm of the phase 3 study received >/=1 panitumumab dose in this extension study. Panitumumab was well tolerated. The most frequent treatment-related adverse events were skin toxic effects. Three (2%) patients had a grade 4 treatment-related adverse event. There were no infusion reactions. One (0.6%) patient had a complete response; 19 (11%) patients had a partial response; and 58 (33%) patients had stable disease. Median progression-free survival time was 9.4 [95% confidence interval (CI): 8.0-13.4) weeks. Median overall survival time was 6.3 (95% CI: 5.1-6.8) months. Anti-panitumumab antibodies were detected in 3 (4.2%) of 71 patients with a post-baseline sample. CONCLUSIONS: These findings are comparable to those from the phase 3 study and support panitumumab monotherapy for chemorefractory colorectal cancer.

Gemcitabine with a specific conformal 3D 5FU radiochemotherapy technique is safe and effective in the definitive management of locally advanced pancreatic cancer.

The aim of this phase II study was to assess the feasibility and efficacy of a specific three-dimensional conformal radiotherapy technique with concurrent continuous infusion of 5-fluorouracil (CI 5FU) sandwiched between gemcitabine chemotherapy in patients with locally advanced pancreatic cancer. Patients with inoperable cancer in the pancreatic head or body without metastases were given gemcitabine at 1000 mg m(-2) weekly for 3 weeks followed by a 1-week rest and a 6-week period of radiotherapy and concurrent CI 5FU (200 mg m(-2) day(-1)). The defined target volume was treated to 54 Gy in 30 daily fractions of 1.8 Gy. After 4 weeks' rest, gemcitabine treatment was re-initiated for three cycles (days 1, 8, 15, q28). Forty-one patients were enrolled. At the end of radiotherapy, one patient (2.4%) had a complete response and four patients (9.6%) had a partial response; at the end of treatment, three patients (7.3%) had a complete response and two patients (4.9%) had a partial response. Median survival time was 11.7 months, median time to progression was 7.1 months, and median time to failure of local control was 11.9 months. The 1- and 2-year survival rates were 46.3 and 9.8%, respectively. Treatment-related grade 3 and 4 toxicities were reported by 16 (39.0%) and four (9.8%) patients, respectively. Sixteen out of 41 patients did not complete the planned treatment and nine due to disease progression. This approach to treatment of locally advanced pancreatic cancer is safe and promising, with good local control for a substantial proportion of patients, and merits testing in a randomised trial.

Feasibility study on the MammoSite in early-stage breast cancer: initial experience.

The aims of this study were to evaluate the feasibility, practicality, efficacy and safety of the delivery of accelerated partial breast irradiation using the MammoSite for the boost phase. Six patients aged 53-69 years with stage T1N0, T2N0, Grade I-II invasive ductal carcinoma received 9-10 Gy prescribed at 1 cm from the MammoSite balloon surface in two fractions of 4.5-5 Gy 6 h apart. The MammoSite was inserted 20-37 days postoperatively. External beam radiation therapy to the whole breast commenced 1-5 days after accelerated partial breast irradiation. The maximum skin dose ranged from 3 to 9 Gy. The skin-cavity distance ranged from 7 to 19 mm. Local discomfort resolved as the scar healed spontaneously within 3-5 days. No Grade III or higher acute toxicity or local infection was recorded. The ease of insertion and accuracy of dosimetry makes the MammoSite suitable for use in properly selected women with early-stage breast cancer in a trial setting.

Gemcitabine and carboplatin in carcinoma of unknown primary site: a phase 2 Adelaide Cancer Trials and Education Collaborative study.

Cancer of unknown primary site (CUP) represents up to 5% of all cancer diagnoses and is associated with poor survival. We have performed a prospective multicentre phase 2 trial to evaluate efficacy and toxicity of the combination of gemcitabine (G) and carboplatin (C) for patients with CUP. Patients with histologically confirmed metastatic carcinoma in which the primary site of cancer was not evident after prospectively designated investigation and who had ECOG performance status 0-2 were treated with G 1000 mg m(-2) intravenously (i.v.) days 1 and 8, and C AUC 5 i.v. on day 8 every 3 weeks to a maximum of nine cycles. The primary end points were response rate, and toxicity, with secondary end points of progression-free survival and overall survival. Fifty-one (23 male, 27 female) patients were enrolled (one patient ineligible), with a median age of 69 years (range 41-83 years). Fifty patients were evaluable for toxicity and 46 patients were evaluable for efficacy. The overall response rate to the GC regimen was 30.5%. With a median follow-up of 24 months, the median progression-free survival was 18 weeks (4.2 months) and the median overall survival was 34 weeks (7.8 months). The frequency of grade 3 or 4 toxicity was low. Nausea/vomiting was the most common side effect, but was usually only mild in severity. Uncomplicated neutropenia (14%), thrombocytopenia (10%) and anaemia (8%) were the most common causes of grade 3-4 toxicity. The regimen was very well tolerated, particularly in the elderly. The GC regimen is an active regimen in CUP with excellent tolerability and should be considered particularly for elderly patients with CUP.

Darbepoetin alfa administered every 3 weeks alleviates anaemia in patients with solid tumours receiving chemotherapy; results of a double-blind, placebo-controlled, randomised study.

This dose-finding, placebo-controlled study evaluated the safety and efficacy of darbepoetin alfa administered every 3 weeks (Q3W) to anaemic patients receiving chemotherapy. In part A, patients (haemoglobin 50% of patients achieving a haematopoietic response. Administration of darbepoetin alfa Q3W has a tolerable safety profile and effectively ameliorates anaemia due to chemotherapy.

Molecular detection of blood-borne epithelial cells in colorectal cancer patients and in patients with benign bowel disease.

In colorectal cancer (CRC), a proportion of patients with early stage disease still die of metastatic or recurrent disease within 5 years of "curative" resection. Detection of carcinoma cells in the peripheral circulation at presentation may identify a subgroup of patients with micro-metastatic disease who may benefit from adjuvant chemotherapy or radiotherapy. Our aim was to determine the presence and clinical significance of colon carcinoma cells in peripheral blood at the time of surgery. Preoperative peripheral blood samples were collected from 94 patients with CRC and 64 patients undergoing bowel resection for benign conditions (adenoma, diverticular disease or Crohn's colitis). Blood was also obtained from 20 normal donors not undergoing bowel surgery. Immunomagnetic beads were used to isolate epithelial cells followed by reverse transcription-polymerase chain reaction (RT-PCR) analysis of expression of cytokeratin (CK) 19, CK 20, mucin (MUC) 1 and MUC 2. Nineteen of 94 (20%) CRC patients were positive for epithelial cells in preoperative blood, including 6 with early stage disease. Kaplan-Meier survival analysis showed that detection of epithelial cells in preoperative blood was associated with reduced disease-free and overall survival (log-rank test, p = 0.0001). Surprisingly, circulating epithelial cells were detected in 3/30 (10%) patients resected for adenoma, and in 4/34 (12%) patients resected for benign inflammatory conditions, suggesting that cells from nonmalignant colonic epithelium may also gain entry into the bloodstream in the presence of bowel pathology. All 20 normal control bloods were negative for epithelial cells.

Multicycle high-dose chemotherapy and filgrastim-mobilized peripheral-blood progenitor cells in women with high-risk stage II or III breast cancer: five-year follow-up.

PURPOSE: To determine the safety and efficacy of multiple cycles of dose-intensive, nonablative chemotherapy in women with poor-prognosis breast cancer. PATIENTS AND METHODS: Women with stage II breast cancer and 10 or more involved nodes or four or more involved nodes and estrogen receptor-negative tumors and women with stage III disease received three cycles of epirubicin 200 mg/m2 and cyclophosphamide 4 g/m2, with progenitor cell and filgrastim support every 28 days (n = 79) or 21 days (n = 20). Patients were reviewed at least twice yearly thereafter. Twenty-six patients had bone marrow and apheresis collections assessed for the presence of micrometastatic tumor cells. RESULTS: Ninety-nine women (median age, 43 years; range, 24 to 60 years) were treated. Ninety-two completed all three cycles of chemotherapy. The major toxicity was severe, reversible myelosuppression that was more prolonged with successive cycles, and this did not differ between patients given treatment every 28 days and those treated every 21 days. Febrile neutropenia occurred in 176 (61%) of 287 cycles. Severe mucositis (grade 3 or 4) occurred in 23% of cycles but tended to be short-lived and was reversible. The cardiac ejection fraction fell by a median of 4% during treatment, and three patients developed evidence of cardiac failure after chemotherapy. Two patients (2%) died of acute toxicity. Three of 26 patients had evidence of circulating micrometastatic tumor cells. The actuarial distant disease-free and overall survival rates at 60-month follow-up were 64% (95% confidence interval [CI], 53% to 75%) and 67% (95% CI, 56% to 78%), respectively. CONCLUSION: Multiple cycles of dose-intensive, nonablative chemotherapy is a feasible and safe approach. Disease control and survival are similar to those in other studies of myeloablative chemotherapy in poor-prognosis breast cancer. The regimen is being evaluated in a randomized trial of the International Breast Cancer Study Group.

Treatment of metastatic breast cancer with continuous infusional 5 fluorouracil.

BACKGROUND: Single agent continuous infusional 5 fluorouracil (CI-5FU) via a central venous catheter (CVC) is usually reserved for breast cancer patients who have previously failed one or more chemotherapy regimens. The patients are usually heavily pre-treated with later stage disease. Previously published studies of CI-5FU have reported response rates as high as 54%. It is considered an approach with an acceptable side effect profile in such patients. AIMS: To evaluate the efficacy and toxicity of CI-5FU in previously treated metastatic breast cancer. METHODS: A retrospective review of advanced breast cancer patients treated with CI-5FU between October 1992 and October 1996 was performed. Response to treatment, toxicity, CVC complications and patient survival were analysed. RESULTS: Twenty-four patients with metastatic breast cancer were treated with CI-5FU. All had received previous chemotherapy, including 19 patients (79%) with prior 5FU exposure and eight patients (33%) who had previous high dose chemotherapy with autologous stem cell transplantation. The median duration of CI-5FU treatment was 3.1 months. Nineteen patients had evaluable disease, three (16%) of whom demonstrated a partial response and four patients had stable disease. There were no complete responses. All responses occurred in soft tissue sites with no objective evidence of response in liver or bone metastases. The survival rate at one year was 21% (five of 24) and the median survival of all patients was 6.1 months. Five patients (21%) stopped treatment due to treatment related morbidity (two CVC complications and three CI-5FU side effects). Diarrhoea, nausea, and palmar-plantar erythrodysaesthesia were the major side effects of chemotherapy. CVC complications requiring intervention, the most notable of which were infection and thrombosis, occurred in 11 patients (46%). There were no treatment related deaths. CONCLUSIONS: Single agent CI-5FU has modest activity in women with previously treated advanced breast cancer. The efficacy is lower than in previously published series. This may reflect patient selection factors. The toxicity was mainly related to CVC complications. Important issues relating to quality of life need to be objectively measured in future studies of CI-5FU.

Effect of high-dose chemotherapy on intestinal permeability in humans.

1. Mucositis is a common side-effect of chemotherapy which is difficult to assess except by invasive means such as upper gastrointestinal endoscopy. Differential absorption of mono- and di-saccharides, such as rhamnose and lactulose, is a non-invasive measure of intestinal damage. 2. The purpose of the study was to assess the duration and severity of intestinal damage in patients undergoing high-dose chemotherapy and autologous blood stem-cell transplantation for malignant disease. 3. Thirty-five patients were studied before treatment and at 7, 28, 60 and 90 days after treatment. 4. The median lactulose/rhamnose ratios before treatment and at 7 and 90 days post-treatment were 0.09, 0.62 and 0.06 respectively. Altered permeability was due to both increased lactulose permeation and decreased rhamnose absorption. These abnormalities suggest a defect in tight-junction integrity as well as a decrease in surface area of small bowel. 5. We conclude that chemotherapy given for malignant disease is associated with a transient abnormality in intestinal sugar permeability, which peaks at 7 days after treatment and is composed of both mono- and di-saccharide absorption abnormalities.

Immunobead RT-PCR: a sensitive method for detection of circulating tumor cells.

Detection of circulating tumor cells and micrometastases in patients with cancer should prove useful in determining prognosis and in planning and monitoring systemic therapies. We have developed immunomagnetic isolation of carcinoma cells followed by reverse transcription polymerase chain reaction (immunobead RT-PCR) as a method for identifying very small numbers of breast cancer cells in blood. The expression of cytokeratin 19 (K19) was used as the marker by which the isolated tumor cells were identified. The immunobead RT-PCR technique allowed detection of one tumor cell per 10(6) leukocytes in whole blood. Immunobead RT-PCR is a highly sensitive method of detecting cancer cells in a hematopoietic environment.

Detection of circulating tumor cells in colorectal cancer by immunobead-PCR is a sensitive prognostic marker for relapse of disease.

BACKGROUND: Recurrent and metastatic carcinoma of the colorectum remains a major problem, with survival at 5 years post curative resection still only about 50%. Moreover, up to 30% of patients who present with early stage disease also relapse and die within 5 years, suggesting the presence of micrometastatic disease at diagnosis. One route of metastatic spread is via the blood stream, hence the detection of tumor cells in blood is likely to provide an important predictive tool with respect to relapse of disease. We have developed a sensitive molecular technique to identify tumor cells in blood using mutations in codon 12 of the K-ras gene as a marker. MATERIALS AND METHODS: Twenty-seven patients whose tumor carried a mutation in codon 12 of K-ras were studied for the presence of tumor cells in perioperative peripheral blood samples. Immunomagnetic beads, labeled with an epithelial-specific antibody, were used to harvest epithelial cells from blood. K-ras mutations were identified in this selected population using a polymerase chain reaction (PCR)-based analysis (immunobead-PCR). RESULTS: Circulating K-ras mutant cells were detected in 9 or 27 patients; seven of these nine patients have since died due to recurrent or metastatic disease. Mutant cells were not detected in 18 patients, and 16 or 18 have remained disease free (median follow-up: 16 months; range: 7-42 months). Kaplan-Meier analysis showed that detection of K-ras mutant cells in bloods was associated with significantly reduced disease-free survival (p = 0.0001). CONCLUSION: This study indicates that detection of circulating tumor cells perioperatively by immunobead-PCR provides a sensitive prognostic marker for recurrent and metastatic colorectal cancer.

Significance of molecular marker-positive cells after autologous peripheral-blood stem-cell transplantation for non-Hodgkin's lymphoma.

PURPOSE: To evaluate the significance of molecular marker-positive cells in a cohort of non-Hodgkin's lymphoma (NHL) patients undergoing high-dose chemotherapy and autologous peripheral-blood stem-cell transplantation (PBSCT). PATIENTS AND METHODS: Twenty-eight PBSC transplants have been performed in 24 patients with poor-prognosis NHL. Molecular analysis of the t(14;18) (q32;q21) translocation (bcl-2/immunoglobulin [Ig] heavy-chain joining locus [JH] fusion) or antigen receptor gene rearrangements was performed to determine the presence of lymphoma cells at presentation, in PBSC harvests, and before and after autologous PBSCT. Kaplan-Meier estimates of survival and Cox regression analyses were used to test the effect of bone marrow involvement, tumor-cell contamination of PBSCs, disease stage, and chemotherapy sensitivity at transplantation, and presence of marker-positive cells post-PBSCT on disease-free and overall survival. RESULTS: Thirteen of 24 patients (54%) are alive following PBSCT at a median follow-up time of 654 days (range, 193 to 1,908). Nine patients are in complete remission (CR) at day 216 to 1,799 (median, 805) and four are alive following relapse (day 440, 573, 1,188, and 1,908). Eleven patients (46%) have died: three of transplant-related complications at day 0, 1, and 13, and eight of recurrent disease (day 132 to 1,330; median, 451). Longitudinal marker studies post-PBSCT showed that of 16 relapse events, 13 (81%) were positive for the lymphoma marker at or before clinically documented relapse. Marker studies became negative post-PBSCT in nine of nine patients who entered and remained in CR. Disease-free survival (DFS) was significantly shortened in patients in whom marker-positive cells were detected in serial samples posttransplantation (P = .006). Cox regression analysis showed that patients in this group had a 24 times higher risk of relapse (P = .03). CONCLUSION: The results show that the reappearance or persistence of marker-positive cells after autologous PBSCT is strongly associated with relapse.

Dose intensive therapy with autologous blood stem cell transplantation in breast cancer.

BACKGROUND: Breast cancer is the commonest form of cancer in Australian women. Although approximately 50% of women with breast cancer achieve long term survival by current management methods, recurrent or metastatic disease is generally incurable. In addition, women with Stage II disease with > 10 positive axillary lymph nodes and also women with locally advanced disease (Stage III) have a poor survival even with adjuvant therapy. AIMS: To assess the toxicity and efficacy of high-dose chemotherapy with autologous peripheral blood stem cell (PBSC) transplantation in women with both metastatic and poor prognosis primary breast cancer. METHODS: Twenty-eight women with either metastatic (15) or poor prognosis (13) primary breast cancer were enrolled in the study between November 1988 to January 1993. PBSC were harvested using high-dose cyclophosphamide (Cy) with or without granulocyte-colony stimulating factor (G-CSF) and a myeloablative regimen of Cy, melphalan and carboplatin (CMCp) was used in the transplantation phase. RESULTS: Optimum numbers of stem cells were harvested in 85% of patients. The use of five G/m2 Cy plus G-CSF resulted in better PBSC yields and a significant reduction in haematologic morbidity when compared to mobilisation with Cy alone. Twenty-two women underwent 23 PBSC transplants (PBSCT). There have been two early deaths due to sepsis. The predominant morbidities observed following high dose chemotherapy and transplantation have been nausea, mucositis and diarrhoea. The median number of days to discharge following infusion of PBSC was 15 (range 11-21). At a median follow up time of 1.1 years (range 0 months-3.6 years), 8/22 (36%) evaluable patients remain alive and disease free while 14/22 (64%) have relapsed or progressed or died. CONCLUSION: High-dose chemotherapy and autologous PBSCT is a potentially highly effective treatment of women with metastatic and poor prognosis primary breast cancer. Randomised studies are required to compare this form of therapy to more standard forms of treatment in breast cancer.