K. D. Bagshawe

Comparison of ondansetron and ondansetron plus dexamethasone as antiemetic prophylaxis during cisplatin-containing chemotherapy

Ondansetron, a serotonin antagonist, is effective in controlling the emesis associated with cancer chemotherapy; however, emesis in patients receiving high-dose cisplatin is poorly controlled by ondansetron alone. Dexamethasone is an effective antiemetic with no known interaction with serotonin receptors and was thus chosen for study in combination with ondansetron. 31 patients (30 male, 1 female; median age 28.5 years, range 18-49) receiving a 4-day course of a chemotherapy regimen containing cisplatin (100-120 mg/m2) for metastatic germ-cell tumours were entered in a randomised, double-blind, cross-over trial comparing oral ondansetron plus placebo with oral ondansetron plus dexamethasone as antiemetic prophylaxis. Ondansetron (8 mg every 8 h) was given to all patients for 8 days from the start of chemotherapy. Patients were given 8 mg of dexamethasone or placebo every 8 h starting 2 h before cisplatin (on day 4) and continuing for six doses (ie, for 2 days only). A second course of chemotherapy began 14 days after the start of the first, during which patients crossed over to the alternative antiemetic regimen. Results were available from 27 patients. In the 24-48 h after cisplatin 78% of patients taking ondansetron plus dexamethasone reported complete or major control of emesis compared with 30% of those taking ondansetron plus placebo (p = 0.001). Cross-over analysis showed a significant advantage for ondansetron plus dexamethasone in the control of nausea (p = 0.013) and emesis (p less than 0.001) over the 8-day study. 24 of 26 patients expressed a preference for the combination therapy (p less than 0.001). Ondansetron plus dexamethasone is effective antiemetic prophylaxis for high-dose cisplatin chemotherapy, has few side effects, and is active when given orally.

The First Bagshawe lecture. Towards generating cytotoxic agents at cancer sites.

Several years of experience have now accumulated in the targeting of anti-cancer agents so that we can take stock, identify problems and look for ways round them. Three major obstacles seem to limit present approaches. These are heterogeneity in the distribution of target molecules within the cancer cell population, the pharmacokinetic characteristics of macromolecules and host antibody response to foreign protein. An approach which we have been investigating uses antibodies or other vectors to carry enzymes which have no close human homologue to tumour sites. After clearing residual enzyme activity from the blood by one of several possible techniques, a relatively non-toxic prodrug is given. This prodrug is a substrate for the tumour located enzyme which results in the generation of a highly toxic molecule able to penetrate the tumour mass and cross cell membranes. Genetic engineering methods now offer the prospect of human immunoglobulins with tumour binding and catalytic sites having the potential to minimise host response. Whether this can be achieved depends on having antibodies with adequate specificity and our ability to develop enzyme-prodrug systems with the required characteristics. Early results encourage us to think progress can be made in this direction.

The role of low‐dose methotrexate and folinic acid in gestational trophoblastic tumours (GTT)

Summary. Between 1964 and 1986, 487 patients with gestational trophoblastic tumour (GTT) were treated with methotrexate and folinic acid. The patients comprise two groups: between 1964 and 1974, 126 patients were treated but were not systematically stratified using a prognostic score before the start of treatment. These patients formed part of the 317 women who were analysed to identify a number of prognostic variables (Bagshawe 1976). Retrospective analysis of these 126 patients using these prognostic factors showed that in the true low‐risk group 85/88 (96%) are alive while 20/22 (91%) of the medium‐risk group and only 5/16 (31%) of the high‐risk group are alive. Overall the survival was 110/126 (87%) with a minimum follow‐up of 14 years. From 1974 all patients were stratified on admission into prognostic groups. Of the true low‐risk patients 347/348 survived (99·7%); 13 patients were underscored and treated as low risk when they should have been treated as medium risk, 12 (92%) of these are alive, but nine (69%) needed to change treatment because of drug resistance. While the overall survival in the 1974–1986 group was 359/361 (99%) with a minimum follow‐up of 16 months, the survival in all patients (1964–1986) was 469/487 (96%). Although the survival in these patients is excellent it should be noted that 69/348 (20%) low‐risk patients had to change treatment because of the development of drug resistance, and a further 23 (6%) needed to change treatment because of drug‐induced toxicity. Low‐dose methotrexate and folinic acid given in the schedule described remains the treatment of choice in patients with low‐risk GTT because of its effectiveness and minimal short‐ and long‐term toxicity. Patients presenting with medium‐ and high‐risk GTT need to start with combination chemotherapy from the outset to avoid the development of drug resistance.

Repeated antitumour antibody therapy in man with suppression of the host response by Cyclosporin A

Antibody targeted therapy of cancer results in anti-antibody production which prevents repeated treatment. Cyclosporin A (CsA) has been used to suppress this response in patients treated with a radiolabelled antibody to carcinoembryonic antigen (CEA). Patients with CEA producing tumours received a minimum of two courses consisting of an injection of radiolabelled antibody and CsA, 24 mg kg-1 day-1, for 6 days; each course was given at 2 week intervals. Two weeks after the completion of the second course the mean human antimouse antibody (HAMA) levels were 3.5 micrograms ml-1 (s.d. 2.7) in 3 patients receiving CsA and 1,998 micrograms ml-1 (s.d. 387) in 3 patients not receiving the drug. Clearance of antitumour antibody was accelerated and tumour localisation absent when HAMA levels exceeded 30 micrograms ml-1. With lower levels of HAMA in the CsA-treated patients, further antitumour antibody accumulated in the tumour after each dose. Further therapy with antitumour antibody and CsA lead to the development of HAMA, but this was less than 25% of the amount in patients not given CsA. In this preliminary study up to 4 times as many doses of antitumour antibody could be usefully given when CsA was used. This increases the potential for effective antibody targeted therapy of cancer.

Chemotherapy for gestational trophoblastic tumours hastens menopause by 3 years

Chemotherapy may induce acute ovarian failure, but in women who retain gonadal function throughout chemotherapy, the late effects upon ovarian function are unknown. A retrospective controlled survey was performed to ascertain whether chemotherapy for gestational trophoblastic tumours (GTT) results in premature menopause. Questionnaires were sent to 1,489 women diagnosed between 1971 and 1990 with GTT, including 1089 who had received chemotherapy and 400 who had not received chemotherapy (controls). Responses were obtained from 972 chemotherapy-treated patients and 327 controls. 124 women were not evaluable for menopause date as they had undergone hysterectomy as part of the treatment for GTT or had developed permanent amenorrhoea during chemotherapy. Overall, 172 women reported that they were postmenopausal, including 157 women who had received chemotherapy. The median age at menopause for the evaluable population was 50 years (range 25-56 years). The age at menopause was significantly earlier in the treated arm (median 50, range 25-56 years) than in the controls (median 53, range 40-57 years) (logrank test chi 2 = 12.6, P = 0.0004). Menopause occurred significantly earlier in women treated with combination chemotherapy (median 49, range 25-56 years) compared with single agent methotrexate (median 51, range 25-56 years) (logrank test chi 2 = 8.3, P = 0.004). However, the age at completion of chemotherapy in the treated arm did not influence the age of menopause (proportional Hazards chi 2 = 1.99, P = 0.16). Chemotherapy for GTT induces menopause 3 years earlier than it occurs in women with GTT who do not receive chemotherapy. Although the difference is statistically significant, the magnitude is modest and most women can be reassured that neither fertility nor postmenopausal osteoporosis will be greatly affected.

Antibody distribution and dosimetry in patients receiving radiolabelled antibody therapy for colorectal cancer

The distribution of iodine-131 (131I) labelled antibody to carcinoembryonic antigen (CEA) has been studied in 16 patients with colorectal cancer. Levels of tumour and normal tissue radioactivity were measured by serial gamma-camera imaging and counting of blood and urine. Maximum concentrations were found in tumour 8 h after administration and varied up to 9-fold in different patients. Higher levels were found on average in tumour than in any other tissue. Liver, lung and blood were the other tissues in which antibody was concentrated relative to the rest of the body. Antibody cleared from all these tissues over 1 week. Second antibody directed against the antitumour (first) antibody was given 24 h after first antibody in order to accelerate clearance from the blood. This increased the tumour to blood ratio but had little effect on other tissues. Cumulative radiation dose to tumour and normal tissue was estimated. In patients with the most efficient localisation the tumour to body ratio was 20:1 and tumour to blood ratio 5:1. This may be sufficient for effective therapy of cancer in patients selected for efficient antibody localisation. The data may be used to estimate the effect of different therapeutic strategies. For instance, in the time after second antibody administration the average tumour to blood ratio of radiation dose was 11:1, suggesting that two phase systems in which the therapeutic modality is given after a good tumour to normal tissue ratio is obtained may be effective for the majority of patients.

Long-term outcome in patients with germ cell tumours treated with POMB/ACE chemotherapy: comparison of commonly used classification systems of good and poor prognosis.

We analysed outcome in 206 consecutive male patients treated for metastatic non-seminomatous germ cell tumour (NSGCT) of testicular or extragonadal origin treated with the POMB/ACE (cisplatin, vincristine, methotrexate, bleomycin, actinomycin D, cyclophosphamide, etoposide) regimen after division into prognostic groups by commonly used clinical classification systems and definitions of adverse prognosis. The adverse prognostic groups of all classification systems and definitions examined showed similar, but only moderate, sensitivity (71-81%) and specificity (52-56%) in predicting death. A simple definition of poor prognosis based on raised initial levels of serum tumour markers alpha fetoprotein (aFP) and human chorionic gonadotrophin (hCG) proved at least as useful (sensitivity 80%, specificity 55%) as other more complicated systems in predicting failure to achieve long-term survival. Comparison of survival between ultra-high dose cisplatin-based combination chemotherapy and patients treated with POMB/ACE shows no advantage from this more toxic approach. This suggests that good results in adverse prognosis patients can be achieved using conventional dose regimens administered intensively.

Gestational trophoblastic disease with liver metastases: the Charing Cross experience.

Objective To define management options for women presenting with gestational trophoblastic disease (GTD) which had already metastasised to the liver.

Bone disease in testicular and extragonadal germ cell tumours

Of 297 patients with metastatic testicular and extragonadal germ cell tumours (GCT), bone involvement was detected clinically in 3% (7/251) of those at first presentation and in 9% (4/46) of relapsed cases. This difference was not statistically significant (95% confidence limits -2%; +14%). Concurrent systemic metastases, commonly involving lung (7/11 cases) and para-aortic lymph nodes (6/11), were present in all patients with bone disease. All affected patients had localized bone pain and lumbar spine was the most frequent site involved (9/11). Spinal cord compression occurred in two patients while a third developed progressive vertebral collapse after chemotherapy and required extensive surgical reconstruction. At median follow-up of 4 years, survival among patients presenting with bone disease (6/7) was similar to overall survival in the whole group (84%) and appeared better than in those with liver (18/26, 69%) or central nervous system (6/9) metastases at presentation. Back pain in metastatic germ cell tumours is often due to retroperitoneal lymphadenopathy but lumbar spine osseus metastases must be recognized early if severe potential complications, such as spinal cord compression, are to be avoided. In this series, bone metastases were not seen in the absence of widespread systemic disease suggesting all solitary bony lesions in GCT patients should be biopsied.

Urinary gonadotrophin peptide--isolation and purification, and its immunohistochemical distribution in normal and neoplastic tissues.

A urinary gonadotrophin peptide (UGP) was isolated and purified from semi-purified human chorionic gonadotrophin (hCG), prepared from pregnancy urine. The peptide showed hCG-B subunit activity and no hCG-alpha subunit activity as demonstrated by binding studies with the relevant antibodies. It had a molecular weight significantly less than hCG-B subunit. The peptide was linked to thyroglobulin and this conjugate used to immunise rabbits and mice. A radioimmunoassay (RIA) using 125I-UGP and the rabbit antiserum (AK12) was used to monitor chromatographed urine fractions from patients with ovarian carcinoma, seminoma and hydatidiform mole. UGP was also found in the urine extract of a healthy male, but at a much lower level. In each case the UGP detected had the same molecular weight as the pregnancy preparation and appeared to be the main gonadotrophin constituent in those urine samples. Initial immunohistochemical screening of normal and neoplastic tissues with the rabbit antibody (AK12) showed reactivity with some tumours including carcinomas of the lung, ovary, cervix and breast as well as trophoblastic and germ cell tumours. Reactions with non-neoplastic tissues were confined to some specialised epithelia and macrophage populations. A more comprehensive immunohistochemical study was made using a monoclonal antibody to UGP (2C2), with a monoclonal antibody to conformational hCG (INN 13) and another monoclonal antibody to free B subunit (1E5) as controls. Similar patterns of reactivity were produced by the AK12 and 2C2 antibodies in both neoplastic and non-neoplastic tissues. Additional tissues were investigated with the three monoclonal antibodies. The 2C2 antibody reacted with 93% (77/83) of tumours examined; the INN 13 antibody reacted with only the syncytiotrophoblast cells of choriocarcinoma, hydatidiform mole, placental site trophoblastic tumour, and in one case of seminoma; the 1E5 reactivity was confined to only choriocarcinoma syncytiotrophoblast cells.