Randomised Phase I/II trial assessing the safety and efficacy of radiolabelled anti-carcinoembryonic antigen I¹³¹KAb201 antibodies given intra-arterially or intravenously in patients with unresectable pancreatic adenocarcinoma

Patients with locally advanced or metastatic adenocarcinoma of the head of the pancreas were eligible. The inclusion criteria were age > 18 years, histological or cytological proof, at least one confirmed and measurable tumour site in the head of pancreas, Karnofsky performance status (KPS) ≥ 70 and life expectancy of at least three months. Patients who had undergone prior treatment were enrolled into the trial, provided there was a month's gap between the radiotherapy/chemotherapy (preceding six weeks for nitrosoureas).

Patients were excluded if there was haematological impairment, worsening hepatic impairment or significant renal dysfunction. Other exclusion criteria were known immunological reactions to previously administered antibodies, proteins or iodine, previous external beam radiotherapy to maximal tolerable levels to any critical organ and treatment with any other clinical trial medication within the preceding three months.

Following confirmation of eligibility, patients were randomised to receive the study drug by either the intra-arterial or intravenous route, using computer generated random numbers. The study was not blinded.

KAb201 is a human-sheep chimaeric monoclonal antibody [22] against CEA, produced by a pharmaceutical company (and study sponsor) Xenova Biomedix [23]. The linking of the radioisotope to the MAb helps detect and treat potential sites of disease. Iodine¹³¹ was chosen as its half-life is close to the expected retention of the MAb at the tumour site, its β emission can kill tumours over a range of up to 40 cell diameters and its γ emission can help in imaging the biodistribution.

An earlier phase I study carried out on 10 patients with metastatic colorectal cancer using KAb201 with I¹³¹ found the study drug to be well tolerated, with no drug related adverse events, good localisation at the tumour site and no antigen response in 9 patients [24].

Following this pilot study, the current phase I/II trial was designed for pancreatic cancer after an in vitro study, which yielded a sensitivity of 83% and specificity of 95% for staining with Kab 201. In vitro studies had been conducted by Xenova Biomedix and these showed specificity of the antibody for tumour tissue but not for normal tissue.

Radiolabelling for the pretherapy dose was done at the local Nuclear Medicine department, using the Iodogen method. Each 1 mg of KAb201 was labelled with 10 mCi of I¹³¹, prepared up to 24 hours before administration. Quality control was assessed using Mini TLC, aiming for labelling efficiency of > 60%. The therapeutic dose was prepared by Vrije University, Amsterdam, Netherland. The activity of the I¹³¹ KAb 201 was assessed by the local Nuclear Medicine Department prior to administration. The stability of the radio labelling was > 90%.

Following informed consent, patients underwent clinical evaluation, blood tests for haematological and biochemical assessment, tumour markers (CEA and CA19-9 levels), baseline evaluation of anti-sheep (HASA) and anti-chimaeric (HACA) antibody and radioactivity levels. Tumour size and extent were assessed by multi-detector computed tomography (CT) of the chest, abdomen and pelvis and/or F¹⁸ positron emission tomography (PET).

Pre therapy dose with 2 mCi of I¹³¹-Kab201 was administered via the intended therapeutic route [25]. Intra-arterial drug delivery was either through a temporary catheter inserted at angiography using a 2.5 French prograde microcatheter (Terumo, Belgium) into the gastroduodenal artery or via a permanent catheter, a Port-A-Cath single-lumen implantable vascular access system (SIMS Deltec, Inc., St. Paul, Minnesota, USA), which was inserted into the gastroduodenal artery in patients who required palliative bypass surgery. Intravenous drug delivery was via a standard intravenous line (20 or 22 gauge Venflon). Twenty-four hours later, gamma camera static and single photon emission computerised tomography (SPECT) scans were performed to check for uptake of I¹³¹ in the tumour. This pre therapy check scan was done to assess for uptake of I¹³¹ KAb201 in the primary and/or secondary tumour. If the scan was positive, the patient received the therapy dose 5–7 days later and if there was no uptake, the patient was withdrawn from the trial.

At each dose level, six patients were to be treated, three patients per administration route (intravenous or intra-arterial). Following the therapy dose, patients were isolated for at least five days, in keeping with local radiation safety measures. Repeat gamma camera and SPECT scans were performed a week after treatment to assess localisation of I¹³¹ KAb201 post-therapy.

Patients were followed up 2, 5, 7, 11, 14, 28, 42, 60 and 90 days post- treatment. The blood tests undertaken during follow up are summarised in Table 1. Analyses of blood results and urinalysis were done at a single independent laboratory (Pivotal Laboratory, York) except for radioactivity levels, which were evaluated in the local Nuclear Medicine Department of the Royal Liverpool University Hospital.

CT and PET scans were repeated at 1 and 3 months post treatment. Response was evaluated using the Response Evaluation Criteria in Solid Tumours (RECIST) criteria [2, 26]. If progressive disease was seen on the one month post treatment scans, patients were allowed to withdraw from the trial treatment and be referred for palliative chemotherapy. These patients were followed up for ongoing toxicity, if any, as well as survival.

The primary endpoints were safety, tolerability and level at which dose limiting toxicity occurred. The secondary endpoints were to assess the pharmacokinetics, antigenicity, efficacy and overall survival.

Safety was tolerability were assessed by clinical evaluation, Karnofsky performance status, urinalysis and blood tests (Table 1) and adverse events (AE) were graded by the Common Toxicity Criteria (CTC) version 2 [27].

Dose limiting toxicity (DLT) was defined as grade 4 neutropenia lasting for 7 or more days, febrile neutropenia, platelet count < 25 × 10⁹/L or counts between 25 × 10⁹/L to 50 × 10⁹/L associated with bleeding requiring transfusion, grade 3 or greater nausea despite adequate anti-emetics and any other grade 3 or 4 adverse events. If a DLT occurred in one of three patients, then a further three patients were treated at that dose level and route. Dose escalation was stopped if DLTs were seen in two or more patients.

Pharmacokinetics of both the monoclonal antibody KAb201 (frozen serum samples sent for analyses to Huntingdon Life Sciences, Cambridgeshire, UK) and radioiodine were studied (count done at local Nuclear Medicine department, using a gamma camera calibrated for I¹³¹), while antigenicity was assessed by estimating the Human anti-sheep antibody (HASA) and Human anti-chimaeric antibody (HACA) levels (frozen serum samples sent for analyses to Huntingdon Life Sciences, Cambridgeshire, UK).

The number (and percentage) of patients experiencing each adverse event were tabulated along with the number (and percentage) of occurrences of each event. Time to occurrence of first haematological toxicity was calculated from date of pre therapy dose to date of haematological toxicity (all grades) or the date of the last follow-up if censored. Overall survival was calculated from the date of randomisation to death or the censor date. Time to event data were analysed using the method of Kaplan and Meier [28] and presented graphically with median (95% confidence interval) data. The log rank test was used to assess differences across groups according to the route of administration, KPS, tumour stage (locally advanced versus metastatic disease) and prior treatment. A post-hoc analysis exploring the effect of baseline body surface area (BSA in cm²) on time to haematological toxicity (all grades) was undertaken using a Cox proportional hazards regression model.

Pharmacokinetics of KAb201 and I¹³¹ were assessed using a single compartment model with instantaneous intravenous bolus and first order elimination rate to model both the therapeutic dose and radioactivity plasma concentrations. Area under the curve, Cmax, volume of distribution and rate of elimination were estimated.

Overall response rate (any partial or complete response) was summarised as a number/percentage.

Written informed consent was obtained from all patients prior to entry to the study. The study was conducted according to International Conference on Harmonisation (ICH) on the topic Good Clinical Practice (GCP) guidelines. Ethical approval to carry out this study was given by the Liverpool Local Research Ethics Committee. The trial was conducted to conform to the principles of the Declaration of Helsinki.

Twenty five patients were screened between February 2003 and July 2005. Six patients were screening failures and the other 19 patients were randomised, nine to the intravenous arm and 10 to the intra-arterial arm (Table 2). There was no loss to follow up.

Four patients had received prior chemotherapy, three of whom had progressed on this prior treatment. One patient had regression of disease following chemotherapy (degree of response not evaluated) prior to enrolment onto the trial. Median time from diagnosis to entry into trial was 77 days (range = 25 to 177 days), and in those with previous treatment, this was 247 days (range = 199 to 311 days).

Seventeen patients received the pre therapy dose, nine by the intravenous route and eight by intra-arterial route (three via Port-A-Cath and five via a temporary catheter placed at angiography). Two patients could not undergo pre therapy dose due to problems with I¹³¹ availability. Pre therapy scan uptake was seen in 16 out of 17 patients with lack of uptake occurring in one patient in the intra-arterial group. This patient was hence not given the therapy dose and was excluded from toxicity and efficacy analysis. Post-therapy uptake was seen in all 18 patients.

Dose limiting toxicity (DLT) was observed in none out of three patients at 45 mCi and two out of six patients at 50 mCi in the intra-arterial arm and one out of six patients at 50 mCi and none out of three patients at 75 mCi in the intravenous arm (Figure 1).

The most common adverse event was haematological toxicity (all grades), seen in 17/18 patients (8/9 in the intra-arterial arm and 9/9 in the intravenous arm). One patient received treatment with granulocyte monocyte colony stimulating factor for febrile neutropenia. The median (95% Confidence Interval) time to occurrence of haematological toxicity was 24 (12 to 40) days, with an earlier onset in the intra-arterial arm (21 (12 to 34) days) compared to the intravenous arm (35 (12 to 42) days), although this difference was not statistically significant (log rank p = 0.92). A Cox regression model suggested a significant (p = 0.022) relationship between body surface area in cm² and time to first haematological toxicity (all grades), with a hazard ratio (HR) of 0.97 (95% CI = 0.949, 0.997) for a unit increase in cm² of BSA i.e. a higher BSA had a reduced hazard of haematological toxicity at any time. Some of the other grade 3 or 4 adverse events are detailed in Table 3.